Abstract: Objective To explore the expression of serum soluble triggering receptor expressed on myeloid cells-1 (sTREM-1) and Clara cell protein (CC16) in neonates with and without mechanical ventilation and their predictive value in ventilator-associated pneumonia (VAP). 
Methods A total of 183 neonates in neonatal intensive care unit (NICU) were selected and categorized into two groups based on their utilization of mechanical ventilation: the mechanical ventilation group (n=138) and the non-mechanical ventilation group (n=45). Clinical data, as well as serum levels of sTREM-1 and CC16 at different time periods were detected in both groups. Subsequently, the mechanical ventilation group was further stratified into the VAP group (n=63) and the non-VAP group (n=75). Comparative analysis of clinical data and serum sTREM-1 and CC16 was conducted between these subgroups. Logistic regression analysis was employed to determine the factors influencing VAP occurrence, while receiver operating characteristic (ROC) curve analysis was employed to evaluate the diagnostic performance of various indicators for VAP. 
Results The mechanical ventilation group exhibited significantly lower Apgar scores compared with the non-mechanical ventilation group, while lactate (Lac) levels were significantly higher in the mechanical ventilation group (P＜0.05). Serum CC16 levels displayed a declining trend over time in both groups, with the mechanical ventilation group demonstrating lower levels than the non-mechanical ventilation group. Conversely, serum sTREM-1 levels initially increased and subsequently decreased in both groups, with the mechanical ventilation group exhibiting higher levels than the non-mechanical ventilation group. These differences in interaction between groups, time points, and time points between groups were statistically significant (P＜0.05). Within the VAP subgroup, Apgar scores were lower compared with the non-VAP subgroup, while C-reactive protein (CRP) levels were higher (P＜0.05). Similarly, both subgroups displayed a declining trend in serum CC16 levels over time, with the VAP subgroup demonstrating lower levels than the non-VAP subgroup. Serum sTREM-1 levels initially increased and subsequently decreased in both subgroups, with the VAP subgroup exhibiting higher levels than the non-VAP subgroup. The differences in interaction between groups, time points, and time points between groups were statistically significant (P＜0.05). Logistic regression analysis indicated that serum CC16 levels at 48 h, 72 h, and at 48 h post-extubation, as well as serum sTREM-1 levels at 48 h, 72 h, and 48 h post-extubation, were significant factors influencing the occurrence of VAP (P＜0.05). Moreover, regression analysis demonstrated significantly higher sensitivity and specificity in diagnosing VAP compared with individual indicator detection (P＜0.05). 
Conclusion Serum sTREM-1 and CC16 exhibit aberrant expression in mechanically ventilated pediatric patients, and their levels at different time periods during mechanical ventilation hold promise for predicting the occurrence of VAP. Moreover, the combined use of these biomarkers demonstrated enhanced diagnostic accuracy, emphasizing the necessity for further clinical investigations in this domain.

Key words: respiration, artificial, infant, newborn, pneumonia, ventilator-associated