Natriuretic Peptide Levels Correlate with Pulmonary Pressures and Prospective Mortality in SCD: Use of This Biomarker To Identify Prevalence and Mortality of Pulmonary Hypertension in the MSH Cohort.

Pulmonary hypertension [PH-tricuspid regurgitant jet velocity (TRV) ≥2.5 m/s] is a common complication of sickle cell disease associated with high mortality. Identification of biomarkers of PH and mortality could facilitate screening and risk stratification in this population. Validated biomarkers would provide methods for retrospective evaluation of the prevalence and prognosis of PH in large historical cohorts of patients such as the Multicenter Study of Hydroxyurea in Sickle Cell Anemia(MSH). Because brain natriuretic peptide(BNP) is released from the ventricles during pressure strain, we hypothesized that BNP levels would correlate with the severity of PH and prospective risk of death in patients with SCD. BNP was measured in 45 African-American control subjects and 230 patients with SCD. Median (interquartile range) BNP(pg/ml) was higher in patients with PH than patients without PH or controls[+PH: 206(81–701),-PH: 47(26–104), C: 29, P<0.001]. BNP levels directly correlated with age (R=0.32, P<0.001), creatinine (R=0.22, P<0.001), LDH(R=0.31, P<0.001), TRV (R=0.5, P<0.001), pulmonary vascular resistance (R=0.5, P=0.001); and inversely with hemoglobin(R=0.41, P<0.001), cardiac output(R=0.47, P= 0.003) and 6-minute walk distance(R=0.51, P=0.001). The area under the ROC for BNP and the diagnosis of pulmonary hypertension was 0.84 (P<0.001). A cutoff value of 160 pg/ml (corresponding to the 75th percentile for the population) had 58% sensitivity and 98% specificity for the diagnosis of PH. Cox proportional hazards regression identified BNP as an independent predictor of mortality(RR 2.17,95% CI 1.2–3.8, P =0.001) with clear mortality break point at the 75th percentile(160 pg/ml).

To independently explore the prevalence and associated risk of PH in patients with sickle cell disease, a BNP value of 160 pg/ml was used as an indicator of PH. BNP levels were then measured in plasma samples collected in 121 patients who were enrolled in the MSH patient’s follow-up study that started in 1996. These patients had received hydroxyurea or placebo for two years, had moderately severe disease based on study entry criteria, and had 9-years of comprehensive follow-up. An abnormal BNP level ≥160 pg/ml was present in 30% of patients in the MSH cohort. BNP levels correlated directly with age(R=0.35, P<0.001) and creatinine (R=0.24, P<0.001), and inversely with hemoglobin(R=−0.54, P<0.001). There was no correlation between BNP and rate of painful episodes or acute chest syndrome, use of hydroxyurea or leukocyte count. A high BNP level in the MSH cohort was associated with mortality by logistic regression(OR 3.04,95% CI 1.2–7.6, P = 0.018) and Cox proportional hazards regression analysis(RR 2.87, P=0.017). The relationship remained significant for continuous log- transformed BNP values and after adjustment for other covariates. These studies confirm that PH is common, mechanistically linked to hemolytic anemia and the major risk factor for death in SCD. Provocatively, the MSH analysis suggests that rates of pain episodes in this small sample of seriously ill patients were unrelated to risk of death: this risk was largely determined by a high BNP level, which is probably explained by undiagnosed hemolysis-associated PH.