Background Extremely low and very low gestational age (ELGA and VLGA) constitutes a risk factor for development even in absence of cerebral damage, as an immature central nervous system is exposed to invasive and inadequate stimulation We tested the hypothesis that GA impacts developmental outcomes and trajectories of preterms without major cerebral damage in the first 2 years of life, expecting poorer developmental outcomes and higher rate of impairment with the decreasing of GA We also evaluated whether GA, together with developmental outcomes in the first year of life, was related to developmental outcomes at 24 months Methods Eighty-eight infants, divided into three GA groups (ELGA <= 28 weeks, VLGA 29-32 weeks, full term >37 weeks) were assessed longitudinally at 6 12, 18 and 24 months using the Griffiths Mental Development Scales Results Use of a repeated measure multivariate analysis of variance resulted in several significant findings GA was associated with the developmental quotient (DQ) scores (P = 0 006) and locomotor (P < 0 001), eye and hand co-ordination (P = 0 016) and performance (P = 0 040) sub-scale quotient (SQ) scores, age of evaluation was also associated with DQ scores (P = 0002), and locomotor (P < 0 001) and performance (P < 0 001) SQ scores In particular, ELGAs exhibited lower DQ and SQ scores compared with the VLGA and full-term groups, some ELGAs showed mild, moderate or severe cognitive impairments, while few VLGAs mild impairments Linear regression analysis showed that GA (P = 0 034) and 12-month developmental outcome (P < 0 001) were related to 24-month developmental outcome Conclusions Different developmental trajectories emerged in relation to GA, with poorer developmental outcomes and higher rates of impairment in ELGAs and few mild impairments in VLGAs The relevance of taking into account both GA and repeated assessments in the first 2 years of life was shown