OBJECTIVES : Peru presents high prevalence of overweight and obesity, 36% and 18%, respectively. Economic determinants play an important role in this context, but it’s unclear what’s the effect of the income over the BMI. Mostly, because there's a setting dependency in the association. With the aim to have a comprehensive analysis we estimate the income-elasticity of BMI using a categorical variable of wealth as proxy of the economic strata, to assess effect modification, and subsequently estimate the strata-specific elasticities.

METHODS : We used the National Household Survey (ENAHO) 2008, including sociodemographic and economic information at the household level, as well as weight and height at the individual level for adults over 15 years old. We used the DHS methodology to estimate the wealth index and later categorize it in five groups. Income was obtained from the expenditures approach at the household level. Weight, height, and income were transformed to their natural logarithms to be included in the regression model. Data was fit with GLM multilevel models to account for the sampling design.

RESULTS : We found the BMI has an income-elasticity of 0.0053 (p<0.001). This means, an increase of 1% in the income is associated with an increase of 0.005% in the BMI. We found that wealth is an effect modifier in the main association (p<0.001). Then, we fit separate models for each wealth category. In four out five quintiles, the income elasticity was 0.004, while in the fifth one there was no association.

CONCLUSIONS : We hope our findings help to increase the knowledge of how this association behaves in different economic context of the Peruvian population, and therefore identify population at risk to design and implement prevention interventions for obesity and overall nutrition problems.