Macrophage Migration Inhibitory Factor Predicts Outcome in Complex Aortic Surgery

Abstract

The perioperative inflammatory response is associated with outcome after complex aortic repair. Macrophage migration inhibitory factor (MIF) shows protective effects in ischemia-reperfusion (IR), but also adverse pro-inflammatory effects in acute inflammation, potentially leading to adverse outcome, which should be investigated in this trial. This prospective study enrolled 52 patients, of whom 29 (55.7%) underwent open repair (OR) and 23 (44.3%) underwent endovascular repair (ER) between 2014 and 2015. MIF serum levels were measured until 72 h post-operatively. We used linear mixed models and ROC analysis to analyze the MIF time-course and its diagnostic ability. Compared to ER, OR induced higher MIF release perioperatively; at 12 h after ICU admission, MIF levels were similar between groups. MIF course was significantly influenced by baseline MIF level (P = 0.0016) and acute physiology and chronic health evaluation (APACHE) II score (P = 0.0005). MIF level at 24 h after ICU admission showed good diagnostic value regarding patient survival [sensitivity, 80.0% (28.4-99.5%); specificity, 51.2% (35.1-67.1%); AUC, 0.688 (0.534-0.816)] and discharge modality [sensitivity, 87.5% (47.3-99.7%); specificity, 73.7% (56.9-86.6%), AUC, 0.789 (0.644-0.896)]. Increased perioperative MIF-levels are related to an increased risk of adverse outcome in complex aortic surgery and may represent a biomarker for risk stratification in complex aortic surgery.

Keywords: intensive care unit; macrophage migration inhibitory factor; survival; thoracoabdominal aortic aneurysm.

Figure 1

Macrophage migration inhibitory factor (MIF) release during aortic surgery. Compared to endovascular repair, the open procedure induced higher perioperative MIF release, and this difference persisted up to the first hours after intensive care unit (ICU) admission. Endovascular repair induced only a moderate perioperative MIF increase; MIF values are presented as median with interquartile range (ng/mL).

Figure 2

ROC analysis of perioperative MIF levels and patient survival. ROC analysis was performed to evaluate the diagnostic capacities of (A) perioperative MIF levels and (B) MIF levels during ICU admission with regard to survival. If an elevated MIF value indicates that the patient is likely to die after surgery, the ROC curve should be farther from the bisecting line (Sensitivity = 1 − Specificity). (C) Sensitivity (Se), specificity (Sp), likelihood ratios (LR +/−), and area under the curve (AUC) are reported for either the Youden optimal cut-off (maximize Se + Sp − 1) or for a sensitivity cut-off of at least 75%. * Good-to-moderate diagnostic quality: LQ+ of >3 and LQ− of <0.3. * Excellent diagnostic quality: LQ+ of >10 and LQ− of >0.1.

Figure 3

ROC analysis of perioperative MIF levels and patient discharge modality. ROC analysis was performed to evaluate the diagnostic capacities of (A) perioperative MIF levels and (B) MIF levels during ICU admission with regard to discharge modality. If an elevated MIF value indicates an adverse discharge modality, the ROC curve should be farther from the bisecting line (Sensitivity = 1 − Specificity). (C) Sensitivity (Se), specificity (Sp), likelihood ratios (LR +/−), and area under the curve (AUC) are reported for either the Youden optimal cut-off (maximize Se + Sp − 1) or for a sensitivity cut-off of at least 75%. * Good-to-moderate diagnostic quality: LQ+ of >3; LQ− of <0.3. * Excellent diagnostic quality: LQ+ of >10 and LQ− of >0.1.

Figure 4

ROC analysis of perioperative MIF levels and acute kidney injury. ROC analysis was performed to evaluate the diagnostic capacities of (A) perioperative MIF levels and (B) MIF levels during ICU admission with regard to acute kidney injury (AKI) based on a serum creatinine increase of >50% within 48 h after ICU admission. If an elevated MIF value indicates AKI, the ROC curve should be farther from the bisecting line (Sensitivity = 1 − Specificity). (C) Sensitivity (Se), specificity (Sp), likelihood ratios (LR +/−), and area under the curve (AUC) are reported for either the Youden optimal cut-off (maximize Se + Sp − 1) or for a sensitivity cut-off of at least 75%. * Good-to-moderate diagnostic quality: LQ+ of >3 and LQ− of <0.3. * Excellent diagnostic quality: LQ+ of >10 and LQ− of >0.1.