Context

Low-carbohydrate—high-protein diets have become popular in developed countries as an aid to weight loss. However, uncertainties about their effectiveness in weight loss and concerns about long-term adverse impacts on cardiovascular health remain. Several sys-tematic reviews1 2 have concluded that, with the increas-ing prevalence of obesity, there is a need for long-term evidence/studies of low-carbohydrate diets. The authors investigate the long-term consequences of low-carbohydrate diets (LCHD) on cardiovascular health in a female Swedish population. Methods Women aged 30–49 recruited from the Swedish Women’s Lifestyle and Health cohort were followed for a mean of 15.7 years. A self-administered questionnaire was used to assess lifestyle, physical activity and disease-related vari-ables; dietary intake was assessed with a validated food frequency questionnaire. Participants were then assigned an energy-adjusted score for protein intake and a score for carbohydrate intake; these were used to assign each participant with a composite low-carbohydrate–high-protein score ranging from two (high-carbohydrate–low-protein) to 20 (low-carbohydrate–high-protein). The primary outcome of interest was a first diagnosis of car-diovascular disease (CVD). Survival analyses were con-ducted to generate incidence rate ratios for cardiovascular disease; these looked at the age-adjusted incidence rate ratio stratified by a number of cardiovas-cular conditions, individual non-dietary (including age at enrolment, length of education, height, body mass index, physical activity, smoking at enrolment, alcohol intake and hypertension) and dietary (energy, saturated fat and unsaturated fat intake) variables. Findings There were 1270 incident cardiovascular events recorded over the 15.7-year period. A 5 % increase in the inci-dence of cardiovascular events per 2 unit increase in the 20-unit low-carbohydrate score was identified. Despite the lack of controlling for potential confounding (other than age) and time to event, the authors suggest that this relationship is consistent with existing evidence