PubMed1996
OBJECTIVES: To assess risk factors affecting viability and analyze the effects of mechanical ventilation (MV) on neonatal outcome and resource use among extremely premature infants. DESIGN: Inception cohort study. SETTING: Neonatal intensive care units of the 12-center National Institute of Child Health and Human Development Neonatal Research Network. PARTICIPANTS: A total of 1126 infants with a birth weight of 501 to 800 g born in network centers between January 1, 1994, and December 31, 1995. MAIN OUTCOME MEASURES: Observed survival; maximum estimated survival (assuming the same survival among infants who died without MV as among infants in the same risk category who received MV); observed and maximum estimated survival without severe brain injury (either interventricular echodensity with ventricular dilation or parenchymal echodensity); hospital stay; resource investment. RESULTS: Overall mortality was 43%; mortality in infants without MV was 93%. A total of 15% of all the infants died without MV. Females, small-for-gestational-age infants, and infants whose mothers received antenatal steroids had an advantage in survival with MV equivalent to an increase in birth weight of 90 g, 57 g, and 67 g, respectively. The corresponding advantage of these infants in survival without severe brain injury was 107 g, 97 g, and 64 g, respectively. Females in the lowest birth-weight group were more likely to die without MV than were larger males with a similar estimated likelihood of survival with MV. Mean hospital stay was 115 days for the survivors, values much greater than the 17.9-day standard for 501- to 800-g survivors under the diagnosis related group system. Resource investment was considerable (127 hospital days per survivor and 148 days per survivor without severe brain injury), but, like outcome, varied markedly between risk categories. Had MV been used for all infants who died, we estimate a substantial increase in resource use and a maximum of 8 additional survivors (no more than 6 without severe brain injury per 100 infants with a birth weight of 501 to 800 g. CONCLUSIONS: Although recommendations to initiate or forgo MV for extremely premature infants have often focused on 1 factor (birth weight or gestational age), multiple factors should be considered. Other factors being equal, our analyses support use of MV for females at a minimum birth weight approximately 100 g lower than that for males. The current diagnosis related group reimbursement system can be expected to compromise resources for 501- to 800-g infants who would benefit from MV. Such care entails considerable resource use, although the cost per life-year gained is likely to be considerably less than that for many adults given intensive care. Our findings can be used to facilitate more appropriate treatment decisions, determine adequate resources, and better inform the debate about the benefits and burdens of intensive care for extremely premature newborns.