Hasnain and Vieweg raise some important limitations regarding observational studies such as ours. As stressed in our article,1 we fully agree that our study cannot confirm a causal relationship between APD use and hyperglycemia because of the potential role of unmeasured confounders. However, they also argue that the dramatic difference between first-time and prevalent users contradicts an acute hyperglycemic effect of APDs, which is untrue. Some adverse drug effects may occur early in susceptible individuals. Offending drugs tend to be discontinued in these patients, and these “susceptibles” are depleted from the population leaving lower-risk “prevalent” users.2,3 The absolute risk of an acute effect may be low because there may be relatively few susceptible individuals, but their risk would be higher relative to less vulnerable patients. Because we found significant risks among both first-time and prevalent users, our study suggests that APDs can have both acute and chronic metabolic effects. While our study cannot prove a direct effect of antipsychotic agents on hyperglycemia, it indicates that the initiation of an APD represents a critical period during which diabetic patients are particularly vulnerable to serious glycemic deterioration.
Lorraine L. Lipscombe, Andrea Gruneir, Linda E. Levesque, Sudeep S. Gill, Paula A. Rochon. Antipsychotic Use and Risk for Hyperglycemia in Elderly Patients With DM—Reply. Arch Intern Med. 2009;169(22):2162–2166. doi:10.1001/archinternmed.2009.450