Tyler Stevens, MD

The treatment of early esophageal, gastric, and colorectal cancer is changing.¹ For many years, surgery was the mainstay of treatment for early-stage gastrointestinal cancer. Unfortunately, surgery leads to significant loss of function of the organ, resulting in increased morbidity and decreased quality of life.²

Endoscopic techniques, particularly endoscopic mucosal resection (EMR) and endoscopic submucosal dissection (ESD), have been developed and are widely used in Japan, where gastrointestinal cancer is more common than in the West. This article reviews the indications, complications, and outcomes of ESD for early gastrointestinal neoplasms, so that readers will recognize the subset of patients who would benefit from ESD in a Western setting.

ENDOSCOPIC MUCOSAL RESECTION AND SUBMUCOSAL DISSECTION

Since the first therapeutic polypectomy was performed in Japan in 1974, several endoscopic techniques for tumor resection have been developed.³

EMR, one of the most successful and widely used techniques, involves elevating the lesion either with submucosal injection of a solution or with cap suction, and then removing it with a snare.⁴ Most lesions smaller than 20 mm can be removed in one piece (en bloc).⁵ Larger lesions are removed in multiple pieces (ie, piecemeal). Unfortunately, some fibrotic lesions, which are usually difficult to lift, cannot be completely removed by EMR.

ESD was first performed in the late 1990s with the aim of overcoming the limitations of EMR in resecting large or fibrotic tumors en bloc.^6,7 Since then, ESD technique has been standardized and training centers have been created, especially in Asia, where it is widely used for treatment of early gastric cancer.^3,8–10 Since 2012 it has been covered by the Japanese National Health Insurance for treatment of early gastric cancer, and since 2014 for treatment of colorectal malignant tumors measuring 2 to 5 cm.¹¹

Adoption of ESD has been slow in Western countries, where many patients are still referred for surgery or undergo EMR for removal of superficial neoplasms. Reasons for this slow adoption are that gastric cancer is much less common in Western countries, and also that ESD demands a high level of technical skill, is difficult to learn, and is expensive.^3,12,13 However, small groups of Western endoscopists have become interested and are advocating it, first studying it on their own and then training in a Japanese center and learning from experts performing the procedure.

Therefore, in a Western setting, ESD should be performed in specialized endoscopy centers and offered to selected patients.¹  

CANDIDATES SHOULD HAVE EARLY-STAGE, SUPERFICIAL TUMORS

Ideal candidates for endoscopic resection are patients who have early cancer with a negligible risk of lymph node metastasis, such as cancer limited to the mucosa (stage T1a).⁷ Therefore, to determine the best treatment for a patient with a newly diagnosed gastrointestinal neoplasm, it is mandatory to estimate the depth of invasion.

The depth of invasion is directly correlated with lymph node involvement, which is ultimately the main predictive factor for long-term adverse outcomes of gastrointestinal tumors.^4,14–17 Accurate multidisciplinary preprocedure estimations are mandatory, as incorrect evaluations may result in inappropriate therapy and residual cancer.¹⁸

Other factors that have been used to predict lymph node involvement include tumor size, macroscopic appearance, histologic differentiation, and lymphatic and vascular involvement.¹⁹ Some of these factors can be assessed by special endoscopic techniques (chromoendoscopy and narrow-band imaging with magnifying endoscopy) that allow accurate real-time estimation of the depth of invasion of the lesion.^5,17,20–27 Evaluation of microsurface and microvascular arrangements is especially useful for determining the feasibility of ESD in gastric tumors, evaluation of intracapillary loops is useful in esophageal lesions, and assessment of mucosal pit patterns is useful for colorectal lesions.^21–29

Endoscopic ultrasonography is another tool that has been used to estimate the depth of the tumor. Although it can differentiate between definite intramucosal and definite submucosal invasive cancers, its ability to confirm minute submucosal invasion is limited. Its use as the sole tumor staging modality is not encouraged, and it should always be used in conjunction with endoscopic evaluation.¹⁸

Though the aforementioned factors help stratify patients, pathologic staging is the best predictor of lymph node metastasis. ESD provides adequate specimens for accurate pathologic evaluation, as it removes lesions en bloc.³⁰

All patients found to have risk factors for lymph node metastasis on endoscopic, ultrasonographic, or pathologic analysis should be referred for surgical evaluation.^9,19,31,32

ENDOSCOPIC SUBMUCOSAL DISSECTION

Before the procedure, the patient’s physicians need to do the following:

Determine the best type of intervention (EMR, ESD, ablation, surgery) for the specific lesion.³ A multidisciplinary approach is encouraged, with involvement of the internist, gastroenterologist, and surgeon.

Plan for anesthesia, additional consultations, pre- and postprocedural hospital admission, and need for special equipment.³³

During the procedure

Define the lateral extent of the lesion using magnification chromoendoscopy or narrow-band imaging. In the stomach, a biopsy sample should be taken from the worst-looking segment and from normal-looking mucosa. Multiple biopsies should be avoided to prevent subsequent fibrosis.³³ In the colon, biopsy should be avoided.³⁴

Identify and circumferentially mark the target lesion. Cautery or argon plasma coagulation can be used for making markings at a distance of 5 to 10 mm from the edges.³³ This is done to recognize the borders of the lesion, because they can become distorted after submucosal injection.¹⁴ This step is unnecessary in colorectal cases, as tumor margins can be adequately visualized after chromoendoscopy.^16,35

Lift the lesion by injecting saline, 0.5% hyaluronate, or glycerin to create a submucosal fluid cushion.^19,33

Perform a circumferential incision lateral to the mucosal margins to allow for a normal tissue margin.³³ Partial incision is performed for esophageal and colorectal ESD to avoid fluid leakage from the submucosal layer, achieving a sustained submucosal lift and safer dissection.¹⁶

Submucosal dissection. The submucosal layer is dissected with an electrocautery knife until the lesion is completely removed. Dissection should be done carefully to keep the submucosal plane.³³ Hemoclips or hemostat forceps can be used to control visible bleeding. The resected specimen is then stretched and fixed to a board using small pins for further histopathologic evaluation.³⁵

Postprocedural monitoring.  All patients should be admitted for overnight observation. Those who undergo gastric ESD should receive high-dose acid suppression, and the next day they can be started on a liquid diet.¹⁹

STOMACH CANCER

Indications for ESD for stomach cancer in the East

The incidence of gastric cancer is higher in Japan and Korea, where widespread screening programs have led to early identification and early treatment of this disease.³⁶

Pathology studies³⁷ of samples from patients with gastric cancer identified the following as risk factors for lymph node metastasis, which would make ESD unsuitable:

• Undifferentiated type
• Tumors larger than 2 cm
• Lymphatic or venous involvement
• Submucosal invasion
• Ulcerative change.

Based on these findings, the situations in which there was no risk of lymph node involvement (ie, when none of the above factors are present) were accepted as absolute indications for endoscopic resection of early gastric cancer.³⁸ Further histologic studies identified a subset of patients with lesions with very low risk of lymph node metastasis, which outweighed the risk of surgery. Based on these findings, expanded criteria for gastric ESD were proposed,^39,40 and the Japanese gastric cancer treatment guidelines now include these expanded preoperative indications^9,17 (Table 1).

Based on information from the Japanese Gastric Cancer Association, reference 9.

The Japanese Gastric Cancer Association has proposed a treatment algorithm based on the histopathologic evaluation after resection (Figure 2).⁹

Outcomes

In the largest series of patients who underwent curative ESD for early gastric cancer, the 5-year survival rate was 92.6%, the 5-year disease-specific survival rate was 99.9%, and the 5-year relative survival rate was 105%.⁴¹

Similarly, in a Japanese population-based survival analysis, the relative 5-year survival rate for localized gastric cancer was 94.4%.⁴² Rates of en bloc resection and complete resection with ESD are higher than those with EMR, resulting in a lower risk of local recurrence in selected patients who undergo ESD.^8,43,44

Although rare, local recurrence after curative gastric ESD has been reported.⁴⁵ The annual incidence of local recurrence has been estimated to be 0.84%.⁴⁶

ESD entails a shorter hospital stay and requires fewer resources than surgery, resulting in lower medical costs (Table 2).⁴⁴ Additionally, as endoscopic resection is associated with less morbidity, fewer procedure-related adverse events, and fewer complications, ESD could be used as the standard treatment for early gastric cancer.^47,48

The Western perspective on endoscopic submucosal dissection for gastric cancer

Since the prevalence of gastric cancer in Western countries is significantly lower than in Japan and Korea, local data and experience are scarce. However, experts performing ESD in the West have adopted the indications of the Japan Gastroenterological Endoscopy Society. The European Society of Gastrointestinal Endoscopy recommends ESD for excision of most superficial gastric neoplasms, with EMR being preferred only in lesions smaller than 15 mm, Paris classification 0 or IIA.^5,32

Patients with gastric lesions measuring 15 mm or larger should undergo high-quality endoscopy, preferably chromoendoscopy, to evaluate the mucosal patterns and determine the depth of invasion. If superficial involvement is confirmed, other imaging techniques are not routinely recommended.⁵ A surgery consult is also recommended.

ESOPHAGEAL CANCER

Indications for ESD for esophageal cancer in the East

Due to the success of ESD for early gastric cancer, this technique is now also used for superficial esophageal neoplasms.^19,49 It should be done in a specialized center, as it is more technically difficult than gastric ESD: the esophageal lumen is narrow, the wall is thin, and the esophagus moves with respiration and heartbeat.⁵⁰ A multidisciplinary approach including an endoscopist, a surgeon, and a pathologist is highly recommended for evaluation and treatment.

EMR is preferred for removal of mucosal cancer, in view of its safety profile and success rates. ESD can be considered in cases of lesions larger than 15 mm, poorly lifting tumors, and those with the possibility of submucosal invasion (Table 3).^5,45,49,51

Circumference involvement is critical when determining eligible candidates, as a defect involving more than three-fourths of the esophageal circumference can lead to esophageal strictures.⁵² Controlled prospective studies have shown promising results from giving intralesional and oral steroids to prevent stricture after ESD, which could potentially overcome this size limitation.^53,54

Outcomes for esophageal cancer

ESD has been shown to be safe and effective, achieving en bloc resection in 85% to 100% of patients.^19,51 Its advantages over EMR include en bloc resection, complete resection, and high curative rates, resulting in higher recurrence-free survival.^2,55,56 Although the incidence of complications such as bleeding, perforation, and stricture formation are higher with ESD, patients usually recover uneventfully.^2,19,20

ESD in the esophagus: The Western perspective

As data on the efficacy of EMR vs ESD for the treatment of Barrett esophagus with adenocarcinoma are limited, EMR is the gold standard endoscopic technique for removal of visible esophageal dysplastic lesions.^5,51,57 ESD can be considered for tumors larger than 15 mm, for poorly lifting lesions, and if there is suspicion of submucosal invasion.⁵

Patients should be evaluated by an experienced endoscopist, using an advanced imaging technique such as narrow-band imaging or chromoendoscopy. If suspicious features are found, endoscopic ultrasonography should be considered to confirm submucosal invasion or lymph node involvement.⁵

Indications for ESD for colorectal cancer in the East

Colon cancer is one of the leading causes of cancer-related deaths worldwide.⁵⁸ Since ESD has been found to be effective and safe in treating gastric cancer, it has also been used to remove large colorectal tumors.⁵⁹ However, ESD is not universally accepted in the treatment of colorectal neoplasms due to its greater technical difficulty, longer procedural time, and higher risk of perforating the thinner colonic wall compared with EMR.^21,60

Outcomes for colorectal cancer

Tumor size of 50 mm or larger is a risk factor for complications, while a high procedure volume at the center is a protective factor.⁶⁰

Endoscopic treatment of colorectal cancer: The Western perspective

EMR is the gold standard for removal of superficial colorectal lesions. However, ESD can be considered if there is suspicion of superficial submucosal invasion, especially for lesions larger than 20 mm that cannot be resected en bloc by EMR.³² ESD can also be used for fibrotic lesions not amenable to complete EMR removal, or as a salvage procedure after recurrence after EMR.⁶⁷ Proper selection of cases is critical.¹

Patients who have a superficial colonic lesion should be evaluated by means of high-definition endoscopy and chromoendoscopy to assess the mucosal pattern and establish feasibility of endoscopic resection. If submucosal invasion is suspected, staging with endoscopic ultrasonography or magnetic resonance imaging should be considered.⁵

FOLLOW-UP AFTER ESD

Endoscopic surveillance after the procedure is recommended, given the persistent risk of metachronous cancer after curative ESD due to its organ-sparing quality.⁶⁸ Surveillance endoscopy aims to achieve early detection and subsequent endoscopic resection of metachronous lesions.

Histopathologic evaluation assessing the presence of malignant cells in the margins of a resected sample is mandatory for determining the next step in treatment. If margins are negative, follow-up endoscopy can be done every 6 to 12 months. If margins are positive, the approach includes surgery, reattempting ESD or endoscopic surveillance in 3 or 6 months.^3,32 Although the surveillance strategy varies according to individual risk of metachronous cancer, it should be continued indefinitely.⁶⁸

COMPLICATIONS OF ESD

The most common procedure-related complications of ESD are bleeding, perforation, and stricture. Most intraprocedural adverse events can be managed endoscopically.⁶⁹

Bleeding

Most bleeding occurs during the procedure or early after it and can be controlled with electrocautery.^49,69 No episodes of massive bleeding, defined as causing clinical symptoms and requiring transfusion or surgery, have been reported.^20,43,55

In gastric ESD, delayed bleeding rates have ranged from 0 to 15.6%.⁶⁹ Bleeding may be prevented with endoscopic coagulation of visible vessels after dissection has been completed and by proton pump inhibitor therapy.^70,71 Excessive coagulation should be avoided to lower the risk of perforation.³³

In colorectal ESD the bleeding rate has been reported to be 2.2%; applying coagulation to an area where a blood vessel is suspected before cutting (precoagulation) may prevent subsequent bleeding.²¹

Perforation

For gastric ESD, perforation rates range from 1.2% to 5.2%.⁶⁹ Esophageal perforation rates can be up to 4%.⁴⁹ In colorectal ESD, perforation rates have been reported to be 1.6% to 6.6%.^60,72

Although most of the cases were successfully managed with conservative treatment, some required emergency surgery.^60,73

Strictures

In a case series of 532 patients undergoing gastric ESD, stricture was reported in 5 patients, all of whom presented with obstructive symptoms.⁷⁴ Risk factors for post-ESD gastric stenosis are a mucosal defect with a circumferential extent of more than three-fourths or a longitudinal extent of more than 5 cm.⁷⁵

Strictures are common after esophageal ESD, with rates ranging from 2% to 26%. The risk is higher when longer segments are removed or circumferential resection is performed. As previously mentioned, this complication may be reduced with ingestion or injection of steroids  after the procedure.^53,54

Surprisingly, ESD of large colorectal lesions involving more than three-fourths of the circumference of the rectum is rarely complicated by stenosis.⁷⁶

ESD requires a high level of technical skill, is time-consuming, and has a higher rate of complications than conventional endoscopic resection. A standardized ESD training system is needed, as the procedure is more difficult than EMR. Training in porcine models has been shown to confer competency in ESD in a Western setting.^13,16,33

Colorectal ESD is an even more challenging procedure, given the potential for complications related to its anatomy. Training centers in Japan usually have their trainees first master gastric ESD, then assist in more than 20 colorectal ESDs conducted by experienced endoscopists, and accomplish 30 cases before performing the procedure safely and independently.

As the incidence of gastric cancer is low in Western countries, trainees may also begin with lower rectal lesions, which are easier to remove.⁷⁷ Incorporation of ESD in the West would require a clear treatment algorithm. It is a complex procedure, with higher rates of complications, a prolonged learning curve, and prolonged procedure time. Therefore, it should be performed in specialized centers and under the special situations discussed here to ensure that the benefits for the patients outweigh the risks.

VALUE OF ENDOSCOPIC SUBMUCOSAL DISSECTION

The optimal method for resecting gastrointestinal neoplasms should be safe, cost-effective, and quick and should also completely remove the lesion. The best treatment strategy takes into account the characteristics of the lesion and the comorbidities and wishes of the patient. Internists should be aware of the multiple options available to achieve the best outcome for the patient.¹

Endoscopic resection of superficial gastrointestinal neoplasms, including EMR and ESD, has been a subject of increasing interest due to its minimally invasive and potentially curative character. However, cancer can recur after endoscopic resection because the procedure is organ-sparing.

ESD allows resection of early gastrointestinal tumors with a minimally invasive technique. It can achieve higher curative resection rates and lower recurrence rates compared with EMR. Compared with surgery, ESD leads to less morbidity, fewer procedure-related complications, and lower medical costs. Indications should be rigorously followed to achieve successful treatments in selected patients.

Multiple variables have to be taken into account when deciding which treatment is best, such as tumor characteristics, the patient’s baseline condition, physician expertise, and hospital resources.⁴⁸ Less-invasive treatments may improve the prognosis of patients. No matter the approach, patients should be treated in specialized treatment centers.

Internal medicine physicians should be aware of the advances in treatments for early gastrointestinal cancer so appropriate options can be considered.

The treatment of early esophageal, gastric, and colorectal cancer is changing.¹ For many years, surgery was the mainstay of treatment for early-stage gastrointestinal cancer. Unfortunately, surgery leads to significant loss of function of the organ, resulting in increased morbidity and decreased quality of life.²

Endoscopic techniques, particularly endoscopic mucosal resection (EMR) and endoscopic submucosal dissection (ESD), have been developed and are widely used in Japan, where gastrointestinal cancer is more common than in the West. This article reviews the indications, complications, and outcomes of ESD for early gastrointestinal neoplasms, so that readers will recognize the subset of patients who would benefit from ESD in a Western setting.

ENDOSCOPIC MUCOSAL RESECTION AND SUBMUCOSAL DISSECTION

Since the first therapeutic polypectomy was performed in Japan in 1974, several endoscopic techniques for tumor resection have been developed.³

EMR, one of the most successful and widely used techniques, involves elevating the lesion either with submucosal injection of a solution or with cap suction, and then removing it with a snare.⁴ Most lesions smaller than 20 mm can be removed in one piece (en bloc).⁵ Larger lesions are removed in multiple pieces (ie, piecemeal). Unfortunately, some fibrotic lesions, which are usually difficult to lift, cannot be completely removed by EMR.

ESD was first performed in the late 1990s with the aim of overcoming the limitations of EMR in resecting large or fibrotic tumors en bloc.^6,7 Since then, ESD technique has been standardized and training centers have been created, especially in Asia, where it is widely used for treatment of early gastric cancer.^3,8–10 Since 2012 it has been covered by the Japanese National Health Insurance for treatment of early gastric cancer, and since 2014 for treatment of colorectal malignant tumors measuring 2 to 5 cm.¹¹

Adoption of ESD has been slow in Western countries, where many patients are still referred for surgery or undergo EMR for removal of superficial neoplasms. Reasons for this slow adoption are that gastric cancer is much less common in Western countries, and also that ESD demands a high level of technical skill, is difficult to learn, and is expensive.^3,12,13 However, small groups of Western endoscopists have become interested and are advocating it, first studying it on their own and then training in a Japanese center and learning from experts performing the procedure.

Therefore, in a Western setting, ESD should be performed in specialized endoscopy centers and offered to selected patients.¹  

CANDIDATES SHOULD HAVE EARLY-STAGE, SUPERFICIAL TUMORS

Ideal candidates for endoscopic resection are patients who have early cancer with a negligible risk of lymph node metastasis, such as cancer limited to the mucosa (stage T1a).⁷ Therefore, to determine the best treatment for a patient with a newly diagnosed gastrointestinal neoplasm, it is mandatory to estimate the depth of invasion.

The depth of invasion is directly correlated with lymph node involvement, which is ultimately the main predictive factor for long-term adverse outcomes of gastrointestinal tumors.^4,14–17 Accurate multidisciplinary preprocedure estimations are mandatory, as incorrect evaluations may result in inappropriate therapy and residual cancer.¹⁸

Other factors that have been used to predict lymph node involvement include tumor size, macroscopic appearance, histologic differentiation, and lymphatic and vascular involvement.¹⁹ Some of these factors can be assessed by special endoscopic techniques (chromoendoscopy and narrow-band imaging with magnifying endoscopy) that allow accurate real-time estimation of the depth of invasion of the lesion.^5,17,20–27 Evaluation of microsurface and microvascular arrangements is especially useful for determining the feasibility of ESD in gastric tumors, evaluation of intracapillary loops is useful in esophageal lesions, and assessment of mucosal pit patterns is useful for colorectal lesions.^21–29

Endoscopic ultrasonography is another tool that has been used to estimate the depth of the tumor. Although it can differentiate between definite intramucosal and definite submucosal invasive cancers, its ability to confirm minute submucosal invasion is limited. Its use as the sole tumor staging modality is not encouraged, and it should always be used in conjunction with endoscopic evaluation.¹⁸

Though the aforementioned factors help stratify patients, pathologic staging is the best predictor of lymph node metastasis. ESD provides adequate specimens for accurate pathologic evaluation, as it removes lesions en bloc.³⁰

All patients found to have risk factors for lymph node metastasis on endoscopic, ultrasonographic, or pathologic analysis should be referred for surgical evaluation.^9,19,31,32

ENDOSCOPIC SUBMUCOSAL DISSECTION

Before the procedure, the patient’s physicians need to do the following:

Determine the best type of intervention (EMR, ESD, ablation, surgery) for the specific lesion.³ A multidisciplinary approach is encouraged, with involvement of the internist, gastroenterologist, and surgeon.

Plan for anesthesia, additional consultations, pre- and postprocedural hospital admission, and need for special equipment.³³

During the procedure

Define the lateral extent of the lesion using magnification chromoendoscopy or narrow-band imaging. In the stomach, a biopsy sample should be taken from the worst-looking segment and from normal-looking mucosa. Multiple biopsies should be avoided to prevent subsequent fibrosis.³³ In the colon, biopsy should be avoided.³⁴

Identify and circumferentially mark the target lesion. Cautery or argon plasma coagulation can be used for making markings at a distance of 5 to 10 mm from the edges.³³ This is done to recognize the borders of the lesion, because they can become distorted after submucosal injection.¹⁴ This step is unnecessary in colorectal cases, as tumor margins can be adequately visualized after chromoendoscopy.^16,35

Lift the lesion by injecting saline, 0.5% hyaluronate, or glycerin to create a submucosal fluid cushion.^19,33

Perform a circumferential incision lateral to the mucosal margins to allow for a normal tissue margin.³³ Partial incision is performed for esophageal and colorectal ESD to avoid fluid leakage from the submucosal layer, achieving a sustained submucosal lift and safer dissection.¹⁶

Submucosal dissection. The submucosal layer is dissected with an electrocautery knife until the lesion is completely removed. Dissection should be done carefully to keep the submucosal plane.³³ Hemoclips or hemostat forceps can be used to control visible bleeding. The resected specimen is then stretched and fixed to a board using small pins for further histopathologic evaluation.³⁵

Postprocedural monitoring.  All patients should be admitted for overnight observation. Those who undergo gastric ESD should receive high-dose acid suppression, and the next day they can be started on a liquid diet.¹⁹

STOMACH CANCER

Indications for ESD for stomach cancer in the East

The incidence of gastric cancer is higher in Japan and Korea, where widespread screening programs have led to early identification and early treatment of this disease.³⁶

Pathology studies³⁷ of samples from patients with gastric cancer identified the following as risk factors for lymph node metastasis, which would make ESD unsuitable:

• Undifferentiated type
• Tumors larger than 2 cm
• Lymphatic or venous involvement
• Submucosal invasion
• Ulcerative change.

Based on these findings, the situations in which there was no risk of lymph node involvement (ie, when none of the above factors are present) were accepted as absolute indications for endoscopic resection of early gastric cancer.³⁸ Further histologic studies identified a subset of patients with lesions with very low risk of lymph node metastasis, which outweighed the risk of surgery. Based on these findings, expanded criteria for gastric ESD were proposed,^39,40 and the Japanese gastric cancer treatment guidelines now include these expanded preoperative indications^9,17 (Table 1).

Based on information from the Japanese Gastric Cancer Association, reference 9.

The Japanese Gastric Cancer Association has proposed a treatment algorithm based on the histopathologic evaluation after resection (Figure 2).⁹

Outcomes

In the largest series of patients who underwent curative ESD for early gastric cancer, the 5-year survival rate was 92.6%, the 5-year disease-specific survival rate was 99.9%, and the 5-year relative survival rate was 105%.⁴¹

Similarly, in a Japanese population-based survival analysis, the relative 5-year survival rate for localized gastric cancer was 94.4%.⁴² Rates of en bloc resection and complete resection with ESD are higher than those with EMR, resulting in a lower risk of local recurrence in selected patients who undergo ESD.^8,43,44

Although rare, local recurrence after curative gastric ESD has been reported.⁴⁵ The annual incidence of local recurrence has been estimated to be 0.84%.⁴⁶

ESD entails a shorter hospital stay and requires fewer resources than surgery, resulting in lower medical costs (Table 2).⁴⁴ Additionally, as endoscopic resection is associated with less morbidity, fewer procedure-related adverse events, and fewer complications, ESD could be used as the standard treatment for early gastric cancer.^47,48

The Western perspective on endoscopic submucosal dissection for gastric cancer

Since the prevalence of gastric cancer in Western countries is significantly lower than in Japan and Korea, local data and experience are scarce. However, experts performing ESD in the West have adopted the indications of the Japan Gastroenterological Endoscopy Society. The European Society of Gastrointestinal Endoscopy recommends ESD for excision of most superficial gastric neoplasms, with EMR being preferred only in lesions smaller than 15 mm, Paris classification 0 or IIA.^5,32

Patients with gastric lesions measuring 15 mm or larger should undergo high-quality endoscopy, preferably chromoendoscopy, to evaluate the mucosal patterns and determine the depth of invasion. If superficial involvement is confirmed, other imaging techniques are not routinely recommended.⁵ A surgery consult is also recommended.

ESOPHAGEAL CANCER

Indications for ESD for esophageal cancer in the East

Due to the success of ESD for early gastric cancer, this technique is now also used for superficial esophageal neoplasms.^19,49 It should be done in a specialized center, as it is more technically difficult than gastric ESD: the esophageal lumen is narrow, the wall is thin, and the esophagus moves with respiration and heartbeat.⁵⁰ A multidisciplinary approach including an endoscopist, a surgeon, and a pathologist is highly recommended for evaluation and treatment.

EMR is preferred for removal of mucosal cancer, in view of its safety profile and success rates. ESD can be considered in cases of lesions larger than 15 mm, poorly lifting tumors, and those with the possibility of submucosal invasion (Table 3).^5,45,49,51

Circumference involvement is critical when determining eligible candidates, as a defect involving more than three-fourths of the esophageal circumference can lead to esophageal strictures.⁵² Controlled prospective studies have shown promising results from giving intralesional and oral steroids to prevent stricture after ESD, which could potentially overcome this size limitation.^53,54

Outcomes for esophageal cancer

ESD has been shown to be safe and effective, achieving en bloc resection in 85% to 100% of patients.^19,51 Its advantages over EMR include en bloc resection, complete resection, and high curative rates, resulting in higher recurrence-free survival.^2,55,56 Although the incidence of complications such as bleeding, perforation, and stricture formation are higher with ESD, patients usually recover uneventfully.^2,19,20

ESD in the esophagus: The Western perspective

As data on the efficacy of EMR vs ESD for the treatment of Barrett esophagus with adenocarcinoma are limited, EMR is the gold standard endoscopic technique for removal of visible esophageal dysplastic lesions.^5,51,57 ESD can be considered for tumors larger than 15 mm, for poorly lifting lesions, and if there is suspicion of submucosal invasion.⁵

Patients should be evaluated by an experienced endoscopist, using an advanced imaging technique such as narrow-band imaging or chromoendoscopy. If suspicious features are found, endoscopic ultrasonography should be considered to confirm submucosal invasion or lymph node involvement.⁵

Indications for ESD for colorectal cancer in the East

Colon cancer is one of the leading causes of cancer-related deaths worldwide.⁵⁸ Since ESD has been found to be effective and safe in treating gastric cancer, it has also been used to remove large colorectal tumors.⁵⁹ However, ESD is not universally accepted in the treatment of colorectal neoplasms due to its greater technical difficulty, longer procedural time, and higher risk of perforating the thinner colonic wall compared with EMR.^21,60

Outcomes for colorectal cancer

Tumor size of 50 mm or larger is a risk factor for complications, while a high procedure volume at the center is a protective factor.⁶⁰

Endoscopic treatment of colorectal cancer: The Western perspective

EMR is the gold standard for removal of superficial colorectal lesions. However, ESD can be considered if there is suspicion of superficial submucosal invasion, especially for lesions larger than 20 mm that cannot be resected en bloc by EMR.³² ESD can also be used for fibrotic lesions not amenable to complete EMR removal, or as a salvage procedure after recurrence after EMR.⁶⁷ Proper selection of cases is critical.¹

Patients who have a superficial colonic lesion should be evaluated by means of high-definition endoscopy and chromoendoscopy to assess the mucosal pattern and establish feasibility of endoscopic resection. If submucosal invasion is suspected, staging with endoscopic ultrasonography or magnetic resonance imaging should be considered.⁵

FOLLOW-UP AFTER ESD

Endoscopic surveillance after the procedure is recommended, given the persistent risk of metachronous cancer after curative ESD due to its organ-sparing quality.⁶⁸ Surveillance endoscopy aims to achieve early detection and subsequent endoscopic resection of metachronous lesions.

Histopathologic evaluation assessing the presence of malignant cells in the margins of a resected sample is mandatory for determining the next step in treatment. If margins are negative, follow-up endoscopy can be done every 6 to 12 months. If margins are positive, the approach includes surgery, reattempting ESD or endoscopic surveillance in 3 or 6 months.^3,32 Although the surveillance strategy varies according to individual risk of metachronous cancer, it should be continued indefinitely.⁶⁸

COMPLICATIONS OF ESD

The most common procedure-related complications of ESD are bleeding, perforation, and stricture. Most intraprocedural adverse events can be managed endoscopically.⁶⁹

Bleeding

Most bleeding occurs during the procedure or early after it and can be controlled with electrocautery.^49,69 No episodes of massive bleeding, defined as causing clinical symptoms and requiring transfusion or surgery, have been reported.^20,43,55

In gastric ESD, delayed bleeding rates have ranged from 0 to 15.6%.⁶⁹ Bleeding may be prevented with endoscopic coagulation of visible vessels after dissection has been completed and by proton pump inhibitor therapy.^70,71 Excessive coagulation should be avoided to lower the risk of perforation.³³

In colorectal ESD the bleeding rate has been reported to be 2.2%; applying coagulation to an area where a blood vessel is suspected before cutting (precoagulation) may prevent subsequent bleeding.²¹

Perforation

For gastric ESD, perforation rates range from 1.2% to 5.2%.⁶⁹ Esophageal perforation rates can be up to 4%.⁴⁹ In colorectal ESD, perforation rates have been reported to be 1.6% to 6.6%.^60,72

Although most of the cases were successfully managed with conservative treatment, some required emergency surgery.^60,73

Strictures

In a case series of 532 patients undergoing gastric ESD, stricture was reported in 5 patients, all of whom presented with obstructive symptoms.⁷⁴ Risk factors for post-ESD gastric stenosis are a mucosal defect with a circumferential extent of more than three-fourths or a longitudinal extent of more than 5 cm.⁷⁵

Strictures are common after esophageal ESD, with rates ranging from 2% to 26%. The risk is higher when longer segments are removed or circumferential resection is performed. As previously mentioned, this complication may be reduced with ingestion or injection of steroids  after the procedure.^53,54

Surprisingly, ESD of large colorectal lesions involving more than three-fourths of the circumference of the rectum is rarely complicated by stenosis.⁷⁶

ESD requires a high level of technical skill, is time-consuming, and has a higher rate of complications than conventional endoscopic resection. A standardized ESD training system is needed, as the procedure is more difficult than EMR. Training in porcine models has been shown to confer competency in ESD in a Western setting.^13,16,33

Colorectal ESD is an even more challenging procedure, given the potential for complications related to its anatomy. Training centers in Japan usually have their trainees first master gastric ESD, then assist in more than 20 colorectal ESDs conducted by experienced endoscopists, and accomplish 30 cases before performing the procedure safely and independently.

As the incidence of gastric cancer is low in Western countries, trainees may also begin with lower rectal lesions, which are easier to remove.⁷⁷ Incorporation of ESD in the West would require a clear treatment algorithm. It is a complex procedure, with higher rates of complications, a prolonged learning curve, and prolonged procedure time. Therefore, it should be performed in specialized centers and under the special situations discussed here to ensure that the benefits for the patients outweigh the risks.

VALUE OF ENDOSCOPIC SUBMUCOSAL DISSECTION

The optimal method for resecting gastrointestinal neoplasms should be safe, cost-effective, and quick and should also completely remove the lesion. The best treatment strategy takes into account the characteristics of the lesion and the comorbidities and wishes of the patient. Internists should be aware of the multiple options available to achieve the best outcome for the patient.¹

Endoscopic resection of superficial gastrointestinal neoplasms, including EMR and ESD, has been a subject of increasing interest due to its minimally invasive and potentially curative character. However, cancer can recur after endoscopic resection because the procedure is organ-sparing.

ESD allows resection of early gastrointestinal tumors with a minimally invasive technique. It can achieve higher curative resection rates and lower recurrence rates compared with EMR. Compared with surgery, ESD leads to less morbidity, fewer procedure-related complications, and lower medical costs. Indications should be rigorously followed to achieve successful treatments in selected patients.

Multiple variables have to be taken into account when deciding which treatment is best, such as tumor characteristics, the patient’s baseline condition, physician expertise, and hospital resources.⁴⁸ Less-invasive treatments may improve the prognosis of patients. No matter the approach, patients should be treated in specialized treatment centers.

Internal medicine physicians should be aware of the advances in treatments for early gastrointestinal cancer so appropriate options can be considered.

The treatment of early esophageal, gastric, and colorectal cancer is changing.¹ For many years, surgery was the mainstay of treatment for early-stage gastrointestinal cancer. Unfortunately, surgery leads to significant loss of function of the organ, resulting in increased morbidity and decreased quality of life.²

Endoscopic techniques, particularly endoscopic mucosal resection (EMR) and endoscopic submucosal dissection (ESD), have been developed and are widely used in Japan, where gastrointestinal cancer is more common than in the West. This article reviews the indications, complications, and outcomes of ESD for early gastrointestinal neoplasms, so that readers will recognize the subset of patients who would benefit from ESD in a Western setting.

ENDOSCOPIC MUCOSAL RESECTION AND SUBMUCOSAL DISSECTION

Since the first therapeutic polypectomy was performed in Japan in 1974, several endoscopic techniques for tumor resection have been developed.³

EMR, one of the most successful and widely used techniques, involves elevating the lesion either with submucosal injection of a solution or with cap suction, and then removing it with a snare.⁴ Most lesions smaller than 20 mm can be removed in one piece (en bloc).⁵ Larger lesions are removed in multiple pieces (ie, piecemeal). Unfortunately, some fibrotic lesions, which are usually difficult to lift, cannot be completely removed by EMR.

ESD was first performed in the late 1990s with the aim of overcoming the limitations of EMR in resecting large or fibrotic tumors en bloc.^6,7 Since then, ESD technique has been standardized and training centers have been created, especially in Asia, where it is widely used for treatment of early gastric cancer.^3,8–10 Since 2012 it has been covered by the Japanese National Health Insurance for treatment of early gastric cancer, and since 2014 for treatment of colorectal malignant tumors measuring 2 to 5 cm.¹¹

Adoption of ESD has been slow in Western countries, where many patients are still referred for surgery or undergo EMR for removal of superficial neoplasms. Reasons for this slow adoption are that gastric cancer is much less common in Western countries, and also that ESD demands a high level of technical skill, is difficult to learn, and is expensive.^3,12,13 However, small groups of Western endoscopists have become interested and are advocating it, first studying it on their own and then training in a Japanese center and learning from experts performing the procedure.

Therefore, in a Western setting, ESD should be performed in specialized endoscopy centers and offered to selected patients.¹  

CANDIDATES SHOULD HAVE EARLY-STAGE, SUPERFICIAL TUMORS

Ideal candidates for endoscopic resection are patients who have early cancer with a negligible risk of lymph node metastasis, such as cancer limited to the mucosa (stage T1a).⁷ Therefore, to determine the best treatment for a patient with a newly diagnosed gastrointestinal neoplasm, it is mandatory to estimate the depth of invasion.

The depth of invasion is directly correlated with lymph node involvement, which is ultimately the main predictive factor for long-term adverse outcomes of gastrointestinal tumors.^4,14–17 Accurate multidisciplinary preprocedure estimations are mandatory, as incorrect evaluations may result in inappropriate therapy and residual cancer.¹⁸

Other factors that have been used to predict lymph node involvement include tumor size, macroscopic appearance, histologic differentiation, and lymphatic and vascular involvement.¹⁹ Some of these factors can be assessed by special endoscopic techniques (chromoendoscopy and narrow-band imaging with magnifying endoscopy) that allow accurate real-time estimation of the depth of invasion of the lesion.^5,17,20–27 Evaluation of microsurface and microvascular arrangements is especially useful for determining the feasibility of ESD in gastric tumors, evaluation of intracapillary loops is useful in esophageal lesions, and assessment of mucosal pit patterns is useful for colorectal lesions.^21–29

Endoscopic ultrasonography is another tool that has been used to estimate the depth of the tumor. Although it can differentiate between definite intramucosal and definite submucosal invasive cancers, its ability to confirm minute submucosal invasion is limited. Its use as the sole tumor staging modality is not encouraged, and it should always be used in conjunction with endoscopic evaluation.¹⁸

Though the aforementioned factors help stratify patients, pathologic staging is the best predictor of lymph node metastasis. ESD provides adequate specimens for accurate pathologic evaluation, as it removes lesions en bloc.³⁰

All patients found to have risk factors for lymph node metastasis on endoscopic, ultrasonographic, or pathologic analysis should be referred for surgical evaluation.^9,19,31,32

ENDOSCOPIC SUBMUCOSAL DISSECTION

Before the procedure, the patient’s physicians need to do the following:

Determine the best type of intervention (EMR, ESD, ablation, surgery) for the specific lesion.³ A multidisciplinary approach is encouraged, with involvement of the internist, gastroenterologist, and surgeon.

Plan for anesthesia, additional consultations, pre- and postprocedural hospital admission, and need for special equipment.³³

During the procedure

Define the lateral extent of the lesion using magnification chromoendoscopy or narrow-band imaging. In the stomach, a biopsy sample should be taken from the worst-looking segment and from normal-looking mucosa. Multiple biopsies should be avoided to prevent subsequent fibrosis.³³ In the colon, biopsy should be avoided.³⁴

Identify and circumferentially mark the target lesion. Cautery or argon plasma coagulation can be used for making markings at a distance of 5 to 10 mm from the edges.³³ This is done to recognize the borders of the lesion, because they can become distorted after submucosal injection.¹⁴ This step is unnecessary in colorectal cases, as tumor margins can be adequately visualized after chromoendoscopy.^16,35

Lift the lesion by injecting saline, 0.5% hyaluronate, or glycerin to create a submucosal fluid cushion.^19,33

Perform a circumferential incision lateral to the mucosal margins to allow for a normal tissue margin.³³ Partial incision is performed for esophageal and colorectal ESD to avoid fluid leakage from the submucosal layer, achieving a sustained submucosal lift and safer dissection.¹⁶

Submucosal dissection. The submucosal layer is dissected with an electrocautery knife until the lesion is completely removed. Dissection should be done carefully to keep the submucosal plane.³³ Hemoclips or hemostat forceps can be used to control visible bleeding. The resected specimen is then stretched and fixed to a board using small pins for further histopathologic evaluation.³⁵

Postprocedural monitoring.  All patients should be admitted for overnight observation. Those who undergo gastric ESD should receive high-dose acid suppression, and the next day they can be started on a liquid diet.¹⁹

STOMACH CANCER

Indications for ESD for stomach cancer in the East

The incidence of gastric cancer is higher in Japan and Korea, where widespread screening programs have led to early identification and early treatment of this disease.³⁶

Pathology studies³⁷ of samples from patients with gastric cancer identified the following as risk factors for lymph node metastasis, which would make ESD unsuitable:

• Undifferentiated type
• Tumors larger than 2 cm
• Lymphatic or venous involvement
• Submucosal invasion
• Ulcerative change.

Based on these findings, the situations in which there was no risk of lymph node involvement (ie, when none of the above factors are present) were accepted as absolute indications for endoscopic resection of early gastric cancer.³⁸ Further histologic studies identified a subset of patients with lesions with very low risk of lymph node metastasis, which outweighed the risk of surgery. Based on these findings, expanded criteria for gastric ESD were proposed,^39,40 and the Japanese gastric cancer treatment guidelines now include these expanded preoperative indications^9,17 (Table 1).

Based on information from the Japanese Gastric Cancer Association, reference 9.

The Japanese Gastric Cancer Association has proposed a treatment algorithm based on the histopathologic evaluation after resection (Figure 2).⁹

Outcomes

In the largest series of patients who underwent curative ESD for early gastric cancer, the 5-year survival rate was 92.6%, the 5-year disease-specific survival rate was 99.9%, and the 5-year relative survival rate was 105%.⁴¹

Similarly, in a Japanese population-based survival analysis, the relative 5-year survival rate for localized gastric cancer was 94.4%.⁴² Rates of en bloc resection and complete resection with ESD are higher than those with EMR, resulting in a lower risk of local recurrence in selected patients who undergo ESD.^8,43,44

Although rare, local recurrence after curative gastric ESD has been reported.⁴⁵ The annual incidence of local recurrence has been estimated to be 0.84%.⁴⁶

ESD entails a shorter hospital stay and requires fewer resources than surgery, resulting in lower medical costs (Table 2).⁴⁴ Additionally, as endoscopic resection is associated with less morbidity, fewer procedure-related adverse events, and fewer complications, ESD could be used as the standard treatment for early gastric cancer.^47,48

The Western perspective on endoscopic submucosal dissection for gastric cancer

Since the prevalence of gastric cancer in Western countries is significantly lower than in Japan and Korea, local data and experience are scarce. However, experts performing ESD in the West have adopted the indications of the Japan Gastroenterological Endoscopy Society. The European Society of Gastrointestinal Endoscopy recommends ESD for excision of most superficial gastric neoplasms, with EMR being preferred only in lesions smaller than 15 mm, Paris classification 0 or IIA.^5,32

Patients with gastric lesions measuring 15 mm or larger should undergo high-quality endoscopy, preferably chromoendoscopy, to evaluate the mucosal patterns and determine the depth of invasion. If superficial involvement is confirmed, other imaging techniques are not routinely recommended.⁵ A surgery consult is also recommended.

ESOPHAGEAL CANCER

Indications for ESD for esophageal cancer in the East

Due to the success of ESD for early gastric cancer, this technique is now also used for superficial esophageal neoplasms.^19,49 It should be done in a specialized center, as it is more technically difficult than gastric ESD: the esophageal lumen is narrow, the wall is thin, and the esophagus moves with respiration and heartbeat.⁵⁰ A multidisciplinary approach including an endoscopist, a surgeon, and a pathologist is highly recommended for evaluation and treatment.

EMR is preferred for removal of mucosal cancer, in view of its safety profile and success rates. ESD can be considered in cases of lesions larger than 15 mm, poorly lifting tumors, and those with the possibility of submucosal invasion (Table 3).^5,45,49,51

Circumference involvement is critical when determining eligible candidates, as a defect involving more than three-fourths of the esophageal circumference can lead to esophageal strictures.⁵² Controlled prospective studies have shown promising results from giving intralesional and oral steroids to prevent stricture after ESD, which could potentially overcome this size limitation.^53,54

Outcomes for esophageal cancer

ESD has been shown to be safe and effective, achieving en bloc resection in 85% to 100% of patients.^19,51 Its advantages over EMR include en bloc resection, complete resection, and high curative rates, resulting in higher recurrence-free survival.^2,55,56 Although the incidence of complications such as bleeding, perforation, and stricture formation are higher with ESD, patients usually recover uneventfully.^2,19,20

ESD in the esophagus: The Western perspective

As data on the efficacy of EMR vs ESD for the treatment of Barrett esophagus with adenocarcinoma are limited, EMR is the gold standard endoscopic technique for removal of visible esophageal dysplastic lesions.^5,51,57 ESD can be considered for tumors larger than 15 mm, for poorly lifting lesions, and if there is suspicion of submucosal invasion.⁵

Patients should be evaluated by an experienced endoscopist, using an advanced imaging technique such as narrow-band imaging or chromoendoscopy. If suspicious features are found, endoscopic ultrasonography should be considered to confirm submucosal invasion or lymph node involvement.⁵

Indications for ESD for colorectal cancer in the East

Colon cancer is one of the leading causes of cancer-related deaths worldwide.⁵⁸ Since ESD has been found to be effective and safe in treating gastric cancer, it has also been used to remove large colorectal tumors.⁵⁹ However, ESD is not universally accepted in the treatment of colorectal neoplasms due to its greater technical difficulty, longer procedural time, and higher risk of perforating the thinner colonic wall compared with EMR.^21,60

Outcomes for colorectal cancer

Tumor size of 50 mm or larger is a risk factor for complications, while a high procedure volume at the center is a protective factor.⁶⁰

Endoscopic treatment of colorectal cancer: The Western perspective

EMR is the gold standard for removal of superficial colorectal lesions. However, ESD can be considered if there is suspicion of superficial submucosal invasion, especially for lesions larger than 20 mm that cannot be resected en bloc by EMR.³² ESD can also be used for fibrotic lesions not amenable to complete EMR removal, or as a salvage procedure after recurrence after EMR.⁶⁷ Proper selection of cases is critical.¹

Patients who have a superficial colonic lesion should be evaluated by means of high-definition endoscopy and chromoendoscopy to assess the mucosal pattern and establish feasibility of endoscopic resection. If submucosal invasion is suspected, staging with endoscopic ultrasonography or magnetic resonance imaging should be considered.⁵

FOLLOW-UP AFTER ESD

Endoscopic surveillance after the procedure is recommended, given the persistent risk of metachronous cancer after curative ESD due to its organ-sparing quality.⁶⁸ Surveillance endoscopy aims to achieve early detection and subsequent endoscopic resection of metachronous lesions.

Histopathologic evaluation assessing the presence of malignant cells in the margins of a resected sample is mandatory for determining the next step in treatment. If margins are negative, follow-up endoscopy can be done every 6 to 12 months. If margins are positive, the approach includes surgery, reattempting ESD or endoscopic surveillance in 3 or 6 months.^3,32 Although the surveillance strategy varies according to individual risk of metachronous cancer, it should be continued indefinitely.⁶⁸

COMPLICATIONS OF ESD

The most common procedure-related complications of ESD are bleeding, perforation, and stricture. Most intraprocedural adverse events can be managed endoscopically.⁶⁹

Bleeding

Most bleeding occurs during the procedure or early after it and can be controlled with electrocautery.^49,69 No episodes of massive bleeding, defined as causing clinical symptoms and requiring transfusion or surgery, have been reported.^20,43,55

In gastric ESD, delayed bleeding rates have ranged from 0 to 15.6%.⁶⁹ Bleeding may be prevented with endoscopic coagulation of visible vessels after dissection has been completed and by proton pump inhibitor therapy.^70,71 Excessive coagulation should be avoided to lower the risk of perforation.³³

In colorectal ESD the bleeding rate has been reported to be 2.2%; applying coagulation to an area where a blood vessel is suspected before cutting (precoagulation) may prevent subsequent bleeding.²¹

Perforation

For gastric ESD, perforation rates range from 1.2% to 5.2%.⁶⁹ Esophageal perforation rates can be up to 4%.⁴⁹ In colorectal ESD, perforation rates have been reported to be 1.6% to 6.6%.^60,72

Although most of the cases were successfully managed with conservative treatment, some required emergency surgery.^60,73

Strictures

In a case series of 532 patients undergoing gastric ESD, stricture was reported in 5 patients, all of whom presented with obstructive symptoms.⁷⁴ Risk factors for post-ESD gastric stenosis are a mucosal defect with a circumferential extent of more than three-fourths or a longitudinal extent of more than 5 cm.⁷⁵

Strictures are common after esophageal ESD, with rates ranging from 2% to 26%. The risk is higher when longer segments are removed or circumferential resection is performed. As previously mentioned, this complication may be reduced with ingestion or injection of steroids  after the procedure.^53,54

Surprisingly, ESD of large colorectal lesions involving more than three-fourths of the circumference of the rectum is rarely complicated by stenosis.⁷⁶

ESD requires a high level of technical skill, is time-consuming, and has a higher rate of complications than conventional endoscopic resection. A standardized ESD training system is needed, as the procedure is more difficult than EMR. Training in porcine models has been shown to confer competency in ESD in a Western setting.^13,16,33

Colorectal ESD is an even more challenging procedure, given the potential for complications related to its anatomy. Training centers in Japan usually have their trainees first master gastric ESD, then assist in more than 20 colorectal ESDs conducted by experienced endoscopists, and accomplish 30 cases before performing the procedure safely and independently.

As the incidence of gastric cancer is low in Western countries, trainees may also begin with lower rectal lesions, which are easier to remove.⁷⁷ Incorporation of ESD in the West would require a clear treatment algorithm. It is a complex procedure, with higher rates of complications, a prolonged learning curve, and prolonged procedure time. Therefore, it should be performed in specialized centers and under the special situations discussed here to ensure that the benefits for the patients outweigh the risks.

VALUE OF ENDOSCOPIC SUBMUCOSAL DISSECTION

The optimal method for resecting gastrointestinal neoplasms should be safe, cost-effective, and quick and should also completely remove the lesion. The best treatment strategy takes into account the characteristics of the lesion and the comorbidities and wishes of the patient. Internists should be aware of the multiple options available to achieve the best outcome for the patient.¹

Endoscopic resection of superficial gastrointestinal neoplasms, including EMR and ESD, has been a subject of increasing interest due to its minimally invasive and potentially curative character. However, cancer can recur after endoscopic resection because the procedure is organ-sparing.

ESD allows resection of early gastrointestinal tumors with a minimally invasive technique. It can achieve higher curative resection rates and lower recurrence rates compared with EMR. Compared with surgery, ESD leads to less morbidity, fewer procedure-related complications, and lower medical costs. Indications should be rigorously followed to achieve successful treatments in selected patients.

Multiple variables have to be taken into account when deciding which treatment is best, such as tumor characteristics, the patient’s baseline condition, physician expertise, and hospital resources.⁴⁸ Less-invasive treatments may improve the prognosis of patients. No matter the approach, patients should be treated in specialized treatment centers.

Internal medicine physicians should be aware of the advances in treatments for early gastrointestinal cancer so appropriate options can be considered.

Acute pancreatitis accounted for more than 300,000 admissions and $2.6 billion in associated healthcare costs in the United States in 2012.¹ First-line management is early aggressive fluid resuscitation and analgesics for pain control. Guidelines recommend estimating the clinical severity of each attack using a validated scoring system such as the Bedside Index of Severity in Acute Pancreatitis.² Clinically severe pancreatitis is associated with necrosis.

Acute pancreatitis results from inappropriate activation of zymogens and subsequent auto­digestion of the pancreas by its own enzymes. Though necrotizing pancreatitis is thought to be an ischemic complication, its pathogenesis is not completely understood. Necrosis increases the morbidity and mortality risk of acute pancreatitis because of its association with organ failure and infectious complications. As such, patients with necrotizing pancreatitis may need admission to the intensive care unit, nutritional support, antibiotics, and radiologic, endoscopic, or surgical interventions.

Here, we review current evidence regarding the diagnosis and management of necrotizing pancreatitis.

PROPER TERMINOLOGY HELPS COLLABORATION

Managing necrotizing pancreatitis requires the combined efforts of internists, gastroenterologists, radiologists, and surgeons. This collaboration is aided by proper terminology.

A classification system was devised in Atlanta, GA, in 1992 to facilitate communication and interdisciplinary collaboration.³ Severe pancreatitis was differentiated from mild by the presence of organ failure or the complications of pseudocyst, necrosis, or abscess.

The original Atlanta classification had several limitations. First, the terminology for fluid collections was ambiguous and frequently misused. Second, the assessment of clinical severity required either the Ranson score or the Acute Physiology and Chronic Health Evaluation II score, both of which are complex and have other limitations. Finally, advances in imaging and treatment have rendered the original Atlanta nomenclature obsolete.

In 2012, the Acute Pancreatitis Classification Working Group issued a revised Atlanta classification that modernized the terminology pertaining to natural history, severity, imaging features, and complications. It divides the natural course of acute pancreatitis into early and late phases.⁴

Early vs late phase

In the early phase, findings on computed tomography (CT) neither correlate with clinical severity nor alter clinical management.⁶ Thus, early imaging is not indicated unless there is diagnostic uncertainty, lack of response to appropriate treatment, or sudden deterioration.

Moderate pancreatitis describes patients with pancreatic necrosis with or without transient organ failure (organ dysfunction for ≤ 48 hours).

Severe pancreatitis is defined by pancreatic necrosis and persistent organ dysfunction.⁴ It may be accompanied by pancreatic and peripancreatic fluid collections; bacteremia and sepsis can occur in association with infection of necrotic collections.

Interstitial edematous pancreatitis vs necrotizing pancreatitis

The revised Atlanta classification maintains the original classification of acute pancreatitis into 2 main categories: interstitial edematous pancreatitis and necrotizing pancreatitis.

Necrotizing pancreatitis is further divided into 3 subtypes based on extent and location of necrosis:

• Parenchymal necrosis alone (5% of cases)
• Necrosis of peripancreatic fat alone (20%)
• Necrosis of both parenchyma and peripancreatic fat (75%).

Peripancreatic involvement is commonly found in the mesentery, peripancreatic and distant retroperitoneum, and lesser sac.

Of the three subtypes, peripancreatic necrosis has the best prognosis. However, all of the subtypes of necrotizing pancreatitis are associated with poorer outcomes than interstitial edematous pancreatitis.

Fluid collections

Acute pancreatic fluid collections contain exclusively nonsolid components without an inflammatory wall and are typically found in the peripancreatic fat. These collections often resolve without intervention as the patient recovers. If they persist beyond 4 weeks and develop a nonepithelialized, fibrous wall, they become pseudocysts. Intervention is generally not recommended for pseudocysts unless they are symptomatic.

Radiographic imaging is not usually necessary to diagnose acute pancreatitis. However, it can be a valuable tool to clarify an ambiguous presentation, determine severity, and identify complications.

The timing and appropriate type of imaging are integral to obtaining useful data. Any imaging obtained in acute pancreatitis to evaluate necrosis should be performed at least 3 to 5 days from the initial symptom onset; if imaging is obtained before 72 hours, necrosis cannot be confidently excluded.⁸

COMPUTED TOMOGRAPHY

CT is the imaging test of choice when evaluating acute pancreatitis. In addition, almost all percutaneous interventions are performed with CT guidance. The Balthazar score is the most well-known CT severity index. It is calculated based on the degree of inflammation, acute fluid collections, and parenchymal necrosis.⁹ However, a modified severity index incorporates extrapancreatic complications such as ascites and vascular compromise and was found to more strongly correlate with outcomes than the standard Balthazar score.¹⁰

Contrast-enhanced CT is performed in 2 phases:

The pancreatic parenchymal phase

The pancreatic parenchymal or late arterial phase is obtained approximately 40 to 45 seconds after the start of the contrast bolus. It is used to detect necrosis in the early phase of acute pancreatitis and to assess the peripancreatic arteries for pseudoaneurysms in the late phase of acute pancreatitis.¹¹

Pancreatic necrosis appears as an area of decreased parenchymal enhancement, either well-defined or heterogeneous. The normal pancreatic parenchyma has a postcontrast enhancement pattern similar to that of the spleen. Parenchyma that does not enhance to the same degree is considered necrotic. The severity of necrosis is graded based on the percentage of the pancreas involved (< 30%, 30%–50%, or > 50%), and a higher percentage correlates with a worse outcome.^12,13

Peripancreatic necrosis is harder to detect, as there is no method to assess fat enhancement as there is with pancreatic parenchymal enhancement. In general, radiologists assume that heterogeneous peripancreatic changes, including areas of fat, fluid, and soft tissue attenuation, are consistent with peripancreatic necrosis. After 7 to 10 days, if these changes become more homogeneous and confluent with a more mass-like process, peripancreatic necrosis can be more confidently identified.^12,13

The portal venous phase

The later, portal venous phase of the scan is obtained approximately 70 seconds after the start of the contrast bolus. It is used to detect and characterize fluid collections and venous complications of the disease.

Drawbacks of CT

A drawback of CT is the need for iodinated intravenous contrast media, which in severely ill patients may precipitate or worsen pre-existing acute kidney injury.

Further, several studies have shown that findings on CT rarely alter the management of patients in the early phase of acute pancreatitis and in fact may be an overuse of medical resources.¹⁴ Unless there are confounding clinical signs or symptoms, CT should be delayed for at least 72 hours.^9,10,14,15

MAGNETIC RESONANCE IMAGING

Magnetic resonance imaging (MRI) is not a first-line imaging test in this disease because it is not as available as CT and takes longer to perform—20 to 30 minutes. The patient must be evaluated for candidacy, as it is difficult for acutely ill patients to tolerate an examination that takes this long and requires them to hold their breath multiple times.

MRI is an appropriate alternative in patients who are pregnant or who have severe iodinated-contrast allergy. While contrast is necessary to detect pancreatic necrosis with CT, MRI can detect necrosis without the need for contrast in patients with acute kidney injury or severe chronic kidney disease. Also, MRI may be better in complicated cases requiring repeated imaging because it does not expose the patient to radiation.

On MRI, pancreatic necrosis appears as a heterogeneous area, owing to its liquid and solid components. Liquid components appear hyperintense, and solid components hypointense, on T2 fluid-weighted imaging. This ability to differentiate the components of a walled-off pancreatic necrosis can be useful in determining whether a collection requires drainage or debridement. MRI is also more sensitive for hemorrhagic complications, best seen on T1 fat-weighted images.^12,16

Magnetic resonance cholangiopancreatography is an excellent method for ductal evaluation through heavily T2-weighted imaging. It is more sensitive than CT for detecting common bile duct stones and can also detect pancreatic duct strictures or extravasation into fluid collections.¹⁶

SUPPORTIVE MANAGEMENT OF EARLY NECROTIZING PANCREATITIS

In the early phase of necrotizing pancreatitis, management is supportive with the primary aim of preventing intravascular volume depletion. Aggressive fluid resuscitation in the first 48 to 72 hours, pain control, and bowel rest are the mainstays of supportive therapy. Intensive care may be necessary if organ failure and hemodynamic instability accompany necrotizing pancreatitis.

Prophylactic antibiotic and antifungal therapy to prevent infected necrosis has been controversial. Recent studies of its utility have not yielded supportive results, and the American College of Gastroenterology and the Infectious Diseases Society of America no longer recommend it.^9,17 These medications should not be given unless concomitant cholangitis or extrapancreatic infection is clinically suspected.

Early enteral nutrition is recommended in patients in whom pancreatitis is predicted to be severe and in those not expected to resume oral intake within 5 to 7 days. Enteral nutrition most commonly involves bedside or endoscopic placement of a nasojejunal feeding tube and collaboration with a nutritionist to determine protein-caloric requirements.

Compared with enteral nutrition, total parenteral nutrition is associated with higher rates of infection, multiorgan dysfunction and failure, and death.¹⁸

Necrotizing pancreatitis is a defining complication of acute pancreatitis, and its presence alone indicates greater severity. However, superimposed complications may further worsen outcomes.

Infected pancreatic necrosis

Infection occurs in approximately 20% of patients with necrotizing pancreatitis and confers a mortality rate of 20% to 50%.¹⁹ Infected pancreatic necrosis occurs when gut organisms translocate into the nearby necrotic pancreatic and peripancreatic tissue. The most commonly identified organisms include Escherichia coli and Enterococcus species.²⁰

This complication usually manifests 2 to 4 weeks after symptom onset; earlier onset is uncommon to rare. It should be considered when the systemic inflammatory response syndrome persists or recurs after 10 days to 2 weeks. Systemic inflammatory response syndrome is also common in sterile necrotizing pancreatitis and sometimes in interstitial pancreatitis, particularly during the first week. However, its sudden appearance or resurgence, high spiking fevers, or worsening organ failure in the later phase (2–4 weeks) of pancreatitis should heighten suspicion of infected pancreatic necrosis.

Imaging may also help diagnose infection, and the presence of gas within a collection or region of necrosis is highly specific. However, the presence of gas is not completely sensitive for infection, as it is seen in only 12% to 22% of infected cases.

Before minimally invasive techniques became available, the diagnosis of infected pancreatic necrosis was confirmed by percutaneous CT-guided aspiration of the necrotic mass or collection for Gram stain and culture.

Antibiotic therapy is indicated in confirmed or suspected cases of infected pancreatic necrosis. Antibiotics with gram-negative coverage and appropriate penetration such as carbapenems, metronidazole, fluoroquinolones, and selected cephalosporins are most commonly used. Meropenem is the antibiotic of choice at our institution.

CT-guided fine-needle aspiration is often done if suspected infected pancreatic necrosis fails to respond to empiric antibiotic therapy.

Debridement or drainage. Generally, the diagnosis or suspicion of infected pancreatic necrosis (suggestive signs are high fever, elevated white blood cell count, and sepsis) warrants an intervention to debride or drain infected pancreatic tissue and control sepsis.²¹

While source control is integral to the successful treatment of infected pancreatic necrosis, antibiotic therapy may provide a bridge to intervention for critically ill patients by suppressing bacteremia and subsequent sepsis. A 2013 meta-analysis found that 324 of 409 patients with suspected infected pancreatic necrosis were successfully stabilized with antibiotic treatment.^21,22 The trend toward conservative management and promising outcomes with antibiotic therapy alone or with minimally invasive techniques has lessened the need for diagnostic CT-guided fine-needle aspiration.

Hemorrhage

Spontaneous hemorrhage into pancreatic necrosis is a rare but life-threatening complication. Because CT is almost always performed with contrast enhancement, this complication is rarely identified with imaging. The diagnosis is made by noting a drop in hemoglobin and hematocrit.

Hemorrhage into the retroperitoneum or the peritoneal cavity, or both, can occur when an inflammatory process erodes into a nearby artery. Luminal gastrointestinal bleeding can occur from gastric varices arising from splenic vein thrombosis and resulting left-sided portal hypertension, or from pseudoaneurysms. These can also bleed into the pancreatic duct (hemosuccus pancreaticus). Pseudoaneurysm is a later complication that occurs when an arterial wall (most commonly the splenic or gastroduodenal artery) is weakened by pancreatic enzymes.²³

Prompt recognition of hemorrhagic events and consultation with an interventional radiologist or surgeon are required to prevent death.

Inflammation and abdominal compartment syndrome

Inflammation from necrotizing pancreatitis can cause further complications by blocking nearby structures. Reported complications include jaundice from biliary compression, hydronephrosis from ureteral compression, bowel obstruction, and gastric outlet obstruction.

Abdominal compartment syndrome is an increasingly recognized complication of acute pancreatitis. Abdominal pressure can rise due to a number of factors, including fluid collections, ascites, ileus, and overly aggressive fluid resuscitation.²⁴ Elevated abdominal pressure is associated with complications such as decreased respiratory compliance, increased peak airway pressure, decreased cardiac preload, hypotension, mesenteric and intestinal ischemia, feeding intolerance, and lower-extremity ischemia and thrombosis.

Patients with necrotizing pancreatitis who have abdominal compartment syndrome have a mortality rate 5 times higher than patients without abdominal compartment syndrome.²⁵

Abdominal pressures should be monitored using a bladder pressure sensor in critically ill or ventilated patients with acute pancreatitis. If the abdominal pressure rises above 20 mm Hg, medical and surgical interventions should be offered in a stepwise fashion to decrease it. Interventions include decompression by nasogastric and rectal tube, sedation or paralysis to relax abdominal wall tension, minimization of intravenous fluids, percutaneous drainage of ascites, and (rarely) surgical midline or subcostal laparotomy.

The treatment of necrotizing pancreatitis has changed rapidly, thanks to a growing experience with minimally invasive techniques.

Indications for intervention

Infected pancreatic necrosis is the primary indication for surgical, percutaneous, or endoscopic intervention.

In sterile necrosis, the threshold for intervention is less clear, and intervention is often reserved for patients who fail to clinically improve or who have intractable abdominal pain, gastric outlet obstruction, or fistulating disease.²⁶

In asymptomatic cases, intervention is almost never indicated regardless of the location or size of the necrotic area.

In walled-off pancreatic necrosis, less-invasive and less-morbid interventions such as endoscopic or percutaneous drainage or video-assisted retroperitoneal debridement can be done.

Timing of intervention

In the past, delaying intervention was thought to increase the risk of death. However, multiple studies have found that outcomes are often worse if intervention is done early, likely due to the lack of a fully formed fibrous wall or demarcation of the necrotic area.²⁷

If the patient remains clinically stable, it is best to delay intervention until at least 4 weeks after the index event to achieve optimal outcomes. Delay can often be achieved by antibiotic treatment to suppress bacteremia and endoscopic or percutaneous drainage of infected collections to control sepsis.

Open surgery

The gold-standard intervention for infected pancreatic necrosis or symptomatic sterile walled-off pancreatic necrosis is open necrosectomy. This involves exploratory laparotomy with blunt debridement of all visible necrotic pancreatic tissue.

Methods to facilitate later evacuation of residual infected fluid and debris vary widely. Multiple large-caliber drains can be placed to facilitate irrigation and drainage before closure of the abdominal fascia. As infected pancreatic necrosis carries the risk of contaminating the peritoneal cavity, the skin is often left open to heal by secondary intention. An interventional radiologist is frequently enlisted to place, exchange, or downsize drainage catheters.

Infected pancreatic necrosis or symptomatic sterile walled-off pancreatic necrosis often requires more than one operation to achieve satisfactory debridement.

The goals of open necrosectomy are to remove nonviable tissue and infection, preserve viable pancreatic tissue, eliminate fistulous connections, and minimize damage to local organs and vasculature.

Minimally invasive techniques

Video-assisted retroperitoneal debridement has been described as a hybrid between endoscopic and open retroperitoneal debridement.²⁸ This technique requires first placing a percutaneous catheter into the necrotic area through the left flank to create a retroperitoneal tract. A 5-cm incision is made and the necrotic space is entered using the drain for guidance. Necrotic tissue is carefully debrided under direct vision using a combination of forceps, irrigation, and suction. A laparoscopic port can also be introduced into the incision when the procedure can no longer be continued under direct vision.^29,30

Although not all patients are candidates for minimal-access surgery, it remains an evolving surgical option.

Endoscopic transmural debridement is another option for infected pancreatic necrosis and symptomatic walled-off pancreatic necrosis. Depending on the location of the necrotic area, an echoendoscope is passed to either the stomach or duodenum. Guided by endoscopic ultrasonography, a needle is passed into the collection, allowing subsequent fistula creation and stenting for internal drainage or debridement. In the past, this process required several steps, multiple devices, fluoroscopic guidance, and considerable time. But newer endoscopic lumen-apposing metal stents have been developed that can be placed in a single step without fluoroscopy. A slimmer endoscope can then be introduced into the necrotic cavity via the stent, and the necrotic debris can be debrided with endoscopic baskets, snares, forceps, and irrigation.^9,31

Similar to surgical necrosectomy, satisfactory debridement is not often obtained with a single procedure; 2 to 5 endoscopic procedures may be needed to achieve resolution. However, the luminal approach in endoscopic necrosectomy avoids the significant morbidity of major abdominal surgery and the potential for pancreaticocutaneous fistulae that may occur with drains.

In a randomized trial comparing endoscopic necrosectomy vs surgical necrosectomy (video-assisted retroperitoneal debridement and exploratory laparotomy),³² endoscopic necrosectomy showed less inflammatory response than surgical necrosectomy and had a lower risk of new-onset organ failure, bleeding, fistula formation, and death.³²

Selecting the best intervention for the individual patient

Given the multiple available techniques, selecting the best intervention for individual patients can be challenging. A team approach with input from a gastroenterologist, surgeon, and interventional radiologist is best when determining which technique would best suit each patient.

Surgical necrosectomy is still the treatment of choice for unstable patients with infected pancreatic necrosis or multiple, inaccessible collections, but current evidence suggests a different approach in stable infected pancreatic necrosis and symptomatic sterile walled-off pancreatic necrosis.

The Dutch Pancreatitis Group²⁸ randomized 88 patients with infected pancreatic necrosis or symptomatic walled-off pancreatic necrosis to open necrosectomy or a minimally invasive “step-up” approach consisting of up to 2 percutaneous drainage or endoscopic debridement procedures before escalation to video-assisted retroperitoneal debridement. The step-up approach resulted in lower rates of morbidity and death than surgical necrosectomy as first-line treatment. Furthermore, some patients in the step-up group avoided the need for surgery entirely.³⁰

Necrosis significantly increases rates of morbidity and mortality in acute pancreatitis. Hospitalists, general internists, and general surgeons are all on the front lines in identifying severe cases and consulting the appropriate specialists for optimal multidisciplinary care. Selective and appropriate timing of radiologic imaging is key, and a vital tool in the management of necrotizing pancreatitis.

While the primary indication for intervention is infected pancreatic necrosis, additional indications are symptomatic walled-off pancreatic necrosis secondary to intractable abdominal pain, bowel obstruction, and failure to thrive. As a result of improving technology and inpatient care, these patients may present with intractable symptoms in the outpatient setting rather than the inpatient setting. The onus is on the primary care physician to maintain a high level of suspicion and refer these patients to subspecialists as appropriate.

Open surgical necrosectomy remains an important approach for care of infected pancreatic necrosis or patients with intractable symptoms. A step-up approach starting with a minimally invasive procedure and escalating if the initial intervention is unsuccessful is gradually becoming the standard of care.

Acute pancreatitis accounted for more than 300,000 admissions and $2.6 billion in associated healthcare costs in the United States in 2012.¹ First-line management is early aggressive fluid resuscitation and analgesics for pain control. Guidelines recommend estimating the clinical severity of each attack using a validated scoring system such as the Bedside Index of Severity in Acute Pancreatitis.² Clinically severe pancreatitis is associated with necrosis.

Acute pancreatitis results from inappropriate activation of zymogens and subsequent auto­digestion of the pancreas by its own enzymes. Though necrotizing pancreatitis is thought to be an ischemic complication, its pathogenesis is not completely understood. Necrosis increases the morbidity and mortality risk of acute pancreatitis because of its association with organ failure and infectious complications. As such, patients with necrotizing pancreatitis may need admission to the intensive care unit, nutritional support, antibiotics, and radiologic, endoscopic, or surgical interventions.

Here, we review current evidence regarding the diagnosis and management of necrotizing pancreatitis.

PROPER TERMINOLOGY HELPS COLLABORATION

Managing necrotizing pancreatitis requires the combined efforts of internists, gastroenterologists, radiologists, and surgeons. This collaboration is aided by proper terminology.

A classification system was devised in Atlanta, GA, in 1992 to facilitate communication and interdisciplinary collaboration.³ Severe pancreatitis was differentiated from mild by the presence of organ failure or the complications of pseudocyst, necrosis, or abscess.

The original Atlanta classification had several limitations. First, the terminology for fluid collections was ambiguous and frequently misused. Second, the assessment of clinical severity required either the Ranson score or the Acute Physiology and Chronic Health Evaluation II score, both of which are complex and have other limitations. Finally, advances in imaging and treatment have rendered the original Atlanta nomenclature obsolete.

In 2012, the Acute Pancreatitis Classification Working Group issued a revised Atlanta classification that modernized the terminology pertaining to natural history, severity, imaging features, and complications. It divides the natural course of acute pancreatitis into early and late phases.⁴

Early vs late phase

In the early phase, findings on computed tomography (CT) neither correlate with clinical severity nor alter clinical management.⁶ Thus, early imaging is not indicated unless there is diagnostic uncertainty, lack of response to appropriate treatment, or sudden deterioration.

Moderate pancreatitis describes patients with pancreatic necrosis with or without transient organ failure (organ dysfunction for ≤ 48 hours).

Severe pancreatitis is defined by pancreatic necrosis and persistent organ dysfunction.⁴ It may be accompanied by pancreatic and peripancreatic fluid collections; bacteremia and sepsis can occur in association with infection of necrotic collections.

Interstitial edematous pancreatitis vs necrotizing pancreatitis

The revised Atlanta classification maintains the original classification of acute pancreatitis into 2 main categories: interstitial edematous pancreatitis and necrotizing pancreatitis.

Necrotizing pancreatitis is further divided into 3 subtypes based on extent and location of necrosis:

• Parenchymal necrosis alone (5% of cases)
• Necrosis of peripancreatic fat alone (20%)
• Necrosis of both parenchyma and peripancreatic fat (75%).

Peripancreatic involvement is commonly found in the mesentery, peripancreatic and distant retroperitoneum, and lesser sac.

Of the three subtypes, peripancreatic necrosis has the best prognosis. However, all of the subtypes of necrotizing pancreatitis are associated with poorer outcomes than interstitial edematous pancreatitis.

Fluid collections

Acute pancreatic fluid collections contain exclusively nonsolid components without an inflammatory wall and are typically found in the peripancreatic fat. These collections often resolve without intervention as the patient recovers. If they persist beyond 4 weeks and develop a nonepithelialized, fibrous wall, they become pseudocysts. Intervention is generally not recommended for pseudocysts unless they are symptomatic.

Radiographic imaging is not usually necessary to diagnose acute pancreatitis. However, it can be a valuable tool to clarify an ambiguous presentation, determine severity, and identify complications.

The timing and appropriate type of imaging are integral to obtaining useful data. Any imaging obtained in acute pancreatitis to evaluate necrosis should be performed at least 3 to 5 days from the initial symptom onset; if imaging is obtained before 72 hours, necrosis cannot be confidently excluded.⁸

COMPUTED TOMOGRAPHY

CT is the imaging test of choice when evaluating acute pancreatitis. In addition, almost all percutaneous interventions are performed with CT guidance. The Balthazar score is the most well-known CT severity index. It is calculated based on the degree of inflammation, acute fluid collections, and parenchymal necrosis.⁹ However, a modified severity index incorporates extrapancreatic complications such as ascites and vascular compromise and was found to more strongly correlate with outcomes than the standard Balthazar score.¹⁰

Contrast-enhanced CT is performed in 2 phases:

The pancreatic parenchymal phase

The pancreatic parenchymal or late arterial phase is obtained approximately 40 to 45 seconds after the start of the contrast bolus. It is used to detect necrosis in the early phase of acute pancreatitis and to assess the peripancreatic arteries for pseudoaneurysms in the late phase of acute pancreatitis.¹¹

Pancreatic necrosis appears as an area of decreased parenchymal enhancement, either well-defined or heterogeneous. The normal pancreatic parenchyma has a postcontrast enhancement pattern similar to that of the spleen. Parenchyma that does not enhance to the same degree is considered necrotic. The severity of necrosis is graded based on the percentage of the pancreas involved (< 30%, 30%–50%, or > 50%), and a higher percentage correlates with a worse outcome.^12,13

Peripancreatic necrosis is harder to detect, as there is no method to assess fat enhancement as there is with pancreatic parenchymal enhancement. In general, radiologists assume that heterogeneous peripancreatic changes, including areas of fat, fluid, and soft tissue attenuation, are consistent with peripancreatic necrosis. After 7 to 10 days, if these changes become more homogeneous and confluent with a more mass-like process, peripancreatic necrosis can be more confidently identified.^12,13

The portal venous phase

The later, portal venous phase of the scan is obtained approximately 70 seconds after the start of the contrast bolus. It is used to detect and characterize fluid collections and venous complications of the disease.

Drawbacks of CT

A drawback of CT is the need for iodinated intravenous contrast media, which in severely ill patients may precipitate or worsen pre-existing acute kidney injury.

Further, several studies have shown that findings on CT rarely alter the management of patients in the early phase of acute pancreatitis and in fact may be an overuse of medical resources.¹⁴ Unless there are confounding clinical signs or symptoms, CT should be delayed for at least 72 hours.^9,10,14,15

MAGNETIC RESONANCE IMAGING

Magnetic resonance imaging (MRI) is not a first-line imaging test in this disease because it is not as available as CT and takes longer to perform—20 to 30 minutes. The patient must be evaluated for candidacy, as it is difficult for acutely ill patients to tolerate an examination that takes this long and requires them to hold their breath multiple times.

MRI is an appropriate alternative in patients who are pregnant or who have severe iodinated-contrast allergy. While contrast is necessary to detect pancreatic necrosis with CT, MRI can detect necrosis without the need for contrast in patients with acute kidney injury or severe chronic kidney disease. Also, MRI may be better in complicated cases requiring repeated imaging because it does not expose the patient to radiation.

On MRI, pancreatic necrosis appears as a heterogeneous area, owing to its liquid and solid components. Liquid components appear hyperintense, and solid components hypointense, on T2 fluid-weighted imaging. This ability to differentiate the components of a walled-off pancreatic necrosis can be useful in determining whether a collection requires drainage or debridement. MRI is also more sensitive for hemorrhagic complications, best seen on T1 fat-weighted images.^12,16

Magnetic resonance cholangiopancreatography is an excellent method for ductal evaluation through heavily T2-weighted imaging. It is more sensitive than CT for detecting common bile duct stones and can also detect pancreatic duct strictures or extravasation into fluid collections.¹⁶

SUPPORTIVE MANAGEMENT OF EARLY NECROTIZING PANCREATITIS

In the early phase of necrotizing pancreatitis, management is supportive with the primary aim of preventing intravascular volume depletion. Aggressive fluid resuscitation in the first 48 to 72 hours, pain control, and bowel rest are the mainstays of supportive therapy. Intensive care may be necessary if organ failure and hemodynamic instability accompany necrotizing pancreatitis.

Prophylactic antibiotic and antifungal therapy to prevent infected necrosis has been controversial. Recent studies of its utility have not yielded supportive results, and the American College of Gastroenterology and the Infectious Diseases Society of America no longer recommend it.^9,17 These medications should not be given unless concomitant cholangitis or extrapancreatic infection is clinically suspected.

Early enteral nutrition is recommended in patients in whom pancreatitis is predicted to be severe and in those not expected to resume oral intake within 5 to 7 days. Enteral nutrition most commonly involves bedside or endoscopic placement of a nasojejunal feeding tube and collaboration with a nutritionist to determine protein-caloric requirements.

Compared with enteral nutrition, total parenteral nutrition is associated with higher rates of infection, multiorgan dysfunction and failure, and death.¹⁸

Necrotizing pancreatitis is a defining complication of acute pancreatitis, and its presence alone indicates greater severity. However, superimposed complications may further worsen outcomes.

Infected pancreatic necrosis

Infection occurs in approximately 20% of patients with necrotizing pancreatitis and confers a mortality rate of 20% to 50%.¹⁹ Infected pancreatic necrosis occurs when gut organisms translocate into the nearby necrotic pancreatic and peripancreatic tissue. The most commonly identified organisms include Escherichia coli and Enterococcus species.²⁰

This complication usually manifests 2 to 4 weeks after symptom onset; earlier onset is uncommon to rare. It should be considered when the systemic inflammatory response syndrome persists or recurs after 10 days to 2 weeks. Systemic inflammatory response syndrome is also common in sterile necrotizing pancreatitis and sometimes in interstitial pancreatitis, particularly during the first week. However, its sudden appearance or resurgence, high spiking fevers, or worsening organ failure in the later phase (2–4 weeks) of pancreatitis should heighten suspicion of infected pancreatic necrosis.

Imaging may also help diagnose infection, and the presence of gas within a collection or region of necrosis is highly specific. However, the presence of gas is not completely sensitive for infection, as it is seen in only 12% to 22% of infected cases.

Before minimally invasive techniques became available, the diagnosis of infected pancreatic necrosis was confirmed by percutaneous CT-guided aspiration of the necrotic mass or collection for Gram stain and culture.

Antibiotic therapy is indicated in confirmed or suspected cases of infected pancreatic necrosis. Antibiotics with gram-negative coverage and appropriate penetration such as carbapenems, metronidazole, fluoroquinolones, and selected cephalosporins are most commonly used. Meropenem is the antibiotic of choice at our institution.

CT-guided fine-needle aspiration is often done if suspected infected pancreatic necrosis fails to respond to empiric antibiotic therapy.

Debridement or drainage. Generally, the diagnosis or suspicion of infected pancreatic necrosis (suggestive signs are high fever, elevated white blood cell count, and sepsis) warrants an intervention to debride or drain infected pancreatic tissue and control sepsis.²¹

While source control is integral to the successful treatment of infected pancreatic necrosis, antibiotic therapy may provide a bridge to intervention for critically ill patients by suppressing bacteremia and subsequent sepsis. A 2013 meta-analysis found that 324 of 409 patients with suspected infected pancreatic necrosis were successfully stabilized with antibiotic treatment.^21,22 The trend toward conservative management and promising outcomes with antibiotic therapy alone or with minimally invasive techniques has lessened the need for diagnostic CT-guided fine-needle aspiration.

Hemorrhage

Spontaneous hemorrhage into pancreatic necrosis is a rare but life-threatening complication. Because CT is almost always performed with contrast enhancement, this complication is rarely identified with imaging. The diagnosis is made by noting a drop in hemoglobin and hematocrit.

Hemorrhage into the retroperitoneum or the peritoneal cavity, or both, can occur when an inflammatory process erodes into a nearby artery. Luminal gastrointestinal bleeding can occur from gastric varices arising from splenic vein thrombosis and resulting left-sided portal hypertension, or from pseudoaneurysms. These can also bleed into the pancreatic duct (hemosuccus pancreaticus). Pseudoaneurysm is a later complication that occurs when an arterial wall (most commonly the splenic or gastroduodenal artery) is weakened by pancreatic enzymes.²³

Prompt recognition of hemorrhagic events and consultation with an interventional radiologist or surgeon are required to prevent death.

Inflammation and abdominal compartment syndrome

Inflammation from necrotizing pancreatitis can cause further complications by blocking nearby structures. Reported complications include jaundice from biliary compression, hydronephrosis from ureteral compression, bowel obstruction, and gastric outlet obstruction.

Abdominal compartment syndrome is an increasingly recognized complication of acute pancreatitis. Abdominal pressure can rise due to a number of factors, including fluid collections, ascites, ileus, and overly aggressive fluid resuscitation.²⁴ Elevated abdominal pressure is associated with complications such as decreased respiratory compliance, increased peak airway pressure, decreased cardiac preload, hypotension, mesenteric and intestinal ischemia, feeding intolerance, and lower-extremity ischemia and thrombosis.

Patients with necrotizing pancreatitis who have abdominal compartment syndrome have a mortality rate 5 times higher than patients without abdominal compartment syndrome.²⁵

Abdominal pressures should be monitored using a bladder pressure sensor in critically ill or ventilated patients with acute pancreatitis. If the abdominal pressure rises above 20 mm Hg, medical and surgical interventions should be offered in a stepwise fashion to decrease it. Interventions include decompression by nasogastric and rectal tube, sedation or paralysis to relax abdominal wall tension, minimization of intravenous fluids, percutaneous drainage of ascites, and (rarely) surgical midline or subcostal laparotomy.

The treatment of necrotizing pancreatitis has changed rapidly, thanks to a growing experience with minimally invasive techniques.

Indications for intervention

Infected pancreatic necrosis is the primary indication for surgical, percutaneous, or endoscopic intervention.

In sterile necrosis, the threshold for intervention is less clear, and intervention is often reserved for patients who fail to clinically improve or who have intractable abdominal pain, gastric outlet obstruction, or fistulating disease.²⁶

In asymptomatic cases, intervention is almost never indicated regardless of the location or size of the necrotic area.

In walled-off pancreatic necrosis, less-invasive and less-morbid interventions such as endoscopic or percutaneous drainage or video-assisted retroperitoneal debridement can be done.

Timing of intervention

In the past, delaying intervention was thought to increase the risk of death. However, multiple studies have found that outcomes are often worse if intervention is done early, likely due to the lack of a fully formed fibrous wall or demarcation of the necrotic area.²⁷

If the patient remains clinically stable, it is best to delay intervention until at least 4 weeks after the index event to achieve optimal outcomes. Delay can often be achieved by antibiotic treatment to suppress bacteremia and endoscopic or percutaneous drainage of infected collections to control sepsis.

Open surgery

The gold-standard intervention for infected pancreatic necrosis or symptomatic sterile walled-off pancreatic necrosis is open necrosectomy. This involves exploratory laparotomy with blunt debridement of all visible necrotic pancreatic tissue.

Methods to facilitate later evacuation of residual infected fluid and debris vary widely. Multiple large-caliber drains can be placed to facilitate irrigation and drainage before closure of the abdominal fascia. As infected pancreatic necrosis carries the risk of contaminating the peritoneal cavity, the skin is often left open to heal by secondary intention. An interventional radiologist is frequently enlisted to place, exchange, or downsize drainage catheters.

Infected pancreatic necrosis or symptomatic sterile walled-off pancreatic necrosis often requires more than one operation to achieve satisfactory debridement.

The goals of open necrosectomy are to remove nonviable tissue and infection, preserve viable pancreatic tissue, eliminate fistulous connections, and minimize damage to local organs and vasculature.

Minimally invasive techniques

Video-assisted retroperitoneal debridement has been described as a hybrid between endoscopic and open retroperitoneal debridement.²⁸ This technique requires first placing a percutaneous catheter into the necrotic area through the left flank to create a retroperitoneal tract. A 5-cm incision is made and the necrotic space is entered using the drain for guidance. Necrotic tissue is carefully debrided under direct vision using a combination of forceps, irrigation, and suction. A laparoscopic port can also be introduced into the incision when the procedure can no longer be continued under direct vision.^29,30

Although not all patients are candidates for minimal-access surgery, it remains an evolving surgical option.

Endoscopic transmural debridement is another option for infected pancreatic necrosis and symptomatic walled-off pancreatic necrosis. Depending on the location of the necrotic area, an echoendoscope is passed to either the stomach or duodenum. Guided by endoscopic ultrasonography, a needle is passed into the collection, allowing subsequent fistula creation and stenting for internal drainage or debridement. In the past, this process required several steps, multiple devices, fluoroscopic guidance, and considerable time. But newer endoscopic lumen-apposing metal stents have been developed that can be placed in a single step without fluoroscopy. A slimmer endoscope can then be introduced into the necrotic cavity via the stent, and the necrotic debris can be debrided with endoscopic baskets, snares, forceps, and irrigation.^9,31

Similar to surgical necrosectomy, satisfactory debridement is not often obtained with a single procedure; 2 to 5 endoscopic procedures may be needed to achieve resolution. However, the luminal approach in endoscopic necrosectomy avoids the significant morbidity of major abdominal surgery and the potential for pancreaticocutaneous fistulae that may occur with drains.

In a randomized trial comparing endoscopic necrosectomy vs surgical necrosectomy (video-assisted retroperitoneal debridement and exploratory laparotomy),³² endoscopic necrosectomy showed less inflammatory response than surgical necrosectomy and had a lower risk of new-onset organ failure, bleeding, fistula formation, and death.³²

Selecting the best intervention for the individual patient

Given the multiple available techniques, selecting the best intervention for individual patients can be challenging. A team approach with input from a gastroenterologist, surgeon, and interventional radiologist is best when determining which technique would best suit each patient.

Surgical necrosectomy is still the treatment of choice for unstable patients with infected pancreatic necrosis or multiple, inaccessible collections, but current evidence suggests a different approach in stable infected pancreatic necrosis and symptomatic sterile walled-off pancreatic necrosis.

The Dutch Pancreatitis Group²⁸ randomized 88 patients with infected pancreatic necrosis or symptomatic walled-off pancreatic necrosis to open necrosectomy or a minimally invasive “step-up” approach consisting of up to 2 percutaneous drainage or endoscopic debridement procedures before escalation to video-assisted retroperitoneal debridement. The step-up approach resulted in lower rates of morbidity and death than surgical necrosectomy as first-line treatment. Furthermore, some patients in the step-up group avoided the need for surgery entirely.³⁰

Necrosis significantly increases rates of morbidity and mortality in acute pancreatitis. Hospitalists, general internists, and general surgeons are all on the front lines in identifying severe cases and consulting the appropriate specialists for optimal multidisciplinary care. Selective and appropriate timing of radiologic imaging is key, and a vital tool in the management of necrotizing pancreatitis.

While the primary indication for intervention is infected pancreatic necrosis, additional indications are symptomatic walled-off pancreatic necrosis secondary to intractable abdominal pain, bowel obstruction, and failure to thrive. As a result of improving technology and inpatient care, these patients may present with intractable symptoms in the outpatient setting rather than the inpatient setting. The onus is on the primary care physician to maintain a high level of suspicion and refer these patients to subspecialists as appropriate.

Open surgical necrosectomy remains an important approach for care of infected pancreatic necrosis or patients with intractable symptoms. A step-up approach starting with a minimally invasive procedure and escalating if the initial intervention is unsuccessful is gradually becoming the standard of care.

Acute pancreatitis accounted for more than 300,000 admissions and $2.6 billion in associated healthcare costs in the United States in 2012.¹ First-line management is early aggressive fluid resuscitation and analgesics for pain control. Guidelines recommend estimating the clinical severity of each attack using a validated scoring system such as the Bedside Index of Severity in Acute Pancreatitis.² Clinically severe pancreatitis is associated with necrosis.

Acute pancreatitis results from inappropriate activation of zymogens and subsequent auto­digestion of the pancreas by its own enzymes. Though necrotizing pancreatitis is thought to be an ischemic complication, its pathogenesis is not completely understood. Necrosis increases the morbidity and mortality risk of acute pancreatitis because of its association with organ failure and infectious complications. As such, patients with necrotizing pancreatitis may need admission to the intensive care unit, nutritional support, antibiotics, and radiologic, endoscopic, or surgical interventions.

Here, we review current evidence regarding the diagnosis and management of necrotizing pancreatitis.

PROPER TERMINOLOGY HELPS COLLABORATION

Managing necrotizing pancreatitis requires the combined efforts of internists, gastroenterologists, radiologists, and surgeons. This collaboration is aided by proper terminology.

A classification system was devised in Atlanta, GA, in 1992 to facilitate communication and interdisciplinary collaboration.³ Severe pancreatitis was differentiated from mild by the presence of organ failure or the complications of pseudocyst, necrosis, or abscess.

The original Atlanta classification had several limitations. First, the terminology for fluid collections was ambiguous and frequently misused. Second, the assessment of clinical severity required either the Ranson score or the Acute Physiology and Chronic Health Evaluation II score, both of which are complex and have other limitations. Finally, advances in imaging and treatment have rendered the original Atlanta nomenclature obsolete.

In 2012, the Acute Pancreatitis Classification Working Group issued a revised Atlanta classification that modernized the terminology pertaining to natural history, severity, imaging features, and complications. It divides the natural course of acute pancreatitis into early and late phases.⁴

Early vs late phase

In the early phase, findings on computed tomography (CT) neither correlate with clinical severity nor alter clinical management.⁶ Thus, early imaging is not indicated unless there is diagnostic uncertainty, lack of response to appropriate treatment, or sudden deterioration.

Moderate pancreatitis describes patients with pancreatic necrosis with or without transient organ failure (organ dysfunction for ≤ 48 hours).

Severe pancreatitis is defined by pancreatic necrosis and persistent organ dysfunction.⁴ It may be accompanied by pancreatic and peripancreatic fluid collections; bacteremia and sepsis can occur in association with infection of necrotic collections.

Interstitial edematous pancreatitis vs necrotizing pancreatitis

The revised Atlanta classification maintains the original classification of acute pancreatitis into 2 main categories: interstitial edematous pancreatitis and necrotizing pancreatitis.

Necrotizing pancreatitis is further divided into 3 subtypes based on extent and location of necrosis:

• Parenchymal necrosis alone (5% of cases)
• Necrosis of peripancreatic fat alone (20%)
• Necrosis of both parenchyma and peripancreatic fat (75%).

Peripancreatic involvement is commonly found in the mesentery, peripancreatic and distant retroperitoneum, and lesser sac.

Of the three subtypes, peripancreatic necrosis has the best prognosis. However, all of the subtypes of necrotizing pancreatitis are associated with poorer outcomes than interstitial edematous pancreatitis.

Fluid collections

Acute pancreatic fluid collections contain exclusively nonsolid components without an inflammatory wall and are typically found in the peripancreatic fat. These collections often resolve without intervention as the patient recovers. If they persist beyond 4 weeks and develop a nonepithelialized, fibrous wall, they become pseudocysts. Intervention is generally not recommended for pseudocysts unless they are symptomatic.

Radiographic imaging is not usually necessary to diagnose acute pancreatitis. However, it can be a valuable tool to clarify an ambiguous presentation, determine severity, and identify complications.

The timing and appropriate type of imaging are integral to obtaining useful data. Any imaging obtained in acute pancreatitis to evaluate necrosis should be performed at least 3 to 5 days from the initial symptom onset; if imaging is obtained before 72 hours, necrosis cannot be confidently excluded.⁸

COMPUTED TOMOGRAPHY

CT is the imaging test of choice when evaluating acute pancreatitis. In addition, almost all percutaneous interventions are performed with CT guidance. The Balthazar score is the most well-known CT severity index. It is calculated based on the degree of inflammation, acute fluid collections, and parenchymal necrosis.⁹ However, a modified severity index incorporates extrapancreatic complications such as ascites and vascular compromise and was found to more strongly correlate with outcomes than the standard Balthazar score.¹⁰

Contrast-enhanced CT is performed in 2 phases:

The pancreatic parenchymal phase

The pancreatic parenchymal or late arterial phase is obtained approximately 40 to 45 seconds after the start of the contrast bolus. It is used to detect necrosis in the early phase of acute pancreatitis and to assess the peripancreatic arteries for pseudoaneurysms in the late phase of acute pancreatitis.¹¹

Pancreatic necrosis appears as an area of decreased parenchymal enhancement, either well-defined or heterogeneous. The normal pancreatic parenchyma has a postcontrast enhancement pattern similar to that of the spleen. Parenchyma that does not enhance to the same degree is considered necrotic. The severity of necrosis is graded based on the percentage of the pancreas involved (< 30%, 30%–50%, or > 50%), and a higher percentage correlates with a worse outcome.^12,13

Peripancreatic necrosis is harder to detect, as there is no method to assess fat enhancement as there is with pancreatic parenchymal enhancement. In general, radiologists assume that heterogeneous peripancreatic changes, including areas of fat, fluid, and soft tissue attenuation, are consistent with peripancreatic necrosis. After 7 to 10 days, if these changes become more homogeneous and confluent with a more mass-like process, peripancreatic necrosis can be more confidently identified.^12,13

The portal venous phase

The later, portal venous phase of the scan is obtained approximately 70 seconds after the start of the contrast bolus. It is used to detect and characterize fluid collections and venous complications of the disease.

Drawbacks of CT

A drawback of CT is the need for iodinated intravenous contrast media, which in severely ill patients may precipitate or worsen pre-existing acute kidney injury.

Further, several studies have shown that findings on CT rarely alter the management of patients in the early phase of acute pancreatitis and in fact may be an overuse of medical resources.¹⁴ Unless there are confounding clinical signs or symptoms, CT should be delayed for at least 72 hours.^9,10,14,15

MAGNETIC RESONANCE IMAGING

Magnetic resonance imaging (MRI) is not a first-line imaging test in this disease because it is not as available as CT and takes longer to perform—20 to 30 minutes. The patient must be evaluated for candidacy, as it is difficult for acutely ill patients to tolerate an examination that takes this long and requires them to hold their breath multiple times.

MRI is an appropriate alternative in patients who are pregnant or who have severe iodinated-contrast allergy. While contrast is necessary to detect pancreatic necrosis with CT, MRI can detect necrosis without the need for contrast in patients with acute kidney injury or severe chronic kidney disease. Also, MRI may be better in complicated cases requiring repeated imaging because it does not expose the patient to radiation.

On MRI, pancreatic necrosis appears as a heterogeneous area, owing to its liquid and solid components. Liquid components appear hyperintense, and solid components hypointense, on T2 fluid-weighted imaging. This ability to differentiate the components of a walled-off pancreatic necrosis can be useful in determining whether a collection requires drainage or debridement. MRI is also more sensitive for hemorrhagic complications, best seen on T1 fat-weighted images.^12,16

Magnetic resonance cholangiopancreatography is an excellent method for ductal evaluation through heavily T2-weighted imaging. It is more sensitive than CT for detecting common bile duct stones and can also detect pancreatic duct strictures or extravasation into fluid collections.¹⁶

SUPPORTIVE MANAGEMENT OF EARLY NECROTIZING PANCREATITIS

In the early phase of necrotizing pancreatitis, management is supportive with the primary aim of preventing intravascular volume depletion. Aggressive fluid resuscitation in the first 48 to 72 hours, pain control, and bowel rest are the mainstays of supportive therapy. Intensive care may be necessary if organ failure and hemodynamic instability accompany necrotizing pancreatitis.

Prophylactic antibiotic and antifungal therapy to prevent infected necrosis has been controversial. Recent studies of its utility have not yielded supportive results, and the American College of Gastroenterology and the Infectious Diseases Society of America no longer recommend it.^9,17 These medications should not be given unless concomitant cholangitis or extrapancreatic infection is clinically suspected.

Early enteral nutrition is recommended in patients in whom pancreatitis is predicted to be severe and in those not expected to resume oral intake within 5 to 7 days. Enteral nutrition most commonly involves bedside or endoscopic placement of a nasojejunal feeding tube and collaboration with a nutritionist to determine protein-caloric requirements.

Compared with enteral nutrition, total parenteral nutrition is associated with higher rates of infection, multiorgan dysfunction and failure, and death.¹⁸

Necrotizing pancreatitis is a defining complication of acute pancreatitis, and its presence alone indicates greater severity. However, superimposed complications may further worsen outcomes.

Infected pancreatic necrosis

Infection occurs in approximately 20% of patients with necrotizing pancreatitis and confers a mortality rate of 20% to 50%.¹⁹ Infected pancreatic necrosis occurs when gut organisms translocate into the nearby necrotic pancreatic and peripancreatic tissue. The most commonly identified organisms include Escherichia coli and Enterococcus species.²⁰

This complication usually manifests 2 to 4 weeks after symptom onset; earlier onset is uncommon to rare. It should be considered when the systemic inflammatory response syndrome persists or recurs after 10 days to 2 weeks. Systemic inflammatory response syndrome is also common in sterile necrotizing pancreatitis and sometimes in interstitial pancreatitis, particularly during the first week. However, its sudden appearance or resurgence, high spiking fevers, or worsening organ failure in the later phase (2–4 weeks) of pancreatitis should heighten suspicion of infected pancreatic necrosis.

Imaging may also help diagnose infection, and the presence of gas within a collection or region of necrosis is highly specific. However, the presence of gas is not completely sensitive for infection, as it is seen in only 12% to 22% of infected cases.

Before minimally invasive techniques became available, the diagnosis of infected pancreatic necrosis was confirmed by percutaneous CT-guided aspiration of the necrotic mass or collection for Gram stain and culture.

Antibiotic therapy is indicated in confirmed or suspected cases of infected pancreatic necrosis. Antibiotics with gram-negative coverage and appropriate penetration such as carbapenems, metronidazole, fluoroquinolones, and selected cephalosporins are most commonly used. Meropenem is the antibiotic of choice at our institution.

CT-guided fine-needle aspiration is often done if suspected infected pancreatic necrosis fails to respond to empiric antibiotic therapy.

Debridement or drainage. Generally, the diagnosis or suspicion of infected pancreatic necrosis (suggestive signs are high fever, elevated white blood cell count, and sepsis) warrants an intervention to debride or drain infected pancreatic tissue and control sepsis.²¹

While source control is integral to the successful treatment of infected pancreatic necrosis, antibiotic therapy may provide a bridge to intervention for critically ill patients by suppressing bacteremia and subsequent sepsis. A 2013 meta-analysis found that 324 of 409 patients with suspected infected pancreatic necrosis were successfully stabilized with antibiotic treatment.^21,22 The trend toward conservative management and promising outcomes with antibiotic therapy alone or with minimally invasive techniques has lessened the need for diagnostic CT-guided fine-needle aspiration.

Hemorrhage

Spontaneous hemorrhage into pancreatic necrosis is a rare but life-threatening complication. Because CT is almost always performed with contrast enhancement, this complication is rarely identified with imaging. The diagnosis is made by noting a drop in hemoglobin and hematocrit.

Hemorrhage into the retroperitoneum or the peritoneal cavity, or both, can occur when an inflammatory process erodes into a nearby artery. Luminal gastrointestinal bleeding can occur from gastric varices arising from splenic vein thrombosis and resulting left-sided portal hypertension, or from pseudoaneurysms. These can also bleed into the pancreatic duct (hemosuccus pancreaticus). Pseudoaneurysm is a later complication that occurs when an arterial wall (most commonly the splenic or gastroduodenal artery) is weakened by pancreatic enzymes.²³

Prompt recognition of hemorrhagic events and consultation with an interventional radiologist or surgeon are required to prevent death.

Inflammation and abdominal compartment syndrome

Inflammation from necrotizing pancreatitis can cause further complications by blocking nearby structures. Reported complications include jaundice from biliary compression, hydronephrosis from ureteral compression, bowel obstruction, and gastric outlet obstruction.

Abdominal compartment syndrome is an increasingly recognized complication of acute pancreatitis. Abdominal pressure can rise due to a number of factors, including fluid collections, ascites, ileus, and overly aggressive fluid resuscitation.²⁴ Elevated abdominal pressure is associated with complications such as decreased respiratory compliance, increased peak airway pressure, decreased cardiac preload, hypotension, mesenteric and intestinal ischemia, feeding intolerance, and lower-extremity ischemia and thrombosis.

Patients with necrotizing pancreatitis who have abdominal compartment syndrome have a mortality rate 5 times higher than patients without abdominal compartment syndrome.²⁵

Abdominal pressures should be monitored using a bladder pressure sensor in critically ill or ventilated patients with acute pancreatitis. If the abdominal pressure rises above 20 mm Hg, medical and surgical interventions should be offered in a stepwise fashion to decrease it. Interventions include decompression by nasogastric and rectal tube, sedation or paralysis to relax abdominal wall tension, minimization of intravenous fluids, percutaneous drainage of ascites, and (rarely) surgical midline or subcostal laparotomy.

The treatment of necrotizing pancreatitis has changed rapidly, thanks to a growing experience with minimally invasive techniques.

Indications for intervention

Infected pancreatic necrosis is the primary indication for surgical, percutaneous, or endoscopic intervention.

In sterile necrosis, the threshold for intervention is less clear, and intervention is often reserved for patients who fail to clinically improve or who have intractable abdominal pain, gastric outlet obstruction, or fistulating disease.²⁶

In asymptomatic cases, intervention is almost never indicated regardless of the location or size of the necrotic area.

In walled-off pancreatic necrosis, less-invasive and less-morbid interventions such as endoscopic or percutaneous drainage or video-assisted retroperitoneal debridement can be done.

Timing of intervention

In the past, delaying intervention was thought to increase the risk of death. However, multiple studies have found that outcomes are often worse if intervention is done early, likely due to the lack of a fully formed fibrous wall or demarcation of the necrotic area.²⁷

If the patient remains clinically stable, it is best to delay intervention until at least 4 weeks after the index event to achieve optimal outcomes. Delay can often be achieved by antibiotic treatment to suppress bacteremia and endoscopic or percutaneous drainage of infected collections to control sepsis.

Open surgery

The gold-standard intervention for infected pancreatic necrosis or symptomatic sterile walled-off pancreatic necrosis is open necrosectomy. This involves exploratory laparotomy with blunt debridement of all visible necrotic pancreatic tissue.

Methods to facilitate later evacuation of residual infected fluid and debris vary widely. Multiple large-caliber drains can be placed to facilitate irrigation and drainage before closure of the abdominal fascia. As infected pancreatic necrosis carries the risk of contaminating the peritoneal cavity, the skin is often left open to heal by secondary intention. An interventional radiologist is frequently enlisted to place, exchange, or downsize drainage catheters.

Infected pancreatic necrosis or symptomatic sterile walled-off pancreatic necrosis often requires more than one operation to achieve satisfactory debridement.

The goals of open necrosectomy are to remove nonviable tissue and infection, preserve viable pancreatic tissue, eliminate fistulous connections, and minimize damage to local organs and vasculature.

Minimally invasive techniques

Video-assisted retroperitoneal debridement has been described as a hybrid between endoscopic and open retroperitoneal debridement.²⁸ This technique requires first placing a percutaneous catheter into the necrotic area through the left flank to create a retroperitoneal tract. A 5-cm incision is made and the necrotic space is entered using the drain for guidance. Necrotic tissue is carefully debrided under direct vision using a combination of forceps, irrigation, and suction. A laparoscopic port can also be introduced into the incision when the procedure can no longer be continued under direct vision.^29,30

Although not all patients are candidates for minimal-access surgery, it remains an evolving surgical option.

Endoscopic transmural debridement is another option for infected pancreatic necrosis and symptomatic walled-off pancreatic necrosis. Depending on the location of the necrotic area, an echoendoscope is passed to either the stomach or duodenum. Guided by endoscopic ultrasonography, a needle is passed into the collection, allowing subsequent fistula creation and stenting for internal drainage or debridement. In the past, this process required several steps, multiple devices, fluoroscopic guidance, and considerable time. But newer endoscopic lumen-apposing metal stents have been developed that can be placed in a single step without fluoroscopy. A slimmer endoscope can then be introduced into the necrotic cavity via the stent, and the necrotic debris can be debrided with endoscopic baskets, snares, forceps, and irrigation.^9,31

Similar to surgical necrosectomy, satisfactory debridement is not often obtained with a single procedure; 2 to 5 endoscopic procedures may be needed to achieve resolution. However, the luminal approach in endoscopic necrosectomy avoids the significant morbidity of major abdominal surgery and the potential for pancreaticocutaneous fistulae that may occur with drains.

In a randomized trial comparing endoscopic necrosectomy vs surgical necrosectomy (video-assisted retroperitoneal debridement and exploratory laparotomy),³² endoscopic necrosectomy showed less inflammatory response than surgical necrosectomy and had a lower risk of new-onset organ failure, bleeding, fistula formation, and death.³²

Selecting the best intervention for the individual patient

Given the multiple available techniques, selecting the best intervention for individual patients can be challenging. A team approach with input from a gastroenterologist, surgeon, and interventional radiologist is best when determining which technique would best suit each patient.

Surgical necrosectomy is still the treatment of choice for unstable patients with infected pancreatic necrosis or multiple, inaccessible collections, but current evidence suggests a different approach in stable infected pancreatic necrosis and symptomatic sterile walled-off pancreatic necrosis.

The Dutch Pancreatitis Group²⁸ randomized 88 patients with infected pancreatic necrosis or symptomatic walled-off pancreatic necrosis to open necrosectomy or a minimally invasive “step-up” approach consisting of up to 2 percutaneous drainage or endoscopic debridement procedures before escalation to video-assisted retroperitoneal debridement. The step-up approach resulted in lower rates of morbidity and death than surgical necrosectomy as first-line treatment. Furthermore, some patients in the step-up group avoided the need for surgery entirely.³⁰

Necrosis significantly increases rates of morbidity and mortality in acute pancreatitis. Hospitalists, general internists, and general surgeons are all on the front lines in identifying severe cases and consulting the appropriate specialists for optimal multidisciplinary care. Selective and appropriate timing of radiologic imaging is key, and a vital tool in the management of necrotizing pancreatitis.

While the primary indication for intervention is infected pancreatic necrosis, additional indications are symptomatic walled-off pancreatic necrosis secondary to intractable abdominal pain, bowel obstruction, and failure to thrive. As a result of improving technology and inpatient care, these patients may present with intractable symptoms in the outpatient setting rather than the inpatient setting. The onus is on the primary care physician to maintain a high level of suspicion and refer these patients to subspecialists as appropriate.

Open surgical necrosectomy remains an important approach for care of infected pancreatic necrosis or patients with intractable symptoms. A step-up approach starting with a minimally invasive procedure and escalating if the initial intervention is unsuccessful is gradually becoming the standard of care.

For some patients with chronic pancreatitis, the best option is to remove the entire pancreas. This does not necessarily doom the patient to diabetes mellitus, because we can harvest the islet cells and reinsert them so that, lodged in the liver, they can continue making insulin. However, this approach is underemphasized in the general medical literature and is likely underutilized in the United States.

Here, we discuss chronic pancreatitis, the indications for and contraindications to this procedure, its outcomes, and the management of patients who undergo it.

CHRONIC PANCREATITIS IS PROGRESSIVE AND PAINFUL

Chronic pancreatitis is a progressive condition characterized by chronic inflammation, irreversible fibrosis, and scarring, resulting in loss of both exocrine and endocrine tissue.

According to a National Institutes of Health database, pancreatitis is the seventh most common digestive disease diagnosis on hospitalization, with annual healthcare costs exceeding $3 billion.¹ Although data are scarce, by some estimates the incidence of chronic pancreatitis ranges from 4 to 14 per 100,000 person-years, and the prevalence ranges from 26.4 to 52 per 100,000.^2–4 Moreover, a meta-analysis⁵ found that acute pancreatitis progresses to chronic pancreatitis in 10% of patients who have a first episode of acute pancreatitis and in 36% who have recurrent episodes.

Historically, alcoholism was and still is the most common cause of chronic pancreatitis, contributing to 60% to 90% of cases in Western countries.^6,7 However, cases due to nonalcoholic causes have been increasing, and in more than one-fourth of patients, no identifiable cause is found.^6,8 Smoking is an independent risk factor.^6,8,9 Some cases can be linked to genetic abnormalities, particularly in children.¹⁰

The clinical manifestations of chronic pancreatitis include exocrine pancreatic insufficiency (leading to malnutrition and steatorrhea), endocrine insufficiency (causing diabetes mellitus), and intractable pain.¹¹ Pain is the predominant clinical symptom early in the disease and is often debilitating and difficult to manage. Uncontrolled pain has a devastating impact on quality of life and may become complicated by narcotic dependence.

The pain of chronic pancreatitis is often multifactorial, with mechanisms that include increased intraductal pressure from obstruction of the pancreatic duct, pancreatic ischemia, neuronal injury, and neuroimmune interactions between neuronal processes and chronic inflammation.¹²

Treatment: Medical and surgical

In chronic pancreatitis, the aim of treatment is to alleviate deficiencies of exocrine and endocrine function and mitigate the pain. Patients who smoke or drink alcohol should be strongly encouraged to quit.

Loss of exocrine function is mainly managed with oral pancreatic enzyme supplements, and diabetes control is often attained with insulin therapy.¹³ Besides helping digestion, pancreatic enzyme therapy in the form of nonenteric tablets may also reduce pain and pancreatitis attacks.¹⁴ The mechanism may be by degrading cholecystokinin-releasing factor in the duodenum, lowering cholecystokinin levels and thereby reducing pain.¹²

Nonnarcotic analgesics are often the first line of therapy for pain management, but many patients need narcotic analgesics. Along with narcotics, adjunctive agents such as tricyclic antidepressants, serotonin-norepinephrine reuptake inhibitors, selective serotonin reuptake inhibitors, and gabapentinoids have been used to treat chronic pancreatitis pain, but with limited success.¹⁵

In patients for whom medical pain management has failed, one can consider another option, such as nerve block, neurolysis, or endoscopic or surgical therapy. Neuromodulators are often prescribed by pain clinics.¹⁵ Percutaneous and endoscopic celiac ganglion blocks can be an option but rarely achieve substantial or permanent pain relief, and the induced transient responses (on average 2 to 4 months) often cannot be repeated.^14–17

Surgical options to relieve pain try to preserve pancreatic function and vary depending on the degree of severity and nature of pancreatic damage. In broad terms, the surgical procedures can be divided into two types:

• Drainage procedures (eg, pseudocyst drainage; minimally invasive endoscopic duct drainage via sphincterotomy or stent placement, or both; pancreaticojejunostomy)
• Resectional procedures (eg, distal pancreatectomy, isolated head resection, pancreaticoduodenectomy, Whipple procedure, total pancreatectomy).

In carefully selected patients, total pancreatectomy can be offered to remove the cause of the pain.¹⁸ This procedure is most often performed in patients who have small-duct disease or a genetic cause or for whom other surgical procedures have failed.¹¹

HISTORY OF THE PROCEDURE

Islet cell transplantation grew out of visionary work by Paul Lacy and David Scharp at the University of Washington at Seattle, whose research focused on isolating and transplanting islet cells in rodent models. The topic has been reviewed by Jahansouz et al.¹⁹ In the 1970s, experiments in pancreatectomized dogs showed that transplanting unpurified pancreatic islet tissue that was dispersed by collagenase digestion into the spleen or portal vein could prevent diabetes.^20,21 In 1974, the first human trials of transplanting islet cells were conducted, using isolated islets from cadaveric donors to treat diabetes.¹⁹

In the past, pancreatectomy was performed to treat painful chronic pancreatitis, but it was viewed as undesirable because removing the gland would inevitably cause insulin-dependent diabetes.²² That changed in 1977 at the University of Minnesota, with the first reported islet cell autotransplant after pancreatectomy. The patient remained pain-free and insulin-independent long-term.²³ This seminal case showed that endocrine function could be preserved by autotransplant of islets prepared from the excised pancreas.²⁴

In 1992, Pyzdrowski et al²⁵ reported that intrahepatic transplant of as few as 265,000 islets was enough to prevent the need for insulin therapy. Since this technique was first described, there have been many advances, and now more than 30 centers worldwide do it.

PRIMARY INDICATION: INTRACTABLE PAIN

Interest has been growing in using total pancreatectomy and islet autotransplant to treat recurrent acute pancreatitis, chronic pancreatitis, and hereditary pancreatitis. The rationale is that removing the offending tissue eliminates pancreatitis, pain, and cancer risk, while preserving and replacing the islet cells prevents the development of brittle diabetes with loss of insulin and glucagon.²⁶

Proposed criteria for total pancreatectomy and islet autotransplant

Bellin et al¹⁴ proposed five criteria for patient selection for this procedure based on imaging studies, pancreatic function tests, and histopathology to detect pancreatic fibrosis. Patients must fulfill all five of the following criteria:

Criterion 1. Diagnosis of chronic pancreatitis, based on chronic abdominal pain lasting more than 6 months with either at least one of the following:

• Pancreatic calcifications on computed tomography
• At least two of the following: four or more of nine criteria on endoscopic ultrasonography described by Catalano et al,²⁷ a compatible ductal or parenchymal abnormality on secretin magnetic resonance cholangiopancreatography; abnormal endoscopic pancreatic function test (peak HCO₂ ≤ 80 mmol/L)
• Histopathologically confirmed diagnosis of chronic pancreatitis
• Compatible clinical history and documented hereditary pancreatitis (PRSS1 gene mutation)

OR

• History of recurrent acute pancreatitis (more than one episode of characteristic pain associated with imaging diagnostic of acute pancreatitis or elevated serum amylase or lipase > 3 times the upper limit of normal).

Criterion 2. At least one of the following:

• Daily narcotic dependence
• Pain resulting in impaired quality of life, which may include inability to attend school, recurrent hospitalizations, or inability to participate in usual age-appropriate activities.

Criterion 3. Complete evaluation with no reversible cause of pancreatitis present or untreated.

Criterion 4. Failure to respond to maximal medical and endoscopic therapy.

Criterion 5. Adequate islet cell function (nondiabetic or C-peptide-positive). Patients with C-peptide-negative diabetes meeting criteria 1 to 4 are candidates for total pancreatectomy alone.

The primary goal is to treat intractable pain and improve quality of life in selected patients with chronic pancreatitis or recurrent acute pancreatitis when endoscopic and prior surgical therapies have failed, and whose impairment due to pain is substantial enough to accept the risk of postoperative insulin-dependent diabetes and lifelong commitment to pancreatic enzyme replacement therapy.^15,26 Patients with a known genetic cause of chronic pancreatitis should be offered special consideration for the procedure, as their disease is unlikely to remit.

CONTRAINDICATIONS

Total pancreatectomy and islet autotransplant should not be performed in patients with active alcoholism, illicit drug use, or untreated or poorly controlled psychiatric illnesses that could impair the patient’s ability to adhere to a complicated postoperative medical regimen.

A poor support network may be a relative contraindication in view of the cost and complexity of diabetic and pancreatic enzyme replacement therapy.^18,26

Islet cell autotransplant is contraindicated in patients with conditions such as C-peptide-negative or type 1 diabetes or a history of portal vein thrombosis, portal hypertension, significant liver disease, high-risk cardiopulmonary disease, or pancreatic cancer (Table 1).²⁶

WHEN TO CONSIDER REFERRAL FOR THIS PROCEDURE

The choice of total pancreatectomy and islet autotransplant vs conventional surgery must be individualized on the basis of each patient’s anatomy, comorbidities, symptom burden, presence or degree of diabetes, and rate of disease progression. The most important factors to consider in determining the need for and timing of this procedure are the patient’s pain, narcotic requirements, and impaired ability to function.²⁶

Sooner rather than later?

An argument can be made for performing this procedure sooner in the course of the disease rather than later when all else has failed. First, prolonged pain can result in central sensitization, in which the threshold for perceiving pain is lowered by damage to the nociceptive neurons from repeated stimulation and inflammation.²⁸

Further, prolonged opioid therapy can lead to opioid-induced hyperalgesia, which may also render patients more sensitive to pain and aggravate their preexisting pain.^26,28

In addition, although operative drainage procedures and partial resections are often considered the gold standard for chronic pancreatitis management, patients who undergo partial pancreatectomy or lateral pancreaticojejunostomy (Puestow procedure) have fewer islet cells left to harvest (about 50% fewer) if they subsequently undergo total pancreatectomy and islet cell autotransplant.^22,26

Therefore, performing this procedure earlier may help the patient avoid chronic pain syndromes and complications of chronic opioid use, including hyperalgesia, and give the best chance of harvesting enough islet cells to prevent or minimize diabetes afterward.¹¹

REMOVING THE PANCREAS, RETURNING THE ISLET CELLS

During this procedure, the blood supply to the pancreas must be preserved until just before its removal to minimize warm ischemia of the islet cells.^18,29 Although there are several surgical variations, a pylorus-preserving total pancreatectomy with duodenectomy is typically performed, usually with splenectomy to preserve perfusion to the body and tail.³⁰

The resected pancreas is taken to the islet isolation laboratory. There, the pancreatic duct is cannulated to fill the organ with a cold collagenase solution, followed by gentle mechanical dispersion using the semiautomated Ricordi method,³¹ which separates the islet cells from the exocrine tissue.³²

The number of islet cells is quantified as islet equivalents; 1 islet equivalent is equal to the volume of an islet with a diameter of 150 µm. Islet equivalents per kilogram of body weight is the unit commonly used to report the graft amount transplanted.³³

After digestion, the islet cells can be purified or partially purified by a gradient separation method using a Cobe 2991 cell processor (Terumo Corporation, Tokyo, Japan),³⁴ or can be transplanted as an unpurified preparation. In islet cell autotransplant for chronic pancreatitis, purification is not always necessary due to the small tissue volume extracted from the often atrophic and fibrotic pancreas.³² The decision to purify depends on the postdigest tissue volume; usually, a tissue volume greater than 0.25 mL/kg body weight is an indication to at least partially purify.^18,35

The final preparation is returned to the operating room, and after heparin is given, the islets are infused into the portal system using a stump of the splenic vein, or alternatively through direct puncture of the portal vein or cannulation of the umbilical vein.^32,36 If the portal vein pressure reaches 25 cm H₂O, the infusion is stopped and the remaining islets can be placed in the peritoneal cavity or elsewhere.¹⁸ Transplant of the islets into the liver or peritoneum allows the islets to secrete insulin into the hepatic portal circulation, which is the route used by the native pancreas.²²

CONTROLLING GLUCOSE DURING AND AFTER THE PROCEDURE

Animal studies have shown that hyperglycemia impairs islet revascularization,³⁷ and glucose toxicity may cause dysfunction and structural lesions of the transplanted islets.^11,38

Therefore, during and after the procedure, most centers maintain euglycemia by an intravenous insulin infusion and subsequently move to subcutaneous insulin when the patient starts eating again. Some centers continue insulin at discharge and gradually taper it over months, even in patients who can possibly achieve euglycemia without it.

OUTCOMES

Many institutions have reported their clinical outcomes in terms of pain relief, islet function, glycemic control, and improvement of quality of life. The largest series have been from the University of Minnesota, Leicester General Hospital, the University of Cincinnati, and the Medical University of South Carolina.

Insulin independence is common but wanes with time

The ability to achieve insulin independence after islet autotransplant appears to be related to the number of islets transplanted, with the best results when more than 2,000 or 3,000 islet equivalents/kg are transplanted.^39,40

Sutherland et al¹⁸ reported that of 409 patients who underwent islet cell autotransplant at the University of Minnesota (the largest series reported to date), 30% were insulin-independent at 3 years, 33% had partial graft function (defined by positive C-peptide), and 82% achieved a mean hemoglobin A_1c of less than 7%. However, in the subset who received more than 5,000 islet equivalents/kg, nearly three-fourths of patients were insulin-independent at 3 years.

The Leicester General Hospital group presented long-term data on 46 patients who underwent total pancreatectomy and islet cell autotransplant. Twelve of the 46 had shown periods of insulin independence for a median of 16.5 months, and 5 remained insulin-free at the time of the publication.⁴¹ Over the 10 years of follow-up, insulin requirements and hemoglobin A_1c increased notably. However, all of the patients tested C-peptide-positive, suggesting long-lasting graft function.

Most recently, the University of Cincinnati group reported long-term data on 116 patients. The insulin independence rate was 38% at 1 year, decreasing to 27% at 5 years. The number of patients with partial graft function was 38% at 1 year and 35% at 5 years.⁴²

Thus, all three institutions confirmed that the autotransplanted islets continue to secrete insulin long-term, but that function decreases over time.

Pancreatectomy reduces pain

Multiple studies have shown that total pancreatectomy reduces pain in patients with chronic pancreatitis. Ahmad et al⁴³ reported a marked reduction in narcotic use (mean morphine equivalents 206 mg/day before surgery, compared with 90 mg after), and a 58% reduction in pain as demonstrated by narcotic independence.

In the University of Minnesota series, 85% of the 409 patients had less pain at 2 years, and 59% were able to stop taking narcotics.¹⁸

The University of Cincinnati group reported a narcotic independence rate of 55% at 1 year, which continued to improve to 73% at 5 years.⁴²

Although the source of pain is removed, pain persists or recurs in 10% to 20% of patients after total pancreatectomy and islet cell autotransplant, showing that the pathogenesis of pain is complex, and some uncertainty exists about it.²⁶

Quality of life

Reports evaluating health-related quality of life after total pancreatectomy and islet autotransplant are limited.

The University of Cincinnati group reported the long-term outcomes of quality of life as measured by the Short Form 36 Health Survey.⁴² Ninety-two percent of patients reported overall improvement in their health at 1 year, and 85% continued to report improved health more than 5 years after the surgery.

In a series of 20 patients, 79% to 90% reported improvements in the seven various domains of the Pain Disability Index. In addition, 60% showed improvement in depression and 70% showed improvement in anxiety. The greatest improvements were in those who had not undergone prior pancreatic surgery, who were younger, and in those with higher levels of preoperative pain.³⁰

Similarly, in a series of 74 patients, the Medical University of South Carolina group reported significant improvement in physical and mental health components of the Short Form 12 Health Survey and an associated decrease in daily narcotic requirements. Moreover, the need to start or increase the dose of insulin after the surgery was not associated with a lower quality of life.⁴⁴

OFF-SITE ISLET CELL ISOLATION

Despite the positive outcomes in terms of pain relief and insulin independence in many patients after total pancreatectomy and islet cell autotransplant, few medical centers have an on-site islet-processing facility. Since the mid-1990s, a few centers have been able to circumvent this limitation by working with off-site islet cell isolation laboratories.^45,46

Whether and when to consider this procedure must be individualized

The University of California, Los Angeles, first reported on a series of nine patients who received autologous islet cells after near-total or total pancreatectomy using a remote islet cell isolation facility, with results comparable to those of other large institutions.⁴⁵

Similarly, the procedure has been performed at Cleveland Clinic since 2007 with the collaboration of an off-site islet cell isolation laboratory at the University of Pittsburgh. A cohort study from this collaboration published in 2015 showed that in 36 patients (mean follow-up 28 months, range 3–26 months), 33% were insulin-independent, with a C-peptide-positive rate of 70%. This is the largest cohort to date from a center utilizing an off-site islet isolation facility.⁴⁷

In view of the positive outcomes at these centers, lack of a local islet-processing facility should no longer be a barrier to total pancreatectomy and islet cell autotransplant.

PATIENT CARE AFTER THE PROCEDURE

A multidisciplinary team is an essential component of the postoperative management of patients who undergo total pancreatectomy and islet cell autotransplant.

For patients who had been receiving narcotics for a long time before surgery or who were receiving frequent doses, an experienced pain management physician should be involved in the patient’s postoperative care.

Because islet function can wane over time, testing for diabetes should be done at least annually for the rest of the patient’s life and should include fasting plasma glucose, hemoglobin A_1c, and C-peptide, along with self-monitored blood glucose.²⁶

All patients who have surgically induced exocrine insufficiency are at risk of malabsorption and fat-soluble vitamin deficiencies.⁴⁸ Hence, lifelong pancreatic enzyme replacement therapy is mandatory. Nutritional monitoring should include assessment of steatorrhea, body composition, and fat-soluble vitamin levels (vitamins A, D, and E) at least every year.²⁶ Patients with chronic pancreatitis are at increased risk for low bone density from malabsorption of vitamin D and calcium; therefore, it is recommended that a dual-energy x-ray absorptiometry bone density scan be obtained.^26,49

Patients who undergo splenectomy as part of their procedure will require appropriate precautions and ongoing vaccinations as recommended by the US Centers for Disease Control and Prevention.^26,50,51

WHAT TO EXPECT FOR THE FUTURE

The National Institute of Diabetes and Digestive and Kidney Diseases has reviewed the potential future research directions for total pancreatectomy and islet autotransplant.¹⁵

The more islet cells transplanted, the better the chance of insulin independence

Patient selection remains challenging despite the availability of criteria¹⁵ and guidelines.²⁶ More research is needed to better assess preoperative beta-cell function and to predict postoperative outcomes. Mixed meal-tolerance testing is adopted by most clinical centers to predict posttransplant beta-cell function. The use of arginine instead of glucagon in a stimulation test for insulin and C-peptide response has been validated and may allow more accurate assessment.^52,53

Another targeted area of research is the advancement of safety and metabolic outcomes. Techniques to minimize warm ischemic time and complications are being evaluated. Islet isolation methods that yield more islets, reduce beta-cell apoptosis, and can isolate islets from glands with malignancy should be further investigated.⁵⁴ Further, enhanced islet infusion methods that achieve lower portal venous pressures and minimize portal vein thrombosis are needed.

Unfortunately, the function of transplanted islet grafts declines over time. This phenomenon is at least partially attributed to the immediate blood-mediated inflammatory response,^55,56 along with islet hypoxia,⁵⁷ leading to islet apoptosis. Research on different strategies is expanding our knowledge in islet engraftment and posttransplant beta-cell apoptosis, with the expectation that the transplanted islet lifespan will increase. Alternative transplant sites with low inflammatory reaction, such as the omental pouch,⁵⁸ muscle,⁵⁹ and bone marrow,⁶⁰ have shown encouraging data. Other approaches, such as adjuvant anti-inflammatory agents and heparinization, have been proposed.¹⁵

Research into complications is also of clinical importance. There is growing attention to hypoglycemia unrelated to exogenous insulin use in posttransplant patients. One hypothesis is that glucagon secretion, a counterregulatory response to hypoglycemia, is defective if the islet cells are transplanted into the liver, and that implanting them into another site may avoid this effect.⁶¹

For some patients with chronic pancreatitis, the best option is to remove the entire pancreas. This does not necessarily doom the patient to diabetes mellitus, because we can harvest the islet cells and reinsert them so that, lodged in the liver, they can continue making insulin. However, this approach is underemphasized in the general medical literature and is likely underutilized in the United States.

Here, we discuss chronic pancreatitis, the indications for and contraindications to this procedure, its outcomes, and the management of patients who undergo it.

CHRONIC PANCREATITIS IS PROGRESSIVE AND PAINFUL

Chronic pancreatitis is a progressive condition characterized by chronic inflammation, irreversible fibrosis, and scarring, resulting in loss of both exocrine and endocrine tissue.

According to a National Institutes of Health database, pancreatitis is the seventh most common digestive disease diagnosis on hospitalization, with annual healthcare costs exceeding $3 billion.¹ Although data are scarce, by some estimates the incidence of chronic pancreatitis ranges from 4 to 14 per 100,000 person-years, and the prevalence ranges from 26.4 to 52 per 100,000.^2–4 Moreover, a meta-analysis⁵ found that acute pancreatitis progresses to chronic pancreatitis in 10% of patients who have a first episode of acute pancreatitis and in 36% who have recurrent episodes.

Historically, alcoholism was and still is the most common cause of chronic pancreatitis, contributing to 60% to 90% of cases in Western countries.^6,7 However, cases due to nonalcoholic causes have been increasing, and in more than one-fourth of patients, no identifiable cause is found.^6,8 Smoking is an independent risk factor.^6,8,9 Some cases can be linked to genetic abnormalities, particularly in children.¹⁰

The clinical manifestations of chronic pancreatitis include exocrine pancreatic insufficiency (leading to malnutrition and steatorrhea), endocrine insufficiency (causing diabetes mellitus), and intractable pain.¹¹ Pain is the predominant clinical symptom early in the disease and is often debilitating and difficult to manage. Uncontrolled pain has a devastating impact on quality of life and may become complicated by narcotic dependence.

The pain of chronic pancreatitis is often multifactorial, with mechanisms that include increased intraductal pressure from obstruction of the pancreatic duct, pancreatic ischemia, neuronal injury, and neuroimmune interactions between neuronal processes and chronic inflammation.¹²

Treatment: Medical and surgical

In chronic pancreatitis, the aim of treatment is to alleviate deficiencies of exocrine and endocrine function and mitigate the pain. Patients who smoke or drink alcohol should be strongly encouraged to quit.

Loss of exocrine function is mainly managed with oral pancreatic enzyme supplements, and diabetes control is often attained with insulin therapy.¹³ Besides helping digestion, pancreatic enzyme therapy in the form of nonenteric tablets may also reduce pain and pancreatitis attacks.¹⁴ The mechanism may be by degrading cholecystokinin-releasing factor in the duodenum, lowering cholecystokinin levels and thereby reducing pain.¹²

Nonnarcotic analgesics are often the first line of therapy for pain management, but many patients need narcotic analgesics. Along with narcotics, adjunctive agents such as tricyclic antidepressants, serotonin-norepinephrine reuptake inhibitors, selective serotonin reuptake inhibitors, and gabapentinoids have been used to treat chronic pancreatitis pain, but with limited success.¹⁵

In patients for whom medical pain management has failed, one can consider another option, such as nerve block, neurolysis, or endoscopic or surgical therapy. Neuromodulators are often prescribed by pain clinics.¹⁵ Percutaneous and endoscopic celiac ganglion blocks can be an option but rarely achieve substantial or permanent pain relief, and the induced transient responses (on average 2 to 4 months) often cannot be repeated.^14–17

Surgical options to relieve pain try to preserve pancreatic function and vary depending on the degree of severity and nature of pancreatic damage. In broad terms, the surgical procedures can be divided into two types:

• Drainage procedures (eg, pseudocyst drainage; minimally invasive endoscopic duct drainage via sphincterotomy or stent placement, or both; pancreaticojejunostomy)
• Resectional procedures (eg, distal pancreatectomy, isolated head resection, pancreaticoduodenectomy, Whipple procedure, total pancreatectomy).

In carefully selected patients, total pancreatectomy can be offered to remove the cause of the pain.¹⁸ This procedure is most often performed in patients who have small-duct disease or a genetic cause or for whom other surgical procedures have failed.¹¹

HISTORY OF THE PROCEDURE

Islet cell transplantation grew out of visionary work by Paul Lacy and David Scharp at the University of Washington at Seattle, whose research focused on isolating and transplanting islet cells in rodent models. The topic has been reviewed by Jahansouz et al.¹⁹ In the 1970s, experiments in pancreatectomized dogs showed that transplanting unpurified pancreatic islet tissue that was dispersed by collagenase digestion into the spleen or portal vein could prevent diabetes.^20,21 In 1974, the first human trials of transplanting islet cells were conducted, using isolated islets from cadaveric donors to treat diabetes.¹⁹

In the past, pancreatectomy was performed to treat painful chronic pancreatitis, but it was viewed as undesirable because removing the gland would inevitably cause insulin-dependent diabetes.²² That changed in 1977 at the University of Minnesota, with the first reported islet cell autotransplant after pancreatectomy. The patient remained pain-free and insulin-independent long-term.²³ This seminal case showed that endocrine function could be preserved by autotransplant of islets prepared from the excised pancreas.²⁴

In 1992, Pyzdrowski et al²⁵ reported that intrahepatic transplant of as few as 265,000 islets was enough to prevent the need for insulin therapy. Since this technique was first described, there have been many advances, and now more than 30 centers worldwide do it.

PRIMARY INDICATION: INTRACTABLE PAIN

Interest has been growing in using total pancreatectomy and islet autotransplant to treat recurrent acute pancreatitis, chronic pancreatitis, and hereditary pancreatitis. The rationale is that removing the offending tissue eliminates pancreatitis, pain, and cancer risk, while preserving and replacing the islet cells prevents the development of brittle diabetes with loss of insulin and glucagon.²⁶

Proposed criteria for total pancreatectomy and islet autotransplant

Bellin et al¹⁴ proposed five criteria for patient selection for this procedure based on imaging studies, pancreatic function tests, and histopathology to detect pancreatic fibrosis. Patients must fulfill all five of the following criteria:

Criterion 1. Diagnosis of chronic pancreatitis, based on chronic abdominal pain lasting more than 6 months with either at least one of the following:

• Pancreatic calcifications on computed tomography
• At least two of the following: four or more of nine criteria on endoscopic ultrasonography described by Catalano et al,²⁷ a compatible ductal or parenchymal abnormality on secretin magnetic resonance cholangiopancreatography; abnormal endoscopic pancreatic function test (peak HCO₂ ≤ 80 mmol/L)
• Histopathologically confirmed diagnosis of chronic pancreatitis
• Compatible clinical history and documented hereditary pancreatitis (PRSS1 gene mutation)

OR

• History of recurrent acute pancreatitis (more than one episode of characteristic pain associated with imaging diagnostic of acute pancreatitis or elevated serum amylase or lipase > 3 times the upper limit of normal).

Criterion 2. At least one of the following:

• Daily narcotic dependence
• Pain resulting in impaired quality of life, which may include inability to attend school, recurrent hospitalizations, or inability to participate in usual age-appropriate activities.

Criterion 3. Complete evaluation with no reversible cause of pancreatitis present or untreated.

Criterion 4. Failure to respond to maximal medical and endoscopic therapy.

Criterion 5. Adequate islet cell function (nondiabetic or C-peptide-positive). Patients with C-peptide-negative diabetes meeting criteria 1 to 4 are candidates for total pancreatectomy alone.

The primary goal is to treat intractable pain and improve quality of life in selected patients with chronic pancreatitis or recurrent acute pancreatitis when endoscopic and prior surgical therapies have failed, and whose impairment due to pain is substantial enough to accept the risk of postoperative insulin-dependent diabetes and lifelong commitment to pancreatic enzyme replacement therapy.^15,26 Patients with a known genetic cause of chronic pancreatitis should be offered special consideration for the procedure, as their disease is unlikely to remit.

CONTRAINDICATIONS

Total pancreatectomy and islet autotransplant should not be performed in patients with active alcoholism, illicit drug use, or untreated or poorly controlled psychiatric illnesses that could impair the patient’s ability to adhere to a complicated postoperative medical regimen.

A poor support network may be a relative contraindication in view of the cost and complexity of diabetic and pancreatic enzyme replacement therapy.^18,26

Islet cell autotransplant is contraindicated in patients with conditions such as C-peptide-negative or type 1 diabetes or a history of portal vein thrombosis, portal hypertension, significant liver disease, high-risk cardiopulmonary disease, or pancreatic cancer (Table 1).²⁶

WHEN TO CONSIDER REFERRAL FOR THIS PROCEDURE

The choice of total pancreatectomy and islet autotransplant vs conventional surgery must be individualized on the basis of each patient’s anatomy, comorbidities, symptom burden, presence or degree of diabetes, and rate of disease progression. The most important factors to consider in determining the need for and timing of this procedure are the patient’s pain, narcotic requirements, and impaired ability to function.²⁶

Sooner rather than later?

An argument can be made for performing this procedure sooner in the course of the disease rather than later when all else has failed. First, prolonged pain can result in central sensitization, in which the threshold for perceiving pain is lowered by damage to the nociceptive neurons from repeated stimulation and inflammation.²⁸

Further, prolonged opioid therapy can lead to opioid-induced hyperalgesia, which may also render patients more sensitive to pain and aggravate their preexisting pain.^26,28

In addition, although operative drainage procedures and partial resections are often considered the gold standard for chronic pancreatitis management, patients who undergo partial pancreatectomy or lateral pancreaticojejunostomy (Puestow procedure) have fewer islet cells left to harvest (about 50% fewer) if they subsequently undergo total pancreatectomy and islet cell autotransplant.^22,26

Therefore, performing this procedure earlier may help the patient avoid chronic pain syndromes and complications of chronic opioid use, including hyperalgesia, and give the best chance of harvesting enough islet cells to prevent or minimize diabetes afterward.¹¹

REMOVING THE PANCREAS, RETURNING THE ISLET CELLS

During this procedure, the blood supply to the pancreas must be preserved until just before its removal to minimize warm ischemia of the islet cells.^18,29 Although there are several surgical variations, a pylorus-preserving total pancreatectomy with duodenectomy is typically performed, usually with splenectomy to preserve perfusion to the body and tail.³⁰

The resected pancreas is taken to the islet isolation laboratory. There, the pancreatic duct is cannulated to fill the organ with a cold collagenase solution, followed by gentle mechanical dispersion using the semiautomated Ricordi method,³¹ which separates the islet cells from the exocrine tissue.³²

The number of islet cells is quantified as islet equivalents; 1 islet equivalent is equal to the volume of an islet with a diameter of 150 µm. Islet equivalents per kilogram of body weight is the unit commonly used to report the graft amount transplanted.³³

After digestion, the islet cells can be purified or partially purified by a gradient separation method using a Cobe 2991 cell processor (Terumo Corporation, Tokyo, Japan),³⁴ or can be transplanted as an unpurified preparation. In islet cell autotransplant for chronic pancreatitis, purification is not always necessary due to the small tissue volume extracted from the often atrophic and fibrotic pancreas.³² The decision to purify depends on the postdigest tissue volume; usually, a tissue volume greater than 0.25 mL/kg body weight is an indication to at least partially purify.^18,35

The final preparation is returned to the operating room, and after heparin is given, the islets are infused into the portal system using a stump of the splenic vein, or alternatively through direct puncture of the portal vein or cannulation of the umbilical vein.^32,36 If the portal vein pressure reaches 25 cm H₂O, the infusion is stopped and the remaining islets can be placed in the peritoneal cavity or elsewhere.¹⁸ Transplant of the islets into the liver or peritoneum allows the islets to secrete insulin into the hepatic portal circulation, which is the route used by the native pancreas.²²

CONTROLLING GLUCOSE DURING AND AFTER THE PROCEDURE

Animal studies have shown that hyperglycemia impairs islet revascularization,³⁷ and glucose toxicity may cause dysfunction and structural lesions of the transplanted islets.^11,38

Therefore, during and after the procedure, most centers maintain euglycemia by an intravenous insulin infusion and subsequently move to subcutaneous insulin when the patient starts eating again. Some centers continue insulin at discharge and gradually taper it over months, even in patients who can possibly achieve euglycemia without it.

OUTCOMES

Many institutions have reported their clinical outcomes in terms of pain relief, islet function, glycemic control, and improvement of quality of life. The largest series have been from the University of Minnesota, Leicester General Hospital, the University of Cincinnati, and the Medical University of South Carolina.

Insulin independence is common but wanes with time

The ability to achieve insulin independence after islet autotransplant appears to be related to the number of islets transplanted, with the best results when more than 2,000 or 3,000 islet equivalents/kg are transplanted.^39,40

Sutherland et al¹⁸ reported that of 409 patients who underwent islet cell autotransplant at the University of Minnesota (the largest series reported to date), 30% were insulin-independent at 3 years, 33% had partial graft function (defined by positive C-peptide), and 82% achieved a mean hemoglobin A_1c of less than 7%. However, in the subset who received more than 5,000 islet equivalents/kg, nearly three-fourths of patients were insulin-independent at 3 years.

The Leicester General Hospital group presented long-term data on 46 patients who underwent total pancreatectomy and islet cell autotransplant. Twelve of the 46 had shown periods of insulin independence for a median of 16.5 months, and 5 remained insulin-free at the time of the publication.⁴¹ Over the 10 years of follow-up, insulin requirements and hemoglobin A_1c increased notably. However, all of the patients tested C-peptide-positive, suggesting long-lasting graft function.

Most recently, the University of Cincinnati group reported long-term data on 116 patients. The insulin independence rate was 38% at 1 year, decreasing to 27% at 5 years. The number of patients with partial graft function was 38% at 1 year and 35% at 5 years.⁴²

Thus, all three institutions confirmed that the autotransplanted islets continue to secrete insulin long-term, but that function decreases over time.

Pancreatectomy reduces pain

Multiple studies have shown that total pancreatectomy reduces pain in patients with chronic pancreatitis. Ahmad et al⁴³ reported a marked reduction in narcotic use (mean morphine equivalents 206 mg/day before surgery, compared with 90 mg after), and a 58% reduction in pain as demonstrated by narcotic independence.

In the University of Minnesota series, 85% of the 409 patients had less pain at 2 years, and 59% were able to stop taking narcotics.¹⁸

The University of Cincinnati group reported a narcotic independence rate of 55% at 1 year, which continued to improve to 73% at 5 years.⁴²

Although the source of pain is removed, pain persists or recurs in 10% to 20% of patients after total pancreatectomy and islet cell autotransplant, showing that the pathogenesis of pain is complex, and some uncertainty exists about it.²⁶

Quality of life

Reports evaluating health-related quality of life after total pancreatectomy and islet autotransplant are limited.

The University of Cincinnati group reported the long-term outcomes of quality of life as measured by the Short Form 36 Health Survey.⁴² Ninety-two percent of patients reported overall improvement in their health at 1 year, and 85% continued to report improved health more than 5 years after the surgery.

In a series of 20 patients, 79% to 90% reported improvements in the seven various domains of the Pain Disability Index. In addition, 60% showed improvement in depression and 70% showed improvement in anxiety. The greatest improvements were in those who had not undergone prior pancreatic surgery, who were younger, and in those with higher levels of preoperative pain.³⁰

Similarly, in a series of 74 patients, the Medical University of South Carolina group reported significant improvement in physical and mental health components of the Short Form 12 Health Survey and an associated decrease in daily narcotic requirements. Moreover, the need to start or increase the dose of insulin after the surgery was not associated with a lower quality of life.⁴⁴

OFF-SITE ISLET CELL ISOLATION

Despite the positive outcomes in terms of pain relief and insulin independence in many patients after total pancreatectomy and islet cell autotransplant, few medical centers have an on-site islet-processing facility. Since the mid-1990s, a few centers have been able to circumvent this limitation by working with off-site islet cell isolation laboratories.^45,46

Whether and when to consider this procedure must be individualized

The University of California, Los Angeles, first reported on a series of nine patients who received autologous islet cells after near-total or total pancreatectomy using a remote islet cell isolation facility, with results comparable to those of other large institutions.⁴⁵

Similarly, the procedure has been performed at Cleveland Clinic since 2007 with the collaboration of an off-site islet cell isolation laboratory at the University of Pittsburgh. A cohort study from this collaboration published in 2015 showed that in 36 patients (mean follow-up 28 months, range 3–26 months), 33% were insulin-independent, with a C-peptide-positive rate of 70%. This is the largest cohort to date from a center utilizing an off-site islet isolation facility.⁴⁷

In view of the positive outcomes at these centers, lack of a local islet-processing facility should no longer be a barrier to total pancreatectomy and islet cell autotransplant.

PATIENT CARE AFTER THE PROCEDURE

A multidisciplinary team is an essential component of the postoperative management of patients who undergo total pancreatectomy and islet cell autotransplant.

For patients who had been receiving narcotics for a long time before surgery or who were receiving frequent doses, an experienced pain management physician should be involved in the patient’s postoperative care.

Because islet function can wane over time, testing for diabetes should be done at least annually for the rest of the patient’s life and should include fasting plasma glucose, hemoglobin A_1c, and C-peptide, along with self-monitored blood glucose.²⁶

All patients who have surgically induced exocrine insufficiency are at risk of malabsorption and fat-soluble vitamin deficiencies.⁴⁸ Hence, lifelong pancreatic enzyme replacement therapy is mandatory. Nutritional monitoring should include assessment of steatorrhea, body composition, and fat-soluble vitamin levels (vitamins A, D, and E) at least every year.²⁶ Patients with chronic pancreatitis are at increased risk for low bone density from malabsorption of vitamin D and calcium; therefore, it is recommended that a dual-energy x-ray absorptiometry bone density scan be obtained.^26,49

Patients who undergo splenectomy as part of their procedure will require appropriate precautions and ongoing vaccinations as recommended by the US Centers for Disease Control and Prevention.^26,50,51

WHAT TO EXPECT FOR THE FUTURE

The National Institute of Diabetes and Digestive and Kidney Diseases has reviewed the potential future research directions for total pancreatectomy and islet autotransplant.¹⁵

The more islet cells transplanted, the better the chance of insulin independence

Patient selection remains challenging despite the availability of criteria¹⁵ and guidelines.²⁶ More research is needed to better assess preoperative beta-cell function and to predict postoperative outcomes. Mixed meal-tolerance testing is adopted by most clinical centers to predict posttransplant beta-cell function. The use of arginine instead of glucagon in a stimulation test for insulin and C-peptide response has been validated and may allow more accurate assessment.^52,53

Another targeted area of research is the advancement of safety and metabolic outcomes. Techniques to minimize warm ischemic time and complications are being evaluated. Islet isolation methods that yield more islets, reduce beta-cell apoptosis, and can isolate islets from glands with malignancy should be further investigated.⁵⁴ Further, enhanced islet infusion methods that achieve lower portal venous pressures and minimize portal vein thrombosis are needed.

Unfortunately, the function of transplanted islet grafts declines over time. This phenomenon is at least partially attributed to the immediate blood-mediated inflammatory response,^55,56 along with islet hypoxia,⁵⁷ leading to islet apoptosis. Research on different strategies is expanding our knowledge in islet engraftment and posttransplant beta-cell apoptosis, with the expectation that the transplanted islet lifespan will increase. Alternative transplant sites with low inflammatory reaction, such as the omental pouch,⁵⁸ muscle,⁵⁹ and bone marrow,⁶⁰ have shown encouraging data. Other approaches, such as adjuvant anti-inflammatory agents and heparinization, have been proposed.¹⁵

Research into complications is also of clinical importance. There is growing attention to hypoglycemia unrelated to exogenous insulin use in posttransplant patients. One hypothesis is that glucagon secretion, a counterregulatory response to hypoglycemia, is defective if the islet cells are transplanted into the liver, and that implanting them into another site may avoid this effect.⁶¹

For some patients with chronic pancreatitis, the best option is to remove the entire pancreas. This does not necessarily doom the patient to diabetes mellitus, because we can harvest the islet cells and reinsert them so that, lodged in the liver, they can continue making insulin. However, this approach is underemphasized in the general medical literature and is likely underutilized in the United States.

Here, we discuss chronic pancreatitis, the indications for and contraindications to this procedure, its outcomes, and the management of patients who undergo it.

CHRONIC PANCREATITIS IS PROGRESSIVE AND PAINFUL

Chronic pancreatitis is a progressive condition characterized by chronic inflammation, irreversible fibrosis, and scarring, resulting in loss of both exocrine and endocrine tissue.

According to a National Institutes of Health database, pancreatitis is the seventh most common digestive disease diagnosis on hospitalization, with annual healthcare costs exceeding $3 billion.¹ Although data are scarce, by some estimates the incidence of chronic pancreatitis ranges from 4 to 14 per 100,000 person-years, and the prevalence ranges from 26.4 to 52 per 100,000.^2–4 Moreover, a meta-analysis⁵ found that acute pancreatitis progresses to chronic pancreatitis in 10% of patients who have a first episode of acute pancreatitis and in 36% who have recurrent episodes.

Historically, alcoholism was and still is the most common cause of chronic pancreatitis, contributing to 60% to 90% of cases in Western countries.^6,7 However, cases due to nonalcoholic causes have been increasing, and in more than one-fourth of patients, no identifiable cause is found.^6,8 Smoking is an independent risk factor.^6,8,9 Some cases can be linked to genetic abnormalities, particularly in children.¹⁰

The clinical manifestations of chronic pancreatitis include exocrine pancreatic insufficiency (leading to malnutrition and steatorrhea), endocrine insufficiency (causing diabetes mellitus), and intractable pain.¹¹ Pain is the predominant clinical symptom early in the disease and is often debilitating and difficult to manage. Uncontrolled pain has a devastating impact on quality of life and may become complicated by narcotic dependence.

The pain of chronic pancreatitis is often multifactorial, with mechanisms that include increased intraductal pressure from obstruction of the pancreatic duct, pancreatic ischemia, neuronal injury, and neuroimmune interactions between neuronal processes and chronic inflammation.¹²

Treatment: Medical and surgical

In chronic pancreatitis, the aim of treatment is to alleviate deficiencies of exocrine and endocrine function and mitigate the pain. Patients who smoke or drink alcohol should be strongly encouraged to quit.

Loss of exocrine function is mainly managed with oral pancreatic enzyme supplements, and diabetes control is often attained with insulin therapy.¹³ Besides helping digestion, pancreatic enzyme therapy in the form of nonenteric tablets may also reduce pain and pancreatitis attacks.¹⁴ The mechanism may be by degrading cholecystokinin-releasing factor in the duodenum, lowering cholecystokinin levels and thereby reducing pain.¹²

Nonnarcotic analgesics are often the first line of therapy for pain management, but many patients need narcotic analgesics. Along with narcotics, adjunctive agents such as tricyclic antidepressants, serotonin-norepinephrine reuptake inhibitors, selective serotonin reuptake inhibitors, and gabapentinoids have been used to treat chronic pancreatitis pain, but with limited success.¹⁵

In patients for whom medical pain management has failed, one can consider another option, such as nerve block, neurolysis, or endoscopic or surgical therapy. Neuromodulators are often prescribed by pain clinics.¹⁵ Percutaneous and endoscopic celiac ganglion blocks can be an option but rarely achieve substantial or permanent pain relief, and the induced transient responses (on average 2 to 4 months) often cannot be repeated.^14–17

Surgical options to relieve pain try to preserve pancreatic function and vary depending on the degree of severity and nature of pancreatic damage. In broad terms, the surgical procedures can be divided into two types:

• Drainage procedures (eg, pseudocyst drainage; minimally invasive endoscopic duct drainage via sphincterotomy or stent placement, or both; pancreaticojejunostomy)
• Resectional procedures (eg, distal pancreatectomy, isolated head resection, pancreaticoduodenectomy, Whipple procedure, total pancreatectomy).

In carefully selected patients, total pancreatectomy can be offered to remove the cause of the pain.¹⁸ This procedure is most often performed in patients who have small-duct disease or a genetic cause or for whom other surgical procedures have failed.¹¹

HISTORY OF THE PROCEDURE

Islet cell transplantation grew out of visionary work by Paul Lacy and David Scharp at the University of Washington at Seattle, whose research focused on isolating and transplanting islet cells in rodent models. The topic has been reviewed by Jahansouz et al.¹⁹ In the 1970s, experiments in pancreatectomized dogs showed that transplanting unpurified pancreatic islet tissue that was dispersed by collagenase digestion into the spleen or portal vein could prevent diabetes.^20,21 In 1974, the first human trials of transplanting islet cells were conducted, using isolated islets from cadaveric donors to treat diabetes.¹⁹

In the past, pancreatectomy was performed to treat painful chronic pancreatitis, but it was viewed as undesirable because removing the gland would inevitably cause insulin-dependent diabetes.²² That changed in 1977 at the University of Minnesota, with the first reported islet cell autotransplant after pancreatectomy. The patient remained pain-free and insulin-independent long-term.²³ This seminal case showed that endocrine function could be preserved by autotransplant of islets prepared from the excised pancreas.²⁴

In 1992, Pyzdrowski et al²⁵ reported that intrahepatic transplant of as few as 265,000 islets was enough to prevent the need for insulin therapy. Since this technique was first described, there have been many advances, and now more than 30 centers worldwide do it.

PRIMARY INDICATION: INTRACTABLE PAIN

Interest has been growing in using total pancreatectomy and islet autotransplant to treat recurrent acute pancreatitis, chronic pancreatitis, and hereditary pancreatitis. The rationale is that removing the offending tissue eliminates pancreatitis, pain, and cancer risk, while preserving and replacing the islet cells prevents the development of brittle diabetes with loss of insulin and glucagon.²⁶

Proposed criteria for total pancreatectomy and islet autotransplant

Bellin et al¹⁴ proposed five criteria for patient selection for this procedure based on imaging studies, pancreatic function tests, and histopathology to detect pancreatic fibrosis. Patients must fulfill all five of the following criteria:

Criterion 1. Diagnosis of chronic pancreatitis, based on chronic abdominal pain lasting more than 6 months with either at least one of the following:

• Pancreatic calcifications on computed tomography
• At least two of the following: four or more of nine criteria on endoscopic ultrasonography described by Catalano et al,²⁷ a compatible ductal or parenchymal abnormality on secretin magnetic resonance cholangiopancreatography; abnormal endoscopic pancreatic function test (peak HCO₂ ≤ 80 mmol/L)
• Histopathologically confirmed diagnosis of chronic pancreatitis
• Compatible clinical history and documented hereditary pancreatitis (PRSS1 gene mutation)

OR

• History of recurrent acute pancreatitis (more than one episode of characteristic pain associated with imaging diagnostic of acute pancreatitis or elevated serum amylase or lipase > 3 times the upper limit of normal).

Criterion 2. At least one of the following:

• Daily narcotic dependence
• Pain resulting in impaired quality of life, which may include inability to attend school, recurrent hospitalizations, or inability to participate in usual age-appropriate activities.

Criterion 3. Complete evaluation with no reversible cause of pancreatitis present or untreated.

Criterion 4. Failure to respond to maximal medical and endoscopic therapy.

Criterion 5. Adequate islet cell function (nondiabetic or C-peptide-positive). Patients with C-peptide-negative diabetes meeting criteria 1 to 4 are candidates for total pancreatectomy alone.

The primary goal is to treat intractable pain and improve quality of life in selected patients with chronic pancreatitis or recurrent acute pancreatitis when endoscopic and prior surgical therapies have failed, and whose impairment due to pain is substantial enough to accept the risk of postoperative insulin-dependent diabetes and lifelong commitment to pancreatic enzyme replacement therapy.^15,26 Patients with a known genetic cause of chronic pancreatitis should be offered special consideration for the procedure, as their disease is unlikely to remit.

CONTRAINDICATIONS

Total pancreatectomy and islet autotransplant should not be performed in patients with active alcoholism, illicit drug use, or untreated or poorly controlled psychiatric illnesses that could impair the patient’s ability to adhere to a complicated postoperative medical regimen.

A poor support network may be a relative contraindication in view of the cost and complexity of diabetic and pancreatic enzyme replacement therapy.^18,26

Islet cell autotransplant is contraindicated in patients with conditions such as C-peptide-negative or type 1 diabetes or a history of portal vein thrombosis, portal hypertension, significant liver disease, high-risk cardiopulmonary disease, or pancreatic cancer (Table 1).²⁶

WHEN TO CONSIDER REFERRAL FOR THIS PROCEDURE

The choice of total pancreatectomy and islet autotransplant vs conventional surgery must be individualized on the basis of each patient’s anatomy, comorbidities, symptom burden, presence or degree of diabetes, and rate of disease progression. The most important factors to consider in determining the need for and timing of this procedure are the patient’s pain, narcotic requirements, and impaired ability to function.²⁶

Sooner rather than later?

An argument can be made for performing this procedure sooner in the course of the disease rather than later when all else has failed. First, prolonged pain can result in central sensitization, in which the threshold for perceiving pain is lowered by damage to the nociceptive neurons from repeated stimulation and inflammation.²⁸

Further, prolonged opioid therapy can lead to opioid-induced hyperalgesia, which may also render patients more sensitive to pain and aggravate their preexisting pain.^26,28

In addition, although operative drainage procedures and partial resections are often considered the gold standard for chronic pancreatitis management, patients who undergo partial pancreatectomy or lateral pancreaticojejunostomy (Puestow procedure) have fewer islet cells left to harvest (about 50% fewer) if they subsequently undergo total pancreatectomy and islet cell autotransplant.^22,26

Therefore, performing this procedure earlier may help the patient avoid chronic pain syndromes and complications of chronic opioid use, including hyperalgesia, and give the best chance of harvesting enough islet cells to prevent or minimize diabetes afterward.¹¹

REMOVING THE PANCREAS, RETURNING THE ISLET CELLS

During this procedure, the blood supply to the pancreas must be preserved until just before its removal to minimize warm ischemia of the islet cells.^18,29 Although there are several surgical variations, a pylorus-preserving total pancreatectomy with duodenectomy is typically performed, usually with splenectomy to preserve perfusion to the body and tail.³⁰

The resected pancreas is taken to the islet isolation laboratory. There, the pancreatic duct is cannulated to fill the organ with a cold collagenase solution, followed by gentle mechanical dispersion using the semiautomated Ricordi method,³¹ which separates the islet cells from the exocrine tissue.³²

The number of islet cells is quantified as islet equivalents; 1 islet equivalent is equal to the volume of an islet with a diameter of 150 µm. Islet equivalents per kilogram of body weight is the unit commonly used to report the graft amount transplanted.³³

After digestion, the islet cells can be purified or partially purified by a gradient separation method using a Cobe 2991 cell processor (Terumo Corporation, Tokyo, Japan),³⁴ or can be transplanted as an unpurified preparation. In islet cell autotransplant for chronic pancreatitis, purification is not always necessary due to the small tissue volume extracted from the often atrophic and fibrotic pancreas.³² The decision to purify depends on the postdigest tissue volume; usually, a tissue volume greater than 0.25 mL/kg body weight is an indication to at least partially purify.^18,35

The final preparation is returned to the operating room, and after heparin is given, the islets are infused into the portal system using a stump of the splenic vein, or alternatively through direct puncture of the portal vein or cannulation of the umbilical vein.^32,36 If the portal vein pressure reaches 25 cm H₂O, the infusion is stopped and the remaining islets can be placed in the peritoneal cavity or elsewhere.¹⁸ Transplant of the islets into the liver or peritoneum allows the islets to secrete insulin into the hepatic portal circulation, which is the route used by the native pancreas.²²

CONTROLLING GLUCOSE DURING AND AFTER THE PROCEDURE

Animal studies have shown that hyperglycemia impairs islet revascularization,³⁷ and glucose toxicity may cause dysfunction and structural lesions of the transplanted islets.^11,38

Therefore, during and after the procedure, most centers maintain euglycemia by an intravenous insulin infusion and subsequently move to subcutaneous insulin when the patient starts eating again. Some centers continue insulin at discharge and gradually taper it over months, even in patients who can possibly achieve euglycemia without it.

OUTCOMES

Many institutions have reported their clinical outcomes in terms of pain relief, islet function, glycemic control, and improvement of quality of life. The largest series have been from the University of Minnesota, Leicester General Hospital, the University of Cincinnati, and the Medical University of South Carolina.

Insulin independence is common but wanes with time

The ability to achieve insulin independence after islet autotransplant appears to be related to the number of islets transplanted, with the best results when more than 2,000 or 3,000 islet equivalents/kg are transplanted.^39,40

Sutherland et al¹⁸ reported that of 409 patients who underwent islet cell autotransplant at the University of Minnesota (the largest series reported to date), 30% were insulin-independent at 3 years, 33% had partial graft function (defined by positive C-peptide), and 82% achieved a mean hemoglobin A_1c of less than 7%. However, in the subset who received more than 5,000 islet equivalents/kg, nearly three-fourths of patients were insulin-independent at 3 years.

The Leicester General Hospital group presented long-term data on 46 patients who underwent total pancreatectomy and islet cell autotransplant. Twelve of the 46 had shown periods of insulin independence for a median of 16.5 months, and 5 remained insulin-free at the time of the publication.⁴¹ Over the 10 years of follow-up, insulin requirements and hemoglobin A_1c increased notably. However, all of the patients tested C-peptide-positive, suggesting long-lasting graft function.

Most recently, the University of Cincinnati group reported long-term data on 116 patients. The insulin independence rate was 38% at 1 year, decreasing to 27% at 5 years. The number of patients with partial graft function was 38% at 1 year and 35% at 5 years.⁴²

Thus, all three institutions confirmed that the autotransplanted islets continue to secrete insulin long-term, but that function decreases over time.

Pancreatectomy reduces pain

Multiple studies have shown that total pancreatectomy reduces pain in patients with chronic pancreatitis. Ahmad et al⁴³ reported a marked reduction in narcotic use (mean morphine equivalents 206 mg/day before surgery, compared with 90 mg after), and a 58% reduction in pain as demonstrated by narcotic independence.

In the University of Minnesota series, 85% of the 409 patients had less pain at 2 years, and 59% were able to stop taking narcotics.¹⁸

The University of Cincinnati group reported a narcotic independence rate of 55% at 1 year, which continued to improve to 73% at 5 years.⁴²

Although the source of pain is removed, pain persists or recurs in 10% to 20% of patients after total pancreatectomy and islet cell autotransplant, showing that the pathogenesis of pain is complex, and some uncertainty exists about it.²⁶

Quality of life

Reports evaluating health-related quality of life after total pancreatectomy and islet autotransplant are limited.

The University of Cincinnati group reported the long-term outcomes of quality of life as measured by the Short Form 36 Health Survey.⁴² Ninety-two percent of patients reported overall improvement in their health at 1 year, and 85% continued to report improved health more than 5 years after the surgery.

In a series of 20 patients, 79% to 90% reported improvements in the seven various domains of the Pain Disability Index. In addition, 60% showed improvement in depression and 70% showed improvement in anxiety. The greatest improvements were in those who had not undergone prior pancreatic surgery, who were younger, and in those with higher levels of preoperative pain.³⁰

Similarly, in a series of 74 patients, the Medical University of South Carolina group reported significant improvement in physical and mental health components of the Short Form 12 Health Survey and an associated decrease in daily narcotic requirements. Moreover, the need to start or increase the dose of insulin after the surgery was not associated with a lower quality of life.⁴⁴

OFF-SITE ISLET CELL ISOLATION

Despite the positive outcomes in terms of pain relief and insulin independence in many patients after total pancreatectomy and islet cell autotransplant, few medical centers have an on-site islet-processing facility. Since the mid-1990s, a few centers have been able to circumvent this limitation by working with off-site islet cell isolation laboratories.^45,46

Whether and when to consider this procedure must be individualized

The University of California, Los Angeles, first reported on a series of nine patients who received autologous islet cells after near-total or total pancreatectomy using a remote islet cell isolation facility, with results comparable to those of other large institutions.⁴⁵

Similarly, the procedure has been performed at Cleveland Clinic since 2007 with the collaboration of an off-site islet cell isolation laboratory at the University of Pittsburgh. A cohort study from this collaboration published in 2015 showed that in 36 patients (mean follow-up 28 months, range 3–26 months), 33% were insulin-independent, with a C-peptide-positive rate of 70%. This is the largest cohort to date from a center utilizing an off-site islet isolation facility.⁴⁷

In view of the positive outcomes at these centers, lack of a local islet-processing facility should no longer be a barrier to total pancreatectomy and islet cell autotransplant.

PATIENT CARE AFTER THE PROCEDURE

A multidisciplinary team is an essential component of the postoperative management of patients who undergo total pancreatectomy and islet cell autotransplant.

For patients who had been receiving narcotics for a long time before surgery or who were receiving frequent doses, an experienced pain management physician should be involved in the patient’s postoperative care.

Because islet function can wane over time, testing for diabetes should be done at least annually for the rest of the patient’s life and should include fasting plasma glucose, hemoglobin A_1c, and C-peptide, along with self-monitored blood glucose.²⁶

All patients who have surgically induced exocrine insufficiency are at risk of malabsorption and fat-soluble vitamin deficiencies.⁴⁸ Hence, lifelong pancreatic enzyme replacement therapy is mandatory. Nutritional monitoring should include assessment of steatorrhea, body composition, and fat-soluble vitamin levels (vitamins A, D, and E) at least every year.²⁶ Patients with chronic pancreatitis are at increased risk for low bone density from malabsorption of vitamin D and calcium; therefore, it is recommended that a dual-energy x-ray absorptiometry bone density scan be obtained.^26,49

Patients who undergo splenectomy as part of their procedure will require appropriate precautions and ongoing vaccinations as recommended by the US Centers for Disease Control and Prevention.^26,50,51

WHAT TO EXPECT FOR THE FUTURE

The National Institute of Diabetes and Digestive and Kidney Diseases has reviewed the potential future research directions for total pancreatectomy and islet autotransplant.¹⁵

The more islet cells transplanted, the better the chance of insulin independence

Patient selection remains challenging despite the availability of criteria¹⁵ and guidelines.²⁶ More research is needed to better assess preoperative beta-cell function and to predict postoperative outcomes. Mixed meal-tolerance testing is adopted by most clinical centers to predict posttransplant beta-cell function. The use of arginine instead of glucagon in a stimulation test for insulin and C-peptide response has been validated and may allow more accurate assessment.^52,53

Another targeted area of research is the advancement of safety and metabolic outcomes. Techniques to minimize warm ischemic time and complications are being evaluated. Islet isolation methods that yield more islets, reduce beta-cell apoptosis, and can isolate islets from glands with malignancy should be further investigated.⁵⁴ Further, enhanced islet infusion methods that achieve lower portal venous pressures and minimize portal vein thrombosis are needed.

Unfortunately, the function of transplanted islet grafts declines over time. This phenomenon is at least partially attributed to the immediate blood-mediated inflammatory response,^55,56 along with islet hypoxia,⁵⁷ leading to islet apoptosis. Research on different strategies is expanding our knowledge in islet engraftment and posttransplant beta-cell apoptosis, with the expectation that the transplanted islet lifespan will increase. Alternative transplant sites with low inflammatory reaction, such as the omental pouch,⁵⁸ muscle,⁵⁹ and bone marrow,⁶⁰ have shown encouraging data. Other approaches, such as adjuvant anti-inflammatory agents and heparinization, have been proposed.¹⁵

Research into complications is also of clinical importance. There is growing attention to hypoglycemia unrelated to exogenous insulin use in posttransplant patients. One hypothesis is that glucagon secretion, a counterregulatory response to hypoglycemia, is defective if the islet cells are transplanted into the liver, and that implanting them into another site may avoid this effect.⁶¹

Endoscopic ultrasonography (EUS) is a minimally invasive test that provides high-resolution imaging of the pancreas.^1,2 As such, it is proving useful.

Accurate diagnosis and timely intervention are essential in managing acute and chronic pancreatitis, which are often encountered in the clinic and the hospital. However, the cause of acute pancreatitis is not always easy to determine. Furthermore, recurrent bouts can progress to chronic pancreatitis if the cause is not identified and eliminated. EUS has been studied extensively in the evaluation of both acute and chronic pancreatitis, as it can identify obstructive and biliary causes of acute pancreatitis and early structural features of chronic pancreatitis.

This article will review the indications and evidence for EUS in the evaluation of acute and chronic pancreatitis.

SPECIALIZED TRAINING REQUIRED

EUS involves passage of a specialized endoscope through the esophagus and stomach and into the duodenum. The scope has a very small ultrasound probe at the tip, allowing detailed imaging of the upper gastrointestinal tract and surrounding organs.

There are two types of EUS endoscope: radial and linear. A radial scope provides a 360° range of view perpendicular to the long axis of the scope. A linear scope provides a 150° view parallel to the long axis of the scope. Many endosonographers favor linear EUS for imaging the pancreas because it permits fine-needle aspiration biopsy of masses, cysts, and lymph nodes.

Specialized training beyond the gastroenterology fellowship is usually required to become proficient in performing EUS, in recognizing the anatomy it reveals, and in performing fine-needle aspiration biopsy.

ENDOSCOPIC ULTRASONOGRAPHY IN ACUTE PANCREATITIS

Finding the cause of acute pancreatitis can be challenging in patients who do not have typical risk factors, eg, those who do not drink substantial amounts of alcohol and in whom transabdominal ultrasonography fails to reveal gallstones.

Several studies have evaluated the role of EUS in recurrent “idiopathic” pancreatitis.^3–5 Causes of acute pancreatitis detectable with EUS included gallbladder and bile duct microlithiasis (stones smaller than 3 mm), cysts, intraductal papillary mucinous neoplasms, ampullary neoplasms, pancreas divisum, and pancreatic masses.

Stones, sludge. Transabdominal ultrasonography is often performed in the workup of acute pancreatitis to rule out gallbladder stones and biliary dilation. Unfortunately, it does a poor job of imaging the distal common bile duct, where culprit stones may reside.

EUS provides a high-quality view of the bile duct from the ampulla of Vater to the region of the hepatic hilum and is safer than endoscopic retrograde cholangiopancreatography (ERCP). The available evidence supports the use of EUS as a diagnostic test for bile duct stones.^3–7 In fact, using ERCP as the reference standard, EUS has been found to be more sensitive than transabdominal ultrasonography for bile duct stones.⁴

The yield of EUS for finding biliary sludge and stones may be high in patients with unexplained pancreatitis. EUS detected sludge, microlithiasis, or both in 33 of 35 patients with idiopathic acute pancreatitis who underwent transabdominal ultrasonography with negative results.⁸ Furthermore, most were symptom-free at an average of 10 months after cholecystectomy, suggesting that microlithiasis was the cause of the “idiopathic” pancreatitis.

EUS can also decrease the number of unnecessary ERCP procedures in patients with suspected biliary pancreatitis. In these patients, EUS can be performed as an initial diagnostic test to exclude retained biliary stones. If a stone is present, the endoscopist can proceed to ERCP for sphincterotomy and stone removal during the same endoscopic session. If EUS is negative, the endoscopy can be concluded without cannulating the bile duct and putting the patient at risk of acute pancreatitis. In one report, this approach eliminated the need for ERCP in five of six patients with suspected biliary pancreatitis.⁶

Tumors and other causes of bile duct obstruction can also cause recurrent acute pancreatitis and may be difficult to detect with cross-sectional imaging. EUS, on the other hand, can detect small pancreatic masses (< 2 cm), which may be missed by conventional computed tomography. Also, a linear EUS scope, with its forward oblique view, can image the duodenum and ampulla, where obstructing inflammation, tumors, and polyps may be found. One should strongly suspect occult malignancy in elderly patients with unexplained acute pancreatitis. In those patients, repeat imaging with high-resolution dual-phase computed tomography or with EUS should be considered after a few weeks once the acute inflammation resolves.

Pancreas divisum is a relatively common congenital abnormality in which the dorsal and ventral pancreatic ducts do not properly fuse during embryonic development. To rule out pancreas divisum, the endosonographer must carefully trace the pancreatic duct from the dorsal pancreas into the ventral pancreas, where it connects with the bile duct at the duodenal wall.

In summary, EUS appears to be safe and accurate for diagnosing bile duct stones and other structural causes of idiopathic acute pancreatitis.

ENDOSCOPIC ULTRASONOGRAPHY IN CHRONIC PANCREATITIS

Chronic pancreatitis, a relatively common and sometimes debilitating cause of chronic upper abdominal pain, may be difficult to diagnose using noninvasive imaging tests. Minimal-change chronic pancreatitis is defined as a syndrome of pancreatic abdominal pain with no or slight structural changes detected on imaging but with histologic inflammation and fibrosis diagnostic of chronic pancreatitis.⁹

A clinical rationale for trying to detect chronic pancreatitis early in its course is that interventions can be started earlier. These include abstinence from alcohol, giving exogenous pancreatic enzymes, and advanced interventions such as celiac plexus blocks for pain control. Some patients may even benefit from resection of the pancreas if pain is severe and resistant to conservative measures.

EUS can detect both parenchymal and ductal changes that correlate with histologic fibrosis.¹⁰ Parenchymal changes include hyperechoic foci, hyperechoic strands, lobularity, cysts, and shadowing calcifications. Ductal changes include dilation of the main pancreatic duct, irregularity, hyperechoic duct margins, and visible side branches.

Several studies have evaluated the ability of EUS to diagnose early chronic pancreatitis.^9,11–15 Reference standards used to determine the accuracy of EUS have included histology,^10,16–18 pancreatic function testing,^19–22 and ERCP.^11,15,23,24

The best diagnostic test may be pancreatic histology. However, biopsy of the pancreas is impractical and exposes patients to high risk. In addition, the patchy and focal distribution of histologic changes may decrease its reliability. Fortunately, the histologic findings of fibrosis have been shown to correlate with EUS criteria in patients undergoing EUS before surgical resection in three recent studies.^16–18 A threshold of four or more criteria out of a possible nine was found to provide the optimal sensitivity and specificity for histologic pancreatic fibrosis.^16,17 The criteria used were four parenchymal features (hyperechoic foci, strands, hypoechoic lobules, cysts) and five ductal features (irregularity of the main pancreatic duct, dilation, hyperechoic duct walls, visible side branches, and calcifications or stones).

EUS is sensitive for chronic pancreatitis, but ‘true’ accuracy is impossible to know

Unfortunately, greater sensitivity may come at the expense of worse specificity. Certain demographic variables may alter the EUS appearance of the pancreas. A multivariate analysis²⁵ found several variables that predicted abnormalities on EUS even in the absence of clinically evident pancreatitis; the strongest were heavy ethanol use (odds ratio [OR] 5.1, 95% confidence interval [CI] 3.1–8.5), male sex (OR 1.8, 95% CI 1.3–2.55), clinical suspicion of pancreatic disease (OR 1.7, 95% CI 1.2–2.3), and heavy smoking (OR 1.7, 95% CI 1.2–2.4). More prospective studies are needed to further differentiate true disease from false-positive findings of chronic pancreatitis.

Also, traditional EUS scoring symptoms have counted features in an unweighted fashion and assigned an arbitrary cut point (eg, four or more features) for diagnosis. This approach fails to account for the greater importance of some features (eg, calcifications) compared with others.

Interobserver variability is another important limitation of EUS in diagnosing chronic pancreatitis.^26,27 In one multicenter study of EUS interpretation, the overall kappa (agreement beyond chance) was only 0.45 for overall chronic pancreatitis diagnosis and worse for many individual criteria for chronic pancreatitis. The endosonographers disagreed most about hyperechoic strands and foci, main pancreatic duct irregularity, and visible side branches (kappa < 0.4).

The Rosemont classification

These limitations led a group of experts to meet in Chicago, IL, to develop a consensus-based and weighted EUS scoring system for the diagnosis of chronic pancreatitis, termed the Rosemont classification.

In this system, the previous parenchymal and ductal features are assigned stricter definitions and reclassified as major and minor criteria. Based on the presence of major and minor features, EUS results are stratified as “normal,” “indeterminate for chronic pancreatitis,” “suggestive of chronic pancreatitis,” or “most consistent with chronic pancreatitis.”^15,28

Further validation of this scoring system is needed before it can be used widely.

ENDOSCOPIC ULTRASONOGRAPHY PLUS PANCREATIC FUNCTION TESTING

The best way to diagnose minimal-change chronic pancreatitis may be a combination of sensitive structural and functional testing. Although clinically apparent steatorrhea typically occurs late in the course of chronic pancreatitis, mild exocrine insufficiency may occur early and is detectable with hormone-stimulated pancreatic function testing. Therefore, pancreatic function tests are considered sensitive for diagnosing chronic pancreatitis.^20,21,29

Endoscopic pancreatic function testing involves injecting secretin intravenously and then collecting duodenal aspirates through the endoscope. The duodenal fluid is analyzed for bicarbonate concentration as a measure of exocrine function.²⁹

We have studied combined EUS and endoscopic pancreatic function testing in the diagnosis of chronic pancreatitis.¹⁶ The combination gives a simultaneous structural and functional assessment of the pancreas and may optimize sensitivity for detecting minimal-change chronic pancreatitis. In a small study, we found the combination had 100% sensitivity for noncalcific chronic pancreatitis compared with a histologic reference standard.¹⁶

Endoscopic ultrasonography (EUS) is a minimally invasive test that provides high-resolution imaging of the pancreas.^1,2 As such, it is proving useful.

Accurate diagnosis and timely intervention are essential in managing acute and chronic pancreatitis, which are often encountered in the clinic and the hospital. However, the cause of acute pancreatitis is not always easy to determine. Furthermore, recurrent bouts can progress to chronic pancreatitis if the cause is not identified and eliminated. EUS has been studied extensively in the evaluation of both acute and chronic pancreatitis, as it can identify obstructive and biliary causes of acute pancreatitis and early structural features of chronic pancreatitis.

This article will review the indications and evidence for EUS in the evaluation of acute and chronic pancreatitis.

SPECIALIZED TRAINING REQUIRED

EUS involves passage of a specialized endoscope through the esophagus and stomach and into the duodenum. The scope has a very small ultrasound probe at the tip, allowing detailed imaging of the upper gastrointestinal tract and surrounding organs.

There are two types of EUS endoscope: radial and linear. A radial scope provides a 360° range of view perpendicular to the long axis of the scope. A linear scope provides a 150° view parallel to the long axis of the scope. Many endosonographers favor linear EUS for imaging the pancreas because it permits fine-needle aspiration biopsy of masses, cysts, and lymph nodes.

Specialized training beyond the gastroenterology fellowship is usually required to become proficient in performing EUS, in recognizing the anatomy it reveals, and in performing fine-needle aspiration biopsy.

ENDOSCOPIC ULTRASONOGRAPHY IN ACUTE PANCREATITIS

Finding the cause of acute pancreatitis can be challenging in patients who do not have typical risk factors, eg, those who do not drink substantial amounts of alcohol and in whom transabdominal ultrasonography fails to reveal gallstones.

Several studies have evaluated the role of EUS in recurrent “idiopathic” pancreatitis.^3–5 Causes of acute pancreatitis detectable with EUS included gallbladder and bile duct microlithiasis (stones smaller than 3 mm), cysts, intraductal papillary mucinous neoplasms, ampullary neoplasms, pancreas divisum, and pancreatic masses.

Stones, sludge. Transabdominal ultrasonography is often performed in the workup of acute pancreatitis to rule out gallbladder stones and biliary dilation. Unfortunately, it does a poor job of imaging the distal common bile duct, where culprit stones may reside.

EUS provides a high-quality view of the bile duct from the ampulla of Vater to the region of the hepatic hilum and is safer than endoscopic retrograde cholangiopancreatography (ERCP). The available evidence supports the use of EUS as a diagnostic test for bile duct stones.^3–7 In fact, using ERCP as the reference standard, EUS has been found to be more sensitive than transabdominal ultrasonography for bile duct stones.⁴

The yield of EUS for finding biliary sludge and stones may be high in patients with unexplained pancreatitis. EUS detected sludge, microlithiasis, or both in 33 of 35 patients with idiopathic acute pancreatitis who underwent transabdominal ultrasonography with negative results.⁸ Furthermore, most were symptom-free at an average of 10 months after cholecystectomy, suggesting that microlithiasis was the cause of the “idiopathic” pancreatitis.

EUS can also decrease the number of unnecessary ERCP procedures in patients with suspected biliary pancreatitis. In these patients, EUS can be performed as an initial diagnostic test to exclude retained biliary stones. If a stone is present, the endoscopist can proceed to ERCP for sphincterotomy and stone removal during the same endoscopic session. If EUS is negative, the endoscopy can be concluded without cannulating the bile duct and putting the patient at risk of acute pancreatitis. In one report, this approach eliminated the need for ERCP in five of six patients with suspected biliary pancreatitis.⁶

Tumors and other causes of bile duct obstruction can also cause recurrent acute pancreatitis and may be difficult to detect with cross-sectional imaging. EUS, on the other hand, can detect small pancreatic masses (< 2 cm), which may be missed by conventional computed tomography. Also, a linear EUS scope, with its forward oblique view, can image the duodenum and ampulla, where obstructing inflammation, tumors, and polyps may be found. One should strongly suspect occult malignancy in elderly patients with unexplained acute pancreatitis. In those patients, repeat imaging with high-resolution dual-phase computed tomography or with EUS should be considered after a few weeks once the acute inflammation resolves.

Pancreas divisum is a relatively common congenital abnormality in which the dorsal and ventral pancreatic ducts do not properly fuse during embryonic development. To rule out pancreas divisum, the endosonographer must carefully trace the pancreatic duct from the dorsal pancreas into the ventral pancreas, where it connects with the bile duct at the duodenal wall.

In summary, EUS appears to be safe and accurate for diagnosing bile duct stones and other structural causes of idiopathic acute pancreatitis.

ENDOSCOPIC ULTRASONOGRAPHY IN CHRONIC PANCREATITIS

Chronic pancreatitis, a relatively common and sometimes debilitating cause of chronic upper abdominal pain, may be difficult to diagnose using noninvasive imaging tests. Minimal-change chronic pancreatitis is defined as a syndrome of pancreatic abdominal pain with no or slight structural changes detected on imaging but with histologic inflammation and fibrosis diagnostic of chronic pancreatitis.⁹

A clinical rationale for trying to detect chronic pancreatitis early in its course is that interventions can be started earlier. These include abstinence from alcohol, giving exogenous pancreatic enzymes, and advanced interventions such as celiac plexus blocks for pain control. Some patients may even benefit from resection of the pancreas if pain is severe and resistant to conservative measures.

EUS can detect both parenchymal and ductal changes that correlate with histologic fibrosis.¹⁰ Parenchymal changes include hyperechoic foci, hyperechoic strands, lobularity, cysts, and shadowing calcifications. Ductal changes include dilation of the main pancreatic duct, irregularity, hyperechoic duct margins, and visible side branches.

Several studies have evaluated the ability of EUS to diagnose early chronic pancreatitis.^9,11–15 Reference standards used to determine the accuracy of EUS have included histology,^10,16–18 pancreatic function testing,^19–22 and ERCP.^11,15,23,24

The best diagnostic test may be pancreatic histology. However, biopsy of the pancreas is impractical and exposes patients to high risk. In addition, the patchy and focal distribution of histologic changes may decrease its reliability. Fortunately, the histologic findings of fibrosis have been shown to correlate with EUS criteria in patients undergoing EUS before surgical resection in three recent studies.^16–18 A threshold of four or more criteria out of a possible nine was found to provide the optimal sensitivity and specificity for histologic pancreatic fibrosis.^16,17 The criteria used were four parenchymal features (hyperechoic foci, strands, hypoechoic lobules, cysts) and five ductal features (irregularity of the main pancreatic duct, dilation, hyperechoic duct walls, visible side branches, and calcifications or stones).

EUS is sensitive for chronic pancreatitis, but ‘true’ accuracy is impossible to know

Unfortunately, greater sensitivity may come at the expense of worse specificity. Certain demographic variables may alter the EUS appearance of the pancreas. A multivariate analysis²⁵ found several variables that predicted abnormalities on EUS even in the absence of clinically evident pancreatitis; the strongest were heavy ethanol use (odds ratio [OR] 5.1, 95% confidence interval [CI] 3.1–8.5), male sex (OR 1.8, 95% CI 1.3–2.55), clinical suspicion of pancreatic disease (OR 1.7, 95% CI 1.2–2.3), and heavy smoking (OR 1.7, 95% CI 1.2–2.4). More prospective studies are needed to further differentiate true disease from false-positive findings of chronic pancreatitis.

Also, traditional EUS scoring symptoms have counted features in an unweighted fashion and assigned an arbitrary cut point (eg, four or more features) for diagnosis. This approach fails to account for the greater importance of some features (eg, calcifications) compared with others.

Interobserver variability is another important limitation of EUS in diagnosing chronic pancreatitis.^26,27 In one multicenter study of EUS interpretation, the overall kappa (agreement beyond chance) was only 0.45 for overall chronic pancreatitis diagnosis and worse for many individual criteria for chronic pancreatitis. The endosonographers disagreed most about hyperechoic strands and foci, main pancreatic duct irregularity, and visible side branches (kappa < 0.4).

The Rosemont classification

These limitations led a group of experts to meet in Chicago, IL, to develop a consensus-based and weighted EUS scoring system for the diagnosis of chronic pancreatitis, termed the Rosemont classification.

In this system, the previous parenchymal and ductal features are assigned stricter definitions and reclassified as major and minor criteria. Based on the presence of major and minor features, EUS results are stratified as “normal,” “indeterminate for chronic pancreatitis,” “suggestive of chronic pancreatitis,” or “most consistent with chronic pancreatitis.”^15,28

Further validation of this scoring system is needed before it can be used widely.

ENDOSCOPIC ULTRASONOGRAPHY PLUS PANCREATIC FUNCTION TESTING

The best way to diagnose minimal-change chronic pancreatitis may be a combination of sensitive structural and functional testing. Although clinically apparent steatorrhea typically occurs late in the course of chronic pancreatitis, mild exocrine insufficiency may occur early and is detectable with hormone-stimulated pancreatic function testing. Therefore, pancreatic function tests are considered sensitive for diagnosing chronic pancreatitis.^20,21,29

Endoscopic pancreatic function testing involves injecting secretin intravenously and then collecting duodenal aspirates through the endoscope. The duodenal fluid is analyzed for bicarbonate concentration as a measure of exocrine function.²⁹

We have studied combined EUS and endoscopic pancreatic function testing in the diagnosis of chronic pancreatitis.¹⁶ The combination gives a simultaneous structural and functional assessment of the pancreas and may optimize sensitivity for detecting minimal-change chronic pancreatitis. In a small study, we found the combination had 100% sensitivity for noncalcific chronic pancreatitis compared with a histologic reference standard.¹⁶

Endoscopic ultrasonography (EUS) is a minimally invasive test that provides high-resolution imaging of the pancreas.^1,2 As such, it is proving useful.

Accurate diagnosis and timely intervention are essential in managing acute and chronic pancreatitis, which are often encountered in the clinic and the hospital. However, the cause of acute pancreatitis is not always easy to determine. Furthermore, recurrent bouts can progress to chronic pancreatitis if the cause is not identified and eliminated. EUS has been studied extensively in the evaluation of both acute and chronic pancreatitis, as it can identify obstructive and biliary causes of acute pancreatitis and early structural features of chronic pancreatitis.

This article will review the indications and evidence for EUS in the evaluation of acute and chronic pancreatitis.

SPECIALIZED TRAINING REQUIRED

EUS involves passage of a specialized endoscope through the esophagus and stomach and into the duodenum. The scope has a very small ultrasound probe at the tip, allowing detailed imaging of the upper gastrointestinal tract and surrounding organs.

There are two types of EUS endoscope: radial and linear. A radial scope provides a 360° range of view perpendicular to the long axis of the scope. A linear scope provides a 150° view parallel to the long axis of the scope. Many endosonographers favor linear EUS for imaging the pancreas because it permits fine-needle aspiration biopsy of masses, cysts, and lymph nodes.

Specialized training beyond the gastroenterology fellowship is usually required to become proficient in performing EUS, in recognizing the anatomy it reveals, and in performing fine-needle aspiration biopsy.

ENDOSCOPIC ULTRASONOGRAPHY IN ACUTE PANCREATITIS

Finding the cause of acute pancreatitis can be challenging in patients who do not have typical risk factors, eg, those who do not drink substantial amounts of alcohol and in whom transabdominal ultrasonography fails to reveal gallstones.

Several studies have evaluated the role of EUS in recurrent “idiopathic” pancreatitis.^3–5 Causes of acute pancreatitis detectable with EUS included gallbladder and bile duct microlithiasis (stones smaller than 3 mm), cysts, intraductal papillary mucinous neoplasms, ampullary neoplasms, pancreas divisum, and pancreatic masses.

Stones, sludge. Transabdominal ultrasonography is often performed in the workup of acute pancreatitis to rule out gallbladder stones and biliary dilation. Unfortunately, it does a poor job of imaging the distal common bile duct, where culprit stones may reside.

EUS provides a high-quality view of the bile duct from the ampulla of Vater to the region of the hepatic hilum and is safer than endoscopic retrograde cholangiopancreatography (ERCP). The available evidence supports the use of EUS as a diagnostic test for bile duct stones.^3–7 In fact, using ERCP as the reference standard, EUS has been found to be more sensitive than transabdominal ultrasonography for bile duct stones.⁴

The yield of EUS for finding biliary sludge and stones may be high in patients with unexplained pancreatitis. EUS detected sludge, microlithiasis, or both in 33 of 35 patients with idiopathic acute pancreatitis who underwent transabdominal ultrasonography with negative results.⁸ Furthermore, most were symptom-free at an average of 10 months after cholecystectomy, suggesting that microlithiasis was the cause of the “idiopathic” pancreatitis.

EUS can also decrease the number of unnecessary ERCP procedures in patients with suspected biliary pancreatitis. In these patients, EUS can be performed as an initial diagnostic test to exclude retained biliary stones. If a stone is present, the endoscopist can proceed to ERCP for sphincterotomy and stone removal during the same endoscopic session. If EUS is negative, the endoscopy can be concluded without cannulating the bile duct and putting the patient at risk of acute pancreatitis. In one report, this approach eliminated the need for ERCP in five of six patients with suspected biliary pancreatitis.⁶

Tumors and other causes of bile duct obstruction can also cause recurrent acute pancreatitis and may be difficult to detect with cross-sectional imaging. EUS, on the other hand, can detect small pancreatic masses (< 2 cm), which may be missed by conventional computed tomography. Also, a linear EUS scope, with its forward oblique view, can image the duodenum and ampulla, where obstructing inflammation, tumors, and polyps may be found. One should strongly suspect occult malignancy in elderly patients with unexplained acute pancreatitis. In those patients, repeat imaging with high-resolution dual-phase computed tomography or with EUS should be considered after a few weeks once the acute inflammation resolves.

Pancreas divisum is a relatively common congenital abnormality in which the dorsal and ventral pancreatic ducts do not properly fuse during embryonic development. To rule out pancreas divisum, the endosonographer must carefully trace the pancreatic duct from the dorsal pancreas into the ventral pancreas, where it connects with the bile duct at the duodenal wall.

In summary, EUS appears to be safe and accurate for diagnosing bile duct stones and other structural causes of idiopathic acute pancreatitis.

ENDOSCOPIC ULTRASONOGRAPHY IN CHRONIC PANCREATITIS

Chronic pancreatitis, a relatively common and sometimes debilitating cause of chronic upper abdominal pain, may be difficult to diagnose using noninvasive imaging tests. Minimal-change chronic pancreatitis is defined as a syndrome of pancreatic abdominal pain with no or slight structural changes detected on imaging but with histologic inflammation and fibrosis diagnostic of chronic pancreatitis.⁹

A clinical rationale for trying to detect chronic pancreatitis early in its course is that interventions can be started earlier. These include abstinence from alcohol, giving exogenous pancreatic enzymes, and advanced interventions such as celiac plexus blocks for pain control. Some patients may even benefit from resection of the pancreas if pain is severe and resistant to conservative measures.

EUS can detect both parenchymal and ductal changes that correlate with histologic fibrosis.¹⁰ Parenchymal changes include hyperechoic foci, hyperechoic strands, lobularity, cysts, and shadowing calcifications. Ductal changes include dilation of the main pancreatic duct, irregularity, hyperechoic duct margins, and visible side branches.

Several studies have evaluated the ability of EUS to diagnose early chronic pancreatitis.^9,11–15 Reference standards used to determine the accuracy of EUS have included histology,^10,16–18 pancreatic function testing,^19–22 and ERCP.^11,15,23,24

The best diagnostic test may be pancreatic histology. However, biopsy of the pancreas is impractical and exposes patients to high risk. In addition, the patchy and focal distribution of histologic changes may decrease its reliability. Fortunately, the histologic findings of fibrosis have been shown to correlate with EUS criteria in patients undergoing EUS before surgical resection in three recent studies.^16–18 A threshold of four or more criteria out of a possible nine was found to provide the optimal sensitivity and specificity for histologic pancreatic fibrosis.^16,17 The criteria used were four parenchymal features (hyperechoic foci, strands, hypoechoic lobules, cysts) and five ductal features (irregularity of the main pancreatic duct, dilation, hyperechoic duct walls, visible side branches, and calcifications or stones).

EUS is sensitive for chronic pancreatitis, but ‘true’ accuracy is impossible to know

Unfortunately, greater sensitivity may come at the expense of worse specificity. Certain demographic variables may alter the EUS appearance of the pancreas. A multivariate analysis²⁵ found several variables that predicted abnormalities on EUS even in the absence of clinically evident pancreatitis; the strongest were heavy ethanol use (odds ratio [OR] 5.1, 95% confidence interval [CI] 3.1–8.5), male sex (OR 1.8, 95% CI 1.3–2.55), clinical suspicion of pancreatic disease (OR 1.7, 95% CI 1.2–2.3), and heavy smoking (OR 1.7, 95% CI 1.2–2.4). More prospective studies are needed to further differentiate true disease from false-positive findings of chronic pancreatitis.

Also, traditional EUS scoring symptoms have counted features in an unweighted fashion and assigned an arbitrary cut point (eg, four or more features) for diagnosis. This approach fails to account for the greater importance of some features (eg, calcifications) compared with others.

Interobserver variability is another important limitation of EUS in diagnosing chronic pancreatitis.^26,27 In one multicenter study of EUS interpretation, the overall kappa (agreement beyond chance) was only 0.45 for overall chronic pancreatitis diagnosis and worse for many individual criteria for chronic pancreatitis. The endosonographers disagreed most about hyperechoic strands and foci, main pancreatic duct irregularity, and visible side branches (kappa < 0.4).

The Rosemont classification

These limitations led a group of experts to meet in Chicago, IL, to develop a consensus-based and weighted EUS scoring system for the diagnosis of chronic pancreatitis, termed the Rosemont classification.

In this system, the previous parenchymal and ductal features are assigned stricter definitions and reclassified as major and minor criteria. Based on the presence of major and minor features, EUS results are stratified as “normal,” “indeterminate for chronic pancreatitis,” “suggestive of chronic pancreatitis,” or “most consistent with chronic pancreatitis.”^15,28

Further validation of this scoring system is needed before it can be used widely.

ENDOSCOPIC ULTRASONOGRAPHY PLUS PANCREATIC FUNCTION TESTING

The best way to diagnose minimal-change chronic pancreatitis may be a combination of sensitive structural and functional testing. Although clinically apparent steatorrhea typically occurs late in the course of chronic pancreatitis, mild exocrine insufficiency may occur early and is detectable with hormone-stimulated pancreatic function testing. Therefore, pancreatic function tests are considered sensitive for diagnosing chronic pancreatitis.^20,21,29

Endoscopic pancreatic function testing involves injecting secretin intravenously and then collecting duodenal aspirates through the endoscope. The duodenal fluid is analyzed for bicarbonate concentration as a measure of exocrine function.²⁹

We have studied combined EUS and endoscopic pancreatic function testing in the diagnosis of chronic pancreatitis.¹⁶ The combination gives a simultaneous structural and functional assessment of the pancreas and may optimize sensitivity for detecting minimal-change chronic pancreatitis. In a small study, we found the combination had 100% sensitivity for noncalcific chronic pancreatitis compared with a histologic reference standard.¹⁶