Background:
Given the gender differences that permeate Tobacco Use Disorders (TUD), we aimed to identify phenotypes of TUD in female and male in a representative sample of smokers in a developing country.

Methods:
Data came from lifetime weekly smokers ages 18 and older taking part in the São Paulo Megacity Mental Health Survey collected between 2005-2007 (n = 1,386). Latent class analysis (LCA) was performed on nine TUD criteria stratified by gender. Logistic regression models explored the association between latent classes and socio-demographic and psychiatric variables. All analysis were performed using Mplus taking into account sampling weights and complex survey design features.

Results:
The best fitting LCA model had three classes within a severity continuum in both genders: a “non-symptomatic class”(Women[W]:30.2% , Men [M]: 44.4%), a “moderate symptomatic class”(W:45.5% , M:30.7%), and a “high-moderate symptomatic class”(W: 24.3%, M: 24.7%). Respondents in the “moderate symptomatic class” were more likely to have higher household income among female, and high-average education among male than those in the “non-symptomatic class”. Both women and men in the “high-moderate symptomatic class” were more likely to have past-year anxiety, but only men were more likely to have past-year insomnia than those in the “non-symptomatic class”.

Conclusions:
Both men and women smokers are divided into three TUD phenotypes, with approximately a quarter being highly dependent, which is associated with psychiatric comorbidity. However, women are more vulnerable to experiencing lifetime TUD symptomatology: 70%, in contrast to 56% of men. The intermediate symptomatic phenotype in women was associated with high income, which may show an important differential in the smoking behaviors paradigm in this developing country.