Validation of a prospective urinalysis-based prediction model for ICU-resources and outcome of Covid-19 disease: A multicenter cohort study

Abstract Purpose: Identifying preventive strategies in Covid-19 patients helps to improve ICU-resource-allocation and reduce mortality. We recently demonstrated in a post-mortem cohort that SARS-CoV-2 renal tropism was associated with kidney injury, disease severity and mortality. We also proposed an algorithm to predict the need for ICU-resources and the risk of adverse outcomes in Covid-19 patients harnessing urinalysis and protein/coagulation parameters on admission for signs of kidney injury. Here, we aimed to validate this hypothesis in a multicenter cohort. Methods: Patients hospitalized for Covid-19 at four tertiary centers were screened for an available urinalysis, serum albumin (SA) and antithrombin-III activity (AT-III) obtained prospectively within 48h upon admission. The respective presumed risk for an unfavorable course was categorized as “low”, “intermediate” or “high”, depending on a normal urinalysis, an abnormal urinalysis with SA ≥2 g/dl and AT-III ≥70%, or an abnormal urinalysis with at least one SA or AT-III abnormality. Time to ICU or death within ten days served as primary, in-hospital mortality and required organ support served as secondary endpoints.Results: Among a total of N=223 screened patients, N=145 were eligible for enrollment, falling into the low (N=43), intermediate (N=84), and high risk (N=18) categories. The risk for ICU transfer or death was 100% in the high risk group and significantly elevated in the composite of high and intermediate risk as compared to the low risk group (63.7% vs. 27.9%; HR 2.6; 95%-CI 1.4 to 4.9; P=0.0020). Having an abnormal urinalysis was associated with mortality, need for mechanical ventilation, extra-corporeal membrane oxygenation (ECMO) or renal replacement therapy (RRT). Conclusion: Our data confirm that Covid-19-associated urine abnormalities on admission predict disease aggravation and need for ICU. By engaging a simple urine dipstick on hospital admission our algorithm allows for early preventive measures and appropriate patient stratification. (ClinicalTrials.gov number NCT04347824)