Dose-response relationship between coffee and the risk of pancreas cancer

Abstract

A case-control study was conducted and previous epidemiological data were reviewed in order to investigate the dose-response relationship between coffee and the risk of pancreas cancer. The case-control study was community-based and was carried out in Hokkaido, Japan, employing 141 patients with pancreas cancer and 282 controls (two for each case) matched for sex, age and place of residence. The dose-response relationship between coffee (cups/day) and the relative risk of this disease formed a U-shaped curve. The lowest relative risks (0.18 for male and 0.53 for female) were found among "occasional" drinkers. Epidemiological articles published between 1981 and 1993 were selected from Index Medicus using the two key words "coffee" and "pancreas cancer". In many of the previous case-control studies the curve of the dose-response relationship was also U-shaped, when the relative risks were calculated specifically using four or five levels of coffee dose. The nadirs of the relative risks, most of which ranged from 0.5 to 0.7, were found most frequently at small doses (1-2 or 3-4 cups/day). The results of meta-analysis of these studies formed a U-shaped curve. Studies of other types showed almost the same results. Thus it appears that small amounts of coffee might prevent pancreas cancer, whereas large amounts might cause the disease.