713 - Prediction Model for Extubation Readiness in Extreme Preterm Infants: A Multicenter Cohort Study

Publication Number: 4697.713

Sanjay Chawla, Children's Hospital of Michigan, TROY, MI, United States; Rachel G. Greenberg, Duke Clinical Research Institute, Durham, NC, United States; Matthew Rysavy, McGovern Medical School at the University of Texas Health Science Center at Houston, Houston, TX, United States; Ravi M. Patel, Children's Healthcare of Atlanta and Emory University, Atlanta, GA, United States; Vineet Bhandari, The Children's Regional Hospital at Cooper, Camden, NJ, United States; Michael Cotten, Duke University School of Medicine, Durham, NC, United States; Eduardo Bancalari, University of Miami, MIAMI, FL, United States; Nelson Claure, University of Miami Leonard M. Miller School of Medicine, Miami, FL, United States; Girija Natarajan, Central Michigan University College of Medicine, Detroit, MI, United States; Sanket Jani, Valley Children's hospital, Fresno, CA, United States; Anup Katheria, University of California, San Diego School of Medicine, San Diego, CA, United States; Katherine Coughlin, Sharp Mary Birch Hospital for Women and Newborns, San Diego, CA, United States; Elizabeth Foglia, Children's Hospital of Philadelphia, Swarthmore, PA, United States; Girija G. Konduri, Medical College of Wisconsin, Brookfield, WI, United States; Susan Cohen, CW HOPE Developmental Follow Up Clinic, Milwaukee, WI, United States; Joanne Lagatta, Medical College of Wisconsin, Milwaukee, WI, United States; Lindsey A. Knake, University of Iowa Roy J. and Lucille A. Carver College of Medicine, Iowa City, IA, United States; Amit Mathur, Saint Louis University School of Medicine, Chesterfield, MO, United States; Anand Salem, Children's Healthcare of Atlanta/Emory University School of Medicine, Atlanta, GA, United States; Amir M. Khan, McGovern Medical School at the University of Texas Health Science Center at Houston, Houston, TX, United States; William A. Adams, Cincinnati Children's Hospital Medical Center, Cincinnati, OH, United States

Background: Prolonged invasive mechanical ventilation (IMV) is associated with both short-term morbidities and neurologic impairment in preterm infants. A significant proportion (25%-35%) of infants fail extubation. Extubation practices vary widely. Models predicting extubation success have only used single center data.

Objective: To develop a model to predict successful extubation for extremely preterm infants.

Design/Methods: The MERIT study was a retrospective cohort study conducted to refine and validate an extubation readiness prediction model. The study involved 11 U.S. sites from March 2019 and October 2021. Infants < 28 weeks' gestational age (GA) were eligible if they received IMV for at least 12 h and had an elective extubation attempt within 42 d of birth. The primary outcome was successful extubation, defined as survival without reintubation within 5 days of extubation. We compared demographics and clinical characteristics between infants with and without successful extubation. We used multivariable logistic regression with backward stepwise selection (with p-values 0.2 eligible for removal) to develop a prediction model for successful extubation, with area under the receiver operating characteristic curve (AUC) used to evaluate model performance.

Results: The study cohort included 749 infants. Of these, 568 (76%) had successful extubation and 181 (24%) failed extubation. Among infants who failed extubation, the majority (138/181, 76%) were reintubated within 72 hours (Figure). On unadjusted analysis, successfully extubated infants were more likely to be female and have higher GA at birth and birth weight. At extubation, infants in the successful group had higher postmenstrual age (PMA), weight and blood gas pH (Table 1). On unadjusted analysis, successfully extubated infants were more likely to be females, and having higher GA and birth weights. In adjusted analysis, successful extubation was associated with female sex and higher GA, birth weight, weight at extubation, postnatal age, and blood gas pH and lower pre-extubation Respiratory Severity Score (RSS) (Table 2). The AUC for the prediction model was 0.71.

Conclusion(s): In a large multicenter cohort of extremely preterm infants, we developed a model with 7 factors that could predict extubation success with moderate discrimination. Future work will focus on prospectively testing its clinical utility.

Table 1: Characteristics of Infants in Successful and Failed Extubation Groups
Values are mean (SD), Median (25th, 75th percentile) or n/N (%).
Continuous variables with >4% missingness were Apgar score - 1and -5 minutes, age at extubation and highest RSS on day of life 1


Table 2: Multivariable Model for Successful Extubation
Results from multivariable model fitted to outcome of successful extubation

Kaplan-Meier curve showing proportion of infants remaining extubated over time