Effects of Early Placement of Transjugular Portosystemic Shunts in Patients With High-Risk Acute Variceal Bleeding: a Meta-analysis of Individual Patient Data

BACKGROUND & AIMS

Compared with drugs plus endoscopy, placement of transjugular portosystemic shunt within 72 hours of admission to the hospital (early or preventive transjugular intrahepatic portosystemic shunt [TIPS], also called preemptive TIPS) increases the proportion of high-risk patients with cirrhosis and acute variceal bleeding who survive for 1 year. However, the benefit of preemptive TIPS is less clear for patients with a Child-Pugh score of B and active bleeding (CP-B+AB). We performed an individual data meta-analysis to assess the efficacy of preemptive TIPS in these patients and identify factors associated with reduced survival of patients receiving preemptive TIPS.

METHODS

We searched publication databases for randomized controlled trials and observational studies comparing the effects of preemptive TIPS versus endoscopy plus nonselective beta-blockers in the specific population of high-risk patients with cirrhosis and acute variceal bleeding (CP-B+AB or Child-Pugh C, below 14 points), through December 31, 2019. We performed a meta-analysis of data from 7 studies (3 randomized controlled trials and 4 observational studies), comprising 1327 patients (310 received preemptive TIPS and 1017 received drugs plus endoscopy). We built adjusted models to evaluate risk using propensity score for baseline covariates. Multivariate Cox regression models were used to assess the factors associated with survival time. The primary endpoint was effects of preemptive TIPS versus drugs plus endoscopy on 1-year survival in the overall population as well as CP-B+AB and Child-Pugh C patients.

RESULTS

Overall, preemptive TIPS significantly increased the proportion of high-risk patients with cirrhosis and acute variceal bleeding who survived for 1 year, compared with drugs plus endoscopy (hazard ratio [HR] 0.443; 95% CI 0.323-0.607; P < .001). This effect was observed in CP-B+AB patients (HR 0.524; 95% CI 0.307-0.896; P = .018) and in patients with Child-Pugh C scores below 14 points (HR 0.374; 95% CI 0.253-0.553; P < .001). Preemptive TIPS significantly improved control of bleeding and ascites without increasing risk of hepatic encephalopathy in Child-Pugh C and CP-B+AB patients, compared with drugs plus endoscopy. Cox analysis of patients who received preemptive TIPS showed that patients could be classified into 3 categories for risk of death, based on age, serum level of creatinine, and Child-Pugh score. In each of these risk categories, preemptive TIPS increased the proportion of patients who survived for 1 year, compared with drugs plus endoscopy.

CONCLUSIONS

In a meta-analysis of data from 1327 patients with cirrhosis, acute variceal bleeding, and Child-Pugh score between 10 and 13 points or CP-B+AB, preemptive TIPS increased the proportion who survived for 1 year, in both subgroups separately, compared with drugs plus endoscopy.

Progression of Gastric Mucosal Dysplasia of the Postgastrectomy Stomach

Gastric cancer, if diagnosed at the symptomatic stage, has a poor prognosis, with an overall 5 year survival of about 5%. The surgical treatment of early gastric cancer increases this 5-year survival rate to 90%. In Japan, endoscopic surveillance has increased the proportion of gastric cancer detected at an early stage from 15% in 1960 to 50% in 1985, and the overall 5 year survival has been increased from 35% to 70%. Mass screening in Japan is worthwhile because the incidence of gastric cancer is about 80 cases per 100,000 population per annum (age standardized). But in other countries where the incidence is much lower the case for mass screening is weak and selective screening of those at high risk is advocated.

Progression of gastric mucosal dysplasia of the postgastrectomy stomach

Gastric cancer, if diagnosed at the symptomatic stage, has a poor prognosis, with an overall 5 year survival of about 5%. The surgical treatment of early gastric cancer increases this 5-year survival rate to 90%. In Japan, endoscopic surveillance has increased the proportion of gastric cancer detected at an early stage from 15% in 1960 to 50% in 1985, and the overall 5 year survival has been increased from 35% to 70%. Mass screening in Japan is worthwhile because the incidence of gastric cancer is about 80 cases per 100,000 population per annum (age standardized). But in other countries where the incidence is much lower the case for mass screening is weak and selective screening of those at high risk is advocated.

Transjugular intrahepatic portosystemic shunt versus endoscopic sclerotherapy for the prevention of variceal rebleeding after recent variceal hemorrhage

Variceal hemorrhage continues to be a major cause of morbidity and mortality in cirrhotic patients. Transjugular intrahepatic portosystemic shunt (TIPS) is gaining wide acceptance as a treatment for several complications of portal hypertension. The aim of the current randomized study was to compare the transjugular shunt and endoscopic sclerotherapy (ES) for the prevention of variceal rebleeding (VB) in cirrhotic patients. Forty-six consecutive cirrhotic patients with variceal bleeding were randomly allocated to receive either transjugular shunt (22 patients) or ES (24 patients) 24 hours after control of bleeding. VB (50% vs. 9%) and early (first 6 weeks) VB (33% vs. 5%) were significantly more frequent in sclerotherapy patients; the actuarial probability of being free of VB was higher in the shunt group (P <.002). Eight patients (33%) of the sclerotherapy group and 3 patients (15%) of the shunt group died; the actuarial probability of survival was higher for the shunted patients (P <.05); 6 patients in the sclerotherapy group and none in the shunt group died from VB (P <.05). No difference was found in the proportion of patients with clinically evident hepatic encephalopathy (HE). These results show that the transjugular shunt is more effective than sclerotherapy in the prevention of both early and long-term VB. Moreover, a significant improvement in survival was found in the shunt group.

Is there an irritable bladder in the irritable bowel syndrome?

In this prospective case controlled study 16 premenopausal women with documented irritable bowel syndrome were recruited from the gastroenterology clinic and 16 premenopausal controls without symptoms of irritable bowel syndrome were recruited from the gynaecology clinic. All women answered a standardised bowel and urinary symptom questionnaire and underwent twin channel subtracted cystometry. Women with irritable bowel syndrome also underwent oesophageal balloon distension studies for perception and pain. Oesophageal and bladder sensory thresholds were compared. Urinary frequency and urgency and the urodynamic finding of detrusor instability were significantly more common in women with irritable bowel syndrome (P < 0.05). We were unable to demonstrate a relationship between first sensation of bladder fullness and oesophageal perception or between maximum bladder capacity and oesophageal pain thresholds. These findings suggest that there is an irritable bladder in the irritable bowel syndrome and support the concept that irritable bowel syndrome is part of a generalised disorder of smooth muscle.

Pathogenesis of gastric metaplasia of the human duodenum: role of Helicobacter pylori, gastric acid, and ulceration

BACKGROUND & AIMS

The pathogenesis of gastric metaplasia (GM) in the duodenum is unclear. The aim of this investigation was to study the effect on the extent of GM of ulcer healing, Helicobacter pylori eradication, and acid suppression singly and in combination. The relationship between GM and gastroduodenal inflammation and H. pylori infection density was also studied.

METHODS

Duodenal and gastric antral biopsy specimens were obtained from H. pylori-positive patients with duodenal ulcer and from H. pylori-positive nonulcer subjects. Biopsy procedures from patients with duodenal ulcer were repeated after 6 months of treatment. Nonulcer subjects were treated symptomatically and did not undergo re-endoscopy.

RESULTS

Ulcer healing alone produced no change in GM or in gastroduodenal inflammation. H. pylori eradication produced a 42% reduction in GM and improvement in inflammation. Acid suppression produced a 43% reduction in GM but without a significant change in inflammation. A combination of H. pylori eradication and acid suppression produced an additive effect with a 66% reduction in GM. A positive relationship was detected between the extent of GM and antral H. pylori density, duodenitis score, and antral gastritis score.

CONCLUSIONS

This study shows that the extent of duodenal GM is unrelated to the presence or absence of ulceration but is partly due to H. pylori and partly due to acid.

Gastric DNA content in postgastrectomy patients. Relationship to mucosal dysplasia

BACKGROUND

Gastric mucosal cellular DNA content was assessed in patients who had undergone gastric surgery for peptic ulcer disease more than 20 years previously, with the aim of examining the relationship between abnormal DNA content and gastric mucosal dysplasia, as well as determining the effect of different types of surgery on DNA content.

METHODS

Sixty-five subjects underwent upper gastrointestinal endoscopy. In each, six biopsies were taken from the stoma or antrum and graded for severity of dysplasia. Cellular DNA was quantified using a microprocessor-controlled image analysis system with a fast densitometer card on Feulgen-stained slides. DNA histograms were evaluated using the 2c deviation index (2cDI) for proliferative activity and the 4c exceeding rate (4cER) and the 5c exceeding rate (5cER) as indices of malignant potential.

RESULTS

In subjects with Billroth II operations, all the above DNA criteria were higher than in Billroth I (P < 0.05), vagotomy and pyloroplasty (P < 0.001), and controls (P < 0.0001). DNA values increased as dysplasia progressed in severity (2cDI, Rs = 0.67; 4cER, Rs = 0.61; 5cER, Rs = 0.72; respectively, P < 0.0001). Among subjects with no dysplasia, more aneuploid cells were found in the Billroth II group, (p < 0.005) compared with the other types of operation.

CONCLUSIONS

Cellular DNA content is abnormal at an early stage in dysplasia and may even predate it. Increasing values of abnormal DNA content are related to the severity of dysplasia. DNA analysis may be a useful additional tool in surveillance programs to select high-risk patients for screening.

Expression of trefoil peptides pS2 and human spasmolytic polypeptide in gastric metaplasia at the margin of duodenal ulcers

Duodenal ulcers are associated with gastric metaplasia in the duodenum, both at the ulcer margin and at more distant sites in the duodenal bulb. pS2 and human spasmolytic polypeptide (hSP) are secretory peptides expressed in gastric epithelial cells and in gastric metaplasia. As these peptides may be important in ulcer healing, this study investigated the possibility that the expression of pS2 and hSP is increased in gastric metaplasia at the margin of duodenal ulcers. Duodenal bulb biopsy specimens from 12 duodenal ulcer patients were assessed. Sections were immunostained with monoclonal antibodies for pS2 and hSP. Cytoplasmic stain intensities were measured by an image analysis system and expressed as integrated optical density (IOD) units, In situ hybridisation for pS2 and hSP mRNA was carried out on parallel sections. Duodenal sections were also stained with diatase periodic acid Schiff/alcian blue to localise areas of gastric metaplasia. pS2 antigen staining in the duodenum was restricted to surface epithelial cells, and hSP to acinar and ductular components of Brunner's gland. mRNA localisation corresponded to immunostaining cells. In gastric metaplasia, pS2 expression was greater at the ulcer margin than away from the ulcer, as judged by the intensity of antibody staining (mean IOD units (SEM), 20.6 (3.3) v 9.5 (3.0); p < 0.001). There was a trend towards greater hSP staining at the ulcer margin but this did not achieve statistical significance. These findings support the putative role of pS2 and possible hSP in mucosal healing and providy further evidence for an autocrine 'ulcer-gastric metaplasia-repair' loop involving these trefoil peptides.

Serology for Helicobacter pylori compared with symptom questionnaires in screening before direct access endoscopy

This prospective study aimed to compare serology for Helicobacter pylori with two, symptom questionnaires in screening patients before direct access endoscopy. Methods were compared in terms of the number of endoscopies saved and pathology missed in 315 patients referred to a gastroenterology unit by 65 local GPs. The serology used was based on an acid glycine extract of H pylori. One in-house questionnaire was based on the Glasgow dyspepsia (GLADYS) system and the other questionnaire was that reported by Holdstock et al. A cut off point of 6.3 U/ml for H pylori serology was selected for screening patients (97% sensitive and 75% specific). Serology was combined with a history of NSAID usage in determining who should have endoscopy. For the in-house questionnaire, a cut off score of more than 8 out of a possible maximum of 18 was chosen, after prior evaluation in 118 patients referred for direct access endoscopy (the sensitivity for detection of peptic ulcer was 88%, specificity 61%). A cut off score of more than 412 was used for the Holdstock questionnaire. In patients under 45 years, serology detected more peptic ulcers than the in-house questionnaire and the Holdstock questionnaire (27/28 v 24/28, NS and v 20/28, p < 0.05 respectively). The Holdstock questionnaire saved significantly more endoscopies than the other two methods (76/149 v 57/149 for the in-house questionnaire, p = 0.05 and 59/149 for serology, p = 0.05). In all age groups combined, serology was significantly better than the in-house and Holdstock questionnaires at detecting peptic ulcers and gastric cancer (61/63, 52/63, p<0.02, and 50/63, p<0.01 respectively). But serology saved significantly fewer endoscopies (89/315, 135/315, p<0.005, and 119/315, p<0.05 respectively). Serology was inferior to the Holdstock questionnaire at detecting severe oesophagitis. It is concluded that serology is the method of choice in screening before direct access upper gastrointestinal endoscopy in those under 45 years. It best combines a high sensitivity for peptic ulcer disease with a large reduction in unnecessary negative endoscopies.

Effect of famotidine on oesophageal sensitivity in gastro-oesophageal reflux disease

Gastric acid suppression could improve heartburn by healing oesophagitis or by reduction of oesophageal sensitivity to acid. To independently assess changes in oesophageal sensitivity, it would be necessary to study patients with reflux disease but no oesophagitis. The aim of this study was to investigate the effect of acid suppression on oesophageal sensitivity and to assess the time course of any measured effect. Twenty seven patients were recruited, of whom 25 completed the study (14 men and 11 women, mean (SD) age 50 (15) years). All had classic symptoms of gastro-oesophageal reflux but normal results of upper gastrointestinal endoscopy and oesophageal mucosal histological tests. Each had abnormal 24 hour pH studies and a positive acid perfusion tests. Subjects were assigned double blind to placebo (n = 11) or famotidine 40 mg twice daily (n = 14) for four weeks. Acid perfusion tests were carried out at 0, 4, 5, and 8 weeks and time to heartburn recorded. Time to heartburn (mean (SEM)) was 124 (78) seconds in the famotidine and 187 (154) in the placebo group at week 0 (NS). Compared with baseline, significant increases in time to heartburn was found with famotidine at weeks 4 (383 (102), p < 0.01) and 5 (344 (92), p < 0.01) but not week 8 (336 (90) seconds). No significant effects were found with placebo (219 (41), 146 (23), and 144 (25) seconds for weeks 4, 5, and 8). Heartburn symptom score decreased significantly with famotidine (mean scores 3.6, 1.9, 2.1, and 2.6 at weeks 0, 4, 5, and 8 (p=0.001)) and showed a significant negative correlation with time to heartburn (r(s)=-0.60; p<0.0001). It is concluded that oesophageal sensitivity to acid is reduced by famotidine independent of and effect on oesophagitis; the effect wanes one to four weeks after the end of treatment and correlates with change in heartburn score.

Surface hydrophobicity of gastric mucosa in Helicobacter pylori infection: effect of clearance and eradication

Surface hydrophobicity of the gastric mucosa is reduced in peptic ulcer disease and Helicobacter pylori infection. This abnormality may be caused by H. pylori or may be an inherent defect. The aim of the present study was to clarify the relationship between H. pylori infection and mucosal hydrophobicity by examining the effect of eradication of the organism. H. pylori-positive patients with (n = 42) or without (n = 42) duodenal ulcer were randomized to receive ranitidine, bismuth, or bismuth plus antibiotics. Surface hydrophobicity of gastric mucosa was assessed by measurement of plateau-advancing contact angle. Measurements were performed at presentation, end of treatment, and 1 month later. Contact angle was unchanged after ranitidine (55 degrees vs. 56 degrees) but increased with bismuth (57 degrees-62 degrees; P < 0.05) and bismuth plus antibiotics (56 degrees-67 degrees; P < 0.0001). One month after treatment ended, contact angles in patients in whom H. pylori was not eradicated were not different from those before treatment (56 degrees vs. 56 degrees) but increased to a value similar to H. pylori-negative controls in patients in whom H. pylori was eradicated (56 degrees-69 degrees; P < 0.0001). It is concluded that reduced mucosal hydrophobicity in peptic ulcer disease is secondary to H. pylori infection and that this impaired mucosal defense provides a possible mechanism whereby H. pylori infection predisposes to acid/peptic digestion.

Determinants of pregnancy heartburn

OBJECTIVE

To study the prevalence and severity of reflux symptoms in pregnancy.

DESIGN

Self-administered questionnaire detailing age, race, gestational age, parity, weight, height, symptoms and severity of gastro-oesophageal reflux.

SETTING

An antenatal clinic in a teaching hospital.

SUBJECTS

607 consecutive women at various stages of pregnancy were recruited during attendance at antenatal clinic.

RESULTS

The prevalence of heartburn increased with gestational age (22% in the first, 39% in the second, and 72% in the third trimester; P < 0.0001), as did severity of heartburn (P < 0.0001). Pharyngeal regurgitation as a symptom showed little increase during advancing pregnancy. Symptoms of heartburn rapidly increased towards the end of the second trimester. The decrease in heartburn traditionally expected during the last three weeks of pregnancy (fetal head descent) was not observed. Logistic regression analysis showed increased risk of suffering heartburn with increasing gestational age (P < 0.0001), pre-pregnancy heartburn (P < 0.0001), parity (P < 0.0001) and inversely with maternal age (P < 0.05) but not with body mass index before pregnancy, race, or weight gain in pregnancy.

CONCLUSION

Heartburn, but less so pharyngeal regurgitation, increases in prevalence and severity with gestational age and parity. We speculate that hormone-related impairment of distal oesophageal clearance mechanisms may have a crucial bearing on whether heartburn develops in those individuals at greatest risk.

Thermodynamic effects of bile acids in the stomach

The surface thermodynamic effects of bile acids in the stomach were assessed in 48 subjects who had undergone gastric surgery for peptic ulcer disease and in 52 controls with medically healed ulcers. We derived values for surface tension of gastric mucosa from contact angle using a goniometer and measured the surface tension of gastric juice by the drop-weight method. Subjects with gastric surgery had higher median fasting bile acid concentrations than controls (1.2 vs. 0.1 mmol/L; P less than 0.0001), higher mean mucosal surface tension (51.9 vs. 47.9 mN/m; P less than 0.0001), and lower mean surface tension of gastric juice (43.2 vs. 51.7 mN/m; P less than 0.0001). Subjects who had had a Billroth II gastrectomy (n = 19) had higher bile acid concentrations (5.8 vs. 0.6 mmol/L; P less than 0.01), higher mucosal surface tension (53.7 vs. 50.3 mN/m; P less than 0.05), and lower gastric juice surface tension (41.3 vs. 47.1 mN/m; P less than 0.05) than those who had a vagotomy and drainage procedure (n = 17). Overall, intragastric bile acid concentration correlated directly with surface tension of gastric mucosa (r = 0.51, P less than 0.0001) and inversely with that of gastric juice (r = -0.60, P less than 0.0001). In conclusion, the interfacial energy barrier at the surface of the gastric mucosa is overcome in the presence of intragastric bile acids.

Surface hydrophobicity of gastric mucosa in peptic ulcer disease. Relationship to gastritis and Campylobacter pylori infection

The hydrophobicity of biopsy specimens of gastric mucosa in 228 dyspeptic subjects undergoing diagnostic endoscopy was assessed by measuring the plateau-advancing contact angle of saline drops using a goniometer. Subjects with duodenal ulcers (n = 49) and gastric ulcers (n = 17) had significantly lower mean contact angles than controls (n = 124) without ulcer (57 degrees in duodenal ulcer, 59 degrees in gastric ulcer vs. 66 degrees in controls; p less than 0.0001). There was no change in contact angle after healing with H2-receptor antagonists by comparison with pretreatment (59 degrees vs. 56 degrees for duodenal ulcer, n = 15; 57 degrees vs. 59 degrees for gastric ulcer, n = 5). Controls with gastritis had lower contact angles than those without (61 degrees, n = 50, vs. 70 degrees, n = 63; p less than 0.0001). The presence of Campylobacter pylori was associated with a significant decrease in contact angle in controls (59 degrees, n = 39, vs. 70 degrees, n = 75; p less than 0.0001).

Measurement of the surface hydrophobicity of human gastrointestinal mucosa

A technique has been developed for assessing the surface hydrophobicity of human gastrointestinal mucosa by measuring the plateau contact angle of saline drops applied to endoscopic biopsy specimens. The plateau contact angle was not affected by the mode of drying. The intraobserver and interobserver coefficient of variation was less than 5%. The gastric mucosal surface had a higher mean contact angle than the submucosal surface (69 degrees vs. 47 degrees, p less than 0.001). Glycerol drops gave lower contact angles than saline drops (55 degrees vs. 69 degrees) but gave the same derived values for surface free energy (42 vs. 41 mJ/m2). Regional values for contact angle were as follows: gastric body 70 degrees, antrum 70 degrees, duodenal bulb 62 degrees (p less than 0.01 vs. stomach), distal duodenum 50 degrees (p less than 0.001 vs. stomach and p less than 0.01 vs. bulb), and rectum 57 degrees (p less than 0.001 vs. stomach). We conclude that it is feasible to measure the surface hydrophobicity of human endoscopic biopsy specimens and that the stomach is relatively more hydrophobic than the duodenum and rectum.