•Subclinical thyroid dysfunction is very common in patients with chronic heart failure (∼10%). •There is a U-shaped relation between Thyroid-Stimulating Hormone and mortality. •Subclinical hypothyroidism is associated with worse outcome in univariable analysis, but the association disappears after adjustment for age and natriuretic peptide level. •There is a weak but significant association between increasing Thyroid-Stimulating Hormone and outcome in patients with heart failure with normal ejection fraction. Thyroid dysfunction is common in patients with chronic heart failure (CHF), but there is conflicting evidence regarding its prognostic significance. We investigated the relation between thyroid function and prognosis in a large, well characterized cohort of ambulatory patients with CHF. Heart failure was defined as signs and symptoms of the disease and either left ventricular systolic dysfunction (LVSD) mild or worse (heart failure with reduced ejection fraction [HFrEF]), or no LVSD and raised amino-terminal pro-B-type natriuretic peptide (NT-proBNP) levels (>125 ng/L; heart failure with normal ejection fraction [HFnEF]). Euthyroid state was defined as a thyroid-stimulating hormone (TSH) level between 0.35 and 4.70 mIU/l, hypothyroidism as TSH >4.70 mIU/l, and hyperthyroidism as TSH <0.35 mIU/l. 2997 patients had HFrEF and 1995 patients had HFnEF. 4491 (90%) patients were euthyroid, 312 (6%) were hypothyroid, and 189 (4%) were hyperthyroid. In univariable analysis, both hypothyroid patients (hazard ratio [HR] 1.25, 95% confidence interval [CI] 1.08 to 1.45) and hyperthyroid patients (HR 1.21, 95% CI 1.01 to 1.46) had a greater risk of death compared with euthyroid patients. There was a U-shaped relation between TSH and outcome. Increasing TSH was a predictor of mortality in univariable analysis (HR 1.02, 95% CI 1.01 to 1.03), but the association disappeared in multivariable analysis. The three strongest predictors of adverse outcome were increasing age, increasing NT-proBNP, and higher NYHA class. In conclusion, although thyroid dysfunction is associated with worse survival in patients with CHF, it is not an independent predictor of mortality. Thyroid dysfunction is common in patients with chronic heart failure (CHF), but there is conflicting evidence regarding its prognostic significance. We investigated the relation between thyroid function and prognosis in a large, well characterized cohort of ambulatory patients with CHF. Heart failure was defined as signs and symptoms of the disease and either left ventricular systolic dysfunction (LVSD) mild or worse (heart failure with reduced ejection fraction [HFrEF]), or no LVSD and raised amino-terminal pro-B-type natriuretic peptide (NT-proBNP) levels (>125 ng/L; heart failure with normal ejection fraction [HFnEF]). Euthyroid state was defined as a thyroid-stimulating hormone (TSH) level between 0.35 and 4.70 mIU/l, hypothyroidism as TSH >4.70 mIU/l, and hyperthyroidism as TSH <0.35 mIU/l. 2997 patients had HFrEF and 1995 patients had HFnEF. 4491 (90%) patients were euthyroid, 312 (6%) were hypothyroid, and 189 (4%) were hyperthyroid. In univariable analysis, both hypothyroid patients (hazard ratio [HR] 1.25, 95% confidence interval [CI] 1.08 to 1.45) and hyperthyroid patients (HR 1.21, 95% CI 1.01 to 1.46) had a greater risk of death compared with euthyroid patients. There was a U-shaped relation between TSH and outcome. Increasing TSH was a predictor of mortality in univariable analysis (HR 1.02, 95% CI 1.01 to 1.03), but the association disappeared in multivariable analysis. The three strongest predictors of adverse outcome were increasing age, increasing NT-proBNP, and higher NYHA class. In conclusion, although thyroid dysfunction is associated with worse survival in patients with CHF, it is not an independent predictor of mortality.