Introduction Isoflavones (IFs) are polyphenolic compounds in plants which have: cellular anti-oxidant, pro-apoptotic and anti-proliferative properties. Pancreatic cancer cells cultured in isoflavone-containing media express increased levels of apoptotic and autophagic molecules such as caspases. To date, only one prospective cohort study has investigated dietary IFs intake and the risk of pancreatic cancer (PC), using food frequency questionnaires (FFQ) and 24 hour dietary recall questionnaires, and reported no associations. The aim of this study was to investigate IFs and the risk of pancreatic cancer using, for first time, 7 day food diaries (7-DFDs), the most accurate nutrient measurement method in large epidemiological studies.
Method 23 658 participants in the EPIC-Norfolk Study completed 7-DFDs at recruitment. Participant recorded free text from the 7-DFDs was translated into nutrient values using a specifically designed data base of 55 000 different food items (DINER – Data into nutrients for epidemiological research). The cohort was followed up for 17 years to identify those who developed PC, with cases verified by clinical note review. Dietary IFs intake and their subclasses namely: glycitein ganistein, daidzein and their metabolites: equol and O-desmethylangolensin were divided into quintiles in a sub-cohort of 4 056 participants and hazard ratios estimated using Cox regression adjusted for: age, gender, smoking, diabetes and calorie intake. Sensitivity analyses were done in smokers and non-smokers, as the antioxidant effects of IFs may oppose the action of pro-oxidants in cigarettes.
Results 86 participants (0.36%) developed PC (56% women) with a median age at diagnosis of 73.4 years. For total IFs intake there were no significant associations for any quintiles, including the lowest vs the highest (Q1 vs Q5, HR=1.10, 95% CI=0.60–2.00, p=0.76) and there was no trend across quintiles (HR trend=1.05, 95%CI=0.89–1.23, p=0.62). The findings were similar when adjusted for co-variates (Q1 vs Q5, HR=1.18, 95% CI=0.63–2.23, p=0.60). Similarly, there were no associations with any of the subclasses of IFs or their metabolites with PC. There was a borderline statistically significant trend across quintiles of IFs intake in smokers (HR trend=0.61, 95%CI=0.91–1.45, p<0.10), but not in non-smokers (HR trend=1.14, 95%CI=0.91–1.45, p=0.25).
Conclusion The findings suggest that IFs are not involved in preventing PC in the whole population, although in smokers there is some evidence for a possible protective effect. Further follow-up of this cohort is required to clarify if there is an inverse association in smokers. IF intake should be measured in aetiological studies of PC.
Disclosure of Interest None Declared
- pancreatic cancer