AIM: To characterize temporal pattern of resolution and recurrence of naive choroidal neovascularization (CNV) secondary to wet age-related macular degeneration (AMD) treated with intravitreal bevacizumab on as needed regimen, and to analyze baseline risk factors for CNV resolution or recurrence.METHODS:Ninety-one eyes of 80 patients with newly diagnosed wet AMD were retrospectively studied. All eyes were treated with a round of three monthly intravitreal bevacizumab injections, followed by one additional ‘bonus’ injection after resolution of CNV activity. During follow-up, eyes were monitored with fluorescein angiography, optical coherence tomography, and best-corrected visual acuity (BCVA). In case of recurrences of CNV activity, eyes were retreated with other rounds of bevacizumab injections following the same treatment protocol.RESULTS:Over a median follow-up of 532d, the median resolution time of CNV activity in the first, second, and third treatment round was 98d, 126d, and 111d, respectively. The median recurrence time for the three rounds was 154d, 126d, and 151d, respectively. No significant difference in resolution time (P=0.09) or in recurrence time (P=0.11) was detected among treatment rounds. Age (P=0.0082) and lens status (P=0.035) were found to be associated with CNV resolution; for every 1-year increase in age there was 4% greater chance of CNV resolution; Phakic eyes demonstrated a 33% better chance to experience CNV resolution than pseudophakic eyes. For CNV recurrence, lens status (P=0.0009) and gender (P=0.0446) were found to be predictive; pseudophakic eyes had a 3.69-fold greater risk to experience recurrence of CNV activity compared to phakic eyes; males had a 2.19-fold greater risk to experience recurrence of CNV activity than females. No significant BCVA changes among three treatment rounds were noted (P=0.56).CONCLUSION:Resolution time and recurrence time of CNV activity were not significantly different among treatment rounds, suggesting absence of tachyphylaxis to bevacizumab. A cautious decision should be made upon discontinuing treatment in wet AMD eyes of younger or pseudophakic patients, which showed slower response to bevacizumab. In addition, wet AMD eyes of male or pseudophakic patients should be evaluated more carefully after stopping the treatment, because they may have early reactivation of the CNV. BCVA was preserved by bevacizumab treatment despite multiple recurrences.