Background: The aetiology of pancreatic cancer (PC) has been extensively studied and is the subject of numerous meta-analyses and pooled analyses. We have summarized results from these pooled and meta-analytical studies to estimate the fraction of PCs attributable to each of the identified risk factors.
Methods: Using a comprehensive strategy, we retrieved 117 meta-analytical or pooled reports dealing with the association between specific risk factors and PC risk. We combined estimates of relative risk and estimates of exposure to calculate the fraction of PCs caused or prevented by a particular exposure.
Results: Tobacco smoking ('strong' evidence) and Helicobacter pylori infection ('moderate' evidence) are the major risk factors associated with PC, with respective estimated population attributable fractions of 11-32% and 4-25%. The major protective factors are history of allergy ('strong' evidence) and increasing fruit or folate intake ('moderate' evidence), with respective population preventable fractions of 3-7% and 0-12%.
Conclusions: We summarized results of 117 meta-analytical or pooled data reports dealing with 37 aetiological exposures, to obtain robust information about the suspected causes of PC. By combining these estimates with their prevalences in the population, we calculated population attributable or population preventable fractions. About two-thirds of the major risk factors associated with PC are potentially modifiable, affording a unique opportunity for preventing one of our deadliest cancers.
Keywords: Pancreas cancer; meta-analysis; pooled analysis; population attributable fraction; review; risk factors.
© The Author 2014; all rights reserved. Published by Oxford University Press on behalf of the International Epidemiological Association.