In Reply We agree with Drs Tasci and Safer that the risk of metformin and insulin compared with metformin plus sulfonylurea as a second-line diabetes treatment after failure of metformin monotherapy may differ in certain populations. They point to eFigure 3 that shows a statistically significant increased risk in persons aged 65 years or older and no statistically significant increase in younger persons.
However, the confidence intervals for the HRs in the 2 age groups have considerable overlap and a formal test for interaction between metformin plus insulin and age is not significant (P = .22). Thus, we have not proven significant differences in risk by age. The numerical differences observed could represent real differences or simply chance variation in the estimated associations.
Roumie CL, Greevy RA, Griffin MR. Insulin vs Sulfonylureas for Second-Line Diabetes Treatment—Reply. JAMA. 2014;312(16):1693-1694. doi:10.1001/jama.2014.11120