To assess the prevalence of and risk factors for low bone mineral density in inflammatory bowel disease (IBD), 152 IBD patients and 73 healthy controls were studied. Sixty seven patients had ulcerative colitis, 78 had Crohn's disease (52 of them (66.7%) had ileal disease), and seven had indeterminate colitis. Bone mineral density values (g/cm2) measured by dual energy x ray absorbtiometry at the spine (L2-L4), the femoral neck, Ward's triangle, and the trochanter were 1.177, 0.948, 0.850, and 0.838 in the patients and 1.228 (p = 0.034), 1.001 (p = 0.009), 0.889 (NS), and 0.888 (p = 0.012) in the control group, respectively. The type or extent of the disease or previous small bowel resection did not have any significant effect on the bone mineral density values. There was a weak, but statistically significant negative correlation between bone mineral density and the total lifetime corticosteroid dose (in the lumbar spine r = -0.164, p = 0.04, the femoral neck r = -0.185, p = 0.02, Ward's triangle r = -0.167, p = 0.04, and the trochanter r = -0.237, p = 0.003). The patients whose lifetime corticosteroid dose (prednisone/prednisolone) was more than 10 g had especially low bone mineral density (p < 0.05 compared with the groups with no or less than 5 g of corticosteroid). The patients who had never taken peroral corticosteroids did not have decreased bone mineral density. In conclusion, IBD patients have significantly lower bone mineral density values than healthy controls, but the difference is not so great as has been reported previously. Low bone mineral density values in these patients are related to high lifetime corticosteroid doses.

Omeprazole plus amoxicillin cures Helicobacter pylori infection. The hypothesis was tested that low acidity is a predictor of outcome. Fifty patients with relapsing or complicated, or both H pylori positive duodenal (n = 25) or gastric ulcer (n = 25) were randomly treated with either omeprazole 20 mg twice daily plus amoxicillin 1 g twice daily or with omeprazole 40 mg twice daily plus amoxicillin 1 g twice daily over two weeks. After one week of combined treatment, a 24 hour gastric pH measurement was performed in all patients. H pylori cure rate was 67%. Patients who later turned out to be cured had higher pH values during night time and after meals (p < 0.05). In an explorative analysis drug compliance, smoking, location of the ulcer (duodenum versus stomach), age, and grade of body gastritis were additional predictors of the outcome. Smoking (p = 0.006), compliance (p = 0.037), duodenal ulcer disease (p = 0.065), and young age (p = 0.021) were related to high acidity. In conclusion, the success of eradication treatment with omeprazole and amoxicillin in ulcer patients infected with H pylori depends on intragastric pH. Drug compliance, smoking habits, location of ulcer, age, and activity of body gastritis are other predictors and in part related to intragastric acidity.

Ursodeoxycholic acid (UDCA) leads to biochemical and clinical improvement in many patients with primary biliary cirrhosis (PBC); although, the response is variable. This study compared UDCA treated patients with complete normalisation of biochemical functions to those without such improvement. Of the 65 patients receiving UDCA, 12 (19%) showed normalisation of liver biochemical functions at two years. The remaining 53 patients showed a less complete response. Mean (SD) alkaline phosphatase and total serum bilirubin values were significantly lower at entry in the patients whose liver biochemistry tests normalised (912 (732) U/l v 1417 (1021) U/l, p = 0.003, and 0.7 (12.1 (5.2) mumol/l v 38.9 (48.5) mumol/l, p = 0.0002, respectively), and percentage of UDCA in biliary bile acid was higher (56.3 (9.5)% v 38.3 (21.1)%, p = 0.03). Patients with biochemically and histologically less severe disease, and greater enrichment of biliary bile with UDCA, are more likely to respond favourably to the drug. The main objective of continued study will be to find out if normal liver biochemical functions can retard disease progression. The association of greater UDCA enrichment with complete biochemical responses suggests that higher doses of UDCA should be evaluated.

Primary biliary cirrhosis (PBC) patients have an increased incidence of recurrent urinary tract infection compared with patients with other chronic liver diseases. The course of significant asymptomatic and symptomatic bacteriuria in women with PBC was evaluated: consecutive patients were screened for bacteriuria at their outpatient appointments. Bacteriuric patients who were asymptomatic (n = 21) were randomised to receive antimicrobial therapy (n = 11), or no therapy (n = 10). Bacteriuric patients who were symptomatic (n = 13) were treated. All were followed up by weekly dipslide examination of urine. The course of bacteriuria in the 13 symptomatic and 11 asymptomatic treated patients was similar in terms of the medium interval between successive infective episodes (three and four weeks respectively), the number of relapses (six and seven) and reinfections (14 and 18). Most untreated asymptomatic patients became abacteriuric spontaneously but became reinfected with a different organism during the study period. A separate group of 24 PBC patients with no previous bacteriologically proved urinary tract infection was followed weekly in a similar fashion: seven (29%) became bacteriuric for two to four weeks during a three month period. This study suggests that treatment of recurrent bacteriuric episodes in PBC patients does not alter the natural history of their infection. The long term implication of periodically infected urine in these patients is currently unknown.

The early clinical results are described of a real time, electromagnetic imaging system as an aid to colonoscopy. After gaining experience with the use of the system, one experienced endoscopist was randomised to perform consecutive colonoscopies either with (n = 29) or without (n = 26) the imager view. All procedures were recorded on computer disk and replayed for retrospective analysis. Total colonoscopy was achieved in all patients except one (imager view not available). Comparing intubation time and duration of loop formation per patient, there was no significant difference between the two study groups. The number of attempts taken to straighten the colonoscope pre patient, however, was less when the endoscopist was able to see the imager view, p = 0.03. Hand pressure was also more effective when the endoscopist and endoscopy assistant could see the imager display, p = 0.02. Preliminary experience suggests that real time, electronic imaging of colonoscopy is safe, effective, and will improve the accuracy of the procedure.

Ten patients with ampullary carcinoma, not suitable for surgery, were treated with endoscopic photodynamic therapy (PDT) to evaluate the feasibility and safety of treatment. Patients received 4 mg kg-1 of haematoporphyrin derivative intravenously. Two days later, a duodenoscopy was performed and red (630 nm) light delivered to the tumour at fixed energy densities of 50 J or 200 J cm-1 per application, depending on the type of optical fibre used. The tumours were treated by three or four light applications at each session. Treatment was repeated up to five times at intervals of three to six months. The sole complication of PDT was moderate skin photosensitivity, which occurred in three patients. Tumour size was assessed at four to eight weekly intervals. In the absence of macroscopic tumour, biopsy specimens were taken. In three patients with small tumours confined to the ampulla, remission was obtained for periods ranging from eight to 12 months. In a further four patients with small tumours bulk was greatly reduced. There was little response in three patients with extensive duodenal involvement. Therefore PDT for ampullary carcinoma is both feasible and safe, and with refinement may prove curative for small tumours.

The effect of Helicobacter pylori eradication on ulcer healing and the relapse rate were investigated in a multicentre trial of 239 gastric ulcer patients. Patients with H pylori positive gastric ulcer were randomly assigned to one of three groups: (A) 10 days' treatment with metronidazole and eight weeks' treatment with colloidal bismuth subcitrate (CBS) (84 patients); (B) 10 days' treatment with metronidazole placebo and eight weeks with CBS (73 patients); or (C) ranitidine (82 patients). At 12 weeks in 210 patients, gastric ulcer was present in three (9%) of 35 H pylori negative patients, and in 45 (26%) of 175 H pylori positive patients (p < 0.05). Results after one year of follow up were available for 205 patients. Between 12 and 52 weeks, two (7%) ulcer relapses occurred in 29 H pylori negative patients and in 60 (47%) of 128 H pylori positive patients (p < 0.001). After two weeks of open triple therapy (CBS 120 mg four times daily, amoxicillin 500 mg four times daily, and metronidazole 400 mg three times daily), given to the patients with ulcer relapse, only one (an NSAID user) of 55 successfully treated patients had an ulcer relapse during the one year follow up. Healing of gastric ulcer is rapid and recurrence is infrequent after successful H pylori eradication. H pylori eradication changes the natural history of the gastric ulcer disease.

The effects of octreotide on regional motor function in the human gut are unclear. In a randomised, blinded study the effects of octreotide (50 micrograms, subcutaneously, three times daily) and placebo on gastric, small bowel, and colonic transit, and colonic motility and tone were assessed in 12 healthy volunteers whose colon had been cleansed. Octreotide accelerated initial gastric emptying (p = 0.05), inhibited small bowel transit (p < 0.01), and reduced ileocolonic bolus transfers (p < 0.05). Colonic transit was unaltered by octreotide; the postprandial colonic tonic response was inhibited (p < 0.05 v placebo), whereas colonic phasic pressure activity was increased by octreotide (p < 0.05 v placebo). These data support the use of octreotide in diarrhoeal states but not in diseases that cause small bowel stasis and bacterial overgrowth. Simultaneous measurements of colonic transit, tone, and phasic contractility are valid in studying the effects of pharmacological changes and may be applicable to the study of the human colon in health and disease.

To test the hypothesis that Helicobacter pylori infection is associated with a decreased intragastric acidity during omeprazole therapy, ambulatory 24 hour dual point gastric pH recordings were performed in 18 H pylori positive and 14 H pylori negative subjects. There was a four to six week washout period between the two pH recordings made in each subject after one week courses of placebo or omeprazole, 20 mg daily. During placebo, median 24 hour pH values were not different in the corpus (H pylori positive = 1.5, negative = 1.4; p = 0.9) or antrum (H pylori positive = 1.3, negative = 1.2; p = 0.1). However, during omeprazole treatment, median 24 hour pH values were higher in H pylori positive subjects, both in the corpus (H pylori positive = 5.5, negative = 4.0; p = 0.001) and antrum (H pylori positive = 5.5, negative = 3.5; p = 0.0004). During placebo treatment, the only difference between the two groups was a higher later nocturnal pH in the antrum in the H pylori positive group. During omeprazole treatment, gastric pH was higher both in the corpus and in the antrum in the H pylori positive group for all periods, except for mealtime in the corpus. These data indicate that omeprazole produces a greater decrease in gastric acidity in subjects with H pylori infection than in those who are H pylori negative. It is not, however, known whether there is a causal relationship between H pylori infection and increased omeprazole efficacy.

Prostaglandin analogues of the E-series theoretically offer the ideal antiulcer drugs. Peptic ulcer healing with prostaglandin analogues is, however, no better than would be predicted from their ability to inhibit gastric acid secretion and they are less effective than histamine H2 receptor antagonists in preventing ulcer relapse. It could be that prostaglandin analogues inhibit gastric mucosal synthesis or release of endogenous eicosanoids, thereby abrogating their own effects. This study, therefore, examined how a single therapeutic dose (200 micrograms) of misoprostol, a synthetic analogue of prostaglandin E1, influences gastric mucosal release of endogenous prostaglandin E2 (PGE2), thromboxane B2 (TXB2), and chemotactic leukotriene B4 (LTB4) during basal conditions and in response to gastric luminal acidification (0.1 M HCl; 5 ml/min for 10 minutes). Nine healthy volunteers were studied in a single blind, cross over design. In each subject misoprostol or placebo was instilled in randomised order into the stomach, which was subsequently perfused with isotonic mannitol. Misoprostol significantly decreased basal as well as acid stimulated output of PGE2 and TXB2, without affecting output of LTB4. These data show that misoprostol inhibits gastric mucosal synthesis of prostanoids. Decreased concentrations, or even a changed profile, of native eicosanoids modulating the release of inflammatory mediators from immune cells might explain why prostaglandin analogues have a comparatively poor clinical performance in ulcer healing and prevention.

This study determined the optimal maintenance dose of omeprazole in reflux oesophagitis. One hundred and ninety three patients rendered asymptomatic and healed after four or eight weeks omeprazole were randomised double blind to 10 mg omeprazole once daily (n = 60 evaluable), 20 mg omeprazole once daily (n = 68), or placebo (n = 62) for one year or until symptomatic relapse. Each omeprazole regimen was superior to placebo in preventing both symptomatic relapse (life table analysis, p < 0.001) and endoscopically verified relapse (p < 0.001). At 12 months, the life table endoscopic remission rates (proportions of patients without grade > or = 2 oesophagitis) were: 50% (95% confidence intervals 34 to 66%) with 10 mg omeprazole once daily, 74% (62 to 86%) with 20 mg omeprazole once daily, and 14% (2 to 26%) with placebo. At 12 months, the life table symptomatic remission rates (proportions of patients asymptomatic or with mild symptoms) were: 77% (64 to 89%) with 10 mg omeprazole once daily, 83% (73 to 93%) with 20 mg omeprazole once daily, and 34% (16 to 52%) with placebo. Both 10 mg and 20 mg omeprazole once daily were effective in prolonging the remission of reflux oesophagitis: 10 mg may be appropriate to start longterm treatment, though the existence of a dose response relation means that 20 mg once daily may be effective in patients for whom 10 mg once daily is suboptimal.

It has been suggested that pancreatic ductal hypertension, secondary to pancreatic outflow obstruction, is a cause of pain in chronic pancreatitis. This study investigated the effect of inhibiting pancreatic secretion with octreotide in chronic pancreatitis pain. Ten patients with chronic alcoholic pancreatitis and severe daily pain were included in an intraindividual double blind crossover study. All patients received octreotide (3 x 100 micrograms/day subcutaneously) and placebo (3 x 0.9% saline solution subcutaneously) for three days at random. Between both treatment phases a two day washout period was interposed. Intensity of pain (visual analogue scale) and analgesic consumption were carefully registered. Pancreatic secretion was monitored daily by measuring faecal chymotrypsin concentration. It was found that during the administration of octreotide, pancreatic secretion was strongly inhibited (faecal chymotrypsin mean (SD) 1.7 (0.6) U/g) with respect to placebo (9.6 (4.2) U/g) and washout (7.6 (3.1) U/g) periods (p < 0.001). Pain score (29.6 (4.5) v 28.7 (5.8)) and consumption of analgesics were no different during the octreotide and placebo periods. It is concluded that short term inhibition of pancreatic secretion does not result in pain relief in patients with chronic pancreatitis. This finding is in contrast with the hypothesis that outflow obstruction of pancreatic secretion with consequent ductal hypertension is an important cause of severe persistent pain in chronic pancreatitis.

The effects of hepatitis C virus genotype and viraemia on disease outcome in patients with chronic hepatitis C virus infection were studied. Patients infected with genotype 1 tended to develop more severe disease, and to respond less well to interferon (IFN) treatment, but no pretreatment variable successfully predicted either the severity of the disease or the response to IFN. Failure to eliminate the virus during the first three months of therapy, however, predicted a failure to derive long term benefit from the current IFN regime. Hence pretreatment variables cannot be used to determine whether individual patients will respond to IFN, but observations during the first three months of therapy can be used to decide which patients will not respond to prolonged therapy. In these patients consideration should be given to changing the IFN dosing regime or using alternative treatments.

Test meals are invariably used in the 13C-urea breath test (UBT) but their effect on the intragastric distribution and gastric residence time of urea given in the test is unknown. The site of Helicobacter pylori urease measured in the test is unknown and whether the test measures total or regional gastric urease is uncertain. This study reports the results of paired UBTs with simultaneous gastric distribution studies, one with and one without a fatty test meal, two weeks apart on seven H pylori infected subjects. The test meal did not affect UBT results at 10 minutes, but increased values at 30 minutes and thereafter. The amount of scintigraphic label in the antrum at 10 minutes was also unaffected by the meal but increased at 30 minutes and thereafter, whereas the amount in the body/fundus was greatly increased both at 10 minutes and throughout the test. There was considerable variation in intragastric distribution of urea between subjects, both with and without the test meal. This study shows that a test meal profoundly affects intragastric distribution of urea solution in the UBT, and increases UBT values at 30 minutes and later. Variability between subjects, however, means that accurate measurement of total or regional gastric urease is probably unrealistic.

This study compared the response to adenine arabinoside 5'-monophosphate (ARA AMP) in 60 patients with chronic hepatitis B according to the pretreatment serum hepatitis B virus DNA concentration. The level of hepatitis B virus replication was defined as low (30 patients) or high (30 patients) when serum hepatitis B virus DNA concentration was below or above 100 pg/ml, respectively. Patients received a 28 day course of ARA AMP and a second course of ARA AMP was given six months later to patients with persistent hepatitis B virus replication. At the end of the first course of ARA AMP, 11 of the patients (37%) with low replication and one of the patients (3%) with high replication became negative for hepatitis B virus DNA (p = 0.0012); five of the patients (17%) with low replication and none of the patients with high replication had HBe seroconversion (p = 0.06). Two of these five patients lost HBsAg. Kinetics of serum hepatitis B virus DNA during treatment showed a considerable but transient antiviral effect of ARA AMP. Three of 32 retreated patients became negative for hepatitis B virus DNA and one patient had HBe seroconversion. In conclusion, ARA AMP exerts a considerable but transient antiviral effect on hepatitis B virus. Complete and sustained inhibition of hepatitis B virus replication was only obtained in the patients with low hepatitis B virus replication.

This study evaluates the diagnostic accuracy of a faecal occult blood test and faecal alpha 1-antitrypsin in the investigation of patients with gastrointestinal symptoms or iron deficiency anaemia. One hundred and seventy nine patients with either iron deficiency anaemia (n = 67), changed bowel habit and aged > 39 years (n = 107), or a history suggestive of melaena (n = 5) provided faecal samples. After investigation, 32 patients had a diagnosis of possible gastrointestinal bleeding and 139 patients had no evidence of gastrointestinal bleeding. Eight patients had a cause of enteric protein loss in the absence of gastrointestinal bleeding and were excluded from subsequent analysis. The faecal alpha 1-antitrypsin test was diagnostically more accurate than the guaiac test in identifying probable gastrointestinal bleeding (82% and 72% respectively, p < 0.05). This faecal alpha 1-antitrypsin test was also more specific (83% and 72% respectively, p < 0.05), but was not significantly more sensitive (78% and 72% respectively). The sensitivity of these tests was insufficient to recommend their use for most patients in this study.

The first study of photodynamic therapy in the human gastrointestinal tract using 5 aminolaevulinic acid (ALA) induced protoporphyrin IX as the photosensitising agent is described. Eighteen patients with colorectal, duodenal, and oesophageal tumours were studied. After 30-60 mg/kg of ALA given orally, biopsy specimens of tumour and adjacent normal mucosa were taken 1-72 hours later. These specimens were examined by quantitative fluorescence microscopy for assessment of sensitisation with protoporphyrin IX. Ten patients were given a second dose of ALA a few weeks later and their tumours were treated with red laser light (628 nm). With 30 mg/kg ALA, the highest fluorescence values were detected in the duodenum and oesophagus, and the lowest in the large bowel. Doubling the ALA dose in patients with colorectal tumours gave protoporphyrin IX fluorescence intensities similar to those in patients with upper gastrointestinal lesions and improved the tumour:normal mucosa protoporphyrin IX sensitisation ratio. The treated patients showed superficial mucosal necrosis in the areas exposed to laser light. Six patients had transient rises in serum aspartate aminotransferases, two mild skin photosensitivity reactions, and five mild nausea and vomiting. In conclusion, photodynamic therapy with systemically administered ALA may be a promising technique for the treatment of small tumours and areas of dysplasia such as in Barrett's oesophagus.

Elemental diets are effective in inducing remission in active Crohn's disease, but how they exert this therapeutic effect is unclear. In a previous study a whole protein containing diet proved less effective than one in which food antigens were excluded, suggesting that exclusion of food antigens from the gut was a possible mechanism. This study was designed to test whether an oligopeptide diet of hydrolysed proteins was as effective as an amino acid based diet. These diets were equally antigen free but with different nitrogen sources. Forty four patients with active Crohn's disease were randomised in a controlled trial of amino acid versus oligopeptide diet. The feeds were given by nasogastric tube in equicaloric quantities and were the sole form of nutrition. Treatment was continued for four weeks although failure to improve by day 10 resulted in withdrawal. Quantitative leucocyte scintigraphy was used to investigate the effect of diet treatment on gut inflammation. Clinical and nutritional responses to treatment were also measured. Sixteen patients entered remission (including withdrawal of corticosteroids), six patients could not tolerate the nasogastric tube, and 22 patients failed to respond. The two diets were equally effective. Patients who responded had a rapid drop in clinical index of disease activity and a major reduction in the bowel uptake of leucocytes on scintigraphy. The oligopeptide and amino acid based enteral feeds were equally effective at inducing remission in active Crohn's disease. With both diets clinical improvement was accompanied by a reduction in intestinal inflammation.

Octreotide, an effective treatment for acromegaly, induces gall bladder stones in 13-60% of patients. Because knowledge of stone composition is essential for studies of their pathogenesis, treatment, and prevention, this was investigated by direct and indirect methods in 14 octreotide treated acromegalic patients with gall stones. Chemical analysis of gall stones retrieved at cholecystectomy from two patients, showed that they contained 71% and 87% cholesterol by weight. In the remaining 12 patients, localised computed tomography of the gall bladder showed that eight had stones with maximum attenuation scores of < 100 Hounsfield units (values of < 100 HU predict cholesterol rich, dissolvable stones). Gall bladder bile was obtained by ultrasound guided, fine needle puncture from six patients. All six patients had supersaturated bile (mean (SEM) cholesterol saturation index of 1.19 (0.08) (range 1.01-1.53)) and all had abnormally rapid cholesterol microcrystal nucleation times (< 4 days (range 1-4)), whilst in four, the bile contained cholesterol microcrystals immediately after sampling. Of the 12 patients considered for oral ursodeoxycholic acid (UDCA) treatment, two had a blocked cystic duct and were not started on UDCA while one was lost to follow up. After one year of treatment, five of the remaining nine patients showed either partial (n = 3) or complete (n = 2) gall stone dissolution, suggesting that their stones were cholesterol rich. This corresponds, by actuarial (life table) analysis, to a combined gall stone dissolution rate of 58.3 (15.9%). In conclusion, octreotide induced gall stones are generally small, multiple, and cholesterol rich although, in common with spontaneous gall stone disease, at presentation some patients will have a blocked cystic duct and some gall stones containing calcium.

Patients with liver cirrhosis are often undernourished. In healthy subjects, the pattern of food intake is one of the variables that can influence energy balance and substrate metabolism. The short term (two day) effect of the pattern of food intake in patients with cirrhosis and controls was compared. In a respiration chamber, eight patients with cirrhosis of the liver and 23 controls were fed to estimated energy balance in two meals daily ('gorging' pattern) and four to seven meals daily ('nibbling' pattern). Twenty four hour energy expenditure, expressed as a multiple of the sleeping metabolic rate, was reduced in patients with cirrhosis (1.31 (0.03) v 1.44 (0.02) for controls; p < 0.01) because of an increased sleeping metabolic rate per kg fat free mass in these patients. In both patients and controls, the respiratory quotient was significantly lower during the morning preprandial period (9.00-12.00) on the gorging pattern, reflecting a higher oxidation ratio of fat to carbohydrate compatible with a more catabolic state. For patients with cirrhosis, a nibbling pattern of food intake, including a good breakfast and a late evening meal, would be preferable, in order to have shorter episodes of catabolism during the day.