Trajectories of UCR as a measure of biological mechanisms underlying ethnicity associated outcomes in COVID-19

ABSTRACT

Introduction: Patients with Black and Asian ethnic background have been disproportionately affected by COVID-19 with increased disease severity, organ failure, intensive care admission, and premature mortality. 1-3 The urea-to-creatinine ratio (UCR) has been described as a biochemical signature of persistent critical illness, its hallmark catabolic state and late mortality during prolonged ICU stay.4 Low serum creatinine reflecting reduced muscle mass, which declines rapidly in acute severe illness in combination with net muscle protein breakdown which contribute substrate for increased hepatic urea synthesis, results in markedly elevated UCR. Objectives: To assess UCR as a candidate biological feature driving ethnicity associated outcomes of COVID-19 disease. Methods: Prospective analysis using registry data from all patients aged ≥16 years with an emergency admission to hospitals within Barts Health NHS Trust with SARSCoV-2 infection during 1 January 2020 - 13 May 2020 (wave 1), and 1 September 2020 -17 February 2021 (wave 2). Trajectories of routine haematology and clinical biochemistry blood results during hospital admission were assessed, and distinct phenotypes defined using unsupervised longitudinal clustering techniques using day 0 to 15 results.We determined distribution of identified phenotypes within patients categorised by ethnic group. Multivariable logistic regression accounting for predefined baseline risk factors was used to assess association between ethnicity, phenotypes, and 30-day mortality. All analyses were performed using R software v4.02 and the kml package for clustering. 5 Results are presented as n (%) and adjusted odds ratios (OR) with 95% confidence intervals. Results: We assessed 459 (wave 1) and 1337 (wave 2) patients after excluding those with unknown ethnicity and those with <7 blood results. Three clusters were identified based on trajectories of UCR. In wave 1, 48.1% of patients had persistently low levels of UCR (A), 38.6% had higher but stable levels (B), and 13.3% had the highest levels peaking after day 7 (C). In wave 2, three clusters were identified in similar proportions 42.8% (a), 45.1% (b), 12.1% (c). In wave 1, patients in cluster C compared to A had the highest risk of death at 30 days (OR 4.59 [2.27-9.26], p<0.001). In wave 2, both clusters b (OR 1.58 [1.18-2.12], p< 0.001) and c (OR 3.96 [2.62-5.99], p<0.001) had higher risk of death compared to a. Distribution of cluster membership varied by ethnic category. In both waves, greater proportions of patients within cluster A/a were observed in patients with Black ethnicity (65.5% wave 1, 61.1% wave 2) compared to Asian (50.0% wave 1, 37.3% wave 2) and White (39.7% wave 1, 39.6% wave 2) ethnicity. Black ethnicity patients also had lowest proportions in cluster C/c (6.9% wave 1, 6.3% wave 2) compared to Asian (17.4% wave 1, 14.2% wave 2) and White (13.2% wave 1, 12.9% wave 2) ethnicity. Inclusion of UCR trajectory attenuated the higher risk of death seen in Asian patients in wave 1. Conclusion: Phenotypes based on UCR trajectories during hospital admission are associated with adverse outcomes following COVID-19 infection. Further work is needed to understand phenotypes of prolonged COVID-19 disease muscle wasting and its association with longerterm outcomes.