NT-proBNP in cardiopulmonary resuscitated patients treated with mild therapeutic hypothermia is not independently associated with mortality: a retrospective observational study

Abstract

Background: In spite of the introduction of mild therapeutic hypothermia (MTH), mortality rates remain high in patients with return of spontaneous circulation (ROSC) after cardiac arrest (CA). To date, no accurate and independent biomarker to predict survival in these patients exists. B-type natriuretic peptide (BNP) was found to provide both prognostic and diagnostic value in various cardiovascular diseases, including survival to hospital discharge in patients with ROSC. However, the biologically inactive counterpart of BNP, NT-proBNP, was found to be a more stable and accurate analyte. The current retrospective observational study investigates the value of NT-proBNP to predict 28-day mortality in post-CA patients treated with MTH, as well as the dynamics of NT-proBNP during MTH.

Methods: NT-proBNP levels were measured in post-CA patients cooled via cold intravenous saline infusion and water-circulating body wraps (Medi-Therm®, Gaymar). Plasma samples were obtained before cooling was started, at the start and end of the maintenance phase and at the end of rewarming.

Results: 250 patients, admitted between 2009 and 2013, had NT-proBNP levels measured on ICU admission and were included for the evaluation of NT-proBNP as a prognostic marker. In the 28 days following ICU admission, 114 patients died (46%). Non-survivors had significantly higher NT-proBNP (median 1448 ng/l, IQR 366-4623 vs median 567 ng/1, IQR 148-1899; P < 0.001) levels on ICU admission. Unadjusted odds ratios for 28-day mortality were 1.7 (95% CI 0.8-3.5), 1.6 (0.8-3.3) and 3.6 (1.7-7.5) for increasing quartiles of NT-proBNP as compared to the lowest quartile. Adjusted odds ratios were 1.1 (95% CI 0.5-2.5), 1.1 (0.5-2.5) and 1.6 (0.7-3.8), respectively. A cut-off value of 834 ng/l achieved a sensitivity of 58% and a specificity of 58% to predict 28-day mortality. Of 113 patients, NT-proBNP values of each MTH phase were available and grouped in decreased or increased levels in time. Both decreases and increases of NT-proBNP values were observed during the MTH phases, but presence of either was not associated with outcome.

Conclusions: High NT-proBNP plasma concentrations on ICU admission are associated with high 28-day mortality in post-CA patients treated with MTH in a univariate analysis, but not in a multivariate analysis. Increases or decreases of NT-proBNP levels during MTH appear unrelated to 28 day mortality.

Keywords: Cardiac arrest; Mild therapeutic hypothermia; NT-proBNP; Natriuretic peptides.

Figure 1

NT-proBNP measurement time points. NT-proBNP levels were measured at four different time points: 1) after admission to the ICU, prior to cooling 2) during the maintenance phase, after reaching a target temperature of 32.5°C, 3) at the end of the maintenance phase, before the initiation of rewarming 4) at the end of the rewarming phase.

Figure 2

Odds ratios for 28-day mortality. Unadjusted (A) and adjusted odds ratios (B) for 28-day mortality in increasing quartiles. Adjusted odds ratios were calculated by correcting for age, diabetes mellitus, ventricular fibrillation/tachycardia as the initial rhythms and cardiac cause of arrest. Data are presented as OR (95% Confidence Interval). Abbreviations: OR, Odds Ratio.

Figure 3

28 day survival curves. Patients with higher NT-proBNP levels died earlier and more frequently than patients with lower levels (log rank test for trend, P < 0.01).

Figure 4

NT-proBNP ROC curve. An optimal sensitivity and specificity of respectively 58% and 58% was reached with a baseline NT-proBNP cut-off value of 834 ng/l. Corresponding positive and negative predictive values were 54% and 62%, respectively. When considering maximal specificity, a cut-off of 20101 ng/l yielded a specificity of 99% and a sensitivity of 10%. PPV and NPV were 92% and 57%, respectively. AUC 0.63 (95%CI 0.56 - 0.70; P < 0.0001).