Summary
This retrospective cohort study used US National COVID Cohort Collaborative (N3C) data to assess whether early sotrovimab treatment, compared with no treatment, was associated with post-acute sequelae of COVID-19 (PASC, long COVID) in high-risk patients with COVID-19. The main analysis included 9,504 sotrovimab-treated and 619,668 untreated patients. The observation period ran from day 30 following diagnosis until the earliest of day 365 following diagnosis, loss to follow-up, end of data availability, death, COVID-19 re-infection, or initiation of another mAb, antiviral, or pre-exposure prophylaxis. PASC was defined as at least one PASC symptom at least 30 days after the first COVID-19 diagnosis. PASC occurred in 6,217 (65.4%) treated patients and 324,289 (52.3%) untreated patients. The unweighted HR was 1.42 (95% CI 1.38–1.46), p < 0.001. The doubly robust ATT-weighted HR was 0.92 (0.89–0.96), p < 0.001. With administrative censoring, the HR was 0.95 (0.92–0.98), p < 0.001. With a more stringent PASC definition, the HR was 0.89 (0.74–1.07), p = 0.21, not statistically significant. The adjusted HR from multivariable Cox regression was 0.96 (0.94–0.99), p < 0.001. Subgroup HRs were 0.92 (0.89–0.95) for adults and 0.94 (0.91–0.98) for the highest-risk subgroup; for hospitalization and/or ICU admission during the acute phase, the HR was 1.14 (0.91–1.44), p = 0.25, with no significant difference between cohorts. Harms and withdrawals for harms were not reported. The authors state limitations including observational EHR data, residual or unmeasured confounding, channeling bias, immortal time bias, and heterogeneity in PASC definitions.