Background Diet is an important risk factor for cardiovascular diseases and shows well-established socioeconomic patterning among adults. However, less clear is how socioeconomic inequalities in diet develop across the life course. This study assessed the associations of socioeconomic trajectories (SETs) during early adulthood (16-24 years) with diet quality in mid-adulthood (46 years), independent of childhood socioeconomic position (SEP) and mediated by mid-adulthood SEP.

Methods Participants from the 1970 British Cohort Study with socioeconomic data in early adulthood were included (n=12,434). Diet quality at age 46 years, evaluated using the Mediterranean diet pyramid, was regressed on six previously identified classes of early adulthood SETs, including a Continued Education class entering higher education between ages 16 and 24 years, four occupation-based SET classes defined by occupational social class, and an Economically Inactive class. Causal mediation analyses tested the mediation of the association between early adulthood SETs and adult diet quality by household income and neighbourhood deprivation at age 46 years separately. Models were adjusted for sex, childhood SEP, adolescent diet quality and adolescent health.

Results The early adulthood SET of Continued Education showed the best diet quality at age 46 years, with the predicted mean of the Mediterranean diet score (range 0-15) as 6.50 (95% CI 6.32, 6.67). In contrast, there was little difference in adult diet quality among the other five SET classes, including Managerial Employment (mean 6.03, 95% CI 5.87, 6.18), Skilled Non-manual Employment (mean 5.92, 95% CI 5.76, 6.08), Skilled Manual Employment (mean 5.87, 95% CI 5.71, 6.02), Partly Skilled Employment (mean 5.82, 95% CI 5.66, 5.99), and Economically Inactive (mean 5.76, 95% CI 5.52, 6.00). The observed association of the Continued Education class with adult diet quality was independent of parental SEP in childhood, and was largely not mediated by household and neighbourhood SEP in mid-adulthood. While the Continued Education class was related to higher household income and lower neighbourhood deprivation levels at age 46 years, household income was not and neighbourhood deprivation was only weakly associated with adult diet quality after adjusting for early adulthood SETs.

Conclusion Early adulthood SETs independently contribute to adult diet quality, with continuing education associated with better adherence to Mediterranean diet. Early adulthood is therefore a critical period for intervention to prevent the development of socioeconomic inequalities in diet in later life. Further research is needed to identify the causal process underlying the association of early adulthood SETs with adult diet quality.