We thank Drs Eagles & Robson for their interest in our work and for their insightful comments. The authors discuss that our study did not sufficiently adjust for dietary factors, and further that body mass index (BMI) is a rather imprecise measure for obesity. In the paper the authors refer to, only trunk fat (central obesity) but not whole-body fat nor fat percentage was associated with cardiovascular disease in a cohort of patients with a normal BMI.^{Reference Chen, Arthur, Iyengar, Kamensky, Xue and Wassertheil-Smoller1} We fully agree that our study did not account for central obesity or diet, which could potentially explain part of the association we reported.^{Reference Wium-Andersen, Wium-Andersen, Prescott, Overvad, Jørgensen and Osler2} Consequently, we have looked into our data in order to explore this issue further.

In our study of 93 076 individuals, subsamples of 75 663 and 75 657 had measures of waist circumference and hip circumference, respectively. Both waist circumference and waist–hip ratio have been reported to reflect central obesity. Finally, 55 329 individuals had self-reported information on dietary intake of red meat, fruit, vegetables and fibre, which are key components in the Mediterranean diet.^{Reference Trichopoulou, Costacou, Bamia and Trichopoulos3} When we re-examined the bidirectional associations reported in our paper^{Reference Wium-Andersen, Wium-Andersen, Prescott, Overvad, Jørgensen and Osler2} in these subsamples before or after adjustment for (a) waist circumference; (b) waist–hip ratio; and (c) dietary intake of red meat, fruit, vegetables and dietary fibre, we did not find that these covariates explained the associations (Table 1). As discussed in our paper, we cannot exclude that changes in lifestyle or medication during follow-up could mediate part of the associations, but must conclude that baseline measures of central obesity and diet did not explain the bidirectional association between depression and cardiovascular disease observed in our study.

Table 1 Bidirectional associations between cardiovascular disease and depression before and after adjustment for waist circumference, waist–hip ratio and diet^a

IHS, ischaemic heart disease; WHR, waist–hip ratio.

a. Diet included self-reported information on red meat, fruits, vegetables and dietary fibre in four different variables.