Background: To determine the predictive value of mid-regional pro-adrenomedullin (MR-proADM) compared to routine clinical and laboratory parameters in patients with COVID-19.
Methods: A total of 135 adult patients hospitalized with COVID-19 were included in a prospective single-centre study. In addition to routine parameters, the levels of MR-proADM in blood plasma were measured on the day of hospitalization. The patients were divided into 2 groups: those who survived and were discharged (n = 115, 85%) and those who did not survive (n = 20, 15%). Data are presented as median and interquartile range.
Results: The non-survivors had a statistically significantly greater age (73.4 [63.5-84.8] vs. 62.2 [50.3-71.4] years, P = 0.001), a lower level of haemoglobin oxygen saturation (91 [87-92] vs. 92 [92-93]%, P < 0.001), lower lymphocyte level (13 [7-30] vs. 21 [15-27]%, P = 0.03), higher lactate dehydrogenase (338 [273-480] vs. 280 [233-383] EU L-1, P = 0.04) and aspartate aminotransferase levels (49 [28-72] vs. 33 [23-47] EU L-1, P = 0.03), a higher National Early Warning (NEWS) score (7 [7- 8] vs. 6 [5-7] points, P < 0.001), and higher procalcitonin (0.16 [0.11-0.32] vs. 0.1 [0.07-0.18] ng mL-1, P = 0.006) and MR-proADM levels (1.288 [0.886-1.847] vs. 0.769 [0.6-0.955] nmol L-1, P < 0.001). MR-proADM had the highest predictive value for death during hospital stay (cut-off: 0.895 nmol L-1, AUC ROC 0.78 [95% CI: 0.66-0.90], sensitivity 75%, specificity 69%, OR 6.58 [95% CI: 2.22-19.51]).
Conclusions: Compared with other indicators, MR-proADM has the highest predictive value for in-hospital mortality in patients with COVID-19.
Keywords: COVID-19; mortality; predictors; pro-adrenomedullin; biomarkers.