Continuing on with the remaining two insulin glargine studies:

• Jonasson et al from Sweden looked at 114,481 patients who had prescriptions for insulin glargine in 2006 or 2007. In their analysis, they controlled for gender, age, smoking, BMI, age at onset of diabetes, age at birth of first child, cardiovascular disease, and estrogen use. They found that only the risk for breast cancer was elevated with a RR of 1.29-3.00. Looking specifically at all malignancies, prostate cancer or gastrointestinal cancers did not show any association with insulin glargine use.
• Lastly, Hemkens et al from Germany looked at patients in the state health insurance fund who had received first-time therapy for diabetes with insulin in some form. This included a total of 127,031 patients. The investigators controlled for many, many variables in a statistical analysis that I think most clinicians (present company included) will struggle with. The upshot was, controlling for age, gender and dose of insulin, among other variables, the risk of malignant neoplasms was higher with glargine than with lispro or aspart or human insulins (RR 1.20-1.42). They saw here a dose response curve as well.

So, what do we conclude? The trend seems compelling in retrospective analysis, and there are preclinical data also suggesting that rDNA insulin products may be carcinogenic (see the editorial by Smith and Gale which accompanies the above articles for more details). The main question for me is, if this association is really causation, why such a short time between exposure and cancer? We know however that the data do not appear enough to pull glargine off the market, but certainly larger, confirmatory studies are urgently needed.