We thank Drs Onder, Pedone, and Gambassi for their thoughtful comments on our article. Although we did not provide details, we did perform analyses that adjusted for diabetes, hypertension, and history of myocardial infarction and stroke; none of these adjustments materially affected our relative risk estimates regarding cigarette smoking or alcohol intake. We did not have information regarding depression.
It is difficult to see how these disorders in themselves could function as substantial confounding factors; presumably, the concern is that the risk factors, treatments, or behavioral consequences of these diseases would function in that way. In fact, cigarette smoking is a risk factor for all these disorders1 and may be a large part of the reason that they in turn have been associated with hip fracture or osteoporosis. Smoking is therefore both an exposure under study and a mediator of the relationship between purported confounders and hip fracture. To the extent this is true, adjustment for these disorders would have had the potential to bias our analyses of smoking and hip fracture.2