A prospective cohort study of dynamic cell-free DNA elevation during cardiac surgery with cardiopulmonary bypass

Abstract

Cardiac surgery and cardiopulmonary bypass (CPB) are associated with a systemic inflammatory reaction that occasionally induces a life-threatening organ dysfunction caused by the dysregulated host response to the damage-associated molecular patterns (DAMPs). In severe inflammation, cell-free DNA (cfDNA) and histones are released by inflammatory cells and damaged tissue and act as DAMPs. We sought to characterize the changes in circulating cell-free DNA (cfDNA) levels during CPB. Primary outcomes were renal failure, ventilation time (>18 hr), length of stay (LOS) in the intensive care unit (ICU) (>48hr), hospital LOS (>15 days), and death. We looked for associations with blood tests and comparison to standard scores. In a prospective cohort study, we enrolled 71 patients undergoing non-emergent coronary artery bypass grafting. Blood was drawn at baseline, 20 and 40 minutes on CPB, after cross-clamp removal, and 30 minutes after chest closure. cfDNA was measured by our fast fluorescent method. Baseline cfDNA levels [796 (656-1063) ng/ml] increased during surgery, peaked after cross-clamp removal [2403 (1981-3357) ng/ml] and returned to baseline at recovery. The difference in cfDNA from 20 to 40 minutes on CPB (ΔcfDNA 40-20) inversely correlated with peripheral vascular disease (PVD), longer ventilation time, and longer ICU and hospital length of stay (LOS). Receiver operating characteristic (ROC) curve of ΔcfDNA 40-20 for long ICU-LOS (>48hr) was with an area under the curve (AUC) of 0.738 (p = 0.022). ROC AUC of ΔcfDNA 40-20 to long Hospital LOS (>15 days) was 0.787 (p = 0.006). Correction for time on CPB in a multivariate logistic regression model improved ROC-AUC to 0.854 (p = 0.003) and suggests that ΔcfDNA 40-20 is an independent risk factor. To conclude, of measured parameters, including STS and Euroscore, the predictive power of ΔcfDNA 40-20 was the highest. Thus, measurement of ΔcfDNA 40-20 may enable early monitoring of patients at higher risk. Further studies on the mechanism behind the negative association of ΔcfDNA 40-20 with PVD and outcomes are warranted.

Fig 1. Cell-Free DNA (cfDNA) according to consecutive time points of coronary artery bypass grafting surgery (CABG).

A) Blood was drawn from 71 patients undergoing CABG before skin incision (Baseline), 20 and 40 minutes after initiation of cardiopulmonary bypass (CPB), immediately after cross-clamp removal, and in recovery of 30 minutes after chest closure. cfDNA was measured by a fast fluorescence assay. B) cfDNA levels at 20 and 40 minutes CPB (ΔcfDNA 40–20) according to intensive care unit (ICU) length of stay (LOS). **p<0.01, ****p<0.0001, ns-not significant, tested by non-parametric Friedman test, Kruskal-Wallis test and Wilcoxon test.

Fig 2. Changes in cfDNA and neutrophil numbers at recovery according to the length of stay (LOS) in ICU.

A) ΔcfDNA 40–20 according to ICU LOS. B) ΔcfDNA 40–20 receiver operating characteristic (ROC) curve for long ICU LOS (>48hr). AUC-Area under the curve. C) Neutrophils numbers at recovery according to ICU LOS. *p<0.05, **p<0.01, ****p<0.0001, ns-not significant Kruskal-Wallis test. D) Linear correlation between (ΔcfDNA 40–20) to neutrophils number at recovery from surgery. ΔcfDNA 40–20 correlation to ICU-LOS, Spearman r = -0.365, p = 0.0019 and neutrophils correlation to ICU-LOS r = -0.243, p = 0.0415. E) ΔcfDNA 40–20 according to PVD. Mann-Whitney test, *p<0.05.

Fig 3. ΔcfDNA and scores according to hospital length of stay (LOS), and univariate ROC curves.

A) ΔcfDNA (40–20 minutes), Euroscore and Society of Thoracic Surgeon (STS) score according to LOS. *p<0.05, ns-not significant, Kruskal-Wallis test. B) Scores univariate ROC curves for long LOS (>15 days). AUC (95% CI) for ΔcfDNA = 0.787 (0.64–0.94), p = 0.006, for Euroscore = 0.511 (0.29–0.74), p = 0.917, for STS = 0.626 (0.44–0.81) p = 0.229.

Fig 4. Logistic regression models for prolonged hospital LOS for ΔcfDNA 40–20 adjusted to time on CPB.

Multivariate ROC curve of long hospital LOS (>15 days) according to logistic regression models that include ΔcfDNA 40–20 and time on CPB.