408 Background: The association between Helicobacter pylori infection (HPI) and gastric cancer is well known, but it is unclear whether HPI is also a risk factor for colorectal cancer. Several epidemiological studies on the latter association have yielded conflicting results. The aim of this study is to summarize available evidence on the association between HPI and CRC, evaluating its magnitude and direction in a meta-analysis. Methods: Two reviewers independently conducted a systemic search on Medline, OvidSP and PubMed databases from January 1980 to July 2011 for studies on the association between HPI and CRC. The reference lists of eligible studies were next reviewed for additional studies on the subject. Firstly, a combined analysis including all studies was done. Next, subgroup analysis by study design and country of study (USA Vs Europe Vs Asia) were also performed. All analyses were done using the random effects model. Publication bias was assessed using the Begg’s and Egger’s tests and visual inspection of funnel plot. All analyses were performed using STATA 11. Results: Sixteen studies (14 retrospective and 2 prospective) including 12,892 participants were included in the analysis. Overall, HP was associated with a 49% significantly higher risk of CRC (OR 1.49, 95% CI: 1.22-1.82; P < 0.001). By study design, the association persisted only among the retrospective studies with a pooled OR of 1.43 (95% CI: 1.31-1.56, P = 0.004). The subgroup analysis by study region showed significant associations in Europe (OR 1.35, 95% CI 1.09-1.66) and Asia (RR 1.43, 95% CI 1.29-1.58). Though there was a 17% higher risk noted among studies done in the USA, this was not statistically significant. There was no evidence of publication bias in all the analyses. Conclusions: Current evidence on the association between H. pylori infection and CRC remains inconclusive. The absence of any association among the prospective studies (with less risk of bias) suggests that the association seen among the retrospective studies could be due to residual confounding. Better quality data is required before a conclusive statement on the association between HPI and CRC can be made.

No abstract

4546 Background: There is evidence suggesting that PZD use may be a risk factor for BC, yet these reports are controversial. The aim of this meta-analysis is to review available data and evaluate the association between PZD and BC. Methods: Two independent reviewers conducted a systematic search of the Cochrane Library, OvidSP and PubMed for articles published January 1970 to April 2012. MeSH search terms included PZD, Actos, thiazolidinediones, diabetes mellitus (DM), and BC. Only studies reporting an effect measure for the association between PZD and BC were included. Subgroup analyses by study design and country were performed. Analysis was done using a random effect model after a preliminary review showed evidence of study heterogeneity. This was then assessed using the Cochrane’s Q and I2 statistics. Publication bias was evaluated using the Begg’s and Egger’s tests, and funnel plots. All analyses were performed using REVMAN 5.1. Results: A total of 6 studies (3 cohort and 3 case control) met the inclusion criteria. The overall pooled risk ratio (RR) for the association between PZD and BC was 1.12 (95% CI 1.09, 1.15; P <0.001). The summary RR for cohort and case control studies were 1.13(95% CI 1.07, 1.19; P <0.001) and 1.11(95% CI 1.07, 1.15; P <0.001), respectively. The subgroup analysis showed a significant association with BC in the USA and Europe, but not in Asia (RR 0.98, 95% CI 0.71, 1.34; P = 0.88). The association of PZD and BC was more sensitive for treatment duration (>12 month). Conclusions: Our study analyzed 1,214,071 patients, demonstrating a weak association between PZD and BC. The results were significant with increase in treatment duration, which corresponds to previous studies. This meta-analysis has limitations. Not all studies reported known variables associated with BC, such as smoking or occupational exposure. They did not address patient BMI, time from DM diagnosis or previous drug therapy. These parameters reflect severity of disease and level of insulin-resistance. Future studies should evaluate markers of insulin-resistance with PZD use and correlate with BC. Understanding the link between PZD and BC in the context of DM may allow physicians to determine a more accurate risk of PZD use.