Randomized, controlled trial of nasopharyngeal continuous positive airway pressure in the extubation of very low birth weight infants

J Pediatr. 1994 Mar;124(3):455-60. doi: 10.1016/s0022-3476(94)70375-2.


We conducted a prospective, randomized controlled trial to determine whether extubation of very low birth weight infants was facilitated by the use of nasopharyngeal continuous positive airway pressure (CPAP). Eligible infants included patients weighing 600 to 1500 gm at birth who required tracheal intubation within 48 hours of birth and who met specific predetermined criteria for extubation by day 14 of life. We also sought to determine whether varying the duration of nasopharyngeal CPAP influenced the likelihood of successful extubation. Infants underwent random assignment to receive nasopharyngeal CPAP until resolution of lung disease (n = 40), 6 hours of nasopharyngeal CPAP (n = 42), or oxygen supplementation delivered by hood (n = 42). Extubation failure was predefined as a requirement for > or = 80% oxygen, pH < or = 7.20, severe apnea, or predefined clinical deterioration, and extubation success was predefined as the ability to remain free of a requirement for mechanical ventilation for 7 days and a 66% reduction in the need for supplemental oxygen. Each group was similar with regard to race, sex, and birth weight. Extubation was successful in 62%, 61%, and 60% of infants. After stratification by birth weight, there were no significant differences in the rates of successful extubation among the treatment groups. We conclude that nasopharyngeal CPAP does not improve the likelihood of successful extubation of very low birth weight infants who are ready for extubation within the first 2 weeks of life.

Publication types

  • Clinical Trial
  • Randomized Controlled Trial
  • Research Support, Non-U.S. Gov't

MeSH terms

  • Female
  • Humans
  • Infant, Low Birth Weight*
  • Infant, Newborn
  • Male
  • Oxygen Inhalation Therapy
  • Positive-Pressure Respiration*
  • Prospective Studies
  • Ventilator Weaning / methods*