Nutritional status has evolved in a dual trend worldwide: underweight has become a minor or local issue while overweight or obesity has risen to play a major role in the global burden of disease. In 2014, Brazil was ranked as the third country with the highest absolute number of obese men. Our aim was to estimate trends of underweight and obesity among Brazilian adults using a comprehensive set of surveys from 1974 to 2019. The data used in the study originate from subjects aged ≥ 18 in six Brazilian national surveys, presented in chronological order: Brazilian National Survey on Household Expenses (ENDEF 1974-1975); Brazilian National Survey on Health and Nutrition (PNSN 1989); Brazilian Household Budget Survey (POF 2002-2003, 2008-2009); and Brazilian National Health Survey (PNS 2013 and 2019). All six surveys were designed to sample household complexes that were representative of the Brazilian population. Body mass index was calculated (kg/m2). The nutritional status of individuals was classified following the standards. We have modeled obesity trend according to income and education strata. The trajectories of underweight and obesity over time in Brazil draw the classical “X” of nutrition transition. From 1975 to 2019 underweight has decreased from 9.1% to 2.5% among men and 12.2% to 3.4% among women. On the other hand, obesity trajectories have scaled up from 3% to 22% among men and from 9% to 30% among women. The increase in obesity rate is directly and negatively proportional to income quintiles. Sociodemographic (income and education) improvement is associated with an increase in obesity. All public policies intending to stop the obesity spread in Brazil have been ineffective or too small to be effective.

Chronic diseases, worse socioeconomic conditions and old age can increase infection and hospitalization rate due to Coronavirus disease (COVID-19). We assessed the association between the burden of NCDs and the occurrence of infections and hospitalizations of COVID-19 in Brazil in a large COVID-19 national survey data. We analyzed only data collected between July and November 2020 (n = 1,071,782). The frequencies of positive COVID-19 diagnosis and NCD burden were estimated according to age, sex, socioeconomic strata and skin color categories. We estimated hazard ratios and 95% confidence intervals using Cox regression models. There is a non-linear dose-response inverse association between per capita income and the rates of infection and hospitalization due to COVID-19. The presence of NCDs was associated with a higher incidence of COVID-19 infection (HR1NCD = 1.34; 95% CI: 1.26; 1.43; HR2 or more NCD= 1.54 95% CI: 1.39; 1.71) and incidence of hospitalization (HR1NCD = 3.08 95% CI: 2.26; 4.19; HR 2or more NCD= 6.81 95% CI: 4.88; 9.49).The difference between the risks of infection or hospitalization of COVID-19 attributable to the burden of NCDs is non-linearly associated with the income.