Lifetime Caffeine Intake and the Risk of Epithelial Ovarian Cancer
Ovarian Cancer , Caffeine , Lifestyle
Objectives: To examine the association between lifetime caffeine intake and the risk of epithelial ovarian cancer (EOC) overall, according to tumour behaviour (invasive vs. borderline), and by invasive tumour type (Type I vs. Type II) and to assess menopausal status as an effect modifier. Methods: The PRevention of OVArian Cancer in Quebec (PROVAQ) study is a population-based case-control study that was conducted in Montreal (2011-2016). A total of 497 cases and 904 controls reported lifetime consumption of caffeinated beverages and other variables through in-person interviews. Lifetime average daily caffeine intake values were calculated for each participant and unconditional multivariable logistic regression was used to estimate adjusted odds ratios (with 95% confidence intervals) to describe the association between quartiles of caffeine intake and risk of EOC. A hybrid Directed Acyclic Graph (DAG) and change-in-estimate (CIE) procedure determined variables for statistical adjustment. Polytomous logistic regression was used for associations by tumour behaviour and invasive cancer type. Results: Results did not provide strong evidence of an association between lifetime caffeine consumption and overall EOC risk. The adjusted odds ratios (with 95% CIs) for the highest versus lowest quartile of caffeine intake were 1.21 (0.85-1.70) for overall EOC; 1.32 (0.90-1.94) for invasive EOC; and 0.99 (0.54-1.80) for borderline EOC. Further review indicated a suggestive trend of increasing risk with increasing quartile of caffeine consumption for invasive cases, although no such trend was seen for borderline cases. There was no appreciable difference in risk between Type I and Type II tumours. Stratification according to menopausal status yielded ORs that differed in direction and magnitude, however results were statistically non-significant (interaction p-value = 0.09). Further analyses by specific caffeinated drink types generally indicated non-significant results; yet, cola consumption elevated risk for overall, invasive, and invasive Type II EOC at certain consumption categories. Conclusions: Results did not demonstrate a pattern of association between lifetime caffeine intake and EOC risk and there was no strong evidence of menopausal status acting as an effect modifier. Additional drink type analyses indicate that cola intake may increase EOC risk. Further research with increased sample size may be warranted.