OBJECTIVE: To study the neurodevelopmental outcomes of multiple (twins, triplets, quads) compared with singleton extremely preterm infants <29 weeks gestation. DESIGN: Population-based retrospective cohort study. SETTING: A network of 10 neonatal intensive care units in a geographically defined area of New South Wales and the Australian Capital territory. PATIENTS: 1473 infants <29 weeks gestation born between 1 January 1998 and 31 December 2004. INTERVENTION: At 2-3 years of corrected age, a neurodevelopmental assessment was conducted using either the Griffiths Mental Developmental Scales or the Bayley Scales of Infant Development II. MAIN OUTCOME MEASURE: Moderate-severe functional disability was defined as developmental delay (Griffiths Mental Developmental Scales General Quotient or Bayley Scales of Infant Development-II Mental Development Index >2 SDs below the mean), moderate cerebral palsy (unable to walk without aids), sensorineural or conductive deafness (requiring amplification) or bilateral blindness (visual acuity <6/60 in the better eye). RESULTS: Of the 1081 singletons and 392 multiples followed-up, singletons demonstrated higher rates of systemic infections, steroid treatment for chronic lung disease and birth weight <10th percentile. Moderate-severe functional disability did not differ significantly between singletons and multiples (15.8% vs 17.6%, OR 1.14; 95% CI 0.84 to 1.54; p=0.464). Further subgroup analysis of twins, higher-order gestations, 1st-born multiples, 2nd or higher-born multiples, same and unlike gender multiples, did not demonstrate statistically higher rates of functional disability compared with singletons. CONCLUSIONS: Premature infants from multiple gestation pregnancies appear to have comparable neurodevelopmental outcomes to singletons.