Dynamic procalcitonin trajectories for prognostic prediction in wasp envenomation: A multi-center retrospective
Yonghong Wang1, Junping Zhu1, Pan Wei1
1Department of Emergency, Taihe Hospital, Hubei University of Medicine, Shiyan, 442012, Hubei, China.
Objective:
To explore the relationship between dynamic procalcitonin (PCT) trajectories and clinical outcomes in wasp envenomation patients, and assess the clinical value of PCT trajectories for prognostic prediction.
Methods:
A retrospective multi-center cohort study was conducted on wasp envenomation patients admitted to 3 tertiary hospitals in Hubei Province from January 2020 to December 2023.
Inclusion Criteria:
age≥18, confirmed wasp envenomation history, hospital stay≥24 h, and complete PCT monitoring data (≥3 measurements within 72 h). Group-based trajectory modeling (GBTM) identified PCT patterns, and Cox regression analyzed the association between trajectory groups and 28-day mortality. Supplementary biomarker analysis examined correlations between PCT trajectories and inflammatory cytokines (interleukin-6 (IL-6), tumor necrosis factor-alpha (TNF-α)). Early trajectory identification performance was evaluated at 24-h and 48-h timepoints.
Results:
A total of 826 patients were included (498 males, 60.3 %; mean age 57.4 ± 16.2 years). The 28-day mortality rate was 8.5 % (70 cases). GBTM identified four PCT trajectory patterns: low-stable (37.8 %), moderate elevation-decline (34.7 %), high-persistent (19.6 %), and extremely high-peak (7.9 %). The extremely high-peak group had the highest 28-day mortality rate (41.5 %, P < 0.001). Multivariate Cox regression showed the extremely high-peak trajectory increased mortality risk by 12.8-fold (HR = 12.84, 95 % CI: 6.92-23.81, P < 0.001), and the high-persistent trajectory increased it by 4.2-fold (HR = 4.23, 95 % CI: 2.15-8.31, P < 0.001). The PCT trajectory model had superior predictive accuracy (AUC = 0.867) compared to single-point PCT detection (AUC = 0.743, P < 0.001). Analysis of 623 patients (75.4 %) with complete IL-6 data demonstrated a strong correlation between PCT and IL-6 peaks in the extremely high-peak trajectory group (Spearman r = 0.682, P < 0.001), supporting the cytokine storm hypothesis. Early trajectory identification achieved AUC of 0.721 at 24 h and 0.783 at 48 h.
Conclusion:
PCT trajectory modeling provides superior prognostic value compared to single-point measurements. The extremely high-peak and high-persistent trajectories are independent predictors of 28-day mortality, supporting individualized risk stratification.
