suPAR ≥6 ng/mL for Predicting Severe Respiratory Failure in U.S. Adults Hospitalized With COVID 19 (suPAR US The Michigan Cohort Heterogeneity Study)

The Michigan Medicine COVID-19 Cohort (M2C2) is the largest sub-cohort of the International Study on Inflammation in COVID-19 (ISIC). The M2C2 comprises consecutive, systematically enrolled adults (≥18 years) with confirmed SARS-CoV-2 infection hospitalized specifically for COVID-19 at the University of Michigan from 1 February 2020 to 1 June 2021. Adult patients hospitalized in participating U.S. hospitals with confirmed COVID 19 infection during the study period, who had baseline suPAR measured using the suPARnostic TurbiLatex assay on Roche cobas c501 on plasma samples obtained within 48 hours of admission. The cohort reflects real world U.S. data and includes racially and ethnically diverse populations with typical U.S. burdens of obesity, diabetes, and chronic kidney disease.

SAMPLE SIZE JUSTIFICATION - Since we have a fixed 6 ng/mL threshold and are only validating (not discovering), the analysis is just a 2×2 table.

True sensitivity 94% (matching SPARCOL): N=136 is enough True sensitivity 90% (conservative): N=237 is enough True sensitivity 88% (worst case): N=440 needed SPARCOL showed 93.9%, so N=300 covers you even if U.S. sensitivity drops to ~88% - a generous safety margin.

STATED LIMITATIONS

• N=300 does not support fully adjusted multivariable logistic regression
• Hispanic and Asian subgroups are too small for standalone powered analyses. These subgroups are reported descriptively.
• Formal non-inferiority testing of sensitivity (U.S. vs. SPARCOL) would require a larger sample. The comparison is performed descriptively, with the acceptance criterion applied to the U.S. data independently (lower 95% CI > 80%).