We read Dr Jay's comment with great interest. It is true that body mass index and smoking habits were not recorded in our database; hence, these variables could not be included in the stepwise logistic regression model from which the clinical score described in our article1 was developed. Therefore, it is impossible to know whether our score would have been different if smoking and/or obesity had been included; however, we believe that this is unlikely.
Perrier A, Bounameaux H, Wicki J. Is Smoking a Risk Factor for Pulmonary Embolism?—Reply. Arch Intern Med. 2001;161(15):1925–1926. doi: