Factors Associated with Adverse Covid-19 Outcomes at Levy Mwanawasa University Teaching Hospital

Background COVID-19 caused significant morbidity and mortality in Zambia, where the dual burden of HIV and non-communicable diseases may amplify severe outcomes. Local evidence on biological predictors of adverse COVID-19 outcomes in Zambian tertiary hospitals remains limited. This study aimed to identify factors associated with adverse COVID-19 outcomes among patients admitted to Levy Mwanawasa University Teaching Hospital. Methods A retrospective cross-sectional study was conducted using medical records of 246 adults with laboratory-confirmed COVID-19 admitted to Levy Mwanawasa University Teaching Hospital (LMUTH) between March 2020 and December 2022. The composite adverse outcome comprised ICU admission, mechanical ventilation, or in-hospital death. Predictors included age, sex, comorbidities (hypertension, diabetes, HIV, chronic kidney disease, obesity), and admission biomarkers (oxygen saturation [SpO₂], C-reactive protein [CRP], D-dimer, lymphocyte count). Firth penalized-likelihood multivariable logistic regression was performed, and biomarker discrimination was assessed by ROC curve analysis. All analyses were conducted in R version 4.5. Results Of 246 patients, 106 (43.1%; 95% CI: 37.1–49.3%) experienced an adverse outcome. Crude prevalence was 100% in patients aged 65–79 years, those with HIV (100%), chronic kidney disease (100%), and a comorbidity score of 3 (100%), compared with 0% in those aged 18–39. Hypertension was present in 88.7% of those with adverse outcomes versus 21.4% without (p < 0.001), and diabetes in 60.4% versus 7.9% (p < 0.001). The adverse outcome group had lower mean SpO₂ (85.76 vs. 94.67), higher CRP (110.97 vs. 48.31 mg/L), higher D-dimer (2395.09 vs. 959.04 ng/mL), and lower lymphocyte count (0.76 vs. 1.45 ×10⁹/L; all p < 0.001). In multivariable analysis, CRP independently predicted adverse outcome (adjusted OR [aOR]: 1.31, 95% CI: 1.03–1.66; p = 0.026), as did each additional hospital day (aOR: 2.87, 95% CI: 1.32–6.26; p = 0.008). Age and sex were not independently predictive after adjustment. Conclusion Adverse COVID-19 outcomes at LMUTH were highly prevalent (43.1%) and concentrated in patients with cardiometabolic disease, HIV, and kidney disease. Systemic inflammation, as measured by CRP, was the strongest independent biological predictor. Integrating CRP-guided triage, robust chronic disease management, and vigilant HIV care into COVID-19 protocols is recommended for similar resource-limited settings.