Using ordinal logistic regression to estimate the likelihood of colorectal neoplasia.

The utility of ordinal logistic regression in the prediction of colorectal neoplasia was demonstrated in a group of 461 consecutive patients undergoing colonoscopy in a community practice. One hundred twenty-nine patients had adenomatous polyps and 34 had colorectal adenocarcinoma. An ordinal logistic regression model developed in a random subset (292 patients) identified five predictors of colorectal neoplasia. Colorectal neoplasia risk could be predicted using the patient's age, sex, hematocrit, fecal occult blood test result and indication for colonoscopy. The risk of colorectal neoplasia in the remaining subset of patients (169) could be reliably estimated from the model. Ordinal logistic regression analysis in this select group of patients can accurately estimate the likelihood of colorectal neoplasia. Because the generalizability of our findings are unknown, the model should not be applied to other patients. However, application of this technique to an unselected group of patients not already referred for colonoscopy could provide unbiased estimates of colorectal neoplasia risk in individual patients.

Studies of gastric ulcer disease by community-based gastroenterologists.

A community-based group of gastroenterologists examined 623 patients (541 prospectively and 82 retrospectively) with endoscopically diagnosed gastric ulcer disease during a 12-month period. Patients averaged 60 years of age; the majority were women (62%). Women were less likely to smoke, abuse alcohol, and were more likely to present with abdominal pain (p less than 0.05). Whereas patients presenting with bleeding or requiring transfusion were less likely to complain of pain (p less than 0.05), they were more likely to be taking aspirin or nonsteroidal anti-inflammatory drugs and have prior history of bleeding (p less than 0.05). Patients with a prior history of ulcer disease were more likely to smoke, present with pain and use acetaminophen (p less than 0.05). Patients with large ulcers were more likely to bleed, present with pain, and obstruct (p less than 0.05). Multiple gastric ulcers were seen in patients taking aspirin or nonsteroidal anti-inflammatory drugs (p less than 0.05).

Randomized, double-blind comparison of famotidine with ranitidine in treatment of acute, benign gastric ulcer disease. Community-based study coupled with a patient registry.

A multicenter, double-blind, randomized controlled trial comparing the efficacy and safety of famotidine with ranitidine in the treatment of acute, benign gastric ulcer disease was coupled with a community-based gastric ulcer disease registry. One hundred ninety-five patients with endoscopically documented gastric ulcer disease were enrolled in the trial and randomly allocated to treatment with either famotidine 40 mg at bedtime or ranitidine 150 mg twice a day. Healing rates were similar in both groups: at four weeks 49% vs 48%, at six weeks 71% vs 69%, and at eight weeks 83% vs 81% for famotidine and for ranitidine, respectively. Pain relief, antacid tablet use, and adverse experiences were also similar in the two groups. Only 25% of patients entered in the gastric ulcer registry were enrolled in the trial. Given that patients with more severe or complicated gastric ulcer disease should be excluded from controlled trials of new drugs, the screening criteria used in the present study be excluded from findings being representative of a quarter of the patients seen in these practices. Therefore, coupling a patient registry with a clinical trial helps determine the applicability of its results. Famotidine 40 mg at bedtime is an effective and well-tolerated treatment of acute, benign gastric ulcer disease and is comparable in efficacy and safety to ranitidine 150 mg twice a day.

The natural history of cholelithiasis: the National Cooperative Gallstone Study.

The National Cooperative Gallstone Study, a double-masked, placebo-controlled, therapeutic trial of chenodiol (chenodeoxycholic acid), provided an opportunity to study the natural history of cholelithiasis in patients who choose nonsurgical management. The major component of the study comprised 916 patients, 305 of whom were randomly assigned to receive a placebo for 24 months. Among these 305 patients, the probability of having biliary tract pain during the 24 months of prospective evaluation was significantly increased if the patient had had a history of biliary tract pain in the 12 months before entry into the study (69% versus 31%). Thirty-eight percent of patients had stone growth (greater than 0.5 cm3), and 18% had a spontaneous decrease in stone volume. Despite the high incidence of biliary tract pain, nonelective cholecystectomy was required in only 4% of patients during the 24 months.

Major issues in the organization and implementation of the National Cooperative Gallstone Study (NCGS).

The National Cooperative Gallstone Study (NCGS) was a cooperative, randomized, controlled trial of a drug, chenodiol, for the medical dissolution of gallstones. The design and procedures of the NCGS were complex, having developed as a result of extensive involvement of many experts in the field of gallstone disease and biliary lipids. During the design and implementation of the protocol, many important issues required consideration and resolution. The aim of this article is to review these issues and the deliberations surrounding their resolution and provide personal conclusions and recommendations that may be helpful to other investigators involved in cooperative, controlled trials.

Pretreatment biliary lipid composition in white patients with radiolucent gallstones in the National Cooperative Gallstone Study.

Biliary lipid classes (bile acids, phospholipids, cholesterol) as well as individual biliary bile acids were measured in duodenal bile samples obtained before treatment from 284 white men and 264 white women participating in the National Cooperative Gallstone Study. The patients had radiolucent gallstones present in visualizing gallbladders. Calculated biliary cholesterol saturation was significantly higher in women (143 +/- 43, mean +/- SD, vs. 132 +/- 39 for men). Chenodeoxycholic acid was the major biliary bile acid in both sexes (40.0 +/- 9.9 in men; 38.8 +/- 9.3 in women, NS). Cholic acid was the second most common bile acid, constituting 32.9 +/- 8.8 in men and 31.8 +/- 8.9 in women (NS). When other demographic and clinical characteristics, including serum lipids, were related with biliary lipid composition, only percent ideal body weight correlated significantly. The partial correlation coefficient adjusted for percent ideal body weight indicated that the proportion of chenodeoxycholic acid correlated negatively with the mole fraction of cholesterol in bile in men, but not in women. Multiple regression analyses showed that bile saturation could not be predicted reliably from any clinical, chemical, or radiologic measurement in either sex. Published data for biliary lipid composition in individuals with biliary disease showed considerable overlap with the National Cooperative Gallstone Study data reported here, suggesting that cholesterol gallstone disease is not caused solely by increased biliary cholesterol saturation.

Analysis of patients referred to the gastroenterology service of a VA medical center.

We recorded clinical information over a 12-month period on consecutive consultations to the gastroenterology service of the Durham VA Medical Center. Of 902 consultations, 789 were prospectively collected. Eighty-five percent of the patients were between 40 and 70 years old. Seventy-five percent of the referrals were from the internal medicine service. The most frequent reasons for consultation were abdominal pain (19%), GI bleeding (active, 16%; occult, 9%), abnormal results of liver tests (18%), and request for a procedure (11%). Diseases of the liver (32%) and "peptic diseases" (30%) were the most common diagnoses. One or more procedures were done in 71% of consultations. When these data are compared with those of a practicing gastroenterologist, using an identical instrument, it is apparent that trainees' experience with structurally identifiable gastroenterologic disease and with a variety of procedures was similar in scope. There were, however, differences in that the physicians at the VA saw substantially fewer patients with so-called "functional" illness. If these data are applicable to other VA Medical Centers, then the training of physicians in gastroenterology at a VA Medical Center should probably be broadened.

Analysis of patients referred to a gastroenterologist practicing in a community hospital.

A questionnaire was designed to provide information on consecutive patient referrals over a 12-mo period to a consultant in gastroenterology practicing in a community hospital serving approximately 100,000 people. Data were entered into a computer for ease of storage and analysis. One thousand two hundred seventy-six consultations were seen in 1188 patients. Ages ranged from 1 to 97 yr; most were in the 5th to 7th decade. The majority of consultations were from 11 general internists (66.2%); fewer consultations were from 7 general surgeons (16.0%), 8 general practitioners (10.3%), and 6 pediatricians (2.7%). The most frequent reasons for patient referral were abdominal pain (27.8%), consideration of a procedure (23.7%), and an x-ray abnormality (16.6%). The gastroenterologist performed 1217 procedures as part of 1026 consultations; the most frequently performed was upper gastrointestinal endoscopy (60.1%). The gastroenterologist made 1604 diagnoses in his subspecialty; diagnoses relating to the esophagus were made at 276 consultations, the stomach at 317, the duodenum at 204, the small bowel at 26, the colon-rectum at 321, the liver/biliary tree at 184, and the pancreas at 66. The major objective of this investigation was realized by the prospective collection of the experience of a practicing gastroenterologist. This data collection was accomplished without apparent omissions or encumbrance on the practice routine. These data provide a perspective on the role of a practicing consultant gastroenterologist and appear relevant to considerations of the training of gastroenterologists.

The use of [7alpha-3H]- and [7alpha, 7beta-3H]cholesterol in the enzymic assay of cholesterol 7alpha-hydroxylase.

A tritium release method is described for following the enzymic conversion of cholesterol to 7alpha-hydroxycholesterol. Incubations of rat liver subcellular preparations (containing microsomes) with [7alpha-3H]cholesterol or [7alpha,7beta-3H]cholesterol release the labeled hydrogen in the 7alpha position as 3H2O which, after counting, allows for the determination of the fraction of exogenous cholesterol converted to 7alpha-hydroxycholesterol. These findings document those recently reported by Van Cantfort, Renson, and Gielen (1975. Eur J. Biochem. 55:23). Analysis of incubation mixtures containing both [4-14C]cholesterol and either [7alpha-3H] or [7alpha,7beta-3H]cholesterol demonstrate that one atom of hydrogen (from the 7alpha position) is incorporated into H2O for every molecule of exogenous cholesterol that is converted to 7alpha-hydroxycholesterol. In the case of [7alpha-3H]cholesterol no label is retained by the product. With [7alpha,7beta-3H]cholesterol, one atom is released as 3H2O and one is retained by the product in the 7beta position. Microsomal incubations with [7alpha,7beta-3H]cholesterol were performed, followed by the acetylation of the steroid fractions with [14C]acetic anhydride. If intermixing of exogenous with endogenous cholesterol were complete during the enzymic reaction, one would expect the 3H: 14C ratio of the isolated cholesterol acetate to be four times that observed in the 7alpha-acetoxycholesterol acetate. Average values of 4.23 in one series and 4.03 in a second series indicate that intermixing was sufficiently complete to use the tritium release method as an indicator of mass conversion.