We thank Farooki and Schneider (1) for their comments on our recently published article (2) on diabetes treatments and cancer-related mortality. As we had indicated (2), we agree with Drs. Farooki and Schneider that additional clinical data would aid our understanding of pathophysiologic mechanisms. Clinical intervention studies may be optimal in obtaining such pathophysiologic data. However, such studies may be impractical given the large numbers of patients required and long follow-up time needed to demonstrate statistically significant differences. It is for this reason that we used administrative databases that, in addition to the benefits of quick and inexpensive answers, also have the benefits of population-based estimates and no selection bias.
Drs. Farooki and Schneider point out that 82.4% of the metformin group was also treated with sulfonylureas during the observation period in our retrospective analysis. We did explore the cancer-related mortality in the metformin monotherapy group; however, there was a very small number of events (i.e., cancer-related mortalities) in the metformin monotherapy group (n = 40, 3.3%) compared with the combination therapy group (n = 205, 3.6%) and the sulfonylurea monotherapy group (n = 162, 4.9%). Thus, while it seems that the mortality rate in the metformin monotherapy group had lowest cancer-related mortality, the numbers are too small to allow for meaningful statistical analyses. As such, we combined the metformin monotherapy group and combination therapy group into the metformin users group as presented in the article.
Drs. Farooki and Schneider also states that the premise that insulin is mitogenic for neoplastic cells is not proven. We do agree that insulin resistance and IGF-1 may play a role in cancer incidence and prognosis. Other authors (3,4) have found evidence in the literature that insulin is a growth-promoting hormone that does in fact have mitogenic properties. It may be that hypersinulinemia in the face of insulin resistance, such as may be seen in insulin-treated type 2 diabetic patients, is increasing the risk, and, therefore, we do not want to exclude the potential link between insulin and cancer mortality.