Results
Over the study period (2001-2011), 10 128 children were
born before 30 weeks’ gestation, survived to at least 36 weeks’
postmenstrual age, and were eligible for neonatal follow-up
at 18-22 months’ corrected age in 16 NRN centers. We
excluded 1445 (14%) children because of missing data on
the primary outcome, of whom 29 (2.0%) had tracheostomies.
The remaining 8683 children included 304 children (3.5%) with tracheostomies (Table I). Infants who
underwent tracheostomy had lower average birth weight
and gestational age and were more likely to be male and
have a congenital anomaly. Infants with tracheostomies
had higher incidence of all in-hospital morbidities than
those without tracheostomies (Table I). There was no
significant change in incidence of tracheostomy over the 10
years of this study (Wald c2 P = .29). The proportion of
children undergoing tracheostomy varied significantly
across the 16 participating centers from 1% to 10% (Wald
c2 P < .001), likely reflecting differing referral and practice
patterns. The average corrected age at which developmental
testing was performed in survivors was the same in the 2
groups (median 19 months, IQR 16-22, P = .27).
Primary Outcome
Children with tracheostomies had significantly higher risk of
the primary composite outcome of death or NDI and of all of
the components of NDI (Table II) than did children without
tracheostomies. In an attempt to adjust for risk factors other
than tracheostomy, we performed prespecified analyses
adjusted for 15 factors. Each of these is known to be
associated with both risk for tracheostomy and risk for
adverse neurodevelopmental outcomes, as described earlier.
After adjustment, prevalence of the primary outcome of
death or NDI, NDI, and the components of NDI remained
substantially higher in children with tracheostomies than in
those without tracheostomies (Table II). The “fully
adjusted” OR of death or NDI in children with
tracheostomies, compared with those without, was 3.3
(95% CI 2.4-4.6), and the aOR of NDI in survivors with