Background: Although gastric variceal (GV) bleeding is less common than esophageal variceal bleeding, the severity of GV bleeding is often greater with higher morbidity and mortality rates. Minimally invasive endovascular treatments such as balloon-occluded retrograde transvenous obliteration (BRTO) and transjugular intrahepatic portosystemic shunt (TIPS) have been used for the management of GVs with varying results, and individual and institutional differences exist in the use of BRTO and TIPS. We performed a systematic review and meta-analysis to compare the feasibility, efficacy, and safety of BRTO and TIPS for the treatment of GVs because of portal hypertension. Methods: Searches of PubMed, EMBASE, Google Scholar, and Cochrane Library databases were performed from inception through March 2019. Summary odds ratio (OR) with 95% confidence intervals (CI) was estimated for technical success, hemostasis rate, postprocedural complications, rebleeding rate, incidence of hepatic encephalopathy, and mortality rate at 1 year utilizing a random-effects model. Results: Seven studies with a total of 676 patients (BRTO: 462 and TIPS: 214) were included. There was no difference in pooled technical success rate (OR, 0.87; 95% CI, 0.28-2.73; P=0.81), hemostasis rate (OR, 2.74; 95% CI, 0.61-12.26; P=0.19), and postoperative procedure-related complications (OR, 1.95; 95% CI, 0.44-8.72; P=0.38). However, treatment with BRTO was associated with lower rates of postoperative rebleeding (OR, 0.30; 95% CI, 0.18-0.48; P<0.00001), postoperative encephalopathy (OR, 0.06; 95% CI, 0.02-0.15; P < 0.00001), and mortality at 1 year (OR, 0.43; 95% CI, 0.21-0.87; P=0.02). Conclusions: BRTO was associated with lower rates of rebleeding, postprocedure hepatic encephalopathy, and mortality at 1 year. BRTO should be considered first-line modality for the treatment of GVs because of portal hypertension.

BACKGROUND: Antibiotics are routinely used for diverticulitis irrespective of severity. Current practice guidelines favor against the use of antibiotics for acute uncomplicated diverticulitis. OBJECTIVE: We performed a systematic review and meta-analysis to examine the role of antibiotic use in an episode of uncomplicated diverticulitis. DATA SOURCES: PubMed/Medline, Embase, Scopus, and Cochrane were used. STUDY SELECTION: Eligible studies included those with patients with uncomplicated diverticulitis receiving any antibiotics compared with patients not receiving any antibiotics (or observed alone). MAIN OUTCOME MEASURES: Pooled odds rate of total complications, treatment failure, recurrent diverticulitis, readmission rate, sigmoid resection, mortality rate, and length of stay were measured. RESULTS: Of 1050 citations reviewed, 7 studies were eligible for the analysis. There were total of 2241 patients: 895 received antibiotics (mean age = 59.1 y; 38% men) and 1346 did not receive antibiotics (mean age = 59.4 y; 37% men). Antibiotics were later added in 2.7% patients who initially were observed off antibiotics. Length of hospital stay was not significantly different among either group (no antibiotics = 3.1 d vs antibiotics = 4.5 d; p = 0.20). Pooled rate of recurrent diverticulitis was not significantly different among both groups (pooled OR = 1.27 (95%, CI 0.90–1.79); p = 0.18). Rate of total complications (pooled OR = 1.99 (95% CI, 0.66–6.01); p = 0.22), treatment failure (pooled OR = 0.68 (95% CI, 0.42–1.09); p = 0.11), readmissions (pooled OR = 0.75 (95% CI, 0.44–1.30); p = 0.31). and patients who required sigmoid resection (pooled OR = 3.37 (95% CI, 0.65–17.34); p = 0.15) were not significantly different among patients who received antibiotics and those who did not. Mortality rates were 4 of 1310 (no-antibiotic group) versus 4 of 863 (antibiotic group). LIMITATIONS: Only 2 randomized controlled studies were available and there was high heterogeneity in existing data. CONCLUSIONS: This meta-analysis of current literature shows that patients with uncomplicated diverticulitis can be monitored off antibiotics.

No abstract

Background and study aims Sessile serrated lesion (SSL) detection rate has been variably reported and unlike adenoma detection rate (ADR) is not currently a quality indicator for screening colonoscopy. Composite detection rates of SSL in patients undergoing average risk screening colonoscopy are not available. Methods Electronic database search (Medline, Embase and Cochrane) was conducted for studies reporting detection rates of serrated polyps (SSL, Hyperplastic polyp, traditional serrated adenoma) among average risk subjects undergoing screening colonoscopy. Primary outcomes were pooled SDR (SSL detection rate) and proximal serrated polyp detection rate (PSPDR). Pooled proportion rates were calculated with 95 %CI with assessment of heterogeneity (I2). Publication bias, regression test and 95 %prediction interval were calculated. Results A total of 280,370 screening colonoscopies among average risk subjects that were eligible with 48.9 % males and an average age of 58.7 years (± 3.2). The pooled SDR was available from 16 studies: 2.5 % (1.8 %–3.4 %) with significant heterogeneity (I2 = 98.66 %) and the 95 % prediction interval ranging from 0.6 % to 9.89 %. When analysis was restricted to large (n > 1000) and prospective studies (n = 4), SDR was 2 % (1.1 %–3.3 %). Pooled PSPDR was 10 % (8.5 %–11.8 %; 12 studies). There was evidence of publication bias (P < 0.01). Conclusion Definitions of SSL have been varying over years and there exists significant heterogeneity in prevalence reporting of serrated polyps during screening colonoscopy. Prevalence rate of 2 % for SSL and 10 % for proximal serrated polyps could serve as targets while robust high-quality data is awaited to find a future benchmark showing reduction in colorectal cancer arising from serrated pathway.

SUMMARY Published studies have reported variable results on the association between duration of proton pump inhibitor (PPI) use and the risk of dementia. An extensive literature search was performed in PubMed, Embase, Google Scholar, and Cochrane for studies examining the risk of cognitive decline and dementia among PPI users versus non-PPI users in prospective studies. Retrospective database linkage studies, case reports, case series, editorials, uncontrolled cohort studies, cross-sectional studies, and review articles were excluded. Primary outcome was pooled hazard rate (HR) of any dementia among PPI users compared with non-PPI users. Secondary outcomes were pooled HR of Alzheimer’s dementia (AD) and risk with long-term PPI follow-up (more than 5 years) studies. Meta-analysis outcomes, heterogeneity (I2), and meta-regression (for the effect of covariates) were derived by statistical software R and Open meta-analyst. A total of six studies (one RCT and five prospective) with 308249 subjects, average age of 75.8 ± 5.2 years, and follow-up of 5 (range 1.5–11) years were included in the analysis. Pooled HR of any dementia was 1.16 (n = 6, 95% confidence interval (CI) = 0.86–1.47). Results remained unchanged when only studies with long-term PPI use (more than 5 years) were analyzed (n = 4, pooled HR 1.10, 95% CI 0.66–1.53). Finally, the pooled HR for AD was 1.06 (n = 3, 95% CI 0.70–1.41). There was substantial heterogeneity among inclusion studies (I2 = 93%). Meta-regression did not demonstrate a significant role of age at study start (P = 0.1) or duration of PPI use (P = 0.62) to incident dementia. The results of this systematic review and meta-analysis do not show a significant relationship between PPI use and dementia in prospective studies with at least a 5-year follow-up.