Introduction:
Educational inequalities in smoking start early in life but there still some uncertainty regarding the time they emerge, and what explains their trends. This article investigates whether and how educational inequalities in smoking evolve from pre-adolescence until young adulthood.

Methods:
This study used data from the EPITeen Cohort, that recruited 13-year-old adolescents born in 1990 and enrolled at public and private schools of Porto, Portugal. For this analysis, we considered 1,038 participants, followed across four waves: 2003/2004, 2007/2008, 2011/2013, and 2014/2015. We modelled whether smoking behaviours were influenced by the educational attainment. We computed the odds ratio (OR) for smoking prevalence states (never smoking, experimenter, less than daily smoker, daily smoker and former smoker) and incidence transitions between these smoking states (never smoker to smoking experimenter; non-smoker, experimenter or less than daily smoker to daily smoker; non-smoker, experimenter or daily smoker to less than daily smoker; and daily or less than daily smoker to former smoker) as function of education, age, and adding interactions for age with education. These analysis were done separately for men and women.

Results:
The results showed that men with lower academical achievement were more likely to experiment until 17 years old (OR=0.83 for the interaction with 13 years old, and OR=0.66 for the interaction with 17 years old). However, those with higher academical achievement were more likely to experiment later, at 21 and 24 years old (OR for the interaction was 1.07). The inequalities in the prevalence of daily smoking emerged at 17 years old (OR for interaction with 17 years old was 0.42). Less of participants with higher academical achievement became daily smokers (OR for high education was 0.22 between 13 and 17 years old). Among women, the inequalities followed the same trends observed among men.

Conclusions:
Our results suggested that educational inequalities in daily smoking prevalence emerge at 17 years old and are formed in a cumulative way: by the risk of experimenting from 13 to 17 years old, and by the risk of becoming daily smoker from 17 to 21 years old. In order to tackle inequalities is essential to understand when they emerge across the life-course. These results really highlight that tobacco campaigns should focus not only the middle adolescence but also the late adolescence, from 13 to 21 years old.