We determine the importance of indicators of nutrition including lymphocyte, total protein, albumin, cholinesterase and body mass index, and compare the prognostic significance in chronic heart failure (CHF). We examined consecutive 465 CHF patients (376 males, age 62 ± 14 years) who underwent cardiopulmonary exercise testing, echocardiography and blood examination including indicators of nutrition at the same time in our hospital. The patients were followed up [median period 766 days (interquartile range 500–1060)] to register cardiac deaths and rehospitalization due to worsening heart failure. There were 180 cardiac events during the follow-up periods. Patients with cardiac events had lower cholinesterase level than those without events (P < 0.001). On the receiver operating characteristic analysis, the best cut-off value for cholinesterase was 240 U/l (area under the curve 0.720). In the Kaplan–Meier analysis, patients with cholinesterase <240 U/l had significantly higher cardiac event rates than those with cholinesterase >240 U/l. Multivariable Cox proportional hazards model demonstrated that NYHA class III [hazard ratio (HR): 1.688, 95 % confidence interval (CI) 1.062–2.684, P = 0.027], eGFR (HR: 0.983, 95 % CI 0.971–0.995, P = 0.006), sodium concentration (HR: 0.947, 95 % CI 0.897–0.999, P < 0.046), log BNP (HR: 1.880, 95 % CI 1.509–2.341, P < 0.001), cholinesterase (HR: 0.996, 95 % CI 0.993–0.998, P = 0.002) and exertional periodic breathing (HR: 1.619, 95 % CI 1.098–2.388, P = 0.015) were independent factors to predict adverse clinical outcomes. Serum cholinesterase level was an important prognostic factor in CHF.