The mechanisms underlying the therapeutic function of caffeine on apneas in preterm neonates are not well determined. To better understand these effects, we exposed rat pups from postnatal d 3-12 to chronic intermittent hypoxia (5% O2/100 s every 10 min; 6 cycles/h followed by 1 h at 21% O2, 24 h/d), a model mimicking hypoxemic exposure in apneic neonates. Then, using whole-body plethysmography, we evaluated minute ventilation, apnea frequency, and duration after i.p injection of caffeine citrate (20 mg/kg) or saline under normoxia and in response to either sustained (FiO2 12%, 20 min) or brief (FiO2 5%, 60 s, total 10 episodes of 8 min each) hypoxia. These tests were used to assess peripheral and central components of hypoxic response. The latter also assessed the ventilatory long-term facilitation during recovery (2 h). Caffeine injection increased minute ventilation under baseline and during recovery. This effect was correlated with a decrease in apnea frequency (not duration). On the contrary, caffeine did not change the ventilatory response to sustained or brief hypoxic exposure. These results suggest that the effects of caffeine on apnea depend on increased central normoxic respiratory drive and enhancement of ventilatory long-term facilitation rather than on higher hypoxic ventilatory response.