We read with interest the article by Bowker et al. (1), which suggests that sulfonylurea or insulin therapy may increase cancer-related mortality. The premise that insulin is mitogenic for neoplastic cells is not proven. An expanding body of epidemiologic literature has linked insulin resistance and IGF-I, rather than absolute levels of insulin, with cancer incidence and prognosis.
There are several inherent limitations to this retrospective analysis. A total of 82.4% of the metformin group was also treated with sulfonylureas over the observation period. Also, patients requiring insulin likely had relatively poor glycemic control, longer duration of disease, and more subclinical or overt diabetes complications, perhaps not accounted for by the chronic disease score. These factors might increase mortality.
An area of pathophysiologic common ground between insulin resistance/diabetes and neoplastic diseases is the inflammatory milieu, which likely contributes to both of these disease states. The hyperglycemic, hyperinsulinemic, insulin-resistant state is proinflammatory and is very different than the euglycemic, hyperinsulinemic state achieved with intensive insulin therapy. Insulin itself possesses anti-inflammatory properties (2). Intranuclear levels of nuclear factor κβ, important in cancer progression, as well as plasma levels of several inflammatory cytokines (including vascular endothelial cell growth factor, i.e., vascular endothelial growth factor) are reduced through intravenous insulin infusion in obese nondiabetic subjects (3,4). Reduction in plasma vascular endothelial growth factor levels has also been shown during insulinization of poorly controlled or newly diagnosed diabetic adults and children (5,6). These data suggest that insulin possesses antiangiogenic as well as anti-inflammatory properties, which could benefit oncologic patients. As the authors point out, establishing a cause-and-effect relationship by a retrospective study in such a complex population is not possible. The positive or negative effects of aggressive insulinization in cancer patients remains an opened question that urgently requires randomized controlled clinical trials.
