Prevalence of Pre-obesity and Obesity in Urban Adult Mexicans in Comparison with Other Large Surveys
Periférico Sur 4809 el Arenal Tepepan, Tlalpan 14610, Mexico D.F. MEXICO. Phone (5) 655-9011; Fax (5) 655-8211; E-mail email@example.com
Objective: 1. To estimate the prevalence of pre-obesity and obesity in a 1992 to 1993 national survey of the Mexican urban adult population. 2. To compare our findings with other national surveys and with data for Mexican Americans.
Research Methods and Procedures: The national representative sample of the Mexican urban adult population included 8462 women and 5929 men aged 20 to 69 years from 417 towns of >2500 people. Body mass index (BMI), calculated from measured weight and height, was classified using the World Health Organization categories of underweight (BMI < 18.5 kg/m2), normal weight (BMI 18.5 to 24.9 kg/m2), pre-obesity (PreOB = BMI 25 to 29.9 kg/m2) and obesity (OB = BMI 30+ kg/m2). Estimates for Mexican Americans were calculated from U.S. survey data.
Results: Overall, 38% of the Mexican urban adult population were classified as pre-obese and 21% as obese. Men had a higher prevalence of pre-obesity than women did at all ages, but women had higher values of obesity. Both pre-obesity and obesity increased with age up to the age range brackets of 40 to 49 or 50 to 59 years for both men and women. Both pre-obesity and obesity prevalence estimates were remarkably similar to data for Mexican Americans from 1982 through 1984. Comparison with other large surveys showed that countries differed more in the prevalence of obesity than of pre-obesity, leading to differences in the PreOB/OB ratio, and that countries also differed in the gender ratio (female/male) for both pre-obesity and obesity.
Discussion: Pre-obesity and obesity were high in our population and increased with age. Our approach of characterizing large surveys by PreOB/OB and gender ratios appeared promising.
Urbanization and aging are the population trends that have changed the demographic features of Mexico in this century. The percentage of the population living in localities with more than 2500 people increased from 29% to 71% from 1910 to 1990, and this urban sector is now concentrated in fewer than 1600 localities (<2% of the total localities) (1). Family planning programs, and a general improvement in public health conditions, have substantially modified the age composition of the Mexican population since 1970; in the period from 1970 to 1990, for example, the percentage of the population consisting of people aged 10 years or more increased from 67.1% to 74.5% (1, 2). These trends have contributed to major shifts in the Mexican epidemiological profile. Communicable diseases play a less important role in morbidity and mortality and have been overtaken by cardiovascular diseases and other degenerative pathologies (1, 2).
In the period 1992 through 1993, the Ministry of Health of Mexico conducted the National Survey of Chronic Diseases to estimate the prevalence of obesity, type 2 diabetes, renal pathology, coronary heart disease, dyslipidemia, and hypertension in the urban population. We present here data on the distribution of different categories of body mass index (BMI) obesity, by age and gender, in urban adult Mexicans. In addition, we compare our results with other populations, especially with Mexican Americans (people of Mexican origin or descent living in the United States) as well as with data from other large national surveys.
Research Methods and Procedures
A multistage sampling procedure was used. Mexico is a republic of 31 states and a federal district in which Mexico City is located. The country was divided in three regions (Northern, Central, and Southern) of 10 states each, and half the states were randomly selected for inclusion in our survey. A fourth region was the Metropolitan area with about 20 million people comprising Mexico City and 10 adjacent counties of the State of Mexico. These regions differ in industrialization and mortality indicators (2). A random sample of Basic Geographical Statistical Units was obtained in each state and in the federal district from a database periodically updated by the Instituto Nacional de Geografía y Estadística (National Institute of Geography and Statistics). The sampling was done by considering only towns with more than 2500 people. Neighborhood blocks were selected randomly and all adults (20 to 69 years of age) in all households of the selected blocks were surveyed with the exception of those living in military, religious, health, and other institutions. A total of 417 towns were studied. The sample was representative of the Mexican urban population, which in 1990 constituted 71% of the total population (1). A target of 4730 individuals and 2030 households per region was estimated using the household as the sampling unit and using the average of 2.33 adults per household from the 1990 National Census. Our sample size was considered capable of detecting risk factors having at least a prevalence of 4% with a relative permissible error of estimation of 0.29 and a non-response rate of 30%. A preliminary survey showed that diabetes with a prevalence of 4% was the least frequent of the diseases we were looking for in the survey.
A questionnaire was used to obtain information on age, family history, clinical symptoms, and medical treatment for various chronic diseases of the individuals. Weight and height were measured while the participants wore light clothing with no shoes using a portable scale and a metric tape adhered to a wall. BMI was calculated by dividing the weight (in kilograms) by the square of height (in meters). The World Health Organization criteria for underweight (BMI < 18.5 kg/m2), normal (BMI 18.5 to 24.9 kg/m2), pre-obese (PreOB = BMI 25 to 29.9 kg/m2), and obese (OB = BMI 30+ kg/m2) were used (3). A small number of individuals (N = 12) who did not know their age were excluded from the database.
Comparative data for Mexican Americans were obtained from two U.S. surveys conducted by the National Center for Health Statistics (4, 5). Data on a representative sample of Mexican Americans in the Southwestern United States were collected as part of the Hispanic Health and Nutrition Examination Survey (HHANES; 1982 through 1984). National data for Mexican Americans were also collected as part of the Third National Health and Nutrition Examination Survey (NHANES III; 1988 through 1994). Prevalence estimates for these two surveys were calculated using the same BMI categories and age categories that were used for the Mexican survey to make comparisons between Mexicans and Mexican Americans.
The sample sizes and the prevalence of underweight, normal weight, pre-obesity, and obesity in our sample are shown in Table 1 by age and gender. The prevalence of underweight was very low, both for men and for women. Overall, only 39% of the individuals were normal, 38% were pre-obese, and 21% were obese. In all age groups except the youngest, the prevalence of normal weight was below 50% and more than half of the population was classified as pre-obese or obese. Men had a higher prevalence of pre-obesity than women (41% vs. 36%) but a lower prevalence of obesity (15% vs. 25%). The prevalence of pre-obesity peaked in the 50- to 59-year age bracket for men and in the 40- to 49-year age bracket for women. The prevalence of obesity followed a similar pattern but peaked in the 40- to 49-year age bracket for men and in the 50- to 59-year age bracket for women. There were a smaller number of men than women in our sample. We believe this was mostly due to unavailability of men at their homes during daylight hours and less to temporary migration of more men than women (basically as field and house workers in the United States). On the other hand, some migrant bias may be present in all population studies in Mexico, and it would be of interest to explore this in future studies.
Table 1. Distribution of BMI categories in the Mexican sample, 1992 through 1993
| || || || || || || || || || || |
| || || || || || || || || || || |
In Tables 2 and 3 we show the prevalence of the same BMI categories in two cohorts of Mexican Americans. We will refer to these two cohorts as MEX-AM-I (Hispanic HANES, 1982 through 1984) and MEX-AM-II (NHANES III, 1988 through 1994). As was seen also in the Mexican population, women in both Mexican American cohorts had a lower prevalence of pre-obesity than men but a higher prevalence of obesity. Overall, the prevalence estimates for Mexican Americans from the earlier cohort (1982 through 1984) were remarkably similar to our data for adult urban Mexicans. The prevalence of pre-obesity was 38% in Mexico and 38.7 for Mexican Americans, and the prevalence of obesity was 20.9 in Mexico and 20.6 for Mexican Americans. The age-specific estimates and the trends with age were also very similar between Mexicans and Mexican Americans from the 1982 through 1984 survey.
Table 2. Distribution of BMI categories in Mexican-Americans: Hispanic HANES, 1982 through 1984
| || || || || || || || || || || |
We consider our sample as predominantly urban because 86% of our cases lived in cities with more than 15,000 people and only 14% could be taken as semi-urban living in communities of 2500 to 15,000 people. Our global estimates for the prevalence of pre-obesity plus obesity (BMI ≥ 25) were only slightly higher than those reported recently by Sanchez-Castillo et al. (6) in a 1995 Mexico City sample of 2042 individuals (50% and 58% for males and females by Sanchez-Castillo et al. vs. 56% and 61% in our data). On the other hand, these rates were higher than those of Sepulveda-Amor (7) in a large sample of 19,022 Mexican women, but these were young (aged 12 to 49) and living in urban and rural communities, and thus, not comparable to our urban adult sample.
Table 4 compares the prevalence of pre-obesity and obesity and the PreOB/OB ratio with other large surveys (8, 9, 10, 11, 12, 13, 14, 15, 16) as well as 11 surveys collected by Popkin and Doak (17). These surveys vary by age range, sample design, and methods of measurement and classification, with some surveys using measured weights and heights and a few relying on self-reporting. Nonetheless, they allow for some useful comparisons between countries. The table shows that the prevalence of pre-obesity is highest in Mexicans and Mexican Americans and topped only by South Africans with a 40% prevalence (it is falsely high in the latter because that study also considered women with BMI of 24 to 24.9 as pre-obese). Samoans are the undisputed leaders in obesity, although the surveys from both Samoa and Mauritius exclude individuals below 25 years of age (and Thailand starts at age 35). Age in other studies comprised younger subjects (18 to 20 in most and from age 15 in Italy, Saudi Arabia, and South Africa). It is important to note that the differences between countries were more a result of larger variability in the prevalence of obesity (range 1% to 59%) and less to that of pre-obesity (range 4% to <40%). This led us to incorporate a PreOB/OB ratio in Table 4 that ranged from 0.9 in Kuwait to 15.3 in China. Only one survey (Kuwait-II) had a PreOB/OB ratio below 1. This ratio exhibited an interesting discriminatory power in regard to changes in sequential studies. For example, NHANES-III has a PreOB/OB ratio of 1.4, which is clearly lower than the ratios of 2.2 to 2.4 of the previous surveys indicating a shift in NHANES-III to a larger proportion of obese individuals. This increase in obesity in the U.S. general population has been shown to exist using other analytical strategies (9). The phenomenon was not the same in Mexican Americans who decreased their PreOB/OB ratio from 1.9 in MEX-AM-I to 1.4 in MEX-AM-II. Thus our ratio of 1.8 was more similar to Mexican Americans of 1982 through 1984 (10 years before our study) than to those of 1988 through 1994. As to the terminology, the ratio could be designated as the OW/OB in view of a recent National Heart, Lung & Blood Institute guideline defining overweight (OW) as a BMI of 25 to 29.9 (18).
Table 4. Prevalence (%) of pre-obesity and obesity and the PreOB/Ob ratio in large surveys of different countries (combined data of both genders). Ordered by prevalence of obesity.
|Netherlands||12||Regional||1987–1991||20–59||32,266|| ||8%|| |
|Morocco||17||National||1984–1985||20+|| || ||9%|| |
|Our study|| ||Urban||1992–1993||20–69||14,392||38%||21%||1.8|
|Samoa||16||Regional||1991||25+||1786|| ||59%|| |
In Table 5 we show a gender ratio (female/male [F/M]) in the PreOB and OB categories using the surveys of Table 4 with gender information as well as two additional surveys of Popkin and Doak (17) who only give prevalence ratios of PreOB and OB by gender (Kyrgyzstan and Philippines). The F/M ratio varied more in the obese (range 0.9 to 3.5) than in the pre-obese (range 0.6 to 1.2 plus Tunisia with an unusual ratio of 1.6 in pre-obese). Only one country (Italy) had a gender ratio of <1 in the obese. Our population, together with NHANES and MEX-AM surveys, had a F/M ratio of <1 in the pre-obese and >1 in the obese. It should be noted that the F/M ratio in the obese was always larger than that in the pre-obese, suggesting that women are more prone to obesity than men in practically all countries. We feel it is too early to know the usefulness of this gender ratio as an indicator of the interplay of gender with pre-obesity and obesity in different populations, but the data in Table 5 look promising.
Table 5. Prevalence (%) of pre-obesity and obesity by gender and the female/male ratio for pre-obesity and obesity in the surveys of Table 4
|South Africa*|| || ||39||42||0.9||18||15||1.2|
|Thailand|| || ||21||26||0.8||3||2||1.4|
|Malaysia|| || ||18||24||0.8||8||5||1.7|
|Kuwait-I|| || ||27||31||0.9||30||15||2.0|
|Kyrgyzstan|| || ||24||26||0.9||11||4||2.5|
|Russia†|| || ||32||34||0.9||26||10||2.7|
|Philippines|| || ||12||11||1.1||3||2||2.0|
|China*|| || ||17||14||1.2||6||2||2.6|
|Tunisia|| || ||33||20||1.6||8||2||3.5|
In conclusion, it is clear that the prevalence of pre-obesity plus obesity was above 20% in 17 of the 20 countries in Table 4 (the exceptions were India, Mali, and China with prevalence ratios of 5% to 10%). We believe that further study of the reasons underlying the differences shown here for gender (Table 5) may prove useful in establishing better guidelines to fight what is now being recognized as the obesity epidemic (19).