Background: Shortness of breath, edema, and fatigue are often present in HF and are common among ambulatory care patients. These signs and symptoms (S/S) are predictive of hospital admission for HF but their lack of specificity presents a challenge in risk stratification for HF. Consideration of the cardiac biomarker NT-proBNP may improve the predictive performance of S/S in risk stratification to prevent incident HF hospitalization.
Methods: Shortness of breath, edema, and fatigue and NT-proBNP levels were assessed among ARIC cohort participants during 2nd clinical examination (1990-1992). Degree of lower extremity edema was assessed by examination, and NT-proBNP assayed from serum. Shortness of breath and fatigue were assessed by self-report. Incidence of HF was ascertained from ongoing monitoring of participants’ hospitalization records (ICD-9 code 428.xx in any position) through December 31, 2012. Linear regression was used to examine cross-sectional association of signs and symptoms of HF with NT-proBNP levels. Poisson regression was used to assess age-adjusted mortality rates. Cox proportional hazard models, adjusted for age, gender, race, systolic blood pressure (BP), BP-lowering medication use, prevalent CHD, BMI, smoking status and diabetes were used to examine the effect of NT-proBNP on the association of HF S/S with 10-year risk of incident HF hospitalization.
Results: Of the 13,312 ARIC participant, free of prevalent HF (mean (SD) age 56.9 (5.7) years, 55% female, 24% black), 5,873 (44.1%) reported shortness of breath, fatigue or had bilateral edema. Median NT-proBNP level (pg/ml) was 47.4 (interquartile range (IQR): 25.4, 86.3) in those without HF S/S, 67.3 (IQR: 34.7, 127.5) among those with shortness of breath; 62.4 (IQR: 31.9, 113.8) among those with edema, and 56.2 (IQR: 29.9, 103) among those with fatigue. Incidence of HF hospitalization (age-adjusted rate/1,000 person-years) was associated with presence of HF S/S: 7.6 (95% confidence interval (CI): 7.3, 8.1) among those without any S/S; 22.2 (95% CI: 19.8, 24.6) among those with shortness of breath, 18.3 (95% CI: 15.8, 20.8) among those with edema, and 12.5 (95% CI: 11.5, 13.4) among those with fatigue. The 10-year risk of HF hospitalization was 64% (95% CI: 47%, 83%) higher among those with any of the HF S/S, as compared to those free of HF S/S. Addition of NT-proBNP to HF risk estimation led to net reclassification of 9.2 % (95% CI: 6.9%, 11.8%) of the study participants with respect to their risk of HF hospitalization.
Conclusion: Signs and symptoms frequently related to HF are prevalent in ambulatory care and are associated with increased long term risk of HF hospitalization. Inclusion of these S/S in combination with NT-proBNP in clinical prediction tools may improve the management of patients presenting with signs/symptoms suggestive of HF.