Symptom Clusters Predict Event-Free Survival in Patients With Heart Failure

Physical symptoms are likely to occur in clusters that may be associated with adverse outcome in patients with heart failure (HF). Despite the importance of early recognition of worsening symptoms in HF management, the impact of physical symptoms on adverse outcome has not been explored in the context of symptom clusters.The purposes of this study were to explore which physical symptom clusters occur in HF patients and to determine the impact of symptom clusters on event-free survival.A total of 421 patients (60% male; 62 [SD, 14] years; 72% New York Heart Association class II/III) completed the modified Memorial Symptom Assessment Scale-Heart Failure to measure physical symptoms during an index hospitalization in Seoul, Korea. Times to first event of cardiac rehospitalization and cardiac death were collected for 12 months after discharge and confirmed by review of hospital records. An agglomerative hierarchical clustering approach with Ward's method was used to identify symptom clusters. Hierarchical Cox hazard regression was used to determine the impact of symptom clusters on cardiac rehospitalization and cardiac mortality.Two distinct symptom clusters, the dyspneic and the weary symptom clusters, occurred in patients with HF. Shortness of breath, difficulty breathing when lying flat, and waking up breathless at night comprised the dyspneic symptom cluster. Lack of energy, lack of appetite, and difficulty sleeping comprised the weary symptom cluster. In hierarchical Cox hazard regression, elevated distress from the weary symptom cluster independently predicted cardiac rehospitalization (hazard ratio, 1.45; 95% confidence interval, 1.09-1.93), and increased distress from the dyspneic symptom cluster independently predicted cardiac mortality (hazard ratio, 2.00; 95% confidence interval, 1.16-3.34).The weary and the dyspneic symptom clusters predicted cardiac rehospitalization and cardiac mortality, respectively. Patient education for self-monitoring of symptoms should focus on symptom clusters rather than single symptom.