A description of metabolic syndrome
In 1998, Reaven(Reference Reaven1) defined metabolic syndrome (MS), also known as X syndrome, as a conglomerate of coronary risk factors. Since that time, the definition of MS has been modified by a number of experts(Reference Dandona, Aljada and Chaudhuri2–Reference Alberti, Zimmet and Shaw4). Although the different definitions of MS generally include the same components, they differ in the assignment of each component's importance for the diagnosis of MS (for instance, number of criteria, hierarchy of components).
The definition provided by the Adult Treatment Panel III (ATP III) seems to be more ‘equitable’(Reference Reaven5), given that it does not establish a hierarchal relationship among the five components: abdominal obesity (AO), glucose alterations, hypertension (HT), hypertriacylglycerolaemia and low HDL cholesterol levels. In addition, ATP III is one of the most internationally accepted and used definitions, as well as the most up-to-date as a result of its high use and analysis(Reference Grundy, Cleeman and Daniels3, 6, Reference Grundy, Brewer and Cleeman7). Nevertheless, it is worth mentioning that, in the latter part of 2009, prominent institutions such as the International Diabetes Federation (IDF) and the American Heart Association/National Heart, Lung and Blood Institute standardized their criteria for defining MS(Reference Alberti, Eckel and Grundy8). These updated diagnostic criteria maintained the components defined by the last ATP III modification made in 2005(Reference Grundy, Cleeman and Daniels3), and also contributed with new guidelines for the measurement of waist circumference by recommending specific cut-off points for each ethnic group or population (Table 1).
Table 1 ATP III criteria for clinical diagnosis of metabolic syndrome
ATP III, Adult Treatment Panel III.
Clinical importance of diagnosing metabolic syndrome
In 2007, Grundy(Reference Grundy9) described the reasons why MS analysis was useful from a clinical perspective. To summarize, it mentions that MS detection is important because it allows for easy identification of patients who are at risk of developing atherosclerosis, type 2 diabetes mellitus and/or associated comorbidities and who are subjected to a higher risk of mortality from these causes(Reference Isomaa, Almaren and Tuomi10–Reference Wannamethee, Shaper and Lennon17). In addition, the concept of MS facilitates the understanding of the physiopathological relationships between its different components and other related cardiovascular risk factors. This understanding may also be helpful for carrying out epidemiological and clinical studies associated with the treatment and/or prevention of CVD, which may entail drugs, lifestyle changes(Reference Grundy9, Reference Huang11) or both simultaneously.
Prevalence of metabolic syndrome in the world
The prevalence of MS is high in Europe and USA. In a compilation of observational studies conducted on people from a number of European countries (Switzerland, Spain, the Netherlands, Italy, France, the UK and Denmark), it was reported that non-diabetic subjects under 40 years of age had an MS prevalence of 14–41 %, depending on the age range(Reference Balkau, Charles and Drivsholm18). In Catalonia (Spain), Buckland et al. (Reference Buckland, Salas-Salvadó and Roure19) determined the prevalence of MS in people aged 18–74 years as 24·8 %, ranging from 2·5 % (in 18–24-year-olds) to 51·1 % (in 65–74-year-olds). In the USA, an MS prevalence of 21·8 % was found, which ranged from 6·7 % (in 20–29-year-olds) to 43·5 % (in 60–69-year-olds)(Reference Ford, Mokdad and Giles20).
Results from different cross-sectional studies conducted in countries with emerging economies have reported that the prevalence of MS in these countries is also high. Hence, in studies carried out on adult populations of India and Iran, the prevalence rates of MS varied from 33·7 % to 41 %(Reference Azizi, Salehi and Etemadi21–Reference Ramanchandran, Snehalatha and Satyavani23).
Although there are several studies reporting the prevalence of MS in Latin America, in the majority of them MS is a secondary objective of the study and thus they do not reflect the real prevalence of this syndrome among the general population. This may be because this region has experienced a nutritional transition, having gone from being an area in which malnutrition was a main concern to experiencing problems associated with overeating (obesity, MS and CVD mortality). This situation has probably delayed the development of studies focusing exclusively on MS and its components, as well as the analysis of other precursors to the development of these diseases.
For these reasons, the objective of the present systematic review is to assess the prevalence of MS in Latin America using data already published in the scientific literature.
Materials and methods
General features of searches
Two electronic bibliographic indexes, one of which was discipline specific (PubMed) and another that was multidisciplinary (ISIWeb), were searched, along with two multidisciplinary databases for Ibero-America (SCielo and Redalyc). The terms used to describe MS were taken from the PubMed (MeSH) dictionary. The keywords used were ‘metabolic syndrome x’ (which also included ‘insulin resistance syndrome x’, ‘dysmetabolic syndrome x’, ‘Reaven syndrome x’, ‘metabolic cardiovascular syndrome’) and ‘prevalence’, with their respective translations into Spanish. These terms were searched in titles and/or abstracts. Searches were limited to articles published between January 2000 and March 2010.
Selection of articles
The criteria used and the selection process for the articles are shown in Table 2. A total of 5216 original cross-sectional studies containing data on the prevalence of MS were automatically identified by applying the afore-mentioned search terms. Among the latter, articles were chosen through a sequential application of our five criteria (see Table 2). Finally, a total of twelve articles met the criteria for the present review(Reference Perez, Guzman and Ortiz24–Reference Escobedo, Schargrodsky and Champagne35). Of these articles (eleven cross-sectional(Reference Perez, Guzman and Ortiz24–Reference Salaroli, Barbosa and Mill34) and one multi-centre cross-sectional(Reference Escobedo, Schargrodsky and Champagne35); Tables 3 and 4, respectively), the following data were found: author and year of publication, number and gender of participants, age range of the population, sampling, city and country in which the study was carried out, procedures to determine MS and its components (measurements), the prevalence of MS and its components, as well as other sociodemographic and environmental factors analysed as associations by the authors. In addition, the following variables were analysed: study design, prevalence of MS and its components by gender and age group, as well as adjusted MS prevalence.
Table 2 Flow chart for selection of articles for the present review
MS, metabolic syndrome; ATP III, Adult Treatment Panel III.
*The references for the papers that, finally, were not included in this review (references 46–231) can be read on the online version in the Appendix: Supplementary references.
Table 3 Cross-sectional studies of the prevalence of MS among Latin-American populations included in the present review
MS, metabolic syndrome; W, women; M, men; WC, waist circumference; G, glucose; HDL-C, HDL cholesterol; BP, blood pressure; ATP III, Adult Treatment Panel III; N/A, information not available; NHANES II, Second National Health and Nutrition Examination Survey; AO, abdominal obesity; high G, hyperglycaemia; HT, hypertension; high TAG, hypertriacylglycerolaemia; HOMA-IR, homeostatic model assessment–insulin resistance.
*Only statistically significant associations are shown (P < 0·05).
†Prevalence of age-adjusted MS, only data reported in this article.
‡Reference group for age <30 years.
Table 4 Cross-sectional multi-centre study of the prevalence of MS among Latin-American populations included in the present review
MS, metabolic syndrome; W, women; M, men; WC, waist circumference; G, glucose; HDL-C, HDL cholesterol; BP, blood pressure; ATP III, Adult Treatment Panel III; AO, abdominal obesity; high G, hyperglycaemia; HT, hypertension; high TAG, hypertriacylglycerolaemia.
*Calculated prevalence of population with MS.
The (weighted) mean prevalence rates found in the present review were calculated as follows: sum of the number of cases in all studies considered/sum of the number of participants in all studies considered × 100.
Results
In the first part of this section, we summarize the results of the eleven cross-sectional studies selected for the present review (Table 3). In the last part of this section, the data reported by the CARMELA(Reference Escobedo, Schargrodsky and Champagne35) study are described, because it was the only cross-sectional multi-centre study that did not provide data on the prevalence of the different components of MS within the entire study population and reported only figures from the population that had MS (Table 4).
Prevalence of metabolic syndrome
The eleven studies included in this section encompassed 5635 (40·8 %) men and 8173 (59·2 %) women. In ten studies, the general prevalence of MS(Reference Perez, Guzman and Ortiz24–Reference Florez, Silva and Fernandez30, Reference Marquezine, Oliveira and Pereira32–Reference Salaroli, Barbosa and Mill34) was reported, and in four of these the prevalence adjusted for age was presented(Reference Perez, Guzman and Ortiz24, Reference Medina-Lezama, Zea-Diaz and Morey-Vargas27, Reference Florez, Silva and Fernandez30, Reference Aguilar-Salinas, Rojas and Gomez-Perez31, Reference Velasquez-Melendez, Gazzinelli and Correa-Oliveira33), whereas one presented prevalence adjusted for age and gender(Reference Mujica, Leiva and Icaza25). One study did not present general prevalence and only described prevalence adjusted for age(Reference Aguilar-Salinas, Rojas and Gomez-Perez31). In the studies that mentioned rates of prevalence, the weighted mean for MS general prevalence was 24·9 %(Reference Perez, Guzman and Ortiz24–Reference Florez, Silva and Fernandez30, Reference Marquezine, Oliveira and Pereira32–Reference Salaroli, Barbosa and Mill34), the weighted mean for age-adjusted prevalence was 28·03 %(Reference Perez, Guzman and Ortiz24, Reference Medina-Lezama, Zea-Diaz and Morey-Vargas27, Reference Florez, Silva and Fernandez30, Reference Aguilar-Salinas, Rojas and Gomez-Perez31, Reference Velasquez-Melendez, Gazzinelli and Correa-Oliveira33) and the weighted mean for age- and gender-adjusted prevalence was 29·5 %(Reference Mujica, Leiva and Icaza25).
MS was most prevalent in the city of San Juan (43·3 %)(Reference Perez, Guzman and Ortiz24), in a study carried out on the general population. The city that reported the lowest prevalence of MS was Arequipa, among a mestizo population (18·8 %)(Reference Medina-Lezama, Zea-Diaz and Morey-Vargas27).
Nine articles reported the prevalence of MS by gender(Reference Perez, Guzman and Ortiz24, Reference Mujica, Leiva and Icaza25, Reference Medina-Lezama, Zea-Diaz and Morey-Vargas27–Reference Aguilar-Salinas, Rojas and Gomez-Perez31, Reference Velasquez-Melendez, Gazzinelli and Correa-Oliveira33, Reference Salaroli, Barbosa and Mill34): weighted mean was 25·3 % for women and 23·2 % for men; differences were not statistically significant. In three studies(Reference Medina-Lezama, Zea-Diaz and Morey-Vargas27, Reference Tull, Thurland and LaPorte28, Reference Salaroli, Barbosa and Mill34) (two of which had higher participation by women(Reference Medina-Lezama, Zea-Diaz and Morey-Vargas27, Reference Salaroli, Barbosa and Mill34), whereas one had higher participation by men(Reference Tull, Thurland and LaPorte28)), MS was slightly more frequent among women compared with men, with a difference in prevalence between men and women varying between 0·8 %(Reference Salaroli, Barbosa and Mill34) and 8·89 %(Reference Medina-Lezama, Zea-Diaz and Morey-Vargas27). Furthermore, in the study conducted on the population of Minas Gerais(Reference Velasquez-Melendez, Gazzinelli and Correa-Oliveira33), which contained a higher proportion of women, this difference in prevalence was 25·9 % (33·6 % for women and 7·7 % for men). However, in four other studies(Reference Perez, Guzman and Ortiz24, Reference Mujica, Leiva and Icaza25, Reference Florez, Silva and Fernandez30, Reference Aguilar-Salinas, Rojas and Gomez-Perez31), the prevalence of MS was slightly higher in men than in women: the differences varied between 1·1 %(Reference Mujica, Leiva and Icaza25) and 5·2 %(Reference Florez, Silva and Fernandez30). Likewise, in the study carried out on the population of San Juan(Reference Gomez, Ramirez and Disdier29), a higher prevalence was reported for men (41 %) than for women (26 %).
Nine studies reported prevalence by age(Reference Perez, Guzman and Ortiz24, Reference Medina-Lezama, Zea-Diaz and Morey-Vargas27–Reference Salaroli, Barbosa and Mill34). In six of these(Reference Perez, Guzman and Ortiz24, Reference Medina-Lezama, Zea-Diaz and Morey-Vargas27, Reference Tull, Thurland and LaPorte28, Reference Marquezine, Oliveira and Pereira32–Reference Salaroli, Barbosa and Mill34), the prevalence of MS was observed to increase with age, whereas in the other three studies this was not the case(Reference Gomez, Ramirez and Disdier29–Reference Aguilar-Salinas, Rojas and Gomez-Perez31). The highest prevalence of MS was observed in subjects older than 50 years of age, with the exception of Mexico, in which the population with the highest prevalence (27·5 %) was that of 30–39-year-olds(Reference Aguilar-Salinas, Rojas and Gomez-Perez31). By contrast, the lowest prevalence was generally observed in those under the age of 30 years, with the exception of Mexico, where the lowest prevalence (8·7 %) was found to be among the 60–69-year-old population(Reference Aguilar-Salinas, Rojas and Gomez-Perez31).
Prevalence of the components of metabolic syndrome
The prevalence of the components of MS among the populations analysed varied greatly from one country to another (Table 3). The weighted mean prevalence (range) by component was as follows: low HDL cholesterol, 62·9 (21·5(Reference Mujica, Leiva and Icaza25)–96·1(Reference Pinzon, Serrano and Diaz26))%; hypertriacylglycerolaemia, 46·7 (22·3(Reference Velasquez-Melendez, Gazzinelli and Correa-Oliveira33)–87·3(Reference Medina-Lezama, Zea-Diaz and Morey-Vargas27))%; AO, 45·8 (26·7(Reference Velasquez-Melendez, Gazzinelli and Correa-Oliveira33)–53·6(Reference Pinzon, Serrano and Diaz26))%; HT, 42·6 (16·8(Reference Pinzon, Serrano and Diaz26)–77·2(Reference Gomez, Ramirez and Disdier29))%; and abnormal blood glucose, 17·6 (3·9(Reference Pinzon, Serrano and Diaz26)–49·8(Reference Perez, Guzman and Ortiz24))%. For abnormal blood glucose and dyslipidaemia, none of the studies stated how many subjects were diagnosed because they were being treated or because of concentrations in blood samples (as suggested by ATP III 2005 criteria(Reference Grundy, Cleeman and Daniels3)).
The prevalence of components by gender and age group were not reported in any of the articles. Only seven studies described the presence of MS components by gender(Reference Perez, Guzman and Ortiz24, Reference Pinzon, Serrano and Diaz26, Reference Medina-Lezama, Zea-Diaz and Morey-Vargas27, Reference Gomez, Ramirez and Disdier29, Reference Florez, Silva and Fernandez30, Reference Marquezine, Oliveira and Pereira32, Reference Velasquez-Melendez, Gazzinelli and Correa-Oliveira33). In addition, three studies reported the prevalence of each of the MS components by age group(Reference Medina-Lezama, Zea-Diaz and Morey-Vargas27, Reference Gomez, Ramirez and Disdier29, Reference Marquezine, Oliveira and Pereira32) and another reported these data by age, but only for AO(Reference Tull, Thurland and LaPorte28). In general, a higher prevalence of these components was found in populations older than 49 years, whereas the lower prevalence was found in those under 34 years of age, except for low HDL cholesterol levels.
Abdominal obesity
The prevalence of AO was reported in six studies(Reference Perez, Guzman and Ortiz24, Reference Pinzon, Serrano and Diaz26, Reference Medina-Lezama, Zea-Diaz and Morey-Vargas27, Reference Gomez, Ramirez and Disdier29, Reference Florez, Silva and Fernandez30, Reference Velasquez-Melendez, Gazzinelli and Correa-Oliveira33). The weighted mean prevalence was 45·8 %. AO was shown to be the most prevalent among employees (medical school) in Colombia (53·6 %)(Reference Pinzon, Serrano and Diaz26). The smallest prevalence of AO was found among the black and mestizo population (26·7 %) of Minas Gerais(Reference Velasquez-Melendez, Gazzinelli and Correa-Oliveira33).
In six studies, the prevalence of AO was reported by gender(Reference Perez, Guzman and Ortiz24, Reference Pinzon, Serrano and Diaz26, Reference Medina-Lezama, Zea-Diaz and Morey-Vargas27, Reference Florez, Silva and Fernandez30, Reference Marquezine, Oliveira and Pereira32, Reference Velasquez-Melendez, Gazzinelli and Correa-Oliveira33), with a mean prevalence of 41·6 % for women and 21·7 % for men. In five of them, the prevalence was higher for women than for men(Reference Perez, Guzman and Ortiz24, Reference Medina-Lezama, Zea-Diaz and Morey-Vargas27, Reference Florez, Silva and Fernandez30, Reference Marquezine, Oliveira and Pereira32, Reference Velasquez-Melendez, Gazzinelli and Correa-Oliveira33). Differences in prevalence for women and men varied between 14·4 %(Reference Marquezine, Oliveira and Pereira32) and 23·0 %(Reference Medina-Lezama, Zea-Diaz and Morey-Vargas27), with the exception of one study in which the difference was 42·0 % (46·3 % for women and 4·3 % for men)(Reference Santos, Lopes and Barros36). By contrast, in the study by Pinzón et al. (Reference Pinzon, Serrano and Diaz26), a higher prevalence was reported for men (75·0 %) than for women (28·2 %). In four studies the prevalence of AO was reported by age groups(Reference Medina-Lezama, Zea-Diaz and Morey-Vargas27–Reference Gomez, Ramirez and Disdier29, Reference Marquezine, Oliveira and Pereira32). In three of them, the prevalence of AO increased with age(Reference Medina-Lezama, Zea-Diaz and Morey-Vargas27, Reference Gomez, Ramirez and Disdier29, Reference Marquezine, Oliveira and Pereira32).
Abnormal glycaemia
Seven studies mentioned the prevalence of abnormal glycaemia (AG)(Reference Perez, Guzman and Ortiz24–Reference Medina-Lezama, Zea-Diaz and Morey-Vargas27, Reference Gomez, Ramirez and Disdier29, Reference Florez, Silva and Fernandez30, Reference Velasquez-Melendez, Gazzinelli and Correa-Oliveira33) among the population. In these seven studies, the weighted mean prevalence of AG was 17·6 %. AG had the higher prevalence in the city of San Juan, in the general population (49·8 %)(Reference Perez, Guzman and Ortiz24) and in university employees (31·2 %)(Reference Gomez, Ramirez and Disdier29). The lowest prevalence of AG was found among employees in Bucaramanga (3·9 %)(Reference Pinzon, Serrano and Diaz26).
In six studies, the prevalence of AG was described by gender(Reference Perez, Guzman and Ortiz24, Reference Pinzon, Serrano and Diaz26, Reference Medina-Lezama, Zea-Diaz and Morey-Vargas27, Reference Florez, Silva and Fernandez30, Reference Marquezine, Oliveira and Pereira32, Reference Velasquez-Melendez, Gazzinelli and Correa-Oliveira33) (weighted mean was 14·3 % for women and 17·6 % for men). In one of these, the prevalence was higher in women than in men, with a difference of 1·6 %(Reference Velasquez-Melendez, Gazzinelli and Correa-Oliveira33). In the other five studies(Reference Perez, Guzman and Ortiz24, Reference Pinzon, Serrano and Diaz26, Reference Medina-Lezama, Zea-Diaz and Morey-Vargas27, Reference Florez, Silva and Fernandez30, Reference Marquezine, Oliveira and Pereira32), the prevalence was higher in men than in women, with differences between 0·4 %(Reference Medina-Lezama, Zea-Diaz and Morey-Vargas27) and 16·7 %(Reference Perez, Guzman and Ortiz24). Only three studies reported the prevalence of AG by age(Reference Medina-Lezama, Zea-Diaz and Morey-Vargas27, Reference Gomez, Ramirez and Disdier29, Reference Marquezine, Oliveira and Pereira32), observing an increase with age.
Hypertension
The weighted mean prevalence of HT was found to be 42·6 % in the seven studies that reported on the general prevalence of this component(Reference Perez, Guzman and Ortiz24–Reference Medina-Lezama, Zea-Diaz and Morey-Vargas27, Reference Gomez, Ramirez and Disdier29, Reference Florez, Silva and Fernandez30, Reference Velasquez-Melendez, Gazzinelli and Correa-Oliveira33). HT was most prevalent in San Juan (77·2 %)(Reference Gomez, Ramirez and Disdier29) and the lowest prevalence was, again, in Bucaramanga (16·8 %)(Reference Pinzon, Serrano and Diaz26), both studies conducted among university employees.
Six studies reported the prevalence of HT by gender(Reference Perez, Guzman and Ortiz24, Reference Pinzon, Serrano and Diaz26, Reference Medina-Lezama, Zea-Diaz and Morey-Vargas27, Reference Florez, Silva and Fernandez30, Reference Marquezine, Oliveira and Pereira32, Reference Velasquez-Melendez, Gazzinelli and Correa-Oliveira33) (weighted mean was 40·2 % for women, 38·4 % for men). In one of them, the prevalence was higher for women (47·9 %) than for men (31·4 %)(Reference Florez, Silva and Fernandez30). In the remaining studies(Reference Perez, Guzman and Ortiz24, Reference Pinzon, Serrano and Diaz26, Reference Medina-Lezama, Zea-Diaz and Morey-Vargas27, Reference Florez, Silva and Fernandez30, Reference Marquezine, Oliveira and Pereira32, Reference Velasquez-Melendez, Gazzinelli and Correa-Oliveira33), the prevalence was higher for men than for women, with the reported differences ranging from 3·8 %(Reference Medina-Lezama, Zea-Diaz and Morey-Vargas27) to 18·0 %(Reference Pinzon, Serrano and Diaz26). Only three studies reported the prevalence of HT by age(Reference Medina-Lezama, Zea-Diaz and Morey-Vargas27, Reference Gomez, Ramirez and Disdier29, Reference Marquezine, Oliveira and Pereira32), observing an increase in parallel.
Dyslipidaemia
The same seven studies mentioned the general prevalence of dyslipidaemia(Reference Perez, Guzman and Ortiz24–Reference Medina-Lezama, Zea-Diaz and Morey-Vargas27, Reference Gomez, Ramirez and Disdier29, Reference Florez, Silva and Fernandez30, Reference Velasquez-Melendez, Gazzinelli and Correa-Oliveira33); the weighted mean of hypertriacylglycerolaemia was 46·7 %, and for low levels of HDL cholesterol the weighted mean was 62·9 %.
Hypertriacylglycerolaemia was most prevalent among the general population of the city of Arequipa (87·3 %)(Reference Medina-Lezama, Zea-Diaz and Morey-Vargas27). On the other hand, the lowest prevalence of hypertriacylglycerolaemia was found in Minas Gerais, among the general population (22·3 %)(Reference Velasquez-Melendez, Gazzinelli and Correa-Oliveira33). In five of the six studies that reported the prevalence of hypertriacylglycerolaemia by gender(Reference Perez, Guzman and Ortiz24, Reference Pinzon, Serrano and Diaz26, Reference Medina-Lezama, Zea-Diaz and Morey-Vargas27, Reference Florez, Silva and Fernandez30, Reference Marquezine, Oliveira and Pereira32), it was more prevalent in men than in women, with differences between 12·9 %(Reference Perez, Guzman and Ortiz24) and 16·7 %(Reference Medina-Lezama, Zea-Diaz and Morey-Vargas27), except in one study, in which the difference was 42·5 %(Reference Pinzon, Serrano and Diaz26). In the study by Velasquez-Melendez et al.(Reference Velasquez-Melendez, Gazzinelli and Correa-Oliveira33), the prevalence was slightly higher among women (23·9 %) than among men (20·5 %). The weighted mean prevalence of hypertriacylglycerolaemia was 28·1 % for women and 43·6 % for men.
Low levels of HDL cholesterol were most prevalent among university employees in Bucaramanga (96·1 %)(Reference Pinzon, Serrano and Diaz26), and least prevalent in Talca (general Chilean population, 21·5 %)(Reference Mujica, Leiva and Icaza25). In the six studies that reported on the prevalence of low levels of HDL cholesterol by gender(Reference Perez, Guzman and Ortiz24, Reference Pinzon, Serrano and Diaz26, Reference Medina-Lezama, Zea-Diaz and Morey-Vargas27, Reference Florez, Silva and Fernandez30, Reference Marquezine, Oliveira and Pereira32, Reference Velasquez-Melendez, Gazzinelli and Correa-Oliveira33), a higher prevalence was observed among women than among men (with differences between 2·0 %(Reference Pinzon, Serrano and Diaz26) and 28·4 %(Reference Medina-Lezama, Zea-Diaz and Morey-Vargas27)). The weighted mean prevalence by gender was 63 % for women and 45·8 % for men.
Three studies reported on the prevalence of dyslipidaemia by age group(Reference Medina-Lezama, Zea-Diaz and Morey-Vargas27, Reference Gomez, Ramirez and Disdier29, Reference Marquezine, Oliveira and Pereira32), observing an increase by age. With regard to low HDL cholesterol levels, the prevalence of low HDL cholesterol levels decreased when age increased(Reference Medina-Lezama, Zea-Diaz and Morey-Vargas27, Reference Gomez, Ramirez and Disdier29): it was least frequent among individuals older than 49 years of age and most frequent among those under the age of 34 years. One study did not point to a clear trend by age(Reference Marquezine, Oliveira and Pereira32).
Factors associated with metabolic syndrome
Six studies analysed the possible associations between different risk factors and the presence of MS(Reference Medina-Lezama, Zea-Diaz and Morey-Vargas27, Reference Tull, Thurland and LaPorte28, Reference Florez, Silva and Fernandez30, Reference Marquezine, Oliveira and Pereira32–Reference Salaroli, Barbosa and Mill34). The risk of MS increased significantly with age and with the presence of various conditions such as being Hispanic or indigenous, belonging to a low-income class, insufficient physical activity, excessive alcohol consumption, smoking and a history of HT and of type 2 diabetes in first-degree family members (Table 3). The strongest association was observed with age in the case of adults aged between 60 and 88 years compared with those under 30 (OR = 20·75, 95 % CI 4·98, 86·47)(Reference Velasquez-Melendez, Gazzinelli and Correa-Oliveira33). Insufficient physical activity (sedentary lifestyle) was the second most strongly associated risk factor (OR = 1·92, 95 % CI 1·41, 2·60)(Reference Medina-Lezama, Zea-Diaz and Morey-Vargas27), followed by being Hispanic (OR = 1·82, 95 % CI 1·07, 3·07)(Reference Tull, Thurland and LaPorte28).
No statistically significant protective factors were reported in the studies included in the present review.
The CARMELA study
The CARMELA study is the only multi-centre cross-sectional research aiming to assess cardiovascular risk factors in Latin-American populations(Reference Escobedo, Schargrodsky and Champagne35) (Table 4). The study included a total of 11 502 individuals, with the same ratio of men and women, who were between 25 and 64 years of age and from urban populations of seven cities. The study population was selected by stratified random sample, and measurements were taken using the same protocol in each city. The 2001 ATP III criteria for diagnosing MS were used.
The highest prevalence of MS (27 %) was found in Mexico City, whereas the lowest (14 %) was found in Quito. The average prevalence in the total population was slightly higher for women than for men (22 % v. 20 %, respectively), with the exception of two cities, Buenos Aires and Barquisimeto, where men presented a higher prevalence of MS than did women. The prevalence of MS increased with age.
With respect to the components of MS, the CARMELA study only reported their prevalence among the population that presented MS and not among the total population. Thus, the most prevalent component of these subjects with MS was hypertriacylglycerolaemia (85·9 %), and the least frequent was abnormal glucose levels (31·2 %; Table 4). With regard to gender, the AO and low HDL cholesterol MS components were more frequent in women than in men, with gender differences reported to be 26·8 % and 5·2 %, respectively. The remaining altered components were more prevalent in men than in women. The difference in prevalence between men and women was 14·2 % for HT, 7·6 % for hypertriacylglycerolaemia and 2·5 % for abnormal glucose levels. The prevalence of MS components by age group was not reported in the CARMELA study.
Discussion
The present review provides an overall picture of the presence of MS in Latin America. According to data reviewed, the weighted mean prevalence of MS in the Latin-American countries analysed (24·9 %) was higher than that reported in developed countries(Reference Buckland, Salas-Salvadó and Roure19, Reference Ford, Mokdad and Giles20). In fact, in eight of the studies included in this analysis(Reference Perez, Guzman and Ortiz24–Reference Pinzon, Serrano and Diaz26, Reference Gomez, Ramirez and Disdier29–Reference Marquezine, Oliveira and Pereira32, Reference Salaroli, Barbosa and Mill34), prevalence rates of MS are different in 1·5–20·0 % compared with those reported in the USA (21·8 %) and some European countries: Spain (24·7 %), Portugal (23·9 %) and Italy (24·1 %)(Reference Balkau, Charles and Drivsholm18, Reference Buckland, Salas-Salvadó and Roure19, Reference Ford, Mokdad and Giles20, Reference Santos, Lopes and Barros36, Reference Magi, Stramenga and Morosini37). However, this mean prevalence of MS in Latin America was lower than that reported for some ethnic groups in other parts of the world. MS prevalence among Mexican Americans has been reported at 31·9 %(Reference Ford, Mokdad and Giles20), whereas this figure has been reported at 35·5 % for African or Afro-American people, 32·8 % for Amerindians(Reference Florez, Silva and Fernandez30), 31 % for Yaquis Indians(Reference Hanson, Imperatore and Bennett38) and 44 % for Australian Aborigines(Reference Schutte, Shemesh and Rowley39).
In our results, the most frequent component of MS was low HDL cholesterol (62·9 %), which suggests the need for monitoring it during medical check-ups, because in some Latin-American countries it is common to check only total cholesterol. In other studies, the most prevalent component was AO: 38·6 % in the USA(Reference Ford, Mokdad and Giles20) and 60·3 % in Spain(Reference Buckland, Salas-Salvadó and Roure19). The high prevalence of low HDL cholesterol observed was higher than that reported for indigenous people in Mexico (34·2 %) and for Yaquis Indians (42·4 %)(Reference Rodríguez-Morán, Guerrero-Romero and Brito-Zurita40). In contrast, excess weight and obesity are more frequently reported in various ethnic groups: Yaquis Indians (86·7 %) and Pima Indians (90·4 %)(Reference Meaney, Lara-Esqueda and Ceballos-Reyes41, Reference Schulz, Bennet and Ravussin42).
Differences in MS prevalence as high as 24·5 (range: 18·8–43·3)% among Latin-American countries were observed in the studies analysed in the present review(Reference Perez, Guzman and Ortiz24, Reference Medina-Lezama, Zea-Diaz and Morey-Vargas27). The differences were even greater with respect to the prevalence of the components of MS. This might be explained, at least to some degree, by genetic, environmental and sociodemographic factors in the countries and/or their populations. However, the results of the present review do not make it possible to know the full scope of the problem because data from some countries are not available, and because the methodological differences among the studies published up to the present limit the overall analysis of their results.
One factor that makes comparability among studies difficult is the source and type of individuals studied. Thus, for instance, the population studied was mostly urban, and only in one article was a rural population studied(Reference Velasquez-Melendez, Gazzinelli and Correa-Oliveira33). Nine studies included samples of the general population(Reference Perez, Guzman and Ortiz24, Reference Mujica, Leiva and Icaza25, Reference Medina-Lezama, Zea-Diaz and Morey-Vargas27, Reference Tull, Thurland and LaPorte28, Reference Florez, Silva and Fernandez30, Reference Aguilar-Salinas, Rojas and Gomez-Perez31, Reference Velasquez-Melendez, Gazzinelli and Correa-Oliveira33–Reference Escobedo, Schargrodsky and Champagne35), and the rest included apparently healthy sample populations with particular features: university employees(Reference Pinzon, Serrano and Diaz26, Reference Gomez, Ramirez and Disdier29) or a majority of mulatto subjects(Reference Marquezine, Oliveira and Pereira32). In eight studies(Reference Perez, Guzman and Ortiz24, Reference Mujica, Leiva and Icaza25, Reference Medina-Lezama, Zea-Diaz and Morey-Vargas27, Reference Gomez, Ramirez and Disdier29–Reference Marquezine, Oliveira and Pereira32, Reference Salaroli, Barbosa and Mill34), the presence of women was higher than that of men; in three(Reference Pinzon, Serrano and Diaz26, Reference Tull, Thurland and LaPorte28, Reference Velasquez-Melendez, Gazzinelli and Correa-Oliveira33), the presence of men was higher; and in the multi-centre study(Reference Escobedo, Schargrodsky and Champagne35), the percentages of women and men who took part were the same. In nine studies, the sample included the same number of individuals from each age group studied. In the other three studies(Reference Perez, Guzman and Ortiz24, Reference Medina-Lezama, Zea-Diaz and Morey-Vargas27, Reference Aguilar-Salinas, Rojas and Gomez-Perez31), a different number of individuals from different age groups was included (populations with the highest number of participants were in the age ranges of 30–39(Reference Aguilar-Salinas, Rojas and Gomez-Perez31) and 50–59 years(Reference Perez, Guzman and Ortiz24, Reference Medina-Lezama, Zea-Diaz and Morey-Vargas27)).
Another limitation that makes comparability among studies difficult is the different methods used. The type of sampling was different among the studies selected (Tables 3 and 4). In order to diagnose MS, the majority (eight) of studies used the ATP III 2001 criteria(Reference Tull, Thurland and LaPorte28–Reference Escobedo, Schargrodsky and Champagne35); three were based on the ATP III 2005 criteria(Reference Perez, Guzman and Ortiz24, Reference Mujica, Leiva and Icaza25, Reference Medina-Lezama, Zea-Diaz and Morey-Vargas27) and one was based on both(Reference Pinzon, Serrano and Diaz26) (Table 1). Two studies, in addition to the diagnosis of MS by ATP III criteria, also used IDF criteria(Reference Mujica, Leiva and Icaza25, Reference Pinzon, Serrano and Diaz26). Waist circumference was also not measured by the same method in all studies. In three of them, the method was not stated(Reference Mujica, Leiva and Icaza25, Reference Tull, Thurland and LaPorte28, Reference Gomez, Ramirez and Disdier29); in three other studies, measurements were taken at the mid-point between the lower-most rib and the iliac crest(Reference Marquezine, Oliveira and Pereira32, Reference Velasquez-Melendez, Gazzinelli and Correa-Oliveira33, Reference Escobedo, Schargrodsky and Champagne35). The remainder (six studies) used other techniques: the lesser curvature between the lower-most rib and the iliac crest(Reference Salaroli, Barbosa and Mill34); at the navel level(Reference Florez, Silva and Fernandez30); at the navel level after exhaling normally(Reference Medina-Lezama, Zea-Diaz and Morey-Vargas27); at the highest point of the iliac crest with minimal breathing(Reference Perez, Guzman and Ortiz24); and in accordance with the protocol of the Second National Health and Nutrition Examination Survey(Reference Pinzon, Serrano and Diaz26) or by replacing this variable with BMI measurements(Reference Aguilar-Salinas, Rojas and Gomez-Perez31). Furthermore, for blood pressure measurement, different techniques and instruments were used. A mercury sphygmomanometer was used in six studies(Reference Medina-Lezama, Zea-Diaz and Morey-Vargas27, Reference Tull, Thurland and LaPorte28, Reference Marquezine, Oliveira and Pereira32–Reference Escobedo, Schargrodsky and Champagne35); in another study, a standard aneroid sphygmomanometer was used(Reference Perez, Guzman and Ortiz24); in the rest, the type of instrument used is not mentioned(Reference Mujica, Leiva and Icaza25, Reference Pinzon, Serrano and Diaz26, Reference Gomez, Ramirez and Disdier29–Reference Aguilar-Salinas, Rojas and Gomez-Perez31). Seven studies specify the number of measurements carried out (either one(Reference Tull, Thurland and LaPorte28, Reference Marquezine, Oliveira and Pereira32), two(Reference Florez, Silva and Fernandez30, Reference Aguilar-Salinas, Rojas and Gomez-Perez31, Reference Salaroli, Barbosa and Mill34) or three(Reference Perez, Guzman and Ortiz24, Reference Velasquez-Melendez, Gazzinelli and Correa-Oliveira33)) (Table 3). The majority stated that, before carrying out the first, or the only, measurement, the patient rested in a sitting position for 5 min(Reference Perez, Guzman and Ortiz24, Reference Tull, Thurland and LaPorte28, Reference Florez, Silva and Fernandez30–Reference Marquezine, Oliveira and Pereira32, Reference Salaroli, Barbosa and Mill34). When the second or third blood pressure measurement was taken, the subject rested for between 2(Reference Velasquez-Melendez, Gazzinelli and Correa-Oliveira33) and 30 min(Reference Salaroli, Barbosa and Mill34). Some studies took the average of the measurements as the value for blood pressure(Reference Florez, Silva and Fernandez30, Reference Salaroli, Barbosa and Mill34). Studies by Pinzón et al. (Reference Pinzon, Serrano and Diaz26) and Medina et al. (Reference Medina-Lezama, Zea-Diaz and Morey-Vargas27) used the protocols established by the American Society of Hypertension and the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure, respectively. Two articles gave no detailed explanation of how blood pressure was determined(Reference Mujica, Leiva and Icaza25, Reference Gomez, Ramirez and Disdier29).
A few of the analysed studies described environmental factors associated with the development of MS. It is beneficial to have this information, especially for potentially modifiable factors such as diet and physical activity. Nevertheless, this was not the main objective of the present review, which means that there may be other articles that provide more information on the topic, and hence another systematic review may be called for to examine this topic in more depth.
Morbidity and mortality trends in Latin America have changed in recent decades because of shifts in nutrition and lifestyle habits among the population. Urbanization and technology have transformed work, transportation and lifestyle behaviours, leading to a reduction in daily physical activity. There has also been a rise in the distribution of processed (energy dense) foods, which has had adverse effects on dietary patterns(Reference Drewnowski43). As a result of these factors, the region has gone from being one in which malnutrition was the main concern to one in which cardiovascular and chronic degenerative diseases are now the leading cause of death, as has been the case in developed countries(Reference Popkin44). It is therefore important to increase knowledge about MS and its determinants, as such knowledge could hasten the implementation of measures to reduce the prevalence of MS and its components, which could otherwise continue their upward trend in parallel with cardiovascular mortality(Reference Colin and Loncar45).
Conclusion
To our knowledge, this is the first systematic review of the prevalence of MS among different populations in Latin America. Our data suggest the need for multi-centre studies using current criteria that define MS(Reference Alberti, Eckel and Grundy8) and standardized methods that may make it possible to determine the precise scope of this problem among Latin-American populations. It is also essential to explore possible environmental factors, especially those related to lifestyle. Knowledge of these factors may be helpful in the monitoring of MS and could contribute to planning and prevention strategies to combat this condition, which is associated with a high risk of mortality and increased health-care costs for Latin-American governments and populations. Hence, all efforts to enhance knowledge of and deal better with MS are definitively justified.
Acknowledgements
F.M.-S. obtained a government grant from the CONACYT (Apoyos complementarios para el fortalecimiento de Cuerpos Académicos, modalidad repatriación 2009) programme. G.M.-O. received a scholarship student in a doctoral programme of recognized quality (beca nacional de CONACYT). D.V.-H., a student of Bachelor's degree in Nutrition, received a 6-month PRO-SNI grant from Universidad de Guadalajara. The authors have no conflict of interest to declare. B.V. and F.M.-S. directed the present study; B.V., J.S.-S. and J.F.-B. contributed to the study concept and design; F.M.-S., G.M.-O. and D.V.-H. helped with article searches, review and selection; F.M.-S., G.M.-O. and B.V. contributed to the analysis and interpretation of data and drafting of the manuscript; J.F.-B. and J.S.-S. worked as methodological advisors. All the authors contributed equally to the critical revision of the manuscript. The present paper is the first fruit of the LATINMETS (Latin America Metabolic Syndrome) project and the first collaborative effort by RIBESMET (Red Iberoamericana sobre el estudio del Síndrome Metabólico). The authors thank the following collaborators from the Licenciatura en Nutrición (Bachelor's Degree in Nutrition), who participated in the first stage of the study: Enriqueta Yolanda Castro Acosta, Iris Priscila Olivares Cano and Diana Gámez Romero. They also thank Celia de la Cruz Santana from the Departamento de Reproducción Humana, Crecimiento y Desarrollo Infantil, Centro Universitario de Ciencias de la Salud, Universidad de Guadalajara, and also thank their linguistic advisor, Robert Kimpleton.
