The Combined Use of Cytokine Serum Values with Laboratory Parameters Improves Mortality Prediction of COVID-19 Patients: The Interleukin-15-to-Albumin Ratio

Abstract

Laboratory parameters display limited accuracy in predicting mortality in coronavirus disease 2019 (COVID-19) patients, as with serum albumin. Emerging evidence suggests that cytokine serum values may enhance the predictive capacity of albumin, especially interleukin (IL)-15. We thus investigated whether the use of the IL-15-to-albumin ratio enables improving mortality prediction at hospital admission in a large group of COVID-19 patients. In this prospective cross-sectional study, we enrolled and followed up three hundred and seventy-eight patients with a COVID-19 diagnosis until hospital discharge or death. Two hundred and fifty-five patients survived, whereas one hundred and twenty-three died. Student's T-test revealed that non-survivors had a significant two-fold increase in the IL-15-to-albumin ratio compared to survivors (167.3 ± 63.8 versus 74.2 ± 28.5), a difference that was more evident than that found for IL-15 or albumin separately. Likewise, mortality prediction considerably improved when using the IL-15-to-albumin ratio with a cut-off point > 105.4, exhibiting an area under the receiver operating characteristic curve of 0.841 (95% Confidence Interval, 0.725-0.922, p < 0.001). As we outlined here, this is the first study showing that combining IL-15 serum values with albumin improves mortality prediction in COVID-19 patients.

Keywords: COVID-19; IL-15; SARS-CoV-2; albumin; mortality; prognosis.

Figure 1

Schematic flow chart illustrating the selection process of eligible patients. SpO2, peripheral oxygen saturation; SARS-CoV-2, severe acute respiratory syndrome coronavirus-2; qPCR, quantitative polymerase chain reaction; HIV, human immunodeficiency virus; HCV, hepatitis C virus; HBV, hepatitis B virus; COVID-19, coronavirus disease 2019.

Figure 2

Serum levels of IL-15 and albumin in COVID-19 patients. (A) Serum albumin levels in survivors and non-survivors. (B) Serum IL-15 levels in survivors and non-survivors. (C) The IL-15-to-albumin ratio in survivors and non-survivors. The IL-15-to-albumin ratio resulted from dividing IL-15 serum values by serum albumin. Data are presented as mean ± standard deviation. We considered a p value < 0.05 as significant.

Figure 3

The area under the receiver operating characteristic curves for albumin, IL-15, and the IL-15-to-albumin ratio. (A) AUROC for albumin. (B) AUROC for IL-15. (C) AUROC for the IL-15-to-albumin ratio. We considered a p value < 0.05 as significant. AUROC, area under the receiver operating characteristic curves; AUC, area under the curve; CI, confidence interval.

Figure 4

IL-15 serum values improved the area under the receiver operating characteristic curves of clinical and laboratory parameters. (A) AUROCs for oxygen saturation, the neutrophil count, blood urea, AST, and CRP. (B) AUROCs for the IL-15-to-oxygen saturation ratio, IL-15-to-neutrophil ratio, IL-15-to-urea ratio, IL-15-to-AST ratio, and IL-15-to-CRP ratio. We considered a p value < 0.05 as significant. AUROC, area under the receiver operating characteristic curves; AUC, area under the curve; CI, confidence interval; AST, aspartate aminotransferase; CRP, C-reactive protein.