Hemoglobin and B-type natriuretic peptide preoperative values but not inflammatory markers, are associated with postoperative morbidity in cardiac surgery: a prospective cohort analytic study

INTRODUCTION: Risk stratification in cardiac surgery significantly impacts outcome. This study seeks to define whether there is an independent association between the preoperative serum level of hemoglobin (Hb), leukocyte count (LEUCO), high sensitivity C-reactive protein (hsCRP), or B-type natriuretic peptide (BNP) and postoperative morbidity and mortality in cardiac surgery. METHODS: Prospective, analytic cohort study, with 554 adult patients undergoing cardiac surgery in a tertiary cardiovascular hospital and followed up for 12 months. The cohort was distributed according to preoperative values of Hb, LEUCO, hsCRP, and BNP in independent quintiles for each of these variables. RESULTS: After adjustment for all covariates, a significant association was found between elevated preoperative BNP and the occurrence of low postoperative cardiac output (OR 3.46, 95% CI 1.53–7.80, p = 0.003) or postoperative atrial fibrillation (OR 3.8, 95% CI 1.45–10.38). For the combined outcome (death/acute coronary syndrome/rehospitalization within 12 months), we observed an OR of 1.93 (95% CI 1.00–3.74). An interaction was found between BNP level and the presence or absence of diabetes mellitus. The OR for non-diabetics was 1.26 (95% CI 0.61–2.60) and for diabetics was 18.82 (95% CI 16.2–20.5). Preoperative Hb was also significantly and independently associated with the occurrence of postoperative low cardiac output (OR 0.33, 95% CI 0.13–0.81, p = 0.016). Both Hb and BNP were significantly associated with the lengths of intensive care unit and hospital stays and the number of transfused red blood cells (p < 0.002). Inflammatory markers, although associated with adverse outcomes, lost statistical significance when adjusted for covariates. CONCLUSIONS: High preoperative BNP or low Hb shows an association of independent risk with postoperative outcomes, and their measurement could help to stratify surgical risk. The ability to predict the onset of atrial fibrillation or postoperative low cardiac output has important clinical implications. Our results open the possibility of designing studies that incorporate BNP measurement as a routine part of preoperative evaluation, and this strategy could improve upon the standard evaluation in terms of reducing adverse postoperative events.