To compare neurodevelopmental outcomes of extremely preterm infants diagnosed with patent ductus arteriosus (PDA) who were treated medically or surgically and those who were not diagnosed with PDA or who did not undergo treatment for PDA.This retrospective population-based cohort study used data from a geographically defined area in New South Wales and the Australian Capital Territory served by a network of 10 neonatal intensive care units. Patients included all preterm infants born at <29 completed weeks of gestation between 1998 and 2004. Moderate/severe functional disability at 2-3 years corrected age was defined as developmental delay, cerebral palsy requiring aids, sensorineural or conductive deafness (requiring bilateral hearing aids or cochlear implant), or bilateral blindness (best visual acuity of <6/60).Follow-up information at age 2-3 years was available for 1473 infants (74.8%). Compared with infants not diagnosed with a PDA or who did not receive PDA treatment for PDA, those with medically treated PDA (aOR, 1.622; 95% CI, 1.199-2.196) and those with surgically treated PDA (aOR, 2.001; 95% CI, 1.126-3.556) were at significantly greater risk for adverse neurodevelopmental outcomes at age 2-3 years.Our results demonstrate that treatment for PDA may be associated with a greater risk of adverse neurodevelopmental outcome at age 2-3 years. This was particularly so among infants born at <25 weeks gestation. These results may support permissive tolerance of PDAs; however, reasons for this association remain to be elucidated through carefully designed prospective trials.