Temporal Changes in Inflammatory Markers and Laboratory Parameters in Relation to Risk of Neonatal Sepsis

Background/Objective: Early diagnosis of neonatal sepsis remains a challenge due to the nonspecific nature of clinical symptoms and the delayed dynamics of conventional inflammatory markers. The traditional diagnostic approach is based on static cut-off points, which often do not consider the nonlinear nature of the inflammatory response. The aim of this study was to identify dynamic variables associated with sepsis progression, constituting the foundation for future artificial intelligence (AI) models. Methods: A retrospective analysis was performed using anonymised clinical and laboratory data from a cohort of newborns (n = 122) with suspected sepsis, along with maternal clinical data. Using feature engineering, descriptive statistics parameters were generated for selected parameters, including rates of change in the 24 h and 48 h time windows. This was combined with the verification of clinical and perinatal parameters that affect the assessment of clinical status and diagnosis of sepsis. The correlation analysis was performed using Spearman’s rank correlation coefficient (p < 0.05). Results: Dynamic parameters showed significant associations with sepsis progression than static measurements. The highest correlation was recorded for the rate of CRP growth (CRP_abs_change_rate_max, r = 0.613, p < 0.001), compared with the maximum values (CRP_value_max, r = 0.591, p < 0.001). High correlations were also observed for leukocyte parameters, especially neutrophils (NEUT_%_pct_change_rate_max, r = 0.547, p < 0.001) and basophils (BASO_%_value_max, r = 0.606, p < 0.001), with the strongest correlations observed within the first 24 h (r > 0.70). Conclusions: Neonatal sepsis progression may be more comprehensively characterised by temporal changes in laboratory parameters, not by their absolute values. Rates of change in inflammatory markers may provide clinically relevant complementary information on the course of the disease, reflecting the complex and nonlinear nature of the inflammatory response. The results indicate that the analysis of the trajectory of changes in parameters provides a promising basis for the future development and validation of predictive models based on artificial intelligence for earlier and more accurate diagnosis of neonatal sepsis.