717 - Real-Time Ventilator Compliance Is Associated with pCO₂ in Preterm Infants

Publication Number: 4701.717

Mashfiq Hasan, Golisano Children's Hospital at The University of Rochester Medical Center, Rochester, NY, United States; Andrew M. Dylag, University of Rochester, Golisano Children’s Hospital, Rochester, NY, United States; Rita Dadiz, Golisano Children's Hospital at The University of Rochester Medical Center, Rochester, NY, United States; Nelson Claure, University of Miami Leonard M. Miller School of Medicine, Miami, FL, United States; Abdus Wahed, University of Rochester School of Medicine and Dentistry, Penfield, NY, United States; Divya Chhabra, University of California, Davis, School of Medicine, Sacramento, CA, United States; Patricia Chess, Golisano Children's Hospital at The University of Rochester Medical Center, Rochester, NY, United States

Background: Modern neonatal ventilators record continuous lung mechanics data, but their clinical value for predicting blood gas pCO₂ in preterm infants is unclear.

Objective: To utilize ventilator-derived lung mechanics (VDLM) parameters of compliance and minute ventilation (MV) and test their correlation to blood gas pCO₂.

Design/Methods: We conducted a prospective observational study of nineteen preterm infants < 33 weeks' gestation on volume-targeted ventilation. High-resolution ventilator data (41 variables recorded at 10-second intervals) were collected for up to 31 days or until extubation, whichever came first. Ventilator-derived compliance and MV were selected for primary data analysis given their established relevance in pCO₂ regulation. Arterial, capillary, and venous blood gas measurements were extracted from the electronic medical record. To approximate the physiological state at the time of each blood gas, compliance and MV were averaged over a 10-minute window spanning 5 minutes before and 5 minutes after obtaining a blood gas. We used the Kruskal-Wallis test with Dunn's post-hoc tests and Bonferroni correction to test the association between three pCO₂ categories defined as: low ( < 40 mmHg), normal (40-60 mmHg), and high (>60 mmHg). We applied a multinomial generalized linear mixed model (GLMM) with clustering by patient to account for repeated measures from individual subjects.

Results: Nineteen infants contributed 358 blood gases and 21,480 VDLM data points (Table 1). Compliance was higher in both low (n = 25, p = 0.004) and high (n = 81, p = 0.0087) pCO₂ groups versus normal (n = 252) (Figure 1). In GLMM analyses adjusting for age, FiO₂, leak, and MV, each 0.1-unit increase in compliance raised the odds of low pCO₂ by 32% (aOR 1.32, 95% CI 1.15-1.51, p < 0.001) and high pCO₂ by 13% (aOR 1.13, 95% CI 1.03-1.25, p = 0.014). MV differed between low and normal pCO₂ in Dunn's test (p = 0.017), whereas in the GLMM it differed between high and normal pCO₂ (p < 0.001) (Table 2).

Conclusion(s): In this pilot preterm cohort, ventilator-derived compliance was higher when pCO₂ was outside the normal range. At low pCO₂, higher compliance likely reflects over-ventilation from healthy compliant lungs. Higher compliance at high pCO₂ may reflect increased patient effort reducing intrathoracic pressure. Prior studies have not linked real-time compliance to blood gas pCO₂ and these findings suggest that ventilator-based monitoring can guide respiratory care and reduce blood sampling. Larger studies are needed to further elucidate this relationship and to explore neonatal correlations of other VDLM variables.

Table 1. Summary of study variables across pCO₂ categories ( <40, 40-60, and >60 mmHg).
Data are expressed as median (interquartile range). Abbreviations: GA, gestational age; MV, minute ventilation; FiO₂, fraction of inspired oxygen; Age, postnatal age in days at the time of blood gas sampling; Weight, weight at the time of blood gas measurement; Leak, percent leak.

Figure 1. Association between ventilator-derived compliance and MV and blood gas pCO₂ categories.
Boxplots show the distribution of (A) Compliance (mL/cmH₂O) and (B) Minute Ventilation (mL/kg/min) across low ( <40 mmHg), normal (40-60 mmHg), and high (>60 mmHg) pCO₂ groups. Compliance was significantly higher in both low and high pCO₂ categories compared with normal (Dunn's post-hoc, p < 0.01). Minute ventilation differed significantly only between normal and low groups (p < 0.05).
Significance thresholds: p < 0.05 (*), p < 0.01 (**).

Table 2. Generalized linear mixed model results evaluating factors associated with low and high pCO₂ categories compared with normal pCO₂ category.
Data are presented as adjusted odds ratios (aOR) with 95% confidence intervals (CI) adjusted for minute ventilation, FiO₂, postnatal age, and leak as covariates. Odds ratios for compliance reflect the change in odds per 0.1 mL/cmH₂O increase, minute ventilation reflect the change in odds per 100 mL/kg/min increase, and FiO₂, leak, and age reflect the change in odds per 1-unit increase (1% FiO₂, 1% leak, and 1 day of age, respectively). Abbreviations: MV, minute ventilation; FiO₂, fraction of inspired oxygen; Age, postnatal age in days; Leak, percent leak.