Context: Subclinical thyroid dysfunction is common in older people. However, its clinical importance is uncertain. Objective: Our objective was to determine the extent to which subclinical hyperthyroidism and hypothyroidism influence the risk of heart failure and cardiovascular diseases in older people. Setting and Design: The Prospective Study of Pravastatin in the Elderly at Risk (PROSPER) is an prospective cohort study. Patients: Patients included men and women aged 70–82 yr (n = 5316) with known cardiovascular risk factors or previous cardiovascular disease. Main Outcome Measures: Incidence rate of heart failure hospitalization, atrial fibrillation, and cardiovascular events and mortality according to baseline thyroid status were evaluated. Euthyroid participants (TSH =0.45–4.5 mIU/liter) were compared with those with subclinical hyperthyroidism (TSH <0.45 mIU/liter) and those with subclinical hypothyroidism (TSH ≥4.5 mIU/liter, both with normal free T4). Results: Subclinical hyperthyroidism was present in 71 participants and subclinical hypothyroidism in 199 participants. Over 3.2 yr follow-up, the rate of heart failure was higher for subclinical hyperthyroidism compared with euthyroidism age- and sex-adjusted hazard ratio (HR) = 2.93, 95% confidence interval (CI) = 1.37–6.24, P = 0.005; multivariate-adjusted HR = 3.27, 95% CI = 1.52–7.02, P = 0.002). Subclinical hypothyroidism (only at threshold >10 mIU/liter) was associated with heart failure (age- and sex-adjusted HR = 3.01, 95% CI = 1.12–8.11, P = 0.029; multivariate HR = 2.28, 95% CI = 0.84–6.23). There were no strong evidence of an association between subclinical thyroid dysfunction and cardiovascular events or mortality, except in those with TSH below 0.1 or over 10 mIU/liter and not taking pravastatin. Conclusion: Older people at high cardiovascular risk with low or very high TSH along with normal free T4 appear at increased risk of incident heart failure.