The recent article by Stattin et al. (1) examined whether hyperglycemia was associated with an increased risk of cancer. Using the Västerbotten Intervention Project of northern Sweden, the authors calculated relative risks of incident cancer for levels of fasting and postload glucose. They observed an association between hyperglycemia and total cancer risk in women and men for several cancer sites, suggesting that their results provide further evidence for an association between abnormal glucose levels and risk of cancer.
While the evidence is compelling, we wondered to what extent the authors might have considered potential confounding factors in the observed relationship. It is quite likely that elevated glucose levels are secondary to insulin resistance (2,3) and the resultant hyperinsulinemic state. It is likely that hyperinsulinemia may be the risk factor, or potentially a mechanism, of increased cancer rates in this group of individuals. This relationship has been confirmed in a number of studies (2,3). The authors do not mention the potential hypothesis of hyperinsulinemia and cancer risk in their article (1). We wondered whether they have information on circulating insulin levels in the sample to potentially test this hypothesis.
We also wondered how many individuals went on to develop type 2 diabetes throughout the duration of follow-up. Furthermore, it would be interesting to consider how many patients went on oral hypoglycemic agents or exogenous insulin. Recent research, including our own work, has suggested a link between antidiabetic therapies in patients with type 2 diabetes and cancer (4,5). We observed an increased risk of cancer-related mortality in patients with type 2 diabetes who used sulfonylureas compared with metformin and also an increased risk with exogenous insulin use (4). Similarly, Evans et al. (5) found a decreased risk of cancer in diabetic patients who used metformin compared with insulin secretagoues. These epidemiologic associations are supported by basic research, which suggests a mitogenic effect of insulin (6) and a protective effect of metformin (7). It might, therefore, be interesting for Stattin et al. to consider these potential confounders in their assessment of hyperglycemia as a risk factor for the incidence of cancer.
