There was no significant association of insulin dose with cardiovascular (CV) mortality in a study of intensive treatment of 10,163 patients with type 2 diabetes and high CV risk. The study examined insulin exposure data with a mean follow-up of 5 years and examined associations between CV mortality and total, basal, and prandial insulin does over time. Researchers found:

• Before adjustment for covariates, higher insulin dose was associated with increased risk of CV death (HRs 1.83, 2.29, and 3.36 for total, basal, and prandial insulin, respectively).

• After adjustment for baseline covariates, no significant association of insulin dose with CV death remained.

• Further adjustment for severe hypoglycemia, weight change, attained A1c, and randomized treatment assignment did not materially alter the observation.

Citation: Siraj ES, Rubin DJ, Riddle MC, et al. Insulin dose and cardiovascular mortality in the ACCORD trial. [Published online ahead of print October 13, 2015]. Diabetes Care. doi: 10.2337/dc15-0598.

Commentary: This study is an interesting example of a confounding variable (insulin) and how data can be confusing if not looked at carefully. In the unadjusted analysis, higher insulin doses were associated with increased risk of CV death. When adjusted, this association disappeared. The reason, of course, is that higher insulin doses tend to be needed in patients who have had diabetes for a longer time and who often have more severe end organ damage. When controlled for this, insulin has no effect on CV outcomes. This study, which controlled well for this occurrence, shows insulin, as expected, not to be associated with adverse CV outcomes. ­—Neil Skolnik, MD