Physician groups are expressing concerns regarding a new policy that will allow indication-based formulary design in the Medicare Part D prescription drug benefit.

Kenishirotie/Thinkstock

The Centers for Medicare & Medicaid Services announced the new policy in an Aug. 29 memo to Part D plan sponsors.

According to a fact sheet issued by CMS on the same day, indication-based formulary design “is a formulary management tool that allows health plans to tailor on-formulary coverage of drugs predicated on specific indications.”

Current Part D policy requires plan sponsors to cover all Food and Drug Administration–approved indications for each drug that is on a plan formulary. Sponsors can begin to implement the new indication-based formulary design policy for plans issued in 2020.

The memo notes that if a Part D plan sponsor chooses to opt into this policy, “it must ensure that there is another therapeutically similar drug on formulary for the nonformulary indication. For example, if a tumor necrosis factor (TNF) blocker is FDA-approved for both Crohn’s disease and plaque psoriasis, but the Part D plan will include it on formulary for plaque psoriasis only, the plan must ensure that there is another TNF blocker on formulary that will be covered for Crohn’s disease.”

Beneficiaries can use the exceptions process to get coverage for a drug that has an indication not on the formulary.

“By allowing Medicare’s prescription drug plans to cover the best drug for each patient condition, plans will have more negotiating power with drug companies, which will result in lower prices for Medicare beneficiaries,” CMS Administrator Seema Verma said in a statement.

However, physician groups see this more as something that could create access issues for patients.

The “proposed changes will exacerbate many of the access issues patients currently face with plan usage of existing utilization management practices, such as step therapy,” the American College of Rheumatology said in a statement. “Unlike step therapy, which often delays effective treatments, this proposal would go even further and allow plans to remove therapies from the formulary altogether, leaving patients completely unable to access treatments that doctors and patients choose together. ... We also have concerns on what this would mean for work being done on compendia inclusion to secure off-label drug coverage if plans don’t have to cover all approved FDA-approved indications.”

The ACR called on CMS to clarify the exception process to make sure patients have access to their needed drugs.

The American Medical Association agreed.

“Under the plan, Medicare patients will face increased challenges as they navigate health plans to make sure that their needed drug is on their selected formulary, which can change based on what health conditions they have,” AMA President Barbara McAneny, MD, said in a statement. “While the CMS notice included a statement that plans had to include information addressing indication-based formularies in materials provided to prospective enrollees, that is not much help to patients who are not sophisticated in how formularies work.”

Dr. McAneny added that it will be even more difficult for physicians who are working with patients to get them on the best medicines covered by the patient’s formulary.

“Physicians already lack ready access to accurate formulary information – preferred/tier status, on/off formulary, PA [prior authorization] and step therapy requirements – at the point of care in their EHRs,” she said. “These transparency problems will expand by an order of magnitude by the complications this change introduces.”

She also noted that it could be difficult for patients who have multiple chronic conditions to find a drug plan that covers the range of their medications, although CMS said the move is expected to increase the number of drugs available on a given plan’s formulary.

gtwachtman@mdedge.com

Physician groups are expressing concerns regarding a new policy that will allow indication-based formulary design in the Medicare Part D prescription drug benefit.

Kenishirotie/Thinkstock

The Centers for Medicare & Medicaid Services announced the new policy in an Aug. 29 memo to Part D plan sponsors.

According to a fact sheet issued by CMS on the same day, indication-based formulary design “is a formulary management tool that allows health plans to tailor on-formulary coverage of drugs predicated on specific indications.”

Current Part D policy requires plan sponsors to cover all Food and Drug Administration–approved indications for each drug that is on a plan formulary. Sponsors can begin to implement the new indication-based formulary design policy for plans issued in 2020.

The memo notes that if a Part D plan sponsor chooses to opt into this policy, “it must ensure that there is another therapeutically similar drug on formulary for the nonformulary indication. For example, if a tumor necrosis factor (TNF) blocker is FDA-approved for both Crohn’s disease and plaque psoriasis, but the Part D plan will include it on formulary for plaque psoriasis only, the plan must ensure that there is another TNF blocker on formulary that will be covered for Crohn’s disease.”

Beneficiaries can use the exceptions process to get coverage for a drug that has an indication not on the formulary.

“By allowing Medicare’s prescription drug plans to cover the best drug for each patient condition, plans will have more negotiating power with drug companies, which will result in lower prices for Medicare beneficiaries,” CMS Administrator Seema Verma said in a statement.

However, physician groups see this more as something that could create access issues for patients.

The “proposed changes will exacerbate many of the access issues patients currently face with plan usage of existing utilization management practices, such as step therapy,” the American College of Rheumatology said in a statement. “Unlike step therapy, which often delays effective treatments, this proposal would go even further and allow plans to remove therapies from the formulary altogether, leaving patients completely unable to access treatments that doctors and patients choose together. ... We also have concerns on what this would mean for work being done on compendia inclusion to secure off-label drug coverage if plans don’t have to cover all approved FDA-approved indications.”

The ACR called on CMS to clarify the exception process to make sure patients have access to their needed drugs.

The American Medical Association agreed.

“Under the plan, Medicare patients will face increased challenges as they navigate health plans to make sure that their needed drug is on their selected formulary, which can change based on what health conditions they have,” AMA President Barbara McAneny, MD, said in a statement. “While the CMS notice included a statement that plans had to include information addressing indication-based formularies in materials provided to prospective enrollees, that is not much help to patients who are not sophisticated in how formularies work.”

Dr. McAneny added that it will be even more difficult for physicians who are working with patients to get them on the best medicines covered by the patient’s formulary.

“Physicians already lack ready access to accurate formulary information – preferred/tier status, on/off formulary, PA [prior authorization] and step therapy requirements – at the point of care in their EHRs,” she said. “These transparency problems will expand by an order of magnitude by the complications this change introduces.”

She also noted that it could be difficult for patients who have multiple chronic conditions to find a drug plan that covers the range of their medications, although CMS said the move is expected to increase the number of drugs available on a given plan’s formulary.

gtwachtman@mdedge.com

Physician groups are expressing concerns regarding a new policy that will allow indication-based formulary design in the Medicare Part D prescription drug benefit.

Kenishirotie/Thinkstock

The Centers for Medicare & Medicaid Services announced the new policy in an Aug. 29 memo to Part D plan sponsors.

According to a fact sheet issued by CMS on the same day, indication-based formulary design “is a formulary management tool that allows health plans to tailor on-formulary coverage of drugs predicated on specific indications.”

Current Part D policy requires plan sponsors to cover all Food and Drug Administration–approved indications for each drug that is on a plan formulary. Sponsors can begin to implement the new indication-based formulary design policy for plans issued in 2020.

The memo notes that if a Part D plan sponsor chooses to opt into this policy, “it must ensure that there is another therapeutically similar drug on formulary for the nonformulary indication. For example, if a tumor necrosis factor (TNF) blocker is FDA-approved for both Crohn’s disease and plaque psoriasis, but the Part D plan will include it on formulary for plaque psoriasis only, the plan must ensure that there is another TNF blocker on formulary that will be covered for Crohn’s disease.”

Beneficiaries can use the exceptions process to get coverage for a drug that has an indication not on the formulary.

“By allowing Medicare’s prescription drug plans to cover the best drug for each patient condition, plans will have more negotiating power with drug companies, which will result in lower prices for Medicare beneficiaries,” CMS Administrator Seema Verma said in a statement.

However, physician groups see this more as something that could create access issues for patients.

The “proposed changes will exacerbate many of the access issues patients currently face with plan usage of existing utilization management practices, such as step therapy,” the American College of Rheumatology said in a statement. “Unlike step therapy, which often delays effective treatments, this proposal would go even further and allow plans to remove therapies from the formulary altogether, leaving patients completely unable to access treatments that doctors and patients choose together. ... We also have concerns on what this would mean for work being done on compendia inclusion to secure off-label drug coverage if plans don’t have to cover all approved FDA-approved indications.”

The ACR called on CMS to clarify the exception process to make sure patients have access to their needed drugs.

The American Medical Association agreed.

“Under the plan, Medicare patients will face increased challenges as they navigate health plans to make sure that their needed drug is on their selected formulary, which can change based on what health conditions they have,” AMA President Barbara McAneny, MD, said in a statement. “While the CMS notice included a statement that plans had to include information addressing indication-based formularies in materials provided to prospective enrollees, that is not much help to patients who are not sophisticated in how formularies work.”

Dr. McAneny added that it will be even more difficult for physicians who are working with patients to get them on the best medicines covered by the patient’s formulary.

“Physicians already lack ready access to accurate formulary information – preferred/tier status, on/off formulary, PA [prior authorization] and step therapy requirements – at the point of care in their EHRs,” she said. “These transparency problems will expand by an order of magnitude by the complications this change introduces.”

She also noted that it could be difficult for patients who have multiple chronic conditions to find a drug plan that covers the range of their medications, although CMS said the move is expected to increase the number of drugs available on a given plan’s formulary.

gtwachtman@mdedge.com

The Food and Drug Administration has granted fast-track designation to CX-01 as a treatment for patients older than 60 years receiving induction therapy for newly diagnosed acute myeloid leukemia (AML).

CX-01 also has orphan drug designation from the FDA.

CX-01 is a polysaccharide derived from heparin thought to enhance chemotherapy by disrupting leukemia cell adhesion in bone marrow. Cantex Pharmaceuticals is conducting a randomized, phase 2b study to determine whether CX-01 can improve the efficacy of frontline chemotherapy in patients with AML.

This study builds upon results of a pilot study, which were published in Blood Advances (Blood Adv. 2018 Feb 27;2[4]:381-9). The pilot study enrolled 12 adults with newly diagnosed AML who received CX-01 as a continuous infusion for 7 days, along with standard induction chemotherapy (cytarabine and idarubicin).

The FDA’s fast-track development program is designed to expedite clinical development and submission of applications for products with the potential to treat serious or life-threatening conditions and address unmet medical needs.

jensmith@mdedge.com

The Food and Drug Administration has granted fast-track designation to CX-01 as a treatment for patients older than 60 years receiving induction therapy for newly diagnosed acute myeloid leukemia (AML).

CX-01 also has orphan drug designation from the FDA.

CX-01 is a polysaccharide derived from heparin thought to enhance chemotherapy by disrupting leukemia cell adhesion in bone marrow. Cantex Pharmaceuticals is conducting a randomized, phase 2b study to determine whether CX-01 can improve the efficacy of frontline chemotherapy in patients with AML.

This study builds upon results of a pilot study, which were published in Blood Advances (Blood Adv. 2018 Feb 27;2[4]:381-9). The pilot study enrolled 12 adults with newly diagnosed AML who received CX-01 as a continuous infusion for 7 days, along with standard induction chemotherapy (cytarabine and idarubicin).

The FDA’s fast-track development program is designed to expedite clinical development and submission of applications for products with the potential to treat serious or life-threatening conditions and address unmet medical needs.

jensmith@mdedge.com

The Food and Drug Administration has granted fast-track designation to CX-01 as a treatment for patients older than 60 years receiving induction therapy for newly diagnosed acute myeloid leukemia (AML).

CX-01 also has orphan drug designation from the FDA.

CX-01 is a polysaccharide derived from heparin thought to enhance chemotherapy by disrupting leukemia cell adhesion in bone marrow. Cantex Pharmaceuticals is conducting a randomized, phase 2b study to determine whether CX-01 can improve the efficacy of frontline chemotherapy in patients with AML.

This study builds upon results of a pilot study, which were published in Blood Advances (Blood Adv. 2018 Feb 27;2[4]:381-9). The pilot study enrolled 12 adults with newly diagnosed AML who received CX-01 as a continuous infusion for 7 days, along with standard induction chemotherapy (cytarabine and idarubicin).

The FDA’s fast-track development program is designed to expedite clinical development and submission of applications for products with the potential to treat serious or life-threatening conditions and address unmet medical needs.

jensmith@mdedge.com

Once again, a study has shown that when it comes to managing patients with serious advanced malignancies – in this case, metastatic renal cell carcinoma – experience matters.

A review of data on 41,836 patients with metastatic renal cell carcinoma (mRCC) treated at 1,222 facilities (TFs) showed that across all cohorts, including patients with known liver and lung metastases who received systemic therapies, treatment-center volume was significantly associated with longer survival, reported Daniel M. Geynisman, MD of Fox Chase Cancer Center in Philadelphia, and his colleagues.

“These findings may help define the optimal treatment environment for the management of patients with mRCC. The improved survival outcomes at higher-volume TFs should also be a call to improve mRCC management at lower-volume facilities. Given the negative externalities of care regionalization, focus should shift toward policies that help equalize mRCC management at lower-volume TFs by expanding treatment options, clinical trial access, and specialized resource availability,” the researchers wrote. Their report is in European Urology.

The investigators noted that several studies have demonstrated that patients with localized RCC treated at high-volume centers had better postoperative outcomes and few complications following surgery for renal cancer, but whether treatment volume makes a difference for patients with metastatic disease was less clear.

To get a better understanding of the association between case volume and outcomes for patients with advanced RCC, the investigators searched the National Cancer Database for information on all U.S. patients with mRCC from 2004 through 2013 for whom survival data were available.

To confirm the association with volume, they created five cohorts with increasingly restrictive inclusion criteria, as follows:

• Cohort A: All patients with survival data (41,836 patients).
• Cohort B: Patients with mRCC who received active treatment of any kind (27,557).
• Cohort C: Patients treated with systemic therapy with or without primary surgery (19,138).
• Cohort D: Patients treated with systemic therapy at the reporting facility (12,000).
• Cohort E: Patients with known sites of metastases (4,933).

The investigators also conducted sensitivity analyses on subcohorts of patients who did not receive nephrectomies in cohorts C, D, and E.

They found in a multivariable analysis that increased volume, measured as cases per year, was associated with reduced overall mortality across all cohorts.

For example, in cohort A, the hazard ratio (HR) for overall mortality for TFs caring for a mean of 5 patients per year was 0.92, compared with 0.84 for centers with 10 cases per year, and 0.74 for TFs caring for a mean of 20 patients per year (P less than .001). Similarly, the respective HRs for patients in cohort E were 0.88, 0.79, and 0.72 (P less than .001).

The overall probability of mortality was also significantly lower in higher-volume centers for those patients in cohorts C, D, and E who did not undergo nephrectomy.

The investigators acknowledged that the study was limited by the retrospective nature of the database information, and by the absence of data on treatment regimens used at specific facilities, which may explain mechanisms of the effects they observed.

The investigators did not specify a study funding source. Dr. Geynisman reported having no conflicts of interest.

SOURCE: Joshi SS et al. Eur Urol. 2018 Sep;74[3]:387-93.

Once again, a study has shown that when it comes to managing patients with serious advanced malignancies – in this case, metastatic renal cell carcinoma – experience matters.

A review of data on 41,836 patients with metastatic renal cell carcinoma (mRCC) treated at 1,222 facilities (TFs) showed that across all cohorts, including patients with known liver and lung metastases who received systemic therapies, treatment-center volume was significantly associated with longer survival, reported Daniel M. Geynisman, MD of Fox Chase Cancer Center in Philadelphia, and his colleagues.

“These findings may help define the optimal treatment environment for the management of patients with mRCC. The improved survival outcomes at higher-volume TFs should also be a call to improve mRCC management at lower-volume facilities. Given the negative externalities of care regionalization, focus should shift toward policies that help equalize mRCC management at lower-volume TFs by expanding treatment options, clinical trial access, and specialized resource availability,” the researchers wrote. Their report is in European Urology.

The investigators noted that several studies have demonstrated that patients with localized RCC treated at high-volume centers had better postoperative outcomes and few complications following surgery for renal cancer, but whether treatment volume makes a difference for patients with metastatic disease was less clear.

To get a better understanding of the association between case volume and outcomes for patients with advanced RCC, the investigators searched the National Cancer Database for information on all U.S. patients with mRCC from 2004 through 2013 for whom survival data were available.

To confirm the association with volume, they created five cohorts with increasingly restrictive inclusion criteria, as follows:

• Cohort A: All patients with survival data (41,836 patients).
• Cohort B: Patients with mRCC who received active treatment of any kind (27,557).
• Cohort C: Patients treated with systemic therapy with or without primary surgery (19,138).
• Cohort D: Patients treated with systemic therapy at the reporting facility (12,000).
• Cohort E: Patients with known sites of metastases (4,933).

The investigators also conducted sensitivity analyses on subcohorts of patients who did not receive nephrectomies in cohorts C, D, and E.

They found in a multivariable analysis that increased volume, measured as cases per year, was associated with reduced overall mortality across all cohorts.

For example, in cohort A, the hazard ratio (HR) for overall mortality for TFs caring for a mean of 5 patients per year was 0.92, compared with 0.84 for centers with 10 cases per year, and 0.74 for TFs caring for a mean of 20 patients per year (P less than .001). Similarly, the respective HRs for patients in cohort E were 0.88, 0.79, and 0.72 (P less than .001).

The overall probability of mortality was also significantly lower in higher-volume centers for those patients in cohorts C, D, and E who did not undergo nephrectomy.

The investigators acknowledged that the study was limited by the retrospective nature of the database information, and by the absence of data on treatment regimens used at specific facilities, which may explain mechanisms of the effects they observed.

The investigators did not specify a study funding source. Dr. Geynisman reported having no conflicts of interest.

SOURCE: Joshi SS et al. Eur Urol. 2018 Sep;74[3]:387-93.

Once again, a study has shown that when it comes to managing patients with serious advanced malignancies – in this case, metastatic renal cell carcinoma – experience matters.

A review of data on 41,836 patients with metastatic renal cell carcinoma (mRCC) treated at 1,222 facilities (TFs) showed that across all cohorts, including patients with known liver and lung metastases who received systemic therapies, treatment-center volume was significantly associated with longer survival, reported Daniel M. Geynisman, MD of Fox Chase Cancer Center in Philadelphia, and his colleagues.

“These findings may help define the optimal treatment environment for the management of patients with mRCC. The improved survival outcomes at higher-volume TFs should also be a call to improve mRCC management at lower-volume facilities. Given the negative externalities of care regionalization, focus should shift toward policies that help equalize mRCC management at lower-volume TFs by expanding treatment options, clinical trial access, and specialized resource availability,” the researchers wrote. Their report is in European Urology.

The investigators noted that several studies have demonstrated that patients with localized RCC treated at high-volume centers had better postoperative outcomes and few complications following surgery for renal cancer, but whether treatment volume makes a difference for patients with metastatic disease was less clear.

To get a better understanding of the association between case volume and outcomes for patients with advanced RCC, the investigators searched the National Cancer Database for information on all U.S. patients with mRCC from 2004 through 2013 for whom survival data were available.

To confirm the association with volume, they created five cohorts with increasingly restrictive inclusion criteria, as follows:

• Cohort A: All patients with survival data (41,836 patients).
• Cohort B: Patients with mRCC who received active treatment of any kind (27,557).
• Cohort C: Patients treated with systemic therapy with or without primary surgery (19,138).
• Cohort D: Patients treated with systemic therapy at the reporting facility (12,000).
• Cohort E: Patients with known sites of metastases (4,933).

The investigators also conducted sensitivity analyses on subcohorts of patients who did not receive nephrectomies in cohorts C, D, and E.

They found in a multivariable analysis that increased volume, measured as cases per year, was associated with reduced overall mortality across all cohorts.

For example, in cohort A, the hazard ratio (HR) for overall mortality for TFs caring for a mean of 5 patients per year was 0.92, compared with 0.84 for centers with 10 cases per year, and 0.74 for TFs caring for a mean of 20 patients per year (P less than .001). Similarly, the respective HRs for patients in cohort E were 0.88, 0.79, and 0.72 (P less than .001).

The overall probability of mortality was also significantly lower in higher-volume centers for those patients in cohorts C, D, and E who did not undergo nephrectomy.

The investigators acknowledged that the study was limited by the retrospective nature of the database information, and by the absence of data on treatment regimens used at specific facilities, which may explain mechanisms of the effects they observed.

The investigators did not specify a study funding source. Dr. Geynisman reported having no conflicts of interest.

SOURCE: Joshi SS et al. Eur Urol. 2018 Sep;74[3]:387-93.

Obesity is a major health problem in the United States. The facts are well known:

• Its prevalence has almost tripled since the early 1960s¹
• More than 35% of US adults are obese (body mass index [BMI] ≥ 30 kg/m²)²
• It increases the risk of comorbid conditions including type 2 diabetes mellitus, heart disease, hypertension, obstructive sleep apnea, certain cancers, asthma, and osteoarthritis^3,4
• It decreases life expectancy⁵
• Medical costs are up to 6 times higher per patient.⁶

Moreover, obesity is often not appropriately managed, owing to a variety of factors. In this article, we describe use of shared medical appointments as a strategy to improve the efficiency and effectiveness of treating patients with obesity.

Big benefits from small changes in weight

As little as 3% to 5% weight loss is associated with significant clinical benefits, such as improved glycemic control, reduced blood pressure, and reduced cholesterol levels.^7,8 However, many patients are unable to reach this modest goal using current approaches to obesity management.

This failure is partially related to the complexity and chronic nature of obesity, which requires continued medical management from a multidisciplinary team. We believe this is an area of care that can be appropriately addressed through shared medical appointments.

CURRENT APPROACHES

Interventions for obesity have increased along with the prevalence of the disease. Hundreds of diets, exercise plans, natural products, and behavioral interventions are marketed, all claiming to be successful. More-intense treatment options include antiobesity medications, intra-abdominal weight loss devices, and bariatric surgery. Despite the availability of treatments, rates of obesity have not declined.

Counseling is important, but underused

Lifestyle modifications that encompass nutrition, physical activity, and behavioral interventions are the mainstay of obesity treatment.

Intensive interventions work better than less-intensive ones. In large clinical trials in overweight patients with diabetes, those who received intensive lifestyle interventions lost 3 to 5 kg more (3% to 8% of body weight) than those who received brief diet and nutrition counseling, as is often performed in a physician’s office.^9–12 The US Preventive Services Task Force recommends that patients whose BMI is 30 kg/m² or higher be offered intensive lifestyle intervention consisting of at least 12 sessions in 1 year.¹³

But fewer than half of primary care practitioners consistently provide specific guidance on diet, exercise, or weight control to patients with obesity, including those with a weight-related comorbidity.¹⁴ The rate has decreased since the 1990s despite the increase in obesity.¹⁵

One reason for the underuse is that many primary care practitioners do not have the training or time to deliver the recommended high-intensity obesity treatment.¹⁴ Plus, evidence does not clearly show a weight loss benefit from low-intensity interventions. Even when patients lose weight, most regain it, and only 20% are able to maintain their weight loss 1 year after treatment ends.¹⁶

Drugs and surgery also underused

Antiobesity medications and bariatric surgery are effective when added to lifestyle interventions, but they are also underused.

Bariatric surgery provides the greatest and most durable weight loss—15% to 30% of body weight—along with improvement in comorbidities such as type 2 diabetes, and its benefits are sustained for at least 10 years.¹⁷ However, fewer than 1% of eligible patients undergo bariatric surgery because of its limited availability, invasive nature, potential complications, limited insurance coverage, and high cost.¹⁷

The story is similar for antiobesity drugs. They are useful adjuncts to lifestyle interventions, providing an additional 3% to 7% weight loss,¹⁸ but fewer than 2% of eligible patients receive them.¹⁹ This may be attributed to their modest effectiveness, weight regain after discontinuation, potential adverse effects, and expense due to lack of insurance coverage.

ARE SHARED MEDICAL APPOINTMEMNTS AN ANSWER?

Although treatments have shown some effectiveness at producing weight loss, none has had a widespread impact on obesity. Lifestyle interventions, drugs, and bariatric surgery continue to be underused. Current treatment models are not providing patients with the intensive interventions needed.

Providers often find themselves offering repetitive advice to patients with obesity regarding nutrition and exercise, while simultaneously trying to manage obesity-related comorbidities, all in a 20-minute appointment. Too often, a patient returns home with prescriptions for hypertension or diabetes but no clear plan for weight management.

What can a shared medical appointment do?

A shared medical appointment is a group medical visit in which several patients with a similar clinical diagnosis, such as obesity, see a multidisciplinary team of healthcare providers. Typically, 5 to 10 patients have consultations with providers during a 60- to 90-minute appointment.²⁰

Part of the session is dedicated to education on the patients’ common medical condition with the goal of improving their self-management, but most of the time is spent addressing individual patient concerns.

Each patient takes a turn consulting with a provider, as in a traditional medical appointment, but in a group setting. This allows others in the group to observe and learn from their peers’ experiences. During this consultation, the patient’s concerns are addressed, medications are managed, necessary tests are ordered, and a treatment plan is made.

Patients can continue to receive follow-up care through shared medical appointments at predetermined times, instead of traditional individual medical appointments.

Shared medical appointments could improve patient access, clinical outcomes, and patient and provider satisfaction and decrease costs.^20,21 Since being introduced in the 1990s, their use has dramatically increased. For example, in the first 2 years of conducting shared medical appointments at Cleveland Clinic (2002–2004), there were just 385 shared medical appointments,²¹ but in 2017 there were approximately 12,300. They are used in a variety of medical and surgical specialties, and have been studied most for treating diabetes.^22–24

Increased face time and access

Individual patient follow-up visits typically last 15 to 20 minutes, limiting the provider to seeing a maximum of 6 patients in 90 minutes. In that same time in the setting of a shared appointment, a multidisciplinary team can see up to 10 patients, and the patients receive up to 90 minutes of time with multiple providers.

Additionally, shared medical appointments can improve patient access to timely appointments. In a busy bariatric surgery practice, implementing shared medical appointments reduced patients’ wait time for an appointment by more than half.²⁵ This is particularly important for patients with obesity, who usually require 12 to 26 appointments per year.

Improved patient outcomes

Use of shared medical appointments has improved clinical outcomes compared with traditional care. Patients with type 2 diabetes who attend shared medical appointments are more likely to reach target hemoglobin A_1c and blood pressure levels.^22­­−24 These benefits may be attributed to increased access to care, improved self-management skills, more frequent visits, peer support of the group, and the synergistic knowledge of multiple providers on the shared medical appointment team.

Although some trials reported patient retention rates of 75% to 90% in shared medical appointments, many trials did not report their rates. It is likely that some patients declined randomization to avoid shared medical appointments, which could have led to potential attrition and selection biases.²³

Increased patient and provider satisfaction

Both patients and providers report high satisfaction with shared medical appointments.^22,26 Although patients may initially hesitate to participate, their opinions significantly improve after attending 1 session.²⁶ From 85% to 90% of patients who attend a shared medical appointment schedule their next follow-up appointment as a shared appointment as well.^21,25

In comparative studies, patients who attended shared medical appointments had satisfaction rates equal to or higher than rates in patients who participated in usual care,²² noting better access to care and more sensitivity to their needs.²⁷ Providers report greater satisfaction from working more directly with a team of providers, clearing up a backlogged schedule, and adding variety to their practice.^21,24

Decreased costs

Data on the cost-effectiveness of shared medical appointments are mixed; however, some studies have shown that they are associated with a decrease in hospital admissions and emergency department visits.²² It seems reasonable to assume that, in an appropriate patient population, shared medical appointments can be cost-effective owing to increased provider productivity, but more research is needed to verify this.

CHALLENGES TO STARTING SHARED APPOINTMENTS FOR OBESITY

Despite their potential to provide comprehensive care to patients, shared medical appointments have limitations. These need to be addressed before implementing a shared medical appointment program.

Adequate resources and staff training

To be successful, a shared medical appointment program needs to have intensive physical and staffing resources. You need a space large enough to accommodate the group and access to the necessary equipment (eg, projector, whiteboard) for educational sessions. Larger or armless chairs may better accommodate patients with obesity. Facilitators need training in how to lead the group sessions, including time management and handling conflicts between patients. Schedulers and clinical intake staff need training in answering patient questions regarding these appointments.

Maintaining patient attendance

The benefits of provider efficiency rest on having an adequate number of patients attend the shared appointments.²¹ Patient cancellations and no-shows decrease both the efficiency and cost-effectiveness of this model, and they detract from the peer support and group learning that occurs in the group dynamic. To help minimize patient dropout, a discussion of patient expectations should take place prior to enrollment in shared medical appointments. This should include information on the concept of shared appointments, frequency and duration of appointments, and realistic weight loss goals.

Logistical challenges

A shared medical appointment requires a longer patient time slot and is usually less flexible than an individual appointment. Not all patients can take the time for a prescheduled 60- to 90-minute appointment. However, reduced waiting-room time and increased face time with a provider offset some of these challenges.

Recruiting patients

A shared medical appointment is a novel experience for some, and concerns about it may make it a challenge to recruit patients. Patients might worry that the presence of the group will compromise the patient-doctor relationship. Other concerns include potential irrelevance of other patients’ medical issues and reluctance to participate because of body image and the stigma of obesity.

One solution is to select patients from your existing practice so that the individual patient-provider relationship is established before introducing the concept of shared appointments. You will need to explain how shared appointments work, discuss their pros and cons, stress your expectations about attendance and confidentiality, and address any concerns of the patient. It is also important to emphasize that nearly all patients find shared medical appointments useful.

Once a group is established, it may be a challenge to keep a constant group membership to promote positive group dynamics. In practice, patients may drop out or be added, and facilitators need to be able to integrate new members into the group. It is important to emphasize to the group that obesity is chronic and that patients at all stages and levels of treatment can contribute to group learning.

Despite the advantages of shared medical appointments, some patients may not find them useful, even after attending several sessions. These patients should be offered individual follow-up visits. Also, shared appointments may not be suitable for patients who cannot speak English very well, are hearing-impaired, have significant cognitive impairment, or have acute medical issues.

Maintaining confidentiality of personal and health information in a shared medical appointment is an important concern for patients but can be appropriately managed. In a survey of patients attending pulmonary hypertension shared medical appointments, 24% had concerns about confidentiality before participating, but after a few sessions, this rate was cut in half.²⁸

Patients have reported initially withholding some information, but over time, they usually become more comfortable with the group and disclose more helpful information.²⁹ Strategies to ensure confidentiality include having patients sign a confidentiality agreement at each appointment, providing specific instruction on what characterizes confidentiality breaches, and allowing patients the opportunity to schedule individual appointments as needed.

Ensuring insurance coverage

A shared medical appointment should be billed as an individual medical appointment for level of care, rather than time spent with the provider. This ensures that insurance coverage and copayments are the same as for individual medical appointments.

Lack of insurance coverage is a major barrier to obesity treatment in general. The US Centers for Medicare and Medicaid Services reimburses intensive behavioral obesity treatment delivered by a primary care practitioner, but limits it to 1 year of treatment and requires patients to meet weight loss goals. Some individual and employer-based healthcare plans do not cover dietitian visits, weight management programs, or antiobesity prescriptions.

EVIDENCE OF EFFECTIVENESS IN OBESITY

Few studies have investigated the use of shared medical appointments in obesity treatment. In the pediatric population, these programs significantly decreased BMI and some other anthropometric measurements,^30–32 but they did not consistently involve a prescribing provider. This means they did not manage medications or comorbidities as would be expected in a shared medical appointment.

In adults, reported effects have been encouraging, although the studies are not particularly robust. In a 2-year observational study of a single physician conducting biweekly weight management shared medical appointments, participants lost 1% of their baseline weight, while those continuing with usual care gained 0.8%, a statistically significant difference.³³ However, participation rates were low, with patients attending an average of only 3 shared medical appointments during the study.

In a meta-analysis of 13 randomized controlled trials of shared medical appointments for patients with type 2 diabetes, only 3 studies reported weight outcomes.²³ These results indicated a trend toward weight loss among patients attending shared appointments, but they were not statistically significant.

Positive results also were reported by the Veterans Administration’s MOVE! (Managing Overweight/obesity for Veterans Everywhere) program.³⁴ Participants in shared medical appointments reported that they felt empowered to make positive lifestyle changes, gained knowledge about obesity, were held accountable by their peers, and appreciated the individualized care they received from the multidisciplinary healthcare teams.

A systematic review involving 336 participants in group-based obesity interventions found group treatment produced more robust weight loss than individual treatment.³⁵ However, shared medical appointments are different from weight loss groups in that they combine an educational session and a medical appointment in a peer-group setting, which requires a provider with prescribing privileges to be present. Thus, shared medical appointments can manage medications as well as weight-related comorbidities such as diabetes, hypertension, polycystic ovarian syndrome, and hyperlipidemia.

One more point is that continued attendance at shared medical appointments, even after successful weight loss, may help to maintain the weight loss, which has otherwise been found to be extremely challenging using traditional medical approaches.

WHO SHOULD BE ON THE TEAM?

Because obesity is multifactorial, it requires a comprehensive treatment approach that can be difficult to deliver given the limited time of an individual appointment. In a shared appointment, providers across multiple specialties can meet with patients at the same time to coordinate approaches to obesity treatment.

A multidisciplinary team for shared medical appointments for obesity needs a physician or a nurse practitioner—or ideally, both— who specializes in obesity to facilitate the session. Other key providers include a registered dietitian, an exercise physiologist, a behavioral health specialist, a sleep specialist, and a social worker to participate as needed in the educational component of the appointment or act as outside consultants.

WHAT ARE REALISTIC TARGETS?

• Nutrition
• Physical activity
• Appetite control
• Sleep
• Stress and mood disorders.

Nutrition

A calorie deficit of 500 to 750 calories per day is recommended for weight loss.^7,8 Although there is no consensus on the best nutritional content of a diet, adherence to a diet is a significant predictor of weight loss.³⁶ One reason diets fail to bring about weight loss is that patients tend to underestimate their caloric intake by almost 50%.³⁷ Thus, they may benefit from a structured and supervised diet plan.

A dietitian can help patients develop an individualized diet plan that will promote adherence, which includes specific information on food choices, portion sizes, and timing of meals.

Physical activity

At least 150 minutes of physical activity per week is recommended for weight loss, and 200 to 300 minutes per week is recommended for long-term weight maintenance.^7,8

An exercise physiologist can help patients design a personalized exercise plan to help achieve these goals. This plan should take into account the patient’s cardiac status, activity level, degree of mobility, and lifestyle.

Most patients are not able to achieve the recommended physical activity goals initially, and activity levels need to be gradually increased over a period of weeks to months. Patients who were previously inactive or have evidence of cardiovascular, renal, or metabolic disease may require a cardiopulmonary assessment, including an electrocardiogram and cardiac stress test, before starting an exercise program.

It is very difficult for patients to lose weight without appetite control. Weight loss that results from diet and exercise is often accompanied by a change in weight-regulating hormones (eg, leptin, ghrelin, peptide YY, and cholecystokinin) that promote weight regain.³⁸ Thus, multiple compensatory mechanisms promote weight regain through increases in appetite and decreases in energy expenditure, resisting weight loss efforts.

Antiobesity drugs can help mitigate these adaptive weight-promoting responses through several mechanisms. They are indicated for use with lifestyle interventions for patients with a BMI of at least 30 mg/kg² or a BMI of at least 27 kg/m² with an obesity-related comorbidity.

These drugs promote an additional 3% to 7% weight loss when added to lifestyle interventions.¹⁸ But their effects are limited without appropriate lifestyle interventions.

Sleep

Adequate sleep is an often-overlooked component of obesity treatment. Inadequate sleep is associated with weight gain and an appetite-inducing hormone profile.³⁹ Just 2 days of sleep deprivation in healthy normal-weight adult men was associated with a 70% increase in the ghrelin-to-leptin ratio, which showed a linear relationship with self-reported increased hunger.³⁹ Sleep disorders, especially obstructive sleep apnea, are common in patients with obesity but are often underdiagnosed and undertreated.⁴⁰

Healthy sleep habits and sleep quality should be addressed in shared medical appointments for obesity, as patients may be unaware of the impact that sleep may be having on their obesity treatment. The STOP-BANG questionnaire (­snoring, tiredness, observed apnea, high blood pressure, BMI, age, neck circumference, and male sex) is a simple and reliable tool to screen for obstructive sleep apnea.⁴¹ Patients with symptoms of a sleep disorder should be referred to a sleep specialist for diagnosis and management.

Stress management and mood disorders

Stress and psychiatric disorders are underappreciated contributors to obesity. All patients receiving obesity treatment need to be screened for mood disorders and suicidal ideation.⁸

Chronic stress promotes weight gain through activation of the hypothalamic-pituitary-adrenocortical axis, whereby increased cortisol levels enhance appetite and accumulation of visceral fat.⁴² In addition, obesity is associated with a 25% increased risk of mood disorders, although the mechanism and direction of this association are unclear.⁴³ Weight gain as a side effect of antidepressant or other psychiatric medications is another important consideration.

Management of stress and psychiatric disorders through goal-setting, self-monitoring, and patient education is vital to help patients fully participate in lifestyle changes and maximize weight loss. Patients participating in shared medical appointments usually benefit from consultations with psychiatrists or psychologists to manage psychiatric comorbidities and assist with adherence to behavior modification. 


IN FAVOR OF SHARED MEDICAL APPOINTMENTS FOR OBESITY

Shared medical appointments can be an effective method of addressing the challenges of treating patients with obesity, using a multidisciplinary approach that combines nutrition, physical activity, appetite suppression, sleep improvement, and stress management. In addition, shared appointments allow practitioners to treat the primary problem of excess weight, rather than just its comorbidities, recognizing that obesity is a chronic disease that requires long-term, individualized treatment. Satisfaction rates are high for both patients and providers. Overall, education is essential to implementing and maintaining a successful shared medical appointment program.

Obesity is a major health problem in the United States. The facts are well known:

• Its prevalence has almost tripled since the early 1960s¹
• More than 35% of US adults are obese (body mass index [BMI] ≥ 30 kg/m²)²
• It increases the risk of comorbid conditions including type 2 diabetes mellitus, heart disease, hypertension, obstructive sleep apnea, certain cancers, asthma, and osteoarthritis^3,4
• It decreases life expectancy⁵
• Medical costs are up to 6 times higher per patient.⁶

Moreover, obesity is often not appropriately managed, owing to a variety of factors. In this article, we describe use of shared medical appointments as a strategy to improve the efficiency and effectiveness of treating patients with obesity.

Big benefits from small changes in weight

As little as 3% to 5% weight loss is associated with significant clinical benefits, such as improved glycemic control, reduced blood pressure, and reduced cholesterol levels.^7,8 However, many patients are unable to reach this modest goal using current approaches to obesity management.

This failure is partially related to the complexity and chronic nature of obesity, which requires continued medical management from a multidisciplinary team. We believe this is an area of care that can be appropriately addressed through shared medical appointments.

CURRENT APPROACHES

Interventions for obesity have increased along with the prevalence of the disease. Hundreds of diets, exercise plans, natural products, and behavioral interventions are marketed, all claiming to be successful. More-intense treatment options include antiobesity medications, intra-abdominal weight loss devices, and bariatric surgery. Despite the availability of treatments, rates of obesity have not declined.

Counseling is important, but underused

Lifestyle modifications that encompass nutrition, physical activity, and behavioral interventions are the mainstay of obesity treatment.

Intensive interventions work better than less-intensive ones. In large clinical trials in overweight patients with diabetes, those who received intensive lifestyle interventions lost 3 to 5 kg more (3% to 8% of body weight) than those who received brief diet and nutrition counseling, as is often performed in a physician’s office.^9–12 The US Preventive Services Task Force recommends that patients whose BMI is 30 kg/m² or higher be offered intensive lifestyle intervention consisting of at least 12 sessions in 1 year.¹³

But fewer than half of primary care practitioners consistently provide specific guidance on diet, exercise, or weight control to patients with obesity, including those with a weight-related comorbidity.¹⁴ The rate has decreased since the 1990s despite the increase in obesity.¹⁵

One reason for the underuse is that many primary care practitioners do not have the training or time to deliver the recommended high-intensity obesity treatment.¹⁴ Plus, evidence does not clearly show a weight loss benefit from low-intensity interventions. Even when patients lose weight, most regain it, and only 20% are able to maintain their weight loss 1 year after treatment ends.¹⁶

Drugs and surgery also underused

Antiobesity medications and bariatric surgery are effective when added to lifestyle interventions, but they are also underused.

Bariatric surgery provides the greatest and most durable weight loss—15% to 30% of body weight—along with improvement in comorbidities such as type 2 diabetes, and its benefits are sustained for at least 10 years.¹⁷ However, fewer than 1% of eligible patients undergo bariatric surgery because of its limited availability, invasive nature, potential complications, limited insurance coverage, and high cost.¹⁷

The story is similar for antiobesity drugs. They are useful adjuncts to lifestyle interventions, providing an additional 3% to 7% weight loss,¹⁸ but fewer than 2% of eligible patients receive them.¹⁹ This may be attributed to their modest effectiveness, weight regain after discontinuation, potential adverse effects, and expense due to lack of insurance coverage.

ARE SHARED MEDICAL APPOINTMEMNTS AN ANSWER?

Although treatments have shown some effectiveness at producing weight loss, none has had a widespread impact on obesity. Lifestyle interventions, drugs, and bariatric surgery continue to be underused. Current treatment models are not providing patients with the intensive interventions needed.

Providers often find themselves offering repetitive advice to patients with obesity regarding nutrition and exercise, while simultaneously trying to manage obesity-related comorbidities, all in a 20-minute appointment. Too often, a patient returns home with prescriptions for hypertension or diabetes but no clear plan for weight management.

What can a shared medical appointment do?

A shared medical appointment is a group medical visit in which several patients with a similar clinical diagnosis, such as obesity, see a multidisciplinary team of healthcare providers. Typically, 5 to 10 patients have consultations with providers during a 60- to 90-minute appointment.²⁰

Part of the session is dedicated to education on the patients’ common medical condition with the goal of improving their self-management, but most of the time is spent addressing individual patient concerns.

Each patient takes a turn consulting with a provider, as in a traditional medical appointment, but in a group setting. This allows others in the group to observe and learn from their peers’ experiences. During this consultation, the patient’s concerns are addressed, medications are managed, necessary tests are ordered, and a treatment plan is made.

Patients can continue to receive follow-up care through shared medical appointments at predetermined times, instead of traditional individual medical appointments.

Shared medical appointments could improve patient access, clinical outcomes, and patient and provider satisfaction and decrease costs.^20,21 Since being introduced in the 1990s, their use has dramatically increased. For example, in the first 2 years of conducting shared medical appointments at Cleveland Clinic (2002–2004), there were just 385 shared medical appointments,²¹ but in 2017 there were approximately 12,300. They are used in a variety of medical and surgical specialties, and have been studied most for treating diabetes.^22–24

Increased face time and access

Individual patient follow-up visits typically last 15 to 20 minutes, limiting the provider to seeing a maximum of 6 patients in 90 minutes. In that same time in the setting of a shared appointment, a multidisciplinary team can see up to 10 patients, and the patients receive up to 90 minutes of time with multiple providers.

Additionally, shared medical appointments can improve patient access to timely appointments. In a busy bariatric surgery practice, implementing shared medical appointments reduced patients’ wait time for an appointment by more than half.²⁵ This is particularly important for patients with obesity, who usually require 12 to 26 appointments per year.

Improved patient outcomes

Use of shared medical appointments has improved clinical outcomes compared with traditional care. Patients with type 2 diabetes who attend shared medical appointments are more likely to reach target hemoglobin A_1c and blood pressure levels.^22­­−24 These benefits may be attributed to increased access to care, improved self-management skills, more frequent visits, peer support of the group, and the synergistic knowledge of multiple providers on the shared medical appointment team.

Although some trials reported patient retention rates of 75% to 90% in shared medical appointments, many trials did not report their rates. It is likely that some patients declined randomization to avoid shared medical appointments, which could have led to potential attrition and selection biases.²³

Increased patient and provider satisfaction

Both patients and providers report high satisfaction with shared medical appointments.^22,26 Although patients may initially hesitate to participate, their opinions significantly improve after attending 1 session.²⁶ From 85% to 90% of patients who attend a shared medical appointment schedule their next follow-up appointment as a shared appointment as well.^21,25

In comparative studies, patients who attended shared medical appointments had satisfaction rates equal to or higher than rates in patients who participated in usual care,²² noting better access to care and more sensitivity to their needs.²⁷ Providers report greater satisfaction from working more directly with a team of providers, clearing up a backlogged schedule, and adding variety to their practice.^21,24

Decreased costs

Data on the cost-effectiveness of shared medical appointments are mixed; however, some studies have shown that they are associated with a decrease in hospital admissions and emergency department visits.²² It seems reasonable to assume that, in an appropriate patient population, shared medical appointments can be cost-effective owing to increased provider productivity, but more research is needed to verify this.

CHALLENGES TO STARTING SHARED APPOINTMENTS FOR OBESITY

Despite their potential to provide comprehensive care to patients, shared medical appointments have limitations. These need to be addressed before implementing a shared medical appointment program.

Adequate resources and staff training

To be successful, a shared medical appointment program needs to have intensive physical and staffing resources. You need a space large enough to accommodate the group and access to the necessary equipment (eg, projector, whiteboard) for educational sessions. Larger or armless chairs may better accommodate patients with obesity. Facilitators need training in how to lead the group sessions, including time management and handling conflicts between patients. Schedulers and clinical intake staff need training in answering patient questions regarding these appointments.

Maintaining patient attendance

The benefits of provider efficiency rest on having an adequate number of patients attend the shared appointments.²¹ Patient cancellations and no-shows decrease both the efficiency and cost-effectiveness of this model, and they detract from the peer support and group learning that occurs in the group dynamic. To help minimize patient dropout, a discussion of patient expectations should take place prior to enrollment in shared medical appointments. This should include information on the concept of shared appointments, frequency and duration of appointments, and realistic weight loss goals.

Logistical challenges

A shared medical appointment requires a longer patient time slot and is usually less flexible than an individual appointment. Not all patients can take the time for a prescheduled 60- to 90-minute appointment. However, reduced waiting-room time and increased face time with a provider offset some of these challenges.

Recruiting patients

A shared medical appointment is a novel experience for some, and concerns about it may make it a challenge to recruit patients. Patients might worry that the presence of the group will compromise the patient-doctor relationship. Other concerns include potential irrelevance of other patients’ medical issues and reluctance to participate because of body image and the stigma of obesity.

One solution is to select patients from your existing practice so that the individual patient-provider relationship is established before introducing the concept of shared appointments. You will need to explain how shared appointments work, discuss their pros and cons, stress your expectations about attendance and confidentiality, and address any concerns of the patient. It is also important to emphasize that nearly all patients find shared medical appointments useful.

Once a group is established, it may be a challenge to keep a constant group membership to promote positive group dynamics. In practice, patients may drop out or be added, and facilitators need to be able to integrate new members into the group. It is important to emphasize to the group that obesity is chronic and that patients at all stages and levels of treatment can contribute to group learning.

Despite the advantages of shared medical appointments, some patients may not find them useful, even after attending several sessions. These patients should be offered individual follow-up visits. Also, shared appointments may not be suitable for patients who cannot speak English very well, are hearing-impaired, have significant cognitive impairment, or have acute medical issues.

Maintaining confidentiality of personal and health information in a shared medical appointment is an important concern for patients but can be appropriately managed. In a survey of patients attending pulmonary hypertension shared medical appointments, 24% had concerns about confidentiality before participating, but after a few sessions, this rate was cut in half.²⁸

Patients have reported initially withholding some information, but over time, they usually become more comfortable with the group and disclose more helpful information.²⁹ Strategies to ensure confidentiality include having patients sign a confidentiality agreement at each appointment, providing specific instruction on what characterizes confidentiality breaches, and allowing patients the opportunity to schedule individual appointments as needed.

Ensuring insurance coverage

A shared medical appointment should be billed as an individual medical appointment for level of care, rather than time spent with the provider. This ensures that insurance coverage and copayments are the same as for individual medical appointments.

Lack of insurance coverage is a major barrier to obesity treatment in general. The US Centers for Medicare and Medicaid Services reimburses intensive behavioral obesity treatment delivered by a primary care practitioner, but limits it to 1 year of treatment and requires patients to meet weight loss goals. Some individual and employer-based healthcare plans do not cover dietitian visits, weight management programs, or antiobesity prescriptions.

EVIDENCE OF EFFECTIVENESS IN OBESITY

Few studies have investigated the use of shared medical appointments in obesity treatment. In the pediatric population, these programs significantly decreased BMI and some other anthropometric measurements,^30–32 but they did not consistently involve a prescribing provider. This means they did not manage medications or comorbidities as would be expected in a shared medical appointment.

In adults, reported effects have been encouraging, although the studies are not particularly robust. In a 2-year observational study of a single physician conducting biweekly weight management shared medical appointments, participants lost 1% of their baseline weight, while those continuing with usual care gained 0.8%, a statistically significant difference.³³ However, participation rates were low, with patients attending an average of only 3 shared medical appointments during the study.

In a meta-analysis of 13 randomized controlled trials of shared medical appointments for patients with type 2 diabetes, only 3 studies reported weight outcomes.²³ These results indicated a trend toward weight loss among patients attending shared appointments, but they were not statistically significant.

Positive results also were reported by the Veterans Administration’s MOVE! (Managing Overweight/obesity for Veterans Everywhere) program.³⁴ Participants in shared medical appointments reported that they felt empowered to make positive lifestyle changes, gained knowledge about obesity, were held accountable by their peers, and appreciated the individualized care they received from the multidisciplinary healthcare teams.

A systematic review involving 336 participants in group-based obesity interventions found group treatment produced more robust weight loss than individual treatment.³⁵ However, shared medical appointments are different from weight loss groups in that they combine an educational session and a medical appointment in a peer-group setting, which requires a provider with prescribing privileges to be present. Thus, shared medical appointments can manage medications as well as weight-related comorbidities such as diabetes, hypertension, polycystic ovarian syndrome, and hyperlipidemia.

One more point is that continued attendance at shared medical appointments, even after successful weight loss, may help to maintain the weight loss, which has otherwise been found to be extremely challenging using traditional medical approaches.

WHO SHOULD BE ON THE TEAM?

Because obesity is multifactorial, it requires a comprehensive treatment approach that can be difficult to deliver given the limited time of an individual appointment. In a shared appointment, providers across multiple specialties can meet with patients at the same time to coordinate approaches to obesity treatment.

A multidisciplinary team for shared medical appointments for obesity needs a physician or a nurse practitioner—or ideally, both— who specializes in obesity to facilitate the session. Other key providers include a registered dietitian, an exercise physiologist, a behavioral health specialist, a sleep specialist, and a social worker to participate as needed in the educational component of the appointment or act as outside consultants.

WHAT ARE REALISTIC TARGETS?

• Nutrition
• Physical activity
• Appetite control
• Sleep
• Stress and mood disorders.

Nutrition

A calorie deficit of 500 to 750 calories per day is recommended for weight loss.^7,8 Although there is no consensus on the best nutritional content of a diet, adherence to a diet is a significant predictor of weight loss.³⁶ One reason diets fail to bring about weight loss is that patients tend to underestimate their caloric intake by almost 50%.³⁷ Thus, they may benefit from a structured and supervised diet plan.

A dietitian can help patients develop an individualized diet plan that will promote adherence, which includes specific information on food choices, portion sizes, and timing of meals.

Physical activity

At least 150 minutes of physical activity per week is recommended for weight loss, and 200 to 300 minutes per week is recommended for long-term weight maintenance.^7,8

An exercise physiologist can help patients design a personalized exercise plan to help achieve these goals. This plan should take into account the patient’s cardiac status, activity level, degree of mobility, and lifestyle.

Most patients are not able to achieve the recommended physical activity goals initially, and activity levels need to be gradually increased over a period of weeks to months. Patients who were previously inactive or have evidence of cardiovascular, renal, or metabolic disease may require a cardiopulmonary assessment, including an electrocardiogram and cardiac stress test, before starting an exercise program.

It is very difficult for patients to lose weight without appetite control. Weight loss that results from diet and exercise is often accompanied by a change in weight-regulating hormones (eg, leptin, ghrelin, peptide YY, and cholecystokinin) that promote weight regain.³⁸ Thus, multiple compensatory mechanisms promote weight regain through increases in appetite and decreases in energy expenditure, resisting weight loss efforts.

Antiobesity drugs can help mitigate these adaptive weight-promoting responses through several mechanisms. They are indicated for use with lifestyle interventions for patients with a BMI of at least 30 mg/kg² or a BMI of at least 27 kg/m² with an obesity-related comorbidity.

These drugs promote an additional 3% to 7% weight loss when added to lifestyle interventions.¹⁸ But their effects are limited without appropriate lifestyle interventions.

Sleep

Adequate sleep is an often-overlooked component of obesity treatment. Inadequate sleep is associated with weight gain and an appetite-inducing hormone profile.³⁹ Just 2 days of sleep deprivation in healthy normal-weight adult men was associated with a 70% increase in the ghrelin-to-leptin ratio, which showed a linear relationship with self-reported increased hunger.³⁹ Sleep disorders, especially obstructive sleep apnea, are common in patients with obesity but are often underdiagnosed and undertreated.⁴⁰

Healthy sleep habits and sleep quality should be addressed in shared medical appointments for obesity, as patients may be unaware of the impact that sleep may be having on their obesity treatment. The STOP-BANG questionnaire (­snoring, tiredness, observed apnea, high blood pressure, BMI, age, neck circumference, and male sex) is a simple and reliable tool to screen for obstructive sleep apnea.⁴¹ Patients with symptoms of a sleep disorder should be referred to a sleep specialist for diagnosis and management.

Stress management and mood disorders

Stress and psychiatric disorders are underappreciated contributors to obesity. All patients receiving obesity treatment need to be screened for mood disorders and suicidal ideation.⁸

Chronic stress promotes weight gain through activation of the hypothalamic-pituitary-adrenocortical axis, whereby increased cortisol levels enhance appetite and accumulation of visceral fat.⁴² In addition, obesity is associated with a 25% increased risk of mood disorders, although the mechanism and direction of this association are unclear.⁴³ Weight gain as a side effect of antidepressant or other psychiatric medications is another important consideration.

Management of stress and psychiatric disorders through goal-setting, self-monitoring, and patient education is vital to help patients fully participate in lifestyle changes and maximize weight loss. Patients participating in shared medical appointments usually benefit from consultations with psychiatrists or psychologists to manage psychiatric comorbidities and assist with adherence to behavior modification. 


IN FAVOR OF SHARED MEDICAL APPOINTMENTS FOR OBESITY

Shared medical appointments can be an effective method of addressing the challenges of treating patients with obesity, using a multidisciplinary approach that combines nutrition, physical activity, appetite suppression, sleep improvement, and stress management. In addition, shared appointments allow practitioners to treat the primary problem of excess weight, rather than just its comorbidities, recognizing that obesity is a chronic disease that requires long-term, individualized treatment. Satisfaction rates are high for both patients and providers. Overall, education is essential to implementing and maintaining a successful shared medical appointment program.

Obesity is a major health problem in the United States. The facts are well known:

• Its prevalence has almost tripled since the early 1960s¹
• More than 35% of US adults are obese (body mass index [BMI] ≥ 30 kg/m²)²
• It increases the risk of comorbid conditions including type 2 diabetes mellitus, heart disease, hypertension, obstructive sleep apnea, certain cancers, asthma, and osteoarthritis^3,4
• It decreases life expectancy⁵
• Medical costs are up to 6 times higher per patient.⁶

Moreover, obesity is often not appropriately managed, owing to a variety of factors. In this article, we describe use of shared medical appointments as a strategy to improve the efficiency and effectiveness of treating patients with obesity.

Big benefits from small changes in weight

As little as 3% to 5% weight loss is associated with significant clinical benefits, such as improved glycemic control, reduced blood pressure, and reduced cholesterol levels.^7,8 However, many patients are unable to reach this modest goal using current approaches to obesity management.

This failure is partially related to the complexity and chronic nature of obesity, which requires continued medical management from a multidisciplinary team. We believe this is an area of care that can be appropriately addressed through shared medical appointments.

CURRENT APPROACHES

Interventions for obesity have increased along with the prevalence of the disease. Hundreds of diets, exercise plans, natural products, and behavioral interventions are marketed, all claiming to be successful. More-intense treatment options include antiobesity medications, intra-abdominal weight loss devices, and bariatric surgery. Despite the availability of treatments, rates of obesity have not declined.

Counseling is important, but underused

Lifestyle modifications that encompass nutrition, physical activity, and behavioral interventions are the mainstay of obesity treatment.

Intensive interventions work better than less-intensive ones. In large clinical trials in overweight patients with diabetes, those who received intensive lifestyle interventions lost 3 to 5 kg more (3% to 8% of body weight) than those who received brief diet and nutrition counseling, as is often performed in a physician’s office.^9–12 The US Preventive Services Task Force recommends that patients whose BMI is 30 kg/m² or higher be offered intensive lifestyle intervention consisting of at least 12 sessions in 1 year.¹³

But fewer than half of primary care practitioners consistently provide specific guidance on diet, exercise, or weight control to patients with obesity, including those with a weight-related comorbidity.¹⁴ The rate has decreased since the 1990s despite the increase in obesity.¹⁵

One reason for the underuse is that many primary care practitioners do not have the training or time to deliver the recommended high-intensity obesity treatment.¹⁴ Plus, evidence does not clearly show a weight loss benefit from low-intensity interventions. Even when patients lose weight, most regain it, and only 20% are able to maintain their weight loss 1 year after treatment ends.¹⁶

Drugs and surgery also underused

Antiobesity medications and bariatric surgery are effective when added to lifestyle interventions, but they are also underused.

Bariatric surgery provides the greatest and most durable weight loss—15% to 30% of body weight—along with improvement in comorbidities such as type 2 diabetes, and its benefits are sustained for at least 10 years.¹⁷ However, fewer than 1% of eligible patients undergo bariatric surgery because of its limited availability, invasive nature, potential complications, limited insurance coverage, and high cost.¹⁷

The story is similar for antiobesity drugs. They are useful adjuncts to lifestyle interventions, providing an additional 3% to 7% weight loss,¹⁸ but fewer than 2% of eligible patients receive them.¹⁹ This may be attributed to their modest effectiveness, weight regain after discontinuation, potential adverse effects, and expense due to lack of insurance coverage.

ARE SHARED MEDICAL APPOINTMEMNTS AN ANSWER?

Although treatments have shown some effectiveness at producing weight loss, none has had a widespread impact on obesity. Lifestyle interventions, drugs, and bariatric surgery continue to be underused. Current treatment models are not providing patients with the intensive interventions needed.

Providers often find themselves offering repetitive advice to patients with obesity regarding nutrition and exercise, while simultaneously trying to manage obesity-related comorbidities, all in a 20-minute appointment. Too often, a patient returns home with prescriptions for hypertension or diabetes but no clear plan for weight management.

What can a shared medical appointment do?

A shared medical appointment is a group medical visit in which several patients with a similar clinical diagnosis, such as obesity, see a multidisciplinary team of healthcare providers. Typically, 5 to 10 patients have consultations with providers during a 60- to 90-minute appointment.²⁰

Part of the session is dedicated to education on the patients’ common medical condition with the goal of improving their self-management, but most of the time is spent addressing individual patient concerns.

Each patient takes a turn consulting with a provider, as in a traditional medical appointment, but in a group setting. This allows others in the group to observe and learn from their peers’ experiences. During this consultation, the patient’s concerns are addressed, medications are managed, necessary tests are ordered, and a treatment plan is made.

Patients can continue to receive follow-up care through shared medical appointments at predetermined times, instead of traditional individual medical appointments.

Shared medical appointments could improve patient access, clinical outcomes, and patient and provider satisfaction and decrease costs.^20,21 Since being introduced in the 1990s, their use has dramatically increased. For example, in the first 2 years of conducting shared medical appointments at Cleveland Clinic (2002–2004), there were just 385 shared medical appointments,²¹ but in 2017 there were approximately 12,300. They are used in a variety of medical and surgical specialties, and have been studied most for treating diabetes.^22–24

Increased face time and access

Individual patient follow-up visits typically last 15 to 20 minutes, limiting the provider to seeing a maximum of 6 patients in 90 minutes. In that same time in the setting of a shared appointment, a multidisciplinary team can see up to 10 patients, and the patients receive up to 90 minutes of time with multiple providers.

Additionally, shared medical appointments can improve patient access to timely appointments. In a busy bariatric surgery practice, implementing shared medical appointments reduced patients’ wait time for an appointment by more than half.²⁵ This is particularly important for patients with obesity, who usually require 12 to 26 appointments per year.

Improved patient outcomes

Use of shared medical appointments has improved clinical outcomes compared with traditional care. Patients with type 2 diabetes who attend shared medical appointments are more likely to reach target hemoglobin A_1c and blood pressure levels.^22­­−24 These benefits may be attributed to increased access to care, improved self-management skills, more frequent visits, peer support of the group, and the synergistic knowledge of multiple providers on the shared medical appointment team.

Although some trials reported patient retention rates of 75% to 90% in shared medical appointments, many trials did not report their rates. It is likely that some patients declined randomization to avoid shared medical appointments, which could have led to potential attrition and selection biases.²³

Increased patient and provider satisfaction

Both patients and providers report high satisfaction with shared medical appointments.^22,26 Although patients may initially hesitate to participate, their opinions significantly improve after attending 1 session.²⁶ From 85% to 90% of patients who attend a shared medical appointment schedule their next follow-up appointment as a shared appointment as well.^21,25

In comparative studies, patients who attended shared medical appointments had satisfaction rates equal to or higher than rates in patients who participated in usual care,²² noting better access to care and more sensitivity to their needs.²⁷ Providers report greater satisfaction from working more directly with a team of providers, clearing up a backlogged schedule, and adding variety to their practice.^21,24

Decreased costs

Data on the cost-effectiveness of shared medical appointments are mixed; however, some studies have shown that they are associated with a decrease in hospital admissions and emergency department visits.²² It seems reasonable to assume that, in an appropriate patient population, shared medical appointments can be cost-effective owing to increased provider productivity, but more research is needed to verify this.

CHALLENGES TO STARTING SHARED APPOINTMENTS FOR OBESITY

Despite their potential to provide comprehensive care to patients, shared medical appointments have limitations. These need to be addressed before implementing a shared medical appointment program.

Adequate resources and staff training

To be successful, a shared medical appointment program needs to have intensive physical and staffing resources. You need a space large enough to accommodate the group and access to the necessary equipment (eg, projector, whiteboard) for educational sessions. Larger or armless chairs may better accommodate patients with obesity. Facilitators need training in how to lead the group sessions, including time management and handling conflicts between patients. Schedulers and clinical intake staff need training in answering patient questions regarding these appointments.

Maintaining patient attendance

The benefits of provider efficiency rest on having an adequate number of patients attend the shared appointments.²¹ Patient cancellations and no-shows decrease both the efficiency and cost-effectiveness of this model, and they detract from the peer support and group learning that occurs in the group dynamic. To help minimize patient dropout, a discussion of patient expectations should take place prior to enrollment in shared medical appointments. This should include information on the concept of shared appointments, frequency and duration of appointments, and realistic weight loss goals.

Logistical challenges

A shared medical appointment requires a longer patient time slot and is usually less flexible than an individual appointment. Not all patients can take the time for a prescheduled 60- to 90-minute appointment. However, reduced waiting-room time and increased face time with a provider offset some of these challenges.

Recruiting patients

A shared medical appointment is a novel experience for some, and concerns about it may make it a challenge to recruit patients. Patients might worry that the presence of the group will compromise the patient-doctor relationship. Other concerns include potential irrelevance of other patients’ medical issues and reluctance to participate because of body image and the stigma of obesity.

One solution is to select patients from your existing practice so that the individual patient-provider relationship is established before introducing the concept of shared appointments. You will need to explain how shared appointments work, discuss their pros and cons, stress your expectations about attendance and confidentiality, and address any concerns of the patient. It is also important to emphasize that nearly all patients find shared medical appointments useful.

Once a group is established, it may be a challenge to keep a constant group membership to promote positive group dynamics. In practice, patients may drop out or be added, and facilitators need to be able to integrate new members into the group. It is important to emphasize to the group that obesity is chronic and that patients at all stages and levels of treatment can contribute to group learning.

Despite the advantages of shared medical appointments, some patients may not find them useful, even after attending several sessions. These patients should be offered individual follow-up visits. Also, shared appointments may not be suitable for patients who cannot speak English very well, are hearing-impaired, have significant cognitive impairment, or have acute medical issues.

Maintaining confidentiality of personal and health information in a shared medical appointment is an important concern for patients but can be appropriately managed. In a survey of patients attending pulmonary hypertension shared medical appointments, 24% had concerns about confidentiality before participating, but after a few sessions, this rate was cut in half.²⁸

Patients have reported initially withholding some information, but over time, they usually become more comfortable with the group and disclose more helpful information.²⁹ Strategies to ensure confidentiality include having patients sign a confidentiality agreement at each appointment, providing specific instruction on what characterizes confidentiality breaches, and allowing patients the opportunity to schedule individual appointments as needed.

Ensuring insurance coverage

A shared medical appointment should be billed as an individual medical appointment for level of care, rather than time spent with the provider. This ensures that insurance coverage and copayments are the same as for individual medical appointments.

Lack of insurance coverage is a major barrier to obesity treatment in general. The US Centers for Medicare and Medicaid Services reimburses intensive behavioral obesity treatment delivered by a primary care practitioner, but limits it to 1 year of treatment and requires patients to meet weight loss goals. Some individual and employer-based healthcare plans do not cover dietitian visits, weight management programs, or antiobesity prescriptions.

EVIDENCE OF EFFECTIVENESS IN OBESITY

Few studies have investigated the use of shared medical appointments in obesity treatment. In the pediatric population, these programs significantly decreased BMI and some other anthropometric measurements,^30–32 but they did not consistently involve a prescribing provider. This means they did not manage medications or comorbidities as would be expected in a shared medical appointment.

In adults, reported effects have been encouraging, although the studies are not particularly robust. In a 2-year observational study of a single physician conducting biweekly weight management shared medical appointments, participants lost 1% of their baseline weight, while those continuing with usual care gained 0.8%, a statistically significant difference.³³ However, participation rates were low, with patients attending an average of only 3 shared medical appointments during the study.

In a meta-analysis of 13 randomized controlled trials of shared medical appointments for patients with type 2 diabetes, only 3 studies reported weight outcomes.²³ These results indicated a trend toward weight loss among patients attending shared appointments, but they were not statistically significant.

Positive results also were reported by the Veterans Administration’s MOVE! (Managing Overweight/obesity for Veterans Everywhere) program.³⁴ Participants in shared medical appointments reported that they felt empowered to make positive lifestyle changes, gained knowledge about obesity, were held accountable by their peers, and appreciated the individualized care they received from the multidisciplinary healthcare teams.

A systematic review involving 336 participants in group-based obesity interventions found group treatment produced more robust weight loss than individual treatment.³⁵ However, shared medical appointments are different from weight loss groups in that they combine an educational session and a medical appointment in a peer-group setting, which requires a provider with prescribing privileges to be present. Thus, shared medical appointments can manage medications as well as weight-related comorbidities such as diabetes, hypertension, polycystic ovarian syndrome, and hyperlipidemia.

One more point is that continued attendance at shared medical appointments, even after successful weight loss, may help to maintain the weight loss, which has otherwise been found to be extremely challenging using traditional medical approaches.

WHO SHOULD BE ON THE TEAM?

Because obesity is multifactorial, it requires a comprehensive treatment approach that can be difficult to deliver given the limited time of an individual appointment. In a shared appointment, providers across multiple specialties can meet with patients at the same time to coordinate approaches to obesity treatment.

A multidisciplinary team for shared medical appointments for obesity needs a physician or a nurse practitioner—or ideally, both— who specializes in obesity to facilitate the session. Other key providers include a registered dietitian, an exercise physiologist, a behavioral health specialist, a sleep specialist, and a social worker to participate as needed in the educational component of the appointment or act as outside consultants.

WHAT ARE REALISTIC TARGETS?

• Nutrition
• Physical activity
• Appetite control
• Sleep
• Stress and mood disorders.

Nutrition

A calorie deficit of 500 to 750 calories per day is recommended for weight loss.^7,8 Although there is no consensus on the best nutritional content of a diet, adherence to a diet is a significant predictor of weight loss.³⁶ One reason diets fail to bring about weight loss is that patients tend to underestimate their caloric intake by almost 50%.³⁷ Thus, they may benefit from a structured and supervised diet plan.

A dietitian can help patients develop an individualized diet plan that will promote adherence, which includes specific information on food choices, portion sizes, and timing of meals.

Physical activity

At least 150 minutes of physical activity per week is recommended for weight loss, and 200 to 300 minutes per week is recommended for long-term weight maintenance.^7,8

An exercise physiologist can help patients design a personalized exercise plan to help achieve these goals. This plan should take into account the patient’s cardiac status, activity level, degree of mobility, and lifestyle.

Most patients are not able to achieve the recommended physical activity goals initially, and activity levels need to be gradually increased over a period of weeks to months. Patients who were previously inactive or have evidence of cardiovascular, renal, or metabolic disease may require a cardiopulmonary assessment, including an electrocardiogram and cardiac stress test, before starting an exercise program.

It is very difficult for patients to lose weight without appetite control. Weight loss that results from diet and exercise is often accompanied by a change in weight-regulating hormones (eg, leptin, ghrelin, peptide YY, and cholecystokinin) that promote weight regain.³⁸ Thus, multiple compensatory mechanisms promote weight regain through increases in appetite and decreases in energy expenditure, resisting weight loss efforts.

Antiobesity drugs can help mitigate these adaptive weight-promoting responses through several mechanisms. They are indicated for use with lifestyle interventions for patients with a BMI of at least 30 mg/kg² or a BMI of at least 27 kg/m² with an obesity-related comorbidity.

These drugs promote an additional 3% to 7% weight loss when added to lifestyle interventions.¹⁸ But their effects are limited without appropriate lifestyle interventions.

Sleep

Adequate sleep is an often-overlooked component of obesity treatment. Inadequate sleep is associated with weight gain and an appetite-inducing hormone profile.³⁹ Just 2 days of sleep deprivation in healthy normal-weight adult men was associated with a 70% increase in the ghrelin-to-leptin ratio, which showed a linear relationship with self-reported increased hunger.³⁹ Sleep disorders, especially obstructive sleep apnea, are common in patients with obesity but are often underdiagnosed and undertreated.⁴⁰

Healthy sleep habits and sleep quality should be addressed in shared medical appointments for obesity, as patients may be unaware of the impact that sleep may be having on their obesity treatment. The STOP-BANG questionnaire (­snoring, tiredness, observed apnea, high blood pressure, BMI, age, neck circumference, and male sex) is a simple and reliable tool to screen for obstructive sleep apnea.⁴¹ Patients with symptoms of a sleep disorder should be referred to a sleep specialist for diagnosis and management.

Stress management and mood disorders

Stress and psychiatric disorders are underappreciated contributors to obesity. All patients receiving obesity treatment need to be screened for mood disorders and suicidal ideation.⁸

Chronic stress promotes weight gain through activation of the hypothalamic-pituitary-adrenocortical axis, whereby increased cortisol levels enhance appetite and accumulation of visceral fat.⁴² In addition, obesity is associated with a 25% increased risk of mood disorders, although the mechanism and direction of this association are unclear.⁴³ Weight gain as a side effect of antidepressant or other psychiatric medications is another important consideration.

Management of stress and psychiatric disorders through goal-setting, self-monitoring, and patient education is vital to help patients fully participate in lifestyle changes and maximize weight loss. Patients participating in shared medical appointments usually benefit from consultations with psychiatrists or psychologists to manage psychiatric comorbidities and assist with adherence to behavior modification. 


IN FAVOR OF SHARED MEDICAL APPOINTMENTS FOR OBESITY

Shared medical appointments can be an effective method of addressing the challenges of treating patients with obesity, using a multidisciplinary approach that combines nutrition, physical activity, appetite suppression, sleep improvement, and stress management. In addition, shared appointments allow practitioners to treat the primary problem of excess weight, rather than just its comorbidities, recognizing that obesity is a chronic disease that requires long-term, individualized treatment. Satisfaction rates are high for both patients and providers. Overall, education is essential to implementing and maintaining a successful shared medical appointment program.

The United States has seen a decline in fatal myocardial infarctions, largely thanks to early detection of coronary artery disease. Current guidelines on assessment of cardiovascular risk still rely on the traditional 10-year risk model in clinical practice. However, the predictive value of this approach is only moderate, and many coronary events occur in people considered to be at low or intermediate risk.

See related editorial

Coronary artery calcium scoring has emerged as a means of risk stratification by direct measurement of disease. Primary care providers are either using it or are seeing it used by consulting physicians, and its relatively low cost and ease of performance have contributed to its widespread use. However, downstream costs, radiation exposure, and lack of randomized controlled trials have raised concerns.

This article reviews the usefulness and pitfalls of coronary artery calcium scoring, providing a better understanding of the test, its limitations, and the interpretation of results.

ATHEROSCLEROSIS AND CALCIUM

As the calcium deposits grow, they can be detected by imaging tests such as computed tomography (CT), and quantified to assess the extent of disease.⁴

CALCIFICATION AND CORONARY ARTERY DISEASE

Coronary calcification occurs almost exclusively in atherosclerosis. Several autopsy studies^5,6 and histopathologic studies⁷ have shown a direct relationship between the extent of calcification and atherosclerotic disease.

Sangiorgi et al⁷ performed a histologic analysis of 723 coronary artery segments. The amount of calcium correlated well with the area of plaque:

• r = 0.89, P < .0001 in the left anterior descending artery
• r = 0.7, P < .001 in the left circumflex artery
• r = 0.89, P < .0001 in the right coronary artery.

Coronary artery calcium has also been associated with obstructive coronary artery disease in studies using intravascular ultrasonography and optical coherence tomography.^8,9

TECHNICAL INFORMATION ABOUT THE TEST

First-generation CT scanners used for calcium scoring in the 1980s were electron-beam systems in which a stationary x-ray tube generated an oscillating electron beam, which was reflected around the patient table.¹⁰ A single, stationary detector ring captured the images.

These systems have been replaced by multidetector scanners, in which the x-ray tube and multiple rows of detectors are combined in a gantry that rotates at high speed around the patient.

Coronary calcium is measured by noncontrast CT of the heart. Thus, there is no risk of contrast-induced nephropathy or allergic reactions. Images are acquired while the patient holds his or her breath for 3 to 5 seconds. Electrocardiographic gating is used to reduce motion artifact.^11,12 With modern scanners, the effective radiation dose associated with calcium testing is as low as 0.5 to 1.5 mSv,^13,14 ie, about the same dose as that with mammography. The entire test takes 10 to 15 minutes.

The results fall into 4 categories, which correlate with the severity of coronary artery disease, ranging from no significant disease to severe disease (Table 1). Other scores, which are not commonly used, include the calcium volume score¹⁶ and the calcium mass score.¹⁷Figure 2 shows a screenshot from a coronary artery calcium scoring program. 

Early multicenter studies evaluated the utility of calcium scoring to predict coronary stenosis in patients who underwent both cardiac CT and coronary angiography. The sensitivity of calcium scoring for angiographically significant disease was high (95%), but its specificity was low (about 44%).¹⁸

Budoff et al,¹⁹ reviewing these and subsequent results, concluded that the value of calcium scoring is its high negative predictive value (about 98%); a negative score (no calcification) is strongly associated with the absence of obstructive coronary disease.

Blaha et al²⁰ concluded that a score of 0 would indicate that the patient had a low risk of cardiovascular disease. A test with these characteristics is helpful in excluding cardiovascular disease or at least in determining that it is less likely to be present in a patient deemed to be at intermediate risk.

CALCIUM SCORING AS A PROGNOSTIC TOOL

Early on, investigators recognized the value of calcium scoring in predicting the risk of future cardiovascular events and death.^21–25

Predicting cardiovascular events

Pletcher et al21 performed a meta-analysis of studies that measured calcification in asymptomatic patients with subsequent follow-up. The summary-adjusted relative risk of cardiac events such as myocardial infarction, coronary artery revascularization, and coronary heart disease-related death rose with the calcium score:

• 2.1 (95% confidence interval [CI] 1.6–2.9) with a score of 1 to 100
• 4.2 (95% CI 2.5–7.2) with scores of 101 to 400
• 7.2 (95% CI 3.9-13.0) with scores greater than 400.

The meta-analysis was limited in that it included only 4 studies, which were observational.

Kavousi et al,²² in a subsequent meta-analysis of 6,739 women at low risk of atherosclerotic cardiovascular disease based on the American College of Cardiology/American Heart Association (ACC/AHA) pooled cohort equation (10-year risk < 7.5%), found that 36.1% had calcium scores greater than 0. Compared with those whose score was 0, those with higher scores had a higher risk of atherosclerotic cardiovascular disease events. The incidence rates per 1,000 person-years were 1.41 vs 4.33 (relative risk 2.92, 95% CI 2.02–3.83; multivariable-adjusted hazard ratio 2.04, 95% CI 1.44–2.90). This study was limited because the population was mostly of European descent, making it less generalizable to non-European populations.

Calcium scoring has also been shown to be a strong predictor of incident cardiovascular events across different races beyond traditional risk factors such as hypertension, hyperlipidemia, and tobacco use.

Detrano et al,²³ in a study of 6,722 patients with diverse ethnic backgrounds, found that the adjusted risk of a coronary event was increased by a factor of 7.73 for calcium scores between 101 and 300 and by a factor of 9.67 for scores above 300 (P < .001). A limitation of this study was that the patients and physicians were informed of the scores, which could have led to bias.

Carr et al²⁴ found an association between calcium and coronary heart disease in a younger population (ages 32–46). In 12.5 years of follow-up, the hazard ratio for cardiovascular events increased exponentially with the calcium score:

• 2.6 (95% CI 1.0–5.7, P = .03) with calcium scores of 1 through 19
• 9.8 (95% CI 4.5–20.5, P < .001) with scores greater than 100.

Predicting mortality

Budoff et al,²⁵ in an observational study of 25,253 patients, found coronary calcium to be an independent predictor of mortality in a multivariable model controlling for age, sex, ethnicity, and cardiac risk factors (model chi-square = 2,017, P < .0001). However, most of the patients were already known to have cardiac risk factors, making the study findings less generalizable to the general population.

Nasir et al²⁶ found that mortality rates rose with the calcium score in a study with 44,052 participants. The annualized mortality rates per 1,000 person-years were:

• 0.87 (95% CI 0.72–1.06) with a score of 0
• 2.97 (95% CI 2.61–3.37) with scores of 1–100
• 6.90 (95% CI 6.02–7.90) with scores of 101–400
• 17.68 (95% CI 5.93–19.62) with scores higher than 400.

The mortality rate also rose with the number of traditional risk factors present, ie, current tobacco use, dyslipidemia, diabetes mellitus, hypertension, and family history of coronary artery disease. Interestingly, those with no risk factors but a calcium score greater than 400 had a higher mortality rate than those with no coronary calcium but more than 3 risk factors (16.89 per 1,000 person-years vs 2.72 per 1,000 person years). As in the previous study, the patient population that was analyzed was at high risk and therefore the findings are not generalizable.

Shaw et al²⁷ found that patients without symptoms but with elevated coronary calcium scores had higher all-cause mortality rates at 15 years than those with a score of 0. The difference remained significant after Cox regression was performed, adjusting for traditional risk factors.

Current guidelines on assessing risk still rely on the traditional 10-year risk model in clinical practice.²⁵ Patients are thus classified as being at low, intermediate, or high risk based on their probability of developing a cardiovascular event or cardiovascular disease-related death in the subsequent 10 years.

However, the predictive value of this approach is only moderate,²⁸ and a significant number of cardiovascular events, including sudden cardiac death, occur in people who were believed to be at low or intermediate risk according to traditional risk factor-based predictions. Because risk scores are strongly influenced by age,²⁹ they are least reliable in young adults.³⁰

Akosah et al³¹ reviewed the records of 222 young adults (women age 55 or younger, men age 65 or younger) who presented with their first myocardial infarction, and found that only 25% would have qualified for primary prevention pharmacologic treatments according to the National Cholesterol Education Program III guidelines.^32,33 Similar findings have been reported regarding previous versions of the risk scores.³³

Thus, risk predictions based exclusively on traditional risk factors are not sensitive for detecting young individuals at increased risk, and lead to late treatment of young adults with atherosclerosis, which may be a less effective strategy.³⁴

The reliance on age in risk algorithms also results in low specificity in elderly adults. Using risk scores, elderly adults are systematically stratified in higher risk categories, expanding the indication for statin therapy to almost all men age 65 or older regardless of their actual vascular health, according to current clinical practice guidelines.^35,36

Risk scores are based on self-reported history and single-day measurements, since this kind of information is readily available to the physician in the clinic. Moreover, our knowledge about genetic and epigenetic factors associated with the development of atherosclerosis is still in its infancy, with current guidelines not supporting genetic testing as part of cardiovascular risk assessment.³⁷ Thus, a reliable measure of an individual’s lifelong exposure to a number of environmental and genetic factors that may affect cardiovascular health appears unfeasible.

Atherosclerosis is a process in which interactions between genetic, epigenetic, environmental, and traditional risk factors result in subclinical inflammation that could develop into clinically significant disease. Therefore, subclinical coronary atherosclerosis has been shown to be a strong predictor of future incident cardiovascular disease events and death. Thus, alternative approaches that directly measure disease, such as calcium scoring, may help further refine risk stratification of cardiovascular disease.

The MESA trial (Multi-Ethnic Study of Atherosclerosis), for instance, in 6,814 participants, found coronary calcium to provide better discrimination and risk reclassification than the ankle-brachial index, high sensitivity C-reactive protein level, and family history.³⁸ Coronary calcium also had the highest incremental improvement of the area under the receiver operating curve when added to the Framingham Risk Score (0.623 vs 0.784).

Reclassifying cardiovascular risk also has implications regarding whether to start therapies such as statins and aspirin.

For considering statin therapy

Nasir et al³⁹ showed that, in patients eligible for statin therapy by the pooled cohort equation, the absence of coronary artery calcium reclassified approximately one-half of candidates as not eligible for statin therapy. The number needed to treat to prevent an atherosclerotic cardiovascular event in the population who were recommended a statin was 64 with a calcium score of 0, and 24 with a calcium score greater than 100. In the population for whom a statin was considered, the number needed to treat was 223 with a calcium score of 0 and 46 for those with a score greater than 100. Moreover, 57% of intermediate-risk patients and 41% of high-risk patients based on the Framingham Risk Score were found to have a calcium score of 0, implying that these patients may actually be at a lower risk.

The Society of Cardiovascular Computed Tomography guidelines⁴⁰ say that statin therapy can be considered in patients who have a calcium score greater than 0.

For considering aspirin therapy

Miedema et al⁴¹ studied the role of coronary artery calcium in guiding aspirin therapy in 4,229 participants in the MESA trial who were not taking aspirin at baseline. Those with a calcium score higher than 100 had a number needed to treat of 173 in the group with a Framingham Risk Score less than 10% and 92 with a Framingham Risk Score of 10% or higher. The estimated number needed to harm for a major bleeding event was 442. For those who had a score of 0, the estimated number needed to treat was 2,036 for a Framingham Risk Score less than 10% and 808 for a Framingham Risk Score of 10% or higher, with an estimated number needed to harm of 442 for a major bleeding event.

The Society of Cardiovascular Computed Tomography guidelines⁴⁰ recommend considering aspirin therapy for patients with a coronary calcium score of more than 100.

McClelland et al⁴² developed a MESA risk score to predict 10-year risk of coronary heart disease using the traditional risk factors along with coronary calcium. The score was validated externally with 2 separate longitudinal studies. Thus, this may serve as another tool to help providers further risk-stratify patients.

COST-EFFECTIVENESS OF THE TEST

As coronary calcium measurement began to be widely used, concerns were raised about the lack of data on its cost-effectiveness.

Cost-effectiveness depends not only on patient selection but also on the cost of therapy. For example, if the cost of a generic statin is $85 per year, then calcium scoring would not be beneficial. However, if the cost of a statin is more than $200, then calcium scoring would be much more cost-effective, offering a way to avoid treating some patients who do not need to be treated.⁴³

Hong et al⁴³ showed that coronary calcium testing was cost-effective when the patient and physician share decision-making about initiating statin therapy. This is especially important if the patient has financial limitations, is concerned about side effects, or wants to avoid taking unnecessary medications.

According to some reports, $8.5 billion is spent annually for low-value care.⁴⁴ Many of the 80 million CT scans performed annually in the United States are believed to be unnecessary and may lead to additional testing to investigate incidental findings.⁴⁵

Growing use of coronary calcium measurement has raised similar concerns about radiation exposure, healthcare costs, and increased downstream testing triggered by the detection of incidental noncardiac findings. For instance, Onuma et al⁴⁶ reported that, in 503 patients undergoing CT to evaluate coronary artery disease, noncardiac findings were seen in 58.1% of them, but only 22.7% of the 503 had clinically significant findings.

Some of these concerns have been addressed. Modern scanners can acquire images in only a few seconds, entailing lower radiation doses than in the past.^13,14 The cost of the test is currently less than $100 in many US metropolitan areas.⁴⁷ However, further studies are needed to adequately and cost-effectively guide follow-up imaging of incidental noncardiac findings.⁴⁸

An important limitation of calcium scoring for risk assessment is that no randomized controlled trial has evaluated the impact of preventive interventions guided by calcium scores on hard event outcomes. It can be argued that there have been plenty of observational studies that have shown the benefit of coronary calcium scoring when judiciously done in the appropriate population.⁴⁹ Similarly, no randomized controlled trial has tested the pooled cohort equation and the application of statins based on its use with the current guidelines. The feasibility and cost of a large randomized controlled trial to assess outcomes after coronary artery calcium measurement must also be considered.

Another limitation of coronary calcium scoring is that it cannot rule out the presence of noncalcified atherosclerotic plaque, which often is more unstable and prone to rupture.

In addition, calcification in the coronary vascular bed (even if severe) does not necessarily mean there is clinically relevant coronary stenosis. For instance, an asymptomatic patient could have a coronary artery calcium score higher than 100 and then get a coronary angiogram that reveals only a 30% lesion in the left anterior descending coronary artery. This is because accumulation of (calcified) plaque in the vessel wall is accommodated by expansion of vessel diameter, maintaining luminal dimensions (positive remodeling). By definition, this patient does have coronary artery disease but would be best served by medical management. This could have been determined without an invasive test in an otherwise asymptomatic patient. Thus, performing coronary angiography based on a coronary artery calcium score alone would not have changed this patient’s management and may have exposed the patient to risks of procedural complications, in addition to extra healthcare costs. Therefore, the presence or absence of symptoms should guide the clinician on whether to pursue stress testing for invasive coronary angiography based on the appropriate use criteria.^50,51

WHO SHOULD BE TESTED?

In the ACC/AHA 2013 guidelines,³⁷ coronary calcium scoring has a class IIB recommendation in scenarios where it may appear that the risk-based treatment decision is uncertain after formal risk estimation has been done. As discussed above, a score higher than 100 could be a rationale for starting aspirin therapy, and a score higher than 0 for statin therapy. The current guidelines also mention that the coronary calcium score is comparable to other predictors such as the C-reactive protein level and the ankle-brachial index.

Compared with the ACC/AHA guidelines, the 2016 Society of Cardiovascular Computed Tomography guidelines and expert consensus recently have added more specifics in terms of using this test for asymptomatic patients at intermediate risk (10-year risk of atherosclerotic cardiovascular disease 5%–20%) and in selected patients with a family history of premature coronary artery disease and 10-year risk less than 5%.^40,52 The 2010 ACC/AHA guidelines were more specific, offering a class IIA recommendation for patients who were at intermediate risk (Framingham Risk Score 10%–20%).⁵³

The ACC/AHA cited cost and radiation exposure as reasons they did not give coronary calcium measurement a stronger recommendation.³⁷ However, as data continue to come in, the guidelines may change, especially since low-dose radiation tools are being used for cancer screening (lungs and breast) and since the cost has declined over the past decade.

OUR APPROACH

Given the negative predictive value of the coronary calcium score, our approach has been to use this test in asymptomatic patients who are found to be at intermediate risk of atherosclerotic cardiovascular disease based on the ACC/AHA risk calculation and are reluctant to start pharmacologic therapy, or who want a more personalized measure of coronary artery disease. This is preceded by a lengthy patient-physician discussion about the risks and benefits of the test.⁵⁴

The patient’s risk can then be further clarified and possibly reclassified as either low or high if it doesn’t remain intermediate. A discussion can then take place on potentially starting pharmacologic therapy, intensive lifestyle modifications, or both.^54,55 If an electronic medical record is available, CT results can be shown to the patient in the office to point out coronary calcifications. Seeing the lesions may serve an as additional motivating factor as patients embark on primary preventive efforts.⁵⁶

Below, we describe cases of what we would consider appropriate and inappropriate use of coronary artery calcium scoring.

A 55-year-old man presents for an annual physical and is found to have a 10-year risk of atherosclerotic cardiovascular disease of 7%, placing him in the intermediate-risk category. Despite an extensive conversation about lifestyle modifications and pharmacologic therapy, he is reluctant to initiate these measures. He is otherwise asymptomatic. Would calcium scoring be reasonable?

Yes, it would be reasonable to perform coronary artery calcium scoring in an otherwise asymptomatic man to help reclassify his risk for a coronary vascular event. The objective data provided by the test could motivate the patient to undertake primary prevention efforts or, if his score is 0, to show that he may not need drug therapy.

Example 2

A 55-year-old man who has a family history of coronary artery disease, is an active smoker, and has diabetes mellitus presents to the clinic with 2 months of exertional chest pain that resolves with rest. Would coronary artery calcium scoring be reasonable?

This patient is symptomatic and is at high risk of coronary artery disease. Statin therapy is already indicated in the AHA/ACC guidelines, since he has diabetes. Therefore, calcium scoring would not be helpful, as it would not change this patient’s management. Instead, he would be best served by stress testing or coronary angiography based on the stability of his symptoms and cardiac biomarkers.

Example 3

A 30-year-old woman with no medical history presents with on-and-off chest pain at both exertion and rest. Her electrocardiogram is unremarkable, and cardiac enzyme tests are negative. Would coronary calcium scoring be reasonable?

This young patient’s story is not typical for coronary artery disease. Therefore, she has a low pretest probability of obstructive coronary artery disease. Moreover, calcium scoring may not be helpful because at her young age there has not been enough time for calcification to develop (median age is the fifth decade of life). Thus, she would be exposed to radiation unnecessarily at a young age.

What to do with an elevated calcium score?

Coronary artery calcification is now being incidentally detected as patients undergo CT for other reasons such as screening for lung cancer based on the US Preventive Services Task Force guidelines. Patients may also get the test done on their own and then present to a provider with an elevated score.

It is important to consider the entire clinical scenario in such patients and not just the score. If a patient presents with an elevated calcium score but has no symptoms and falls in the intermediate-risk group, there is evidence to suggest that he or she should be started on statin or aspirin therapy or both.

As mentioned above, an abnormal test result does not mean that the patient should undergo more-invasive testing such as cardiac catheterization or even stress testing, especially if he or she has no symptoms. However, if the patient is symptomatic, then further cardiac evaluation would be recommended.

SUMMARY

Measuring coronary artery calcium has been found to be valuable in detecting coronary artery disease and in predicting cardiovascular events and death. The test is relatively easy to perform, with newer technology allowing for less radiation and cost. It serves as a more personalized measure of disease and can help facilitate patient-physician discussions about starting pharmacologic therapy, especially if a patient is reluctant.

Currently, coronary calcium scoring has a class IIB recommendation in scenarios in which the risk-based treatment decision is uncertain after formal risk estimation has been done according to the ACC/AHA guideline. The Society of Cardiovascular Computed Tomography guideline and expert consensus documents are more specific in recommending the test in asymptomatic patients in the intermediate-risk group.

Limitations of calcium scoring include the possibility of unnecessary cardiovascular testing such as cardiac catheterization or stress testing being driven by the calcium score alone, as well as the impact of incidental findings. With increased reporting of the coronary calcium score in patients undergoing CT for lung cancer screening, the score should be interpreted in view of the entire clinical scenario.

The United States has seen a decline in fatal myocardial infarctions, largely thanks to early detection of coronary artery disease. Current guidelines on assessment of cardiovascular risk still rely on the traditional 10-year risk model in clinical practice. However, the predictive value of this approach is only moderate, and many coronary events occur in people considered to be at low or intermediate risk.

See related editorial

Coronary artery calcium scoring has emerged as a means of risk stratification by direct measurement of disease. Primary care providers are either using it or are seeing it used by consulting physicians, and its relatively low cost and ease of performance have contributed to its widespread use. However, downstream costs, radiation exposure, and lack of randomized controlled trials have raised concerns.

This article reviews the usefulness and pitfalls of coronary artery calcium scoring, providing a better understanding of the test, its limitations, and the interpretation of results.

ATHEROSCLEROSIS AND CALCIUM

As the calcium deposits grow, they can be detected by imaging tests such as computed tomography (CT), and quantified to assess the extent of disease.⁴

CALCIFICATION AND CORONARY ARTERY DISEASE

Coronary calcification occurs almost exclusively in atherosclerosis. Several autopsy studies^5,6 and histopathologic studies⁷ have shown a direct relationship between the extent of calcification and atherosclerotic disease.

Sangiorgi et al⁷ performed a histologic analysis of 723 coronary artery segments. The amount of calcium correlated well with the area of plaque:

• r = 0.89, P < .0001 in the left anterior descending artery
• r = 0.7, P < .001 in the left circumflex artery
• r = 0.89, P < .0001 in the right coronary artery.

Coronary artery calcium has also been associated with obstructive coronary artery disease in studies using intravascular ultrasonography and optical coherence tomography.^8,9

TECHNICAL INFORMATION ABOUT THE TEST

First-generation CT scanners used for calcium scoring in the 1980s were electron-beam systems in which a stationary x-ray tube generated an oscillating electron beam, which was reflected around the patient table.¹⁰ A single, stationary detector ring captured the images.

These systems have been replaced by multidetector scanners, in which the x-ray tube and multiple rows of detectors are combined in a gantry that rotates at high speed around the patient.

Coronary calcium is measured by noncontrast CT of the heart. Thus, there is no risk of contrast-induced nephropathy or allergic reactions. Images are acquired while the patient holds his or her breath for 3 to 5 seconds. Electrocardiographic gating is used to reduce motion artifact.^11,12 With modern scanners, the effective radiation dose associated with calcium testing is as low as 0.5 to 1.5 mSv,^13,14 ie, about the same dose as that with mammography. The entire test takes 10 to 15 minutes.

The results fall into 4 categories, which correlate with the severity of coronary artery disease, ranging from no significant disease to severe disease (Table 1). Other scores, which are not commonly used, include the calcium volume score¹⁶ and the calcium mass score.¹⁷Figure 2 shows a screenshot from a coronary artery calcium scoring program. 

Early multicenter studies evaluated the utility of calcium scoring to predict coronary stenosis in patients who underwent both cardiac CT and coronary angiography. The sensitivity of calcium scoring for angiographically significant disease was high (95%), but its specificity was low (about 44%).¹⁸

Budoff et al,¹⁹ reviewing these and subsequent results, concluded that the value of calcium scoring is its high negative predictive value (about 98%); a negative score (no calcification) is strongly associated with the absence of obstructive coronary disease.

Blaha et al²⁰ concluded that a score of 0 would indicate that the patient had a low risk of cardiovascular disease. A test with these characteristics is helpful in excluding cardiovascular disease or at least in determining that it is less likely to be present in a patient deemed to be at intermediate risk.

CALCIUM SCORING AS A PROGNOSTIC TOOL

Early on, investigators recognized the value of calcium scoring in predicting the risk of future cardiovascular events and death.^21–25

Predicting cardiovascular events

Pletcher et al21 performed a meta-analysis of studies that measured calcification in asymptomatic patients with subsequent follow-up. The summary-adjusted relative risk of cardiac events such as myocardial infarction, coronary artery revascularization, and coronary heart disease-related death rose with the calcium score:

• 2.1 (95% confidence interval [CI] 1.6–2.9) with a score of 1 to 100
• 4.2 (95% CI 2.5–7.2) with scores of 101 to 400
• 7.2 (95% CI 3.9-13.0) with scores greater than 400.

The meta-analysis was limited in that it included only 4 studies, which were observational.

Kavousi et al,²² in a subsequent meta-analysis of 6,739 women at low risk of atherosclerotic cardiovascular disease based on the American College of Cardiology/American Heart Association (ACC/AHA) pooled cohort equation (10-year risk < 7.5%), found that 36.1% had calcium scores greater than 0. Compared with those whose score was 0, those with higher scores had a higher risk of atherosclerotic cardiovascular disease events. The incidence rates per 1,000 person-years were 1.41 vs 4.33 (relative risk 2.92, 95% CI 2.02–3.83; multivariable-adjusted hazard ratio 2.04, 95% CI 1.44–2.90). This study was limited because the population was mostly of European descent, making it less generalizable to non-European populations.

Calcium scoring has also been shown to be a strong predictor of incident cardiovascular events across different races beyond traditional risk factors such as hypertension, hyperlipidemia, and tobacco use.

Detrano et al,²³ in a study of 6,722 patients with diverse ethnic backgrounds, found that the adjusted risk of a coronary event was increased by a factor of 7.73 for calcium scores between 101 and 300 and by a factor of 9.67 for scores above 300 (P < .001). A limitation of this study was that the patients and physicians were informed of the scores, which could have led to bias.

Carr et al²⁴ found an association between calcium and coronary heart disease in a younger population (ages 32–46). In 12.5 years of follow-up, the hazard ratio for cardiovascular events increased exponentially with the calcium score:

• 2.6 (95% CI 1.0–5.7, P = .03) with calcium scores of 1 through 19
• 9.8 (95% CI 4.5–20.5, P < .001) with scores greater than 100.

Predicting mortality

Budoff et al,²⁵ in an observational study of 25,253 patients, found coronary calcium to be an independent predictor of mortality in a multivariable model controlling for age, sex, ethnicity, and cardiac risk factors (model chi-square = 2,017, P < .0001). However, most of the patients were already known to have cardiac risk factors, making the study findings less generalizable to the general population.

Nasir et al²⁶ found that mortality rates rose with the calcium score in a study with 44,052 participants. The annualized mortality rates per 1,000 person-years were:

• 0.87 (95% CI 0.72–1.06) with a score of 0
• 2.97 (95% CI 2.61–3.37) with scores of 1–100
• 6.90 (95% CI 6.02–7.90) with scores of 101–400
• 17.68 (95% CI 5.93–19.62) with scores higher than 400.

The mortality rate also rose with the number of traditional risk factors present, ie, current tobacco use, dyslipidemia, diabetes mellitus, hypertension, and family history of coronary artery disease. Interestingly, those with no risk factors but a calcium score greater than 400 had a higher mortality rate than those with no coronary calcium but more than 3 risk factors (16.89 per 1,000 person-years vs 2.72 per 1,000 person years). As in the previous study, the patient population that was analyzed was at high risk and therefore the findings are not generalizable.

Shaw et al²⁷ found that patients without symptoms but with elevated coronary calcium scores had higher all-cause mortality rates at 15 years than those with a score of 0. The difference remained significant after Cox regression was performed, adjusting for traditional risk factors.

Current guidelines on assessing risk still rely on the traditional 10-year risk model in clinical practice.²⁵ Patients are thus classified as being at low, intermediate, or high risk based on their probability of developing a cardiovascular event or cardiovascular disease-related death in the subsequent 10 years.

However, the predictive value of this approach is only moderate,²⁸ and a significant number of cardiovascular events, including sudden cardiac death, occur in people who were believed to be at low or intermediate risk according to traditional risk factor-based predictions. Because risk scores are strongly influenced by age,²⁹ they are least reliable in young adults.³⁰

Akosah et al³¹ reviewed the records of 222 young adults (women age 55 or younger, men age 65 or younger) who presented with their first myocardial infarction, and found that only 25% would have qualified for primary prevention pharmacologic treatments according to the National Cholesterol Education Program III guidelines.^32,33 Similar findings have been reported regarding previous versions of the risk scores.³³

Thus, risk predictions based exclusively on traditional risk factors are not sensitive for detecting young individuals at increased risk, and lead to late treatment of young adults with atherosclerosis, which may be a less effective strategy.³⁴

The reliance on age in risk algorithms also results in low specificity in elderly adults. Using risk scores, elderly adults are systematically stratified in higher risk categories, expanding the indication for statin therapy to almost all men age 65 or older regardless of their actual vascular health, according to current clinical practice guidelines.^35,36

Risk scores are based on self-reported history and single-day measurements, since this kind of information is readily available to the physician in the clinic. Moreover, our knowledge about genetic and epigenetic factors associated with the development of atherosclerosis is still in its infancy, with current guidelines not supporting genetic testing as part of cardiovascular risk assessment.³⁷ Thus, a reliable measure of an individual’s lifelong exposure to a number of environmental and genetic factors that may affect cardiovascular health appears unfeasible.

Atherosclerosis is a process in which interactions between genetic, epigenetic, environmental, and traditional risk factors result in subclinical inflammation that could develop into clinically significant disease. Therefore, subclinical coronary atherosclerosis has been shown to be a strong predictor of future incident cardiovascular disease events and death. Thus, alternative approaches that directly measure disease, such as calcium scoring, may help further refine risk stratification of cardiovascular disease.

The MESA trial (Multi-Ethnic Study of Atherosclerosis), for instance, in 6,814 participants, found coronary calcium to provide better discrimination and risk reclassification than the ankle-brachial index, high sensitivity C-reactive protein level, and family history.³⁸ Coronary calcium also had the highest incremental improvement of the area under the receiver operating curve when added to the Framingham Risk Score (0.623 vs 0.784).

Reclassifying cardiovascular risk also has implications regarding whether to start therapies such as statins and aspirin.

For considering statin therapy

Nasir et al³⁹ showed that, in patients eligible for statin therapy by the pooled cohort equation, the absence of coronary artery calcium reclassified approximately one-half of candidates as not eligible for statin therapy. The number needed to treat to prevent an atherosclerotic cardiovascular event in the population who were recommended a statin was 64 with a calcium score of 0, and 24 with a calcium score greater than 100. In the population for whom a statin was considered, the number needed to treat was 223 with a calcium score of 0 and 46 for those with a score greater than 100. Moreover, 57% of intermediate-risk patients and 41% of high-risk patients based on the Framingham Risk Score were found to have a calcium score of 0, implying that these patients may actually be at a lower risk.

The Society of Cardiovascular Computed Tomography guidelines⁴⁰ say that statin therapy can be considered in patients who have a calcium score greater than 0.

For considering aspirin therapy

Miedema et al⁴¹ studied the role of coronary artery calcium in guiding aspirin therapy in 4,229 participants in the MESA trial who were not taking aspirin at baseline. Those with a calcium score higher than 100 had a number needed to treat of 173 in the group with a Framingham Risk Score less than 10% and 92 with a Framingham Risk Score of 10% or higher. The estimated number needed to harm for a major bleeding event was 442. For those who had a score of 0, the estimated number needed to treat was 2,036 for a Framingham Risk Score less than 10% and 808 for a Framingham Risk Score of 10% or higher, with an estimated number needed to harm of 442 for a major bleeding event.

The Society of Cardiovascular Computed Tomography guidelines⁴⁰ recommend considering aspirin therapy for patients with a coronary calcium score of more than 100.

McClelland et al⁴² developed a MESA risk score to predict 10-year risk of coronary heart disease using the traditional risk factors along with coronary calcium. The score was validated externally with 2 separate longitudinal studies. Thus, this may serve as another tool to help providers further risk-stratify patients.

COST-EFFECTIVENESS OF THE TEST

As coronary calcium measurement began to be widely used, concerns were raised about the lack of data on its cost-effectiveness.

Cost-effectiveness depends not only on patient selection but also on the cost of therapy. For example, if the cost of a generic statin is $85 per year, then calcium scoring would not be beneficial. However, if the cost of a statin is more than $200, then calcium scoring would be much more cost-effective, offering a way to avoid treating some patients who do not need to be treated.⁴³

Hong et al⁴³ showed that coronary calcium testing was cost-effective when the patient and physician share decision-making about initiating statin therapy. This is especially important if the patient has financial limitations, is concerned about side effects, or wants to avoid taking unnecessary medications.

According to some reports, $8.5 billion is spent annually for low-value care.⁴⁴ Many of the 80 million CT scans performed annually in the United States are believed to be unnecessary and may lead to additional testing to investigate incidental findings.⁴⁵

Growing use of coronary calcium measurement has raised similar concerns about radiation exposure, healthcare costs, and increased downstream testing triggered by the detection of incidental noncardiac findings. For instance, Onuma et al⁴⁶ reported that, in 503 patients undergoing CT to evaluate coronary artery disease, noncardiac findings were seen in 58.1% of them, but only 22.7% of the 503 had clinically significant findings.

Some of these concerns have been addressed. Modern scanners can acquire images in only a few seconds, entailing lower radiation doses than in the past.^13,14 The cost of the test is currently less than $100 in many US metropolitan areas.⁴⁷ However, further studies are needed to adequately and cost-effectively guide follow-up imaging of incidental noncardiac findings.⁴⁸

An important limitation of calcium scoring for risk assessment is that no randomized controlled trial has evaluated the impact of preventive interventions guided by calcium scores on hard event outcomes. It can be argued that there have been plenty of observational studies that have shown the benefit of coronary calcium scoring when judiciously done in the appropriate population.⁴⁹ Similarly, no randomized controlled trial has tested the pooled cohort equation and the application of statins based on its use with the current guidelines. The feasibility and cost of a large randomized controlled trial to assess outcomes after coronary artery calcium measurement must also be considered.

Another limitation of coronary calcium scoring is that it cannot rule out the presence of noncalcified atherosclerotic plaque, which often is more unstable and prone to rupture.

In addition, calcification in the coronary vascular bed (even if severe) does not necessarily mean there is clinically relevant coronary stenosis. For instance, an asymptomatic patient could have a coronary artery calcium score higher than 100 and then get a coronary angiogram that reveals only a 30% lesion in the left anterior descending coronary artery. This is because accumulation of (calcified) plaque in the vessel wall is accommodated by expansion of vessel diameter, maintaining luminal dimensions (positive remodeling). By definition, this patient does have coronary artery disease but would be best served by medical management. This could have been determined without an invasive test in an otherwise asymptomatic patient. Thus, performing coronary angiography based on a coronary artery calcium score alone would not have changed this patient’s management and may have exposed the patient to risks of procedural complications, in addition to extra healthcare costs. Therefore, the presence or absence of symptoms should guide the clinician on whether to pursue stress testing for invasive coronary angiography based on the appropriate use criteria.^50,51

WHO SHOULD BE TESTED?

In the ACC/AHA 2013 guidelines,³⁷ coronary calcium scoring has a class IIB recommendation in scenarios where it may appear that the risk-based treatment decision is uncertain after formal risk estimation has been done. As discussed above, a score higher than 100 could be a rationale for starting aspirin therapy, and a score higher than 0 for statin therapy. The current guidelines also mention that the coronary calcium score is comparable to other predictors such as the C-reactive protein level and the ankle-brachial index.

Compared with the ACC/AHA guidelines, the 2016 Society of Cardiovascular Computed Tomography guidelines and expert consensus recently have added more specifics in terms of using this test for asymptomatic patients at intermediate risk (10-year risk of atherosclerotic cardiovascular disease 5%–20%) and in selected patients with a family history of premature coronary artery disease and 10-year risk less than 5%.^40,52 The 2010 ACC/AHA guidelines were more specific, offering a class IIA recommendation for patients who were at intermediate risk (Framingham Risk Score 10%–20%).⁵³

The ACC/AHA cited cost and radiation exposure as reasons they did not give coronary calcium measurement a stronger recommendation.³⁷ However, as data continue to come in, the guidelines may change, especially since low-dose radiation tools are being used for cancer screening (lungs and breast) and since the cost has declined over the past decade.

OUR APPROACH

Given the negative predictive value of the coronary calcium score, our approach has been to use this test in asymptomatic patients who are found to be at intermediate risk of atherosclerotic cardiovascular disease based on the ACC/AHA risk calculation and are reluctant to start pharmacologic therapy, or who want a more personalized measure of coronary artery disease. This is preceded by a lengthy patient-physician discussion about the risks and benefits of the test.⁵⁴

The patient’s risk can then be further clarified and possibly reclassified as either low or high if it doesn’t remain intermediate. A discussion can then take place on potentially starting pharmacologic therapy, intensive lifestyle modifications, or both.^54,55 If an electronic medical record is available, CT results can be shown to the patient in the office to point out coronary calcifications. Seeing the lesions may serve an as additional motivating factor as patients embark on primary preventive efforts.⁵⁶

Below, we describe cases of what we would consider appropriate and inappropriate use of coronary artery calcium scoring.

A 55-year-old man presents for an annual physical and is found to have a 10-year risk of atherosclerotic cardiovascular disease of 7%, placing him in the intermediate-risk category. Despite an extensive conversation about lifestyle modifications and pharmacologic therapy, he is reluctant to initiate these measures. He is otherwise asymptomatic. Would calcium scoring be reasonable?

Yes, it would be reasonable to perform coronary artery calcium scoring in an otherwise asymptomatic man to help reclassify his risk for a coronary vascular event. The objective data provided by the test could motivate the patient to undertake primary prevention efforts or, if his score is 0, to show that he may not need drug therapy.

Example 2

A 55-year-old man who has a family history of coronary artery disease, is an active smoker, and has diabetes mellitus presents to the clinic with 2 months of exertional chest pain that resolves with rest. Would coronary artery calcium scoring be reasonable?

This patient is symptomatic and is at high risk of coronary artery disease. Statin therapy is already indicated in the AHA/ACC guidelines, since he has diabetes. Therefore, calcium scoring would not be helpful, as it would not change this patient’s management. Instead, he would be best served by stress testing or coronary angiography based on the stability of his symptoms and cardiac biomarkers.

Example 3

A 30-year-old woman with no medical history presents with on-and-off chest pain at both exertion and rest. Her electrocardiogram is unremarkable, and cardiac enzyme tests are negative. Would coronary calcium scoring be reasonable?

This young patient’s story is not typical for coronary artery disease. Therefore, she has a low pretest probability of obstructive coronary artery disease. Moreover, calcium scoring may not be helpful because at her young age there has not been enough time for calcification to develop (median age is the fifth decade of life). Thus, she would be exposed to radiation unnecessarily at a young age.

What to do with an elevated calcium score?

Coronary artery calcification is now being incidentally detected as patients undergo CT for other reasons such as screening for lung cancer based on the US Preventive Services Task Force guidelines. Patients may also get the test done on their own and then present to a provider with an elevated score.

It is important to consider the entire clinical scenario in such patients and not just the score. If a patient presents with an elevated calcium score but has no symptoms and falls in the intermediate-risk group, there is evidence to suggest that he or she should be started on statin or aspirin therapy or both.

As mentioned above, an abnormal test result does not mean that the patient should undergo more-invasive testing such as cardiac catheterization or even stress testing, especially if he or she has no symptoms. However, if the patient is symptomatic, then further cardiac evaluation would be recommended.

SUMMARY

Measuring coronary artery calcium has been found to be valuable in detecting coronary artery disease and in predicting cardiovascular events and death. The test is relatively easy to perform, with newer technology allowing for less radiation and cost. It serves as a more personalized measure of disease and can help facilitate patient-physician discussions about starting pharmacologic therapy, especially if a patient is reluctant.

Currently, coronary calcium scoring has a class IIB recommendation in scenarios in which the risk-based treatment decision is uncertain after formal risk estimation has been done according to the ACC/AHA guideline. The Society of Cardiovascular Computed Tomography guideline and expert consensus documents are more specific in recommending the test in asymptomatic patients in the intermediate-risk group.

Limitations of calcium scoring include the possibility of unnecessary cardiovascular testing such as cardiac catheterization or stress testing being driven by the calcium score alone, as well as the impact of incidental findings. With increased reporting of the coronary calcium score in patients undergoing CT for lung cancer screening, the score should be interpreted in view of the entire clinical scenario.

The United States has seen a decline in fatal myocardial infarctions, largely thanks to early detection of coronary artery disease. Current guidelines on assessment of cardiovascular risk still rely on the traditional 10-year risk model in clinical practice. However, the predictive value of this approach is only moderate, and many coronary events occur in people considered to be at low or intermediate risk.

See related editorial

Coronary artery calcium scoring has emerged as a means of risk stratification by direct measurement of disease. Primary care providers are either using it or are seeing it used by consulting physicians, and its relatively low cost and ease of performance have contributed to its widespread use. However, downstream costs, radiation exposure, and lack of randomized controlled trials have raised concerns.

This article reviews the usefulness and pitfalls of coronary artery calcium scoring, providing a better understanding of the test, its limitations, and the interpretation of results.

ATHEROSCLEROSIS AND CALCIUM

As the calcium deposits grow, they can be detected by imaging tests such as computed tomography (CT), and quantified to assess the extent of disease.⁴

CALCIFICATION AND CORONARY ARTERY DISEASE

Coronary calcification occurs almost exclusively in atherosclerosis. Several autopsy studies^5,6 and histopathologic studies⁷ have shown a direct relationship between the extent of calcification and atherosclerotic disease.

Sangiorgi et al⁷ performed a histologic analysis of 723 coronary artery segments. The amount of calcium correlated well with the area of plaque:

• r = 0.89, P < .0001 in the left anterior descending artery
• r = 0.7, P < .001 in the left circumflex artery
• r = 0.89, P < .0001 in the right coronary artery.

Coronary artery calcium has also been associated with obstructive coronary artery disease in studies using intravascular ultrasonography and optical coherence tomography.^8,9

TECHNICAL INFORMATION ABOUT THE TEST

First-generation CT scanners used for calcium scoring in the 1980s were electron-beam systems in which a stationary x-ray tube generated an oscillating electron beam, which was reflected around the patient table.¹⁰ A single, stationary detector ring captured the images.

These systems have been replaced by multidetector scanners, in which the x-ray tube and multiple rows of detectors are combined in a gantry that rotates at high speed around the patient.

Coronary calcium is measured by noncontrast CT of the heart. Thus, there is no risk of contrast-induced nephropathy or allergic reactions. Images are acquired while the patient holds his or her breath for 3 to 5 seconds. Electrocardiographic gating is used to reduce motion artifact.^11,12 With modern scanners, the effective radiation dose associated with calcium testing is as low as 0.5 to 1.5 mSv,^13,14 ie, about the same dose as that with mammography. The entire test takes 10 to 15 minutes.

The results fall into 4 categories, which correlate with the severity of coronary artery disease, ranging from no significant disease to severe disease (Table 1). Other scores, which are not commonly used, include the calcium volume score¹⁶ and the calcium mass score.¹⁷Figure 2 shows a screenshot from a coronary artery calcium scoring program. 

Early multicenter studies evaluated the utility of calcium scoring to predict coronary stenosis in patients who underwent both cardiac CT and coronary angiography. The sensitivity of calcium scoring for angiographically significant disease was high (95%), but its specificity was low (about 44%).¹⁸

Budoff et al,¹⁹ reviewing these and subsequent results, concluded that the value of calcium scoring is its high negative predictive value (about 98%); a negative score (no calcification) is strongly associated with the absence of obstructive coronary disease.

Blaha et al²⁰ concluded that a score of 0 would indicate that the patient had a low risk of cardiovascular disease. A test with these characteristics is helpful in excluding cardiovascular disease or at least in determining that it is less likely to be present in a patient deemed to be at intermediate risk.

CALCIUM SCORING AS A PROGNOSTIC TOOL

Early on, investigators recognized the value of calcium scoring in predicting the risk of future cardiovascular events and death.^21–25

Predicting cardiovascular events

Pletcher et al21 performed a meta-analysis of studies that measured calcification in asymptomatic patients with subsequent follow-up. The summary-adjusted relative risk of cardiac events such as myocardial infarction, coronary artery revascularization, and coronary heart disease-related death rose with the calcium score:

• 2.1 (95% confidence interval [CI] 1.6–2.9) with a score of 1 to 100
• 4.2 (95% CI 2.5–7.2) with scores of 101 to 400
• 7.2 (95% CI 3.9-13.0) with scores greater than 400.

The meta-analysis was limited in that it included only 4 studies, which were observational.

Kavousi et al,²² in a subsequent meta-analysis of 6,739 women at low risk of atherosclerotic cardiovascular disease based on the American College of Cardiology/American Heart Association (ACC/AHA) pooled cohort equation (10-year risk < 7.5%), found that 36.1% had calcium scores greater than 0. Compared with those whose score was 0, those with higher scores had a higher risk of atherosclerotic cardiovascular disease events. The incidence rates per 1,000 person-years were 1.41 vs 4.33 (relative risk 2.92, 95% CI 2.02–3.83; multivariable-adjusted hazard ratio 2.04, 95% CI 1.44–2.90). This study was limited because the population was mostly of European descent, making it less generalizable to non-European populations.

Calcium scoring has also been shown to be a strong predictor of incident cardiovascular events across different races beyond traditional risk factors such as hypertension, hyperlipidemia, and tobacco use.

Detrano et al,²³ in a study of 6,722 patients with diverse ethnic backgrounds, found that the adjusted risk of a coronary event was increased by a factor of 7.73 for calcium scores between 101 and 300 and by a factor of 9.67 for scores above 300 (P < .001). A limitation of this study was that the patients and physicians were informed of the scores, which could have led to bias.

Carr et al²⁴ found an association between calcium and coronary heart disease in a younger population (ages 32–46). In 12.5 years of follow-up, the hazard ratio for cardiovascular events increased exponentially with the calcium score:

• 2.6 (95% CI 1.0–5.7, P = .03) with calcium scores of 1 through 19
• 9.8 (95% CI 4.5–20.5, P < .001) with scores greater than 100.

Predicting mortality

Budoff et al,²⁵ in an observational study of 25,253 patients, found coronary calcium to be an independent predictor of mortality in a multivariable model controlling for age, sex, ethnicity, and cardiac risk factors (model chi-square = 2,017, P < .0001). However, most of the patients were already known to have cardiac risk factors, making the study findings less generalizable to the general population.

Nasir et al²⁶ found that mortality rates rose with the calcium score in a study with 44,052 participants. The annualized mortality rates per 1,000 person-years were:

• 0.87 (95% CI 0.72–1.06) with a score of 0
• 2.97 (95% CI 2.61–3.37) with scores of 1–100
• 6.90 (95% CI 6.02–7.90) with scores of 101–400
• 17.68 (95% CI 5.93–19.62) with scores higher than 400.

The mortality rate also rose with the number of traditional risk factors present, ie, current tobacco use, dyslipidemia, diabetes mellitus, hypertension, and family history of coronary artery disease. Interestingly, those with no risk factors but a calcium score greater than 400 had a higher mortality rate than those with no coronary calcium but more than 3 risk factors (16.89 per 1,000 person-years vs 2.72 per 1,000 person years). As in the previous study, the patient population that was analyzed was at high risk and therefore the findings are not generalizable.

Shaw et al²⁷ found that patients without symptoms but with elevated coronary calcium scores had higher all-cause mortality rates at 15 years than those with a score of 0. The difference remained significant after Cox regression was performed, adjusting for traditional risk factors.

Current guidelines on assessing risk still rely on the traditional 10-year risk model in clinical practice.²⁵ Patients are thus classified as being at low, intermediate, or high risk based on their probability of developing a cardiovascular event or cardiovascular disease-related death in the subsequent 10 years.

However, the predictive value of this approach is only moderate,²⁸ and a significant number of cardiovascular events, including sudden cardiac death, occur in people who were believed to be at low or intermediate risk according to traditional risk factor-based predictions. Because risk scores are strongly influenced by age,²⁹ they are least reliable in young adults.³⁰

Akosah et al³¹ reviewed the records of 222 young adults (women age 55 or younger, men age 65 or younger) who presented with their first myocardial infarction, and found that only 25% would have qualified for primary prevention pharmacologic treatments according to the National Cholesterol Education Program III guidelines.^32,33 Similar findings have been reported regarding previous versions of the risk scores.³³

Thus, risk predictions based exclusively on traditional risk factors are not sensitive for detecting young individuals at increased risk, and lead to late treatment of young adults with atherosclerosis, which may be a less effective strategy.³⁴

The reliance on age in risk algorithms also results in low specificity in elderly adults. Using risk scores, elderly adults are systematically stratified in higher risk categories, expanding the indication for statin therapy to almost all men age 65 or older regardless of their actual vascular health, according to current clinical practice guidelines.^35,36

Risk scores are based on self-reported history and single-day measurements, since this kind of information is readily available to the physician in the clinic. Moreover, our knowledge about genetic and epigenetic factors associated with the development of atherosclerosis is still in its infancy, with current guidelines not supporting genetic testing as part of cardiovascular risk assessment.³⁷ Thus, a reliable measure of an individual’s lifelong exposure to a number of environmental and genetic factors that may affect cardiovascular health appears unfeasible.

Atherosclerosis is a process in which interactions between genetic, epigenetic, environmental, and traditional risk factors result in subclinical inflammation that could develop into clinically significant disease. Therefore, subclinical coronary atherosclerosis has been shown to be a strong predictor of future incident cardiovascular disease events and death. Thus, alternative approaches that directly measure disease, such as calcium scoring, may help further refine risk stratification of cardiovascular disease.

The MESA trial (Multi-Ethnic Study of Atherosclerosis), for instance, in 6,814 participants, found coronary calcium to provide better discrimination and risk reclassification than the ankle-brachial index, high sensitivity C-reactive protein level, and family history.³⁸ Coronary calcium also had the highest incremental improvement of the area under the receiver operating curve when added to the Framingham Risk Score (0.623 vs 0.784).

Reclassifying cardiovascular risk also has implications regarding whether to start therapies such as statins and aspirin.

For considering statin therapy

Nasir et al³⁹ showed that, in patients eligible for statin therapy by the pooled cohort equation, the absence of coronary artery calcium reclassified approximately one-half of candidates as not eligible for statin therapy. The number needed to treat to prevent an atherosclerotic cardiovascular event in the population who were recommended a statin was 64 with a calcium score of 0, and 24 with a calcium score greater than 100. In the population for whom a statin was considered, the number needed to treat was 223 with a calcium score of 0 and 46 for those with a score greater than 100. Moreover, 57% of intermediate-risk patients and 41% of high-risk patients based on the Framingham Risk Score were found to have a calcium score of 0, implying that these patients may actually be at a lower risk.

The Society of Cardiovascular Computed Tomography guidelines⁴⁰ say that statin therapy can be considered in patients who have a calcium score greater than 0.

For considering aspirin therapy

Miedema et al⁴¹ studied the role of coronary artery calcium in guiding aspirin therapy in 4,229 participants in the MESA trial who were not taking aspirin at baseline. Those with a calcium score higher than 100 had a number needed to treat of 173 in the group with a Framingham Risk Score less than 10% and 92 with a Framingham Risk Score of 10% or higher. The estimated number needed to harm for a major bleeding event was 442. For those who had a score of 0, the estimated number needed to treat was 2,036 for a Framingham Risk Score less than 10% and 808 for a Framingham Risk Score of 10% or higher, with an estimated number needed to harm of 442 for a major bleeding event.

The Society of Cardiovascular Computed Tomography guidelines⁴⁰ recommend considering aspirin therapy for patients with a coronary calcium score of more than 100.

McClelland et al⁴² developed a MESA risk score to predict 10-year risk of coronary heart disease using the traditional risk factors along with coronary calcium. The score was validated externally with 2 separate longitudinal studies. Thus, this may serve as another tool to help providers further risk-stratify patients.

COST-EFFECTIVENESS OF THE TEST

As coronary calcium measurement began to be widely used, concerns were raised about the lack of data on its cost-effectiveness.

Cost-effectiveness depends not only on patient selection but also on the cost of therapy. For example, if the cost of a generic statin is $85 per year, then calcium scoring would not be beneficial. However, if the cost of a statin is more than $200, then calcium scoring would be much more cost-effective, offering a way to avoid treating some patients who do not need to be treated.⁴³

Hong et al⁴³ showed that coronary calcium testing was cost-effective when the patient and physician share decision-making about initiating statin therapy. This is especially important if the patient has financial limitations, is concerned about side effects, or wants to avoid taking unnecessary medications.

According to some reports, $8.5 billion is spent annually for low-value care.⁴⁴ Many of the 80 million CT scans performed annually in the United States are believed to be unnecessary and may lead to additional testing to investigate incidental findings.⁴⁵

Growing use of coronary calcium measurement has raised similar concerns about radiation exposure, healthcare costs, and increased downstream testing triggered by the detection of incidental noncardiac findings. For instance, Onuma et al⁴⁶ reported that, in 503 patients undergoing CT to evaluate coronary artery disease, noncardiac findings were seen in 58.1% of them, but only 22.7% of the 503 had clinically significant findings.

Some of these concerns have been addressed. Modern scanners can acquire images in only a few seconds, entailing lower radiation doses than in the past.^13,14 The cost of the test is currently less than $100 in many US metropolitan areas.⁴⁷ However, further studies are needed to adequately and cost-effectively guide follow-up imaging of incidental noncardiac findings.⁴⁸

An important limitation of calcium scoring for risk assessment is that no randomized controlled trial has evaluated the impact of preventive interventions guided by calcium scores on hard event outcomes. It can be argued that there have been plenty of observational studies that have shown the benefit of coronary calcium scoring when judiciously done in the appropriate population.⁴⁹ Similarly, no randomized controlled trial has tested the pooled cohort equation and the application of statins based on its use with the current guidelines. The feasibility and cost of a large randomized controlled trial to assess outcomes after coronary artery calcium measurement must also be considered.

Another limitation of coronary calcium scoring is that it cannot rule out the presence of noncalcified atherosclerotic plaque, which often is more unstable and prone to rupture.

In addition, calcification in the coronary vascular bed (even if severe) does not necessarily mean there is clinically relevant coronary stenosis. For instance, an asymptomatic patient could have a coronary artery calcium score higher than 100 and then get a coronary angiogram that reveals only a 30% lesion in the left anterior descending coronary artery. This is because accumulation of (calcified) plaque in the vessel wall is accommodated by expansion of vessel diameter, maintaining luminal dimensions (positive remodeling). By definition, this patient does have coronary artery disease but would be best served by medical management. This could have been determined without an invasive test in an otherwise asymptomatic patient. Thus, performing coronary angiography based on a coronary artery calcium score alone would not have changed this patient’s management and may have exposed the patient to risks of procedural complications, in addition to extra healthcare costs. Therefore, the presence or absence of symptoms should guide the clinician on whether to pursue stress testing for invasive coronary angiography based on the appropriate use criteria.^50,51

WHO SHOULD BE TESTED?

In the ACC/AHA 2013 guidelines,³⁷ coronary calcium scoring has a class IIB recommendation in scenarios where it may appear that the risk-based treatment decision is uncertain after formal risk estimation has been done. As discussed above, a score higher than 100 could be a rationale for starting aspirin therapy, and a score higher than 0 for statin therapy. The current guidelines also mention that the coronary calcium score is comparable to other predictors such as the C-reactive protein level and the ankle-brachial index.

Compared with the ACC/AHA guidelines, the 2016 Society of Cardiovascular Computed Tomography guidelines and expert consensus recently have added more specifics in terms of using this test for asymptomatic patients at intermediate risk (10-year risk of atherosclerotic cardiovascular disease 5%–20%) and in selected patients with a family history of premature coronary artery disease and 10-year risk less than 5%.^40,52 The 2010 ACC/AHA guidelines were more specific, offering a class IIA recommendation for patients who were at intermediate risk (Framingham Risk Score 10%–20%).⁵³

The ACC/AHA cited cost and radiation exposure as reasons they did not give coronary calcium measurement a stronger recommendation.³⁷ However, as data continue to come in, the guidelines may change, especially since low-dose radiation tools are being used for cancer screening (lungs and breast) and since the cost has declined over the past decade.

OUR APPROACH

Given the negative predictive value of the coronary calcium score, our approach has been to use this test in asymptomatic patients who are found to be at intermediate risk of atherosclerotic cardiovascular disease based on the ACC/AHA risk calculation and are reluctant to start pharmacologic therapy, or who want a more personalized measure of coronary artery disease. This is preceded by a lengthy patient-physician discussion about the risks and benefits of the test.⁵⁴

The patient’s risk can then be further clarified and possibly reclassified as either low or high if it doesn’t remain intermediate. A discussion can then take place on potentially starting pharmacologic therapy, intensive lifestyle modifications, or both.^54,55 If an electronic medical record is available, CT results can be shown to the patient in the office to point out coronary calcifications. Seeing the lesions may serve an as additional motivating factor as patients embark on primary preventive efforts.⁵⁶

Below, we describe cases of what we would consider appropriate and inappropriate use of coronary artery calcium scoring.

A 55-year-old man presents for an annual physical and is found to have a 10-year risk of atherosclerotic cardiovascular disease of 7%, placing him in the intermediate-risk category. Despite an extensive conversation about lifestyle modifications and pharmacologic therapy, he is reluctant to initiate these measures. He is otherwise asymptomatic. Would calcium scoring be reasonable?

Yes, it would be reasonable to perform coronary artery calcium scoring in an otherwise asymptomatic man to help reclassify his risk for a coronary vascular event. The objective data provided by the test could motivate the patient to undertake primary prevention efforts or, if his score is 0, to show that he may not need drug therapy.

Example 2

A 55-year-old man who has a family history of coronary artery disease, is an active smoker, and has diabetes mellitus presents to the clinic with 2 months of exertional chest pain that resolves with rest. Would coronary artery calcium scoring be reasonable?

This patient is symptomatic and is at high risk of coronary artery disease. Statin therapy is already indicated in the AHA/ACC guidelines, since he has diabetes. Therefore, calcium scoring would not be helpful, as it would not change this patient’s management. Instead, he would be best served by stress testing or coronary angiography based on the stability of his symptoms and cardiac biomarkers.

Example 3

A 30-year-old woman with no medical history presents with on-and-off chest pain at both exertion and rest. Her electrocardiogram is unremarkable, and cardiac enzyme tests are negative. Would coronary calcium scoring be reasonable?

This young patient’s story is not typical for coronary artery disease. Therefore, she has a low pretest probability of obstructive coronary artery disease. Moreover, calcium scoring may not be helpful because at her young age there has not been enough time for calcification to develop (median age is the fifth decade of life). Thus, she would be exposed to radiation unnecessarily at a young age.

What to do with an elevated calcium score?

Coronary artery calcification is now being incidentally detected as patients undergo CT for other reasons such as screening for lung cancer based on the US Preventive Services Task Force guidelines. Patients may also get the test done on their own and then present to a provider with an elevated score.

It is important to consider the entire clinical scenario in such patients and not just the score. If a patient presents with an elevated calcium score but has no symptoms and falls in the intermediate-risk group, there is evidence to suggest that he or she should be started on statin or aspirin therapy or both.

As mentioned above, an abnormal test result does not mean that the patient should undergo more-invasive testing such as cardiac catheterization or even stress testing, especially if he or she has no symptoms. However, if the patient is symptomatic, then further cardiac evaluation would be recommended.

SUMMARY

Measuring coronary artery calcium has been found to be valuable in detecting coronary artery disease and in predicting cardiovascular events and death. The test is relatively easy to perform, with newer technology allowing for less radiation and cost. It serves as a more personalized measure of disease and can help facilitate patient-physician discussions about starting pharmacologic therapy, especially if a patient is reluctant.

Currently, coronary calcium scoring has a class IIB recommendation in scenarios in which the risk-based treatment decision is uncertain after formal risk estimation has been done according to the ACC/AHA guideline. The Society of Cardiovascular Computed Tomography guideline and expert consensus documents are more specific in recommending the test in asymptomatic patients in the intermediate-risk group.

Limitations of calcium scoring include the possibility of unnecessary cardiovascular testing such as cardiac catheterization or stress testing being driven by the calcium score alone, as well as the impact of incidental findings. With increased reporting of the coronary calcium score in patients undergoing CT for lung cancer screening, the score should be interpreted in view of the entire clinical scenario.

Bisphosphonates, especially intravenous zoledronic acid, often cause influenza-like symptoms such as severe musculoskeletal pain, fever, headache, malaise, and fatigue, sometimes accompanied by nausea, vomiting, and diarrhea. As many as 30% of patients experience these symptoms, which are usually transient, last up to 1 week, and, in most patients, only rarely recur with subsequent infusions.

It is essential to counsel and reassure patients about these reactions before starting treatment. We recommend that patients take acetaminophen before intravenous bis­phosphonate infusions, and if an acute-phase reaction occurs, we provide adequate supportive care with acetaminophen or nonsteroidal anti-inflammatory drugs (NSAIDs). If patients report severe musculoskeletal pain, then consider discontinuing the bisphosphonate treatment.

INFLUENZA-LIKE SYMPTOMS

The acute-phase reaction is a transient inflammatory state characterized by influenza-like symptoms such as fever, myalgia, joint pain, and nausea. It often occurs within the first few days after initial exposure to a bisphosphonate. Patients tend to rate the symptoms as mild to moderate. Symptoms may recur with subsequent doses; however, the incidence rate decreases substantially with each subsequent dose.

With intravenous bisphosphonates

Reid et al1 analyzed data from a trial in which 7,765 postmenopausal women with osteoporosis were randomized to receive intravenous zoledronic acid or placebo; 42.4% of the zoledronic acid group experienced symptoms that could be attributed to an acute-phase reaction after the first infusion, compared with 11.7% of the placebo group (P < .0001). Statistically significant differences (P < .0001) in symptoms between the groups included the following:

• Fever 20.3% vs 2.5%
• Musculoskeletal symptoms 19.9% vs 4.7% 
• Gastrointestinal symptoms 7.8% vs 2.1%.

Of the patients describing musculoskeletal symptoms after receiving zoledronic acid, most (79%) described them as generalized pain or discomfort, while about 25% said they were regional, usually localized to the back, neck, chest, and shoulders, 5% described joint stiffness, and 2.5% reported joint swelling.¹

In this and other studies,^1–3 acute-phase reactions most commonly occurred within the first few days after the infusion and were rated as mild to moderate in 90% of cases.^1,2 Patients who reported an acute-phase reaction were not more likely to opt out of subsequent infusions. The authors postulated that this was most likely because acute-phase reactions were mild and transient, and most resolved within 1 week.¹ The incidence decreased with each subsequent infusion of zoledronic acid^1–3; rates of the acute-phase reaction at years 1, 2, and 3 were 30%, 7%, and 3%, respectively.¹

With oral bisphosphonates

The acute-phase reaction is less common with oral bisphosphonates (occurring in 5.6% of patients in a retrospective study4) and is usually less severe.^4,5

Musculoskeletal pain related to the acute-phase reaction is thought to be due to transient release of inflammatory cytokines such as interleukin 6, interferon gamma, and tumor necrosis factor alpha from macrophages, monocytes, and gamma-delta T cells.⁶

Bisphosphonates are taken up by osteoclasts and inhibit their function. But bisphosphonates are not all the same: they can be divided into aminobisphosphonates (eg, alendronate, pamidronate, risedronate, zoledronic acid) and nonaminobisphosphonates (eg, clodronate, etidronate).

Inside the osteoclasts, aminobisphosphonates inhibit farnesyl diphosphate synthase in the meval­onate pathways, thus disrupting cell signaling and leading to apoptosis.⁷ However, inhibition of farnesyl diphosphate synthase also increases intracellular levels of isopentyl pyrophosphate, which induces T-cell activation; this is thought to result in release of inflammatory cytokines, leading to the acute-phase reaction.^7,8

In contrast, nonaminobisphosphonates such as clodronate and etidronate, after being internalized, are metabolized into nonhydrolyzable adenosine triphosphate, which is toxic to the osteoclast. The acute-phase reaction or influenza-like illness is unique to aminobisphosphonates; nonaminobisphosphonates have not been associated with an acute-phase reaction.

TRIALS OF PREVENTIVE TREATMENT

With NSAIDs, acetaminophen

Wark et al⁹ randomized 481 postmenopausal women who had osteopenia but who had never received bisphosphonates to 4 treatment groups:

• Zoledronic acid 5 mg intravenously plus acetaminophen 1,000 mg every 6 hours for 3 days
• Zoledronic acid 5 mg intravenously plus ibuprofen 400 mg every 6 hours for 3 days
• Zoledronic acid 5 mg intravenously plus 2 placebo capsules every 6 hours for 3 days
• Placebo infusion plus 2 placebo capsules every 6 hours for 3 days.

Patients were assessed for fever and worsening symptoms over 3 days after the infusion. The group that got zoledronic acid infusion and placebo capsules had the highest rates of fever (64%) and worsening symptoms (76%); acetaminophen and ibuprofen reduced these rates to an approximately equal extent, to 37% for fever and 46% (acetaminophen) and 49% (ibuprofen) for worsening symptoms. The group that received placebo bisphosphonate infusions had the lowest rates of fever (11%) and worsening symptoms (17%).

Silverman et al¹⁰ found that acetaminophen 650 mg taken 45 minutes before intravenous zoledronic acid infusion and continued every 6 hours for 3 days led to an absolute risk reduction of 21% in the incidence of fever or use of rescue medication compared with placebo.

Trials of other agents

In a study of 60 women,¹¹ 30 received an oral bolus of cholecalciferol 300,000 IU 5 days before zoledronic acid 5 mg infusion plus daily calcium 1,000 mg and vitamin D 800 IU, and 30 received a placebo pill 5 days before the same infusion and vitamin regimen as the other group. The preinfusion oral bolus did not decrease the incidence of acute-phase reactions, although it led to a small decrease in the severity of musculoskeletal pain (the median pain score was reduced from 2 to 1 on a scale of 0 to 10).

Other interventions such as fluvastatin and oral dexamethasone given before intravenous zoledronic acid did not reduce the severity or incidence of the acute-phase reaction.^10,12,13

Before starting bisphosphonate therapy, patients should be counseled about the possibility of acute musculoskeletal pain and other symptoms of the acute-phase reaction.

For intravenous bisphosphonates

We advise all patients scheduled to receive intravenous bisphosphonates to take acetaminophen 650 to 1,000 mg once on the morning of the infusion. We prefer acetaminophen over NSAIDs for prophylaxis to avoid the gastric mucosal and renal toxicity more common with NSAIDs, especially in the elderly.

If the patient has a history of acute musculoskeletal pain or other symptoms of an acute-phase reaction after bisphosphonate infusion, we advise a more aggressive approach to prophylaxis: acetaminophen 650 mg 1 hour before the infusion, then every 6 hours for up to 3 days. This approach, with acetaminophen or NSAIDs, has been shown in large randomized controlled trials to reduce the incidence and severity of the acute-phase reaction.

If an acute-phase reaction occurs, we inform patients that the likelihood decreases and is quite low with subsequent doses. We provide correct and honest information, so that patients who experience an acute-phase reaction can make an informed decision about continuing bisphosphonate treatment or switching to another treatment. If the patient decides to continue with intravenous bisphosphonate treatment, we recommend more-aggressive prophylaxis with acetaminophen or NSAIDs with subsequent infusions.

For oral bisphosphonates

We do not prescribe prophylactic treatment with acetaminophen or NSAIDs with oral bisphosphonates, but we do advise patients to take acetaminophen or NSAIDs as needed for mild to moderate musculoskeletal pain, should this occur.

We try to continue treatment in mild to moderate cases, while monitoring the patient closely to see if the musculoskeletal pain resolves within 1 to 2 weeks.

If the pain is severe or does not resolve in 1 to 2 weeks, we offer switching to another drug class. Since musculoskeletal pain with oral bisphosphonates does not represent an allergic reaction, we have switched patients from oral to intravenous bisphosphonates without recurrence of musculoskeletal pain.

SEVERE MUSCULOSKELETAL PAIN BEYOND THE ACUTE PHASE

Severe musculoskeletal pain that may not be related to the acute-phase reaction, although rare, has been reported.^5,14 From 1995, when alendronate was approved for osteoporosis, through 2002, the US Food and Drug Administration received reports of severe musculoskeletal pain in 117 patients.¹⁵

This severe musculoskeletal pain related to bisphosphonate use remains poorly characterized. It has been reported to occur days or months (median time 14 days, range same day to 52 months) after starting bisphosphonate therapy and to resolve only if the bisphosphonate is stopped.^5,15 It differs from typical acute-phase reactions, which tend to occur with the initial dose (intravenous or oral) and resolve within several days. There are case reports of polyarthritis with synovitis that recurred with each bisphosphonate dose (oral or intravenous) and led to discontinuation of the bisphosphonate.^14,16–18

Clinicians need to be aware of the possibility of severe musculoskeletal pain and consider stopping bisphosphonate treatment in these cases. Besides discontinuation, acetaminophen, NSAIDs, and, in rare cases, glucocorticoids or short-term opiate therapy may be used for symptom control. In patients with a severe or persistent acute-phase reaction or musculoskeletal pain, discontinuation of bisphosphonates is warranted.

Bisphosphonates, especially intravenous zoledronic acid, often cause influenza-like symptoms such as severe musculoskeletal pain, fever, headache, malaise, and fatigue, sometimes accompanied by nausea, vomiting, and diarrhea. As many as 30% of patients experience these symptoms, which are usually transient, last up to 1 week, and, in most patients, only rarely recur with subsequent infusions.

It is essential to counsel and reassure patients about these reactions before starting treatment. We recommend that patients take acetaminophen before intravenous bis­phosphonate infusions, and if an acute-phase reaction occurs, we provide adequate supportive care with acetaminophen or nonsteroidal anti-inflammatory drugs (NSAIDs). If patients report severe musculoskeletal pain, then consider discontinuing the bisphosphonate treatment.

INFLUENZA-LIKE SYMPTOMS

The acute-phase reaction is a transient inflammatory state characterized by influenza-like symptoms such as fever, myalgia, joint pain, and nausea. It often occurs within the first few days after initial exposure to a bisphosphonate. Patients tend to rate the symptoms as mild to moderate. Symptoms may recur with subsequent doses; however, the incidence rate decreases substantially with each subsequent dose.

With intravenous bisphosphonates

Reid et al1 analyzed data from a trial in which 7,765 postmenopausal women with osteoporosis were randomized to receive intravenous zoledronic acid or placebo; 42.4% of the zoledronic acid group experienced symptoms that could be attributed to an acute-phase reaction after the first infusion, compared with 11.7% of the placebo group (P < .0001). Statistically significant differences (P < .0001) in symptoms between the groups included the following:

• Fever 20.3% vs 2.5%
• Musculoskeletal symptoms 19.9% vs 4.7% 
• Gastrointestinal symptoms 7.8% vs 2.1%.

Of the patients describing musculoskeletal symptoms after receiving zoledronic acid, most (79%) described them as generalized pain or discomfort, while about 25% said they were regional, usually localized to the back, neck, chest, and shoulders, 5% described joint stiffness, and 2.5% reported joint swelling.¹

In this and other studies,^1–3 acute-phase reactions most commonly occurred within the first few days after the infusion and were rated as mild to moderate in 90% of cases.^1,2 Patients who reported an acute-phase reaction were not more likely to opt out of subsequent infusions. The authors postulated that this was most likely because acute-phase reactions were mild and transient, and most resolved within 1 week.¹ The incidence decreased with each subsequent infusion of zoledronic acid^1–3; rates of the acute-phase reaction at years 1, 2, and 3 were 30%, 7%, and 3%, respectively.¹

With oral bisphosphonates

The acute-phase reaction is less common with oral bisphosphonates (occurring in 5.6% of patients in a retrospective study4) and is usually less severe.^4,5

Musculoskeletal pain related to the acute-phase reaction is thought to be due to transient release of inflammatory cytokines such as interleukin 6, interferon gamma, and tumor necrosis factor alpha from macrophages, monocytes, and gamma-delta T cells.⁶

Bisphosphonates are taken up by osteoclasts and inhibit their function. But bisphosphonates are not all the same: they can be divided into aminobisphosphonates (eg, alendronate, pamidronate, risedronate, zoledronic acid) and nonaminobisphosphonates (eg, clodronate, etidronate).

Inside the osteoclasts, aminobisphosphonates inhibit farnesyl diphosphate synthase in the meval­onate pathways, thus disrupting cell signaling and leading to apoptosis.⁷ However, inhibition of farnesyl diphosphate synthase also increases intracellular levels of isopentyl pyrophosphate, which induces T-cell activation; this is thought to result in release of inflammatory cytokines, leading to the acute-phase reaction.^7,8

In contrast, nonaminobisphosphonates such as clodronate and etidronate, after being internalized, are metabolized into nonhydrolyzable adenosine triphosphate, which is toxic to the osteoclast. The acute-phase reaction or influenza-like illness is unique to aminobisphosphonates; nonaminobisphosphonates have not been associated with an acute-phase reaction.

TRIALS OF PREVENTIVE TREATMENT

With NSAIDs, acetaminophen

Wark et al⁹ randomized 481 postmenopausal women who had osteopenia but who had never received bisphosphonates to 4 treatment groups:

• Zoledronic acid 5 mg intravenously plus acetaminophen 1,000 mg every 6 hours for 3 days
• Zoledronic acid 5 mg intravenously plus ibuprofen 400 mg every 6 hours for 3 days
• Zoledronic acid 5 mg intravenously plus 2 placebo capsules every 6 hours for 3 days
• Placebo infusion plus 2 placebo capsules every 6 hours for 3 days.

Patients were assessed for fever and worsening symptoms over 3 days after the infusion. The group that got zoledronic acid infusion and placebo capsules had the highest rates of fever (64%) and worsening symptoms (76%); acetaminophen and ibuprofen reduced these rates to an approximately equal extent, to 37% for fever and 46% (acetaminophen) and 49% (ibuprofen) for worsening symptoms. The group that received placebo bisphosphonate infusions had the lowest rates of fever (11%) and worsening symptoms (17%).

Silverman et al¹⁰ found that acetaminophen 650 mg taken 45 minutes before intravenous zoledronic acid infusion and continued every 6 hours for 3 days led to an absolute risk reduction of 21% in the incidence of fever or use of rescue medication compared with placebo.

Trials of other agents

In a study of 60 women,¹¹ 30 received an oral bolus of cholecalciferol 300,000 IU 5 days before zoledronic acid 5 mg infusion plus daily calcium 1,000 mg and vitamin D 800 IU, and 30 received a placebo pill 5 days before the same infusion and vitamin regimen as the other group. The preinfusion oral bolus did not decrease the incidence of acute-phase reactions, although it led to a small decrease in the severity of musculoskeletal pain (the median pain score was reduced from 2 to 1 on a scale of 0 to 10).

Other interventions such as fluvastatin and oral dexamethasone given before intravenous zoledronic acid did not reduce the severity or incidence of the acute-phase reaction.^10,12,13

Before starting bisphosphonate therapy, patients should be counseled about the possibility of acute musculoskeletal pain and other symptoms of the acute-phase reaction.

For intravenous bisphosphonates

We advise all patients scheduled to receive intravenous bisphosphonates to take acetaminophen 650 to 1,000 mg once on the morning of the infusion. We prefer acetaminophen over NSAIDs for prophylaxis to avoid the gastric mucosal and renal toxicity more common with NSAIDs, especially in the elderly.

If the patient has a history of acute musculoskeletal pain or other symptoms of an acute-phase reaction after bisphosphonate infusion, we advise a more aggressive approach to prophylaxis: acetaminophen 650 mg 1 hour before the infusion, then every 6 hours for up to 3 days. This approach, with acetaminophen or NSAIDs, has been shown in large randomized controlled trials to reduce the incidence and severity of the acute-phase reaction.

If an acute-phase reaction occurs, we inform patients that the likelihood decreases and is quite low with subsequent doses. We provide correct and honest information, so that patients who experience an acute-phase reaction can make an informed decision about continuing bisphosphonate treatment or switching to another treatment. If the patient decides to continue with intravenous bisphosphonate treatment, we recommend more-aggressive prophylaxis with acetaminophen or NSAIDs with subsequent infusions.

For oral bisphosphonates

We do not prescribe prophylactic treatment with acetaminophen or NSAIDs with oral bisphosphonates, but we do advise patients to take acetaminophen or NSAIDs as needed for mild to moderate musculoskeletal pain, should this occur.

We try to continue treatment in mild to moderate cases, while monitoring the patient closely to see if the musculoskeletal pain resolves within 1 to 2 weeks.

If the pain is severe or does not resolve in 1 to 2 weeks, we offer switching to another drug class. Since musculoskeletal pain with oral bisphosphonates does not represent an allergic reaction, we have switched patients from oral to intravenous bisphosphonates without recurrence of musculoskeletal pain.

SEVERE MUSCULOSKELETAL PAIN BEYOND THE ACUTE PHASE

Severe musculoskeletal pain that may not be related to the acute-phase reaction, although rare, has been reported.^5,14 From 1995, when alendronate was approved for osteoporosis, through 2002, the US Food and Drug Administration received reports of severe musculoskeletal pain in 117 patients.¹⁵

This severe musculoskeletal pain related to bisphosphonate use remains poorly characterized. It has been reported to occur days or months (median time 14 days, range same day to 52 months) after starting bisphosphonate therapy and to resolve only if the bisphosphonate is stopped.^5,15 It differs from typical acute-phase reactions, which tend to occur with the initial dose (intravenous or oral) and resolve within several days. There are case reports of polyarthritis with synovitis that recurred with each bisphosphonate dose (oral or intravenous) and led to discontinuation of the bisphosphonate.^14,16–18

Clinicians need to be aware of the possibility of severe musculoskeletal pain and consider stopping bisphosphonate treatment in these cases. Besides discontinuation, acetaminophen, NSAIDs, and, in rare cases, glucocorticoids or short-term opiate therapy may be used for symptom control. In patients with a severe or persistent acute-phase reaction or musculoskeletal pain, discontinuation of bisphosphonates is warranted.

Bisphosphonates, especially intravenous zoledronic acid, often cause influenza-like symptoms such as severe musculoskeletal pain, fever, headache, malaise, and fatigue, sometimes accompanied by nausea, vomiting, and diarrhea. As many as 30% of patients experience these symptoms, which are usually transient, last up to 1 week, and, in most patients, only rarely recur with subsequent infusions.

It is essential to counsel and reassure patients about these reactions before starting treatment. We recommend that patients take acetaminophen before intravenous bis­phosphonate infusions, and if an acute-phase reaction occurs, we provide adequate supportive care with acetaminophen or nonsteroidal anti-inflammatory drugs (NSAIDs). If patients report severe musculoskeletal pain, then consider discontinuing the bisphosphonate treatment.

INFLUENZA-LIKE SYMPTOMS

The acute-phase reaction is a transient inflammatory state characterized by influenza-like symptoms such as fever, myalgia, joint pain, and nausea. It often occurs within the first few days after initial exposure to a bisphosphonate. Patients tend to rate the symptoms as mild to moderate. Symptoms may recur with subsequent doses; however, the incidence rate decreases substantially with each subsequent dose.

With intravenous bisphosphonates

Reid et al1 analyzed data from a trial in which 7,765 postmenopausal women with osteoporosis were randomized to receive intravenous zoledronic acid or placebo; 42.4% of the zoledronic acid group experienced symptoms that could be attributed to an acute-phase reaction after the first infusion, compared with 11.7% of the placebo group (P < .0001). Statistically significant differences (P < .0001) in symptoms between the groups included the following:

• Fever 20.3% vs 2.5%
• Musculoskeletal symptoms 19.9% vs 4.7% 
• Gastrointestinal symptoms 7.8% vs 2.1%.

Of the patients describing musculoskeletal symptoms after receiving zoledronic acid, most (79%) described them as generalized pain or discomfort, while about 25% said they were regional, usually localized to the back, neck, chest, and shoulders, 5% described joint stiffness, and 2.5% reported joint swelling.¹

In this and other studies,^1–3 acute-phase reactions most commonly occurred within the first few days after the infusion and were rated as mild to moderate in 90% of cases.^1,2 Patients who reported an acute-phase reaction were not more likely to opt out of subsequent infusions. The authors postulated that this was most likely because acute-phase reactions were mild and transient, and most resolved within 1 week.¹ The incidence decreased with each subsequent infusion of zoledronic acid^1–3; rates of the acute-phase reaction at years 1, 2, and 3 were 30%, 7%, and 3%, respectively.¹

With oral bisphosphonates

The acute-phase reaction is less common with oral bisphosphonates (occurring in 5.6% of patients in a retrospective study4) and is usually less severe.^4,5

Musculoskeletal pain related to the acute-phase reaction is thought to be due to transient release of inflammatory cytokines such as interleukin 6, interferon gamma, and tumor necrosis factor alpha from macrophages, monocytes, and gamma-delta T cells.⁶

Bisphosphonates are taken up by osteoclasts and inhibit their function. But bisphosphonates are not all the same: they can be divided into aminobisphosphonates (eg, alendronate, pamidronate, risedronate, zoledronic acid) and nonaminobisphosphonates (eg, clodronate, etidronate).

Inside the osteoclasts, aminobisphosphonates inhibit farnesyl diphosphate synthase in the meval­onate pathways, thus disrupting cell signaling and leading to apoptosis.⁷ However, inhibition of farnesyl diphosphate synthase also increases intracellular levels of isopentyl pyrophosphate, which induces T-cell activation; this is thought to result in release of inflammatory cytokines, leading to the acute-phase reaction.^7,8

In contrast, nonaminobisphosphonates such as clodronate and etidronate, after being internalized, are metabolized into nonhydrolyzable adenosine triphosphate, which is toxic to the osteoclast. The acute-phase reaction or influenza-like illness is unique to aminobisphosphonates; nonaminobisphosphonates have not been associated with an acute-phase reaction.

TRIALS OF PREVENTIVE TREATMENT

With NSAIDs, acetaminophen

Wark et al⁹ randomized 481 postmenopausal women who had osteopenia but who had never received bisphosphonates to 4 treatment groups:

• Zoledronic acid 5 mg intravenously plus acetaminophen 1,000 mg every 6 hours for 3 days
• Zoledronic acid 5 mg intravenously plus ibuprofen 400 mg every 6 hours for 3 days
• Zoledronic acid 5 mg intravenously plus 2 placebo capsules every 6 hours for 3 days
• Placebo infusion plus 2 placebo capsules every 6 hours for 3 days.

Patients were assessed for fever and worsening symptoms over 3 days after the infusion. The group that got zoledronic acid infusion and placebo capsules had the highest rates of fever (64%) and worsening symptoms (76%); acetaminophen and ibuprofen reduced these rates to an approximately equal extent, to 37% for fever and 46% (acetaminophen) and 49% (ibuprofen) for worsening symptoms. The group that received placebo bisphosphonate infusions had the lowest rates of fever (11%) and worsening symptoms (17%).

Silverman et al¹⁰ found that acetaminophen 650 mg taken 45 minutes before intravenous zoledronic acid infusion and continued every 6 hours for 3 days led to an absolute risk reduction of 21% in the incidence of fever or use of rescue medication compared with placebo.

Trials of other agents

In a study of 60 women,¹¹ 30 received an oral bolus of cholecalciferol 300,000 IU 5 days before zoledronic acid 5 mg infusion plus daily calcium 1,000 mg and vitamin D 800 IU, and 30 received a placebo pill 5 days before the same infusion and vitamin regimen as the other group. The preinfusion oral bolus did not decrease the incidence of acute-phase reactions, although it led to a small decrease in the severity of musculoskeletal pain (the median pain score was reduced from 2 to 1 on a scale of 0 to 10).

Other interventions such as fluvastatin and oral dexamethasone given before intravenous zoledronic acid did not reduce the severity or incidence of the acute-phase reaction.^10,12,13

Before starting bisphosphonate therapy, patients should be counseled about the possibility of acute musculoskeletal pain and other symptoms of the acute-phase reaction.

For intravenous bisphosphonates

We advise all patients scheduled to receive intravenous bisphosphonates to take acetaminophen 650 to 1,000 mg once on the morning of the infusion. We prefer acetaminophen over NSAIDs for prophylaxis to avoid the gastric mucosal and renal toxicity more common with NSAIDs, especially in the elderly.

If the patient has a history of acute musculoskeletal pain or other symptoms of an acute-phase reaction after bisphosphonate infusion, we advise a more aggressive approach to prophylaxis: acetaminophen 650 mg 1 hour before the infusion, then every 6 hours for up to 3 days. This approach, with acetaminophen or NSAIDs, has been shown in large randomized controlled trials to reduce the incidence and severity of the acute-phase reaction.

If an acute-phase reaction occurs, we inform patients that the likelihood decreases and is quite low with subsequent doses. We provide correct and honest information, so that patients who experience an acute-phase reaction can make an informed decision about continuing bisphosphonate treatment or switching to another treatment. If the patient decides to continue with intravenous bisphosphonate treatment, we recommend more-aggressive prophylaxis with acetaminophen or NSAIDs with subsequent infusions.

For oral bisphosphonates

We do not prescribe prophylactic treatment with acetaminophen or NSAIDs with oral bisphosphonates, but we do advise patients to take acetaminophen or NSAIDs as needed for mild to moderate musculoskeletal pain, should this occur.

We try to continue treatment in mild to moderate cases, while monitoring the patient closely to see if the musculoskeletal pain resolves within 1 to 2 weeks.

If the pain is severe or does not resolve in 1 to 2 weeks, we offer switching to another drug class. Since musculoskeletal pain with oral bisphosphonates does not represent an allergic reaction, we have switched patients from oral to intravenous bisphosphonates without recurrence of musculoskeletal pain.

SEVERE MUSCULOSKELETAL PAIN BEYOND THE ACUTE PHASE

Severe musculoskeletal pain that may not be related to the acute-phase reaction, although rare, has been reported.^5,14 From 1995, when alendronate was approved for osteoporosis, through 2002, the US Food and Drug Administration received reports of severe musculoskeletal pain in 117 patients.¹⁵

This severe musculoskeletal pain related to bisphosphonate use remains poorly characterized. It has been reported to occur days or months (median time 14 days, range same day to 52 months) after starting bisphosphonate therapy and to resolve only if the bisphosphonate is stopped.^5,15 It differs from typical acute-phase reactions, which tend to occur with the initial dose (intravenous or oral) and resolve within several days. There are case reports of polyarthritis with synovitis that recurred with each bisphosphonate dose (oral or intravenous) and led to discontinuation of the bisphosphonate.^14,16–18

Clinicians need to be aware of the possibility of severe musculoskeletal pain and consider stopping bisphosphonate treatment in these cases. Besides discontinuation, acetaminophen, NSAIDs, and, in rare cases, glucocorticoids or short-term opiate therapy may be used for symptom control. In patients with a severe or persistent acute-phase reaction or musculoskeletal pain, discontinuation of bisphosphonates is warranted.

If a patient experiences diarrhea, hematochezia, or abdominal pain within the first 6 weeks of therapy with one of the anticancer drugs known as immune checkpoint inhibitors (ICIs), the first step is to rule out infection, especially with Clostridium difficile. The next step is colonosocopy with biopsy or computed tomography.

Patients with mild ICI-associated colitis may need only supportive care, and the ICI can be continued. In moderate or severe cases, the agent may need to be stopped and corticosteroids and other colitis-targeted agents may be needed. Figure 1 shows our algorithm for diagnosing and treating ICI-associated colitis.

POWERFUL ANTICANCER DRUGS

ICIs are monoclonal antibodies used in treating metastatic melanoma, non-small-cell lung cancer, metastatic prostate cancer, Hodgkin lymphoma, renal cell carcinoma, and other advanced malignancies.^1,2 They act by binding to and blocking proteins on T cells, antigen-presenting cells, and tumor cells that keep immune responses in check and prevent T cells from killing cancer cells.¹ For example:

• Ipilimumab blocks cytotoxic T lymphocyte-associated antigen 4
• Nivolumab and pembrolizumab block programmed cell death protein 1
• Atezolizumab blocks programmed death ligand 1.¹

With these proteins blocked, T cells can do their job, often producing dramatic regression of cancer. However, ICIs can cause a range of immune-related adverse effects, including endocrine and cutaneous toxicities, iridocyclitis, lymphadenopathy, neuropathy, nephritis, immune-mediated pneumonitis, pancreatitis, hepatitis, and colitis.³

ICI-ASSOCIATED COLITIS IS COMMON

ICI-associated colitis is common; it is estimated to affect about 30% of patients receiving ipilimumab, for example.⁴ Clinical presentations range from watery bowel movements, blood or mucus in the stool, abdominal cramping, and flatulence to ileus, colectasia, intestinal perforation, and even death.⁵

The incidence appears to increase with the dosage and duration of ICI therapy. The onset of colitis typically occurs 6 to 7 weeks after starting ipilimumab,⁶ and 6 to 18 weeks after starting nivolumab or pembrolizumab.⁷Table 1 lists the incidence of diarrhea and colitis and time of onset to colitis with common ICIs. However, colitis, like other immune-related adverse events, can occur at any point, even after ICI therapy has been discontinued.⁸

It is best to detect side effects of ICIs promptly, as acute inflammation can progress to chronic inflammation within 1 month of onset.⁹ We believe that early intervention and close monitoring may prevent complications and the need for long-term immunosuppressive treatment.

Patients, family members, and caregivers should be informed of possible gastrointestinal along with systemic side effects. Severe gastrointestinal symptoms such as increased stool frequency and change in stool consistency should trigger appropriate investigation and the withholding of ICI therapy.

COLITIS IS A SPECTRUM

The colon appears to be the gastrointestinal organ most affected by ICIs. Of patients with intestinal side effects, including diarrhea, only some develop colitis. The severity of ICI-associated colitis ranges from mild bowel illness to fulminant colitis.

Hodi et al,¹⁰ in a randomized trial in which 511 patients with melanoma received ipilimumab, reported that approximately 30% had mild diarrhea, while fewer than 10% had severe diarrhea, fever, ileus, or peritoneal signs. Five patients (1%) developed intestinal perforation, 4 (0.8%) died of complications, and 26 (5%) required hospitalization for severe enterocolitis.

The pathophysiology of ICI-mediated colitis is unclear. Most cases are diagnosed clinically.

Colitis is graded based on the Montreal classification system¹¹:

Mild colitis is defined as passage of fewer than 4 stools per day (with or without blood) over baseline and absence of any systemic illness.

Moderate is passage of more than 4 stools per day but with minimal signs of systemic toxicity.

Severe is defined as passage of at least 6 stools per day, heart rate at least 90 beats per minute, temperature at least 37.5°C (99.5°F), hemoglobin less than 10.5 g/dL, and erythrocyte sedimentation rate at least 30 mm/h.¹¹

RULE OUT INFECTION

If symptoms such as diarrhea or abdominal pain arise within 6 weeks of starting ICI therapy, then we should check for an infectious cause. The differential diagnosis of suspected ICI-associated colitis includes infections with C difficile, cytomegalovirus, opportunistic organisms, and other bacteria and viruses. ICI-induced celiac disease and immune hyper­thyroidism should also be ruled out.⁴

Once infection is ruled out, colonoscopy should be considered if symptoms persist or are severe. Colonoscopy with biopsy remains the gold standard for diagnosis, and it is also helpful in assessing severity of mucosal inflammation and monitoring response to medical treatment.

Table 2 lists common endoscopic and histologic features of ICI-mediated colitis; however, none of them is specific for this disease.

Common endoscopic features are loss of vascular pattern, edema, friability, spontaneous bleeding, and deep ulcerations.¹² A recent study suggested that colonic ulcerations predict a steroid-refractory course in patients with immune-mediated colitis.⁴

Histologically, ICI-associated colitis is characterized by both acute and chronic changes, including an increased number of neutrophils and lymphocytes in the epithelium and lamina propria, erosions, ulcers, crypt abscess, crypt apoptosis, crypt distortion, and even noncaseating granulomas.¹³ However, transmural disease is rare. Figure 2 compares the histopathologic features of ICI-associated colitis and a normal colon.

COMPUTED TOMOGRAPHY CAN BE USEFUL

Computed tomography (CT) can also be useful for the diagnosis and measurement of severity.

Garcia-Neuer et al1⁴ analyzed 303 patients with advanced melanoma who developed gastrointestinal symptoms while being treated with ipilimumab. Ninety-nine (33%) of them reported diarrhea during therapy, of whom 34 underwent both CT and colonoscopy with biopsy. CT was highly predictive of colitis on biopsy, with a positive predictive value of 96% and a negative likelihood ratio of 0.2.¹⁴

TREATMENT

Supportive care may be enough when treating mild ICI-related colitis. This can include oral and intravenous hydration⁴ and an antidiarrheal drug such as loperamide in a low dose.

Corticosteroids. For moderate ICI-associated colitis with stool frequency of 4 or more per day, patients should be started on an oral corticosteroid such as prednisone 0.5 to 1 mg/kg per day. If symptoms do not improve within 72 hours of starting an oral corticosteroid, the patient should be admitted to the hospital for observation and escalation to higher doses or possibly intravenous corticosteroids.

Infliximab has been used in severe and steroid-refractory cases,¹³ although there has been concern about using anti-tumor necrosis factor (TNF) agents such as this in patients with malignancies, especially melanoma. Since melanoma can be very aggressive and anti-TNF agents may promote it, it is prudent to try not to use this class of agents.

Other biologic agents such as vedolizu­mab, a gut-specific anti-integrin agent, are safer, have theoretic advantages over anti-TNF agents, and can be considered in patients with steroid-dependent or steroid-refractory ICI-associated enterocolitis. A recent study suggested that 2 to 4 infusions of vedolizumab are adequate to achieve steroid-free remission.¹⁵ Results from 6 clinical trials of vedolizumab in 2,830 patients with Crohn disease or ulcerative colitis did not show any increased risk of serious infections or malignancies over placebo.^16,17 A drawback is its slow onset of action.

Surgery is an option for patients with severe colitis refractory to intravenous corticosteroids or biological agents, as severe colitis carries a risk of significant morbidity and even death. The incidence of bowel perforation leading to colectomy or death in patients receiving ICI therapy is 0.5% to 1%.^18,19

Fecal microbiota transplant was associated with mucosal healing after 1 month in a case report of ICI-associated colitis.⁹

Follow-up. In most patients, symptoms resolve with discontinuation of the ICI and brief use of corticosteroids or biological agents. Patients with recurrent or persistent symptoms while on long-term ICI therapy may need periodic endoscopic evaluation, especially if there are chronic structural changes on histologic study.

If patients have recurrent or persistent symptoms along with chronic inflammatory structural changes on histology, a sign of an inflammatory bowel diseaselike condition, long-term maintenance therapy with an anti-inflammatory or immunosuppressant agent may be considered. However, there is no consensus on the treatment of this condition. It can be treated in the same way as classic inflammatory bowel disease in the setting of concurrent or prior history of malignancy, especially melanoma. Certain agents used in inflammatory bowel disease such as methotrexate and vedolizumab carry a lower risk of malignancy than anti-TNF agents and can be considered. A multidisciplinary approach that includes an oncologist, gastroenterologist, infectious disease specialist, and colorectal surgeon is imperative.

If a patient experiences diarrhea, hematochezia, or abdominal pain within the first 6 weeks of therapy with one of the anticancer drugs known as immune checkpoint inhibitors (ICIs), the first step is to rule out infection, especially with Clostridium difficile. The next step is colonosocopy with biopsy or computed tomography.

Patients with mild ICI-associated colitis may need only supportive care, and the ICI can be continued. In moderate or severe cases, the agent may need to be stopped and corticosteroids and other colitis-targeted agents may be needed. Figure 1 shows our algorithm for diagnosing and treating ICI-associated colitis.

POWERFUL ANTICANCER DRUGS

ICIs are monoclonal antibodies used in treating metastatic melanoma, non-small-cell lung cancer, metastatic prostate cancer, Hodgkin lymphoma, renal cell carcinoma, and other advanced malignancies.^1,2 They act by binding to and blocking proteins on T cells, antigen-presenting cells, and tumor cells that keep immune responses in check and prevent T cells from killing cancer cells.¹ For example:

• Ipilimumab blocks cytotoxic T lymphocyte-associated antigen 4
• Nivolumab and pembrolizumab block programmed cell death protein 1
• Atezolizumab blocks programmed death ligand 1.¹

With these proteins blocked, T cells can do their job, often producing dramatic regression of cancer. However, ICIs can cause a range of immune-related adverse effects, including endocrine and cutaneous toxicities, iridocyclitis, lymphadenopathy, neuropathy, nephritis, immune-mediated pneumonitis, pancreatitis, hepatitis, and colitis.³

ICI-ASSOCIATED COLITIS IS COMMON

ICI-associated colitis is common; it is estimated to affect about 30% of patients receiving ipilimumab, for example.⁴ Clinical presentations range from watery bowel movements, blood or mucus in the stool, abdominal cramping, and flatulence to ileus, colectasia, intestinal perforation, and even death.⁵

The incidence appears to increase with the dosage and duration of ICI therapy. The onset of colitis typically occurs 6 to 7 weeks after starting ipilimumab,⁶ and 6 to 18 weeks after starting nivolumab or pembrolizumab.⁷Table 1 lists the incidence of diarrhea and colitis and time of onset to colitis with common ICIs. However, colitis, like other immune-related adverse events, can occur at any point, even after ICI therapy has been discontinued.⁸

It is best to detect side effects of ICIs promptly, as acute inflammation can progress to chronic inflammation within 1 month of onset.⁹ We believe that early intervention and close monitoring may prevent complications and the need for long-term immunosuppressive treatment.

Patients, family members, and caregivers should be informed of possible gastrointestinal along with systemic side effects. Severe gastrointestinal symptoms such as increased stool frequency and change in stool consistency should trigger appropriate investigation and the withholding of ICI therapy.

COLITIS IS A SPECTRUM

The colon appears to be the gastrointestinal organ most affected by ICIs. Of patients with intestinal side effects, including diarrhea, only some develop colitis. The severity of ICI-associated colitis ranges from mild bowel illness to fulminant colitis.

Hodi et al,¹⁰ in a randomized trial in which 511 patients with melanoma received ipilimumab, reported that approximately 30% had mild diarrhea, while fewer than 10% had severe diarrhea, fever, ileus, or peritoneal signs. Five patients (1%) developed intestinal perforation, 4 (0.8%) died of complications, and 26 (5%) required hospitalization for severe enterocolitis.

The pathophysiology of ICI-mediated colitis is unclear. Most cases are diagnosed clinically.

Colitis is graded based on the Montreal classification system¹¹:

Mild colitis is defined as passage of fewer than 4 stools per day (with or without blood) over baseline and absence of any systemic illness.

Moderate is passage of more than 4 stools per day but with minimal signs of systemic toxicity.

Severe is defined as passage of at least 6 stools per day, heart rate at least 90 beats per minute, temperature at least 37.5°C (99.5°F), hemoglobin less than 10.5 g/dL, and erythrocyte sedimentation rate at least 30 mm/h.¹¹

RULE OUT INFECTION

If symptoms such as diarrhea or abdominal pain arise within 6 weeks of starting ICI therapy, then we should check for an infectious cause. The differential diagnosis of suspected ICI-associated colitis includes infections with C difficile, cytomegalovirus, opportunistic organisms, and other bacteria and viruses. ICI-induced celiac disease and immune hyper­thyroidism should also be ruled out.⁴

Once infection is ruled out, colonoscopy should be considered if symptoms persist or are severe. Colonoscopy with biopsy remains the gold standard for diagnosis, and it is also helpful in assessing severity of mucosal inflammation and monitoring response to medical treatment.

Table 2 lists common endoscopic and histologic features of ICI-mediated colitis; however, none of them is specific for this disease.

Common endoscopic features are loss of vascular pattern, edema, friability, spontaneous bleeding, and deep ulcerations.¹² A recent study suggested that colonic ulcerations predict a steroid-refractory course in patients with immune-mediated colitis.⁴

Histologically, ICI-associated colitis is characterized by both acute and chronic changes, including an increased number of neutrophils and lymphocytes in the epithelium and lamina propria, erosions, ulcers, crypt abscess, crypt apoptosis, crypt distortion, and even noncaseating granulomas.¹³ However, transmural disease is rare. Figure 2 compares the histopathologic features of ICI-associated colitis and a normal colon.

COMPUTED TOMOGRAPHY CAN BE USEFUL

Computed tomography (CT) can also be useful for the diagnosis and measurement of severity.

Garcia-Neuer et al1⁴ analyzed 303 patients with advanced melanoma who developed gastrointestinal symptoms while being treated with ipilimumab. Ninety-nine (33%) of them reported diarrhea during therapy, of whom 34 underwent both CT and colonoscopy with biopsy. CT was highly predictive of colitis on biopsy, with a positive predictive value of 96% and a negative likelihood ratio of 0.2.¹⁴

TREATMENT

Supportive care may be enough when treating mild ICI-related colitis. This can include oral and intravenous hydration⁴ and an antidiarrheal drug such as loperamide in a low dose.

Corticosteroids. For moderate ICI-associated colitis with stool frequency of 4 or more per day, patients should be started on an oral corticosteroid such as prednisone 0.5 to 1 mg/kg per day. If symptoms do not improve within 72 hours of starting an oral corticosteroid, the patient should be admitted to the hospital for observation and escalation to higher doses or possibly intravenous corticosteroids.

Infliximab has been used in severe and steroid-refractory cases,¹³ although there has been concern about using anti-tumor necrosis factor (TNF) agents such as this in patients with malignancies, especially melanoma. Since melanoma can be very aggressive and anti-TNF agents may promote it, it is prudent to try not to use this class of agents.

Other biologic agents such as vedolizu­mab, a gut-specific anti-integrin agent, are safer, have theoretic advantages over anti-TNF agents, and can be considered in patients with steroid-dependent or steroid-refractory ICI-associated enterocolitis. A recent study suggested that 2 to 4 infusions of vedolizumab are adequate to achieve steroid-free remission.¹⁵ Results from 6 clinical trials of vedolizumab in 2,830 patients with Crohn disease or ulcerative colitis did not show any increased risk of serious infections or malignancies over placebo.^16,17 A drawback is its slow onset of action.

Surgery is an option for patients with severe colitis refractory to intravenous corticosteroids or biological agents, as severe colitis carries a risk of significant morbidity and even death. The incidence of bowel perforation leading to colectomy or death in patients receiving ICI therapy is 0.5% to 1%.^18,19

Fecal microbiota transplant was associated with mucosal healing after 1 month in a case report of ICI-associated colitis.⁹

Follow-up. In most patients, symptoms resolve with discontinuation of the ICI and brief use of corticosteroids or biological agents. Patients with recurrent or persistent symptoms while on long-term ICI therapy may need periodic endoscopic evaluation, especially if there are chronic structural changes on histologic study.

If patients have recurrent or persistent symptoms along with chronic inflammatory structural changes on histology, a sign of an inflammatory bowel diseaselike condition, long-term maintenance therapy with an anti-inflammatory or immunosuppressant agent may be considered. However, there is no consensus on the treatment of this condition. It can be treated in the same way as classic inflammatory bowel disease in the setting of concurrent or prior history of malignancy, especially melanoma. Certain agents used in inflammatory bowel disease such as methotrexate and vedolizumab carry a lower risk of malignancy than anti-TNF agents and can be considered. A multidisciplinary approach that includes an oncologist, gastroenterologist, infectious disease specialist, and colorectal surgeon is imperative.

If a patient experiences diarrhea, hematochezia, or abdominal pain within the first 6 weeks of therapy with one of the anticancer drugs known as immune checkpoint inhibitors (ICIs), the first step is to rule out infection, especially with Clostridium difficile. The next step is colonosocopy with biopsy or computed tomography.

Patients with mild ICI-associated colitis may need only supportive care, and the ICI can be continued. In moderate or severe cases, the agent may need to be stopped and corticosteroids and other colitis-targeted agents may be needed. Figure 1 shows our algorithm for diagnosing and treating ICI-associated colitis.

POWERFUL ANTICANCER DRUGS

ICIs are monoclonal antibodies used in treating metastatic melanoma, non-small-cell lung cancer, metastatic prostate cancer, Hodgkin lymphoma, renal cell carcinoma, and other advanced malignancies.^1,2 They act by binding to and blocking proteins on T cells, antigen-presenting cells, and tumor cells that keep immune responses in check and prevent T cells from killing cancer cells.¹ For example:

• Ipilimumab blocks cytotoxic T lymphocyte-associated antigen 4
• Nivolumab and pembrolizumab block programmed cell death protein 1
• Atezolizumab blocks programmed death ligand 1.¹

With these proteins blocked, T cells can do their job, often producing dramatic regression of cancer. However, ICIs can cause a range of immune-related adverse effects, including endocrine and cutaneous toxicities, iridocyclitis, lymphadenopathy, neuropathy, nephritis, immune-mediated pneumonitis, pancreatitis, hepatitis, and colitis.³

ICI-ASSOCIATED COLITIS IS COMMON

ICI-associated colitis is common; it is estimated to affect about 30% of patients receiving ipilimumab, for example.⁴ Clinical presentations range from watery bowel movements, blood or mucus in the stool, abdominal cramping, and flatulence to ileus, colectasia, intestinal perforation, and even death.⁵

The incidence appears to increase with the dosage and duration of ICI therapy. The onset of colitis typically occurs 6 to 7 weeks after starting ipilimumab,⁶ and 6 to 18 weeks after starting nivolumab or pembrolizumab.⁷Table 1 lists the incidence of diarrhea and colitis and time of onset to colitis with common ICIs. However, colitis, like other immune-related adverse events, can occur at any point, even after ICI therapy has been discontinued.⁸

It is best to detect side effects of ICIs promptly, as acute inflammation can progress to chronic inflammation within 1 month of onset.⁹ We believe that early intervention and close monitoring may prevent complications and the need for long-term immunosuppressive treatment.

Patients, family members, and caregivers should be informed of possible gastrointestinal along with systemic side effects. Severe gastrointestinal symptoms such as increased stool frequency and change in stool consistency should trigger appropriate investigation and the withholding of ICI therapy.

COLITIS IS A SPECTRUM

The colon appears to be the gastrointestinal organ most affected by ICIs. Of patients with intestinal side effects, including diarrhea, only some develop colitis. The severity of ICI-associated colitis ranges from mild bowel illness to fulminant colitis.

Hodi et al,¹⁰ in a randomized trial in which 511 patients with melanoma received ipilimumab, reported that approximately 30% had mild diarrhea, while fewer than 10% had severe diarrhea, fever, ileus, or peritoneal signs. Five patients (1%) developed intestinal perforation, 4 (0.8%) died of complications, and 26 (5%) required hospitalization for severe enterocolitis.

The pathophysiology of ICI-mediated colitis is unclear. Most cases are diagnosed clinically.

Colitis is graded based on the Montreal classification system¹¹:

Mild colitis is defined as passage of fewer than 4 stools per day (with or without blood) over baseline and absence of any systemic illness.

Moderate is passage of more than 4 stools per day but with minimal signs of systemic toxicity.

Severe is defined as passage of at least 6 stools per day, heart rate at least 90 beats per minute, temperature at least 37.5°C (99.5°F), hemoglobin less than 10.5 g/dL, and erythrocyte sedimentation rate at least 30 mm/h.¹¹

RULE OUT INFECTION

If symptoms such as diarrhea or abdominal pain arise within 6 weeks of starting ICI therapy, then we should check for an infectious cause. The differential diagnosis of suspected ICI-associated colitis includes infections with C difficile, cytomegalovirus, opportunistic organisms, and other bacteria and viruses. ICI-induced celiac disease and immune hyper­thyroidism should also be ruled out.⁴

Once infection is ruled out, colonoscopy should be considered if symptoms persist or are severe. Colonoscopy with biopsy remains the gold standard for diagnosis, and it is also helpful in assessing severity of mucosal inflammation and monitoring response to medical treatment.

Table 2 lists common endoscopic and histologic features of ICI-mediated colitis; however, none of them is specific for this disease.

Common endoscopic features are loss of vascular pattern, edema, friability, spontaneous bleeding, and deep ulcerations.¹² A recent study suggested that colonic ulcerations predict a steroid-refractory course in patients with immune-mediated colitis.⁴

Histologically, ICI-associated colitis is characterized by both acute and chronic changes, including an increased number of neutrophils and lymphocytes in the epithelium and lamina propria, erosions, ulcers, crypt abscess, crypt apoptosis, crypt distortion, and even noncaseating granulomas.¹³ However, transmural disease is rare. Figure 2 compares the histopathologic features of ICI-associated colitis and a normal colon.

COMPUTED TOMOGRAPHY CAN BE USEFUL

Computed tomography (CT) can also be useful for the diagnosis and measurement of severity.

Garcia-Neuer et al1⁴ analyzed 303 patients with advanced melanoma who developed gastrointestinal symptoms while being treated with ipilimumab. Ninety-nine (33%) of them reported diarrhea during therapy, of whom 34 underwent both CT and colonoscopy with biopsy. CT was highly predictive of colitis on biopsy, with a positive predictive value of 96% and a negative likelihood ratio of 0.2.¹⁴

TREATMENT

Supportive care may be enough when treating mild ICI-related colitis. This can include oral and intravenous hydration⁴ and an antidiarrheal drug such as loperamide in a low dose.

Corticosteroids. For moderate ICI-associated colitis with stool frequency of 4 or more per day, patients should be started on an oral corticosteroid such as prednisone 0.5 to 1 mg/kg per day. If symptoms do not improve within 72 hours of starting an oral corticosteroid, the patient should be admitted to the hospital for observation and escalation to higher doses or possibly intravenous corticosteroids.

Infliximab has been used in severe and steroid-refractory cases,¹³ although there has been concern about using anti-tumor necrosis factor (TNF) agents such as this in patients with malignancies, especially melanoma. Since melanoma can be very aggressive and anti-TNF agents may promote it, it is prudent to try not to use this class of agents.

Other biologic agents such as vedolizu­mab, a gut-specific anti-integrin agent, are safer, have theoretic advantages over anti-TNF agents, and can be considered in patients with steroid-dependent or steroid-refractory ICI-associated enterocolitis. A recent study suggested that 2 to 4 infusions of vedolizumab are adequate to achieve steroid-free remission.¹⁵ Results from 6 clinical trials of vedolizumab in 2,830 patients with Crohn disease or ulcerative colitis did not show any increased risk of serious infections or malignancies over placebo.^16,17 A drawback is its slow onset of action.

Surgery is an option for patients with severe colitis refractory to intravenous corticosteroids or biological agents, as severe colitis carries a risk of significant morbidity and even death. The incidence of bowel perforation leading to colectomy or death in patients receiving ICI therapy is 0.5% to 1%.^18,19

Fecal microbiota transplant was associated with mucosal healing after 1 month in a case report of ICI-associated colitis.⁹

Follow-up. In most patients, symptoms resolve with discontinuation of the ICI and brief use of corticosteroids or biological agents. Patients with recurrent or persistent symptoms while on long-term ICI therapy may need periodic endoscopic evaluation, especially if there are chronic structural changes on histologic study.

If patients have recurrent or persistent symptoms along with chronic inflammatory structural changes on histology, a sign of an inflammatory bowel diseaselike condition, long-term maintenance therapy with an anti-inflammatory or immunosuppressant agent may be considered. However, there is no consensus on the treatment of this condition. It can be treated in the same way as classic inflammatory bowel disease in the setting of concurrent or prior history of malignancy, especially melanoma. Certain agents used in inflammatory bowel disease such as methotrexate and vedolizumab carry a lower risk of malignancy than anti-TNF agents and can be considered. A multidisciplinary approach that includes an oncologist, gastroenterologist, infectious disease specialist, and colorectal surgeon is imperative.

An 88-year-old man with a 1-day history of fever and altered mental status was transferred to the emergency department. He had been receiving conservative management for low-risk localized prostate cancer but had no previous cardiovascular or gastrointestinal problems.

Based on these findings, the diagnosis was Fournier gangrene. Despite aggressive treatment, the patient’s condition deteriorated rapidly, and he died 2 hours after admission.

FOURNIER GANGRENE: NECROTIZING FASCIITIS OF THE PERINEUM

Fournier gangrene is a rare but rapidly progressive necrotizing fasciitis of the perineum with a high death rate.

Predisposing factors for Fournier gangrene include older age, diabetes mellitus, morbid obesity, cardiovascular disorders, chronic alcoholism, long-term corticosteroid treatment, malignancy, and human immunodeficiency virus infection.^1,2 Urethral obstruction, instrumentation, urinary extravasation, and trauma have also been associated with this condition.³

In general, organisms from the urinary tract spread along the fascial planes to involve the penis and scrotum.

The differential diagnosis of Fournier gangrene includes scrotal and perineal disorders, as well as intra-abdominal disorders such as cellulitis, abscess, strangulated hernia, pyoderma gangrenosum, allergic vasculitis, vascular occlusion syndromes, and warfarin necrosis.

Delay in the diagnosis of Fournier gangrene leads to an extremely high death rate due to rapid progression of the disease, leading to sepsis, multiple organ failure, and disseminated intravascular coagulation. Immediate diagnosis and appropriate treatment such as broad-spectrum antibiotics and extensive surgical debridement reduce morbidity and control the infection. Antibiotics for methicillin-resistant Staphylococcus aureus should be considered if there is a history of or risk factors for this organism.⁴

Necrotizing fasciitis, including Fournier gangrene, is a common indication for intravenous immunoglobulin, and this treatment has been reported to be effective in a few cases. However, a double-blind, placebo-controlled trial that evaluated the benefit of this treatment was terminated early due to slow patient recruitment.⁵

A delay of even a few hours from suspicion of Fournier gangrene to surgical debridement significantly increases the risk of death.⁶ Thus, when it is suspected, immediate surgical intervention may be necessary to confirm the diagnosis and to treat it. The usual combination of antibiotic therapy for Fournier gangrene includes penicillin for the streptococcal species, a third-generation cephalosporin with or without an aminoglycoside for the gram-negative organisms, and metronidazole for anaerobic bacteria.

An 88-year-old man with a 1-day history of fever and altered mental status was transferred to the emergency department. He had been receiving conservative management for low-risk localized prostate cancer but had no previous cardiovascular or gastrointestinal problems.

Based on these findings, the diagnosis was Fournier gangrene. Despite aggressive treatment, the patient’s condition deteriorated rapidly, and he died 2 hours after admission.

FOURNIER GANGRENE: NECROTIZING FASCIITIS OF THE PERINEUM

Fournier gangrene is a rare but rapidly progressive necrotizing fasciitis of the perineum with a high death rate.

Predisposing factors for Fournier gangrene include older age, diabetes mellitus, morbid obesity, cardiovascular disorders, chronic alcoholism, long-term corticosteroid treatment, malignancy, and human immunodeficiency virus infection.^1,2 Urethral obstruction, instrumentation, urinary extravasation, and trauma have also been associated with this condition.³

In general, organisms from the urinary tract spread along the fascial planes to involve the penis and scrotum.

The differential diagnosis of Fournier gangrene includes scrotal and perineal disorders, as well as intra-abdominal disorders such as cellulitis, abscess, strangulated hernia, pyoderma gangrenosum, allergic vasculitis, vascular occlusion syndromes, and warfarin necrosis.

Delay in the diagnosis of Fournier gangrene leads to an extremely high death rate due to rapid progression of the disease, leading to sepsis, multiple organ failure, and disseminated intravascular coagulation. Immediate diagnosis and appropriate treatment such as broad-spectrum antibiotics and extensive surgical debridement reduce morbidity and control the infection. Antibiotics for methicillin-resistant Staphylococcus aureus should be considered if there is a history of or risk factors for this organism.⁴

Necrotizing fasciitis, including Fournier gangrene, is a common indication for intravenous immunoglobulin, and this treatment has been reported to be effective in a few cases. However, a double-blind, placebo-controlled trial that evaluated the benefit of this treatment was terminated early due to slow patient recruitment.⁵

A delay of even a few hours from suspicion of Fournier gangrene to surgical debridement significantly increases the risk of death.⁶ Thus, when it is suspected, immediate surgical intervention may be necessary to confirm the diagnosis and to treat it. The usual combination of antibiotic therapy for Fournier gangrene includes penicillin for the streptococcal species, a third-generation cephalosporin with or without an aminoglycoside for the gram-negative organisms, and metronidazole for anaerobic bacteria.

An 88-year-old man with a 1-day history of fever and altered mental status was transferred to the emergency department. He had been receiving conservative management for low-risk localized prostate cancer but had no previous cardiovascular or gastrointestinal problems.

Based on these findings, the diagnosis was Fournier gangrene. Despite aggressive treatment, the patient’s condition deteriorated rapidly, and he died 2 hours after admission.

FOURNIER GANGRENE: NECROTIZING FASCIITIS OF THE PERINEUM

Fournier gangrene is a rare but rapidly progressive necrotizing fasciitis of the perineum with a high death rate.

Predisposing factors for Fournier gangrene include older age, diabetes mellitus, morbid obesity, cardiovascular disorders, chronic alcoholism, long-term corticosteroid treatment, malignancy, and human immunodeficiency virus infection.^1,2 Urethral obstruction, instrumentation, urinary extravasation, and trauma have also been associated with this condition.³

In general, organisms from the urinary tract spread along the fascial planes to involve the penis and scrotum.

The differential diagnosis of Fournier gangrene includes scrotal and perineal disorders, as well as intra-abdominal disorders such as cellulitis, abscess, strangulated hernia, pyoderma gangrenosum, allergic vasculitis, vascular occlusion syndromes, and warfarin necrosis.

Delay in the diagnosis of Fournier gangrene leads to an extremely high death rate due to rapid progression of the disease, leading to sepsis, multiple organ failure, and disseminated intravascular coagulation. Immediate diagnosis and appropriate treatment such as broad-spectrum antibiotics and extensive surgical debridement reduce morbidity and control the infection. Antibiotics for methicillin-resistant Staphylococcus aureus should be considered if there is a history of or risk factors for this organism.⁴

Necrotizing fasciitis, including Fournier gangrene, is a common indication for intravenous immunoglobulin, and this treatment has been reported to be effective in a few cases. However, a double-blind, placebo-controlled trial that evaluated the benefit of this treatment was terminated early due to slow patient recruitment.⁵

A delay of even a few hours from suspicion of Fournier gangrene to surgical debridement significantly increases the risk of death.⁶ Thus, when it is suspected, immediate surgical intervention may be necessary to confirm the diagnosis and to treat it. The usual combination of antibiotic therapy for Fournier gangrene includes penicillin for the streptococcal species, a third-generation cephalosporin with or without an aminoglycoside for the gram-negative organisms, and metronidazole for anaerobic bacteria.

A 60-year-old man with hypertension and persistent atrial fibrillation refractory to radiofrequency ablation was brought to the hospital in status epilepticus requiring intubation. His wife said that during the past month he had experienced a number of episodic seizures, but due to his busy work schedule he had not sought medical attention. He had also been hospitalized 3 times during the past week for chills, tremors, and fevers with temperatures up to 101°F (38.3°C), and his symptoms had been ascribed to the amiodarone he had been taking for the past 11 days for atrial fibrillation. The amiodarone dose had been decreased to half a tablet after the first 7 days, but his symptoms had continued.

When the patient was able to speak, he denied intravenous drug abuse and claimed to be up to date with vaccinations. Colonoscopy 10 years earlier had been negative. He has no pets, but says that there are stray cats around his home and that he has had contact with cat feces while gardening. He works as a diesel mechanic and is exposed to motor oil and diesel fuel, but denies any direct exposure to carcinogenic chemicals.

On admission, his temperature was 37.7°C (99.9°F), blood pressure 92/69 mm Hg, heart rate 96 beats per minute, respiratory rate 21 per minute, and oxygen saturation 95% on room air and 100% on oxygen at 2 L per minute.

Decerebrate posturing and forced left visual gaze deviation was observed. Oral examination revealed severe decay of multiple teeth, with some teeth broken down to the level of the gingiva, and moderate generalized periodontal disease with heavy plaque and calculi in the gingiva.

The differential diagnosis for intracranial ring-enhancing lesions includes metastasis, abscess, infection in an immunocompromised state (eg, toxoplasmosis), glioblastoma, subacute infarct, neurocysticercosis, lymphoma, demyelination, and resolving hematoma. In our patient, further testing to narrow the differential included lumbar puncture, with results within normal limits, and transthoracic echocardiography, which was negative for endocarditis. A biopsy obtained by craniotomy confirmed the diagnosis of abscess surrounded by reactive glioses.

During his hospitalization, the patient’s antiseizure regimen was lorazepam 1 to 2 mg as needed, levetiracetam 1,500 mg twice daily, and fosphenytoin infusion at 100 mg phenytoin sodium equivalents per minute. Initial antibiotic therapy included ampicillin 2 g intravenously (IV) 4 times daily.

Because of persistent nocturnal fevers with temperatures ranging from 37.8°C (100°F) to 41.2°C (106.2°F), antibiotic coverage was broadened to meropenem 2 g IV every 8 hours. Testing for Toxoplasma gondii, human immunodeficiency virus, and JC polyomavirus was negative. Cerebrospinal fluid culture and abscess cultures were also negative. Blood cultures were eventually positive for Peptostreptococcus micros and Streptococcus constellatus. Based on review of culture results, antibiotic therapy was switched to ceftriaxone 2 g IV twice daily and metronidazole 500 mg IV 3 times daily.

For the dental infection, the patient underwent surgical irrigation and debridement with full dental extraction for multiple dental abscesses.

His regimen for seizure control was changed to phenytoin and valproic acid, and he was discharged in stable condition on the following drug regimen: ceftriaxone 2 g IV twice daily, metronidazole 500 mg IV 3 times daily for 6 weeks, levetiracetam 1,500 mg twice daily, and valproic acid 750 mg 3 times daily.

At a 3-month follow-up visit, he reported no seizure-like activity but demonstrated persistent neurologic deficits (dysdiadocho­kinesia and mild ataxia).

A LESS COMMON CAUSE OF BRAIN ABSCESS

In the United States, 1,500 to 2,000 cases of brain abscess are diagnosed every year, and this condition is responsible for an estimated 1 in 10,000 hospitalizations. Most patients hospitalized are men over age 60 or children. Most patients with hematogenous or embolic spread of infection from a primary infection source are immunocompromised.

However, the lesions in our patient were not from compromised immunity, but rather from septic hematogenous spread of an odontogenic infection. Odontogenic bacteria are a common cause of pyogenic orofascial infection, including periapical abscess and infection of adjoining fascial spaces of the head and neck.¹

P micros and S constellatus have been commonly found in many types of odontogenic infection, including dentoalveolar infection, periodontitis, and pericoronitis.² Our patient was found to have several periodontal abscesses with bacteremia and spread to the brain. Although transthoracic echocardiography was negative for vegetations or patent foramen ovale, the quality and location of the brain abscesses suggested embolic spread of infection. Most of the suspected septic emboli were in the right hemisphere, consistent with patterns seen with cardioembolic phenomena, and a number of lesions appeared to be within the distribution of the right anterior cerebral artery and the middle cerebral artery.

Empiric antibiotic therapy for local odontogenic infection includes amoxicillin with clavulanic acid and metronidazole.¹ Our patient’s treatment with ceftriaxone and metronidazole was based on the species and sensitivities of the bacteria in blood cultures.

Surgical irrigation with debridement is considered first-line therapy for local dental infection, with antimicrobials as adjunctive therapy. Initiation of antibiotic therapy before surgery has been associated with a shortened duration of infection and a reduced risk of bacteremia.³

First-line therapy for cerebral abscess is typically antibiotics, specifically ceftriaxone and metronidazole as in our patient. Ceftriaxone is selected for coverage against streptococci, enterobacteriacae, and most common anaerobes, whereas metronidazole is chosen for its efficacy against Bacteroides fragilis.

Computed tomography-guided stereotactic aspiration and open drainage are viable options for solitary and surgically accessible abscesses—typically those greater than 2 cm. Our patient had multiple small septic emboli in the right hemisphere, with the largest lesion measuring 1.5 cm, thus limiting the effectiveness of surgical intervention.

Some patients with mass effect or other evidence of increased intracranial pressure may benefit from high doses of a corticosteroid such as dexamethasone. However, since our patient had no clinical or diagnostic findings suggesting elevated intracranial pressure, we opted for nonsurgical management of the brain abscesses, with 6 weeks of intravenous antibiotics, an antiseizure regimen, and plans for repeat imaging in the outpatient setting.

A 60-year-old man with hypertension and persistent atrial fibrillation refractory to radiofrequency ablation was brought to the hospital in status epilepticus requiring intubation. His wife said that during the past month he had experienced a number of episodic seizures, but due to his busy work schedule he had not sought medical attention. He had also been hospitalized 3 times during the past week for chills, tremors, and fevers with temperatures up to 101°F (38.3°C), and his symptoms had been ascribed to the amiodarone he had been taking for the past 11 days for atrial fibrillation. The amiodarone dose had been decreased to half a tablet after the first 7 days, but his symptoms had continued.

When the patient was able to speak, he denied intravenous drug abuse and claimed to be up to date with vaccinations. Colonoscopy 10 years earlier had been negative. He has no pets, but says that there are stray cats around his home and that he has had contact with cat feces while gardening. He works as a diesel mechanic and is exposed to motor oil and diesel fuel, but denies any direct exposure to carcinogenic chemicals.

On admission, his temperature was 37.7°C (99.9°F), blood pressure 92/69 mm Hg, heart rate 96 beats per minute, respiratory rate 21 per minute, and oxygen saturation 95% on room air and 100% on oxygen at 2 L per minute.

Decerebrate posturing and forced left visual gaze deviation was observed. Oral examination revealed severe decay of multiple teeth, with some teeth broken down to the level of the gingiva, and moderate generalized periodontal disease with heavy plaque and calculi in the gingiva.

The differential diagnosis for intracranial ring-enhancing lesions includes metastasis, abscess, infection in an immunocompromised state (eg, toxoplasmosis), glioblastoma, subacute infarct, neurocysticercosis, lymphoma, demyelination, and resolving hematoma. In our patient, further testing to narrow the differential included lumbar puncture, with results within normal limits, and transthoracic echocardiography, which was negative for endocarditis. A biopsy obtained by craniotomy confirmed the diagnosis of abscess surrounded by reactive glioses.

During his hospitalization, the patient’s antiseizure regimen was lorazepam 1 to 2 mg as needed, levetiracetam 1,500 mg twice daily, and fosphenytoin infusion at 100 mg phenytoin sodium equivalents per minute. Initial antibiotic therapy included ampicillin 2 g intravenously (IV) 4 times daily.

Because of persistent nocturnal fevers with temperatures ranging from 37.8°C (100°F) to 41.2°C (106.2°F), antibiotic coverage was broadened to meropenem 2 g IV every 8 hours. Testing for Toxoplasma gondii, human immunodeficiency virus, and JC polyomavirus was negative. Cerebrospinal fluid culture and abscess cultures were also negative. Blood cultures were eventually positive for Peptostreptococcus micros and Streptococcus constellatus. Based on review of culture results, antibiotic therapy was switched to ceftriaxone 2 g IV twice daily and metronidazole 500 mg IV 3 times daily.

For the dental infection, the patient underwent surgical irrigation and debridement with full dental extraction for multiple dental abscesses.

His regimen for seizure control was changed to phenytoin and valproic acid, and he was discharged in stable condition on the following drug regimen: ceftriaxone 2 g IV twice daily, metronidazole 500 mg IV 3 times daily for 6 weeks, levetiracetam 1,500 mg twice daily, and valproic acid 750 mg 3 times daily.

At a 3-month follow-up visit, he reported no seizure-like activity but demonstrated persistent neurologic deficits (dysdiadocho­kinesia and mild ataxia).

A LESS COMMON CAUSE OF BRAIN ABSCESS

In the United States, 1,500 to 2,000 cases of brain abscess are diagnosed every year, and this condition is responsible for an estimated 1 in 10,000 hospitalizations. Most patients hospitalized are men over age 60 or children. Most patients with hematogenous or embolic spread of infection from a primary infection source are immunocompromised.

However, the lesions in our patient were not from compromised immunity, but rather from septic hematogenous spread of an odontogenic infection. Odontogenic bacteria are a common cause of pyogenic orofascial infection, including periapical abscess and infection of adjoining fascial spaces of the head and neck.¹

P micros and S constellatus have been commonly found in many types of odontogenic infection, including dentoalveolar infection, periodontitis, and pericoronitis.² Our patient was found to have several periodontal abscesses with bacteremia and spread to the brain. Although transthoracic echocardiography was negative for vegetations or patent foramen ovale, the quality and location of the brain abscesses suggested embolic spread of infection. Most of the suspected septic emboli were in the right hemisphere, consistent with patterns seen with cardioembolic phenomena, and a number of lesions appeared to be within the distribution of the right anterior cerebral artery and the middle cerebral artery.

Empiric antibiotic therapy for local odontogenic infection includes amoxicillin with clavulanic acid and metronidazole.¹ Our patient’s treatment with ceftriaxone and metronidazole was based on the species and sensitivities of the bacteria in blood cultures.

Surgical irrigation with debridement is considered first-line therapy for local dental infection, with antimicrobials as adjunctive therapy. Initiation of antibiotic therapy before surgery has been associated with a shortened duration of infection and a reduced risk of bacteremia.³

First-line therapy for cerebral abscess is typically antibiotics, specifically ceftriaxone and metronidazole as in our patient. Ceftriaxone is selected for coverage against streptococci, enterobacteriacae, and most common anaerobes, whereas metronidazole is chosen for its efficacy against Bacteroides fragilis.

Computed tomography-guided stereotactic aspiration and open drainage are viable options for solitary and surgically accessible abscesses—typically those greater than 2 cm. Our patient had multiple small septic emboli in the right hemisphere, with the largest lesion measuring 1.5 cm, thus limiting the effectiveness of surgical intervention.

Some patients with mass effect or other evidence of increased intracranial pressure may benefit from high doses of a corticosteroid such as dexamethasone. However, since our patient had no clinical or diagnostic findings suggesting elevated intracranial pressure, we opted for nonsurgical management of the brain abscesses, with 6 weeks of intravenous antibiotics, an antiseizure regimen, and plans for repeat imaging in the outpatient setting.

A 60-year-old man with hypertension and persistent atrial fibrillation refractory to radiofrequency ablation was brought to the hospital in status epilepticus requiring intubation. His wife said that during the past month he had experienced a number of episodic seizures, but due to his busy work schedule he had not sought medical attention. He had also been hospitalized 3 times during the past week for chills, tremors, and fevers with temperatures up to 101°F (38.3°C), and his symptoms had been ascribed to the amiodarone he had been taking for the past 11 days for atrial fibrillation. The amiodarone dose had been decreased to half a tablet after the first 7 days, but his symptoms had continued.

When the patient was able to speak, he denied intravenous drug abuse and claimed to be up to date with vaccinations. Colonoscopy 10 years earlier had been negative. He has no pets, but says that there are stray cats around his home and that he has had contact with cat feces while gardening. He works as a diesel mechanic and is exposed to motor oil and diesel fuel, but denies any direct exposure to carcinogenic chemicals.

On admission, his temperature was 37.7°C (99.9°F), blood pressure 92/69 mm Hg, heart rate 96 beats per minute, respiratory rate 21 per minute, and oxygen saturation 95% on room air and 100% on oxygen at 2 L per minute.

Decerebrate posturing and forced left visual gaze deviation was observed. Oral examination revealed severe decay of multiple teeth, with some teeth broken down to the level of the gingiva, and moderate generalized periodontal disease with heavy plaque and calculi in the gingiva.

The differential diagnosis for intracranial ring-enhancing lesions includes metastasis, abscess, infection in an immunocompromised state (eg, toxoplasmosis), glioblastoma, subacute infarct, neurocysticercosis, lymphoma, demyelination, and resolving hematoma. In our patient, further testing to narrow the differential included lumbar puncture, with results within normal limits, and transthoracic echocardiography, which was negative for endocarditis. A biopsy obtained by craniotomy confirmed the diagnosis of abscess surrounded by reactive glioses.

During his hospitalization, the patient’s antiseizure regimen was lorazepam 1 to 2 mg as needed, levetiracetam 1,500 mg twice daily, and fosphenytoin infusion at 100 mg phenytoin sodium equivalents per minute. Initial antibiotic therapy included ampicillin 2 g intravenously (IV) 4 times daily.

Because of persistent nocturnal fevers with temperatures ranging from 37.8°C (100°F) to 41.2°C (106.2°F), antibiotic coverage was broadened to meropenem 2 g IV every 8 hours. Testing for Toxoplasma gondii, human immunodeficiency virus, and JC polyomavirus was negative. Cerebrospinal fluid culture and abscess cultures were also negative. Blood cultures were eventually positive for Peptostreptococcus micros and Streptococcus constellatus. Based on review of culture results, antibiotic therapy was switched to ceftriaxone 2 g IV twice daily and metronidazole 500 mg IV 3 times daily.

For the dental infection, the patient underwent surgical irrigation and debridement with full dental extraction for multiple dental abscesses.

His regimen for seizure control was changed to phenytoin and valproic acid, and he was discharged in stable condition on the following drug regimen: ceftriaxone 2 g IV twice daily, metronidazole 500 mg IV 3 times daily for 6 weeks, levetiracetam 1,500 mg twice daily, and valproic acid 750 mg 3 times daily.

At a 3-month follow-up visit, he reported no seizure-like activity but demonstrated persistent neurologic deficits (dysdiadocho­kinesia and mild ataxia).

A LESS COMMON CAUSE OF BRAIN ABSCESS

In the United States, 1,500 to 2,000 cases of brain abscess are diagnosed every year, and this condition is responsible for an estimated 1 in 10,000 hospitalizations. Most patients hospitalized are men over age 60 or children. Most patients with hematogenous or embolic spread of infection from a primary infection source are immunocompromised.

However, the lesions in our patient were not from compromised immunity, but rather from septic hematogenous spread of an odontogenic infection. Odontogenic bacteria are a common cause of pyogenic orofascial infection, including periapical abscess and infection of adjoining fascial spaces of the head and neck.¹

P micros and S constellatus have been commonly found in many types of odontogenic infection, including dentoalveolar infection, periodontitis, and pericoronitis.² Our patient was found to have several periodontal abscesses with bacteremia and spread to the brain. Although transthoracic echocardiography was negative for vegetations or patent foramen ovale, the quality and location of the brain abscesses suggested embolic spread of infection. Most of the suspected septic emboli were in the right hemisphere, consistent with patterns seen with cardioembolic phenomena, and a number of lesions appeared to be within the distribution of the right anterior cerebral artery and the middle cerebral artery.

Empiric antibiotic therapy for local odontogenic infection includes amoxicillin with clavulanic acid and metronidazole.¹ Our patient’s treatment with ceftriaxone and metronidazole was based on the species and sensitivities of the bacteria in blood cultures.

Surgical irrigation with debridement is considered first-line therapy for local dental infection, with antimicrobials as adjunctive therapy. Initiation of antibiotic therapy before surgery has been associated with a shortened duration of infection and a reduced risk of bacteremia.³

First-line therapy for cerebral abscess is typically antibiotics, specifically ceftriaxone and metronidazole as in our patient. Ceftriaxone is selected for coverage against streptococci, enterobacteriacae, and most common anaerobes, whereas metronidazole is chosen for its efficacy against Bacteroides fragilis.

Computed tomography-guided stereotactic aspiration and open drainage are viable options for solitary and surgically accessible abscesses—typically those greater than 2 cm. Our patient had multiple small septic emboli in the right hemisphere, with the largest lesion measuring 1.5 cm, thus limiting the effectiveness of surgical intervention.

Some patients with mass effect or other evidence of increased intracranial pressure may benefit from high doses of a corticosteroid such as dexamethasone. However, since our patient had no clinical or diagnostic findings suggesting elevated intracranial pressure, we opted for nonsurgical management of the brain abscesses, with 6 weeks of intravenous antibiotics, an antiseizure regimen, and plans for repeat imaging in the outpatient setting.

Some drugs exhibit a dose-toxicity relationship that is sufficiently predictable to permit drug-level monitoring to limit renal or other toxicity. Others cause ocular or marrow toxicity that can be limited by weight-based dosing and careful monitoring. With azathioprine, some toxicity can be predicted by assessing the activity of the enzyme thiopurine methyltransferase, which metabolizes the drug. Other approaches to using pharmacogenomics have included HLA-B locus haplotyping to detect increased risk of immune-mediated toxicities of drugs such as carbamazepine and allopurinol. Both of these drugs exhibit serious systemic toxicities that are incompletely understood, but these are nascent and significant steps toward providing personalized (precision) medical care.

Adverse effects of some drugs may result from their intracellular effects, which are only partially predictable by drug levels or dosing. Colchicine, hydroxychloroquine, and amiodarone all affect intracellular vacuolar transport and lysosomal processing. Yet, although the footprints of drug effect can be seen in many histopathology samples, only some patients—but maybe more than currently recognized—suffer cardiac and skeletal muscle vacuolar myopathy, axonal neuropathies, or pulmonary or retinal cell toxicity from these drugs. But distinguishing the histopathologic footprints of drug exposure and the biologic effect from true drug toxicity with organ damage is not always straightforward.

Rare adverse effects may only become apparent with frequent use of a drug in the general community. These often remain mechanistically unexplained: Why can minoxidil cause pericardial effusions or a nonsteroidal anti-inflammatory drug cause aseptic meningitis? Some effects may be due to altering of the unique balance of biochemical pathways in a given patient, leading to unexpected drug metabolism with generation of toxic metabolites.

More interesting to me are effects that are seemingly off-target biologic outcomes caused by disrupting normal physiologic homeostasis and stimulating counterregulatory pathways in such a way that unexpected biologic effects occur. Angioedema and cough in some patients who have taken angiotensin-converting enzyme inhibitors are examples, but why the disturbed control mechanisms lead to these effects in only occasional patients is still incompletely elucidated.

Two additional classes of drugs with unique systemic adverse effects are discussed in this issue of the Journal. The “flulike” syndrome after bisphosphonate treatment, presumably resulting from selective cytokine release by macrophages that have ingested certain bisphosphonates, is a not uncommon and significant annoyance to many patients, and in my experience it is a reason patients discontinue the treatment. Lim and Bolster describe the reaction and their approach to its management, and comment on the fairly obscure pathway that may explain its occurrence. Again, it is not clear to me why only relatively few patients experience the reaction. Is there a genetic predisposition? Or is it influenced by the patient’s baseline “inflammatory tone,” as influenced by the state of his or her microbiome or other still uncharacterized factors? And why does this reaction often diminish with repeated dosing of the drug?

Most striking is the description and discussion by Khan et al of the management of autoimmune colitis after administration of immune checkpoint inhibitor anticancer therapies. These drugs represent important advances in the therapy of various cancers. They are novel in that they are not specific to tumor type, although certain drugs within this new class of immunotherapy seem to exhibit more dramatic and enduring responses against one type of cancer than against another. These therapies are not directly tumor-reactive, but act by down-regulating the normal “brakes” or checkpoints of the immune system that normally play a role in reigning in the immune-inflammatory system response to infection once the offending infective agent is neutralized. These checkpoints have also been thought to limit the development of autoimmunity. Many cancers seem to capitalize on the activation of these brakes to evade tumor immunity. That these checkpoint therapies are so effective in some patients with heretofore unresponsive cancers is obviously a major advance. But equally striking is the scientific proof of the immunologic concept that by inhibiting these normal brakes on inflammation there is loss of normal regulation of the immune response and autoimmunity ensues unchecked. Khan et al discuss the colitis that can occur with these therapies, but a host of fascinating and potentially life-threatening organ-specific complications can be invoked by the checkpoint inhibitors, including hypophysitis, myositis, nephritis, and pneumonitis. What determines which patient will suffer these immune complications, which organs will be affected in a given patient, and the relationships between preexisting autoimmune disease, antitumor response, and these autoimmune complications are still being unraveled.

If you have not yet encountered patients with these complications in your practice, it is quite likely you will. The topic is worth reading about now (see the review by June et al¹), and we will provide additional reviews in the future.

Some drugs exhibit a dose-toxicity relationship that is sufficiently predictable to permit drug-level monitoring to limit renal or other toxicity. Others cause ocular or marrow toxicity that can be limited by weight-based dosing and careful monitoring. With azathioprine, some toxicity can be predicted by assessing the activity of the enzyme thiopurine methyltransferase, which metabolizes the drug. Other approaches to using pharmacogenomics have included HLA-B locus haplotyping to detect increased risk of immune-mediated toxicities of drugs such as carbamazepine and allopurinol. Both of these drugs exhibit serious systemic toxicities that are incompletely understood, but these are nascent and significant steps toward providing personalized (precision) medical care.

Adverse effects of some drugs may result from their intracellular effects, which are only partially predictable by drug levels or dosing. Colchicine, hydroxychloroquine, and amiodarone all affect intracellular vacuolar transport and lysosomal processing. Yet, although the footprints of drug effect can be seen in many histopathology samples, only some patients—but maybe more than currently recognized—suffer cardiac and skeletal muscle vacuolar myopathy, axonal neuropathies, or pulmonary or retinal cell toxicity from these drugs. But distinguishing the histopathologic footprints of drug exposure and the biologic effect from true drug toxicity with organ damage is not always straightforward.

Rare adverse effects may only become apparent with frequent use of a drug in the general community. These often remain mechanistically unexplained: Why can minoxidil cause pericardial effusions or a nonsteroidal anti-inflammatory drug cause aseptic meningitis? Some effects may be due to altering of the unique balance of biochemical pathways in a given patient, leading to unexpected drug metabolism with generation of toxic metabolites.

More interesting to me are effects that are seemingly off-target biologic outcomes caused by disrupting normal physiologic homeostasis and stimulating counterregulatory pathways in such a way that unexpected biologic effects occur. Angioedema and cough in some patients who have taken angiotensin-converting enzyme inhibitors are examples, but why the disturbed control mechanisms lead to these effects in only occasional patients is still incompletely elucidated.

Two additional classes of drugs with unique systemic adverse effects are discussed in this issue of the Journal. The “flulike” syndrome after bisphosphonate treatment, presumably resulting from selective cytokine release by macrophages that have ingested certain bisphosphonates, is a not uncommon and significant annoyance to many patients, and in my experience it is a reason patients discontinue the treatment. Lim and Bolster describe the reaction and their approach to its management, and comment on the fairly obscure pathway that may explain its occurrence. Again, it is not clear to me why only relatively few patients experience the reaction. Is there a genetic predisposition? Or is it influenced by the patient’s baseline “inflammatory tone,” as influenced by the state of his or her microbiome or other still uncharacterized factors? And why does this reaction often diminish with repeated dosing of the drug?

Most striking is the description and discussion by Khan et al of the management of autoimmune colitis after administration of immune checkpoint inhibitor anticancer therapies. These drugs represent important advances in the therapy of various cancers. They are novel in that they are not specific to tumor type, although certain drugs within this new class of immunotherapy seem to exhibit more dramatic and enduring responses against one type of cancer than against another. These therapies are not directly tumor-reactive, but act by down-regulating the normal “brakes” or checkpoints of the immune system that normally play a role in reigning in the immune-inflammatory system response to infection once the offending infective agent is neutralized. These checkpoints have also been thought to limit the development of autoimmunity. Many cancers seem to capitalize on the activation of these brakes to evade tumor immunity. That these checkpoint therapies are so effective in some patients with heretofore unresponsive cancers is obviously a major advance. But equally striking is the scientific proof of the immunologic concept that by inhibiting these normal brakes on inflammation there is loss of normal regulation of the immune response and autoimmunity ensues unchecked. Khan et al discuss the colitis that can occur with these therapies, but a host of fascinating and potentially life-threatening organ-specific complications can be invoked by the checkpoint inhibitors, including hypophysitis, myositis, nephritis, and pneumonitis. What determines which patient will suffer these immune complications, which organs will be affected in a given patient, and the relationships between preexisting autoimmune disease, antitumor response, and these autoimmune complications are still being unraveled.

If you have not yet encountered patients with these complications in your practice, it is quite likely you will. The topic is worth reading about now (see the review by June et al¹), and we will provide additional reviews in the future.

Some drugs exhibit a dose-toxicity relationship that is sufficiently predictable to permit drug-level monitoring to limit renal or other toxicity. Others cause ocular or marrow toxicity that can be limited by weight-based dosing and careful monitoring. With azathioprine, some toxicity can be predicted by assessing the activity of the enzyme thiopurine methyltransferase, which metabolizes the drug. Other approaches to using pharmacogenomics have included HLA-B locus haplotyping to detect increased risk of immune-mediated toxicities of drugs such as carbamazepine and allopurinol. Both of these drugs exhibit serious systemic toxicities that are incompletely understood, but these are nascent and significant steps toward providing personalized (precision) medical care.

Adverse effects of some drugs may result from their intracellular effects, which are only partially predictable by drug levels or dosing. Colchicine, hydroxychloroquine, and amiodarone all affect intracellular vacuolar transport and lysosomal processing. Yet, although the footprints of drug effect can be seen in many histopathology samples, only some patients—but maybe more than currently recognized—suffer cardiac and skeletal muscle vacuolar myopathy, axonal neuropathies, or pulmonary or retinal cell toxicity from these drugs. But distinguishing the histopathologic footprints of drug exposure and the biologic effect from true drug toxicity with organ damage is not always straightforward.

Rare adverse effects may only become apparent with frequent use of a drug in the general community. These often remain mechanistically unexplained: Why can minoxidil cause pericardial effusions or a nonsteroidal anti-inflammatory drug cause aseptic meningitis? Some effects may be due to altering of the unique balance of biochemical pathways in a given patient, leading to unexpected drug metabolism with generation of toxic metabolites.

More interesting to me are effects that are seemingly off-target biologic outcomes caused by disrupting normal physiologic homeostasis and stimulating counterregulatory pathways in such a way that unexpected biologic effects occur. Angioedema and cough in some patients who have taken angiotensin-converting enzyme inhibitors are examples, but why the disturbed control mechanisms lead to these effects in only occasional patients is still incompletely elucidated.

Two additional classes of drugs with unique systemic adverse effects are discussed in this issue of the Journal. The “flulike” syndrome after bisphosphonate treatment, presumably resulting from selective cytokine release by macrophages that have ingested certain bisphosphonates, is a not uncommon and significant annoyance to many patients, and in my experience it is a reason patients discontinue the treatment. Lim and Bolster describe the reaction and their approach to its management, and comment on the fairly obscure pathway that may explain its occurrence. Again, it is not clear to me why only relatively few patients experience the reaction. Is there a genetic predisposition? Or is it influenced by the patient’s baseline “inflammatory tone,” as influenced by the state of his or her microbiome or other still uncharacterized factors? And why does this reaction often diminish with repeated dosing of the drug?

Most striking is the description and discussion by Khan et al of the management of autoimmune colitis after administration of immune checkpoint inhibitor anticancer therapies. These drugs represent important advances in the therapy of various cancers. They are novel in that they are not specific to tumor type, although certain drugs within this new class of immunotherapy seem to exhibit more dramatic and enduring responses against one type of cancer than against another. These therapies are not directly tumor-reactive, but act by down-regulating the normal “brakes” or checkpoints of the immune system that normally play a role in reigning in the immune-inflammatory system response to infection once the offending infective agent is neutralized. These checkpoints have also been thought to limit the development of autoimmunity. Many cancers seem to capitalize on the activation of these brakes to evade tumor immunity. That these checkpoint therapies are so effective in some patients with heretofore unresponsive cancers is obviously a major advance. But equally striking is the scientific proof of the immunologic concept that by inhibiting these normal brakes on inflammation there is loss of normal regulation of the immune response and autoimmunity ensues unchecked. Khan et al discuss the colitis that can occur with these therapies, but a host of fascinating and potentially life-threatening organ-specific complications can be invoked by the checkpoint inhibitors, including hypophysitis, myositis, nephritis, and pneumonitis. What determines which patient will suffer these immune complications, which organs will be affected in a given patient, and the relationships between preexisting autoimmune disease, antitumor response, and these autoimmune complications are still being unraveled.

If you have not yet encountered patients with these complications in your practice, it is quite likely you will. The topic is worth reading about now (see the review by June et al¹), and we will provide additional reviews in the future.