Eldon A. Shaffer

Sulfasalazine Revisited: A Meta-Analysis of 5-Aminosalicylic Acid in the Treatment of Ulcerative Colitis

The first major therapeutic advance in ulcerative colitis was the discovery of sulfasalazine (sulfapyridine bonded to 5-aminosalicylic acid [5-ASA]) [1], which was the first effective therapeutic agent for active ulcerative colitis [2] and for maintenance of remission [3]. Currently, all patients with ulcerative colitis in remission take sulfasalazine indefinitely [4]. The maximum dose of sulfasalazine is limited by the frequency of adverse drug reactions [5]. Approximately 30% of patients taking sulfasalazine report adverse reactions; many are idiosyncratic but others are dose related. Azad Khan [6] and others [7] showed that 5-ASA was the active moiety of sulfasalazine and that sulfapyridine functioned only as a carrier. Many investigators believed that the development of a new delivery system would allow the use of higher doses of 5-ASA, which might provide additional therapeutic benefits to patients. Currently five new preparations are available. They can be broadly classified into three categories: pH-dependent formulations; microsphere formulations; and those formulations that bind 5-ASA to another carrier, which requires splitting of the diazo bond by bacteria. Although generic names are generally preferable, in this situation proprietary names are required to identify each medication by its release mechanisms. The first 5-ASA delayed-release, pH-sensitive preparation (Asacol) (Proctor & Gamble Pharmaceuticals, Inc., Norwich, New York) is a pellet of 5-ASA coated with a resin (Eudragit S) that dissolves when the pH of the fecal contents is more than 7 [8]. A second pH-dependent, 5-ASA preparation (Claversal, Salofalk, Mesasal, Rowasa) is coated with another resin (Eudragit L) that dissolves at a fecal pH that is more than 6 [9]. The only currently available microsphere preparation of 5-ASA (Pentasa) consists of 5-ASA coated with an ethylcellulose that allows for slow release of the medication beginning in the duodenum and extending into the colon [9]. The alternative-carrier approach includes binding 5-ASA to another 5-ASA molecule, olsalazine (Dipentum, Kabi Pharmacia, Piscataway, New Jersey), or to various benzoic acid derivatives, such as balsalazide [10] or benzalazine [11]. The newer products are more expensive for patients, and their superiority over sulfasalazine has not yet been proven. To better define the role and efficacy of these new 5-ASA-containing medications, we did a meta-analysis of the randomized, controlled trials comparing them with either placebo or sulfasalazine. Methods A literature search was done using the MEDLINE and BIOS data bases for January 1981 to May 1992 and using the MESH headings ulcerative colitis and aminosalicylates. Citations were limited to those published in English. A manual search sought additional citations using references in articles retrieved from the computerized search as well as review articles, proceedings from symposia, and textbooks. Studies of oral 5-ASA preparations for the acute and maintenance treatment of ulcerative colitis were accepted for analysis if they were randomized, controlled, of parallel design, and had a treatment duration of a minimum of 4 weeks for acute therapy and of 6 months for maintenance therapy. Studies accepted for analysis were reviewed independently by two of the authors (LRS and GRM), and the results of each study were recorded. For reports of acute therapy, the numbers of patients with clinical, endoscopic, or histologic remission as defined in each study were recorded. For studies of maintenance therapy, the number of patients remaining in remission clinically or endoscopically were recorded. Results were recorded on an intention-to-treat basis. No uniform definition of a response existed across the studies. Where possible the numbers of patients having adverse effects or withdrawing from therapy because of adverse events were recorded. Original investigators were contacted when necessary to clarify points regarding inclusion criteria or data. We did a blinded quality analysis of each study we selected, using the criteria and methods outlined by Chalmers and colleagues [12] (Appendix Table 1, Appendix Table 2, and Appendix Table 3). Any discrepancies between reviewers about the retrieved data or quality assessment were settled by consensus. Appendix Table 1. Quality Assessment Scoring System: Study Protocol* Appendix Table 2. Quality Assessment Scoring System: Statistical Analysis Appendix Table 3. Quality Assessment Scoring System: Presentation of Results Studies were separated into three groups: 5-ASA compared with placebo for acute therapy, 5-ASA compared with sulfasalazine for acute therapy, and 5-ASA compared with sulfasalazine for maintenance therapy. Each group was analyzed separately. For each group, homogeneity was tested using the method outlined by DerSimonian and Laird [13]. Pooled odds ratios with 95% confidence intervals (CIs) were calculated using a commercially available, computerized biostatistical program [14]. To confirm the degree of homogeneity, all of the individual trial 95% CIs were also graphed to confirm that a substantial overlap existed for all of the individual CIs (see Figures 1, 2, and 3). (Note that if the CIs for a pooled odds ratio include unity, the evidence supporting an actual difference is weak or inconclusive.) If possible, results were also subgrouped according to the dose of 5-ASA and the specific 5-ASA preparation used; these subgroups were analyzed in a similar manner. Adverse effects and withdrawals were analyzed by pooled odds ratio and by combining the adverse effects for each group; the overall results were analyzed using chi-square tests. The Fisher exact test was used when appropriate for small counts. Figure 1. Randomized, controlled studies of various 5-aminosalicylic acid preparations (all study arms) compared with placebo in active ulcerative colitis. Figure 2. Randomized, controlled studies of various 5-aminosalicylic acid (5-ASA) preparations (all study arms) compared with sulfasalazine in active ulcerative colitis. Figure 3. Randomized, controlled studies of various 5-aminosalicylic acid (5-ASA) preparations (all study arms) compared with sulfasalazine for maintenance of remission of ulcerative colitis. Results From the more than 150 citations collected by the literature search, 59 trials of oral 5-ASA therapy for mild-to-moderate ulcerative colitis as defined by Truelove and Witts [15] were identified. Twenty-seven trials (45.7%) Table 1 and Table 2 met the inclusion criteria and were included in the analysis [11,16-41]. Thirty-two studies were excluded. Fifteen of these studies were open label and were excluded due to lack of randomization [8, 42-55]. Of the remaining 17 studies, the reasons for exclusion included design other than controlled, with parallel design in 3 studies [56-58]; duration less than 4 weeks in 2 studies [59, 60]; single-center reports of patients who were abstracted from larger multicenter trials in 2 [61, 62]; data presented in an inappropriate form for our analysis in 3 abstracts [63-65]; 5 abstracts subsequently published [66-70]; and 2 papers also being published in symposia [71, 72]. Table 1. Characteristics of Each Trial Included in the Meta-Analysis Table 2. Treatment Regimens and Outcome for Each Trial Included in the Meta-Analysis Quality scores, from Chalmers criteria [12], ranged from 0.25 to 0.80, with a median of 0.62 (see Table 1). Scores were normally distributed (Shapiro-Wilks test for normality = 0.920, P = 0.13). Excellent correlation existed between observers, with a correlation coefficient of 0.88 (P < 0.001). Most studies lost points for omissions about study design and use of statistics. No apparent effect of study quality on the results existed within each group. 5-Aminosalicylic Acid Compared with Placebo for Acute Therapy Eight trials, comprising 1007 patients, examined oral 5-ASA compared with placebo for acute mild-to-moderate ulcerative colitis [18, 23-25, 27, 33, 37, 39]. Studies were homogeneous (P > 0.2). The pooled odds ratio for complete remission combining all 5-ASA preparations with all doses was 2.02 (CI, 1.50 to 2.72) (Figure 1). Results were not significantly different if abstracts were excluded from the analysis. If the results were grouped according to the dose of 5-ASA used, a dose-dependent trend was apparent (Table 3). For patients who received less than 2 g of 5-ASA, the response was not different from placebo. The pooled odds ratio, for patients receiving at least 2 g/d of 5-ASA, showed improvement compared with placebo. Table 3. Pooled Odds Ratios for Studies of 5-Aminosalicylic Acid Compared with Placebo by Dose No significant differences in response could be shown between any of the 5-ASA preparations (Table 4). Only three studies [24, 27, 39], all with Dipentum, had sufficient data to determine an endoscopic response to therapy. The pooled odds ratio for endoscopic response was 4.97 (CI, 2.50 to 9.85). Table 4. Pooled Odds Ratios for 5-Aminosalicylic Acid Therapy for Ulcerative Colitis by Preparation 5-Aminosalicylic Acid Compared with Sulfasalazine for Acute Therapy Eight trials, involving 553 patients, compared 5-ASA with sulfasalazine for the acute treatment of ulcerative colitis [11, 16, 19, 29-31, 34, 35]. Studies were homogeneous (P > 0.2). The pooled odds ratio for complete remission, after combining all 5-ASA preparations and comparing them with sulfasalazine, was 1.15 (CI, 0.83 to 1.61) (Figure 2). The result was similar if patients entering into complete remission were combined with those showing improvement; pooled odds ratio was 1.20 (CI, 0.86 to 1.69). The results did not differ significantly when abstracts were excluded. If studies were subgrouped according to the dose of 5-ASA used, no evidence existed for a dose-response relationship. Trials using 2 g/d of 5-ASA had a pooled odds ratio of 1.24 (CI, 0.79 to 1.94), [11, 16, 29, 31, 34, 35], whereas those using less than 2 g/d h

Gallbladder polyps: Epidemiology, natural history and management

Polypoid lesions of the gallbladder affect approximately 5% of the adult population. Most affected individuals are asymptomatic, and their gallbladder polyps are detected during abdominal ultrasonography performed for unrelated conditions. Although the majority of gallbladder polyps are benign, most commonly cholesterol polyps, malignant transformation is a concern. The differentiation of benign from malignant lesions can be challenging. Several features, including patient age, polyp size and number, and rapid growth of polyps, are important discriminating features between benign and malignant polyps. Based on the evidence highlighted in this review, the authors recommend resection in symptomatic patients, as well as in asymptomatic individuals over 50 years of age, or those whose polyps are solitary, greater than 10 mm in diameter, or associated with gallstones or polyp growth on serial ultrasonography. Novel imaging techniques, including endoscopic ultrasonography and enhanced computed tomography, may aid in the differential diagnosis of these lesions and permit expectant management.

Changes in gallbladder motility and gallstone formation following laparoscopic gastric banding for morbid obesity

UNLABELLEDnMorbid obesity is associated with cholesterol gallstone formation, a risk compounded by rapid weight loss. Laparoscopic gastric banding allows for a measured rate of weight loss, but the subsequent risk for developing gallstones is unknown.nnnMETHODnTwenty-six normal-weight volunteers (body mass index [BMI] less than 30) were compared with 14 morbidly obese patients (BMI greater than 40). Gallbladder volumes were measured ultrasonographically, after fasting and following stimulation with intravenous cholecystokinin-octapeptide (CCK-8)nnnRESULTSnPreoperatively, fasting gallbladder volume and residual volume after CCK stimulation were both two times greater in the obese group (P<0.02 versus controls). Per cent gallbladder emptying was not different. Gallbladder refilling was four times higher in the obese patients (P<0.01). By six weeks postoperatively, the obese patients lost 1.4+/-0.1% body weight per week. Gallbladder emptying decreased 18.4% (80.3+/-3.9% to 65.5+/-6.9%; P<0.05); residual volume rose one-third (not significant), and refilling fell 60.5% (0.43+/-0.09 to 0.26+/-0.04 mL/min; P=0.07). Three patients with weight losses of greater than 1.7% per week developed gallstones; gallbladder emptying fell outside the 95 percentile. By six months, weight loss slowed to 0.5+/-0.1% per week; gallbladder motility improved modestly. No further stones developed.nnnCONCLUSIONnRapid weight loss following laparoscopic gastric banding impairs gallbladder emptying and when pronounced, gallstones form by six weeks postoperatively. The accompanying reduction in gallbladder emptying, increased gallbladder residual volume and decreased refilling promote gallbladder stasis and hence stone formation.

The Hot Air and Cold Facts of Dietary Fibre

Division of Gastroenterology, Faculty of Medicine, University of Calgary, Calgary, Alberta Correspondence: Dr Eldon A Shaffer, Division of Gastroenterology, Faculty of Medicine, University of Calgary, 3330 Hospital Drive South West, Calgary, Alberta T2N 4N1. Telephone 403-210-9363, fax 403-210-9358, e-mail shaffer@ucalgary.ca Received for publication July 28, 2005. Accepted August 2, 2005 America is a constipated nation.... If you pass small stools, you have to have large hospitals – Denis Burkitt

Nonalcoholic steatohepatitis: more than just being fat.

318 Fatty liver (steatosis) is an excessive accumulation of lipids within hepatocytes, most commonly in the form of large droplets of triglyceride (macrovesicular fat). The other form of fat accumulation, microvesicular steatosis with tiny fat droplets (such as fatty liver of pregnancy and Reyex92s syndrome), is quite rare and often rapidly fatal. In contrast, simple macrovesicular fatty liver was once considered quite benign and unimportant x96 an incidental finding with an obscure pathogenesis. Long considered a rather innocuous entity associated with obesity and diabetes, this form of steatosis was thought to progress rarely to advanced liver disease, particularly when not associated with alcohol use (1-3). Over the past two decades, it has become apparent that some of these patients have (or develop) a wide spectrum of lesions. Nonalcoholic fatty liver disease (NAFLD) is now recognized as a clinicopathological entity that extends beyond uncomplicated steatosis to steatohepatitis (fat plus inflammation, histologically resembling alcoholic hepatitis), steatonecrosis, advanced fibrosis, liver failure and, in some cases, hepatocellular carcinoma (2-7). In fact, nonalcoholic steatohepatitis (NASH) may frequently be a precursor of cryptogenic cirrhosis (8). NASH is probably the crucial stage in the progression of NAFLD to advanced liver disease. NASH is usually asymptomatic. Frequently, the only clues to its presence are smooth hepatomegaly and/or elevated transaminase levels (6,9). The alanine aminotransferase (ALT) level exceeds the aspartate aminotransferase (AST) level in most cases, unless cirrhosis has supervened. Ultrasonography may reveal only the x91brightx92 appearance of the fatty liver. None of these findings is diagnostic. Of fundamental importance is the absence of excessive alcohol intake, which is best defined as more than 20 g/day for women or more than 30 g/day for men, although this may be difficult to confirm in practice (6). Hepatitis C and other liver diseases also must be excluded. Liver biopsy is necessary to distinguish NASH from simple steatosis, to detect fibrosis or cirrhosis, and to exclude other forms of liver disease. Ludwig et al (10) coined the term, x91NASHx92 and pointed out its histological similarity to alcoholic hepatitis. The minimum diagnostic requirements are macrovesicular fat and lobular inflammation. Stricter criteria include evidence of hepatic degeneration and/or fibrosis (11,12). Hepatocellular injury manifests as ballooning degeneration (fluid accumulates and causes hepatocytes to swell) or, less commonly, as acidophilic degeneration (Malloryx92s hyaline bodies) (9,12). A stronger definition would facilitate epidemiological and clinical investigations. The true prevalence of NASH is unknown because the need for liver biopsy inherently creates a selection bias. NASH has been found in up to 11% of liver biopsy specimens and 6% of autopsies (10,13,14). It occurs more frequently in obese women who have type II diabetes or hypertriglyceridemia (6). Although it is often not recogNonalcoholic steatohepatitis: More than just being fat

Insurance Rating of Patients with Inflammatory Bowel Disease: Report of a Conference on Morbidity and Mortality

Patient members reported to the Crohn’s and Colitis Foundation of Canada (CCFC) about their difficulties to obtain insurance. In 1991, the Lay Board of the CCFC requested its Medical Advisory Board (MAB) to investigate this problem. At that time, insurance ratings could be illustrated by the 1985 edition of Brackenridge’s monograph on life risks. The MAB found that data on mortality were outdated. A conference on morbidity and mortality of inflammatory bowel disease (IBD) was organized by the authors and held in May 1992. Based on questionnaires to patients, evidence provided by invited speakers and the results of small group conferences, it was concluded that patients with IBD have difficulties in obtaining insurance, even though the quality of life and mortality of IBD patients is not very different from that of the general population. However, the mortality rate of the healthy insured population is lower than that of the general population, and thus much lower than that of IBD patients. Patients have a better chance to obtain insurance if there is a close cooperation between the treating physician and the medical officer of the insurance company. Changes have occurred since the conference held in May 1992. The recent edition of Brackenridge’s text (1992) provides a better prognosis but unfortunately unchanged rating for patients with IBD than did the 1985 edition. Close cooperation between the Patient Advisory Committee of the CCFC and the Executives of the Canadian Life Insurance Medical Officers Association may further improve the insurance rating of patients with IBD.

S1982 Rising Incidence of Eosinophilic Esophagitis in the Calgary Health Region: A Population-Based Study

with abnormal 24-hour pH monitoring (p= .08). In patients who were not taking a PPI, there was no significant difference in number of patients with pathologic reflux or distribution of eosinophils. On multivariate analysis, age, gender and presence of pathologic reflux were not factors associated with distal esophageal eosinophils. There was no significant difference in median eosinophils found in distal biopsies of patients with endoscopic reflux when compared to those with normal EGD and normal esophageal pHmonitoring. CONCLUSIONS We found no causal interaction between reflux and distal esophageal eosinophils, by histologic or pathologic studies. The histologic finding of distal esophageal eosinophils is not necessarily related to GER and may be diagnostic of EoE. Biopsies from multiple sites in the esophagus, as well as formal testing for pathologic reflux should be performed in all patients in whom the diagnosis of EoE is a consideration.

Acalculous Biliary Pain: Motility Dysfunction and Functional Pain

satisfactory outcome with relief of symptoms following surgery (1). Such a high response may in part be attributable to high patient expectations that their symptoms will worsen without surgery, their assumption being that surgery will provide a cure. Presumably this biliary-type pain originates either from obstruction of the gallbladder contracting against a fixed (eg, gallstone) or functional (eg, muscle spasm) obstruction at the cystic duct, or at the level of the sphincter of Oddi (SO), or from the inflammation (cholecystitis, cholangitis) that results. The basis of such pain is less clear if gallbladder smooth muscle contractility is defective, as occurs in cholesterol gallstone disease (2). Although biliary pain heralds cholelithiasis and cholecystitis, most (80%) people harbouring gallstones never experience pain (3). Gallstones and abdominal pain are thus not necessarily synonymous. It is, therefore, not surprising that chronic abdominal pain persists in up to 50% following cholecystectomy (4). Many of these complaints are nonspecific, but 14% have true biliary pain attributable to a definable cause, eg, SO dysfunction (5). The figures for surgical failures might be higher with the advent of laparoscopic cholecystectomy, which has dramatically increased the rate of surgeries over the past decade. Predictors of this ‘postcholecystectomy syndrome’ include psychological vulnerability (6), chronic symptoms before cholecystectomy and pain six weeks after cholecystectomy. The situation becomes even less clear when no structural abnormality is evident. Such ‘functional’ biliary type pain (7) (acalculous gallbladder disease) requires careful investigation to eliminate all structural abnormalities, particularly very small gallstones, as the basis for recurrent biliary-type pain. This includes repeating transabdominal ultrasound, which detects stones larger than 3 to 5 mm in diameter, performing an endoscopic ultrasound for tiny stones smaller than 3 mm in diameter and using accurate microscopy of gallbladder bile to detect microlithiasis (8). The accompanying article by Janowitz et al (pages 363-366) demonstrates that the lipid composition of human gallbladder bile remains stable even when frozen and stored. Although it is best to centrifuge bile immediately and use microscopy to detect cholesterol microcrystals and/or bilirubin granules, if a delay is unavoidable, bile can be frozen for later analysis. The cholecystokinin (CCK)-provocation test can cause pain in some normal individuals, depending on the rate of CCK infusion. It does not, however, predict a symptomatic benefit from cholecystectomy and should be abandoned when evaluating acalculous biliary pain (9). More accurate is the quantitative measurement of gallbladder emptying under controlled conditions with a slow infusion of CCK – CCK cholescintigraphy (7). In well defined patients with acalculous gallbladder disease and low ejection fractions on CCK ultrasonography, 59% to 75% continue to have symptoms (10,11). Conversely, some 90% experience complete or substantial relief following cholecystectomy (11-13). Here, impaired gallbladder emptying appears to be a marker of acalculous biliary pain. The basis for such impaired gallbladder emptying is a defect in the contractile machinery (14). Again, the dilemma is, what causes the pain? The biliary tract is a low-pressure conduit in which the gallbladder acts as a reservoir to decompress and regulate its pressure. Bile ducts lack a smooth muscle layer, leaving the SO a prime EDITORIAL

Simmering innards: Does irritable bowel syndrome have an immunological basis?

Not so long ago, physicians construed the irritable bowel syndrome (IBS) as being a neurotic trait: it was all in the head. Today most clinicians believe that the main abnormality lies in the brain (and spinal cord), which reacts abnormally to stimuli from the gut. Recent studies are identifying a basis for these neural changes - low grade inflammation in the gut - which may play a key role in IBS.

Too many authors spoil the credit.

Single authorship was the norm eons ago. According to rabbinical tradition, Moses wrote the five books of the Jewish bible, the Torah, meaning ‘Instruction’ or ‘Law’. Christian scholars credit God with breathing out the Bible. Before the mid-20th century, landmark scientific works from giants, such as Newton, Einstein and Fermi, were single authored. Although single authors wrote the vast majority (>98%) of important medical articles a century ago, this has become a rarity; <5% are now single authored. Meanwhile, the number of multi-authored articles has escalated, many of which list individuals who made insignificant contributions (1). At times, the list of authors reaches astronomical numbers, occupying as much space as the corresponding abstract. Extreme examples include a report in the physical sciences on the Large Hadron Collider listing almost 3000 authors, and a clinical trial published in The New England Journal of Medicine listing 974 authors (2). The basis for this rise in multiple authors does not simply reside in the complexities of current medical bioscience or the need for large multicentred clinical trials. Unmerited authorship is rampant. Twenty to thirty percent of medical science authors do not contribute substantially to the eventual peer-reviewed publication, particularly in large, multi-authored articles (1,3). One facet of undeserved authorship increasingly occurs in the form of ‘honorary authorship’ (ie, named authors who do not significantly contribute), which is granted either to chairs of departments as a convention or to more senior authors to boost the paper (4). Another is ‘ghost writers’ (unnamed authors who do contribute), employed by some biotechnology companies, aiming to lever ‘key opinion leaders’ and, so, portray the publication as originating in the academic domain, unsullied by commercial interests (5). n nTo assess original articles appropriately, scientists and clinicians must know the proper authorship and the origin and execution of the study, devoid of ghost writing or other bias. For the academic, authorship represents the means by which peers perceive their scholarly work. This establishes their reputation, productivity, grant support and opportunity for promotion. n nThe order in which authors are listed quantitatively identifies credit. Although there are some differences among disciplines, most have the authors listed according to the magnitude of their involvement in the work, placing the principal investigator last. Because some journals, such as Gastroenterology, limit authorship lists in the references to contain a maximum of three followed by the near afterthought ‘et al’, the senior author may select the third position so as to appear in any subsequent citations. Many publications, however, allow up to six authors before ‘et al’ (eg, Canadian Journal of Gastroenterology and Hepatology, Hepatology) to be listed in the references. Thus, this practice has disappeared. n nTo accommodate the growing number of authors per article, institutions and journals have established guidelines that seek to clarify the role, involvement and responsibility of each author. Since 1985, a voluntary, closed-membership group of select general medical editors (the International Committee of Medical Journal Editors) has provided guidance on authorship that continually evolves with time. This committee of editors provides clarification on accountability, roles and responsibilities of authors, fraud, conflict of interest and clinical trial registration. The International Committee of Medical Journal Editors “recommends that authorship be based on the following 4 criteria: substantial contributions to the conception or design of the work; or the acquisition, analysis, or interpretation of data for the work; AND drafting the work or revising it critically for important intellectual content; AND final approval of the version to be published; AND agreement to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved” (6). In its instructions to authors, the Canadian Journal of Gastroenterology and Hepatology fulfills some of these criteria: requesting a letter to indicate that all authors have participated in the research, and have reviewed and agree with the contents of the article. Few articles, however, clarify the actual contributions of the authors. n nAuthorship assigns responsibility and attributes credit. Substantive contribution must be a primary criterion. This should include significant involvement in the three components of any scientific publication: its original conception and design; implementation of the study including data collection and analysis; and, finally, writing major sections of the manuscript while being accountable for all of its content. Such decisions about authorship and order in the publication is best identified by the research team engaging in an open conversation. This should begin with the design of the study and continue throughout its implementation and manuscript submission. Recognizing individuals whose role represents a limited contribution may best be communicated through an ‘Acknowledgement’ section. n nMultiple authors are necessary in this increasingly complex biomedical world. Even in biblical times, 40 different authors from three continents, writing in three different languages, created the Bible. Moses required help to complete Deuteronomy as the last portion covered a time after his death.