daw
03-08-2005, 12:39 PM
From Nature Clinical Practice Gastroenterology & Hepatology
The Value of Surveillance and Other Unresolved Issues in the Management of Barrett's Esophagus
Raymond Playford
The basic requirements for diagnosing Barrett's esophagus have changed over time. Previously, the affected segment had to cover >/=3 cm, but length is no longer thought to have a major influence on the risk of developing carcinoma.[1] Similarly, histologic confirmation previously only required the presence of columnar cells, but most definitions now also require identification of intestinal metaplasia. Virtually all patients with a segment of columnar-lined esophagus >3 cm will have areas of intestinal metaplasia if enough biopsy samples are taken, although it is unclear if this holds true for shorter segments. As esophageal adenocarcinoma probably develops from areas of intestinal metaplasia, there is likely to be a higher risk if more intestinal metaplasia is present. Barrett's esophagus should not be confused with the presence of intestinal metaplasia at the gastroesophageal Z line. The relevance of this condition, which is sometimes termed ultrashort Barrett's esophagus, is uncertain.
Reflux of gastroduodenal contents is probably the major contributor to the development of Barrett's esophagus, although the relative contribution of acid versus duodenal contents, particularly bile, is unclear. Depending on how Barrett's esophagus is defined, 0.25–2% of the population and about 10% of patients who undergo endoscopy for symptoms of longstanding reflux have the condition[2], although some studies suggest that this might be an underestimate.
Barrett's esophagus is thought to be the major risk factor for developing esophageal adenocarcinoma. Cancer registry data indicate that there has been a fourfold rise in the incidence of esophageal adenocarcinoma in the Western world over the last 25 years, albeit from a very low baseline of about 0.7 per 100,000 person years.[3] Published figures for the level of risk of an individual with Barrett's esophagus developing adenocarcinoma have varied widely, although a recent analysis of the literature suggests the true figure is around 1 per 200 person years.[4]
The prognosis for esophageal adenocarcinoma is poor (5-year survival rate is generally accepted to be 5–10%). It is important that the healthcare burden of this disease is put into a wider context because there are only finite resources for all services, including endoscopy. In the US, in 1997, there were about 6,000 new cases of esophageal adenocarcinoma whereas about 131,000 people developed colonic adenocarcinoma.[3]
Most patients who are diagnosed with esophageal adenocarcinoma have not had a previous endoscopy; current surveillance programs are having minimal effects on population health. Should we therefore introduce screening of asymptomatic individuals? In addition to the question of money and resources, key to this decision is that screening is only logical if regular histologic inspection of the esophageal mucosa results in earlier detection of abnormalities and increases the chances of curative treatment. The evidence for the benefits of surveillance programs is, however, surprisingly weak. The major evidence comes indirectly from reports that compare the stage of disease and the survival of patients identified in a surveillance program with patients presenting from the community.[5] Although these studies have usually suggested a beneficial effect, they are often not randomized, not analyzed on an intention-to-treat basis and fail to statistically separate high grade dysplasia from cancer.
There are relatively few critical reviews of actual surveillance programs. A 10-year retrospective review indicated that five cancers developed in patients who were surveyed. Only in one patient was cancer identified as a direct result of surveillance; others were identified during an additional endoscopy for clinical symptoms.[6]
Results from cost-benefit MARCOVIAN-TYPE MODELS have suggested that surveillance of patients with Barrett's esophagus (without dysplasia) could be cost-effective if done at about 5-year intervals. The cost-benefit conclusion is, however, highly sensitive to the assumed malignant conversion rate. Unfortunately, because of the scarcity of data about the value of surveillance, and even the natural history of Barrett's mucosa, many of the fundamental assumptions on which these models are based could be flawed (e.g. that surveillance will be effective at reducing death related to esophageal cancer in 75% of patients). Several clinical studies have shown that patient participation is often <5 years, owing to withdrawal from surveillance programs for reasons such as development of serious co-morbidities.[6]
As the amount of data on the value of surveillance programs are limited, central organisations, such as the UK's National Institute for Clinical Excellence (NICE), should insist that large multicenter trials be undertaken if funding is to be continued for the surveillance of Barrett's esophagus. This would overcome the reluctance of centers who state that such trials are impractical, and who believe that it is inappropriate to withhold surveillance if Barrett's esophagus is found.
Follow-up studies of patients with Barrett's esophagus have shown that most patients die of unrelated causes and, even in the few who do develop esophageal adenocarcinoma, many will die from other pathologies.[7]As most individuals who enter into a surveillance program will, therefore, not benefit from it, there is an urgent need for biomarkers to be developed that can indicate those individuals who are most at risk of progression from the non- or low-grade dyplasia stage to severe dysplasia or cancer. More time and resources could then be focused on these individuals. The utility of several markers such as ANEUPLOIDY, and abnormal expression of molecules like CYCLIN D1 and P53 is currently under review.[8]
We know very little about the natural history of Barrett's mucosa. Most patients will not progress to malignancy but, in those that do, the sequence may run nondysplastic → low grade dysplasia → high grade dysplasia → carcinoma, in a similar way to colonic carcinoma, although for Barrett's mucosa this sequence of events can stop at any stage and might even reverse while still at the dysplastic stage. Mechanisms to reduce the continuing reflux of gastric/duodenal contents onto the Barrett's mucosa have, until recently, consisted of the use of PPIs or surgical (antireflux) intervention. To date, the evidence that these maneuvers change the natural history or risk of subsequent malignant development is poor, although partial histologic conversion to squamous mucosa can occur if profound acid suppression is achieved. More drastically, PHOTODYNAMIC THERAPY (PDT) with long-term acid suppression has been shown to reverse much of the affected mucosa and is likely to form a therapeutic alternative to surgery for subgroups of patients with high grade dysplasia.[9] Use of PDT in people with low grade dysplasia, and even nondysplastic Barrett's mucosa is, however, more controversial as there are associated serious side effects, risks and costs. As Barrett's mucosa contains quite large amounts of cyclooxygenase (COX),[8] trials are currently underway to examine whether long-term administration of aspirin or COX-2 inhibitors, with and without acid suppressants, can reduce the risk of malignant progression.
Both UK and US national guidelines advise that surveillance should be undertaken. Many UK gastroenterologists are not following these guidelines[10] and of those that are, many profess to doing so for fear of litigation. Similarly, US guidelines recommending screening have recently been criticised by an expert panel because of the lack of data on which this recommendation is based. As the fundamental problem is lack of robust data, rather than continuing to debate old information, we need to produce new data on the true value of surveillance, development of biomarkers and testing the efficacy of pharmacological and mechanical interventions.
http://www.medscape.com/viewarticle/499714?src=mp
The Value of Surveillance and Other Unresolved Issues in the Management of Barrett's Esophagus
Raymond Playford
The basic requirements for diagnosing Barrett's esophagus have changed over time. Previously, the affected segment had to cover >/=3 cm, but length is no longer thought to have a major influence on the risk of developing carcinoma.[1] Similarly, histologic confirmation previously only required the presence of columnar cells, but most definitions now also require identification of intestinal metaplasia. Virtually all patients with a segment of columnar-lined esophagus >3 cm will have areas of intestinal metaplasia if enough biopsy samples are taken, although it is unclear if this holds true for shorter segments. As esophageal adenocarcinoma probably develops from areas of intestinal metaplasia, there is likely to be a higher risk if more intestinal metaplasia is present. Barrett's esophagus should not be confused with the presence of intestinal metaplasia at the gastroesophageal Z line. The relevance of this condition, which is sometimes termed ultrashort Barrett's esophagus, is uncertain.
Reflux of gastroduodenal contents is probably the major contributor to the development of Barrett's esophagus, although the relative contribution of acid versus duodenal contents, particularly bile, is unclear. Depending on how Barrett's esophagus is defined, 0.25–2% of the population and about 10% of patients who undergo endoscopy for symptoms of longstanding reflux have the condition[2], although some studies suggest that this might be an underestimate.
Barrett's esophagus is thought to be the major risk factor for developing esophageal adenocarcinoma. Cancer registry data indicate that there has been a fourfold rise in the incidence of esophageal adenocarcinoma in the Western world over the last 25 years, albeit from a very low baseline of about 0.7 per 100,000 person years.[3] Published figures for the level of risk of an individual with Barrett's esophagus developing adenocarcinoma have varied widely, although a recent analysis of the literature suggests the true figure is around 1 per 200 person years.[4]
The prognosis for esophageal adenocarcinoma is poor (5-year survival rate is generally accepted to be 5–10%). It is important that the healthcare burden of this disease is put into a wider context because there are only finite resources for all services, including endoscopy. In the US, in 1997, there were about 6,000 new cases of esophageal adenocarcinoma whereas about 131,000 people developed colonic adenocarcinoma.[3]
Most patients who are diagnosed with esophageal adenocarcinoma have not had a previous endoscopy; current surveillance programs are having minimal effects on population health. Should we therefore introduce screening of asymptomatic individuals? In addition to the question of money and resources, key to this decision is that screening is only logical if regular histologic inspection of the esophageal mucosa results in earlier detection of abnormalities and increases the chances of curative treatment. The evidence for the benefits of surveillance programs is, however, surprisingly weak. The major evidence comes indirectly from reports that compare the stage of disease and the survival of patients identified in a surveillance program with patients presenting from the community.[5] Although these studies have usually suggested a beneficial effect, they are often not randomized, not analyzed on an intention-to-treat basis and fail to statistically separate high grade dysplasia from cancer.
There are relatively few critical reviews of actual surveillance programs. A 10-year retrospective review indicated that five cancers developed in patients who were surveyed. Only in one patient was cancer identified as a direct result of surveillance; others were identified during an additional endoscopy for clinical symptoms.[6]
Results from cost-benefit MARCOVIAN-TYPE MODELS have suggested that surveillance of patients with Barrett's esophagus (without dysplasia) could be cost-effective if done at about 5-year intervals. The cost-benefit conclusion is, however, highly sensitive to the assumed malignant conversion rate. Unfortunately, because of the scarcity of data about the value of surveillance, and even the natural history of Barrett's mucosa, many of the fundamental assumptions on which these models are based could be flawed (e.g. that surveillance will be effective at reducing death related to esophageal cancer in 75% of patients). Several clinical studies have shown that patient participation is often <5 years, owing to withdrawal from surveillance programs for reasons such as development of serious co-morbidities.[6]
As the amount of data on the value of surveillance programs are limited, central organisations, such as the UK's National Institute for Clinical Excellence (NICE), should insist that large multicenter trials be undertaken if funding is to be continued for the surveillance of Barrett's esophagus. This would overcome the reluctance of centers who state that such trials are impractical, and who believe that it is inappropriate to withhold surveillance if Barrett's esophagus is found.
Follow-up studies of patients with Barrett's esophagus have shown that most patients die of unrelated causes and, even in the few who do develop esophageal adenocarcinoma, many will die from other pathologies.[7]As most individuals who enter into a surveillance program will, therefore, not benefit from it, there is an urgent need for biomarkers to be developed that can indicate those individuals who are most at risk of progression from the non- or low-grade dyplasia stage to severe dysplasia or cancer. More time and resources could then be focused on these individuals. The utility of several markers such as ANEUPLOIDY, and abnormal expression of molecules like CYCLIN D1 and P53 is currently under review.[8]
We know very little about the natural history of Barrett's mucosa. Most patients will not progress to malignancy but, in those that do, the sequence may run nondysplastic → low grade dysplasia → high grade dysplasia → carcinoma, in a similar way to colonic carcinoma, although for Barrett's mucosa this sequence of events can stop at any stage and might even reverse while still at the dysplastic stage. Mechanisms to reduce the continuing reflux of gastric/duodenal contents onto the Barrett's mucosa have, until recently, consisted of the use of PPIs or surgical (antireflux) intervention. To date, the evidence that these maneuvers change the natural history or risk of subsequent malignant development is poor, although partial histologic conversion to squamous mucosa can occur if profound acid suppression is achieved. More drastically, PHOTODYNAMIC THERAPY (PDT) with long-term acid suppression has been shown to reverse much of the affected mucosa and is likely to form a therapeutic alternative to surgery for subgroups of patients with high grade dysplasia.[9] Use of PDT in people with low grade dysplasia, and even nondysplastic Barrett's mucosa is, however, more controversial as there are associated serious side effects, risks and costs. As Barrett's mucosa contains quite large amounts of cyclooxygenase (COX),[8] trials are currently underway to examine whether long-term administration of aspirin or COX-2 inhibitors, with and without acid suppressants, can reduce the risk of malignant progression.
Both UK and US national guidelines advise that surveillance should be undertaken. Many UK gastroenterologists are not following these guidelines[10] and of those that are, many profess to doing so for fear of litigation. Similarly, US guidelines recommending screening have recently been criticised by an expert panel because of the lack of data on which this recommendation is based. As the fundamental problem is lack of robust data, rather than continuing to debate old information, we need to produce new data on the true value of surveillance, development of biomarkers and testing the efficacy of pharmacological and mechanical interventions.
http://www.medscape.com/viewarticle/499714?src=mp