It has been hypothesised that increased asthma prevalence in westernised countries is associated with changes in lifestyle factors, including a poorer diet. However, little is known regarding the association between diet quality and asthma. In the diet–asthma association, the role of BMI as a potential mediator needs clarification; moreover, potential effect modification by non-diet sources of oxidants, such as smoking, merits investigation. We investigated the association between diet quality and change in asthma symptoms, as well as assessed effect modification by smoking, while accounting for BMI as a potential mediator. Using data from the French prospective Epidemiological study on the Genetics and Environment of Asthma study, we assessed diet quality using the Alternate Healthy Eating Index 2010 (AHEI-2010) at baseline and change in asthma symptoms (stable (reference), worsening, improved; mean follow-up time: 7 years). Mediation analysis was used to disentangle total and direct effects and the indirect effect mediated by BMI. The analyses included 969 adults (mean age 43 years; 49 % men; 42 % ever asthma). We observed a significant interaction between smoking and AHEI-2010 on change in asthma symptoms (P
for interaction=0·04). Among never smokers (n 499), we observed a positive total effect (multivariable OR 1·39; 95 % CI 1·07, 1·80) and a positive direct effect (OR 1·41; 95 % CI 1·09, 1·80) of the AHEI-2010 (per ten-point increment) on improved symptoms. No indirect effect mediated through BMI was observed (OR 0·99; 95 % CI 0·91, 1·07). Among former and current smokers, all effects were statistically non-significant. Better diet quality was associated with improved asthma symptoms over time in never smokers, independently of BMI.