Overweight and Obesity in Brazilian Immigrants in Massachusetts, USA: A Time Series Analysis (2009–2020)

Obesity has increased worldwide and is now considered a global epidemic [1‒4]. According to the World Health Organization (WHO) [5], the worldwide prevalence of this condition has almost tripled between 1975 and 2016. In 2016, about 39% of adults over 18 years old were overweight and 13% obese. Developed countries remain the leaders in the rise of obesity, although similar upward trends are also observed in low- and middle-income countries [2, 4].

According to Malik, Willett, and Hu [2, 4], some low- and middle-income countries have experienced an increase in the trend toward overweight and obesity in recent years, driven by global free trade, economic growth, and urbanization, which favor the creation of obesogenic environments due to reduced physical activity and increased consumption of unhealthy foods such as refined grains, artificially sweetened beverages, processed products with low nutritional quality and high calorie content, and animal products. Obesity increases the risk of morbidity as well as mortality from chronic noncommunicable diseases, such as cardiovascular disease, diabetes mellitus, musculoskeletal diseases, and some types of cancer. The worldwide mortality associated with overweight and obesity is higher than that associated with underweight [3, 5].

As a country with low restrictions concerning food access or availability, the USA was one of the first countries to show signs that obesity became an epidemic, since it more than doubled between 1980 and 2010. In recent years, it has become stable at around 35%. However, its prevalence remains alarming and is now increasing among Hispanics and non-Hispanic blacks, indicating a disproportionate increase in different ethnic/racial groups [2, 3].

Immigrants are the fastest growing segments in the USA [6, 7]. It is, therefore, necessary to conduct studies that elucidate concerns related to their health, as well as responses aimed at public health planning and management, given the high costs that increasing overweight and obesity represent due to associated comorbidities [8].

Some studies demonstrate the high prevalence of overweight and obesity among Latinos, significantly associated with acculturation or integration in host societies, time of residence, and number of generations in the USA, as immigrants with shorter periods of residence in the country tend to present lower body mass index (BMI) when compared to Latinos who have lived in the country longer or who were born in the USA [6, 7, 9]. The emigration of Brazilians to the USA is currently increasing, while the states of Massachusetts and Florida are the main destinations for these immigrants [10]. As the Brazilian immigrant community is one of the largest in Massachusetts and overweight and obesity conditions have also increased in Brazil in recent decades [11], studies aimed at Brazilians are paramount. However, studies concerning Latino immigrants [6, 7, 9, 12] do not consider or do not analyze Brazilian immigrants as a separate group. Brazil is a large country, with wide cultural variations throughout its territory and colonized by the Portuguese, unlike other Latin American countries. Thus, they should be evaluated separately from Latinos due to relevant cultural differences (e.g., diets) based on country of origin.

Our study contributes to create knowledge on the health of Brazilian immigrants in the USA by describing the temporal trend of overweight and obesity among Brazilian patients cared for by the Cambridge Health Alliance (CHA) healthcare system in Massachusetts from 2009 to 2020. The main question guiding the study was what is the trend of overweight and obesity over 12 years in Brazilian immigrants in Massachusetts? Our hypothesis was that overweight and obesity trends in Brazilian immigrants cared for by the CHA healthcare system have increased over the last few decades.

This is an observational epidemiological, ecological time series study. The study population was Brazilian immigrant patients of the CHA network. This state has the second largest population of Brazilian immigrants in the USA, totaling 92,957 immigrants according to the 2019 American Community Survey [13]. The study sample consisted of 128,206 records of Brazilian immigrant patients aged 18–60 cared for at one of the CHA network units from 2009 to 2020. This article is part of a larger study that seeks to analyze metabolic syndrome and sleep components in Brazilian immigrants in Massachusetts. Thus, a cutoff point of 60 years old was established, considering the changes that occur in ontogenesis and metabolism related to aging.

The analyzed data were obtained from the patient CHA network database. The CHA healthcare system [14] is a network of hospitals and healthcare centers formed in 1996 from the union of two Massachusetts hospitals, the Cambridge Hospital and the Somerville Hospital. In 2001, the acquisition of an additional hospital (formerly Whidden Memorial Hospital, now CHA Everett Hospital) and associated services in Malden and Revere expanded this network, currently present in five Massachusetts cities (Cambridge, Somerville, Everett, Malden, and Revere). The CHA healthcare system has offered bilingual and bicultural English/Portuguese outpatient, hospital, emergency, and routine care for thousands of Brazilians for many years by.

Data were grouped considering a single consultation for each patient in each year, as the same patient’s appointments in different years could not be identified. The dependent variable was patient’s BMI, obtained from hospital records and categorized as normal weight, overweight, and obesity. The BMI classification followed the WHO [1]: eutrophy for BMI values between 18.50 and 24.99 kg/m², overweight for values between 25.00 and 29.99 kg/m², and obesity for values ≥30 kg/m². The independent variable time (attendance year) was used to identify potential trends during the study period; sex and age were covariates.

Descriptive analyses were initially performed to calculate the mean and standard deviation (SD) for the continuous variables, and the absolute and relative frequencies for categorical variables. Next, normal weight, overweight, and obesity rates were calculated, both gross and stratified by sex and age, for each year of the study period. For age stratification, the continuous variable was categorized into 18–32 years old, 33–47 years old, and 48–60 years old. Rates were obtained by applying the following formulas:

For the trend analysis, rate dispersion diagrams were initially plotted, followed by polynomial regression models when potential diagram relationships were observed. The rates in the polynomial regression models were included as the dependent variable, and time (year of visit) was the independent variable. Polynomial regression models frequently suffer from autocorrelation between the equation terms [15]. To reduce this effect, the variable “year” was transformed into “centralized year” by subtracting the midpoint of the study period (2014) from each year [15]. To reduce rate fluctuations and facilitate main trend interpretation, the rates were smoothed by applying the centralized moving average of order 3 [15].

Next, simple (Y = β0 + β1X), second degree (Y = β0 + β1X + β2X²), and third degree (Y = β0 + β1X + β2X² + β3X³) linear regression models were estimated. The choice of the best model was based on the best coefficient of determination (adjusted R²), followed by the statistical significance of the model and by the residual analysis to confirm the homoscedasticity assumption of the model. The lowest order model was chosen when two or more models were statistically similar. The trend was considered significant when p ≤ 0.05. Data were organized and coded in Excel spreadsheets for further statistical analyses. All analyses were conducted using SPSS® Statistics, version 23.0. All potential patient identifiers were removed prior to data access.

During the study period, a total of 128,206 records of Brazilian immigrants who sought care at CHA were analyzed. The population characteristics are presented in Table 1. Immigrants from 48 to 60 years old displayed the highest prevalence of both overweight and obesity, followed by those 33 and 47 years old and 18–32 years old.

We observed an increase in overweight and obesity rates among Brazilians who sought care between 2009 and 2020, with small fluctuations over the years, totaling a 7.7% increase for the overweight and 25.5% for the obese, while a 20.1% decrease was observed for normal weight during this period (Table 1). When considering the sex and age (suppl. File) of immigrants who sought care between 2009 and 2020, we observed eutrophy decreases for both men and women from 18 to 32 years old (5.9% for men and 16.1% for women) and from 33 to 47 years old (27.8% for men and 16.1% for women), while increases were noted for men and women from 48 to 60 years old (0.1% for men and 18.6% for women).

Regarding overweight rates, we found a decrease among men for all ages (18–32 years: 11.1%; 33–47 years: 6.6%; 48–60 years: 6.0%), while 19.0% and 14.2% increases were observed in the age groups from 18 to 32 and 33–47 years old, respectively, for women, with a 7.4% reduction noted from the 48–60 age group. On the other hand, obesity rates increased for both sexes in all age groups (men: 18–32 years: 29.3%; 33–47 years: 34.3%; 48–60 years: 8.9%; women: 18–32 years: 28.4%; and 33–47 years: 14.7%), with the exception of women aged 48–60 years, for whom there was a 4.7% decrease (suppl. File).

Regarding eutrophy, overweight, and obesity rates for every 1,000 records of Brazilian immigrants who sought care during the study period (Table 2; Fig. 1), we found a decreasing and constant trend for eutrophy rate, with a mean 7.2 decrease/year, and a statistically significant decreasing trend of overweight with a non-constant average annual increment of 3.0/year. As there is a negative acceleration of 0.6/year, the increment became negative at the end of the period. A statistically significant increasing trend in obesity was observed over the period, with a mean annual increase of 4.6/year, and a positive acceleration of 0.6/year.

When stratified for sex and age (Table 2; Fig. 2), eutrophy rates decreased for men and women of all age groups. Rather than a growing trend, overweight rates decreased for men from all age groups (18–32 years: mean annual reduction of 4.1/year; 33–47 years: mean annual reduction of 1.4/year, with negative acceleration of 1.4/year; 48–60 years old: mean annual reduction of 2.7/year), while there was an increasing trend for the 18–32 (mean annual increase of 4.2/year) and 33–47 (average annual increment of 3.8/year, with positive acceleration of 0.6/year) age groups for women, and decreased for the 48–60 age group (mean annual reduction of 3.4/year) (Table 2; Fig. 2).

An increasing trend in obesity rates for both men and women of all age groups was observed, except for the 48–60 age group. For men, the increase in the 18–32 age group was constant, while for the other age groups it was not constant, with an average annual increment of 5.3/year for the 18–32 age group, an average annual increment of 8.0/year and positive acceleration of 1.7/year for the 33–47 age group, and a negative annual increment of 1.0/year, but with positive acceleration of 1.4/year for the 48–60 age group. For women, the obesity trend increased and was statistically significant and not constant for all age groups (18–32: average annual increase of 3.7/year, with positive acceleration of 0.5/year; 33–47: average annual increase of 1.3/year, with positive acceleration of 0.7/year; 48–60: average annual reduction of 1.2/year, but with positive acceleration of 0.8/year) (Table 2; Fig. 2).

This study raises hypotheses on the increasing overweight and obesity trend over the years observed in a population of Brazilian immigrant patients of the CHA healthcare system. We chose the main overweight and obesity trends because they were the most significant in the time series. Curve fluctuations may reflect the duration of the time series.

Between 2009 and 2020, there was an increase in the prevalence of overweight and obesity conditions among Brazilian patients. Increases in overweight and obesity also happened in national and state statistics in both the USA and Brazil. In Massachusetts, Centers for Disease Control and Prevention (CDC) data [16] show an increase in overweight and obesity for both the overall state population (1.1% overweight and 11.9% obesity increases, respectively) and Latinos (0.9% overweight and 16.0% obesity increases, respectively) from 2009 to 2020. During this period, the average overweight prevalence for the Massachusetts population and Latinos was 35.7% (SD = 0.7) and 36.7% (SD = 3.0), respectively, while the average obesity prevalence was 23.9% (SD = 1.2) for the state population and 29.6% (SD = 4.2) for Latinos.

According to Vigitel Brazil data¹, the prevalence of overweight and obesity has increased in the country from 2009 to 2020, 23.4% for overweight and 54.7% for obesity, respectively [11, 17]. In addition, the mean overweight and obesity prevalence during the same period was 52.3% (SD = 3.4) (CI = 46.6–57.5%) and 18.0% (SD = 2.2) (CI = 13.9–21.5%), respectively.

The percentage increases in overweight (7.7%) and obesity (25.5%) rates for Brazilian patients during the study period are higher than those observed for the Latino population in Massachusetts and lower than those for the Brazilian population. However, the mean overweight (38.1%, SD = 1.1) and obesity (24.8%, SD = 2.0) prevalence approached the means for the US population. The average overweight prevalence was higher than that for both the general Massachusetts population and Latinos, while the average obesity rate is close to that noted for the Massachusetts population, although lower than that of Latinos. This finding is in line with other studies conducted on other Latino immigrant populations, where lower obesity rates in immigrants compared to Latinos in the USA are described [6, 7, 9, 12].

There was higher obesity prevalence in the Brazilian immigrant patient population at CHA compared to Brazil in the time series, which may indicate a change from the Brazilian to the US lifestyle norms, such as changes in eating habits and physical activity [6, 18]. Martínez et al. [19] highlight the predisposition for weight gain that may occur among Latino immigrants due to the nutritional transition that many Latin American countries, including Brazil, have shown in recent decades. Thus, the shift from a country where increasing overweight and obesity rates are intense and constant, to another where the prevalence of obesity is even greater, may increase the probability of acculturation to obesogenic food environments and sedentary lifestyles.

The 2019 National Health Survey [20] conducted by the Brazilian Institute of Geography and Statistics (IBGE) and CDC data for Massachusetts [16] in the same year also showed similarity between the estimated prevalence of obesity for the Brazilian and Massachusetts populations and that estimated for Brazilian patients, and higher prevalence for Latinos in Massachusetts compared to Brazilian patients. Brazilians emigrate to the USA in search of better living conditions for themselves and their families [10], a reason historically recognized as an immigration driver. To improve their income, they accumulate jobs and often send remittances to family members who remain in Brazil [21]. This situation contributes to limiting sleep time, physical activity, and the ability to maintain a healthy diet, due to the higher cost of unprocessed foods in the USA.

The overweight and obesity prevalence among men and women patients suggests that overweight, followed by obesity, was more prevalent among men, while women showed a more prevalent normal weight, followed by overweight and obesity. This differs from findings reported in other studies [7, 22] where obesity was the highest prevalence for women. However, our findings may indicate differences between men and women regarding concern and care for their own health [23, 24].

The trends observed in this study indicate decreasing eutrophic and overweight rates and increasing obesity rates. However, when comparing the percentage increases for the two rates, obesity was higher. This finding is similar to the global trend of increasing obesity [25] and may reflect changes in lifestyle in a society that works 24 h a day [26].

There was a decreasing eutrophy trend for both men and women in all age groups, as well as an increasing overweight trend for women between 18 and 47 years old and a decreasing trend for men in all age groups. In addition, there was increasing obesity for both sexes in all age groups, though more intense for men aged 33–47 years. Those trends suggest that the transition from eutrophic to overweight and obese in both men and women is consistent with findings from other studies [12, 18]. Furthermore, they indicate the need for studies with more complex designs that identify individual influencing factors as well as the implementation of interventions aimed at reducing overweight and obesity in this age group.

The lower prevalence as well as the slower increase in obesity among women may indicate that women are more concerned with their own health compared to men and seek help to maintain and/or lose weight. A similar result was found in another study [12] that investigated the association between acculturation and obesity in immigrants from Mexico. Chrisman et al. [12] noted a lower prevalence of obesity in women displaying higher levels of acculturation, which assumes that the greater the acculturation among women, the more they adopt the US norm about ideal weight and find ways to reduce their own weight.

National statistics also indicate overweight reduction among men (a 0.5% reduction) and increase among women (increase of 2.1%) for the Massachusetts population during the study period, with a mean prevalence of 42.8% (SD = 1.3) for men and 28.7% (SD = 1.0) for women. The data reveal an increase in obesity among both men (increase of 2.8%) and women (increase of 22.5%), although more intense among women, contrary to what we found in our study, where the average obesity prevalence was 25.5% for men (SD = 1.2) and 22.3% for women (SD = 1, 6) [16].

In Brazil, Vigitel data from 2009 to 2020 also indicate increases in both overweight and obesity for men (overweight: 15.5%, obesity: 48.2%) and women (overweight: 32.9%, obesity: 61.4%). However, the increases were more intense and the annual prevalence higher for obesity among women. Among men, the mean overweight and obesity rates were 55.6% (SD = 2.6) and 17.4% (SD = 2.0), respectively, while for women the mean overweight prevalence was 49.4% (SD = 4.2) and 18.5% for obesity (SD = 2.5) [11, 17].

The male and female obesity prevalence average for immigrants is higher than that of both Brazilian men and women. The decreasing trends in overweight observed for both women and men aged 48–60 may reflect better healthcare due to aging when other chronic diseases, such as diabetes and hypertension, are associated with weight gain and lead to changes in lifestyle [27]. Prasad, Sung, and Aggarwal [27] emphasize that chronic diseases generally appear during middle age as a result of long-term exposure to unhealthy lifestyle behaviors, such as smoking, alcohol, unhealthy food consumption, sedentary life, and stress.

However, the curves observed for both men and women in this age group may be inconsistent considering the increasing trend of obesity, the decrease of eutrophy, and the increase in the prevalence of overweight and obesity with aging, as immigrants between 33 and 60 years showed the highest prevalence. This finding is consistent with the literature [28, 29], which indicates that weight gain in adults tends to occur progressively during middle age. In the USA, the average weight gain among adults ranges between 0.5 and 1 kg per year.

The overweight prevalence for the population of Massachusetts [16] increased in all age groups², ranging from 0.5% to 5.9%, with the exception of the age group from 35 to 44 years that had a 0.3% decrease. Obesity also increased in all age groups, with the exception of individuals from the 25–34 age group who displayed a 3.6% decrease. The greatest increase observed in the state was for the 35–44 age group (22.5%). In contrast, the Vigitel data [11, 17] indicate both overweight and obesity increases in all age groups from 2009 to 2020 (overweight variation of 8.9% (55-54 years old) to 30.0% (25–34 years) and obesity variation from 26.6% (55–64 years) to 43.5% (18–24 years old)), except for overweight for the 18–24 age group, which had a 1.9% decrease. The obesity increases with aging we found corroborate both data from Massachusetts [16] and Vigitel Brazil [11, 17], which also indicate a gradual increase with aging.

Our study had a few weaknesses. First, data from the studied population were grouped, which makes it impossible to control for individual factors that could justify the results, such as length of US residence, immigrant generation, and age at the time of migration, often associated with overweight and obesity increases in immigrants, as well as the effect of healthy immigrants in this population [6, 7, 9, 12]. Second, the self-reported data from the national surveys employed for data comparisons may lead to weight underestimation among women and height overestimation among men [30], unlike our findings, which were based on medical records. Finally, the Berkson bias may have occurred, as the investigated individuals are a hospital patient population. Because we included a hospital population, with possible interaction of other health conditions, this may not be representative of the general population.

Our study identified decreasing eutrophy and overweight and increasing obesity trends in Brazilian immigrant patients in Massachusetts from 2009 to 2020. Obesity rates were higher than the prevalence observed in Brazil and close to that observed for the Massachusetts population, confirming our initial hypothesis. The same trend was observed in the sex stratified analyses. Male patients showed higher overweight and obesity prevalence than women. We also found a gradual increase in obesity with aging and a higher prevalence in the age group 48–60 years old.

Our findings describe the evolution of overweight and obesity over time in a large population of Brazilian immigrants. It raises new hypotheses to be tested in future studies focused on the individual level, where factors such as immigrant acculturation to host societies may be explored. Finally, it may contribute for the development of health promotion and prevention actions for the Brazilian immigrant population in the USA.

The study followed the precepts of Ethics in Research with Human Beings (National Health Council of Brazil, Resolution 466/2012 [31]), was approved by the ENSP Research Ethics Committee (CEP ENSP) (CAAE 15012319.3.0000.5240), in Rio de Janeiro, and was granted an exemption from requiring ethics approval by the Institutional Review Board (IRB) of the CHA network in Cambridge, MA. Since the study used secondary data and has not directly interviewed any subject, the written informed consent was not required.

The authors have no conflicts of interest to declare.

This study was partially supported by the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior – Brasil (CAPES) – Finance Code 001.

Talita Monsores Paixão developed the aim of the study, received data access, performed data analysis, discussed the data, and wrote and approved the final version of the article; Carlos Eduardo Gomes Siqueira developed the aim of the study, articulated obtaining access to the database, discussed the data, assisted in writing, and approved the final version of the article; Elisa Tristan-Cheever and Kirsten Meisinger articulated obtaining access to the database, discussed the data, helped in writing the article, and approved the final version of the article; Frida Marina Fischer, Maria Carmen Martinez, and Marcelo Augusto Christoffolete guided the statistical analysis and data interpretation, discussed the data, and approved the final version of the article; Liliane Reis Teixeira developed the aim of the study, guided the statistical analysis and data interpretation, discussed the data, helped in writing the article, and approved the final version of the article.

All data generated or analyzed during this study are included in this article. Further inquiries can be directed to the corresponding author.