OBJECTIVE To evaluate the effectiveness and safety of different doses of caffeine in treatment of primary apnea in preterm infants. METHOD A total of 164 preterm infants (<32 weeks gestation), presented with primary apnea, were recruited in Tianjin Central Hospital of Gynecology and Obstetrics from October 2013 to December 2014. The patients were prospectively allocated into low-dose (loading 20 mg/kg and maintenance of 5 mg/(kg·d) after 24 h, n=82) and high-dose (loading 20 mg/kg and maintenance of 15 mg/(kg·d) after 24 h, n=82) groups of caffeine citrate treatment by using a random number table. The treatment effects, side effects of caffeine, and the clinical outcome of the preterm infants were compared between groups by χ(2) test or nonparametric test. RESULT The patients in low-dose group had birth weight of (1,237 ± 338) g, male gender of 43 (52%) and gestational age of (29.8 ± 3.4) weeks. The patients in high-dose group had birth weight of (1 262 ± 296) g, male gender of 45 (55%) and gestational age of (29.9 ± 2.7) weeks. The baseline characteristics including birth weight, gender and gestational age were comparable between the two groups. Frequency of apnea was significantly lower in high-dose group compared with low-dose group (10 (8, 15) vs.18 (13, 22), Z = -2.610, P = 0.009), and the success rate of removal of the ventilator was significantly higher in high-dose group compared with low-dose group (85% (70/82) vs.70% (57/82), χ(2) = 5.898, P = 0.015). The effective rate of caffeine treatment was significantly higher in high-dose group compared with low-dose group (82% (67/82) vs.61% (50/82), χ(2)=8.619, P = 0.003). No significant differences were observed concerning the incidence of caffeine-associated side effects including tachycardia, irritability, difficulty in feeding, hyperglycemia, hypertension, digestive disorders and electrolyte disturbances between two groups (P all > 0.05). There were no significant differences in the clinical outcomes of the preterm infants including death during hospitalization, chronic lung disease, other complications and duration of hospital stay between two groups (P all > 0.05). CONCLUSION A therapeutic regimen consisting of a loading dose of 20 mg/kg and maintenance dose of 15 mg/(kg·d) of caffeine citrate could improve the treatment effects and keep safety for primary apnea in preterm infants, and will not cause more adverse events.