Folate Intake and the Risk of Upper Gastrointestinal Cancers

A Systematic Review and Meta-analysis

Martin Tio; Juliana Andrici; Michael R Cox; Guy D Eslick


J Gastroenterol Hepatol. 2014;29(2):250-258. 

In This Article


Our systematic review found inconsistent associations between folate intake and different upper gastrointestinal cancers. While we found an association between dietary folate and a decreased risk of esophageal and pancreatic cancers, we did not observe an association between dietary folate and gastric cancer.

The meta-analysis of dietary folate and esophageal cancer shows a significant inverse association between dietary folate intake and both histological subtypes of esophageal cancer. This association is further supported by evidence that polymorphisms in folate metabolism genes that result in lower circulating levels of folate also increase esophageal cancer risk.[11,54] Although we found a homogenous and statistically significant association, there are a number of limitations to the interpretation of this result. Firstly, our search failed to find any prospective studies, and so our result is open to being affected by recall bias and selection bias. Secondly, no studies were found that analyzed total folate intake, which comprises of the combination of folate from the diet along with folate from supplements.

In developed countries, the prevalence of regular supplementation of the diet with vitamins and minerals is as high as 49%.[55] A number of studies attempted to control for supplement use by either comparing "ever" users versus "never" users, or by comparing "high" versus "low" supplement use, but no study attempted to quantify the actual folate intake from supplements. Given that seven from the total of nine studies were conducted in developed countries, it is likely that the true folate intake is being underestimated in our meta-analysis. Additionally, given that there is evidence from clinical trial data that folate supplementation may increase the overall risk of cancer,[56] further investigation of total folate intake is warranted.

No association was found between dietary folate intake and gastric cancer, although this result was highly heterogeneous. Stratifying our results into analyses of prospective and retrospective studies reduced the heterogeneity, but did not change the overall result. We were not able to perform meta-analyses for total folate intake or for plasma/serum folate levels, as only one study was found for each folate measurement type. However, both studies similarly reported no association between total folate intake or plasma folate levels and the risk of gastric cancer.[35,52] A significant limitation of our result is that the majority of studies did not control for Helicobacter pylori, a major risk factor for the development of gastric cancer.[57] Of the two studies that did control for H. pylori, there was no clear trend to the results, with one study finding no association,[39] while the other found a decreased risk of gastric cancer.[34]

This lack of association between folate intake and risk of gastric cancer contrasts with a number of meta-analyses of genetic polymorphisms implicating folate deficiency in certain subgroups as a risk factor for gastric cancer.[10,12,58] Our meta-analysis only looked at studies that examined elevated risk in the setting of the general population, and so it is likely that any elevated risk in these subgroups would not have been detected. At this stage though, it appears unlikely that folate intake plays a major role in gastric cancer risk, with the possible exception in individuals with genetic susceptibilities.

The meta-analysis of folate intake and pancreatic cancer found a decreased risk of pancreatic cancer with increased dietary folate, but not with increased total folate. One possible explanation is that since the total folate meta-analysis had fewer studies, it may simply have had insufficient statistical power to detect a decreased risk. Our dietary folate results are additionally complicated by the fact that when the dietary folate analysis was stratified by study design, the prospective analysis was only statistically significant when the outlier study, Keszei et al.,[33] was removed. Keszei et al. had significantly lower levels of folate intake compared with the other included studies, with a mean high level of folate intake at approximately 246 μg/day, compared levels of 350–700 μg/day in the other studies. This low level of folate intake may explain why Keszei et al. contributed to the heterogeneity of the pooled analysis.

A recent meta-analysis of 14 cohort studies that included unpublished data on folate and pancreatic cancer risk found no association with dietary or total folate intake.[59] This meta-analysis had a similar number of pancreatic cancer cases to our overall pancreatic cancer meta-analysis, but had more than twice the number of cases compared with our meta-analysis stratified by prospective studies. This raises the possibility that our dietary folate result may be a spurious result influenced by publication bias. Although we found no statistical evidence of publication bias in our dietary folate analysis, we cannot rule out a false negative result because of the small number of studies analyzed. The fact that our meta-analysis of plasma folate levels and pancreatic cancer risk found no significant association adds further support to the possibility that folate may not play a major role in the development of pancreatic cancer.

Our systematic review had a number of strengths. We followed the MOOSE guidelines on reporting meta-analyses of observational studies. We searched through multiple databases, and were not limited by language. We only performed pre-specified subgroup analyses.