Differentiating the preterm phenotype: Distinct profiles of cognitive and behavioural development following late and moderately preterm birth.

Objectives: To explore patterns of comorbidity in cognitive and behavioural outcomes at two years corrected age among children born late and moderately preterm (LMPT) and to identify predictors of different patterns of comorbidity in this population. Study design: Geographical prospective population-based cohort study of 1139 LMPT (32+0 to 36+6 weeks’ gestation) and 1255 term-born (37+0 to 42+6 weeks’ gestation) babies. Parent questionnaires were used to identify impaired cognitive and language development, behaviour problems, delayed social-emotional competence, autistic features and clinicallysignificant eating difficulties at 24 months corrected age for 638 (57%) LMPT and 765 (62%) term-born children. Results: Latent Class Analysis revealed two classes of outcomes among the term group: optimal outcome (Class I: 84%) and non-optimal outcome (Class 2: 16%). In contrast, three classes were identified in the LMPT group: optimal outcome (Class 1: 67%), non-optimal outcome (Class 2: 26%), and an additional preterm phenotype (Class 3: 7%). Non-white ethnicity, socio-economic risk and not receiving breast milk at hospital discharge were risk factors for non-optimal outcome in both groups. Male sex, higher gestational age and preeclampsia were only associated with the preterm phenotype. Conclusions: Only a small proportion of LMPT born children have cognitive and behavioural problems that are consistent with the very preterm phenotype and which are likely to have arisen through a preterm pathway. A larger proportion have a profile of problems that correspond with those observed in children born at term. This study advances understanding of the long term risks attached to birth at late and moderately preterm gestations.