Associations of fats and carbohydrates with cardiovascular disease and mortality—PURE and simple? – Authors' reply

We concur with Jim Mann and colleagues that dietary recommendations should be informed with all available evidence, and that ideally large randomised trials on clinical outcomes should be used. In the absence of randomised trials, large, well designed observational studies on clinical outcomes provide the next best type of evidence. Our results are consistent with previous observational studies1Wang DD Li Y Chiuve SE et al.Association of...