We acknowledge several additional limitations of this large observational study. First, attrition bias is possible because we are unable to ascertain the primary outcome in 14% of eligible children and attrition was higher among children without tracheostomies. Selection bias may have been introduced either by clinical decisions about which infants received tracheostomies or by the inclusion of only survivors to 36 weeks in the current study. Last, we emphasize again that we can only describe an association between the need for tracheostomy and adverse outcomes. The child who ultimately requires tracheostomy placement will have accumulated multiple risk factors for adverse outcome, some of which we were able to include in our analyses and some of which must be considered unmeasured confounders. Thus, the need for tracheostomy must suggest to clinicians that a patient's severity of illness puts that patient at extremely high risk for adverse long-term outcomes.
In conclusion, this is the first large, multicenter study to describe the long-term neurodevelopmental outcomes of preterm infants with tracheostomies with comparison with the general preterm population. Tracheostomy does not mitigate the significant risk for adverse neurodevelopment that is associated with multiple neonatal morbidities, major malformations, or severe BPD. However, if tracheostomy is to be performed, earlier surgery may allow opportunities for enhanced neurodevelopment. Information about the association between tracheostomy placement and outcomes in this population may facilitate a shared decision-making process between parents and clinicians about whether and when to pursue tracheostomy placement.