The study’s primary objective was to evaluate how well admission procalcitonin levels could distinguish patients with confirmed bacterial infections from those without. The results were sobering. The area under the receiver operating characteristic curve, a standard measure of diagnostic discrimination where 1.0 represents perfect accuracy and 0.5 represents no better than chance, was just 0.56 with a 95 percent confidence interval of 0.45 to 0.68. At the optimal cutoff of 0.465 nanograms per milliliter, the test achieved a sensitivity of only 57 percent and a specificity of 60 percent. In practical terms, this means that procalcitonin measured at admission missed more than four in ten true bacterial infections while simultaneously misclassifying a substantial proportion of patients without bacterial infection as positive.

This poor diagnostic performance stands in stark contrast to the biomarker’s prognostic capabilities. When the researchers assessed procalcitonin’s ability to identify patients meeting the study-defined quick Sequential Organ Failure Assessment, or qSOFA, criteria for sepsis, the results were markedly different. The area under the curve rose to 0.83 with a confidence interval of 0.738 to 0.928, indicating good discrimination. The optimal cutoff for identifying sepsis was 2.65 nanograms per milliliter, yielding a sensitivity of 82 percent and a specificity of 85 percent. These figures suggest that while procalcitonin may struggle to confirm the presence of bacterial infection itself, it performs considerably better at flagging patients whose infections have progressed to a more dangerous systemic state.

Procalcitonin is a peptide precursor of the hormone calcitonin, produced primarily by thyroid C cells under normal conditions. However, during systemic bacterial infections, its expression is upregulated in extra-thyroidal tissues, particularly the liver, in response to pro-inflammatory cytokines such as interleukin-1 beta, tumor necrosis factor-alpha, and interleukin-6. Blood levels typically begin rising within two to four hours of a bacterial insult and peak at six to twelve hours, making it theoretically attractive as an early marker of bacterial rather than viral or non-infectious inflammation. Clinical guidelines in Europe and North America have increasingly incorporated procalcitonin into antibiotic stewardship protocols, using falling levels to justify discontinuing antimicrobial therapy in patients whose condition is improving.

The Ethiopian study also examined whether serial procalcitonin measurements could guide antibiotic discontinuation decisions. Among the 50 participants who underwent repeat procalcitonin testing on day three of hospitalization, a decline in the biomarker was significantly associated with clinician-directed antibiotic discontinuation. Day-three procalcitonin demonstrated excellent discriminatory performance for this outcome, with an area under the curve of 0.926 and both sensitivity and specificity of 90 percent. This finding, while promising, comes with important caveats. The researchers emphasize that the serial procalcitonin analysis was exploratory in nature and limited to a selected subgroup of patients, meaning the results should be interpreted cautiously and require validation in larger, more diverse cohorts.

The overall in-hospital mortality rate among the study participants was 6.7 percent, with 14 of the 210 enrolled patients dying during their hospital stay. In an exploratory multivariable analysis that controlled for potential confounding variables, meeting the qSOFA-based sepsis classification was associated with higher odds of mortality, although the researchers note that this estimate was imprecise due to the small number of deaths in the cohort. This limitation underscores the challenges of conducting prognostic research in settings where patient volumes and follow-up capacity may constrain statistical power.

The study employed a hospital-based longitudinal design, enrolling admitted patients at Jimma Medical Center between May and November 2025. Participants were selected consecutively until the expected sample size was achieved, a method that reduces selection bias compared to convenience sampling. Data collection used structured questionnaires covering sociodemographic characteristics, clinical information, and laboratory parameters. The mean admission procalcitonin level across the cohort was 1.65 nanograms per milliliter with a standard deviation of 2.1, indicating considerable variability in baseline biomarker levels among patients presenting with suspected bacterial infection. Statistical analysis was performed using SPSS version 27, with receiver operating characteristic curves used to determine optimal cutoff values and multivariate regression models employed to adjust for confounders.

The findings carry significant implications for clinical practice in low- and middle-income countries, where bacterial infections remain a leading cause of hospitalization and death. Ethiopia, like many sub-Saharan African nations, faces a high burden of infectious diseases alongside rising rates of antimicrobial resistance. If procalcitonin cannot reliably distinguish bacterial from non-bacterial illness in this population, clinicians may need to rely on other diagnostic strategies, including microbiological cultures, clinical scoring systems, and imaging, to guide initial antibiotic decisions. At the same time, the strong performance of procalcitonin for sepsis identification and antibiotic stewardship suggests the biomarker retains value as a prognostic and monitoring tool rather than a standalone diagnostic test.

The disconnect between diagnostic and prognostic performance observed in this study aligns with a growing body of international literature suggesting that procalcitonin’s utility depends heavily on the clinical question being asked. Previous meta-analyses have reported variable diagnostic accuracy for procalcitonin across different patient populations, infection types, and healthcare settings, with performance generally declining in patients with localized infections, early disease, or conditions that elevate procalcitonin through non-bacterial mechanisms such as trauma, surgery, or severe inflammation. The Ethiopian data add an important geographic dimension to this literature, as most existing evidence derives from European and North American cohorts with different epidemiological profiles, comorbidity burdens, and access to pre-hospital care.

The researchers acknowledge several limitations of their work. The single-center design at Jimma Medical Center may limit generalizability to other healthcare settings in Ethiopia and beyond. The modest sample size of 210 patients, while adequate for the primary diagnostic analysis, constrained the precision of mortality-related estimates and limited the serial procalcitonin subgroup to 50 participants. The study also relied on clinician-directed antibiotic discontinuation as an outcome, which reflects real-world practice patterns but may be influenced by factors beyond biomarker values, including clinical judgment, radiological findings, and microbiological results. Despite these constraints, the study represents one of the most detailed prospective evaluations of procalcitonin performance in a sub-Saharan African hospital population and provides a foundation for larger multi-center validation studies that could inform evidence-based antimicrobial stewardship guidelines tailored to resource-limited settings.

Subject of Research: Diagnostic and prognostic performance of the biomarker procalcitonin in hospitalized adults with suspected bacterial infection in Ethiopia

Article Title: Diagnostic and prognostic performance of procalcitonin among hospitalized adults with suspected bacterial infection in Southwest Ethiopia: a prospective observational study

Article References: Abadiko, E. A., Ashine, A. H., Ahimed, M. D., Herego, N. S., Shifa, A. B., Mengistu, F. G., Woyimo, T. G., & Gudina, E. K. (2026). Diagnostic and prognostic performance of procalcitonin among hospitalized adults with suspected bacterial infection in Southwest Ethiopia: a prospective observational study. BMC Infectious Diseases. https://doi.org/10.1186/s12879-026-14655-y

Image Credits: AI Generated

DOI: 10.1186/s12879-026-14655-y

Keywords: procalcitonin, bacterial infection, sepsis, biomarker, antimicrobial stewardship, diagnostic accuracy, Ethiopia, qSOFA, antibiotic discontinuation, hospitalized patients, ROC analysis, infectious diseases