Thinking fast and slow: lactate and MELD-XI (model for end-stage liver disease excluding INR) are useful for estimating mortality after cardiopulmonary resuscitation

Abstract

Background: Intensivists keep getting challenged with prognostication after cardiopulmonary resuscitation (CPR). The model for end-stage liver disease, excluding INR (MELD-XI) Score has proven valuable for assessing illness severity. Serum lactate is a readily available and established indicator of general stress and tissue hypoxia. We aimed to evaluate the prognostic value of MELD-XI combined with serum lactate in patients after CPR.

Methods: A retrospective analysis on 106 patients after CPR was performed. Multivariable Cox regression was performed to evaluate associations with 30-day mortality and neurological outcome by means of cerebral performance category (CPC). An optimal cut-off was calculated by means of the Youden Index. Patients were then divided into subgroups based on the optimal cut-offs for MELD-XI and serum lactate.

Results: MELD-XI and lactate were independently associated with mortality. The respective cut-offs were MELD-XI>12 and lactate ≥2.5 mmol/L. Patients were split into three groups: lactate <2.5 mmol/L and MELD-XI≤12 (low-risk; N.=32), lactate ≥2.5 mmol/L or MELD-XI>12 (medium-risk; N.=39), and lactate ≥2.5 mmol/L and MELD-XI >12 (high-risk; N.=33). The mortality rates were 6%, 26% and 61% in the low, medium and high-risk group. This combined model yielded in the highest predictive abilities (AUC=0.78 95%CI: 0.68-0.85; P=0.03 vs. AUC=0.66 for SOFA Score). Worse neurological outcome (CPC 3 or 4) was more common in the medium and high-risk group (6.25%, 10.3% and 9.1%).

Conclusions: The combination of MELD-XI and lactate concentration at ICU admission was superior to the more complex SOFA Score for prediction of mortality after CPR.