Introduction: Chronic oxygen dependency (COD) is a common adverse outcome of very premature birth. It is, therefore, important to develop an accurate and simple predictive test to facilitate targeting of interventions to prevent COD. Our aim was to determine if a simple score based on respiratory support requirements predicted COD development.
Methods: A retrospective study of 136 infants, median gestation age (GA) 28 weeks (range: 23-33 weeks) and a prospective study of 75 infants, median GA 30 weeks (range: 23-32 weeks), were performed. The score was calculated by multiplying the inspired oxygen concentration by the level of respiratory support (mechanical ventilation: 2.5; continuous positive airway pressure: 1.5; nasal cannula or head box oxygen or air: 1.0). Scores were calculated on data from days 2 and 7, and their predictive ability compared to that of the maximum inspired oxygen concentration at those ages and (retrospective study) the results of lung volume measurement.
Results: Infants that were oxygen dependent at 28 days and 36 weeks post-menstrual age (PMA) had higher scores on days 2 (p<0.0001, p<0.0001, respectively) and 7 (p<0.0001, p<0.0001, respectively) than the non-oxygen dependent infants in both the retrospective and prospective cohorts. Construction of receiver operator characteristic curves demonstrated the score performed better than the inspired oxygen level and lung volume measurement results. A score on day 7 >0.323 had 95% specificity and 78% sensitivity in predicting COD at 28 days, and 80% specificity and 73% sensitivity in predicting COD at 36 weeks PMA.
Conclusion: Chronic oxygen dependency can be predicted using a simple scoring system.