The aims of this study were to determine, in a cohort of extremely premature infants, the prevalence of iron deficiency identified by zinc protoporphyrin/heme ratio (ZPP) testing, and its association with neurodevelopmental problems and dietary risk factors for iron deficiency. Infants of less than 29 weeks' gestation or less than 1000 g birth weight were studied prospectively at a multidisciplinary follow-up clinic. Assessments were made at a corrected age of either 12 months (n=72) or 2 years (n=69). Physical examination, Griffiths Developmental Scale, and neurosensory-motor assessment were administered, information on diet and behaviour was obtained by questionnaire, and a fingerprick ZPP ratio was performed to identify iron deficiency. 18.4% of infants had positive ZPP tests. There was no significant association between a positive ZPP test result and dietary risk factors, or symptoms of lethargy, irritability or poor attention. In children without cerebral palsy, there was no difference on Griffiths scores or neurosensory-motor assessment between ZPP-positive and ZPP-negative groups. The diagnosis of cerebral palsy (n=12) was significantly associated with both a positive ZPP test and a lower Griffiths general quotient (GQ) score. Iron deficiency occurs commonly in extremely low birth weight (ELBW) children in early childhood, and is not predicted by dietary risk factors. The prevalence of iron deficiency is increased in ELBW children with cerebral palsy. Non-anaemic iron deficiency (NAID) does not impair development or significantly affect behaviour of ELBW subjects who do not have cerebral palsy.