Abstract. OBJECTIVE: To examine the success rate of our current practice of discontinuing caffeine at 33 0/7-35 6/7 weeks post menstrual age (PMA), as well as factors that predict the success or failure of discontinuation. STUDY DESIGN: A retrospective chart review of infants born before 34 weeks gestational age between 2006–2012. Data collected included demographics, and other comorbidities mainly complications of prematurity. RESULTS: 647 treated infants had caffeine discontinued at 33–35 PMA before discharge or transfer. 64 (10%) infants failed discontinuation and had caffeine restarted. Most (77%) of those who failed started having monitor alarms within 7 days of discontinuation. BPD and Hispanic ethnicity were predictive of weaning failure (p < 0.05). CONCLUSION: Caffeine can be discontinued at 33–35 weeks PMA with a failure rate of 10%. BPD and Hispanic ethnicity are predictive of failure. It is generally safe to discharge infants seven days after the caffeine was discontinued if no significant monitor events occur during that time.