Clin Chem Lab Med. 2025 Nov 12. doi: 10.1515/cclm-2025-0778. Online ahead of print.
ABSTRACT
OBJECTIVES: Procalcitonin (PCT) is increasingly used to support sepsis diagnosis, but its role in predicting outcomes remains uncertain, particularly in low-resource settings. We evaluated whether single and serial PCT measurements were associated with 28-day mortality in critically ill adults with suspected sepsis.
METHODS: We conducted a retrospective study of adult intensive care unit (ICU) patients with suspected sepsis and at least one PCT measurement at a tertiary hospital in South Africa (August 2022-July 2023). Baseline PCT was analysed using multivariable logistic regression. Among patients with ≥2 PCT values, we examined the association between PCT slope (from patient-level linear regression) and mortality. Latent class mixed models (LCMM) were used to identify PCT trajectory subgroups.
RESULTS: Of 371 patients, 119 (32 %) died within 28 days. Higher baseline log-PCT was independently associated with increased mortality (adjusted odds ratio [aOR] 1.58; 95 % CI 1.01-2.50). A rising PCT slope trended toward higher mortality (aOR 3.56 per unit/day; p=0.06). LCMM identified three trajectory classes with distinct mortality risks (class 2: aOR 4.53; class 3: aOR 4.35, vs. reference). These models were based entirely on laboratory data and did not assess clinical scoring systems.
CONCLUSIONS: Both baseline and serial PCT measurements predicted mortality in ICU patients with suspected sepsis. Modelling approaches based on routine laboratory data may offer scalable tools for early risk stratification in resource-limited settings.
PMID:41217295 | DOI:10.1515/cclm-2025-0778