Total fat consumption and pancreatic cancer risk: a meta-analysis of epidemiologic studies

Fat consumption has been hypothesized to influence pancreatic cancer risk, but the results of epidemiologic studies have been controversial. We conducted a systematic review and meta-analysis of case-control and cohort studies to investigate this issue. Relevant published studies were identified by searching MEDLINE (PubMed) through February 2014. Two authors (Q.-W.S. and Q.-Y.Y.) independently assessed eligibility and the extracted data. Study-specific relative risks (RRs) were pooled using a random-effects model. We also carried out heterogeneity and publication bias analyses. Six cohort and 13 case-control studies with 6159 pancreatic cancer cases and 1,068,476 noncases were included in this meta-analysis. The summary RR for pancreatic cancer for the highest versus lowest intake was 1.04 [95% confidence interval (CI)=0.90-1.20, I²=57.3%, P for heterogeneity=0.001] for total fat. In addition, when separately analyzed by study design, case-control (RR=1.03, 95% CI=0.83-1.27, I²=55.8%, P for heterogeneity=0.007) and cohort studies (RR=1.05, 95% CI=0.85-1.29, I²=66.7%, P for heterogeneity=0.010) yielded similar results. Furthermore, no statistically significant associations were observed in the subgroup analyses on the basis of fat source, geographic location, whether using energy-adjusted models, and whether adjusted for several potential confounders and important risk factors. There was no evidence of publication bias or significant heterogeneity between subgroups on meta-regression analyses. The results of this meta-analysis do not support an independent association between diets high in total fat and pancreatic cancer risk.