Systematic Review and Meta-Analysis of Metabolic Syndrome and Its Components in Latino Immigrants to the USA

The Metabolic Syndrome (MetS) is an increasingly prevalent condition globally. Latino populations in the USA have shown an alarming increase in factors associated with MetS in recent years. The objective of the present systematic review was to determine the prevalence of MetS and its risk factors in immigrant Latinos in the USA and perform a meta-analysis of those prevalence. The review included cross-sectional, cohort, or case–control studies involving adult immigrant Latinos in the USA, published during the period 1980–2020 in any language. Studies involving individuals who were pregnant, aged <18 years, immigrant non-Latinos, published outside the 1980–2020 period, or with other design types were excluded. The Pubmed, Web of Science, Embase, Lilacs, Scielo, and Google Scholar databases were searched. The risk of bias was assessed using the checklists of the Joanna Briggs Institute. The review included 60 studies, and the meta-analysis encompassed 52 studies. The pooled prevalence found for hypertension, diabetes, general obesity, and abdominal obesity were 28% (95% Confidence Interval (CI): 23–33%), 17% (95% CI: 14–20%), 37% (95% CI: 33–40%), and 54% (95% CI: 48–59%), respectively. The quality of the evidence of the primary studies was classified as low or very low. Few studies including immigrants from South America were identified. Further studies of those immigrants are needed due to the cultural, dietary, and language disparities among Latin American countries. The research protocol was registered with the Open Science Framework (OSF).


Introduction
The last few decades have seen a shift in the morbidity-mortality profile of the population globally, first in developed countries and then in developing nations. The epidemiological transition theory was first put forward by Omran [1] in 1971. According to this theory, the transition stems from long-term changes in patterns of mortality, and illnesses caused by infectious diseases are gradually giving way to an increased occurrence of chronic-degenerative diseases as the leading cause of morbidity and death.
According to Omran [1], the transition from the predominance of infectious-contagious diseases to chronic diseases involves three main groups of determinants: ecobiological determinants, socioeconomic, political and cultural determinants, and public health determinants.
There is currently a steady increase in cardiovascular disease, diabetes, cancers, and obesity. Regarding obesity, it is now considered a global epidemic, with the United States (USA) numbering as one of the first countries to show that rising obesity was becoming Immigrants from developing countries undergoing a process of nutritional transition who emigrate to developed countries with a highly obesogenic environment are subject to an acceleration in the process of nutritional transition, favoring the development of obesity and its associated factors [17]. Nonetheless, when assessing such factors, it is important that cultural, language, economic, and social characteristics of the immigrants be taken into account, as well as origin.
Another key point is the very high likelihood that Brazilian immigrants were not included in samples of most studies conducted in the USA, owing to the classification adopted by the USA Census for the Latino population. Individuals with Cuban, Mexican, or Puerto Rican background/origin are usually classified as Hispanic or Latinos, whereas those from other Central and South American countries are identified as "others with Hispanic, Latino, or Spanish origin" [20]. Thus, the generalization for Brazilian immigrants of health information identified in studies of immigrants in the USA, using the term "Latino" as a reference, is not appropriate unless one can ascertain that Brazilians were included in the data sources.
Given the growth in diseases associated with MetS in immigrant Latinos in the USA and the epidemiological transition that has taken place and is still underway globally, in which chronic diseases feature as the leading cause of morbidity-mortality, identifying the publications involving this population to accurately determine the prevalence of MetS and its components is crucial.
However, the vast majority of publications on health conditions of immigrant Latinos in the USA address only immigrants from Mexico and Central America. Thus, how or whether South American immigrants are included in those studies remains unclear. Taking into account the cultural and economic heterogeneity that exists across Latin America, where Brazil has different language, culture, customs, and eating habits from other Spanishspeaking nations, analyses centering on the Brazilian population are important.
Therefore, the objective of our systematic review was to determine the prevalence of MetS and its risk factors among immigrant Latinos in the USA and perform a meta-analysis of those rates. This review also sought to ascertain which Latino groups are included and the extent to which Brazilians feature in those published studies.

Protocol
Our systematic review was carried out according to the PICO (Population, Intervention/Exposure, Comparator(s)/Control and Outcomes) strategy, where: • Population: Immigrants from Latin America, residing in the USA, aged >18 years; • Intervention/Exposure: Immigration; • Comparator(s)/control: US-born Latino population; • Outcomes: MetS and/or its components (primary) and sleep disorders (secondary).
The study protocol was based on the criteria established by PROSPERO, an International Prospective Register of Systematic Reviews, and registered with the Open Science Framework (OSF) before starting the review (https://doi.org/10.17605/OSF.IO/JFM7G (accessed on 7 January 2023)). The systematic review yielded sufficient data to perform a meta-analysis of prevalence, which was subsequently incorporated into the review. The systematic review was reported according to Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 statement [21] (Tables S1 and S2).

Definition of Disease and Disease Codes
The primary outcomes included in our review were the MetS and/or its risk factors: high blood pressure or diagnosis/treatment of arterial hypertension; high fasting blood glucose or diagnosis/treatment of type 2 diabetes mellitus; low HDL-c or treatment for low HDL-c; high triglycerides or treatment for high triglycerides; and high waist circumference or BMI > 30 kg/m 2 [7,9]. Sleep disorders were the secondary outcomes.
Based on the codes of the International Statistical Classification of Diseases and Related Health Problems (ICD-10) [22], the relevant categories for the review were: E88.8-other specified metabolic disorders (MetS), I10-essential hypertension (primary), E78-disorders of lipoprotein metabolism and other lipidemias, E11-type 2 diabetes mellitus, E66-obesity, and G47-sleep disorders.

Search Strategy
The search strategy was devised to include all important descriptors needed to retrieve the relevant studies for the review (Table S3).
The PubMed, Web of Science, Embase, Lilacs, and Scielo databases were searched, along with Google Scholar as a complementary source, with no restrictions regarding language of publication. The search strategy was applied for the last time on 3 June 2020.

Study Selection
Studies that met the inclusion criteria listed below were selected for inclusion in the review: • Cross-sectional, cohort, or case-control type study; • Involving adults (>18 years); • Investigating immigrant Latinos residing in the USA; • Published during the period 1980-2020 in any language.
Studies involving pregnant women, immigrants from countries other than Latin America, children and adolescents (<18 years), published outside the 1980-2020 period or based on data collected prior to 1980, qualitative or case studies, and those with non-original data were excluded from the review.
Two reviewers (TP, DS) independently screened titles and abstracts of the articles retrieved using the search strategy by applying the Rayyan [23] app and Excel ® spreadsheets. The full texts of the articles selected were then obtained, reviewed, and categorized as "included" or "excluded" by the same reviewers (TP, DS) in a double-blind analysis, according to the predefined inclusion criteria. The reasons for exclusion were documented using the Rayyan app [23]. At both stages, any discrepancies were checked by a third reviewer (LT).

Data Extraction
For all articles included in the previous stage, three reviewers (TP, DS, CN) extracted the data independently using an extraction spreadsheet previously created and validated by the research team, containing the following parameters: •

Risk of Bias
The risk of bias of the studies included was assessed using the checklists of the Joanna Briggs Institute (JBI) titled "Critical Appraisal Checklist for Analytical Cross-Sectional Studies, Checklist for Case-Control Studies, and Checklist for Cohort Studies" [24]. Located within the Faculty of Health Sciences of Adelaide University, the JBI is an international research and development organization world-renowned for evidence-based healthcare [24].
Given that systematic reviews represent a summary of core evidence, the JBI developed processes for the critical evaluation and synthesis of evidence to aid decision-making in health. The eight domains of risk of bias included in the JBI Critical Appraisal Checklist for Analytical Cross-Sectional Studies are: (i) clearly defined criteria for inclusion; (ii) description of study subjects and setting; (iii) valid and reliable measurement of exposure; (iv) use of objective, standard criteria; (v) identification of confounding factors; (vi) strategies for dealing with confounding factors; (vii) measurement of outcomes; and (viii) appropriate statistical analysis [24].
The JBI Critical Appraisal Checklist for Case Control Studies has 10 domains: (i) comparability of the groups; (ii) appropriateness of matching of cases and controls; (iii) criteria used for identification of cases and controls; (iv) valid and reliable measurement of exposure; (v) identification of confounding factors; (vi) identification of confounding factors; (vii) strategies for dealing with confounding factors; (viii) measurement of outcomes; (ix) length of exposure period of interest; and (x) appropriate statistical analysis [24].
The JBI Critical Appraisal Checklist for Cohort Studies has 11 domains: (i) similarity between groups and recruitment in the same population; (ii) similarity in measurement of exposures in assigning participants to the exposed and unexposed groups; (iii) valid and reliable measurement of exposure; (iv) identification of confounding factors; (v) strategies for dealing with confounding factors; (vi) outcome-free groups or participants at study baseline or exposure; (vii) measurement of outcomes; (viii) appropriate follow-up time; (ix) description of conclusion of follow-up or reason for loss to follow-up; (x) use of strategies to address incomplete follow-up; and (xi) appropriate statistical analysis [24].
For each study included the questions from the respective JBI Checklist were answered with a "yes", "no", "unclear", or "not applicable". The general risk of bias for each study was determined according to the following cut-off points [24]: • ≥70% "yes" answers: low risk of bias; • 50-69% "yes" answers: moderate risk of bias; • <50% "yes" answers: high risk of bias.
Three of the study's reviewers (TP, DS, CN) independently rated the risk of bias for each study included in the review, and a fourth reviewer (LT) settled any disagreements or discrepancies.

Analysis and Presentation of Results
The results obtained in the systematic review were first presented as a map of evidence, in which all articles reviewed were included according to study design, number of participants, characteristics of population, study site, and outcome for later analysis and discussion of findings.
A meta-analysis of the prevalence of risk factors for the MetS was then conducted. The clinical heterogeneity of studies that had the same outcome was assessed independently by two authors, with a third author settling any disagreements. Since studies assessed many different, yet related, effects, they were pooled using the inverse variance technique employing random-effect models. The software Stata version 15.0 was used for statistical analysis. The I 2 statistic was used to analyze the statistical heterogeneity of the studies. Pooled effect estimates were presented on the basis of the meta-analysis, even when the statistical heterogeneity was high (I 2 > 70%), because high levels of statistical heterogeneity were expected. A subgroup and sensitivity analysis was performed comparing immigrant Latinos with US-born Latinos-including sample size, extreme estimates, and study design for the outcomes arterial hypertension or high blood pressure, type-2 diabetes mellitus or high blood glucose, and general obesity-to check their influence on the pooled estimate. We could not conduct sensitivity analysis for the other outcomes due to the limited number of articles included in the meta-analysis.
The quality of the evidence was analyzed using the assessment criteria of the Grading of Recommendations Assessment, Development and Evaluation (GRADE) system, the most widely used instrument for assessing the quality of evidence of studies included in systematic reviews [25]. GRADE differs from other evaluation instruments by separating quality of evidence and strength of recommendation, rating quality of evidence for each outcome of interest, and allowing the level of quality of evidence of observational studies to be rated up if three criteria are met. The latter are large effect size, presence of dose-response gradient, and residual confounding factors, which increase confidence in estimates [25]. Forest plots were used in the meta-analyses for outcomes rated as very low quality of evidence, although combined estimates are not shown. The results of the synthesis of the evidence are shown in the summary table of findings.

Study Selection
Of the 2497 unique studies retrieved after applying the search strategy, 60 studies met the criteria for inclusion in the systematic review. Of this total, 52 studies were subsequently included in the meta-analysis, according to the model established by the PRISMA statement [21] (Figure 1). The eight studies excluded from the meta-analysis lacked the parameters required to estimate prevalence.
Participant age ranged from 0 to 91 years. Only one study included children and adolescents in its sample [56], age group 0 to ≥75 years. However, the analysis was stratified by age group, allowing data for subjects aged >18 years to be extracted.
The exposures assessed in the studies were grouped under four categories related to immigration; immigrant; health, diet and lifestyle; and community. The distribution of exposures for the different categories is presented in Table 1.

Risk of Bias
For the 55 cross-sectional studies included in the systematic review, the overall scoring of risk of bias ranged from 25.5 to 100% for the eight domains of the JBI Critical Appraisal Checklist for Analytical Cross-Sectional Studies (Table 3). Most of the studies (50) had low risk of bias, three had moderate risk, and two had high risk of bias (Table 3).
Were the study subjects and the setting described in detail?
3. Was the exposure measured in a valid and reliable way?
Were the outcomes measured in a valid and reliable way?
Were the study subjects and the setting described in detail?
3. Was the exposure measured in a valid and reliable way?
Were strategies to deal with confounding factors stated?
Were the outcomes measured in a valid and reliable way?
Were the study subjects and the setting described in detail?
3. Was the exposure measured in a valid and reliable way?
Were objective, standard criteria used for measurement of the condition?
Were the outcomes measured in a valid and reliable way?
Were the study subjects and the setting described in detail?
3. Was the exposure measured in a valid and reliable way?
The prevalence of arterial hypertension or high blood pressure in the studies reviewed in Latinos was 28% (95% CI: 23%-33%, I 2 : 99.6%). Of the studies included in the meta-analysis, 23 were heterogeneous (Figure 2). Within this group, some studies [16,39,50,53,68,70,78,82] differed more with respect to population size, measurement approach (objective or subjective), design (one case-control study) and characteristics of the population (ethnicity, age, and country of origin).
Additionally, a meta-analysis of the prevalence of arterial hypertension or high blood pressure for non-US-born Latinos (immigrants) and US-born Latinos was carried out (Figure 3). We found slightly higher prevalence for US-born Latinos (32% (95% CI: 19-45%, I 2 : 99.7%)), than for immigrant Latinos (28% (95% CI: 22-33%, I 2 : 99.3%)). The seven studies of immigrant Latinos and four of US-born Latinos included in the meta-analysis were heterogeneous. Greater differences were identified in two studies of immigrant Latinos [39,54] and two of US-born immigrants [14,54], possibly explained by population size, study design, or outcome measurement method.
A total of 24 studies were included in the meta-analysis of the prevalence of type 2 diabetes mellitus or high blood glucose (Figure 4), although those were heterogeneous. The prevalence of those conditions in Latinos was 17% (95% CI: 14-20%, I 2 : 99.3%). Notably, a number of studies [16,37,39,43,52,65,70,78,80,82] exhibited greater heterogeneity compared with the others regarding population size, outcome measurement method, study design, and country of origin of immigrant Latinos.
The prevalence of arterial hypertension or high blood pressure in the studies reviewed in Latinos was 28% (95% CI: 23%-33%, I 2 : 99.6%). Of the studies included in the meta-analysis, 23 were heterogeneous (Figure 2). Within this group, some studies [16,39,50,53,68,70,78,82] differed more with respect to population size, measurement approach (objective or subjective), design (one case-control study) and characteristics of the population (ethnicity, age, and country of origin). Additionally, a meta-analysis of the prevalence of arterial hypertension or high blood pressure for non-US-born Latinos (immigrants) and US-born Latinos was carried out (Figure 3). We found slightly higher prevalence for US-born Latinos (32% (95% CI: 19-45%, I 2 : 99.7%)), than for immigrant Latinos (28% (95% CI: 22-33%, I 2 : 99.3%)). The seven studies of immigrant Latinos and four of US-born Latinos included in the metaanalysis were heterogeneous. Greater differences were identified in two studies of immigrant Latinos [39,54] and two of US-born immigrants [14,54], possibly explained by population size, study design, or outcome measurement method.
In our analyses of general obesity rates in immigrant Latinos and US-born Latinos, 13 studies were included in the meta-analysis for immigrants and 10 for US-born Latinos ( Figure 8). We observed a higher prevalence of general obesity in immigrants (23%, 95% CI: 19-26%, I 2 : 99.7%) compared with US-born Latinos (15%, 95% CI: 13-16%, I 2 : 99%). All articles included in the meta-analysis involving immigrants, as well as those of US-born immigrants, were heterogeneous. However, five studies of immigrant Latinos [30,37,43,51,72] and one of US-born Latinos [72] exhibited greater heterogeneity for population size, age, and country of origin.

US-born Latinos
A total of 24 studies were included in the meta-analysis of the prevalence of type 2 diabetes mellitus or high blood glucose (Figure 4), although those were heterogeneous. The prevalence of those conditions in Latinos was 17% (95% CI: 14-20%, I 2 : 99.3%). Notably, a number of studies [16,37,39,43,52,65,70,78,80,82] exhibited greater heterogeneity compared with the others regarding population size, outcome measurement method, study design, and country of origin of immigrant Latinos. In the further analyses, a meta-analysis of the prevalence of type 2 diabetes mellitus or high blood glucose was conducted on seven studies for immigrant Latinos and five for US-born Latinos ( Figure 5). However, population size and outcome measurement methods differed in three of the studies for immigrant Latinos [39,43,54] and three for US-born Latinos [38,43,54]. A higher prevalence of the conditions was found for US-born Latinos (25% (95% CI: 16-33%, I 2 : 99.6%)) compared with immigrant Latinos (19% (95% CI: 14-24%, I 2 : 99.3%)).
Despite the greater heterogeneity of two studies [40,65], attributed to population size and country of origin of immigrant Latinos, they were included, together with three other studies in the meta-analysis of the prevalence rate of MetS in Latinos (Figure 11), which results showed a prevalence of 39% (95% CI: 32-45%, I 2 : 93.2%).
Of the six studies assessing the MetS, five used objective measures for obtaining data [40,60,64,65,77], while one collected data using self-reports [16].
Despite the greater heterogeneity of two studies [40,65], attributed to population size and country of origin of immigrant Latinos, they were included, together with three other studies in the meta-analysis of the prevalence rate of MetS in Latinos (Figure 11), which results showed a prevalence of 39% (95% CI: 32-45%, I 2 : 93.2%).

Further Sensitivity Analyses
Sensitivity analyses considering sample size, extreme estimates, and study design did not show substantial changes in the results of the meta-analyses for almost all analyses of the outcomes hypertension or high blood pressure, type 2 diabetes mellitus or high blood glucose, and overall obesity. However, a statistically significant difference was observed between groups (data not shown) when considering study design (longitudinal and cross-sectional) for type 2 diabetes mellitus or high blood glucose (p = 0.02, prevalence of 34% for longitudinal and 15% for cross-sectional) and sample size (≤1000 participants and >1000 participants) for general obesity (p = 0.007, prevalence of 45% for ≤1000 and 32% for >1000). The difference observed for the study design may have occurred because longitudinal studies have greater methodological rigor and control for potential confounders than cross-sectional ones.
Further sensitivity analyses including sex, age, country of birth, length of residence in the US, migration status, occupation, and health insurance, which could explain other differences, could not be conducted because insufficient studies had the necessary data for the estimates.

Quality of Evidence
After applying the GRADE system criteria (Table 7), the categories risk of bias (heterogeneity) and imprecision were downgraded by one point for the outcomes arterial hypertension, type 2 diabetes mellitus, general obesity, and abdominal obesity. Publication bias was not downgraded, not even for indirect evidence, since the surrogate outcomes observed (high blood pressure and glucose, and BMI and waist measures) were strongly associated with the outcomes of interest.

Further Sensitivity Analyses
Sensitivity analyses considering sample size, extreme estimates, and study design did not show substantial changes in the results of the meta-analyses for almost all analyses of the outcomes hypertension or high blood pressure, type 2 diabetes mellitus or high blood glucose, and overall obesity. However, a statistically significant difference was observed between groups (data not shown) when considering study design (longitudinal and crosssectional) for type 2 diabetes mellitus or high blood glucose (p = 0.02, prevalence of 34% for longitudinal and 15% for cross-sectional) and sample size (≤1000 participants and >1000 participants) for general obesity (p = 0.007, prevalence of 45% for ≤1000 and 32% for >1000). The difference observed for the study design may have occurred because longitudinal studies have greater methodological rigor and control for potential confounders than cross-sectional ones.
Further sensitivity analyses including sex, age, country of birth, length of residence in the USA, migration status, occupation, and health insurance, which could explain other differences, could not be conducted because insufficient studies had the necessary data for the estimates.

Quality of Evidence
After applying the GRADE system criteria (Table 7), the categories risk of bias (heterogeneity) and imprecision were downgraded by one point for the outcomes arterial hypertension, type 2 diabetes mellitus, general obesity, and abdominal obesity. Publication bias was not downgraded, not even for indirect evidence, since the surrogate outcomes observed (high blood pressure and glucose, and BMI and waist measures) were strongly associated with the outcomes of interest.
The quality of evidence for the assessment of high triglycerides and low HDL cholesterol was downgraded by one point for risk of bias, inconsistency (heterogeneity), and publication bias, and by two points for imprecision, because we found only a few studies assessing those outcomes. Most studies showed major differences in sample size and broad confidence intervals in the meta-analysis. No downgrades for indirect evidence were made since no surrogate outcomes were identified. The available evidence is sufficient to determine the prevalence, but confidence in the estimate is limited. As more information becomes available, the observed prevalence could change, and this change may be large enough to alter the conclusion. The available evidence is sufficient to determine the prevalence, but confidence in the estimate is limited. As more information becomes available, the observed prevalence could change, and this change may be large enough to alter the conclusion.

Obesity (BMI > 30 kg/m 2 )
37 (33)(34)(35)(36)(37)(38)(39)(40) 237.035 ⊕⊕ a,b,c,g,h low The available evidence is sufficient to determine the prevalence, but confidence in the estimate is limited. As more information becomes available, the observed prevalence could change, and this change may be large enough to alter the conclusion.

Abdominal Obesity
The available evidence is sufficient to determine the prevalence, but confidence in the estimate is limited. As more information becomes available, the observed prevalence could change, and this change may be large enough to alter the conclusion. The available evidence is insufficient to determine a reliable prevalence, and confidence in the estimate is limited. More information may allow for a more accurate estimation.

CI: confidence interval
GRADE quality of evidence ratings High quality: We are very confident that the effect in the study reflects the actual effect. Moderate quality: We are quite confident that the effect in the study is close to the true effect, but it is also possible that it is substantially different.
Low quality: The true effect may differ significantly from the estimate. Very low quality: The true effect is likely to be substantially different from the estimated effect.
x Because we are very uncertain regarding the effect estimate, we do not present it in Table 7. a Downgraded by one level (−1) for serious concerns with risk of bias. b Downgraded by one level (−1) for serious concerns with inconsistency. c Downgraded by one level (−1) for serious concerns with imprecision. d Downgraded by two levels (−2) for very serious concerns with imprecision. e Downgraded by two levels (−2) for very serious concerns with inconsistency. f Downgraded by one level (−1) for serious concerns with publication bias. g The observed surrogate endpoints are strongly associated with the outcome, so we did not reduce the evidence. h Most included studies were cross-sectional, which may lead to further reduction in the quality of the evidence.
For MetS, the risk of bias and publication bias categories were each downgraded by one level, whereas the inconsistency (heterogeneity) and imprecision categories were each downgraded by two levels. The quality of evidence was downgraded for publication bias because few studies were found assessing the MetS in immigrant Latinos, which may have led to overestimation of the measure.
The studies included in the review were expected to be heterogeneous, chiefly owing to cultural disparities, such as dietary habits and different lifestyles, among immigrant populations. Those differences can impact the outcomes observed, confirmed by I 2 results exceeding 90%. Other factors, such as socioeconomic differences, migration status, education, acculturation, and length of residence in the USA, can also affect heterogeneity. Those factors, however, could not be explored further in our review because few studies reported those characteristics. The downgrading of quality by one level due to risk of bias for all outcomes observed was based on the individual assessment of risk of bias by the JBI.
In conclusion, the quality of evidence for the outcomes of interest of this systematic review and meta-analysis was defined as low for systemic arterial hypertension, type 2 diabetes mellitus, general obesity, and abdominal obesity and as very low for high triglycerides, low HDL cholesterol, and MetS. It is likely that further research may impact this conclusion and change it.

Summary of Evidence
A total of 60 sixty studies were included in this systematic review, of which 52 were included in the meta-analysis. The overall population of the 52 studies was 436,654 immigrant Latinos in the US. The pooled prevalence obtained for arterial hypertension, type 2 diabetes mellitus, general obesity, and abdominal obesity were 28% (95% CI: 23-33%), 17% (95% CI: 14-20%), 37% (95% CI: 33-40%) and 54% (95% CI: 48-59%), respectively ( Table 7). The prevalence for the other outcomes are not shown because of uncertainty regarding those outcomes in immigrant Latinos and the very low quality of the evidence.
Higher prevalence for the outcomes arterial hypertension or high blood pressure and type 2 diabetes mellitus or high blood glucose were found in US-born Latinos compared with immigrant Latinos. By contrast, the prevalence for general obesity was higher in immigrants. Data from the USA Centers for Disease Control and Prevention (CDC) [12] show an alarming rise in obesity, hypertension, and diabetes in the Latino population. The evidence found in this review demonstrates an increased cardiovascular risk in Latino populations in the USA and supports the notion of increased weight and chronic conditions in this group as a result of the accelerated process of nutrition transition occurring in many countries of Latin America and other developing countries. Secondly, the evidence suggests that exposure to a highly obesogenic environment such as the USA should be taken into account in studies on obesity and associated factors [37,51]. Thirdly, the evidence supports the potential association of socioeconomic status, variations in physical activity behavior, and significant cultural variations among Latin American countries with the development of chronic diseases.
Comparison of the prevalence found in our meta-analysis with those of the CDC for the Latino population in the six states with the largest concentration of Latino and Brazilians (Florida, Massachusetts, New Jersey, New York, California, and Connecticut) [12], based on 2019 American Community Survey (ACS) data [83], revealed higher prevalence in our meta-analysis for hypertension (CDC prevalence range: 22.4-27.4%), diabetes (CDC range: 8.3-13.9%), and high cholesterol (CDC range: 27.3-31.8% based on meta-analysis rates for triglycerides and HDL cholesterol). The meta-analysis prevalence rate for obesity was similar to the 2020 CDC prevalence [12] for general obesity (range: 27.1-41.4%) but lower only for the rate observed in California.

Comparison with Previous Systematic Reviews and Meta-Analyses
To the best of our knowledge, there have been no previous systematic reviews or meta-analyses assessing factors associated with MetS in immigrant Latinos in the USA, precluding any meaningful comparisons. However, two systematic reviews assessing obesity alone [17,18] in immigrant populations were conducted. Those reviews suggested a potential relationship between increased obesity and higher level of acculturation, sex, nativity, length of residence in the USA, and generational status of immigrants. Nevertheless, the prevalence of obesity in this population was not addressed in those reviews, preventing any comparisons with obesity data we found in our review and meta-analysis.
The review by Delavari et al. [17] explored the relationship between acculturation and overweight/obesity in adult immigrants from low-and middle-income countries. Overall, of the nine studies reviewed, seven identified a positive association between level of acculturation and bodyweight variables. Among the studies reviewed by those authors involving immigrant Latino populations (Mexican-Americans) in the USA, significant BMI differences were found according to acculturation levels, sex, nativity, length of residence in the country, and generational status of immigrants. That review [17] showed that higher levels of acculturation were associated with greater BMI in immigrant Latinos, and also identified an association of sex, immigrant generational status, and nativity with increased obesity.
In the analyses stratified by sex, women immigrants residing in the USA for over 15 years had a mean BMI that was 2.38 kg/m 2 greater than the average found for women with less than 5 years of residence. A similar result was found for men living in the USA for over 15 years compared with those residing in the country for less than 5 years (1.10 kg/m 2 higher mean BMI in former group). To assess the risk of developing obesity associated with increased level of acculturation, the same review found a 4% and 3% higher risk of obesity in immigrant men and women with a high level of acculturation, respectively. With regard to the generational status of immigrants, Delavari et al. [17] found that second and third generations of Mexican-Americans had higher BMIs than their first generation counterparts.
The review by Oza-Frank and Cunningham [18] investigated the relationship between length of residence of immigrants in the USA and BMI. For the majority of the studies reviewed, the authors identified a positive association between time of residence in the USA and BMI. Some of the study results showed that a length of residence of up to 10 years promoted no major change in the BMI of immigrants, whereas others identified peak BMI growth after 21 years of residence for men and 15 years for women. The same review [18] included four studies involving immigrant Latinos, all of which reported a significant association between BMI and length of residence in the US.

Limitations
Our systematic review and meta-analysis have some limitations. First, there is high statistical heterogeneity for all outcomes. Immigrant populations differ in many ways, e.g., country of origin, migration status, degree of acculturation, and educational and socioeconomic levels, and those differences might explain the high inconsistencies found. Thus, the pooled estimates in this study should be regarded as suggestive as opposed to conclusive.
Second, most studies reviewed specifying the origin of the immigrant Latino population included only immigrants from Mexico or Central America. As a result, much of the evidence found in our review is limited to a subpopulation of immigrants from Latin America. There is scant scientific evidence on components of the MetS in immigrants from South America. Thus, the evidence found in this review and meta-analysis may not be representative of those countries which exhibit major sociocultural and economic differences that can impact the development of chronic diseases. Third, it was not possible to stratify analyses for other key characteristics, such as sex, age, country of birth, length of residence in the USA, migration status, occupation, and health insurance, because not all studies reported the detailed information needed to calculate such estimates.
Fourth, some studies reported results for more than one outcome of interest. However, it was not possible to extract information for all outcomes in some cases, as not all studies provided the data needed for that.
Fifth, some studies used measures derived from self-reports, particularly for general obesity, which are known to be subject to information bias, typically leading to an underestimation of own weight by women and overestimation by men [84].
Lastly, the studies were pooled in the meta-analyses, irrespective of risk of individual bias associated with them in order to prevent selection bias in meta-analyses due to stratification by quality of study [85].

Conclusions
This systematic review and meta-analysis were conducted in collaboration with researchers from the USA and Brazil to compile the available evidence on MetS and its risk factors in the immigrant Latino population. A large body of evidence was identified encompassing numerous studies, particularly pertaining to hypertension, type 2 diabetes, and both general and abdominal obesity. However, the evidence was classified as low or very low depending on the outcome. This rating suggests that future reviews and meta-analyses may reach different conclusions when more studies assessing MetS and its components become available, particularly HDL cholesterol and triglycerides, factors which featured less in the studies reviewed.
The production of prevalence estimates of the MetS and its risk factors in the immigrant population appears to be justified and evidence-based, considering the cultural, dietary, and lifestyle differences encountered by migrants, which promote increased obesity (strongly associated with MetS) and chronic diseases associated with the condition.
However, few studies were available that included, or analyzed separately, information on immigrant Latinos from South America, including Brazil, a region for which only two studies were found. Therefore, further studies are needed addressing MetS and its factors in immigrants from South America, particularly Brazil, in view of the particularities regarding the culture, language, and diet compared with other Latin American nations.