Comparison of preoperative NT-proBNP and simple cardiac risk scores for predicting postoperative morbidity after non-cardiac surgery with intermediate or high surgical risk.

ABSTRACT

BACKGROUND: Chronic heart failure (HF) is frequent in elderly patients undergoing non-cardiac surgery. Preoperative risk stratification is vital and can be achieved using simple clinical risk scores or preoperative N-terminal prohormone of brain natriuretic peptide (NT-proBNP) measurement. This study aimed to compare the predictivity of the revised cardiac risk index (RCRI), the American University of Beirut cardiovascular risk index (AUB-HAS2), and a score proposed by Andersson et al. for postoperative 30-day morbidity to preoperative NT-proBNP. METHODS: Preoperative NT-proBNP was measured in 199 consecutive patients aged ≥ 65 years undergoing elective non-cardiac surgery with intermediate or high surgical risk. The areas under the receiver operating characteristic curve (AUCROC) for the composite morbidity endpoint (CME) comprising the incidence of any rehospitalisation, acute decompensated HF, acute kidney injury, and any infection at postoperative day 30 were assessed. Multivariable logistic regression analysis derived new scores from the simple risk scores and the NT-proBNP cut-off of 450 pg/mL. RESULTS: AUB-HAS2, but not RCRI or Andersson score, significantly predicted the CME (AUB-HAS2 AUCROC 0.646, p < 0.001; RCRI AUCROC 0.560, p = 0.126; Andersson AUCROC 0.487, p = 0.760). The AUCROC was comparable between preoperative NT-proBNP (0.679, p < 0.001) and AUB-HAS2 (p = 0.334). Multivariable analyses revealed a preoperative NT-proBNP ≥ 450 pg/mL to be the strongest predictor of CME among the individual score components (p < 0.001). Adding preoperative NT-proBNP improved the predictive value of AUB-HAS2 and RCRI (modified AUB-HAS2 AUCROC 0.703, p < 0.001; modified RCRI AUCROC 0.679, p < 0.001; both p < 0.001 vs original scores). The predictive value of the modified RCRI and AUB-HAS2 was comparable to preoperative NT-proBNP alone (p = 0.988 vs modified RCRI, p = 0.367 vs modified AUB-HAS2). CONCLUSIONS: The predictive value of postoperative morbidity varies significantly between the available simple perioperative risk scores and can be enhanced by preoperative NT-proBNP. New scores, including preoperative NT-proBNP, should be evaluated in large multicentre cohorts. TRIAL REGISTRATION German Clinical Trials Register DRKS00027871.