Heart failure (HF) treatment recommendations are centered on New York Heart Association (NYHA) classification, such that most apparently asymptomatic patients are not eligible for disease-modifying therapies. To assess within-patient variation in NYHA classification over time, the association between NYHA class and an objective measure of HF severity (N-terminal pro-B-type natriuretic peptide [NT-proBNP] level), and their association with long-term prognosis in the PARADIGM-HF trial. All patients in PARADIGM-HF were in NYHA class II or higher at baseline and were treated with sacubitril-valsartan during a 6- to 10-week run-in period before randomization. Patients classified as NYHA class I, II, and III in PARADIGM-HF were compared at randomization. NYHA class at randomization after 6 to 10 weeks of the run-in period. Primary outcome was cardiovascular death or first HF hospitalization. Logistic regression models, areas under the receiver operating characteristic curve (AUC), kernel density estimation overlaps, and Cox proportional hazards models were used. The analysis included 8326 patients with known NYHA classification at randomization. Of 389 patients in NYHA class I, 228 (58%) changed functional class during the first year after randomization. Level of NT-proBNP was a poor discriminator of NYHA classification: for NYHA class I vs II, the AUC was 0.51 (95% CI, 0.48-0.54). For NT-proBNP level, estimated kernel density overlap was 93% between NYHA class I vs II, 79% between NYHA I vs III, and 83% between NYHA II vs III. Patients classified as NYHA III displayed a distinctively higher rate of cardiovascular events (NYHA III vs I, hazard ratio [HR], 1.84; 95% CI, 1.44-2.37; NYHA III vs II, HR, 1.49; 95% CI, 1.35-1.64). Patients in NYHA class I and II revealed lower event rates (NYHA II vs I, HR, 1.24; 95% CI, 0.97-1.58). Stratification by NT-proBNP level (<1600 pg/mL or ≥1600 pg/mL) identified subgroups with distinctive risk, such that NYHA class I patients with high NT-proBNP levels (n = 175) had a numerically higher event rate than patients with low NT-proBNP levels from any NYHA class (vs I, HR, 3.43; 95% CI, 2.03-5.87; vs II, HR, 2.12; 95% CI, 1.58-2.86; vs III, HR, 1.37; 95% CI, 1.00-1.88). In this study, patients in NYHA class I and II overlapped substantially in objective measures and long-term prognosis. Physician-defined "asymptomatic" functional class concealed patients who were at substantial risk for adverse outcomes. NYHA classification might be limited to differentiate mild forms of HF. ClinicalTrials.gov Identifier: NCT01035255.