The authors of this paper, Kerstin Rödström and colleagues, have reported some errors in their article (BMJ 1999;319:890-3). In two places the paper referred to “multivariate” modelling when the authors intended to say “age adjusted.” These instances occur in the second sentence of the results section of the abstract and in the heading of the final column of table 2. In a third place—the “multivariate analysis” section of the results—the final sentence should have said that central obesity (measured as waist to hip ratio and waist circumference) “and the body mass index were no longer significant in the multivariate model” [not “seemed slightly more predictive than body mass index, whereas the reverse had previously been the case in the age adjusted model”].