6 Prognostic value of malnutrition screening tools in patients with chronic heart failure

Abstract

Background: Cachexia in heart failure (HF) is associated with adverse outcomes. Malnutrition is common in HF patients and contributes to cachexia but it is not part of the routine assessment of patients. There is no consensus on how best to screen for malnutrition in patients with HF. Purpose: To evaluate the prevalence, clinical associations and prognosis of malnutrition using three common multi-dimensional malnutrition indices in a large cohort of well-characterised patients with chronic heart failure (CHF). Methods: Consecutive referrals to a community HF clinic between 2000 and 2016 were analysed. HF was defined as signs or symptoms and evidence of cardiac dysfunction, either a reduced left ventricular ejection fraction at echocardiography (LVEF, <50%) or raised NTproBNP (>125 ng/L). Patients were screened for malnutrition using the geriatric nutritional risk index (GNRI), the controlling nutritional status (CONUT) score and the prognostic nutritional index (PNI). Results: 3386 patients had CHF (61% males, median age 75 (interquartile range (IQR): 67-81) years, median NTproBNP 1573 (IQR: 702-2799) ng/L). Of these, 1198 (35%) and 1458 (43%) patients had HF with reduced (HeFREF, LVEF<40%) and normal (HeFNEF, LVEF≥50%) ejection fraction, respectively. According to the GNRI (≤91), CONUT score (>4) and PNI (≤38), 6.7%, 10.0% and 7.5% patients were moderately to severely malnourished, respectively. Patients with HeFREF were at higher risk of malnutrition than those with HeFNEF. (8-12% vs 5-8%, respectively) Compared to normal patients, malnourished patients were older, had lower body mass index (BMI), worse symptoms and renal function; more likely to have atrial fibrillation, anaemia and reduced mobility. Worsening malnutrition correlated with increasing NTProBNP levels regardless of HF phenotypes.(R2=0.36-0.45(CONUT), 0.28-0.43(GNRI), 0.37-0.43(PNI), all p < 0.001) During a median follow-up of 1573 days (IQR: 702-2799 days), 1723 (50.9%) patients died. We created 25 prognostic Cox regression models using K-fold cross validation (k=25). CONUT and GNRI category (worse outcome with worsening malnutrition categories) were independently associated with outcome in 100% of the models while PNI category was an independent predictor in 19 models (76%). A base model for predicting mortality including variables with an inclusion frequency > 18 of 25 (>70%) models: age, sex, diastolic BP, heart rate, NYHA class III& IV vs I&II, urea, logNTProBNP, stroke and peripheral vascular disease had a C statistics of 0.72. Amongst the malnutrition indices, GNRI improved model performance most compared with base model. Patients who were severely malnourished according to GNRI had over 6 times greater mortality risk than normal patients. (Figure) Conclusion: Malnutrition is common amongst patients with CHF and is strongly related to mortality. Screening for malnutrition using GNRI should be routine adopted during assessment of HF patients to guide management. (Figure presented).