We electively compared the distal splenorenal ("selective") shunt with the end-to-side portacaval shunt in 80 prospectively randomized patients with variceal bleeding. Selective shunts required more operative time (3.9 vs. 2.8 h) and blood replacement (4.6 vs. 2.5 U) and postoperative mortality was slightly higher (5 of 38 selective vs. 0 of 40 portacaval). Postoperative complication rates were similar. After 65-mo mean follow-up, both shunts have protected well against late gastrointestinal bleeding (5 selective, 4 portacaval episodes). However, after selective shunts, spontaneous encephalopathy occurred less often (23% vs. 40% of patients), was severe in fewer patients (12% vs. 33%), and precipitated fewer hospital admissions (6 admissions in 4 selective patients vs. 26 admissions in 13 portacaval patients). Furthermore, selective shunt patients remained longer without functional disability (83% vs. 70% of postoperative patient months). Long-term survival was not significantly different in the two groups (5-yr survival: selective 51%, portacaval 56%).

Although penicillamine is used in the treatment of primary biliary cirrhosis, its mechanism of action in this disease is unknown. As an immunologic action had been attributed to the drug, we investigated whether penicillamine might alter serum immunoglobulin levels or immune complex-reactive material in patients with primary biliary cirrhosis. Immunoglobulin levels and immune complex reactivity were measured and clinical tests were performed in 53 consecutive patients entering a double-blind randomized trial of 750 mg vs. 250 mg of penicillamine. Measurement of immune complex reactivity was determined by laser nephelometry, 125I-C1q binding, and Raji cell assays. Immune complex reactivity was detected by at least one assay in 75% of patients tested before treatment. Sixty-two percent were positive in the C1q assay, 28% in the Raji cell assay, and 39% by nephelometry. After therapy with either dose, we found no change in immune complex-reactive material by any assay. Concentrations of immunoglobulins G and M fell (p less than 0.05) after 12 mo of therapy. Concentrations of immunoglobulin A decreased (p less than 0.05) only in the high-dose group. Correlation was not consistent between results of immune complex assays and clinical liver tests. Although immunoglobulin levels fell during penicillamine therapy, no decrease in immune complex-reactive material was detected. The effect of penicillamine in primary biliary cirrhosis is not mediated through alteration of immune complex-reactive material.

The purpose of this study was to investigate the effects of cimetidine and ranitidine on human lower esophageal sphincter pressure (LESP) and plasma levels of gastrin in all phases of the interdigestive motor complex and after a test meal. In a random, double-blind manner, placebo, cimetidine (1.0 mg/kg X h), and ranitidine (0.16 mg/kg X h) were administered by intravenous infusion to nine healthy volunteers. By using a sleeve catheter assembly, LESP was constantly monitored, as were esophageal, fundic, antral, and duodenal pressures. Considerable minute-to-minute and interdigestive motor phase-related LESP variations were observed. Cimetidine and ranitidine decreased the interdigestive LESP, but did not abolish the gradual increase in LESP from phase I to phase III. During the first 2 h after the meal, cimetidine and ranitidine had no significant effect on LESP. Plasma gastrin levels were increased by cimetidine and ranitidine, both in the interdigestive and in the postprandial state. The results indicate that the effect of H2-blockers on LESP is not gastrin-mediated. The results further indicate that, in studies on the effects of drugs on LESP, prolonged recording of LESP in all motor states is a prerequisite.

Prostaglandins protect against aspirin-induced damage to the gastrointestinal tract. This study tested the ability of enprostil, a synthetic analog of prostaglandin E2, given concurrently to prevent gastroduodenal injury. Twenty-four healthy subjects were randomly assigned to one of three groups. All received aspirin 650 mg q.i.d. for 5 days. One group received placebo and the other groups were given either 7 or 70 micrograms of enprostil b.i.d. for 5 days. Upper endoscopy was performed at entry and 2 h after the final dose of aspirin. Enprostil 70 micrograms b.i.d. afforded significant protection of both the antral and duodenal mucosa. The 7-micrograms dose protected only the antral mucosa. Side effects were not observed with the lower dose of enprostil. Serum salicylate levels did not differ significantly between the groups.

Pharmacodynamic and pharmacokinetic studies of omeprazole, a new gastric antisecretory agent, were undertaken in 8 healthy subjects. The drug was administered orally as an encapsulated enteric-coated granulate (40 mg daily at 9 am or 9 pm for 5 days), and its effect on the integrated 24-h gastric pH was determined, together with its apparent bioavailability. The pretreatment 24-h median pH was 1.9 (interquartile range 1.4-2.9). After 5 days of treatment, the median pH had risen to 5.0 (3.7-6.0) (p less than 0.01) with morning dosage and 4.5 (3.0-5.6) (p less than 0.01) with evening dosage. This corresponded to a greater than 99% reduction in 24-h median hydrogen ion activity, with morning dosage having a greater effect (from 9 am to 8 pm) (p less than 0.01) than evening dosage. The relative bioavailability of omeprazole increased twofold from day 1 to day 5 of treatment with morning dosage (p less than 0.02) and threefold with evening dosage (p less than 0.02), suggesting that increased absorption of this acid-labile drug occurs with increasing inhibition of acid secretion. We conclude that this formulation of omeprazole presently being used in clinical trials is a highly potent antisecretory agent in humans, although its optimal effect may not be observed for several days.

Our aim was to construct an index that accurately predicts the degree of benefit or harm that prednisone therapy holds for patients with liver cirrhosis. The admission and survival data of 488 patients with cirrhosis who participated in a controlled clinical trial of prednisone in a dosage of 10-15 mg daily (251 patients) versus placebo (237 patients) and who were observed for up to 12 yr were analyzed using Cox's multiple regression model. Four variables each provided significant therapeutic information: antinuclear factor (p = 0.02) and large piecemeal necroses (p = 0.02) were associated with a beneficial effect, whereas ascites (p = 0.0004) and large regenerative nodules (p = 0.0007) were associated with a harmful effect of prednisone. From these four variables a therapeutic index was constructed. For a given patient the therapeutic index is a measure of how big the effect will be if prednisone is given. The gain in survival time obtained by administering prednisone according to the therapeutic index was estimated to be 349 yr, mainly confined to 217 patients with a significant positive (121) or negative (96) therapeutic index. The therapeutic index may prove useful for the optimal administration of prednisone treatment in new patients with cirrhosis.

Primary sclerosing cholangitis and primary biliary cirrhosis are chronic cholestatic syndromes that may be difficult to differentiate clinically. Destructive cholangitis occurs in both diseases and leads to similar clinical and biochemical abnormalities. Therefore, we compared the clinical, biochemical, immunologic, radiologic, and hepatic histologic features of these syndromes in two large groups of patients prospectively selected by predefined criteria. Primary biliary cirrhosis (n = 258) occurred predominantly in middle-aged women who were usually symptomatic with fatigue and pruritus, commonly had keratoconjunctivitis sicca, and often were hyperpigmented. Tests for antimitochondrial antibodies were always positive, usually in very high titer. Although the extrahepatic bile ducts were normal radiographically, smooth tapering and narrowing of the intrahepatic bile ducts was occasionally noted. Hepatic histology was diagnostic when a florid duct lesion was present. In contrast, primary sclerosing cholangitis (n = 60) occurred primarily in young men who were usually symptomatic with fatigue and pruritus and frequently had chronic ulcerative colitis. Tests for antimitochondrial antibodies were nearly always negative and cholangiography demonstrated abnormalities of the extrahepatic and intrahepatic bile ducts in all cases. Although hepatic histology was often compatible with the diagnosis, it was usually not diagnostic, and considerable overlap existed with the abnormalities seen in primary biliary cirrhosis. Likewise, biochemical tests of copper metabolism were similar in both syndromes. These results call attention to the differences and similarities in the clinicopathologic features of these two cholestatic syndromes and provide a basis for a rational diagnostic strategy.

This study was conducted to determine whether taurocholate alters human gastric function and structure at a neutral pH, when ionized, and to contrast this with its effect at an acid intraluminal pH, when pronated. Five fasted healthy subjects were studied on 4 days in random order. Net ion fluxes, mucosal damage (as quantitated endoscopically), and potential difference were measured. The control solution instilled into the stomach in the first, third, and fourth 15-min periods contained 200 ml of 100 mM HCl, 54 mM mannitol, and [14C]polyethylene glycol. Taurocholate (10 mM) was added to the control solution (pH 1.1) or citrate buffer (pH 7.0) during the second 15-min period. The effect of citrate buffer alone or control solution alone was also tested. Because hydrogen and sodium fluxes could not be quantitated at pH 7 in the presence of citrate buffer, the net ion fluxes during the 15 min immediately after exposure to the test agent were measured. At both pH 1.1 and 7.0 taurocholate produced similar and significant increases in net hydrogen ion flux (-1.7 +/- 0.4 and -1.8 +/- 0.3 mmol/15 min, respectively), net sodium ion flux (1.8 +/- 0.4 and 1.7 +/- 0.2 mmol/15 min, respectively), decreases in potential difference, and mucosal erosions. The net hydrogen ion fluxes were significantly greater than occurred after citrate buffer alone or the HCl control. The net sodium fluxes after taurocholate in citrate were significantly greater than the pH 1.1 acid control, but not citrate buffer alone. These findings indicate that pronated (pH 1.1) or ionized (pH 7.0) taurocholate significantly damaged the in vivo human gastric mucosa. Taurocholate at pH 7 could in part be responsible for the gastric mucosal injury that occurs in patients with bile reflux gastritis.

To determine the frequency and prognosis of histologic cirrhosis developing during or after corticosteroid therapy of hepatitis B surface antigen-negative chronic active hepatitis, we followed 83 patients for 90 +/- 5 mo after administration of corticosteroids. Thirty-three patients satisfied histologic criteria for cirrhosis after 30 +/- 5 mo. In 25 patients, cirrhosis developed during treatment; in 8 patients, cirrhosis eventuated after remission and cessation of therapy. The probability of developing histologic findings of cirrhosis was 59% if remission had not been achieved after 3 yr of continuous therapy. Longer requirements for treatment and deterioration during therapy characterized these patients. Once remission was achieved, the mean annual incidence of cirrhosis was only 2.6%. Patients who manifested evidence of cirrhosis in their biopsy specimens could not be distinguished by initial clinical, biochemical, or histologic findings. Ascites, encephalopathy, and esophageal varices developed infrequently; 5-yr survival after documentation of cirrhosis was 93%. We conclude that histologic features of cirrhosis develop commonly during therapy, especially if remission is not achieved quickly. After remission, cirrhosis develops infrequently. The development of histologic cirrhosis does not influence immediate morbidity and 5-yr life expectancy.

The addition of microbial beta-galactosidases directly to milk at mealtime represents a potential "enzyme replacement therapy" for primary lactase deficiency. We used the hydrogen breath test as the index of incomplete carbohydrate absorption to assess the efficacy of two enzymes--one from yeast, Kluyveromyces lactis (LactAid), and the other from the fungus Aspergillus niger (Lactase N)--to assist in the hydrolysis of 18 g of lactose in 360 ml (12 oz) of whole milk when consumed by an adult lactose malabsorber. Graded amounts of Lactase N produced, at best, a 53% relative reduction in breath hydrogen excretion, whereas quantitative elimination of excess hydrogen excretion was produced by 1 and 1.5 g of LactAid. A double-blind, controlled, crossover trial was subsequently performed in 50 healthy, unselected Mexican adults, to whom 360 ml of cow's milk was presented in the three forms in a randomized order: intact milk, prehydrolyzed milk, and milk to which 1 g of LactAid was added immediately before consumption. Among the 25 subjects with incomplete carbohydrate absorption with intact milk, adding enzyme 5-min before consumption produced a 62% reduction in breath hydrogen excretion, and symptoms of intolerance were significantly reduced. The feasibility of effective enzyme replacement therapy with a beta-galactosidase from K. lactis is demonstrated.

Labyrinthine stimulation and cold pain inhibit feeding antral pressure activity, delay gastric emptying, and increase blood concentrations of beta-endorphin and norepinephrine. Further, labyrinthine stimulation induces, in approximately one-third of healthy individuals, a migrating burst of motor activity in the proximal intestine that interrupts the normal fed pattern. Our hypothesis was that endogenous opiates and catecholamines act as mediators of such disruptive effects of centrally acting stressful stimuli on gut motility. Thus, we studied feeding gastrointestinal pressure activity in healthy volunteers who were exposed to labyrinthine stimulation or cold pressure test, or both (both stimuli being either in their active or in their control forms), while receiving an intravenous infusion of either placebo (saline), or an opioid blocker (naloxone), or a combination of alpha- and beta-adrenergic blockers (phentolamine and propranolol), or all the drugs together. Neither opioid nor adrenergic blockers affected motility during control stimulations. Active stressful stimuli (labyrinthine stimulation, cold pain, or both) significantly inhibited antral feeding activity (p less than 0.05), but these effects were prevented by concomitant infusion of naloxone (p less than 0.05). Adrenergic blockade also prevented the antral motor inhibition caused by stress (p less than 0.05), but it was more effective for cold pain than for labyrinthine stimulation, and, when performed concomitantly with opiate blockade, the preventive effects disappeared. Furthermore, during adrenergic blockade labyrinthine stimulation invariably induced the appearance of a migrating duodenal burst of motor activity. Neither opioid nor adrenergic blockers modified the stress-induced rise of plasma beta-endorphin and norepinephrine. Our results suggest that opioids and catecholamines are involved in the mediation of the disruptive effects induced by centrally acting stressful stimuli on postprandial motor activity in the proximal human gut.

An efficacy trial of hepatitis B immune globulin in the prevention of perinatal transmission of hepatitis B virus infection was conducted in Los Angeles. Ten infants born to hepatitis B surface antigen (HBsAg)-and hepatitis B e antigen (HBeAg)-positive carrier mothers as well as 4 infants born to mothers with acute hepatitis B in the third trimester of pregnancy received hepatitis B immune globulin without randomization. Twenty infants born to HBsAg- and antibody to hepatitis B e antigen (anti-HBe)-positive carrier mothers were randomized to receive hepatitis B immune globulin or a placebo. All infants were followed for up to 18 mo. Four infants (2 born to HBeAg carrier mothers, 1 to an anti-HBe carrier, and 1 to a mother with acute hepatitis) became HBsAg-positive carriers, whereas another infant, whose mother was an HBeAg carrier, developed a transient anicteric hepatitis B infection. All infants who became infected did so after 9 mo of age as hepatitis B immune globulin protection waned. In infants born to HBsAg-positive carrier mothers, active immunoprophylaxis with the hepatitis B vaccine must be used in conjunction with hepatitis B immune globulin.

Measurements of biliary lipid secretion rates were performed in 10 patients with radiolucent gallstones before and after 4 wk of administration of chenodeoxycholic acid and ursodeoxycholic acid (1 g/day) in a randomized crossover study. The results of both bile acid feeding periods were similar in many respects: expansion of the bile acid pool, increase in bile acid and phospholipid secretion, reduction in cholesterol output, and decrease in percent saturation of hepatic bile, which was more pronounced with ursodeoxycholic than chenodeoxycholic acid therapy. Despite these similarities, the mechanisms by which these two litholytic bile acids induced these changes were quite different. Ursodeoxycholic acid, in contrast to chenodeoxycholic acid, only partially suppresses bile acid synthesis. During chenodeoxycholic acid feeding, the ratio of phospholipids to bile acids increased from 0.264 to 0.307 (p less than 0.05), indicating an increased coupling of phospholipids by chenodeoxycholic acid, whereas ursodeoxycholic acid did not alter this ratio. The molar ratio of cholesterol to bile acid during the chenodeoxycholic- and ursodeoxycholic-acid periods decreased significantly from 0.073 to 0.058 and 0.041, respectively. However, this ratio during the ursodeoxycholic-acid period was unchanged when the amount of ursodeoxycholic acid was subtracted from total bile acid (0.069), indicating that UDCA has little, if any, effect on the mobilization of hepatic cholesterol into bile.

We studied 20 critically ill patients receiving ventilatory support to determine both their metabolic requirements and the effect of providing energy substrate regimens containing different lipid to glucose calorie ratios on whole-body protein economy. The measurements used included indirect calorimetry, substrate hormone profile, and whole body protein turnover by [14C]leucine. Measurements were done while patients were receiving all their nonprotein calories as dextrose ( D100 ) and were compared with results obtained when they received all their nonprotein calories as a combination of dextrose and lipid in a calorie ratio either of 3:1 ( D75 ) or 1:3 ( D25 ). To maintain euglycemia, exogenous insulin was infused by attending physicians not cognizant of the total parenteral nutrition regimen used. Energy expenditure before receiving total parenteral nutrition was only 4.6% above basal, and did not rise significantly during any of the regimens. The insulin infusion rate, plasma insulin, CO2 production, and serum lactate were significantly higher with D100 than with D25 , but not with D75 . Correspondingly, plasma free fatty acids were significantly lower with D100 when compared with D25 but not with D75 . Despite this, there were no significant differences in whole-body protein synthesis, breakdown, or net synthesis (synthesis - breakdown) and, in general, the patients in all groups were close to zero protein balance. These data suggest that critically ill patients are not severely hypermetabolic, and that they can maintain protein balance with a modest excess of calories while using a wide range of fuel mixtures.

Aspirin and ethanol damage the gastric mucosa in humans, whereas acetaminophen does not. Acetaminophen increases prostacyclin activity in animals and thus may increase endogenous prostaglandin synthesis. The effect of acetaminophen was examined in five healthy subjects after intragastric aspirin or ethanol. Hydrogen and sodium ion fluxes were measured in the pylorus-occluded stomach that prevents duodenogastric reflux and gastric fluid losses. Studies were in random order and on 8 separate days. Potential difference was measured throughout and mucosal erosions were endoscopically quantitated (endoscopist masked, no premedication). The isoosmolar test solution (200 ml, 100 mM HC1, 54 mM mannitol, [14C]polyethylene glycol) was instilled into the stomach and removed 15 min later for four 15-min periods. Either aspirin (1300 mg) or ethanol (20% vol/vol) was added to the test solution only during the second 15 min. Acetaminophen (2600 mg) was given orally 1 h before aspirin or ethanol administration. To determine if the effect of acetaminophen was related to prostaglandin synthesis, prostaglandin production was inhibited with indomethacin (50 mg orally 12 h and 1 h before acetaminophen). Aspirin and ethanol alone each produced significant changes in net hydrogen and sodium fluxes, potential difference, and endoscopic changes. Acetaminophen significantly inhibited hydrogen ion flux, change in potential difference, and endoscopic damage; however, it did not totally abolish these changes. The protective effect of acetaminophen was abolished by pretreatment with indomethacin, suggesting that the effect of acetaminophen is likely to be prostaglandin-mediated. Indomethacin alone was without effect. These results demonstrate that oral acetaminophen protects the human gastric mucosa against the damaging effects of aspirin and ethanol.

The deleterious effects of aspirin on gastric mucosa have been well documented in experimental and clinical studies. Prostaglandins offer a potential method by which this injury may be prevented. In these studies, we developed a single-dose endoscopic assay system of aspirin-induced gastric mucosal injury in normal volunteers. With this system, 27 of 30 volunteers (90%) demonstrated severe mucosal injury after ingestion of aspirin. Subsequently, we evaluated whether pretreatment with 15-R-15 methyl prostaglandin E2 prevented severe injury after ingestion of aspirin. Following an initial dose-response study, a double-blind crossover trial was performed using pretreatment with placebo or with 10-micrograms doses of 15-R-15 methyl prostaglandin E2 for 24 h before treatment with aspirin. The results of this trial indicate that prostaglandin pretreatment significantly prevented the occurrence of endoscopically visible severe gastric mucosal injury after single-dose aspirin administration.

A multicenter double-blind study of the effectiveness of sulfasalazine and 6-methylprednisolone, alone and in combination, was conducted on 452 patients with Crohn's disease. One hundred sixty patients were previously untreated; 292 patients were previously treated. The Crohn's disease activity index (CDAI) was used to determine whether a patient had active (CDAI greater than or equal to 150, n = 215) or quiescent disease (CDAI less than 150, n = 237). Treatment of active disease consisted of high-dose 6-methylprednisolone, 6-methylprednisolone combined with 3 g of sulfasalazine, 3 g of sulfasalazine alone, or placebo, and lasted 6 wk. Patients in remission received maintenance doses of one of these drug regimens for periods of up to 2 yr. One hundred ninety-two patients completed the 2-yr study period. Results were evaluated using life-table analysis and outcome ranking. These methods showed 6-methylprednisolone to be the most effective drug in overall comparison of all patients (p less than 0.001); in previously treated patients (p less than 0.001); and in subgroups: active disease (p less than 0.001), only small bowel disease (p less than 0.05), and both small bowel and colon disease (p less than 0.05). Combination of 6-methylprednisolone and sulfasalazine was the most effective regimen in previously untreated patients (p less than 0.05) and when disease was localized in the colon (p less than 0.001). Sulfasalazine alone was least effective in overall comparison of all patients (p less than 0.05) and in all strata. Drug treatment was of no significant benefit to patients with quiescent disease. Continuous administration of low doses of 6-methylprednisolone, or the combination regimen, was beneficial in patients who responded initially to treatment of active disease. The addition of sulfasalazine, however, offered no advantage.

We studied the effects of the calcium-channel blocker nifedipine on esophageal smooth muscle function in 10 normal volunteers. Lower esophageal sphincter pressure and relaxation and esophageal contraction amplitude, peristalsis, velocity, and duration after wet swallows were determined before and for 120 min after the sublingual/buccal administration of placebo and of nifedipine in doses of 10, 20, 30, and 40 mg. Blood samples for measurement of plasma nifedipine concentration were obtained at baseline and every 30 min during this 120-min period. Nifedipine led to decreases in sphincter pressure of 13.3%, 29.9%, 34.3%, and 35.1% as the dose was increased from 10 mg to 40 mg. These changes were significantly (p less than 0.05) different from baseline and placebo for the 20-, 30-, and 40-mg doses and were more sustained with the higher doses, lasting as long as 90 min. Contraction amplitude fell 5.3%, 5.9%, 13.5%, and 19.6% at the corresponding doses. These changes were significantly (p less than 0.05) different from baseline and placebo only for the 30- and 40-mg doses, with the effect lasting up to 60 min. Peak plasma nifedipine concentration ranged from 28.7 +/- 3.7 ng/ml (mean +/- SEM) after 10 mg to 138.7 +/- 43.7 ng/ml after 40 mg of the drug, and occurred at either the 30- or 60-min measurement. The mean percent of decrease in sphincter pressure and contraction amplitude in the esophageal body correlated (p less than 0.001) with plasma nifedipine levels. There were no changes in sphincter relaxation or in peristalsis, velocity, or duration of contraction with any dose of nifedipine. It is concluded that (a) nifedipine significantly decreases lower esophageal sphincter pressure and contraction amplitude in the body of the esophagus, (b) the effect on sphincter pressure requires a lower dose of nifedipine and is more marked than that on contraction amplitude, and (c) the effects on both sphincter pressure and contraction amplitude correlate with plasma nifedipine levels. Calcium-channel blockers such as nifedipine may have a role in the treatment of motility disorders of the lower esophageal sphincter or esophageal body, and further controlled clinical studies are indicated.