Association between trajectory of heparin-binding protein and extubation in ARDS patients: A longitudinal study
Introduction
Acute respiratory distress syndrome (ARDS) is a critical condition characterized by severe hypoxemia and bilateral chest radiographical opacities due to non-cardiogenic pulmonary edema. It has a high mortality rate, making accurate prediction of extubation success crucial. The pathophysiology of ARDS involves injury to pulmonary vascular endothelial cells and alveolar epithelia, leading to increased capillary permeability and impaired pulmonary function. Heparin-binding protein (HBP), a neutrophil-derived biomarker, plays a pivotal role in vascular leakage and pulmonary edema by mediating proinflammatory cytokine release and immune cell chemotaxis. Elevated HBP levels are strongly associated with the development and severity of ARDS.
However, research on the relationship between HBP dynamics and extubation outcomes in ARDS patients is limited. This study aims to investigate the association between different HBP trajectories and extubation outcomes using group-based trajectory modeling (GBTM). The findings will contribute to a predictive tool for guiding ventilator weaning strategies and optimizing clinical decision-making.
Methods
ARDS patients admitted to the ICU of Ningbo Medical Center Lihuli Hospital between September 2023 and March 2025 were enrolled. The diagnosis of ARDS was based on global criteria. Informed consent was waived for this retrospective study. The inclusion criteria were: staying in the ICU for more than six days, complete clinical data, and age > 18 years. Exclusion criteria included extubation within 6 days after enrollment, tracheostomy, and patients requesting discharge during treatment.
HBP levels were measured daily from 1-6 days after enrollment. The primary outcome was extubation success, defined as sustained spontaneous breathing for more than 48 hours without mechanical ventilation. Secondary outcomes included ventilator-free days, length of ICU and hospital stays, and ICU mortality.
Statistical analysis involved the Kolmogorov-Smirnov test for normality. Continuous variables were expressed as means ± SD or [M (P25, P75)] medians (interquartile ranges) and compared using t-tests or Mann-Whitney U-tests. Categorical variables were presented as numbers (%) and compared using the Chi-square test.
HBP trajectories were categorized using GBTM, implemented in R with the 'gbmt' package. The optimal number of trajectory classes was determined using Bayesian Information Criterion (BIC) and Akaike Information Criterion (AIC) values, average posterior probability of assignment (AvePP), odds of correct classification (OCC), minimal class sizes, and clinical experience.
A two-step regression analysis was conducted to identify factors associated with extubation success. Univariate logistic regression was performed for different HBP patterns and clinical variables. Multivariate logistic regression was then used to assess the relationship between HBP trajectories and extubation success, adjusting for potential confounders.
Results
267 patients were included, with 156 successfully extubated. The cohort was predominantly male (73.78%) with a median age of 72 years [IQR: 63-80] and a median APACHE II score of 21 (IQR: 16.00-24.00). The primary disease was pulmonary disease. HBP levels on the first day were lower in successfully extubated patients (P < 0.05).
Five distinct HBP trajectories were identified: traj1 (persistently extremely low HBP), traj2 (persistently low HBP), traj3 (high-to-low transitional HBP), traj4 (persistently moderate HBP), and traj5 (persistently high HBP). The extubation success rate decreased significantly from traj1 to traj5 (P < 0.001). Ventilator-free days and ICU mortality also showed significant differences among the groups.
Association Between HBP Trajectories and Extubation Success
Logistic regression analysis revealed a significant association between HBP trajectories and extubation success. Traj2, traj3, traj4, and traj5 were associated with decreased extubation success rates compared to traj1. These associations persisted after adjusting for confounders.
Subgroup Analysis
Subgroup analysis by age and SOFA score showed non-significant interaction effects. However, robust associations were observed in patients aged ≥ 72 years and with SOFA scores < 7 or ≥ 7.
Discussion
This study identified five HBP trajectories in ARDS patients. Higher HBP levels were associated with a higher risk of extubation failure. These findings may help identify high-risk populations and guide extubation decisions.
HBP, a key component of neutrophil granules, plays a crucial role in initiating inflammation and enhancing vascular permeability. It binds to endothelial cell surfaces, activating protein kinase C and Rho kinase pathways, leading to increased vascular permeability and the release of proinflammatory mediators. Elevated HBP levels in ARDS patients correlate with lung injury severity.
The study's findings suggest that sustained moderate or high HBP levels may perpetuate alveolar-capillary barrier dysfunction, driving pulmonary edema and more severe lung injury. Dynamic HBP trajectory monitoring provides enhanced prognostic utility for extubation decision-making.
Subgroup analyses revealed persistent associations between HBP trajectory patterns and extubation outcomes, particularly in patients aged ≥ 72 years and with SOFA scores < 7 or ≥ 7.
Limitations
The study had a moderate sample size and was conducted at a single center. As a retrospective study, it cannot establish a causal relationship between HBP trajectories and extubation outcomes. Larger, multi-center studies are needed for further validation.
Conclusion
The study identified distinct HBP trajectories strongly associated with extubation success in ARDS, providing a potential tool for risk stratification. Monitoring HBP dynamics may help identify high-risk patients, enabling closer monitoring and tailored weaning protocols. Future research should focus on validating these trajectories and establishing specific HBP cut-off values for clinical decision-making.
