Diabetes–hypertension multimorbidity and admission hypoxaemia marked every in-hospital COVID-19 death during the Delta wave in a Lagos isolation unit: a retrospective case series

Abstract Background : Patient-level clinical data from sub-Saharan Africa on hospitalised COVID-19 during the Delta-driven third wave are limited. We describe the characteristics, management and outcomes of consecutive patients admitted to a dedicated COVID-19 isolation unit in Lagos, Nigeria, and identify bedside features associated with in-hospital death. Methods : Retrospective case series of all consecutive patients admitted to the COVID-19 isolation unit of the Federal Medical Centre, Ebute-Metta, Lagos, between 1 August and 30 September 2021. Infection was confirmed by rapid antigen testing and RT-PCR. Admission demographics, comorbidities, room-air oxygen saturation (SpO₂), full blood count, in-hospital interventions and outcome were abstracted from the unit register. Survivors and non-survivors were compared using exact tests. Results : Fifteen patients were included (median age 57 years, range 16–81; 10 male); nine had at least one comorbidity. In-hospital mortality was 26.7% (4/15; 95% CI 7.8–55.1). All four deaths occurred in patients with coexisting diabetes and hypertension (4/4 vs 0/11 without this combination; p = 0.001), all of whom were hypoxaemic on admission (SpO₂ 84–89%). Mortality was 80% (4/5) among patients with admission SpO₂ <90% and 0% (0/10) among those with SpO₂ ≥90% (p = 0.004). Non-survivors had a lower lymphocyte fraction (median 8.6% vs 27.3%; p = 0.040) and a higher neutrophil-to-lymphocyte ratio (10.8 vs 2.4; p = 0.056). Dexamethasone, ivermectin and vitamin supplements were each given to 11 of 14 patients with treatment data. Conclusions : In this small series, every in-hospital COVID-19 death occurred in patients who presented with diabetes–hypertension multimorbidity and hypoxaemia. Pulse oximetry and a full blood count at admission may help identify a high-risk phenotype for early escalation in resource-limited settings. The frequent use of therapies without proven benefit underlines the need for evidence-aligned prescribing in future epidemic responses. These hypothesis-generating findings require confirmation in larger, multicentre cohorts.