Prognostic value of B-type natriuretic peptide for nursing- and healthcare-associated pneumonia and aspiration pneumonia in comparison with procalcitonin and A-DROP score: a prospective cohort study

Background: Pneumonia in elderly patients who require nursing care is becoming more and more common among the aging populations of developed countries, and treatment is an important topic worldwide. A simpler prognostic indicator would be expected to improve the treatment of pneumonia. This single-center, prospective cohort study aimed to compare the usefulness of B-type natriuretic peptide (BNP), procalcitonin (PCT), and the A-DROP score, which have been reported to correlate with pneumonia prognoses, such as aspiration pneumonia (AP).

Methods: We included patients who were admitted to the Kanazawa Medical University Himi Municipal Hospital with a diagnosis of either nursing- and healthcare-associated pneumonia (NHCAP) or AP between January 1, 2012 and July 31, 2019. We collected demographic, clinical, and laboratory characteristics and outcome data from electronic medical records, and calculated A-DROP scores. The primary outcome was 30-day mortality. We evaluated correlations with the primary outcome using the chi-square test, Fisher exact test, t-test, Cox-regression analysis, and receiver operating characteristic curve analysis.

Results: Of the 1,215 patients with pneumonia, 297 were eligible for the study, of whom 37 (12%) died whin 30 days. After univariate analysis, we performed Cox proportional-hazards analysis for BNP, PCT, A-DROP score, albumin, C-reactive protein, and disseminated intravascular coagulation, which were significantly correlated with the primary outcome in univariate analysis. As a result, only BNP showed a significant correlation (P=0.008, 95% CI: 1.30-6.06). No significant correlation was obtained in PCT (P=0.529) and A-DROP score (P=0.107). Furthermore, we generated receiver operating characteristic curve to estimate the prognostic cut-off values of BNP for the primary outcome of NHCAP and AP. The optimal cut-off value of BNP for predicting death was 179.3 pg/mL (sensitivity 62.2%, specificity 76.2%, negative likelihood ratio 0.50%, positive likelihood ratio 2.61%). And, BNP yielded the highest area under the curve (0.72) in comparison with PCT (0.67) and A-DROP score (0.69).

Conclusions: BNP may be a more clinically useful prognostic factor for NHCAP and AP than PCT or A-DROP score, and should be considered as a routine test at the beginning of these treatments.