Objective To assess the effects of caffeine use in apnea of prematurity (AOP) in the early stage (caffeine was used within 3 days after birth) versus late stage (caffeine was used 4 to 10 days after birth ) . Methods Medline, Science Direct, Elsevier, Embase, CBMdisc and Wanfang databases were retrieved to incorporate studies that met the inclusion criteria. The retrieval time limit was from the establishment of the databases to November, 2015. References that were in line with the inclusion criteria were selected and relevant conference data were collected by manual retrieval. Two researchers conducted meta-analysis using software RevMan 5. 2 after independent data selection, information extraction and quality evaluation according to the inclusion and exclusion criteria. Results One randomized controlled trial and four retrospective cohort studies were included. A total of 59 288 patients were involved. Meta-analysis suggested that, compared with caffeine use in the late stage, the early stage group showed significant lower incidence ( P < 0. 05 ) of mortality, bronchopulmonary dysplasia, intraventricular hemorrhage, periventricular leukomalacia, retinopathy of prematurity and surgical intervention of patent ductus arteriosus. No significant difference (OR = 0. 98, 95% CI 0. 72 - 1. 33, P = 0. 881) was observed between the 2 groups in terms of the incidence of necrotic enterocolitis. Conclusions Patients with AOP should start caffeine treatment as early as possible, which significantly decreases mortality rate and the occurrence of other complications with little adverse reactions and good clinical tolerance.