Apnea of prematurity (AOP) has recently received increased attention in the impact it has on length of stay and timing of discharge. The use of caffeine as both prophylaxis and treatment of AOP has become widespread despite the fact that there is a lack of consensus on how AOP should be managed. We prospectively followed, via chart review, all infants placed on caffeine in our NICU over a 10 month period to define our pattern of usage. Initiation, dose adjustment, and discontinuation of caffeine were up to the attending neonatologist. 83 infants were begun on caffeine during the observation period: birthweight of 1144±398 gms and gestational age of 28±2 wks. Caffeine therapy was initiated prior to the onset of symptomatic AOP in 58% of the infants. 43% of all infants while on caffeine were apnea-free. Infants received caffeine for 31±16 days with a maximal dose of 7±2 mg/kg (range 5 to 13.3 mg/kg/day). Levels were obtained on only 69% of the patients with a maximal level of 18±5 mg/L (range 6.5 to 29.6 mg/L). Therapy was discontinued at 34±2 wks postconceptual age (PCA). Apneic events were classified based on the need for intervention vs. self-resolution with the last event requiring intervention at 34±2 wks PCA and the last self-resolving event at 35±2 wks PCA. Seven infants were discharged home on a monitor: 4 for unresolved AOP and 3 for other clinical indications. Resolution of AOP was the limiting factor for discharge in 32% of the infants sent home from our NICU. Conclusions: In our patients treated with caffeine, half were started prior to becoming symptomatic and half of those on therapy remained asymptomatic, indicating possible overuse of the medication. The maximal dose of caffeine was highly variable, as were the serum levels, implying significant interpatient and interphysician variability in the amount of medication needed and how used. Nearly one-third of the patients were managed without serum drug levels indicating that levels are not needed for all patients. Discontinuation of caffeine and the last event requiring intervention both occured at 34 wks PCA, suggesting a possible delay in discontinuing therapy. A more cost-effective approach to the use of caffeine for AOP might include: better defined guidelines for the initiation and discontinuation of caffeine; and less routine measurement of serum levels.