Clinical characteristics and factors associated with long COVID among post-acute COVID-19 clinic patients in Zambia, August 2020 to January 2023: A cross-sectional and longitudinal study design.

Introduction: A number of seroprevalence studies in Zambia document the extent of spread of acute SARS-CoV-2 infection, yet knowledge gaps still exist on symptoms and conditions that continue or develop after acute COVID-19 (long COVID). This is an important gap given the estimated prevalence of long COVID in other African countries. We assessed factors associated with long COVID at the initial visit to a post-acute COVID-19 (PAC-19) clinic and longitudinally among a cohort of patients with ≥2 review visits.

Methods: We implemented a cross-sectional and longitudinal analysis of PAC-19 clinic patients from Aug-2020 to Jan-2023. The study outcome was long COVID; defined as the presence of new, relapsing, or persistent COVID-19 symptoms that interfere with the ability to function at home or work. Explanatory variables were demographic and clinical characteristics of patients which included sex, age group, presence of new onset medical conditions, presence of pre-existing comorbidities, vaccination status and acute COVID-19 episode details. We fitted logistic and mixed effects regression models to assess for associated factors and considered statistical significance at p<0.05.

Results: Out of a total 1,359 PAC-19 clinic patients in the cross-sectional analysis, 548 (40.3%) patients with ≥2 PAC-19 clinic visits were in the longitudinal analysis. Patients' median age was 53 (interquartile range [IQR]: 41-63) years, 919 (67.6%) were hospitalized for acute COVID-19, and of whom 686 (74.6%) had severe acute COVID-19. Overall, 377 (27.7%) PAC-19 clinic patients had long COVID. Patients with hospital length of stay ≥15 days (adjusted odds ratio [aOR]: 5.37; 95% confidence interval [95% CI]: 2.99-10.0), severe acute COVID-19 (aOR: 3.22; 95% CI: 1.68-6.73), and comorbidities (aOR:1.50; 95% CI: 1.02-2.21) had significantly higher chance of long COVID. Longitudinally, long COVID prevalence significantly (p<0.001) declined from 75.4% at the initial PAC-19 visit to 26.0% by the final visit. The median follow-up time was 7 (IQR: 4-12) weeks.

Conclusion: Factors associated with long COVID in Zambia were consistent both cross-sectionally at the initial visit to PAC-19 clinics and longitudinally across subsequent review visits. This highlights the importance of ongoing monitoring and tailored interventions for patients with comorbidities and severe COVID-19 to mitigate the long-term impacts of COVID-19.