Background
Sepsis remains a leading cause of death worldwide, with higher burden and diagnostic challenges in low-resource settings. In sub-Saharan Africa, limited ICU capacity and lab infrastructure complicate timely risk stratification. This study evaluates whether low-cost white blood cell (WBC) indices and procalcitonin (PCT) can predict mortality among adults with suspected sepsis admitted to two national referral hospitals in Rwanda, offering a pragmatic approach for context-appropriate triage.
Study design and setting
In this prospective cohort, adult patients (≥18 years) with suspected sepsis, identified using Sepsis-2 criteria, were enrolled from Butare University Teaching Hospital (CHUB) and Rwanda Military Referral and Teaching Hospital (RMRTH). A total of 125 patients were followed for 40 days post-ICU admission. Biomarkers were measured within 24 hours of admission: PCT in serum and a complete blood count (CBC) to derive neutrophil counts, lymphocytes, monocytes, and the neutrophil-to-lymphocyte ratio (NLR).
Biomarkers and analytic approach
The study focused on three low-cost metrics with broad availability in resource-limited ICUs: neutrophil count (WBCNeut), total white blood cell count (TotalWBC), and the neutrophil-to-lymphocyte ratio (NLR), alongside procalcitonin (PCT). Natural log transformations were applied to skewed biomarkers, and Cox proportional hazards models assessed independent associations with ICU mortality. Time-dependent area under the curve (AUC) analyses evaluated short-term discriminative power at days 6, 10, and 15.
Key findings
Critically, elevated neutrophil counts emerged as the strongest independent predictor of mortality after adjustment for age and sex (adjusted hazard ratio ~2.0 per threefold increase in neutrophils, p<0.001). The NLR also demonstrated strong discriminatory ability, with higher ratios observed among non-survivors (median NLR 12.5 vs 4.0, p<0.001). While PCT levels were significantly higher in non-survivors, PCT did not retain independent significance in the final multivariable model when neutrophil metrics were included, suggesting that neutrophil-centric indices capture core inflammatory risk in this setting. Time-dependent ROC analysis showed that models incorporating neutrophil count or total WBC alongside clinical factors yielded the best early discrimination (AUC ~0.68–0.71 at days 6–10), underscoring the value of simple hematology markers for rapid risk stratification in the ICU.
Implications for triage and management
In Rwanda’s resource-constrained ICUs, using neutrophil counts and NLR—readily available from routine CBCs—could guide escalation of care, antibiotic stewardship, and timely initiation of supportive therapies. Combining these markers with clinical data and, where feasible, PCT, may enhance early identification of high-risk patients who require aggressive management, even before advanced diagnostics are accessible.
Context and interpretation
The study aligns with global evidence linking neutrophilia and elevated NLR to poorer sepsis outcomes, while highlighting regional immunologic patterns. Notably, survivors tended to retain preserved lymphocyte counts, contrasting some U.S.-based findings of lymphopenia as a mortality predictor. These differences warrant further exploration into population-specific sepsis phenotypes and the role of comorbidities such as HIV in shaping immune response and prognosis.
Limitations and future directions
Limitations include a modest sample size and the exclusion of certain immunocompromised groups, which may affect generalizability. Future research should validate these findings across diverse African ICU settings, with emphasis on integrating cost-effective biomarker panels into validated, locally relevant scoring systems such as Kigali-based sepsis scores to optimize triage and resource allocation.
Conclusion
Among adults with suspected sepsis in Rwanda, simple WBC indices—particularly neutrophil counts and NLR—provide meaningful, independent prognostic information for ICU mortality. PCT adds value in risk stratification but did not independently predict outcomes when neutrophil metrics were considered. Implementing these low-cost biomarkers could support timely, evidence-informed care in resource-limited African ICUs.
