Introduction: Sepsis in resource-limited settings
Sepsis remains a leading cause of mortality worldwide, with sub-Saharan Africa facing unique challenges due to limited ICU capacity and diagnostic resources. In Rwanda, where ICU beds are scarce and advanced laboratory testing is not universally accessible, practical biomarkers are essential for early risk stratification and targeted therapy. This prospective ICU study investigates whether simple, inexpensive biomarkers—neutrophil counts, neutrophil-to-lymphocyte ratio (NLR), and procalcitonin (PCT)—can predict mortality among adults with suspected sepsis.
Study overview
Researchers enrolled 125 adults with suspected sepsis in two national referral hospitals in Rwanda (Butare CHUB and Rwanda Military Hospital). Sepsis-2 criteria based on SIRS were used, and patients were followed for 40 days after ICU admission. Blood samples collected within 24 hours of admission provided PCT levels and a full blood count (CBC). The aim was to determine whether neutrophil counts, NLR, total WBC, and PCT independently or in combination could forecast all-cause mortality during the ICU stay.
Key biomarkers and what they tell us
Among the biomarkers studied, elevated neutrophil counts emerged as a robust, independent predictor of ICU mortality. Specifically, a roughly threefold increase in neutrophils was associated with almost double the risk of death (adjusted hazard ratio ≈ 2.0). The NLR also showed strong discriminative value, with fatal cases exhibiting markedly higher ratios than survivors. Procalcitonin levels were higher in non-survivors, including very high values in those who died early, reinforcing PCT’s role as a marker of sepsis severity. However, in multivariable models, PCT did not retain independent significance, suggesting its predictive utility may be enhanced when combined with WBC-derived indices.
Modeling and short-term discrimination
The study used Cox proportional hazards models, time-to-event analyses, and time-dependent ROC analyses to assess predictive performance. Models that included neutrophil counts or total WBC alongside clinical data consistently achieved higher short-term discrimination (AUCs around 0.68–0.71 within the first 10–15 days). The best-performing models combined PCT with neutrophil counts and/or NLR, suggesting a synergistic approach could optimize early risk stratification in ICU settings where resources are constrained.
Clinical implications for Rwanda and similar settings
Given the constraints on costly diagnostics, neutrophil counts and NLR—routinely available from CBC—and PCT (when accessible) offer a pragmatic means to identify high-risk patients upon ICU admission. Implementing biomarker-guided triage could prioritize escalated care, earlier antibiotic optimization, and targeted supportive therapies, potentially reducing mortality. The study also highlights that age and sex were not independently predictive in the final models, underscoring the value of dynamic biomarker information over demographic factors in this context.
Limitations and future directions
While informative, the study reflects data from two hospitals using Sepsis-2 criteria and may not generalize to all sub-Saharan populations. Excluding advanced HIV/AIDS patients and the specific local epidemiology could influence biomarker performance. Future multicenter studies, adopting Sepsis-3 criteria where feasible and exploring immune response patterns in African populations, are warranted to refine risk-stratification tools.
Concluding remarks
In resource-limited ICUs, simple, low-cost haematologic indices alongside PCT present a practical strategy for mortality risk prediction in sepsis. Neutrophil counts and NLR demonstrated robust prognostic value, with PCT adding incremental information when used in combination. This approach can support timely clinical decisions, optimize antibiotic stewardship, and guide resource allocation in Rwanda and similar settings.
