Abstract 17795: Symptoms as a Predicator of Mortality in Heart Failure

Abstract

Background and Significance: The American Heart Association guidelines for Heart Failure (HF) recommend that palliative care be addressed as an “ongoing key component of the plan of care”. Yet, clinicians are reluctant to discuss palliative care because the course of HF is unpredictable.

Purpose: To identify symptoms that may prognosticate end of life in HF patients.

Methods: We recruited subjects with systolic HF (N=194) from 2 inner city medical center HF clinics and asked them to rank their experience related to 50 general symptoms with HF symptoms embedded (0= not present to 4=severe). We also invited subjects to complete the same survey instrument at 3 month intervals for 1 year. We used latent class cluster analysis to identify symptom clusters at each data collection point and applied a multinomial logistic regression model to explain the probability of changing symptoms between periods.

Results: Subjects were primarily African American (77%), male (73%), mean age was 62.5 years (SD = 15.13, range 20 - 92). Modal NYHA classification was 3 (range 1 - 4). Four different clusters of symptoms were identified (Wald = 88.58, p<. 0001). The most prominent features of all 4 clusters were similar (weight gain, SOB with exertion, decreased urination, swelling in feet or ankles, joint pain and stiffness, leg pain, fatigue, cough at night, difficulty keeping thoughts straight, lower back pain, difficulty falling asleep, and frequent daytime dozing). Intensity of symptoms differentiated the clusters. Patients who entered the study with higher intensity symptoms based on total symptom burden score were 1.72 times more likely to die during the study period (p=.033). Additionally, intensity of difficulty sleeping increased sleeping, increased sensation of bloating and decreased appetite was proportionally greater in the patients who died during the course of the study.

Conclusions: Four different clusters of symptoms for HF were identified that differed primarily by intensity of symptoms and varied over time. Patients with higher symptom burden at study enrollment were more likely to die within one year. An understanding how HF symptom burden is related to mortality would assist practitioners in being more proactive related to initiation of palliative care.