N-terminal pro-brain natriuretic peptide levels and risk of death in sickle cell disease.

CONTEXT: Thirty percent of patients with sickle cell disease (SCD) develop pulmonary hypertension, a major risk factor for death in this population. A validated blood biomarker of pulmonary hypertension in SCD could provide important prognostic and diagnostic information and allow the exploration of the prevalence of pulmonary hypertension in participants in the 1996 Multicenter Study of Hydroxyurea in Sickle Cell Anemia (MSH) Patients' Follow-up Study. Levels of N-terminal pro-brain natriuretic peptide (NT-proBNP) provide such information in patients with idiopathic pulmonary arterial hypertension.

OBJECTIVE: To determine the relationship between NT-proBNP levels and severity of pulmonary hypertension and prospective mortality in patients with SCD.

DESIGN, SETTING, AND PARTICIPANTS: NT-proBNP levels were measured in 230 participants in the National Institutes of Health (NIH) Sickle Cell Disease-Pulmonary Hypertension Screening Study (enrollment between February 2001 and March 2005) and in 121 samples from patients enrolled starting in 1996 in the MSH Patients' Follow-up Study. A threshold level predictive of high pulmonary artery pressure and mortality was identified in the NIH Sickle Cell Disease-Pulmonary Hypertension Screening Study and used to define an a priori analytical plan to determine the prevalence and associated mortality of pulmonary hypertension in the MSH follow-up study.

MAIN OUTCOME MEASURES: Severity of pulmonary hypertension and risk of all-cause mortality.

RESULTS: NT-proBNP levels were higher in patients with sickle cell pulmonary hypertension and correlated directly with tricuspid regurgitant jet velocity in the NIH cohort (R = 0.50, P<.001). An NT-proBNP level of 160 pg/mL or greater had a 78% positive predictive value for the diagnosis of pulmonary hypertension and was an independent predictor of mortality (21 deaths at 31 months' median follow-up; risk ratio, 5.1; 95% confidence interval, 2.1-12.5; P<.001; 19.5% absolute increase in risk of death). In the MSH cohort, 30% of patients had an NT-proBNP level of 160 pg/mL or greater. An NT-proBNP level of 160 pg/mL or greater in the MSH cohort was independently associated with mortality by Cox proportional hazards regression analysis (24 deaths at 47 months' median follow-up; risk ratio, 2.87; 95% confidence interval, 1.2-6.6; P = .02; 11.9% absolute increase in risk of death).

CONCLUSIONS: Pulmonary hypertension, as indicated by an NT-proBNP level of 160 pg/mL or greater, was very common in patients in the NIH study and in the MSH cohort. The MSH analysis suggests that rates of vaso-occlusive pain episodes in these patients were unrelated to risk of death; this risk was largely determined by occult hemolytic anemia-associated pulmonary hypertension.