Clinical background
The New York Heart Association (NYHA) functional classification is the most widely used instrument for grading symptom-related functional limitation in heart failure. It meets two clinical needs: capturing the severity of the disease at a given point in time and following changes over time in relation to treatment adjustments. Since heart failure is a condition in which symptom burden and objective cardiac function often correlate poorly, a structured way of systematising what the patient actually experiences is needed.
The NYHA class is moreover used as an inclusion criterion and stratification variable in virtually all major heart failure trials, which makes it a common language between clinic and research. Treatment guidelines from both the European Society of Cardiology and the American Heart Association routinely refer to the NYHA class when formulating indications for pharmacological and device therapy.
Determining the NYHA functional class
Unlike score-based risk instruments, the NYHA classification has no mathematical formula. The functional class is determined from the history according to the patient's symptoms at different levels of activity:
| Functional class | Criterion |
|---|---|
| Class I | No symptoms with ordinary physical activity |
| Class II | Slight limitation; comfortable at rest, ordinary activity causes symptoms |
| Class III | Marked limitation; comfortable at rest, less than ordinary activity causes symptoms |
| Class IV | Symptoms at rest; any physical activity increases discomfort |
The assessment is made clinically by a physician or nurse after a structured history. No objective measurement of cardiac function, biomarkers or exercise capacity forms part of the classification itself.
The classification originates from the Criteria Committee of the New York Heart Association and was first published in 1928. It has been revised several times, but the four-level symptomatic scale has been unchanged since the 1960s. There is no derivation cohort in the modern sense; the system is consensus-based, not statistically derived from a patient population.
Interpretation in practice
The NYHA class is not a treatment directive in itself, but it forms part of several treatment decisions. Any individual classification should always be interpreted together with the ejection fraction, biomarkers and comorbidity.
| Functional class | Clinical action |
|---|---|
| Class I | Heart failure diagnosed but the patient is asymptomatic. Optimised baseline treatment should be maintained. Focus on prevention and follow-up to detect deterioration early. |
| Class II | Mildly symptomatic. Full guideline-directed pharmacotherapy should be pursued. Assess whether doses are optimal and whether the patient tolerates target doses. |
| Class III | Marked symptomatic limitation. Requires intensified management, often through a heart failure clinic. Consider assessment for device therapy if not already done. Assess the need for diuretic adjustment and specialist care. |
| Class IV | Symptoms at rest. Advanced heart failure. Requires urgent or subacute specialist assessment. Consider mechanical circulatory support, transplantation or palliative care depending on prognosis and the patient's wishes. |
At follow-up over time, a change of one class is in principle clinically meaningful, but the sensitivity of the instrument to change is limited (see below). An unchanged class does not exclude a change in the patient's actual functional capacity.
Validation and performance
Since the NYHA classification is a history-based instrument, not a statistical model, traditional measures such as a c-statistic and calibration do not apply. Its performance has instead been evaluated through correlation with objective measures of function and through prognostic validation.
Correlation with objective measures. A systematic review of 37 studies with a total of 5,678 patients examined the relationship between NYHA class and the six-minute walk distance (6MWD) [1]. There was an inverse correlation: a higher NYHA class was associated with a shorter walking distance. The difference in mean 6MWD was significant between classes II and III (393 versus 321 m, p = 0.014) and between III and IV (321 versus 224 m, p = 0.027). By contrast, no significant difference was seen between classes I and II (420 versus 393 m, p = 0.416), indicating that the instrument has difficulty distinguishing asymptomatic from mildly symptomatic patients. There was substantial heterogeneity in 6MWD within each individual NYHA class, confirming that two patients in the same class can have very different objective functional capacity [1].
Another systematic review, of 38 studies and 2,645 patients, correlated NYHA class with peak oxygen uptake (pVO2) at cardiopulmonary exercise testing [2]. Here too there was a general inverse correlation, but with significant heterogeneity within each class. The difference in pVO2 between classes III and IV was not significant, further indicating that the instrument loses resolution at the severe end of the spectrum.
Prognostic value. A meta-analysis of 22 studies and 29,064 heart failure patients showed that the NYHA class was an independent predictor of mortality [3]. With class I as the reference, the hazard ratio for mortality was 1.54 (95% CI 1.16 to 2.04) for class II, 2.08 (95% CI 1.57 to 2.77) for class III and 2.53 (95% CI 1.25 to 5.12) for class IV. The association persisted after adjustment for demographics, comorbidity and the aetiology of the heart failure. Heterogeneity was low for classes II and III (I² = 0 and 2% respectively) but high for class IV (I² = 73%), explained in part by one study with a disproportionate effect [3].
Sensitivity to change. In a cohort study from the CHAMP-HF registry of 2,872 American heart failure patients with reduced ejection fraction, the NYHA class was compared with a patient-reported outcome (KCCQ-OS) at baseline and after 12 months [4]. During follow-up, 35% of patients had a change of at least one NYHA class, while 75% had a clinically meaningful change (at least 5 points) in KCCQ-OS. An improvement in KCCQ-OS was independently associated with reduced mortality (HR 0.59, 95% CI 0.44 to 0.80), but an improvement in NYHA class was not significantly associated with subsequent outcomes. This suggests that the NYHA class is less sensitive to clinically relevant change than structured symptom questionnaires [4].
Inter-rater reliability. A validation study of a structured aid for NYHA classification reported moderate agreement between assessors (a kappa of approximately 0.46) [5]. This accords with earlier observations that reliability is limited by the fact that "ordinary physical activity" and "less than ordinary activity" are subjective concepts interpreted differently by different clinicians.
Limitations
The NYHA class is based entirely on symptoms and does not reflect objective cardiac function. A patient with a severely reduced ejection fraction may be in class I, and a patient with a preserved ejection fraction may be in class IV. The classification should therefore never be used alone to judge the severity of the disease or to make decisions on device therapy or advanced treatment.
The most important limitations:
- Poor discrimination between adjacent classes. The boundary between classes I and II in particular is hard to draw, and objective measures such as the 6MWD and pVO2 do not differ significantly between these groups [1, 2]. Between classes III and IV, too, the instrument loses resolution.
- Moderate inter-rater reliability. The same patient may be classified differently by different assessors, particularly between classes II and III, where the boundary between "ordinary" and "less than ordinary" activity is hard to define [5]. A trend over time in the same patient, ideally assessed by the same clinician, is more informative than a single classification.
- Low sensitivity to change. Only a third of patients with a clinically relevant improvement or deterioration over a year are captured by a change of class, compared with three-quarters with the KCCQ-OS [4]. For serial follow-up of the treatment response, a patient-reported symptom questionnaire should be considered as a complement.
- Dependence on the patient's activity level. An inactive patient may report few symptoms not because the heart failure is well treated but because their physical activity is so low that symptoms are not provoked. This can lead to underestimation of severity.
- Not applicable outside heart failure. The classification was constructed for heart failure and should not be used for other cardiac conditions, even though it has historically been applied to valvular disease and other conditions. In valvular disease a modified version is sometimes used, but its interpretation and implications differ.
References
- Yap J, Lim FY, Gao F, et al. Correlation of the New York Heart Association Classification and the 6-Minute Walk Distance: A Systematic Review. Clinical Cardiology 2015. PMID: 26442458
- Lim FY, Yap J, Gao F, et al. Correlation of the New York Heart Association classification and the cardiopulmonary exercise test: A systematic review. International Journal of Cardiology 2018. PMID: 29678511
- Ali A, Siddiqui AA, Shahid I, et al. Prognostic value of quality of life and functional status in patients with heart failure: a systematic review and meta-analysis. Egyptian Heart Journal 2024. PMID: 39101961
- Greene SJ, Butler J, Spertus JA, et al. Comparison of New York Heart Association Class and Patient-Reported Outcomes for Heart Failure With Reduced Ejection Fraction. JAMA Cardiology 2021. PMID: 33760037
- Prasun MA, Stamp KD, McCoy TP, et al. The Reliability and Validity of Consistently Assess Signs and Symptoms of Heart Failure (CLASS-HF) Guide. Journal of Cardiovascular Nursing 2026. PMID: 41693041
