Cleveland Clinic Journal of Medicine

A striking change in cancer medicine over the past several decades has been the rising amplitude of the voices of cancer patients and survivors and their loved ones. Increasingly, they have organized as advocates for better cancer treatment, better research, and better attention to the experience of those directly affected by cancer.

At the same time, despite a somewhat delirious period in the 1970s when it was expected that cancer could be cured, or at least that the mortality rate could be cut in half, it has become clear that progress in eliminating the disease is a long, slow slog with no guarantee of success. Nowhere is this lack of major progress clearer than in the US Food and Drug Administration’s decision a few years ago to approve a drug for pancreatic cancer based on a 10-day increase in median survival.¹

These two factors, the rising voices of those affected by cancer and the failure of cancer research to deliver a cure, have helped fuel a dramatic increase in the attention paid to the symptoms caused by cancer and its treatments. Improving quality of life has become recognized as an additional important goal worthy of rigorous study. Alleviating symptoms is often the most we have to offer patients with advanced cancer, and palliative medicine services are now found at many if not all major cancer centers. To help ensure a supply of well-trained palliative medicine doctors, accredited palliative medicine fellowships have been started at institutions around the country. These programs produce physicians with a higher level of expertise in managing pain, nausea, constipation, fatigue, psychosocial distress, dyspnea, and a wide variety of other symptoms. And while the needs of cancer patients have helped accelerate the growth of palliative medicine, the specialty has a role to play with almost any patient with intractable symptoms, regardless of the nature of the underlying disease.

With the growing recognition of a need to better manage cancer patients’ symptoms, research in palliative care has grown rapidly, and an evidence-based approach to symptom management has become possible. Meanwhile, a variety of substantial advances has occurred. For instance, modern antiemetics have dramatically reduced chemotherapy-related vomiting, and long-acting narcotics have allowed patients to achieve better pain control with milder side effects.

In other areas such as cancer-related fatigue or chemotherapy-induced neuropathy, there is very limited evidence that our interventions are effective at alleviating symptoms or improving quality of life. In these and a number of other areas, more research and better treatments are urgently needed.

In order to keep our readers up to date on the progress that is being made in palliative medicine for cancer patients, the Cleveland Clinic Journal of Medicine is presenting a series of articles on the topic. The series begins this issue with an article on the principles of symptom management and a review of the current state of knowledge about alleviating fatigue, nausea, constipation, and dyspnea. Future articles will focus on pain and bowel obstruction.

Our goal is to give our readers practical information that will help them provide better symptom management for their patients, particularly their cancer patients. This series will deal with one of the most important problems of cancer medicine.

A striking change in cancer medicine over the past several decades has been the rising amplitude of the voices of cancer patients and survivors and their loved ones. Increasingly, they have organized as advocates for better cancer treatment, better research, and better attention to the experience of those directly affected by cancer.

At the same time, despite a somewhat delirious period in the 1970s when it was expected that cancer could be cured, or at least that the mortality rate could be cut in half, it has become clear that progress in eliminating the disease is a long, slow slog with no guarantee of success. Nowhere is this lack of major progress clearer than in the US Food and Drug Administration’s decision a few years ago to approve a drug for pancreatic cancer based on a 10-day increase in median survival.¹

These two factors, the rising voices of those affected by cancer and the failure of cancer research to deliver a cure, have helped fuel a dramatic increase in the attention paid to the symptoms caused by cancer and its treatments. Improving quality of life has become recognized as an additional important goal worthy of rigorous study. Alleviating symptoms is often the most we have to offer patients with advanced cancer, and palliative medicine services are now found at many if not all major cancer centers. To help ensure a supply of well-trained palliative medicine doctors, accredited palliative medicine fellowships have been started at institutions around the country. These programs produce physicians with a higher level of expertise in managing pain, nausea, constipation, fatigue, psychosocial distress, dyspnea, and a wide variety of other symptoms. And while the needs of cancer patients have helped accelerate the growth of palliative medicine, the specialty has a role to play with almost any patient with intractable symptoms, regardless of the nature of the underlying disease.

With the growing recognition of a need to better manage cancer patients’ symptoms, research in palliative care has grown rapidly, and an evidence-based approach to symptom management has become possible. Meanwhile, a variety of substantial advances has occurred. For instance, modern antiemetics have dramatically reduced chemotherapy-related vomiting, and long-acting narcotics have allowed patients to achieve better pain control with milder side effects.

In other areas such as cancer-related fatigue or chemotherapy-induced neuropathy, there is very limited evidence that our interventions are effective at alleviating symptoms or improving quality of life. In these and a number of other areas, more research and better treatments are urgently needed.

In order to keep our readers up to date on the progress that is being made in palliative medicine for cancer patients, the Cleveland Clinic Journal of Medicine is presenting a series of articles on the topic. The series begins this issue with an article on the principles of symptom management and a review of the current state of knowledge about alleviating fatigue, nausea, constipation, and dyspnea. Future articles will focus on pain and bowel obstruction.

Our goal is to give our readers practical information that will help them provide better symptom management for their patients, particularly their cancer patients. This series will deal with one of the most important problems of cancer medicine.

A striking change in cancer medicine over the past several decades has been the rising amplitude of the voices of cancer patients and survivors and their loved ones. Increasingly, they have organized as advocates for better cancer treatment, better research, and better attention to the experience of those directly affected by cancer.

At the same time, despite a somewhat delirious period in the 1970s when it was expected that cancer could be cured, or at least that the mortality rate could be cut in half, it has become clear that progress in eliminating the disease is a long, slow slog with no guarantee of success. Nowhere is this lack of major progress clearer than in the US Food and Drug Administration’s decision a few years ago to approve a drug for pancreatic cancer based on a 10-day increase in median survival.¹

These two factors, the rising voices of those affected by cancer and the failure of cancer research to deliver a cure, have helped fuel a dramatic increase in the attention paid to the symptoms caused by cancer and its treatments. Improving quality of life has become recognized as an additional important goal worthy of rigorous study. Alleviating symptoms is often the most we have to offer patients with advanced cancer, and palliative medicine services are now found at many if not all major cancer centers. To help ensure a supply of well-trained palliative medicine doctors, accredited palliative medicine fellowships have been started at institutions around the country. These programs produce physicians with a higher level of expertise in managing pain, nausea, constipation, fatigue, psychosocial distress, dyspnea, and a wide variety of other symptoms. And while the needs of cancer patients have helped accelerate the growth of palliative medicine, the specialty has a role to play with almost any patient with intractable symptoms, regardless of the nature of the underlying disease.

With the growing recognition of a need to better manage cancer patients’ symptoms, research in palliative care has grown rapidly, and an evidence-based approach to symptom management has become possible. Meanwhile, a variety of substantial advances has occurred. For instance, modern antiemetics have dramatically reduced chemotherapy-related vomiting, and long-acting narcotics have allowed patients to achieve better pain control with milder side effects.

In other areas such as cancer-related fatigue or chemotherapy-induced neuropathy, there is very limited evidence that our interventions are effective at alleviating symptoms or improving quality of life. In these and a number of other areas, more research and better treatments are urgently needed.

In order to keep our readers up to date on the progress that is being made in palliative medicine for cancer patients, the Cleveland Clinic Journal of Medicine is presenting a series of articles on the topic. The series begins this issue with an article on the principles of symptom management and a review of the current state of knowledge about alleviating fatigue, nausea, constipation, and dyspnea. Future articles will focus on pain and bowel obstruction.

Our goal is to give our readers practical information that will help them provide better symptom management for their patients, particularly their cancer patients. This series will deal with one of the most important problems of cancer medicine.

We believe the available evidence does not support routine antimicrobial prophylaxis before dental procedures in patients who have undergone total joint replacement, even though the practice is very common¹ and even though professional societies recommend it in patients at high risk,² or even in all patients.³

On the other hand, good oral hygiene prevents dental disease and decreases the frequency of bacteremia from routine daily activities, and thus should be especially encouraged in patients with prosthetic joints or in those undergoing total joint arthroplasty.

AN UNCOMMON BUT SERIOUS PROBLEM

By 2030, an estimated 4 million total hip or total knee replacements per year will be performed in the United States.⁴ Most patients have a satisfactory outcome, but in a small percentage the prosthesis fails prematurely.

Prosthetic joint infection is the second most common cause of prosthetic failure leading to loss of joint function, after aseptic loosening.⁵ Its treatment often requires removal of the infected prosthesis and prolonged intravenous antimicrobial therapy. The cost incurred with each episode of prosthetic joint infection is estimated to exceed $50,000.¹

Because of the morbidity and substantial cost associated with managing this condition, investigators have focused on identifying preventable risk factors for it.

RISK FACTORS FOR PROSTHETIC JOINT INFECTION

Factors associated with a higher risk of prosthetic joint infection include prior joint surgery, failure to give antimicrobial prophylaxis during surgery, immunosuppression, perioperative wound complications, a high American Society of Anesthesiologists (ASA) score, prolonged operative time, and a history of prosthetic joint infection.^6,7 The primary predisposing factors are related to the foreign body itself and to the opportunities for and the degree of exposure of the prosthesis to microorganisms during surgery. Bacteremia, especially with Staphylococcus aureus, has been recognized as a risk factor for hematogenous prosthetic joint infection.⁶

Whether dental procedures pose a risk of prosthetic joint infection has been debated for decades. Common daily activities such as toothbrushing and chewing can cause transient bacteremia in up to 40% of episodes.⁸

Extrapolating from the guidelines for preventing endocarditis, the American Dental Association (ADA)² and the American Academy of Orthopaedic Surgeons (AAOS)³ have issued guidelines favoring antimicrobial prophylaxis in patients with prosthetic joints. However, given the significant differences in the pathophysiology, microbiology, and anatomy of infection between infective endocarditis and prosthetic joint infection, extrapolating the recommendations may not be valid.

MICROBIOLOGY OF PROSTHETIC JOINT INFECTION AND DENTAL FLORA

Staphylococci, the most common cause of prosthetic joint infection, are uncommon commensals of the oral flora and have been rarely implicated in bacteremia occurring after dental procedures.⁹ In contrast, viridans-group streptococci constitute most of the facultative oral flora and are the most common cause of transient bacteremia after dental procedures that result in trauma to the gingival or oral mucosa.¹⁰ However, viridans-group streptococci account for only 2% of all hematogenous prosthetic joint infections.⁹

DO DENTAL PROCEDURES INCREASE THE RISK OF PROSTHETIC JOINT INFECTION?

Prolonged or high-grade bacteremia is associated with prosthetic joint infection. On the other hand, data are scant on the association between low-grade or transient bacteremia and prosthetic joint infection.

After dental procedures, bacteria can be found in the blood, but at much lower levels (< 10⁴ cfu/mL) than that needed for hematogenous seeding of prostheses in animal studies (3–5 × 10⁸ cfu/mL).¹¹ Transient, low-grade bacteremia occurs not only after dental procedures but also, as mentioned, after common activities such as chewing, brushing, or flossing.¹ The cumulative exposure to transient bacteremia through these daily activities is several times higher than the single exposure that a person is subjected to during dental procedures.¹²

WHAT IS THE EVIDENCE?

Most of the current evidence linking dental procedures or dental manipulation to prosthetic joint infection is based on reports of single cases of infections that occurred after dental procedures.

In two retrospective reviews, late hematogenous prosthetic joint infection associated with a dental source occurred after 0.2% of primary knee arthroplasties¹¹ and 6% of primary hip arthroplasties.¹³

Ainscow and Denham¹⁴ followed 1,000 patients who underwent total joint replacement over 6 years. Of these, 226 subsequently underwent dental procedures without receiving antimicrobial prophylaxis, and none developed a prosthetic joint infection.

In a recently published case-control study,¹ our group assessed 339 patients with prosthetic joint infection and 339 patients with prosthetic joints that did not become infected. In this study, neither low-risk nor high-risk dental procedures were associated with an increased risk of prosthetic knee or hip infections (odds ratio [OR] 0.8; 95% confidence interval [CI] 0.4–1.6). Moreover, prophylactic use of antimicrobials before dental procedures was not associated with a lower risk.

However, a factor that was associated with a lower risk of prosthetic joint infection was good oral hygiene (OR 0.7; 95% CI 0.5–1.03). Good oral hygiene and prevention of dental disease could potentially decrease the frequency of bacteremia from daily activities and may even protect against prosthetic joint infection. Further study of the association of poor dental health and the risk of prosthetic joint infection is warranted.

GUIDELINES AND RECOMMENDATIONS

Despite the lack of evidence suggesting an association between prosthetic joint infection and dental procedure, surveys of orthopedists, dentists, infectious disease specialists, and other health care professionals show that a significant number of them recommend antimicrobial prophylaxis for patients with a prosthetic joint prior to a dental procedure.¹

In 2003, a consensus panel of the AAOS and the ADA recommended routine consideration of antimicrobial prophylaxis in patients at high risk due to both patient factors and the type of dental procedure.² Patient factors thought to confer high risk are immunosuppression, diabetes, malnourishment, human immunodeficiency virus infection, prior prosthetic joint infection, hemophilia, malignancy, and a prosthesis less than 2 years old. High-risk dental procedures are tooth extractions, periodontal procedures, root canal surgery, and dental cleaning in which bleeding is anticipated.

In a recent statement, the AAOS recommended antimicrobial prophylaxis in all patients with prosthetic joints.³

Concerns about promoting antimicrobial resistance and about adverse reactions from antimicrobial use may outweigh any hypothetic benefit related to prevention of prosthetic joint infection. Analyses of cost, risks, and benefits argue against this practice.³

In summary, the current evidence does not support the use of antimicrobial therapy to prevent prosthetic joint infection in patients with total joint replacement undergoing dental procedures. However, good oral hygiene should be encouraged to prevent dental disease and to decrease the frequency of bacteremia from routine daily activities in patients who have undergone or will be undergoing total joint arthroplasty.

We believe the available evidence does not support routine antimicrobial prophylaxis before dental procedures in patients who have undergone total joint replacement, even though the practice is very common¹ and even though professional societies recommend it in patients at high risk,² or even in all patients.³

On the other hand, good oral hygiene prevents dental disease and decreases the frequency of bacteremia from routine daily activities, and thus should be especially encouraged in patients with prosthetic joints or in those undergoing total joint arthroplasty.

AN UNCOMMON BUT SERIOUS PROBLEM

By 2030, an estimated 4 million total hip or total knee replacements per year will be performed in the United States.⁴ Most patients have a satisfactory outcome, but in a small percentage the prosthesis fails prematurely.

Prosthetic joint infection is the second most common cause of prosthetic failure leading to loss of joint function, after aseptic loosening.⁵ Its treatment often requires removal of the infected prosthesis and prolonged intravenous antimicrobial therapy. The cost incurred with each episode of prosthetic joint infection is estimated to exceed $50,000.¹

Because of the morbidity and substantial cost associated with managing this condition, investigators have focused on identifying preventable risk factors for it.

RISK FACTORS FOR PROSTHETIC JOINT INFECTION

Factors associated with a higher risk of prosthetic joint infection include prior joint surgery, failure to give antimicrobial prophylaxis during surgery, immunosuppression, perioperative wound complications, a high American Society of Anesthesiologists (ASA) score, prolonged operative time, and a history of prosthetic joint infection.^6,7 The primary predisposing factors are related to the foreign body itself and to the opportunities for and the degree of exposure of the prosthesis to microorganisms during surgery. Bacteremia, especially with Staphylococcus aureus, has been recognized as a risk factor for hematogenous prosthetic joint infection.⁶

Whether dental procedures pose a risk of prosthetic joint infection has been debated for decades. Common daily activities such as toothbrushing and chewing can cause transient bacteremia in up to 40% of episodes.⁸

Extrapolating from the guidelines for preventing endocarditis, the American Dental Association (ADA)² and the American Academy of Orthopaedic Surgeons (AAOS)³ have issued guidelines favoring antimicrobial prophylaxis in patients with prosthetic joints. However, given the significant differences in the pathophysiology, microbiology, and anatomy of infection between infective endocarditis and prosthetic joint infection, extrapolating the recommendations may not be valid.

MICROBIOLOGY OF PROSTHETIC JOINT INFECTION AND DENTAL FLORA

Staphylococci, the most common cause of prosthetic joint infection, are uncommon commensals of the oral flora and have been rarely implicated in bacteremia occurring after dental procedures.⁹ In contrast, viridans-group streptococci constitute most of the facultative oral flora and are the most common cause of transient bacteremia after dental procedures that result in trauma to the gingival or oral mucosa.¹⁰ However, viridans-group streptococci account for only 2% of all hematogenous prosthetic joint infections.⁹

DO DENTAL PROCEDURES INCREASE THE RISK OF PROSTHETIC JOINT INFECTION?

Prolonged or high-grade bacteremia is associated with prosthetic joint infection. On the other hand, data are scant on the association between low-grade or transient bacteremia and prosthetic joint infection.

After dental procedures, bacteria can be found in the blood, but at much lower levels (< 10⁴ cfu/mL) than that needed for hematogenous seeding of prostheses in animal studies (3–5 × 10⁸ cfu/mL).¹¹ Transient, low-grade bacteremia occurs not only after dental procedures but also, as mentioned, after common activities such as chewing, brushing, or flossing.¹ The cumulative exposure to transient bacteremia through these daily activities is several times higher than the single exposure that a person is subjected to during dental procedures.¹²

WHAT IS THE EVIDENCE?

Most of the current evidence linking dental procedures or dental manipulation to prosthetic joint infection is based on reports of single cases of infections that occurred after dental procedures.

In two retrospective reviews, late hematogenous prosthetic joint infection associated with a dental source occurred after 0.2% of primary knee arthroplasties¹¹ and 6% of primary hip arthroplasties.¹³

Ainscow and Denham¹⁴ followed 1,000 patients who underwent total joint replacement over 6 years. Of these, 226 subsequently underwent dental procedures without receiving antimicrobial prophylaxis, and none developed a prosthetic joint infection.

In a recently published case-control study,¹ our group assessed 339 patients with prosthetic joint infection and 339 patients with prosthetic joints that did not become infected. In this study, neither low-risk nor high-risk dental procedures were associated with an increased risk of prosthetic knee or hip infections (odds ratio [OR] 0.8; 95% confidence interval [CI] 0.4–1.6). Moreover, prophylactic use of antimicrobials before dental procedures was not associated with a lower risk.

However, a factor that was associated with a lower risk of prosthetic joint infection was good oral hygiene (OR 0.7; 95% CI 0.5–1.03). Good oral hygiene and prevention of dental disease could potentially decrease the frequency of bacteremia from daily activities and may even protect against prosthetic joint infection. Further study of the association of poor dental health and the risk of prosthetic joint infection is warranted.

GUIDELINES AND RECOMMENDATIONS

Despite the lack of evidence suggesting an association between prosthetic joint infection and dental procedure, surveys of orthopedists, dentists, infectious disease specialists, and other health care professionals show that a significant number of them recommend antimicrobial prophylaxis for patients with a prosthetic joint prior to a dental procedure.¹

In 2003, a consensus panel of the AAOS and the ADA recommended routine consideration of antimicrobial prophylaxis in patients at high risk due to both patient factors and the type of dental procedure.² Patient factors thought to confer high risk are immunosuppression, diabetes, malnourishment, human immunodeficiency virus infection, prior prosthetic joint infection, hemophilia, malignancy, and a prosthesis less than 2 years old. High-risk dental procedures are tooth extractions, periodontal procedures, root canal surgery, and dental cleaning in which bleeding is anticipated.

In a recent statement, the AAOS recommended antimicrobial prophylaxis in all patients with prosthetic joints.³

Concerns about promoting antimicrobial resistance and about adverse reactions from antimicrobial use may outweigh any hypothetic benefit related to prevention of prosthetic joint infection. Analyses of cost, risks, and benefits argue against this practice.³

In summary, the current evidence does not support the use of antimicrobial therapy to prevent prosthetic joint infection in patients with total joint replacement undergoing dental procedures. However, good oral hygiene should be encouraged to prevent dental disease and to decrease the frequency of bacteremia from routine daily activities in patients who have undergone or will be undergoing total joint arthroplasty.

We believe the available evidence does not support routine antimicrobial prophylaxis before dental procedures in patients who have undergone total joint replacement, even though the practice is very common¹ and even though professional societies recommend it in patients at high risk,² or even in all patients.³

On the other hand, good oral hygiene prevents dental disease and decreases the frequency of bacteremia from routine daily activities, and thus should be especially encouraged in patients with prosthetic joints or in those undergoing total joint arthroplasty.

AN UNCOMMON BUT SERIOUS PROBLEM

By 2030, an estimated 4 million total hip or total knee replacements per year will be performed in the United States.⁴ Most patients have a satisfactory outcome, but in a small percentage the prosthesis fails prematurely.

Prosthetic joint infection is the second most common cause of prosthetic failure leading to loss of joint function, after aseptic loosening.⁵ Its treatment often requires removal of the infected prosthesis and prolonged intravenous antimicrobial therapy. The cost incurred with each episode of prosthetic joint infection is estimated to exceed $50,000.¹

Because of the morbidity and substantial cost associated with managing this condition, investigators have focused on identifying preventable risk factors for it.

RISK FACTORS FOR PROSTHETIC JOINT INFECTION

Factors associated with a higher risk of prosthetic joint infection include prior joint surgery, failure to give antimicrobial prophylaxis during surgery, immunosuppression, perioperative wound complications, a high American Society of Anesthesiologists (ASA) score, prolonged operative time, and a history of prosthetic joint infection.^6,7 The primary predisposing factors are related to the foreign body itself and to the opportunities for and the degree of exposure of the prosthesis to microorganisms during surgery. Bacteremia, especially with Staphylococcus aureus, has been recognized as a risk factor for hematogenous prosthetic joint infection.⁶

Whether dental procedures pose a risk of prosthetic joint infection has been debated for decades. Common daily activities such as toothbrushing and chewing can cause transient bacteremia in up to 40% of episodes.⁸

Extrapolating from the guidelines for preventing endocarditis, the American Dental Association (ADA)² and the American Academy of Orthopaedic Surgeons (AAOS)³ have issued guidelines favoring antimicrobial prophylaxis in patients with prosthetic joints. However, given the significant differences in the pathophysiology, microbiology, and anatomy of infection between infective endocarditis and prosthetic joint infection, extrapolating the recommendations may not be valid.

MICROBIOLOGY OF PROSTHETIC JOINT INFECTION AND DENTAL FLORA

Staphylococci, the most common cause of prosthetic joint infection, are uncommon commensals of the oral flora and have been rarely implicated in bacteremia occurring after dental procedures.⁹ In contrast, viridans-group streptococci constitute most of the facultative oral flora and are the most common cause of transient bacteremia after dental procedures that result in trauma to the gingival or oral mucosa.¹⁰ However, viridans-group streptococci account for only 2% of all hematogenous prosthetic joint infections.⁹

DO DENTAL PROCEDURES INCREASE THE RISK OF PROSTHETIC JOINT INFECTION?

Prolonged or high-grade bacteremia is associated with prosthetic joint infection. On the other hand, data are scant on the association between low-grade or transient bacteremia and prosthetic joint infection.

After dental procedures, bacteria can be found in the blood, but at much lower levels (< 10⁴ cfu/mL) than that needed for hematogenous seeding of prostheses in animal studies (3–5 × 10⁸ cfu/mL).¹¹ Transient, low-grade bacteremia occurs not only after dental procedures but also, as mentioned, after common activities such as chewing, brushing, or flossing.¹ The cumulative exposure to transient bacteremia through these daily activities is several times higher than the single exposure that a person is subjected to during dental procedures.¹²

WHAT IS THE EVIDENCE?

Most of the current evidence linking dental procedures or dental manipulation to prosthetic joint infection is based on reports of single cases of infections that occurred after dental procedures.

In two retrospective reviews, late hematogenous prosthetic joint infection associated with a dental source occurred after 0.2% of primary knee arthroplasties¹¹ and 6% of primary hip arthroplasties.¹³

Ainscow and Denham¹⁴ followed 1,000 patients who underwent total joint replacement over 6 years. Of these, 226 subsequently underwent dental procedures without receiving antimicrobial prophylaxis, and none developed a prosthetic joint infection.

In a recently published case-control study,¹ our group assessed 339 patients with prosthetic joint infection and 339 patients with prosthetic joints that did not become infected. In this study, neither low-risk nor high-risk dental procedures were associated with an increased risk of prosthetic knee or hip infections (odds ratio [OR] 0.8; 95% confidence interval [CI] 0.4–1.6). Moreover, prophylactic use of antimicrobials before dental procedures was not associated with a lower risk.

However, a factor that was associated with a lower risk of prosthetic joint infection was good oral hygiene (OR 0.7; 95% CI 0.5–1.03). Good oral hygiene and prevention of dental disease could potentially decrease the frequency of bacteremia from daily activities and may even protect against prosthetic joint infection. Further study of the association of poor dental health and the risk of prosthetic joint infection is warranted.

GUIDELINES AND RECOMMENDATIONS

Despite the lack of evidence suggesting an association between prosthetic joint infection and dental procedure, surveys of orthopedists, dentists, infectious disease specialists, and other health care professionals show that a significant number of them recommend antimicrobial prophylaxis for patients with a prosthetic joint prior to a dental procedure.¹

In 2003, a consensus panel of the AAOS and the ADA recommended routine consideration of antimicrobial prophylaxis in patients at high risk due to both patient factors and the type of dental procedure.² Patient factors thought to confer high risk are immunosuppression, diabetes, malnourishment, human immunodeficiency virus infection, prior prosthetic joint infection, hemophilia, malignancy, and a prosthesis less than 2 years old. High-risk dental procedures are tooth extractions, periodontal procedures, root canal surgery, and dental cleaning in which bleeding is anticipated.

In a recent statement, the AAOS recommended antimicrobial prophylaxis in all patients with prosthetic joints.³

Concerns about promoting antimicrobial resistance and about adverse reactions from antimicrobial use may outweigh any hypothetic benefit related to prevention of prosthetic joint infection. Analyses of cost, risks, and benefits argue against this practice.³

In summary, the current evidence does not support the use of antimicrobial therapy to prevent prosthetic joint infection in patients with total joint replacement undergoing dental procedures. However, good oral hygiene should be encouraged to prevent dental disease and to decrease the frequency of bacteremia from routine daily activities in patients who have undergone or will be undergoing total joint arthroplasty.

Vascular catheters are very common in everyday inpatient and, increasingly, outpatient care. Nearly 300 million catheters are estimated to be used annually in the United States, and approximately 3 million of these are central venous catheters (CVCs).¹

Although significant gains have been made in preventing CVC-related bloodstream infections, these infections continue to occur, with estimated rates ranging from 1.3 per 1,000 catheter days on inpatient medical-surgical wards to 5.6 per 1,000 catheter days in intensive care burn units.²

CVCs are classified as either long-term or short-term. Long-term CVCs are surgically implanted or tunneled and used for prolonged chemotherapy, home infusion therapy, or hemodialysis. Short-term CVCs do not require surgical implantation. They are more common than long-term CVCs and account for most CVC-related bloodstream infections. Given the frequency of short-term CVC use, a growing number of health care providers from mid-level practitioners to intensivists are faced with deciding how to manage bloodstream infection related to short-term CVCs.

At baseline, management decisions about bloodstream infections from short-term CVCs can be challenging. Questions that regularly arise include:

• Should a potentially infected catheter be removed?
• Which empiric antibiotic therapy should be started pending a microbiologic diagnosis?
• How should therapy be tailored (eg, antibiotic choice and course and whether to remove or retain the catheter) based on the specific pathogen identified?

Adding to the complexity of these decisions are increasingly resistant microorganisms, heterogeneity of affected patient populations, and variability in the quality and availability of evidence.

This review provides a concise guide to managing bloodstream infections related to short-term CVCs in adults, based on updated guidelines from the Infectious Diseases Society of America (IDSA).³

DEFINITION AND DIAGNOSTIC CRITERIA

We have adapted the following definition and diagnostic criteria from the general definition and diagnostic criteria for catheter-related bloodstream infections proposed by the IDSA.

Bloodstream infection related to a short-term CVC is defined as bacteremia or fungemia in a patient with the CVC in place, clinical manifestations of infection (eg, fever, chills, hypotension), and no apparent source of the bloodstream infection aside from the catheter. At least one of the three diagnostic criteria should be met:

• Cultures of the catheter tip and of the peripheral blood grow the same organism. Catheter tip culture should be quantitative, with more than 10² colony-forming units (cfu) per catheter segment, or semiquantitative, with more than 15 cfu per catheter segment.
• Blood drawn from the catheter lumen grows the same organism as blood drawn from a peripheral vein (or less optimally, a different lumen), but at three times the amount by quantitative culture.
• Blood drawn simultaneously from the catheter lumen and from a peripheral vein (or less optimally, a different lumen) grows the same organism, and growth from the CVC lumen sample is detected (by automated blood culture system) at least 2 hours before growth from the peripheral vein sample.

MANAGING BLOODSTREAM INFECTIONS IN PATIENTS WITH SHORT-TERM CVCs

When to remove a potentially infected short-term CVC

In a nonneutropenic intensive care population, Bouza et al⁴ found that, of 204 episodes of clinically suspected bloodstream infection from a short-term CVC, only 28 (14%) were confirmed to be catheter-related, 27 (13%) were bloodstream infections that were not catheter-related, 36 (18%) involved catheter-tip colonization with negative blood cultures, and the remainder were cases with negative catheter-tip and blood cultures.

Rijnders et al,⁵ in a study of 100 adult medical-surgical intensive care patients with a clinically suspected bloodstream infection related to a short-term CVC, found only three confirmed cases.

A randomized clinical trial comparing early removal of short-term CVCs and watchful waiting in an adult intensive care population with clinically suspected bloodstream infections showed no difference between treatment groups in length of stay in the intensive care unit or in the mortality rate.⁶ This trial included a low-risk subset of adult medical-surgical intensive care patients (ie, immunocompetent, no intravascular foreign body, no evidence of severe sepsis or septic shock, no evidence of infection at the catheter insertion site, no proven bacteremia or fungemia). These results suggest that a similar subset of patients can be safely monitored without catheter removal while being assessed for possible catheter-related bloodstream infection.

Empiric catheter removal vs watchful waiting has not and likely will not be studied in higher-risk populations. In this group, clinical judgment should outweigh any specific management algorithm. In patients who are in shock or who are otherwise hemodynamically unstable, early catheter removal should be a priority; however, in some circumstances the risks of immediate catheter removal (eg, coagulopathy with risk of bleeding diathesis, or lack of site to replace the catheter) may outweigh the potential benefits.

In order of prevalence, the four most common pathogens are coagulase-negative staphylococci, Staphylococcus aureus, Candida species, and enteric gram-negative bacilli.⁷

Gram-positive pathogens. A recent randomized clinical trial comparing vancomycin and linezolid (Zyvox) treatment for CVC-related bloodstream infections showed that 89 (57%) of 157 S aureus isolates and 95 (80%) of 119 coagulase-negative staphylococcal isolates were resistant to methicillin.⁸ Given the prevalence of gram-positive infections and the regularity of methicillin-resistant isolates, vancomycin should be started empirically in cases of suspected bloodstream infection related to short-term CVCs. In institutions where methicillin-resistant S aureus (MRSA) isolates regularly have a vancomycin minimum inhibitory concentration (MIC) of greater than 2 μg/mL, an alternative agent such as daptomycin (Cubicin) should be used.^9,10

Gram-negative pathogens. Infections due to resistant gram-negative pathogens have become more common in the past 10 years.^11,12 Prospective cohort studies have shown that resistant gram-negative infections and inadequate empiric antimicrobial therapy of bloodstream infections independently predict the risk of death.^13,14 Risk factors for resistant gram-negative infections include critical illness, neutropenia, prior antibiotic therapy, and femoral insertion of the CVC.^15–18 Patients with these risk factors should receive empiric antibiotic therapy for gram-negative bacilli.

No randomized controlled trial has been done to guide the choice of empiric gram-negative antibiotic coverage. The initial choice should be based on local antimicrobial patterns and susceptibility data and on the severity of the patient’s illness. Initial options include fourth-generation cephalosporins, carbapenems, or combined beta-lactam and beta-lactamase inhibitors. Patients with neutropenia, severe sepsis, or known multiple-drug-resistant gram-negative bacilli colonization or prior infection should receive empiric combination therapy with two different classes of antibiotics.

Candida. Risk factors for CVC-related bloodstream infections due to Candida species include total parenteral nutrition, prolonged use of broad-spectrum antibiotics, hematologic malignancy, solid organ or bone marrow transplantation, colonization with Candida species at multiple sites, and femoral catheter insertion. Empiric treatment with an echinocandin is recommended for patients with these risk factors. Fluconazole (Diflucan) can be substituted for an echinocandin in patients without azole exposure in the previous 3 months and in settings where the prevalence of Candida krusei and Candida glabrata is low.

PATHOGEN-SPECIFIC MANAGEMENT: RECOMMENDATIONS

Coagulase-negative staphylococci

Most patients with coagulase-negative staphylococcal infections have a benign clinical course.

Although no randomized trial has evaluated different treatment approaches, most experts recommend removing the catheter and giving a short course of antibiotics (ie, 5–7 days). Longer courses of antibiotics may be required for patients with endovascular hardware in place or persistent fever or bacteremia after catheter removal. The IDSA guidelines recommend 5 to 7 days of antibiotic therapy if the catheter is removed, and 10 to 14 days of systemic antibiotic therapy in combination with “antibiotic lock therapy” if the catheter is retained. Antibiotic lock therapy involves instilling a high concentration of an antibiotic to which the organism is susceptible into the catheter lumen and allowing it to dwell.

Not all patients are good candidates for antibiotic lock therapy, and neither are all organisms. In general, patients should be at low risk (immunocompetent, without hardware in place), and organisms should have a low risk of causing metastatic infection.

Staphylococcus lugdunensis can cause endocarditis and metastatic infections similar to those caused by S aureus and so should be managed similarly to S aureus.¹⁹

Staphylococcus aureus

Short-term CVCs infected with S aureus should be removed immediately. Removal of vascular catheters infected with S aureus has been associated with more rapid clinical response and higher cure rates compared with catheter retention.^20–23S aureus bacteremia results in hematogenous complications in 20% to 30% of patients, and failure to remove or a delay in removing the catheter increases the risk of complications.^21,24–27

There are no data from randomized clinical trials on the optimal duration of antibiotic therapy for S aureus bloodstream infections related to short-term CVCs. Traditionally, 4 weeks have been recommended out of concern for the risk of infective endocarditis,^28,29 and the IDSA recommends 4 to 6 weeks unless patients meet certain low-risk criteria.

Factors associated with a higher risk of hematogenous complications include the presence of a retained foreign body, an intravascular prosthetic device, retained catheter, immune suppression, diabetes, persistent bacteremia at 72 hours despite catheter removal and appropriate antibiotics, skin changes consistent with septic emboli, or evidence of endocarditis or suppurative thrombophlebitis on transesophageal echocardiography (TEE) or ultrasonography, respectively.^21,25–27 TEE is superior to transthoracic echocardiography and is most sensitive when performed 5 to 7 days after the onset of bacteremia.^28,30 Patients who have had the catheter removed and who do not have any of these risk factors, and in whom TEE performed 5 to 7 days after the onset of bacteremia is negative, can be considered for a shorter duration of therapy (but a minimum of 14 days).

Patients with catheters colonized with S aureus (ie, those with positive catheter-tip cultures and negative blood cultures) are at risk of subsequent bacteremia. This risk may be reduced with anti-staphylococcal therapy started within 24 hours of catheter removal.^31,32 Therapy should be continued for 5 to 7 days, and patients should be closely monitored for signs or symptoms of ongoing infection.

Oxacillin or nafcillin should be the first-line therapy for susceptible S aureus isolates. Vancomycin should be used to treat MRSA. Patients with MRSA isolates with a vancomycin MIC greater than 2 μg/mL should receive daptomycin or linezolid, depending on susceptibility data.

Enterococcal species

Up to 10% of nosocomially acquired bloodstream infections are due to enterococci, and many are related to intravascular catheters.^33,34 Although the risk of endocarditis as a complication of enterococcal CVC-related bloodstream infection is relatively low, estimated at 1.5% in a multicenter prospective study, enterococcal bacteremia lasting longer than 4 days has been independently associated with risk of death.^35,36 These observational data support routine removal of short-term CVCs infected with enterococci.

The choice of antibiotics for enterococcal infections depends on the susceptibility of the isolate. Sixty percent of Enterococcus faecium isolates and 2% of Enterococcus faecalis isolates are vancomycin-resistant, and reports of resistance to newer agents, including linezolid, have been published.^34,37,38 Ampicillin is the preferred antibiotic for treatment of ampicillin-susceptible enterococci. Vancomycin should be used if the pathogen is ampicillin-resistant and vancomycin-susceptible. Enterococci resistant to both ampicillin and vancomycin can be treated with linezolid or daptomycin, based on susceptibility data.

For combination therapy with an aminoglycoside, the data are mixed. Retrospective observational studies have shown no difference in outcomes in uncomplicated enterococcal bacteremia with combination therapy vs monotherapy.^39,40 However, in a large series of patients with enterococcal infections in which the catheter was retained, the combination of gentamicin and ampicillin was more effective than monotherapy.⁴¹

No controlled trial has been done to define the optimal duration of antibiotic therapy for enterococcal bloodstream infections related to short-term CVCs, but the IDSA recommends 7 to 14 days. If catheter salvage is attempted, concurrent antimicrobial lock therapy is recommended based on expert opinion. Catheters should be removed if complications arise (eg, insertion site or pocket infection, suppurative thrombophlebitis, sepsis, endocarditis, persistent bacteremia, metastatic infection). Signs and symptoms of endocarditis, persistent bacteremia, or the presence of a prosthetic heart valve should prompt evaluation with TEE.^42,43

Given the propensity of many gram-negative bacilli to form a biofilm, a number of studies have advocated removing CVCs infected with gram-negative bacilli.^15,16,44 Recent studies examining the role of combination systemic antibiotic therapy and antibiotic lock therapy of gram-negative infections have found high success rates.^45,46

The IDSA recommends routine removal of short-term CVCs infected with gram-negative bacilli and 7 to 14 days of systemic antibiotic therapy based on microbial susceptibility data. Antibiotic options generally include fourth-generation cephalosporins, carbapenems, or a combination beta-lactam and beta-lactamase inhibitor. The first-line treatment for Stenotrophomonas maltophilia and Burkholderia cepacia is trimethoprim-sulfamethoxazole (Bactrim). Extended-spectrum beta-lactamase-producing Klebsiella pneumoniae and Escherichia coli should not be treated with cephalosporins or piperacillin-tazobactam (Zosyn) even if the organisms are susceptible in vitro, as doing so has been associated with poor clinical outcomes.^11,47

There is growing concern over multiple-drug-resistant gram-negative bacilli with carbapenemases that confer resistance to carbapenems. No controlled study has evaluated treatment of multiple-drug-resistant gram-negative bacilli that require therapy with polymyxin (Colistin).

Candida species

The benefit of removing the CVC in the setting of candidemia is supported by six prospective studies.^48–53 Patients with catheter-related bloodstream infections due to Candida species should have the catheter removed. C albicans and azole-susceptible candidal strains can be effectively treated with fluconazole at a dosage of 400 mg daily, continued for 14 days following the first negative blood culture.⁵⁴ Echinocandins as first-line therapy and lipid formulations of amphotericin B (Abelcet) as an alternative are both highly effective for the treatment of Candida species with decreased susceptibility to azoles (eg, C glabrata and C krusei).^55–57

Other gram-positive microorganisms

The isolation of Corynebacterium, Bacillus, and Micrococcus species from a single blood culture does not prove bloodstream infection, and confirmation requires at least two positive results drawn from different sites. CVC infections with these organisms are difficult to treat unless the infected catheter is removed.^58,59

ADDITIONAL RECOMMENDATIONS

Infectious disease consultation should be considered for patients with complicated bloodstream infection related to a short-term CVC. Complicated cases include catheter infections in patients with hemodynamic instability, endocarditis, suppurative thrombophlebitis, persistent bloodstream infection despite 72 hours of appropriate antimicrobial therapy, osteomyelitis, active malignancy, or immunosuppression.

Infectious disease consultation should also be sought for assistance with determining if a patient is a candidate for antibiotic lock therapy; for management, dosing, and course of antibiotic lock therapy; for assistance with antibiotic choice and course for multiple-drug-resistant gram-negative bacilli; and for recommendations on management of infections due to uncommon pathogens (eg, Corynebacterium jeikeium, Chryseobacterium species, Malassezia furfur, and Mycobacterium species).

Vascular catheters are very common in everyday inpatient and, increasingly, outpatient care. Nearly 300 million catheters are estimated to be used annually in the United States, and approximately 3 million of these are central venous catheters (CVCs).¹

Although significant gains have been made in preventing CVC-related bloodstream infections, these infections continue to occur, with estimated rates ranging from 1.3 per 1,000 catheter days on inpatient medical-surgical wards to 5.6 per 1,000 catheter days in intensive care burn units.²

CVCs are classified as either long-term or short-term. Long-term CVCs are surgically implanted or tunneled and used for prolonged chemotherapy, home infusion therapy, or hemodialysis. Short-term CVCs do not require surgical implantation. They are more common than long-term CVCs and account for most CVC-related bloodstream infections. Given the frequency of short-term CVC use, a growing number of health care providers from mid-level practitioners to intensivists are faced with deciding how to manage bloodstream infection related to short-term CVCs.

At baseline, management decisions about bloodstream infections from short-term CVCs can be challenging. Questions that regularly arise include:

• Should a potentially infected catheter be removed?
• Which empiric antibiotic therapy should be started pending a microbiologic diagnosis?
• How should therapy be tailored (eg, antibiotic choice and course and whether to remove or retain the catheter) based on the specific pathogen identified?

Adding to the complexity of these decisions are increasingly resistant microorganisms, heterogeneity of affected patient populations, and variability in the quality and availability of evidence.

This review provides a concise guide to managing bloodstream infections related to short-term CVCs in adults, based on updated guidelines from the Infectious Diseases Society of America (IDSA).³

DEFINITION AND DIAGNOSTIC CRITERIA

We have adapted the following definition and diagnostic criteria from the general definition and diagnostic criteria for catheter-related bloodstream infections proposed by the IDSA.

Bloodstream infection related to a short-term CVC is defined as bacteremia or fungemia in a patient with the CVC in place, clinical manifestations of infection (eg, fever, chills, hypotension), and no apparent source of the bloodstream infection aside from the catheter. At least one of the three diagnostic criteria should be met:

• Cultures of the catheter tip and of the peripheral blood grow the same organism. Catheter tip culture should be quantitative, with more than 10² colony-forming units (cfu) per catheter segment, or semiquantitative, with more than 15 cfu per catheter segment.
• Blood drawn from the catheter lumen grows the same organism as blood drawn from a peripheral vein (or less optimally, a different lumen), but at three times the amount by quantitative culture.
• Blood drawn simultaneously from the catheter lumen and from a peripheral vein (or less optimally, a different lumen) grows the same organism, and growth from the CVC lumen sample is detected (by automated blood culture system) at least 2 hours before growth from the peripheral vein sample.

MANAGING BLOODSTREAM INFECTIONS IN PATIENTS WITH SHORT-TERM CVCs

When to remove a potentially infected short-term CVC

In a nonneutropenic intensive care population, Bouza et al⁴ found that, of 204 episodes of clinically suspected bloodstream infection from a short-term CVC, only 28 (14%) were confirmed to be catheter-related, 27 (13%) were bloodstream infections that were not catheter-related, 36 (18%) involved catheter-tip colonization with negative blood cultures, and the remainder were cases with negative catheter-tip and blood cultures.

Rijnders et al,⁵ in a study of 100 adult medical-surgical intensive care patients with a clinically suspected bloodstream infection related to a short-term CVC, found only three confirmed cases.

A randomized clinical trial comparing early removal of short-term CVCs and watchful waiting in an adult intensive care population with clinically suspected bloodstream infections showed no difference between treatment groups in length of stay in the intensive care unit or in the mortality rate.⁶ This trial included a low-risk subset of adult medical-surgical intensive care patients (ie, immunocompetent, no intravascular foreign body, no evidence of severe sepsis or septic shock, no evidence of infection at the catheter insertion site, no proven bacteremia or fungemia). These results suggest that a similar subset of patients can be safely monitored without catheter removal while being assessed for possible catheter-related bloodstream infection.

Empiric catheter removal vs watchful waiting has not and likely will not be studied in higher-risk populations. In this group, clinical judgment should outweigh any specific management algorithm. In patients who are in shock or who are otherwise hemodynamically unstable, early catheter removal should be a priority; however, in some circumstances the risks of immediate catheter removal (eg, coagulopathy with risk of bleeding diathesis, or lack of site to replace the catheter) may outweigh the potential benefits.

In order of prevalence, the four most common pathogens are coagulase-negative staphylococci, Staphylococcus aureus, Candida species, and enteric gram-negative bacilli.⁷

Gram-positive pathogens. A recent randomized clinical trial comparing vancomycin and linezolid (Zyvox) treatment for CVC-related bloodstream infections showed that 89 (57%) of 157 S aureus isolates and 95 (80%) of 119 coagulase-negative staphylococcal isolates were resistant to methicillin.⁸ Given the prevalence of gram-positive infections and the regularity of methicillin-resistant isolates, vancomycin should be started empirically in cases of suspected bloodstream infection related to short-term CVCs. In institutions where methicillin-resistant S aureus (MRSA) isolates regularly have a vancomycin minimum inhibitory concentration (MIC) of greater than 2 μg/mL, an alternative agent such as daptomycin (Cubicin) should be used.^9,10

Gram-negative pathogens. Infections due to resistant gram-negative pathogens have become more common in the past 10 years.^11,12 Prospective cohort studies have shown that resistant gram-negative infections and inadequate empiric antimicrobial therapy of bloodstream infections independently predict the risk of death.^13,14 Risk factors for resistant gram-negative infections include critical illness, neutropenia, prior antibiotic therapy, and femoral insertion of the CVC.^15–18 Patients with these risk factors should receive empiric antibiotic therapy for gram-negative bacilli.

No randomized controlled trial has been done to guide the choice of empiric gram-negative antibiotic coverage. The initial choice should be based on local antimicrobial patterns and susceptibility data and on the severity of the patient’s illness. Initial options include fourth-generation cephalosporins, carbapenems, or combined beta-lactam and beta-lactamase inhibitors. Patients with neutropenia, severe sepsis, or known multiple-drug-resistant gram-negative bacilli colonization or prior infection should receive empiric combination therapy with two different classes of antibiotics.

Candida. Risk factors for CVC-related bloodstream infections due to Candida species include total parenteral nutrition, prolonged use of broad-spectrum antibiotics, hematologic malignancy, solid organ or bone marrow transplantation, colonization with Candida species at multiple sites, and femoral catheter insertion. Empiric treatment with an echinocandin is recommended for patients with these risk factors. Fluconazole (Diflucan) can be substituted for an echinocandin in patients without azole exposure in the previous 3 months and in settings where the prevalence of Candida krusei and Candida glabrata is low.

PATHOGEN-SPECIFIC MANAGEMENT: RECOMMENDATIONS

Coagulase-negative staphylococci

Most patients with coagulase-negative staphylococcal infections have a benign clinical course.

Although no randomized trial has evaluated different treatment approaches, most experts recommend removing the catheter and giving a short course of antibiotics (ie, 5–7 days). Longer courses of antibiotics may be required for patients with endovascular hardware in place or persistent fever or bacteremia after catheter removal. The IDSA guidelines recommend 5 to 7 days of antibiotic therapy if the catheter is removed, and 10 to 14 days of systemic antibiotic therapy in combination with “antibiotic lock therapy” if the catheter is retained. Antibiotic lock therapy involves instilling a high concentration of an antibiotic to which the organism is susceptible into the catheter lumen and allowing it to dwell.

Not all patients are good candidates for antibiotic lock therapy, and neither are all organisms. In general, patients should be at low risk (immunocompetent, without hardware in place), and organisms should have a low risk of causing metastatic infection.

Staphylococcus lugdunensis can cause endocarditis and metastatic infections similar to those caused by S aureus and so should be managed similarly to S aureus.¹⁹

Staphylococcus aureus

Short-term CVCs infected with S aureus should be removed immediately. Removal of vascular catheters infected with S aureus has been associated with more rapid clinical response and higher cure rates compared with catheter retention.^20–23S aureus bacteremia results in hematogenous complications in 20% to 30% of patients, and failure to remove or a delay in removing the catheter increases the risk of complications.^21,24–27

There are no data from randomized clinical trials on the optimal duration of antibiotic therapy for S aureus bloodstream infections related to short-term CVCs. Traditionally, 4 weeks have been recommended out of concern for the risk of infective endocarditis,^28,29 and the IDSA recommends 4 to 6 weeks unless patients meet certain low-risk criteria.

Factors associated with a higher risk of hematogenous complications include the presence of a retained foreign body, an intravascular prosthetic device, retained catheter, immune suppression, diabetes, persistent bacteremia at 72 hours despite catheter removal and appropriate antibiotics, skin changes consistent with septic emboli, or evidence of endocarditis or suppurative thrombophlebitis on transesophageal echocardiography (TEE) or ultrasonography, respectively.^21,25–27 TEE is superior to transthoracic echocardiography and is most sensitive when performed 5 to 7 days after the onset of bacteremia.^28,30 Patients who have had the catheter removed and who do not have any of these risk factors, and in whom TEE performed 5 to 7 days after the onset of bacteremia is negative, can be considered for a shorter duration of therapy (but a minimum of 14 days).

Patients with catheters colonized with S aureus (ie, those with positive catheter-tip cultures and negative blood cultures) are at risk of subsequent bacteremia. This risk may be reduced with anti-staphylococcal therapy started within 24 hours of catheter removal.^31,32 Therapy should be continued for 5 to 7 days, and patients should be closely monitored for signs or symptoms of ongoing infection.

Oxacillin or nafcillin should be the first-line therapy for susceptible S aureus isolates. Vancomycin should be used to treat MRSA. Patients with MRSA isolates with a vancomycin MIC greater than 2 μg/mL should receive daptomycin or linezolid, depending on susceptibility data.

Enterococcal species

Up to 10% of nosocomially acquired bloodstream infections are due to enterococci, and many are related to intravascular catheters.^33,34 Although the risk of endocarditis as a complication of enterococcal CVC-related bloodstream infection is relatively low, estimated at 1.5% in a multicenter prospective study, enterococcal bacteremia lasting longer than 4 days has been independently associated with risk of death.^35,36 These observational data support routine removal of short-term CVCs infected with enterococci.

The choice of antibiotics for enterococcal infections depends on the susceptibility of the isolate. Sixty percent of Enterococcus faecium isolates and 2% of Enterococcus faecalis isolates are vancomycin-resistant, and reports of resistance to newer agents, including linezolid, have been published.^34,37,38 Ampicillin is the preferred antibiotic for treatment of ampicillin-susceptible enterococci. Vancomycin should be used if the pathogen is ampicillin-resistant and vancomycin-susceptible. Enterococci resistant to both ampicillin and vancomycin can be treated with linezolid or daptomycin, based on susceptibility data.

For combination therapy with an aminoglycoside, the data are mixed. Retrospective observational studies have shown no difference in outcomes in uncomplicated enterococcal bacteremia with combination therapy vs monotherapy.^39,40 However, in a large series of patients with enterococcal infections in which the catheter was retained, the combination of gentamicin and ampicillin was more effective than monotherapy.⁴¹

No controlled trial has been done to define the optimal duration of antibiotic therapy for enterococcal bloodstream infections related to short-term CVCs, but the IDSA recommends 7 to 14 days. If catheter salvage is attempted, concurrent antimicrobial lock therapy is recommended based on expert opinion. Catheters should be removed if complications arise (eg, insertion site or pocket infection, suppurative thrombophlebitis, sepsis, endocarditis, persistent bacteremia, metastatic infection). Signs and symptoms of endocarditis, persistent bacteremia, or the presence of a prosthetic heart valve should prompt evaluation with TEE.^42,43

Given the propensity of many gram-negative bacilli to form a biofilm, a number of studies have advocated removing CVCs infected with gram-negative bacilli.^15,16,44 Recent studies examining the role of combination systemic antibiotic therapy and antibiotic lock therapy of gram-negative infections have found high success rates.^45,46

The IDSA recommends routine removal of short-term CVCs infected with gram-negative bacilli and 7 to 14 days of systemic antibiotic therapy based on microbial susceptibility data. Antibiotic options generally include fourth-generation cephalosporins, carbapenems, or a combination beta-lactam and beta-lactamase inhibitor. The first-line treatment for Stenotrophomonas maltophilia and Burkholderia cepacia is trimethoprim-sulfamethoxazole (Bactrim). Extended-spectrum beta-lactamase-producing Klebsiella pneumoniae and Escherichia coli should not be treated with cephalosporins or piperacillin-tazobactam (Zosyn) even if the organisms are susceptible in vitro, as doing so has been associated with poor clinical outcomes.^11,47

There is growing concern over multiple-drug-resistant gram-negative bacilli with carbapenemases that confer resistance to carbapenems. No controlled study has evaluated treatment of multiple-drug-resistant gram-negative bacilli that require therapy with polymyxin (Colistin).

Candida species

The benefit of removing the CVC in the setting of candidemia is supported by six prospective studies.^48–53 Patients with catheter-related bloodstream infections due to Candida species should have the catheter removed. C albicans and azole-susceptible candidal strains can be effectively treated with fluconazole at a dosage of 400 mg daily, continued for 14 days following the first negative blood culture.⁵⁴ Echinocandins as first-line therapy and lipid formulations of amphotericin B (Abelcet) as an alternative are both highly effective for the treatment of Candida species with decreased susceptibility to azoles (eg, C glabrata and C krusei).^55–57

Other gram-positive microorganisms

The isolation of Corynebacterium, Bacillus, and Micrococcus species from a single blood culture does not prove bloodstream infection, and confirmation requires at least two positive results drawn from different sites. CVC infections with these organisms are difficult to treat unless the infected catheter is removed.^58,59

ADDITIONAL RECOMMENDATIONS

Infectious disease consultation should be considered for patients with complicated bloodstream infection related to a short-term CVC. Complicated cases include catheter infections in patients with hemodynamic instability, endocarditis, suppurative thrombophlebitis, persistent bloodstream infection despite 72 hours of appropriate antimicrobial therapy, osteomyelitis, active malignancy, or immunosuppression.

Infectious disease consultation should also be sought for assistance with determining if a patient is a candidate for antibiotic lock therapy; for management, dosing, and course of antibiotic lock therapy; for assistance with antibiotic choice and course for multiple-drug-resistant gram-negative bacilli; and for recommendations on management of infections due to uncommon pathogens (eg, Corynebacterium jeikeium, Chryseobacterium species, Malassezia furfur, and Mycobacterium species).

Vascular catheters are very common in everyday inpatient and, increasingly, outpatient care. Nearly 300 million catheters are estimated to be used annually in the United States, and approximately 3 million of these are central venous catheters (CVCs).¹

Although significant gains have been made in preventing CVC-related bloodstream infections, these infections continue to occur, with estimated rates ranging from 1.3 per 1,000 catheter days on inpatient medical-surgical wards to 5.6 per 1,000 catheter days in intensive care burn units.²

CVCs are classified as either long-term or short-term. Long-term CVCs are surgically implanted or tunneled and used for prolonged chemotherapy, home infusion therapy, or hemodialysis. Short-term CVCs do not require surgical implantation. They are more common than long-term CVCs and account for most CVC-related bloodstream infections. Given the frequency of short-term CVC use, a growing number of health care providers from mid-level practitioners to intensivists are faced with deciding how to manage bloodstream infection related to short-term CVCs.

At baseline, management decisions about bloodstream infections from short-term CVCs can be challenging. Questions that regularly arise include:

• Should a potentially infected catheter be removed?
• Which empiric antibiotic therapy should be started pending a microbiologic diagnosis?
• How should therapy be tailored (eg, antibiotic choice and course and whether to remove or retain the catheter) based on the specific pathogen identified?

Adding to the complexity of these decisions are increasingly resistant microorganisms, heterogeneity of affected patient populations, and variability in the quality and availability of evidence.

This review provides a concise guide to managing bloodstream infections related to short-term CVCs in adults, based on updated guidelines from the Infectious Diseases Society of America (IDSA).³

DEFINITION AND DIAGNOSTIC CRITERIA

We have adapted the following definition and diagnostic criteria from the general definition and diagnostic criteria for catheter-related bloodstream infections proposed by the IDSA.

Bloodstream infection related to a short-term CVC is defined as bacteremia or fungemia in a patient with the CVC in place, clinical manifestations of infection (eg, fever, chills, hypotension), and no apparent source of the bloodstream infection aside from the catheter. At least one of the three diagnostic criteria should be met:

• Cultures of the catheter tip and of the peripheral blood grow the same organism. Catheter tip culture should be quantitative, with more than 10² colony-forming units (cfu) per catheter segment, or semiquantitative, with more than 15 cfu per catheter segment.
• Blood drawn from the catheter lumen grows the same organism as blood drawn from a peripheral vein (or less optimally, a different lumen), but at three times the amount by quantitative culture.
• Blood drawn simultaneously from the catheter lumen and from a peripheral vein (or less optimally, a different lumen) grows the same organism, and growth from the CVC lumen sample is detected (by automated blood culture system) at least 2 hours before growth from the peripheral vein sample.

MANAGING BLOODSTREAM INFECTIONS IN PATIENTS WITH SHORT-TERM CVCs

When to remove a potentially infected short-term CVC

In a nonneutropenic intensive care population, Bouza et al⁴ found that, of 204 episodes of clinically suspected bloodstream infection from a short-term CVC, only 28 (14%) were confirmed to be catheter-related, 27 (13%) were bloodstream infections that were not catheter-related, 36 (18%) involved catheter-tip colonization with negative blood cultures, and the remainder were cases with negative catheter-tip and blood cultures.

Rijnders et al,⁵ in a study of 100 adult medical-surgical intensive care patients with a clinically suspected bloodstream infection related to a short-term CVC, found only three confirmed cases.

A randomized clinical trial comparing early removal of short-term CVCs and watchful waiting in an adult intensive care population with clinically suspected bloodstream infections showed no difference between treatment groups in length of stay in the intensive care unit or in the mortality rate.⁶ This trial included a low-risk subset of adult medical-surgical intensive care patients (ie, immunocompetent, no intravascular foreign body, no evidence of severe sepsis or septic shock, no evidence of infection at the catheter insertion site, no proven bacteremia or fungemia). These results suggest that a similar subset of patients can be safely monitored without catheter removal while being assessed for possible catheter-related bloodstream infection.

Empiric catheter removal vs watchful waiting has not and likely will not be studied in higher-risk populations. In this group, clinical judgment should outweigh any specific management algorithm. In patients who are in shock or who are otherwise hemodynamically unstable, early catheter removal should be a priority; however, in some circumstances the risks of immediate catheter removal (eg, coagulopathy with risk of bleeding diathesis, or lack of site to replace the catheter) may outweigh the potential benefits.

In order of prevalence, the four most common pathogens are coagulase-negative staphylococci, Staphylococcus aureus, Candida species, and enteric gram-negative bacilli.⁷

Gram-positive pathogens. A recent randomized clinical trial comparing vancomycin and linezolid (Zyvox) treatment for CVC-related bloodstream infections showed that 89 (57%) of 157 S aureus isolates and 95 (80%) of 119 coagulase-negative staphylococcal isolates were resistant to methicillin.⁸ Given the prevalence of gram-positive infections and the regularity of methicillin-resistant isolates, vancomycin should be started empirically in cases of suspected bloodstream infection related to short-term CVCs. In institutions where methicillin-resistant S aureus (MRSA) isolates regularly have a vancomycin minimum inhibitory concentration (MIC) of greater than 2 μg/mL, an alternative agent such as daptomycin (Cubicin) should be used.^9,10

Gram-negative pathogens. Infections due to resistant gram-negative pathogens have become more common in the past 10 years.^11,12 Prospective cohort studies have shown that resistant gram-negative infections and inadequate empiric antimicrobial therapy of bloodstream infections independently predict the risk of death.^13,14 Risk factors for resistant gram-negative infections include critical illness, neutropenia, prior antibiotic therapy, and femoral insertion of the CVC.^15–18 Patients with these risk factors should receive empiric antibiotic therapy for gram-negative bacilli.

No randomized controlled trial has been done to guide the choice of empiric gram-negative antibiotic coverage. The initial choice should be based on local antimicrobial patterns and susceptibility data and on the severity of the patient’s illness. Initial options include fourth-generation cephalosporins, carbapenems, or combined beta-lactam and beta-lactamase inhibitors. Patients with neutropenia, severe sepsis, or known multiple-drug-resistant gram-negative bacilli colonization or prior infection should receive empiric combination therapy with two different classes of antibiotics.

Candida. Risk factors for CVC-related bloodstream infections due to Candida species include total parenteral nutrition, prolonged use of broad-spectrum antibiotics, hematologic malignancy, solid organ or bone marrow transplantation, colonization with Candida species at multiple sites, and femoral catheter insertion. Empiric treatment with an echinocandin is recommended for patients with these risk factors. Fluconazole (Diflucan) can be substituted for an echinocandin in patients without azole exposure in the previous 3 months and in settings where the prevalence of Candida krusei and Candida glabrata is low.

PATHOGEN-SPECIFIC MANAGEMENT: RECOMMENDATIONS

Coagulase-negative staphylococci

Most patients with coagulase-negative staphylococcal infections have a benign clinical course.

Although no randomized trial has evaluated different treatment approaches, most experts recommend removing the catheter and giving a short course of antibiotics (ie, 5–7 days). Longer courses of antibiotics may be required for patients with endovascular hardware in place or persistent fever or bacteremia after catheter removal. The IDSA guidelines recommend 5 to 7 days of antibiotic therapy if the catheter is removed, and 10 to 14 days of systemic antibiotic therapy in combination with “antibiotic lock therapy” if the catheter is retained. Antibiotic lock therapy involves instilling a high concentration of an antibiotic to which the organism is susceptible into the catheter lumen and allowing it to dwell.

Not all patients are good candidates for antibiotic lock therapy, and neither are all organisms. In general, patients should be at low risk (immunocompetent, without hardware in place), and organisms should have a low risk of causing metastatic infection.

Staphylococcus lugdunensis can cause endocarditis and metastatic infections similar to those caused by S aureus and so should be managed similarly to S aureus.¹⁹

Staphylococcus aureus

Short-term CVCs infected with S aureus should be removed immediately. Removal of vascular catheters infected with S aureus has been associated with more rapid clinical response and higher cure rates compared with catheter retention.^20–23S aureus bacteremia results in hematogenous complications in 20% to 30% of patients, and failure to remove or a delay in removing the catheter increases the risk of complications.^21,24–27

There are no data from randomized clinical trials on the optimal duration of antibiotic therapy for S aureus bloodstream infections related to short-term CVCs. Traditionally, 4 weeks have been recommended out of concern for the risk of infective endocarditis,^28,29 and the IDSA recommends 4 to 6 weeks unless patients meet certain low-risk criteria.

Factors associated with a higher risk of hematogenous complications include the presence of a retained foreign body, an intravascular prosthetic device, retained catheter, immune suppression, diabetes, persistent bacteremia at 72 hours despite catheter removal and appropriate antibiotics, skin changes consistent with septic emboli, or evidence of endocarditis or suppurative thrombophlebitis on transesophageal echocardiography (TEE) or ultrasonography, respectively.^21,25–27 TEE is superior to transthoracic echocardiography and is most sensitive when performed 5 to 7 days after the onset of bacteremia.^28,30 Patients who have had the catheter removed and who do not have any of these risk factors, and in whom TEE performed 5 to 7 days after the onset of bacteremia is negative, can be considered for a shorter duration of therapy (but a minimum of 14 days).

Patients with catheters colonized with S aureus (ie, those with positive catheter-tip cultures and negative blood cultures) are at risk of subsequent bacteremia. This risk may be reduced with anti-staphylococcal therapy started within 24 hours of catheter removal.^31,32 Therapy should be continued for 5 to 7 days, and patients should be closely monitored for signs or symptoms of ongoing infection.

Oxacillin or nafcillin should be the first-line therapy for susceptible S aureus isolates. Vancomycin should be used to treat MRSA. Patients with MRSA isolates with a vancomycin MIC greater than 2 μg/mL should receive daptomycin or linezolid, depending on susceptibility data.

Enterococcal species

Up to 10% of nosocomially acquired bloodstream infections are due to enterococci, and many are related to intravascular catheters.^33,34 Although the risk of endocarditis as a complication of enterococcal CVC-related bloodstream infection is relatively low, estimated at 1.5% in a multicenter prospective study, enterococcal bacteremia lasting longer than 4 days has been independently associated with risk of death.^35,36 These observational data support routine removal of short-term CVCs infected with enterococci.

The choice of antibiotics for enterococcal infections depends on the susceptibility of the isolate. Sixty percent of Enterococcus faecium isolates and 2% of Enterococcus faecalis isolates are vancomycin-resistant, and reports of resistance to newer agents, including linezolid, have been published.^34,37,38 Ampicillin is the preferred antibiotic for treatment of ampicillin-susceptible enterococci. Vancomycin should be used if the pathogen is ampicillin-resistant and vancomycin-susceptible. Enterococci resistant to both ampicillin and vancomycin can be treated with linezolid or daptomycin, based on susceptibility data.

For combination therapy with an aminoglycoside, the data are mixed. Retrospective observational studies have shown no difference in outcomes in uncomplicated enterococcal bacteremia with combination therapy vs monotherapy.^39,40 However, in a large series of patients with enterococcal infections in which the catheter was retained, the combination of gentamicin and ampicillin was more effective than monotherapy.⁴¹

No controlled trial has been done to define the optimal duration of antibiotic therapy for enterococcal bloodstream infections related to short-term CVCs, but the IDSA recommends 7 to 14 days. If catheter salvage is attempted, concurrent antimicrobial lock therapy is recommended based on expert opinion. Catheters should be removed if complications arise (eg, insertion site or pocket infection, suppurative thrombophlebitis, sepsis, endocarditis, persistent bacteremia, metastatic infection). Signs and symptoms of endocarditis, persistent bacteremia, or the presence of a prosthetic heart valve should prompt evaluation with TEE.^42,43

Given the propensity of many gram-negative bacilli to form a biofilm, a number of studies have advocated removing CVCs infected with gram-negative bacilli.^15,16,44 Recent studies examining the role of combination systemic antibiotic therapy and antibiotic lock therapy of gram-negative infections have found high success rates.^45,46

The IDSA recommends routine removal of short-term CVCs infected with gram-negative bacilli and 7 to 14 days of systemic antibiotic therapy based on microbial susceptibility data. Antibiotic options generally include fourth-generation cephalosporins, carbapenems, or a combination beta-lactam and beta-lactamase inhibitor. The first-line treatment for Stenotrophomonas maltophilia and Burkholderia cepacia is trimethoprim-sulfamethoxazole (Bactrim). Extended-spectrum beta-lactamase-producing Klebsiella pneumoniae and Escherichia coli should not be treated with cephalosporins or piperacillin-tazobactam (Zosyn) even if the organisms are susceptible in vitro, as doing so has been associated with poor clinical outcomes.^11,47

There is growing concern over multiple-drug-resistant gram-negative bacilli with carbapenemases that confer resistance to carbapenems. No controlled study has evaluated treatment of multiple-drug-resistant gram-negative bacilli that require therapy with polymyxin (Colistin).

Candida species

The benefit of removing the CVC in the setting of candidemia is supported by six prospective studies.^48–53 Patients with catheter-related bloodstream infections due to Candida species should have the catheter removed. C albicans and azole-susceptible candidal strains can be effectively treated with fluconazole at a dosage of 400 mg daily, continued for 14 days following the first negative blood culture.⁵⁴ Echinocandins as first-line therapy and lipid formulations of amphotericin B (Abelcet) as an alternative are both highly effective for the treatment of Candida species with decreased susceptibility to azoles (eg, C glabrata and C krusei).^55–57

Other gram-positive microorganisms

The isolation of Corynebacterium, Bacillus, and Micrococcus species from a single blood culture does not prove bloodstream infection, and confirmation requires at least two positive results drawn from different sites. CVC infections with these organisms are difficult to treat unless the infected catheter is removed.^58,59

ADDITIONAL RECOMMENDATIONS

Infectious disease consultation should be considered for patients with complicated bloodstream infection related to a short-term CVC. Complicated cases include catheter infections in patients with hemodynamic instability, endocarditis, suppurative thrombophlebitis, persistent bloodstream infection despite 72 hours of appropriate antimicrobial therapy, osteomyelitis, active malignancy, or immunosuppression.

Infectious disease consultation should also be sought for assistance with determining if a patient is a candidate for antibiotic lock therapy; for management, dosing, and course of antibiotic lock therapy; for assistance with antibiotic choice and course for multiple-drug-resistant gram-negative bacilli; and for recommendations on management of infections due to uncommon pathogens (eg, Corynebacterium jeikeium, Chryseobacterium species, Malassezia furfur, and Mycobacterium species).