Log in with your email address username.


[Correspondence] Body-mass index and all-cause mortality – Authors’ reply

Associations of measured body-mass index (BMI) with mortality within just a few years of the BMI measurement can be strongly distorted by reverse causality (ie, by life-threatening neoplastic, respiratory, vascular, or other diseases having caused weight loss before the BMI was measured). By contrast, associations of measured BMI with mortality several years after the BMI measurement should be little affected by reverse causality, although they can still be strongly affected by confounding, especially in populations in which low BMI is correlated with smoking.