Introduction:
Research indicates a positive association between cigarette smoking and disordered eating. However, much remains unknown about the impact of confounding, the order of cause and effect, and gender and socioeconomic status (SES) differences in the relationship. Thus, this study followed a large, population-based, mixed-gender sample from adolescence to midlife, applying a random intercept cross-lagged panel model (RI-CLPM) approach to examine the bidirectional prospective associations between smoking and disordered eating while ruling out the effect of unmeasured time-invariant confounders.

Material and Methods:
A population-based sample of 2936 participants from Norway (54.16% female) were assessed across five time points during early adolescence (T1; year 1992; Mage = 14.84 years; Cigarettes per day [M = 10.49, SD = 6.44]), late adolescence (T2; year 1994); emerging adulthood (T3; year 1999); adulthood (T4; year 2005); and later adulthood (T5; year 2020) applying RI-CLPMs. Multigroup analyses were used to examine differences by gender and SES.

Results:
The RI-CLPM model presented an excellent fit (Comparative Fit Index [CFI] = 0.99, Tucker-Lewis Index [TLI] = 0.99, and Root Mean Square Error of Approximation [RMSEA] = 0.02). The model differed by gender (Δχ2(21) = 39.687, p =.008) but not by SES (Δχ2(21) = 31.33, p = 0.06). Among women, high levels of disordered eating at T1 and T2 predicted more cigarette smoking at T2 and T3, respectively. Among men, disordered eating at T3 predicted less smoking at T4. No other significant cross-lagged effects emerged.

Conclusions:
The present study provides preliminary evidence for unidirectionality from disordered eating to cigarette smoking. These results suggest that successful interventions to reduce disordered eating may prevent or reduce smoking, regardless of SES, and especially by targeting female adolescents.