Factors associated with in-hospital mortality during the 2023 cholera outbreak in Hammanskraal, South Africa: a cross-sectional study.

backgroundThe 2023 cholera outbreak centred on Hammanskraal, a peri-urban township north of Pretoria, South Africa, resulted in high in-hospital mortality. Hammanskraal had experienced chronic failures in water and sanitation infrastructure for nearly two decades before the outbreak. We aimed to describe the clinical profile of patients managed at the principal case-management facility and to identify factors associated with in-hospital death.

methodsWe conducted a retrospective cross-sectional study based on review of medical records and the hospital cholera admission register at a district hospital in Hammanskraal, covering the period 1 May to 30 June 2023. All patients fulfilling the World Health Organization (WHO) operational case definition for suspected cholera were eligible. There were no comparison arms. Univariate odds ratios (ORs) and multivariable logistic regression were used to examine factors associated with in-hospital mortality. A sensitivity analysis was performed excluding laboratory confirmation as a potential disease-severity proxy. The events-per-variable (EPV) ratio was 4.7; all multivariable findings are considered exploratory.

resultsOf 431 suspected cases, 169 met inclusion criteria (mean age 39.8 years [standard deviation 20.4]; 51.5% female). Twenty-eight patients died, yielding a case fatality ratio (CFR) of 16.6%. Median time from presentation to death was 2 days (range 1-9); 25% of deaths occurred within 24 h. On multivariable logistic regression (n = 166), three variables were independently associated with in-hospital death: age ≥ 40 years (adjusted OR [aOR] 8.06; 95% confidence interval [CI] 2.30-28.21; p = 0.001), laboratory-confirmed Vibrio cholerae O1 or O139 as a disease-severity marker (aOR 9.42; 95% CI 2.65-33.42; p < 0.001), and pre-existing chronic illness (aOR 3.42; 95% CI 1.23-9.53; p = 0.019). In the sensitivity analysis, dehydration on admission was additionally associated with death (aOR 2.82; 95% CI 1.07-7.45; p = 0.036). The observed case fatality proportion among patients with all three risk factors present (age ≥ 40 years, chronic illness, and laboratory-confirmed disease) was 71.4% (10/14), compared with 6.8% (6/88) among patients aged < 40 years without chronic illness, illustrating the clinical gradient of mortality risk within this series.

conclusionsA CFR of 16.6% substantially exceeds the WHO adequately-treated benchmark of < 1%, though this comparison warrants caution: the study hospital received referral-level presentations, and the 28 in-hospital deaths captured represent approximately 60% of all 47 deaths reported nationally during the outbreak, suggesting the sample disproportionately reflects severe outcomes. Older age, chronic comorbidities, and severity markers were independently associated with in-hospital death; given the cross-sectional design, these associations inform clinical risk stratification rather than establishing causality. Prioritised triage of patients aged ≥ 40 years with comorbidities, enhanced rehydration capacity at first contact, and sustained investment in water and sanitation infrastructure are priorities for future outbreak responses in peri-urban South Africa, pending prospective confirmation.