Acid-suppressive therapy is associated with spontaneous bacterial peritonitis in cirrhotic patients: A meta-analysis

Transcription

1 bs_bs_banner doi: /jgh META-ANALYSYS AND SYSTEMATIC REVIEW Acid-suppressive therapy is associated with spontaneous bacterial peritonitis in cirrhotic patients: A meta-analysis Abhishek Deshpande,* 1 Vinay Pasupuleti,* 1 Priyaleela Thota, Chaitanya Pant, Sulaiman Mapara, Sohaib Hassan, David D K Rolston, Thomas J Sferra** and Adrian V Hernandez *Department of Medicine, Division of Infectious Diseases, Case Western Reserve University, Departments of Hospital Medicine, Internal Medicine and Quantitative Health Sciences, Cleveland Clinic, Department of Internal Medicine, St. Vincent Charity Medical Center, **Department of Pediatrics, Case Western Reserve University, Rainbow Babies and Children s Hospital, Cleveland, Ohio, Department of Pediatrics, Oklahoma University Health Sciences Center, Oklahoma City, Oklahoma, and Department of Internal Medicine, Geisinger Medical Center, Danville, Pennsylvania, USA Key words H2 receptor antagonist, meta-analysis, proton pump inhibitor, spontaneous bacterial peritonitis. Accepted for publication 24 October Correspondence Dr Abhishek Deshpande, Case Western Reserve University, Department of Medicine, Division of Infectious Diseases, Euclid Avenue, BRB 10 West, Cleveland, OH , USA. abhishekdp@gmail.com Conflicts of interest: Nothing to declare for all authors. Financial Support: None. 1 Contributed equally to this study. Author Contributions: Conception and design: AD, VP, CP, AVH Analysis and interpretation of data: AD, VP, AVH, TJS, DDKR Drafting of the article: AD, VP, PT, CP, SM, SH, DDKR, TJS, AVH Critical revision of the article for important intellectual content: AD, VP, PT, CP, SM, SH, DDKR, TJS, AVH Final approval of the article: AD, VP, PT, CP, SM, SH, DDKR, TJS, AVH Statistical expertise: AVH Collection and assembly of data: AD, VP, PT, CP, SM, SH, DDKR, TJS, AVH Abstract Background and Aim: Proton pump inhibitors (PPI) and H 2-receptor antagonists (H2RA) are frequently prescribed in hospitalized patients with cirrhosis. There are conflicting reports regarding the role of acid-suppressive therapy in predisposing hospitalized patients with cirrhosis to spontaneous bacterial peritonitis (SBP). The aim of this metaanalysis was to evaluate the association between acid-suppressive therapy and the risk of SBP in hospitalized patients with cirrhosis. Methods: We searched MEDLINE and four other databases for subject headings and text words related to SBP and acid-suppressive therapy. All observational studies that investigated the risk of SBP associated with PPI/H2RA therapy and utilized SBP as an endpoint were considered eligible. Data from the identified studies were combined by means of a random-effects model and odds ratios (ORs) were calculated. Results: Eight studies (n = 3815 patients) met inclusion criteria. The risk of hospitalized cirrhotic patients developing SBP increased when using acid-suppressive therapy. The risk was greater with PPI therapy (n = 3815; OR 3.15, 95% confidence interval ) as compared to those on H2RA therapy (n = 562; OR 1.71, 95% confidence interval ). Conclusions: Pharmacologic acid suppression was associated with a greater risk of SBP in hospitalized patients with cirrhosis. Cirrhotic patients receiving a PPI have approximately three times the risk of developing SBP compared with those not receiving this medication. Prospective studies may help clarify this relationship and shed light on the mechanism(s) by which acid-suppressive therapy increases the risk of SBP in hospitalized patients with cirrhosis. Introduction Spontaneous bacterial peritonitis (SBP) is a frequent complication in cirrhotic patients with ascites and is associated with significant morbidity and increased mortality. 1 3 In hospitalized patients with cirrhosis, SBP accounts for 10 30% of all reported bacterial infections. 1 3 While the mechanism of SBP is unknown, evidence suggests that patients with cirrhosis are predisposed to gastrointestinal bacterial overgrowth resulting from increased intestinal permeability, altered intestinal motility, and medication-induced Journal of Gastroenterology and Hepatology 28 (2013)

2 gastric acid suppression. 4 6 There is a higher prevalence of bacterial overgrowth and impaired small bowel motility in patients with SBP compared with patients who did not have SBP. 7 Bacterial colonization is also facilitated by a weakened immune system in patients with SBP. 8 Proton pump inhibitors (PPIs) and H 2-receptor antagonists (H2RAs) are the most frequently used drugs for acid suppression. However, suppression of gastric acid by these medications has been associated with several potential adverse effects, including other enteric infections (Clostridium difficile, Klebsiella spp., Salmonella spp.) 9,10 and both community-acquired and ventilatorassociated pneumonia. 11,12 Acid-suppressive therapy is also known to predispose to bacterial overgrowth within the gastrointestinal tract and translocation across the epithelial barrier. 7 While there can be many indications for acid-suppressive therapy, the most common indications are gastroesophageal reflux disease, peptic ulcer disease, Barrett s esophagus, Zollinger Ellison syndrome, and nonsteroidal anti-inflammatory drug use. 13 Acid-suppressive therapy is commonly prescribed to cirrhotic patients, although there is limited information available regarding the risks and benefits of this therapy in cirrhosis. At least two studies have reported a high percentage of patients with cirrhosis on acid-suppressive therapy without a documented indication. 6,13 A few observational studies with small sample sizes and one meta-analysis have evaluated PPI therapy and the risk of developing SBP in patients with cirrhosis. The results of the observational studies have been conflicting. The goal of this updated meta-analysis was to determine whether exposure to H2RA and PPI therapy are associated with the development of SBP in patients with cirrhosis and ascites. Methods All procedures used in this meta-analysis were consistent with Meta-analysis of Observational Studies in Epidemiology guidelines. 14 Data sources and searches. Two investigators (VP and AD) systematically searched the literature independently using the following predetermined inclusion criteria: (i) observational study, including case control, and cohort study evaluating the risk of SBP associated with acid-suppressive therapy; (ii) study population comprised adult patients ( 18 years); (iii) SBP (defined as 250 polymorphonuclear leukocytes in the ascitic fluid) was a study endpoint; (iv) hospital- or community-based study; and (v) date of publication between 1966 and 2012 in any language. Our exclusion criteria were (i) no control group of patients; (ii) acidsuppressive therapy usage data (type of therapy and who received the drug) was not available or could not be extracted; and (iii) data were presented based on SBP episodes and not on the number of actual patients. This search was performed in February The following databases were searched: MEDLINE (PubMed) ( ), Web of Science ( 2012), CINAHL ( 2012), and SCOPUS ( 2012). Search terms were SBP bacterial overgrowth, PPI acidsuppressive therapy, proton pump, gastric acid-suppressive agents, H2 receptor antagonists, H2RA, and H2 blockers. Reference lists from included studies and several previously published reviews on SBP and acid-suppressive therapy, and major gastroenterology meeting abstracts were also searched. The electronic search strategy of PubMed is available in supplemental material. Study selection. A list of retrieved articles that met the inclusion criteria was reviewed by two investigators independently (VP and AD). Any disagreement about a particular study was resolved in consensus with a third investigator (AVH). Where more than one article for a single study was found to have been published by the same investigators, we used the most relevant publication and supplemented it, if necessary, with data from the other publications. Data extraction and quality assessment. Two investigators (AD and VP) independently extracted data from the full text of the included studies. Data collected included study design, study population, patient demographics and clinical characteristics, SBP diagnostic criteria, type of acid-suppressive therapy, and dosage and duration of acid-suppressive therapy. Any disagreements or discrepancies were resolved in consensus with a third investigator (AVH). Authors were contacted if the relevant information was not available for a particular study. Quality assessment. The quality of the selected studies was assessed independently by two authors (AD and VP) using the Newcastle Ottawa scale (NOS). 15 The NOS uses two different tools for case-control and cohort studies and consists of three parameters of quality: selection, comparability, and exposure/ outcome assessment. The NOS assigns a maximum of four points for selection, two points for comparability, and three points for exposure or outcome. NOS scores of > 7 were considered as highquality studies, and NOS scores of 5 7 were considered moderate quality. Any discrepancies were addressed by a joint reevaluation of the original article. There was a lack of information on PPI and H2RA exposure for almost all studies, and therefore none of the studies were considered to be at low risk of bias. Also, a methodological quality checklist proposed by the Metaanalysis of Observational Studies in Epidemiology collaboration was used for each included study and verified by two investigators independently (AD and VP). 14 These criteria in the checklist included (i) clear identification of the study population; (ii) clear definition of outcome and outcome assessment; (iii) identification of important confounders and/or prognostic factors; (iv) independent assessment of outcome parameters (i.e. ascertainment of outcomes done by researchers other than the ones involved in the study); and (v) selective loss during follow-up. Each criterion was assessed as a yes or no. If a study did not clearly mention any of the previous criteria, it was assessed as a no. Data synthesis and analysis. The primary outcome is the association between acid-suppressive therapy and the risk of developing SBP. While there were several potential reasons for heterogeneity, we considered one a priori hypothesis to explain potential variability between studies and accordingly performed subgroup analysis based on the type of acid-suppressive therapy (i.e. PPI vs H2RA). Meta-analyses were not stratified by type of PPI or H2RA as they have similar efficacy and can be used interchangeably. 16, Journal of Gastroenterology and Hepatology 28 (2013)

3 DerSimonian and Laird random effects models were used for all meta-analyses. 18 The meta-analysis was performed using the inverse variance method for pooled odds ratios (OR) and 95% confidence intervals (CI). The inverse variance method converts the 95% CI to the standard error on a natural logarithmic scale and back, and the Forest plots may occasionally have rounded up values of 95% CI. 19 We assumed similarity between the odds ratio and other relative measures such as relative risk or rate ratios because SBP events and deaths were rare events. 20 We evaluated statistical heterogeneity using the Cochran chi-square (c 2 ) and the I 2 statistic. 21 I 2 values of 30 60% represented a moderate level of heterogeneity. 22 A P-value of < 0.1 for c 2 was defined as indicating the presence of heterogeneity. While every attempt was made to pool adjusted odds ratios from the primary studies, there were studies where raw data were available, and it used to calculate unadjusted odds ratios. We grouped the studies for the analysis into two subgroups: studies for which only the unadjusted odds ratios were available, with no correction for baseline differences or confounding, and those for which we were able to extract odds ratios adjusted for potential confounders. The number needed to harm (NNH) and its 95% CI was calculated for each study individually and also using the pooled OR from the meta-analysis to the average event rate of the largest study included. The NNH is the number of patients who need to be treated with PPI or H2RA for one additional person to have an adverse outcome. 22 Assessment of risk of bias. We followed the recommendations of the Cochrane Adverse Effects Methods Group and considered participant selection, ascertainment of acid-suppressive therapy exposure, and definition of adverse outcomes, that is, SBP, and statistical adjustment for confounders. 23 To check for publication bias, we generated funnel plots and used Egger s regression asymmetry test. Where asymmetry was detected, we assessed the potential impact of the publication bias using the Duval and Tweedie nonparametric trim-and-fill method. This method recalculates the effect size given the presence of publication bias. 24,25 We used Review Manager (RevMan, version 5.1 for Windows, Oxford, UK; The Cochrane Collaboration, 2008) and Comprehensive Meta-Analysis (Version 2.0, Biostat Inc., Englewood, NJ, USA) for our statistical analyses. Results Study characteristics. Our search identified 405 publications (Fig. 1). After removing duplicates and screening titles of the studies, 38 articles were selected based on its relevance to the study topic. After screening the abstracts of these potentially relevant articles, 22 were selected for full-text review based on relevance to study topic (Fig. 1). Eight articles that reported acid-suppressive therapy and SBP published between January 1990 and February 2012 were included in the meta-analysis. The reasons for exclusion of the remaining 14 articles are listed in Figure 1. Table 1 summarizes the main characteristics of the included studies. All of the included eight articles investigated PPI usage; four of these also investigated H2RA usage. No article studied Figure 1 Study selection process. Journal of Gastroenterology and Hepatology 28 (2013)

4 Table 1 General characteristics of included studies Study reference Study location Sample size Age (year) Male, Study design Adjusted for Type of acid suppression evaluated Year n Mean (SD) n (%) Campbell (2008) 26 USA (10.8) 78 (67.2) CC Age, bilirubin, INR, creatinine, MELD PPI and H2RA score, DM, gender, history of SBP, etiology of liver disease, race Northup (2008) 27 USA 2631 NA NA CC None PPI Bajaj (2009) 6 USA (13.0) 79 (56.4) CC CTP class, age, admission time period PPI and H2RA Bulsiewicz (2009) 28 USA (63.8) CC Antibiotic use, serum sodium, MELD, PPI HIV status Choi (2011) 29 Korea (10.7) 138 (78.4) CC CTP class, MELD score, history of PPI and H2RA esophageal variceal bleeding Goel (2011) 13 USA (11.1) 83 (63.9) CC CTP class PPI and H2RA de Vos (2011) 30 Belgium (60.0) CC None PPI van Vlerken (2012) 31 Netherlands (12.0) 56 (66.7) PC Age, CTP class PPI Antibiotic users excluded. Range. CC, case control; CTP, Child Turcotte Pugh class; DM, diabetes mellitus; H2RA, H2 receptor antagonists; HIV, human immunodeficiency virus; INR, international normalized ratio (prothrombin time); MELD, model for end-stage liver disease; PC, prospective cohort; PPI, proton pump inhibitors. H2RA alone. A total of 3815 patients were included in the metaanalysis. Seven studies were case-control, and one study was a prospective cohort study. Quality assessment. Using the NOS scale, two of the eight studies were identified as high quality, and four of the eight studies were identified as moderate quality studies. All studies clearly identified the study population and defined the outcome and outcome assessment. None of the studies had a selective loss of patients during follow-up. Six studies identified important confounders or prognostic factors and were used for adjustment of the association between acid-suppressive therapy usage and risk of SBP. There was considerable variation in the type of model and the selection of available confounding variables for adjustment. It is possible that a few confounding variables were not fully identified and recorded. The most common confounders adjusted were age, Child Turcotte Pugh class, comorbidities, and model for endstage liver disease. There was limited to no information on the type of PPI/H2RA, dose, and duration of usage before the diagnosis of SBP. Also, there was very little information of the dose and duration of PPI/H2RA after hospital admission. Confirmation of PPI use was variable, and it was obtained from medical records or patient pharmacy databases. Publication bias. As fewer than 10 studies were included in the meta-analysis, the results of the tests for publication bias should be interpreted with caution. SBP-PPI analysis. On visual inspection, the funnel plot appears asymmetric (Fig. 2). As per Egger s method (intercept =-1.87, standard error [SE] = 0.77, P = 0.05, two tailed), there was no evidence of asymmetry. Adjustment for publication bias according to Duval and Tweedie s trim-and-fill procedure resulted in a similar OR 3.15 (95% CI ) with no studies missing. SBP-H2RA analysis. On visual inspection, the funnel plot appears to be asymmetric (Fig. 2). As per Egger s method (intercept =-2.28, SE = 0.66, P = 0.08, two tailed), there was evidence of asymmetry. However, the adjustment for publication bias according to Duval and Tweedie s trim-and-fill procedure resulted in a similar OR 1.71 (95% CI ) with no studies missing. Meta-analyses. Meta-analysis of PPI usage data showed a significantly higher risk of SBP in patients on PPI compared with SBP patients with no PPI (OR 3.15, CI , P < ) (Fig. 3). There was moderate heterogeneity across included studies (I 2 = 57%). When only moderate-high quality studies (NOS score 5, n = 6) were analyzed separately, PPI therapy continued to be associated with a threefold increase in risk for SBP (OR: 2.89; 95% CI , P < ) with lower heterogeneity (I 2 = 33%). Meta-analysis of H2RA usage data showed a risk of SBP in patients on H2RA compared with SBP patients with no H2RA (OR 1.71, CI , P = 0.06) (Fig. 4), although the association was not statistically significant. There was low heterogeneity across included studies (I 2 = 0%). All H2RA associated studies were of moderate-high quality. Number needed to harm. Number needed to harm was calculated for each study individually (Supplemental Table). The table includes several statistically significant NNH figures for PPI (where the CI does not include infinity), ranging in strength from 3 to 6 (by convention, the NNH are always rounded up to the next higher whole number; later instances of rounding up are not specifically noted). If the rates for SBP are the same for both interventions (acid-suppressive therapy or no acid-suppressive therapy), the absolute risk increase is zero, and its reciprocal is represented as infinity ( ). This means that to see a difference in outcome between the two interventions, an infinite number of patients would have to be treated. 238 Journal of Gastroenterology and Hepatology 28 (2013)

5 Figure 2 Funnel plot to assess publication bias. Subgroups: ( ) Unadjusted, ( ) Adjusted. Figure 3 Forest plot of association between PPIs and SBP. SE, standard error; CI, confidence interval; df, degrees of freedom; OR, odds ratio. Among the included studies, in the largest cohort of patients with cirrhosis, the baseline incidence of SBP in non-ppi users was found to be 6.0%. 27 Using the summary OR ( % CI ) from the pooled studies, the use of PPI would result in an NNH of 9 (95% CI 6 to 17) for one episode of SBP. Discussion Our meta-analysis involving eight studies and 3815 cirrhotic patients found an association between SBP and acid-suppressive therapy. As fewer than 10 studies were included in the metaanalysis, the results of the meta-analysis should be interpreted with caution. Overall, cirrhotic patients receiving a PPI have approximately three times the risk of developing SBP compared with those not receiving this medication. However, heterogeneity between the studies implies that the summary OR should be interpreted with caution. Although not statistically significant, cirrhotic patients on H2RA also demonstrated a trend toward an elevated risk of developing SBP compared with non-users. It is therefore possible that the extent of acid suppression is an important mediator of SBP. Journal of Gastroenterology and Hepatology 28 (2013)

6 Figure 4 Forest plot of association between H2RA and SBP. SE, standard error; CI, confidence interval; df, degrees of freedom; OR, odds ratio. SBP is common in cirrhotic patients with significant morbidity and mortality. Gastric acid-suppressive medications can also affect the chances of developing SBP. PPI and H2RA are the most commonly prescribed class of drugs for gastric acid suppression. While a few retrospective case control studies reveal a potential association between the use of gastric acid-suppressive agents and the development of SBP, the association remains controversial with other conflicting results. Campbell et al. reported that PPI use was not associated with SBP, 26 but Gati et al. and Bajaj et al. have reported a significant association. 6,13 The differences between the studies could be because patients with SBP in the two latter studies had more severe liver functional impairment. Goel et al. 13 and Choi et al. 29 have reported that H2RA use was not associated with SBP, but Campbell et al. and Bajaj et al. have reported a significant association. 6,26 Variations in the bacterial strain, type, dose, and duration of H2RA therapy could have contributed to the difference in results. However, the mechanism of such a possible association is not clear. It has been hypothesized that small bowel bacterial colonization and overgrowth of the gastrointestinal tract is more frequent in patients on acid-suppressive therapy. Thus, it is possible that acid-suppressive medications could increase the risk of SBP due to small intestinal overgrowth. Whatever the underlying mechanism, the results from our meta-analysis indicate that patients with cirrhosis receiving PPI therapy have an increased risk of developing SBP compared with those not receiving this therapy, especially as there is evidence that 68% of PPI users who suffered SBP had no documented indications for PPI therapy. 13 However, heterogeneity between the studies may be related to different patient characteristics and dosage of acid-suppressive therapy. While our primary outcome of interest under evaluation was the association between acid-suppressive therapy and the risk of developing SBP, we initially speculated that lesser degrees of acid suppression (as with H2RA compared with PPI) may be less likely to be associated with SBP. This was indeed found to be the case with PPI usage associated with a significantly increased risk of developing SBP. However, while there was a trend toward an association between H2RA use and SBP, the association was not significant. It is possible that H2RAs may be a better alternative instead of PPIs, especially in the intensive care unit setting. However, it will be important that further studies be conducted to evaluate the association of H2RAs and SBP. Using the summary OR from the pooled studies, the use of PPI would result in a NNH of nine for one episode of SBP. The NNH should be interpreted with caution as fewer than 10 studies were included in the meta-analysis. Also, most of the included studies in the meta-analysis have taken place in a hospitalized setting, and given the heterogeneity between baseline SBP rates in various study settings, it is not justifiable to have one overall NNH for this meta-analysis. One recent published meta-analysis has evaluated PPI therapy and the risk of developing SBP. Trikudanathan et al. combined four observational studies with a total of 772 patients in their meta-analysis. 32 Their results indicated an increased risk of SBP among PPI users. The combined OR for all studies together was 2.77 (95% CI , P < 0.001). This study did not include analysis for H2RA. Also, the study had fewer subjects as they did not include the Northup et al. study, 27 and three new studies were published after their meta-analysis was available. Our metaanalysis is larger and a comprehensive study that evaluated the association between acid-suppressive therapy (PPI and H2RA) and the risk of SBP. Because H2RA are also acid-suppressive agents that may potentially predispose cirrhotic patients to SBP, we believe that it is imperative to account for their use in a comprehensive analysis. 240 Journal of Gastroenterology and Hepatology 28 (2013)

7 There is moderate heterogeneity between the included studies, and the sources for them are several and varied. While there was substantial heterogeneity in the estimate, the majority of the studies had a similar direction of effect in demonstrating a significant association. Because of the observational nature of the analyzed studies, it is not possible to establish causality based on this meta-analysis. To implicate causality and to understand the contribution of each factor, a well-designed study is needed. Ideally, a randomized controlled trial that involves withdrawing acidsuppressive therapy in those patients where it is not indicated as suggested by Bajaj et al. would provide more evidence to establish causality. 6 There are limitations to our study. Variations in the bacterial strain, type, dose, and duration of acid-suppression therapy, concomitant illness, and other medications need to be adjusted for each study. It is possible that different PPIs behave differently. The reported data on the majority of these variations were either unavailable or insufficient to permit subgroup analysis. While relatively fewer studies were included in the meta-analysis, it was inclusive based on a thorough literature search that included abstracts presented at conferences, studies from different countries, and a large sample size of more than 3000 subjects. All the studies were observational in nature and are thus subject to residual confounding despite statistical adjustment. Although all studies were observational, using random-effects models, greater reliance could have been placed on smaller studies with potentially inferior data compared with a fixed-effects model. Lastly, it is possible that patients who receive acid-suppressive therapy and those who develop SBP both tend to be sicker at baseline (confounding by indication). In conclusion, while acid-suppressive therapy has documented benefits for many indications, caution should be used when prescribing acid-suppressive therapy in cirrhotic patients with SBP. Future studies are needed to determine if the type and dosage of acid-suppressive therapy use in cirrhotic patients with ascites is associated with a change in the risk for SBP. References 1 Bustamante J, Rimola A, Ventura PJ et al. Prognostic significance of hepatic encephalopathy in patients with cirrhosis. J. Hepatol. 1999; 30: Tandon P, Garcia-Tsao G. Bacterial infections, sepsis, and multiorgan failure in cirrhosis. Semin. Liver Dis. 2008; 28: Wiest R, Krag A, Gerbes A. Spontaneous bacterial peritonitis: recent guidelines and beyond. Gut 2012; 61: Wiest R, Garcia-Tsao G. Bacterial translocation (BT) in cirrhosis. Hepatology 2005; 41: Bauer TM, Steinbruckner B, Brinkmann FE et al. Small intestinal bacterial overgrowth in patients with cirrhosis: prevalence and relation with spontaneous bacterial peritonitis. Am. J. Gastroenterol. 2001; 96: Bajaj JS, Zadvornova Y, Heuman DM et al. Association of proton pump inhibitor therapy with spontaneous bacterial peritonitis in cirrhotic patients with ascites. Am. J. Gastroenterol. 2009; 104: Chang CS, Chen GH, Lien HC, Yeh HZ. Small intestine dysmotility and bacterial overgrowth in cirrhotic patients with spontaneous bacterial peritonitis. Hepatology 1998; 28: Runyon BA. Patients with deficient ascitic fluid opsonic activity are predisposed to spontaneous bacterial peritonitis. Hepatology 1988; 8: Jump RL, Pultz MJ, Donskey CJ. Vegetative Clostridium difficile survives in room air on moist surfaces and in gastric contents with reduced acidity: a potential mechanism to explain the association between proton pump inhibitors and C. difficile-associated diarrhea? Antimicrob. Agents Chemother. 2007; 51: Deshpande A, Pant C, Pasupuleti V et al. Association between proton pump inhibitor therapy and Clostridium difficile infection in a meta-analysis. Clin. Gastroenterol. Hepatol. 2012; 10: Laheij RJ, Sturkenboom MC, Hassing RJ, Dieleman J, Stricker BH, Jansen JB. Risk of community-acquired pneumonia and use of gastric acid-suppressive drugs. JAMA 2004; 292: Prod hom G, Leuenberger P, Koerfer J et al. Nosocomial pneumonia in mechanically ventilated patients receiving antacid, ranitidine, or sucralfate as prophylaxis for stress ulcer. A randomized controlled trial. Ann. Intern. Med. 1994; 120: Goel GA, Deshpande A, Lopez R, Hall GS, van Duin D, Carey WD. Increased rate of spontaneous bacterial peritonitis among cirrhotic patients receiving pharmacologic acid suppression. Clin. Gastroenterol. Hepatol. 2011; 10: Stroup DF, Berlin JA, Morton SC et al. Meta-analysis of observational studies in epidemiology: a proposal for reporting. Meta-analysis of Observational Studies in Epidemiology (MOOSE) group. JAMA 2000; 283: Wells GA, Shea B, O Connell D et al. The Newcastle-Ottawa Scale (NOS) for Assessing the Quality of Nonrandomised Studies in Meta-Analyses Cited December Available from URL: 16 Hellstrom PM, Vitols S. The choice of proton pump inhibitor: does it matter? Basic Clin. Pharmacol. Toxicol. 2004; 94: Klok RM, Postma MJ, van Hout BA, Brouwers JR. Meta-analysis: comparing the efficacy of proton pump inhibitors in short-term use. Aliment. Pharmacol. Ther. 2003; 17: DerSimonian R, Laird N. Meta-analysis in clinical trials. Control. Clin. Trials 1986; 7: Kwok CS, Arthur AK, Anibueze CI, Singh S, Cavallazzi R, Loke YK. Risk of clostridium difficile infection with acid suppressing drugs and antibiotics: meta-analysis. Am. J. Gastroenterol. 2012; 107: Davies HT, Crombie IK, Tavakoli M. When can odds ratios mislead? BMJ 1998; 316: Higgins JP, Thompson SG. Quantifying heterogeneity in a meta-analysis. Stat. Med. 2002; 21: Deeks JJ, Higgins JPT, Altman DG. Analysing data and undertaking meta-analyses. In: Deeks JJ, Higgins JPT, Altman DG, eds. Cochrane Handbook for Systematic Reviews of Interventions, Version Chichester, UK: John Wiley & Sons, 2008; Loke YK, Price D, Herxheimer A. Adverse effects. In: Higgins JPT, Green S, eds. Cochrane Handbook for Systematic Reviews of Interventions. Chichester, UK: John Wiley & Sons, 2008; Egger M, Davey Smith G, Schneider M, Minder C. Bias in meta-analysis detected by a simple, graphical test. BMJ 1997; 315: Duval S, Tweedie R. Trim and fill: a simple funnel-plot-based method of testing and adjusting for publication bias in meta-analysis. Biometrics 2000; 56: Campbell MS, Obstein K, Reddy KR, Yang YX. Association between proton pump inhibitor use and spontaneous bacterial peritonitis. Dig. Dis. Sci. 2008; 53: Northup PG, Argo CL, Berg CL. Chronic proton pump inhibitor use is strongly associated with hepatorenal syndrome and spontaneous Journal of Gastroenterology and Hepatology 28 (2013)

8 bacterial peritonitis in cirrhosis patients. Hepatology 2008; 48: 325A. 28 Bulsiewicz WJ, Scherer JR, Feinglass JM, Howden CW, Flamm SL. 63 Proton Pump Inhibitor (PPI) use is independently associated with spontaneous bacterial peritonitis (SBP) in cirrhotics with ascites. Gastroenterology 2009; 136: A Choi EJ, Lee HJ, Kim KO et al. Association between acid suppressive therapy and spontaneous bacterial peritonitis in cirrhotic patients with ascites. Scand. J. Gastroenterol. 2011; 46: de Vos M, De Vroey B, Kidd F, Henrion J, Deltenre P. Do proton pump inhibitors or beta-blockers influence the occurrence of spontaneous bacterial peritonitis or the prognosis of cirrhotic patients with spontaneous bacterial peritonitis? 19th United European Gastroenterology Week; October 2011, Stockholm, Sweden. 31 van Vlerken LG, Huisman EJ, van Hoek B, etal. Bacterial infections in cirrhosis: role of proton pump inhibitors and intestinal permeability. Eur. J. Clin. Invest. 2012; 42: Trikudanathan G, Israel J, Cappa J, O Sullivan DM. Association between proton pump inhibitors and spontaneous bacterial peritonitis in cirrhotic patients a systematic review and meta-analysis. Int. J. Clin. Pract. 2011; 65: Supporting information Additional Supporting Information may be found in the online version of this article: Appendix S1 Literature search of the PubMed database. Table S1 NNH for PPI and H2RA in included studies. 242 Journal of Gastroenterology and Hepatology 28 (2013)