Background Although early life education for improved long-term health and the amelioration of socioeconomically generated inequalities in chronic disease is advocated in influential policy statements, the evidence base is very modest.

Aims To address this dearth of evidence using data from a representative UK national birth cohort study.

Methods The analytical sample comprised men and women in the 1958 birth cohort study with prospectively gathered data on attendance at nursery or primary school before the age of 5 years who had gone on to participate in social survey at 42 years (n=11 374), or a biomedical survey at 44/5 years of age (n=9210), or had data on vital status from 18 to 55 years (n=17 657).

Results Relative to study members who had not attended nursery, in those who had, there was in fact a higher prevalence of smoking and high alcohol intake in middle age. Conversely, nursery attenders had more favourable levels of lung function and systolic blood pressure in middle age. This apparent association between nursery attendance and lower systolic blood pressure was confined to study members from more deprived social backgrounds of origin (P value for interaction 0.030). There was no apparent link between early school attendance and any behavioural or biological risk factor. Neither nursery nor early school attendance was clearly related to mortality risk.

Conclusions We found no clear evidence for an association of either attendance at nursery or primary school before the age of 5 years and health outcomes around four decades later.