OBJECTIVE: To describe the epidemiology, risk factors, and in-hospital outcomes of tracheostomy in infants in the neonatal intensive care unit. STUDY DESIGN: We analyzed electronic medical records from 348 neonatal intensive care units for the period 1997 to 2012, and evaluated the associations among infant demographics, diagnoses, and pretracheostomy cardiopulmonary support with in-hospital mortality. We also determined the trends in use of infant tracheostomy over time. RESULTS: We identified 885 of 887 910 infants (0.1%) who underwent tracheostomy at a median postnatal age of 72 days (IQR, 27-119 days) and a median postmenstrual age of 42 weeks (IQR, 39-46 weeks). The most common diagnoses associated with tracheostomy were bronchopulmonary dysplasia (396 of 885; 45%), other upper airway anomalies (202 of 885; 23%), and laryngeal anomalies (115 of 885; 13%). In-hospital mortality after tracheostomy was 14% (125 of 885). On adjusted analysis, near-term gestational age (GA), small for GA status, pulmonary diagnoses, number of days of forced fraction of inspired oxygen >0.4, and inotropic support before tracheostomy were associated with increased in-hospital mortality. The proportion of infants requiring tracheostomy increased from 0.01% in 1997 to 0.1% in 2005 (P < .001), but has remained stable since. CONCLUSION: Tracheostomy is not commonly performed in hospitalized infants, but the associated mortality is high. Risk factors for increased in-hospital mortality after tracheostomy include near-term GA, small for GA status, and pulmonary diagnoses.