In-hospital mortality following tracheostomy in very preterm infants with bronchopulmonary dysplasia: A Japanese nationwide retrospective cohort study.

BACKGROUND Very preterm infants with bronchopulmonary dysplasia (BPD) require tracheostomy when withdrawal from invasive ventilation is difficult. However, information regarding in-hospital mortality is limited. The present study aimed to describe the in-hospital courses following tracheostomy and to examine the clinical factors associated with in-hospital death. METHODS This retrospective study used a Japanese national inpatient database. We included all infants born at < 32 weeks' gestational age or with a birthweight <1500 g between July 2010 and March 2020 who underwent tracheotomy secondary to BPD (supplementary oxygen, non-invasive ventilation, or mechanical ventilation at 36 weeks corrected gestational age). Prolonged-steroid-exposed BPD was defined as hydrocortisone or dexamethasone use for ≥7 cumulative days, while severe pulmonary hypertension (PHT) was defined as inhaled nitric oxide or pulmonary vasodilator use for ≥7 cumulative days before tracheostomy. Multivariable Cox regression models were applied to assess factors associated with in-hospital death after tracheostomy. RESULTS Among 18,056 very preterm patients with BPD, 484 underwent tracheostomy. After excluding those with major congenital anomalies, 294 were eligible. The median age at tracheostomy was 174 days (interquartile range, 129-244 days). Of these, 47 (16.0%) patients died, and the median time from tracheostomy to death was 125 days (interquartile range, 59-339 days). Prolonged-steroid-exposed BPD (adjusted hazard ratio [aHR], 3.08; 95% confidence interval [CI], 1.35-6.99) and prolonged-steroid-exposed BPD/PHT (aHR, 5.93; 95% CI, 2.89-12.17), but not corrected gestational age of tracheostomy (aHR, 1.01; 95% CI, 1.00-1.02) were associated with death. CONCLUSIONS In very preterm patients who underwent tracheostomy secondary to BPD, prolonged-steroid-exposed BPD and BPD/PHT before tracheostomy were associated with in-hospital mortality.