Racial Differences In Low Natriuretic Peptide Levels: Implications For Clinical Trials

Background Recent randomized clinical trials (RCT) in heart failure (HF) have increasingly required elevated natriuretic peptide (NP) levels as an inclusion criteria. Some patients with HF have unexpectedly low NP levels (e.g., Black patients, obese patients, patients with HF with preserved ejection fraction [HFpEF]). Currently, it is unclear whether these populations are disproportionately excluded from RCT participation due to the requirement for elevated NPs, with some suggesting utilizing a 20-30% lower threshold for NPs under these circumstances. Here, we investigate demographic and clinical factors associated with unexpectedly low NP levels in a diverse cohort of patients hospitalized with HF, and the potential implications on inclusion of Black patients into a prototype HF RCT. Methods We created a retrospective cohort of 31,704 patients (age 72 ± 16 years, 49% female, 52% Black, 53% HFpEF) hospitalized with acute HF at Emory Healthcare from 2010 to 2020 with B-type natriuretic peptide (BNP) measurements at admission. We quantified the prevalence of patients with unexpectedly low BNP levels (<50 pg/ml). Factors associated with low BNP levels were identified using multivariable logistic regression models, adjusted for age, gender, race, BMI, EF, eGFR, atrial fibrillation, and Social Deprivation Index. We simulated how many patients from our real-world cohort would be eligible for the PARAGON-HF trial using specific inclusion (age ≥ 55, EF ≥ 45 %) and exclusion (Hgb < 10, BMI > 40, SBP ≥ 180, SBP < 110, eGFR < 30, K > 5.2) criteria, as well as varying cut-offs for BNP. Results Unexpectedly low BNP levels were observed in 8.9% of the overall cohort, and Black patients were more likely than White patients to have low BNP levels (10.9% vs. 6.6%, p<.01). On multivariable analysis, factors associated with unexpectedly low BNP levels included HFpEF (aOR 3.76, 95% CI 3.36-4.20), BMI >30 kg/m2 (aOR 1.96, 95% CI 1.73 - 2.21), self-identification as Black (aOR 1.53, 95% CI 1.36 - 1.71), and male gender (aOR 1.45, 95% CI 1.31-1.60). Applying the limited inclusion and exclusion criteria from PARAGON-HF, n=3065 patients were eligible of which 1202 (39.2%) were Black. Adding a criteria of BNP > 50 , >100, and >150 pg/ml further decreased the total number (n=2616, 2177, and 1885) and the proportion of Black patients (36.1%, 35.5%, and 35.6%, all p≤.01) eligible, respectively. Conclusion Unexpectedly low BNP levels were present in nearly 10% of a diverse cohort of patients hospitalized with acute HF. Simulating inclusion into a prototype RCT demonstrated that requiring increasingly elevated NP levels disproportionately excluded Black patients. Recent randomized clinical trials (RCT) in heart failure (HF) have increasingly required elevated natriuretic peptide (NP) levels as an inclusion criteria. Some patients with HF have unexpectedly low NP levels (e.g., Black patients, obese patients, patients with HF with preserved ejection fraction [HFpEF]). Currently, it is unclear whether these populations are disproportionately excluded from RCT participation due to the requirement for elevated NPs, with some suggesting utilizing a 20-30% lower threshold for NPs under these circumstances. Here, we investigate demographic and clinical factors associated with unexpectedly low NP levels in a diverse cohort of patients hospitalized with HF, and the potential implications on inclusion of Black patients into a prototype HF RCT. We created a retrospective cohort of 31,704 patients (age 72 ± 16 years, 49% female, 52% Black, 53% HFpEF) hospitalized with acute HF at Emory Healthcare from 2010 to 2020 with B-type natriuretic peptide (BNP) measurements at admission. We quantified the prevalence of patients with unexpectedly low BNP levels (<50 pg/ml). Factors associated with low BNP levels were identified using multivariable logistic regression models, adjusted for age, gender, race, BMI, EF, eGFR, atrial fibrillation, and Social Deprivation Index. We simulated how many patients from our real-world cohort would be eligible for the PARAGON-HF trial using specific inclusion (age ≥ 55, EF ≥ 45 %) and exclusion (Hgb < 10, BMI > 40, SBP ≥ 180, SBP < 110, eGFR < 30, K > 5.2) criteria, as well as varying cut-offs for BNP. Unexpectedly low BNP levels were observed in 8.9% of the overall cohort, and Black patients were more likely than White patients to have low BNP levels (10.9% vs. 6.6%, p<.01). On multivariable analysis, factors associated with unexpectedly low BNP levels included HFpEF (aOR 3.76, 95% CI 3.36-4.20), BMI >30 kg/m2 (aOR 1.96, 95% CI 1.73 - 2.21), self-identification as Black (aOR 1.53, 95% CI 1.36 - 1.71), and male gender (aOR 1.45, 95% CI 1.31-1.60). Applying the limited inclusion and exclusion criteria from PARAGON-HF, n=3065 patients were eligible of which 1202 (39.2%) were Black. Adding a criteria of BNP > 50 , >100, and >150 pg/ml further decreased the total number (n=2616, 2177, and 1885) and the proportion of Black patients (36.1%, 35.5%, and 35.6%, all p≤.01) eligible, respectively. Unexpectedly low BNP levels were present in nearly 10% of a diverse cohort of patients hospitalized with acute HF. Simulating inclusion into a prototype RCT demonstrated that requiring increasingly elevated NP levels disproportionately excluded Black patients.