We conducted a review of previous cohort studies on the association between a history of diabetes mellitus (DM) and the occurrence of cancer. We limited the papers to those concerning cohort studies on 9 cancer sites, i.e. the kidney, liver, biliary tract, pancreas, colon or rectum, prostate, breast, endometrium, and ovary, in addition to all cancers. With regard to kidney, liver, biliary tract, pancreatic, colorectal, breast, and endometrial cancers, the risk of cancer development has been consistently reported to be positively associated with DM by two or more cohort studies. In contrast, DM was shown to relate negatively to the risk of prostate cancer by two cohort studies. However, there were no cohort studies which showed an either significantly positive or negative association of DM with ovarian cancer. Elevated levels of insulin or IGFs among DM patients have been proposed as a causal mechanism of increased risk for most of the reviewed cancers. In addition, increased estrogen levels in DM patients have been suggested to explain the casual mechanism of increased risk for kidney, breast and endometrial cancers, and decreased risk for prostate cancer. On the other hand, the possibility of detection bias has been suggested in the association of DM with the risk of most of these cancers. Obesity and heavy consumption of alcohol have been indicated as confounding factors in the relationship of DM to the risk for some of them. Thus, further studies are necessary for firm conclusions regarding the association of DM with cancer risk.