Weight loss and mortality risk in patients with chronic heart failure in the candesartan in heart failure: assessment of reduction in mortality and morbidity (CHARM) programme.

AIMS The curiosity that leanness is associated with poor survival in patients with chronic heart failure (CHF) needs further insight by investigating the impact of weight loss on prognosis in a large sample of patients across a broad spectrum of both reduced and preserved left ventricular (LV) systolic function. METHODS AND RESULTS We investigated the change in weight over 6 months in 6933 patients in the Candesartan in Heart failure: Reduction in Mortality and morbidity (CHARM) programme, and its association with subsequent mortality (1435 deaths) over a median 32.9 months follow-up using Cox proportional hazard models to account for the impact of body mass index and other risk predictors. We then used time-updated Cox models to relate each patient's ongoing data on annual weight change to their mortality hazard. The percentage weight loss over 6 months had a highly significant monotonically increasing association with excess mortality, both for cardiovascular and for other causes of death. Patients with 5% or greater weight loss in 6 months had over a 50% increase in hazard compared with those with stable weight. Weight loss carried a particularly high risk in patients who were already lean at study entry. Findings were similar in the presence of dependent oedema, preserved or reduced LV ejection fraction, and treatment with candesartan, although weight loss was significantly less common on candesartan. The time-updated analyses revealed an even stronger link between weight loss and short-term risk of dying, i.e. risk increased more than four-fold for patients whose last recorded annual weight loss exceeded 10%. Weight gain had a more modestly increased short-term mortality risk. Weight loss accelerates in the year prior to death. CONCLUSIONS Weight loss and leanness are important predictors of poor prognosis in CHF. Being lean and losing weight is particularly bad. The detection of weight change, and particularly weight loss, should be considered as an adverse sign prompting further evaluation.