Parametric Survival Analysis of Long COVID Among Hospitalized Patients in Zambia: A Retrospective Cohort Study on the Time to Symptoms Resolving

Long COVID refers to the continuation or development of new symptoms within three months after acute SARS-CoV-2 infection, with these symptoms lasting for at least 2 months. Despite studies on COVID-19 sequelae, gaps remain in understanding the temporal dynamics of COVID-19 symptom resolving. This knowledge is crucial for treatment planning and setting realistic recovery expectations. Using a cohort of patients hospitalized in Zambia for COVID-19, the objective of this study was to evaluate resolution of COVID-19-related symptoms over time and associated factors.

Methods

We conducted a retrospective cohort study of persons discharged after COVID-19 hospitalization and presenting for follow-up care in 13 specialized clinics in Zambia from August-2020 to December-2022. All participants were hospitalized with acute COVID-19. Severe acute COVID-19 was defined as hospitalization further requiring supplemental oxygen therapy, ICU stay, and/or treatment with steroids/remdesivir. We evaluated time-to-symptoms resolving (i.e., survival time) as our primary outcome. We estimated symptoms resolution incidence rate, median survival time (onset-to-resolution), associated factors, and changes in the hazard of symptoms resolving using survival analysis.

Results

Of the 823 study participants, 616 (84.3%) had severe acute COVID-19 illness, half (50.6%) were female, and median age was 54 (interquartile range [IQR]: 43-64) years. Nearly three-quarters (597, 72.5%) had their symptoms resolved at a median survival time of 51 (IQR: 34-104) days. A majority of participants (59.4%) had baseline comorbidities, 16.6% had prior vaccination with ≥1 COVID-19 vaccine dose, and median acute hospitalization duration was 8 (IQR: 4-16) days. Persistent COVID-19 symptoms resolved at a rate of 12.2 per 1,000 person-days. Severe acute COVID-19 (adjusted hazard ratio [aHR]: 0.65, 95% CI: 0.41-0.88) and having ≥1 comorbidity (adjust time ratio [aTR]: 1.86, 95% CI: 1.60-2.17) was associated with slower resolution. Infection during the Omicron-predominant period compared to Alpha (aHR: 4.05, 95% CI: 1.27-12.9) and prior vaccination before illness (aTR: 0.50; 95% CI: 0.26-0.94) were associated with faster symptom resolution. The changes in the hazard rate of symptoms resolving was non-linear, it instead increased to a peak rate of 2.14% per day at 20 days and then decreased.

Conclusion

An increasing then decreasing hazard rate trajectory meant that COVID-19 symptoms resolved relatively faster in the first month. The median survival time occurring nearly a month after peak hazard rate could suggest some patients who do not initially improve in the first month of their post-acute infection were less likely to resolve afterwards. This may likely help in treatment planning and in providing persons with persistent COVID-19 symptoms realistic expectations about their recovery process.