Neutrophil-to-Lymphocyte Ratio as a Predictor of Mortality for COVID-19-Related Acute Respiratory Distress Syndrome (ARDS) Patients Requiring Extracorporeal Membrane Oxygenation Therapy.

BACKGROUND

The neutrophil-to-lymphocyte ratio (NLR) has been studied as an indicator of systemic inflammation and as a prognostic tool in multiple areas of medicine. Previous research has suggested that higher NLR and rapid increase to peak NLR are associated with poorer outcomes in patients with coronavirus disease 2019 (COVID-19), particularly in those experiencing acute respiratory distress syndrome (ARDS). Within vascular surgery, there is data to suggest a positive correlation between elevated pre-extracorporeal membrane oxygenation (ECMO) NLR and higher rates of mortality following major procedures. This study explores the prognostic value of peri-ECMO NLR in patients requiring veno-venous ECMO (VV-ECMO) therapy for COVID-19-related ARDS. The objective of this study was to explore the utility of pre-ECMO NLR as an easily accessible prognostic factor for patients suffering from COVID-19-associated ARDS that require VV-ECMO.

METHODS

This was a retrospective cohort study within a tertiary care hospital conducted between April 2020 and January 2021. Patients requiring VV-ECMO therapy for COVID-19-associated ARDS were included. Peri-ECMO NLR values, length of stay (LOS), duration on VV-ECMO, and discharge status were recorded. Receiver operating characteristic (ROC) curve analysis and Youden's J statistics were performed to calculate a cut-off value of 11.005 for pre-ECMO NLR and 17.616 for on-ECMO NLR. Pre-ECMO and on-ECMO Kaplan-Meyer curves were generated for two groups of patients, those above and below NLR cutoff thresholds. Two-sample T-test was performed to test for significant differences in LOS and duration on VV-ECMO.

RESULTS

Twenty-six patients were included in the study for final analyses. There was an overall mortality of 39% (n = 10). ROC curve analysis and Youden's J statistic revealed an optimal cut-off value of pre-ECMO NLR = 11.005 and on-ECMO NLR = 17.616. Results showed that the patient group placed on VV-ECMO with a pre-ECMO NLR less than 11.005 experienced no mortality (n = 7) and a median LOS of 28 days (IQR = 14.5-64.5 days). The patient group on VV-ECMO with a pre-ECMO NLR greater than 11.005 (n = 19) included all mortality (n = 10) and had a median LOS of 49 days (IQR = 25.5-63.5 days). The patient group with on-ECMO NLR less than 17.616 also conferred a survival advantage. There was no significant difference in LOS or duration on VV-ECMO between the two groups, pre-ECMO or on-ECMO.

CONCLUSIONS

A pre-ECMO NLR cutoff was identified and offered statistically significant prognostic value in predicting mortality. A lower on-ECMO NLR value also indicated a survival advantage. Future studies should include NLR within multivariate models to better discern the effect of NLR and elucidate how it can be factored into clinical decision-making. Importantly, this data can be expanded to assess the predictive value of NLR pertaining to the COVID-19-induced ARDS population and matched cohorts.