Early-Life Metabolic Traits and Physical Fitness in Tarahumara, Mennonite, and Mestizo Adolescents from Northern Mexico

The WHO identifies high BMI, high blood pressure, and high fasting plasma glucose as chronic disease risk factors, whereas physical fitness is identified as a protective behavioral factor. This study responds to the rising interest in assessing metabolic factors and physical activity within young populations of Mestizo, Tarahumara, and Mennonite from Chihuahua Mexico, due to its strong relationship with disease development and low well-being. A cross-sectional study was conducted with 201 teenagers from rural towns in Northern Mexico, and relationships between physical fitness and cardio-metabolic risk related to anthropometric, glycolipid, and vascular function factors were assessed. ANOVA-tested differences among ethnic groups using physical fitness as a grouping variable and measures of cardio-metabolic risks were used as dependent variables. A stepwise regression analysis allowed us to identify the best predictors for physical fitness. Clinical risk factors were analyzed by ethnic group and sex. No differences were found among ethnic groups in physical fitness and cardio-metabolic health risks; sex differentiated higher health risks related to behavioral factors, since young women showed lower physical fitness across ethnicities. Clinically, the Mestizo sample showed higher numbers of individuals with one risk factor. Mennonites showed a high frequency of anthropometric and fitness health risks with low glycolipid and vascular risks. Tarahumara had fewer risk factors as compared with both Mestizo and Mennonite. Rural populations are harder to reach, both for health assessment and intervention; health professionals must work close to local community organizations to gain access.


Introduction
According to the Centers for Disease Control and Prevention (CDC), in the USA, obesity has increased almost three times in adolescents (12-19 years) in the last 30 years [1]. The 2018 National Survey of Health and Nutrition of Mexico reports that 36.8% of adolescents are either overweight or obese. Chihuahua state in Northern Mexico has 9.3% of the total country's population with this phenotype [2], being acknowledged as a population health problem [3]. Overweight and obesity can be defined as a caloric imbalance, which means expending fewer calories compared to the ones consumed [4]. Childhood obesity is defined as a body mass index (BMI, kg/m 2 ) greater than or equal to the 95th percentile according to specific sex-age-related growth charts [5]. Factors contributing to overweight include genetic susceptibility, environment, and behavior, the latter having the advantage of being modified by physical activity and healthy nutrition [6]. Obesity in young people has

Subjects and Study Design
A cross-sectional study was performed in rural high schools, from Chihuahua State, Mexico (cities of Guachochi, Cuauhtémoc, and Carichí).
To increase participation and inclusiveness for ethnicity and sex in all studied communities, we publicly recruited participants gaining access through local authorities. Regular channels of communication in the community were used with written, oral, and personal invitations to participate in both Spanish and Raramuri languages. Local authorities helped to gain access to participation. We offered participation incentives such as individual health reports for every participant; we also offered group reports whenever authorities and community representatives were interested.
All school authorities and parents received information about the study and all requirements for children's voluntary participation. Written consent forms were provided to parents and tutors.
Inclusion criteria for all measures were having written consent from parents, understanding all instructions, fasting for at least 8 h, and voluntary acceptance for each test.
The convenience sample we used reflects those who were available in each community providing signed consent forms. All children with parental authorization also accepted every assessment voluntarily after being informed about each procedure.
Data were processed from 12-to-19-year-old students with complete information for ethnicity, sex, and physical fitness variables. The ethnicity for each participant was identified by social workers in school records.
The resulting sample was 201 individuals (selected by inclusion criteria from a previous main study of 1155 participants) distributed as shown in Table 1. We excluded individuals with missing data for each analysis.

Anthropometric Measures and Body Composition
Barefoot, standing height was measured with a manual stadiometer (Seca, Model 206). Participants wore light clothing and were barefoot when body mass was determined using a calibrated digital scale (Tanita, Model TBF 300, Tokyo, Japan). BMI was calculated. BMI CDC percentile was used for classification [1].

Blood Sample
Peripheral blood was collected in the supine position by venipuncture from the antecubital fossa by trained persons. Blood (4 mL) was collected into EDTA tubes and placed on ice, while another 6 mL was collected into SST tubes and allowed to clot for 20-30 min.
After fasting for blood sample collection and 30-60 min before running the physical fitness tests, scholars had a small breakfast consisting of liquid yogurt, one apple, and a small piece of sweet wheat bread.

Glucose and Lipids
Glucose sampling was performed in triplicate using the glucose GOD-PAP method (YSI 2300 Stat Plus, Yellow Springs Instrument, Yellow Springs, OH, USA). The average of the 3 glucose measurements was used in all analyses, and according to the WHO, normal values for the young population were established as 3.9 to 5.5 mmol/L.
Lipid profile, including measures of total cholesterol (CHOD-PAP method) (normal values less than 4.4 mmol/L), high-density lipoprotein cholesterol (HDL-C, normal values greater than 1.04 mmol/L), low-density lipoprotein cholesterol (LDL-C = CT-(TG/5)-HDL-C or CT-VLDL-HDL-C), and triglycerides (TGs) were determined through enzymatic assays (Human ® , Mannheim, Germany) using an automatized biochemical analyzer (Prestige 24i ® , Tokyo Boeki ® Medical System LTD, Japan) (normal values less than 1.9 mmol/L). The atherogenic index was calculated through the formula Atherogenic Index = total cholesterol/HDL-C (normal values for males > 4.0 and females > 3.5 [17,18], which is used as a predictive cardiovascular risk factor [18]. The same cut-off values for both males and females were used since sex differences among children and adolescents have been observed as non-significant [19].

Physical Fitness
Flexibility: The sit and reach test allowed us to determine the flexibility of the lower back and hamstring muscles. Sitting on the floor, legs stretched out and barefoot placed against the measurement box. The subject reaches as far as possible forward with palms facing downwards and arms extended [20,21].
Strength: Arm, leg, and back strength measurements were carried out with a dynamometer with 660 pounds capacity (Baseline Dynamometer model 12-0403, Wayne, NJ, USA), testing postures and movements were followed as described by Ten Hoor, Musch [22]. The back, arm, and leg measurements were repeated three times and the mean ± standard deviation (SD) was reported. Values were then converted to kg.
Aerobic capacity: The multistage 20-m shuttle run test allowed the estimation of the VO 2 max, with scores having a moderate-to-high criterion-related validity for estimating maximum oxygen uptake (r p = 0.66-0.84); this test is considered a useful field alternative to lab tests for estimating cardiorespiratory fitness [23,24] and it has been and is currently still in use to estimate aerobic fitness [25].
The test required running 20 m back and forward according to a rhythm stipulated by a prerecorded beep sound. Then, the following equation was used to determine the individual aerobic capacity: VO 2 max (mL O 2 min −1 kg −1 ) = 5.85 × Velocity − 19.45, where velocity is the maximal velocity reached at the end of the test.

Statistical Analysis
We present all results stratified by ethnicity and sex. Mean ± SD was used in normally distributed data, while the median was used for not normally distributed data.
We conducted a series of analyses of one-way variance (ANOVA) examining ethnicity and physical fitness as independent variables, and a Tukey test to assess statistical significance. Body composition, heart functions, and metabolic/biochemical measures were used as dependent variables to test differences between ethnic groups.
Comparisons were tested using Chi 2 test values to assess significant differences among means of physical fitness measures by ethnicity, and in percentages on incidence of abnormal values for measured variables among ethnic samples. Simple linear stepwise regression was used to identify the best-predicting equation; as an exploratory technique, it was considered useful to eliminate variables superfluous for physical fitness to tighten future research. Pearson correlations were included for metabolic and physical fitness variables. SPSS v 21.0 Armonk, NY, USA) was used for all statistical analyses.

Physical Fitness
To analyze differences between males and females, and differences among ethnicity by physical fitness, a compound variable was calculated with individuals having complete data on the following measures: strength (arm, leg, and back), flexibility, and VO 2 max. Table 3 shows descriptive statistics for those measures, by sex and ethnicity. Using data from flexibility, arm strength, leg strength, back strength, and VO 2 max, a new variable "physical fitness" was construed. Descriptive statistics for physical fitness are included by ethnicity and sex in Table 4. A Chi 2 test showed no differences among ethnic groups on physical fitness measures (Chi 2 (1, 368) = 372.653; p = 0.423). However, in a physical fitness by sex analysis, males showed significantly higher physical fitness than females across ethnicity (Chi 2 (1; 184) = 248.824; p = 0.001).
Based on the physical fitness score, boundaries were defined for low, normal, and high fitness using 50 and 25± percentiles. Percentages of the sample by fitness condition are shown in Figure 1.    The homogeneity assumption of variances among Mestizo, Tarahumara, and Mennonite was met for systolic blood pressure, minimum cardiac frequency, glucose, cholesterol, TG, HDL-C, LDL-C, VLDL, and Atherogenic Index. However, homogeneity of variance assumption was not observed for DBP (Lavene statistical p = 0.004) or for waist-hip index (p = 0.045).

Distribution of Cases with Normal and Abnormal Variables by Ethnicity and by Sex
Assessment of the OMS three global top risk variables of metabolic risks for years of life lost in 2040 (high blood pressure, high BMI, and high fasting plasma glucose) showed no significant ethnicity differences in our sample.

Distribution of Cases with Normal and Abnormal Variables by Ethnicity and by Sex
Assessment of the OMS three global top risk variables of metabolic risks for years of life lost in 2040 (high blood pressure, high BMI, and high fasting plasma glucose) showed no significant ethnicity differences in our sample.
The male Mestizo group showed the highest percentage (19.35%) of individuals having one, up to three, cardio-metabolic risk variables; fewer Tarahumara individuals (6.44%) showed elevated values on risk variables, as compared to Mennonite (8.86%). Male Mestizos also showed a higher percentage of individuals (35.47%) having abnormal values on additional risk variables, as compared with Tarahumara (14.43%) and Mennonite (21.5%), with Tarahumara having a lower percentage of affected individuals.
Mestizo females affected with one, up to four, risk factors were also on the highest percentage (28.55%), followed by Tarahumara (15.57%), and Mennonite (3.89%) with the lowest percentage of affected females; Mestizo females with additional risk factors, from one, up to four, showed also the highest percentages of affected individuals (49.33%), followed by Tarahumara (15.57%), and Mennonite with the lowest percentage of affected females (3.89%). Individuals affected with additional risk factors were observed in Mestizo females (49.33%), followed by Tarahumara (25.96%), and Mennonite (3.89%) being the least affected ones (Table 5).   Frequencies of individual risk factors were calculated using all variables assessed in this study. We calculated the total number of variables showing abnormal values for each participant to identify clinical variables of cardio-metabolic risk related to anthropometric, glycolipid, and vascular function factors. Table 5 shows percentages of individuals by ethnicity and sex, having from one, up to four, abnormal values on measured variables in this study. Mestizo females affected with one, up to four, risk factors were also on the highest percentage (28.55%), followed by Tarahumara (15.57%), and Mennonite (3.89%) with the lowest percentage of affected females; Mestizo females with additional risk factors, from one, up to four, showed also the highest percentages of affected individuals (49.33%), followed by Tarahumara (15.57%), and Mennonite with the lowest percentage of affected females (3.89%). Individuals affected with additional risk factors were observed in Mestizo females (49.33%), followed by Tarahumara (25.96%), and Mennonite (3.89%) being the least affected ones (Table 5).

Regression Analysis
We used stepwise regression to build a model, not to test one. According to this purpose, the stepwise criteria were probability of F to enter ≤ 0.050, probability of F to remove ≥ 0.100. However, since for this analysis, only cases with complete data were selected, our sample size for linear regression was small (n = 21) (Table 6). Therefore, a two halves analysis was not conducted, limiting our conclusions. , with LDL-C as the best predictor of physical fitness (β = −0.468, p = 0.032).
All other variables (Height in meters, BMI, SBP, DBP, glucose, cholesterol, TG, HDL-C, LDL-C, VLDL, and Atherogenic Index) did not significantly predict physical fitness and were excluded from the exploratory model, as shown in Table 7.
A Pearson correlation coefficient was computed to assess the linear relationship between variables included in the regression analysis.

Discussion and Conclusions
The Sustainable Development Goals defined by the general assembly of the United Nations set Target 3.4 as "By 2030, reduce by one-third premature mortality from noncommunicable diseases through prevention and treatment and promote mental health and well-being" [26]. Three main risk factors associated with premature death were identified in the global population: high blood pressure, high BMI, and high fasting plasma glucose. Moreover, the protective effects of physical fitness have been demonstrated against those health risk factors [27].
The existence of multiple metabolic and behavioral risk factors among the Mexican population demands appropriate prevention strategies to halt the progress of the noncommunicable disease burden within the region.
To improve health in Northern Mexico rural populations, identifying health risk factors is a necessary task. The purpose of this study was to evaluate relationships between physical fitness and health measures, comparing three ethnic samples of youth living in Northern Mexico.
Results showed no significant differences among ethnic groups for physical fitness and cardio-metabolic health risks but sex-differentiated higher health risks related to behavioral factors since young women showed lower physical fitness across ethnicities than young men. From the exploratory model, using measures to assess metabolic health, LDL was the most significant predictor of physical fitness; however, due to the small sample, this result is not conclusive. We need to test this result with a sample having more participants from both sexes.
We Higher percentages of afflicted individuals, with multiple health risks among Mestizos (35.4%), as compared with Mennonites (21.5%) might be related to a combination of lower physical fitness and different and unhealthier feeding practices related to ethnicity. According to this, Tarahumara had a lower percentage of affected individuals (14.43%).
Mestizo females affected with one, up to four, risk factors were also on the highest percentage (28.55%), followed by Tarahumara (15.57%), and Mennonite (3.89%) with the lowest percentage of affected females; Mestizo females with additional risk factors, from one, up to four, showed also the highest percentages of affected individuals (49.33%), followed by Tarahumara (15.57%), and Mennonite with the lowest percentage of affected females (3.89%). Individuals affected with additional risk factors were observed in Mestizo females (49.33%), followed by Tarahumara (25.96%), and Mennonite (3.89%) being the least affected ones. However, those differences might be affected by the small number of female participants in Tarahumara and Mennonite samples.
Elevated cholesterol (39.30% n = 79 individuals) and TG (29.35% n = 59 individuals) were the two most prevalent risk factors. Mestizos showed higher numbers of individuals with one risk factor. Mennonites showed a high frequency of anthropometric and fitness health risks with low glycolipid and vascular risks. Tarahumara participants were the lowest in health risks as compared with both Mestizo and Mennonite individuals.
More than nineteen percent (19.35%) of male Mestizos showed metabolic risk factors, followed by Tarahumara males (6.44%). Less than nine percent (8.86%) of Mennonite individuals have one or more risk factors for chronic diseases.
Our group's main study determined the prevalence of metabolic syndrome (MS) risk factors in 1155 scholars that included 59.7% Mestizos, 24.7% Tarahumara, and 15.7% Mennonite. The prevalence of MS in the sample was higher in Tarahumara adolescents (12.65%) compared to Mestizo (11.95%) and Mennonite (7.15%), according to the ATP III criterion [28,29]. The significant statistical differences in the prevalence of MS in adolescents from the different ethnic groups in Chihuahua may include diet quality, behavior, physical fitness, and genetic (inherited polymorphisms) differences.
In Mexico, the prevalence of dyslipidemia, arterial hypertension, obesity, and diabetes mellitus in the child and adolescent population has increased to a great extent during the last decade [30].
We found different results in terms of overweight and obesity in the total study population (27.6%) to those reported by Cárdenas-Villareal et. al. (2010) conducted with 254 adolescents from the city of Monterrey, Nuevo León, México that showed 21% for overweight and obesity. These data are lower than the reported prevalence of both entities in adolescents by the 2018 National Health and Nutrition Survey, which describes 22.5% overweight and 13.9% obesity in young Mexicans between 12 and 19 years of age [31].
Currently, there is no consensus on the possible impact on public health of the identification of cardiovascular risk components at an early age, both in our country and in our state, and the importance of phenotypic differences among ethnic groups. Therefore, we encourage researchers to continue multidisciplinary studies to improve knowledge and strategies for the well-being of our young populations.

Limitations
This study included only the regions of Cuauhtémoc, Guachochi, and Carichí. The results exposed here could lead to an incomplete understanding of the current health and physical fitness condition of the ethnic groups living in Chihuahua, Mexico. Nutrition and habits were not included in this study, which could give a more integrated view regarding metabolic traits and risk factors.
Since our Mennonite and Tarahumara samples were small, with smaller numbers of females, our data may not be representative of the population in Northern Mexico. Obtaining access to this rural population was a difficult task, as well as getting consent from females. Rural populations are harder to reach, both for health assessment and intervention; health professionals must work close to local community organizations to gain access. It will be interesting to continue this study in collaboration with the local government to design health strategies designed for each community. Including the adult population could give us a better understanding of the health problems that affect each ethnic group, especially those related to metabolic chronic diseases.

Institutional Review Board Statement:
The study was conducted in accordance with the Declaration of Helsinki and approved by the Ethics Board "Comité de Ética en Investigación de la Facultad de Medicina y Ciencias Biomédicas de la Universidad Autónoma de Chihuahua", with a registration number FM-FM-A269/12 and CI-040-17"; date of approval 15 June 2009, for studies involving humans.
Informed Consent Statement: Informed consent was obtained from all subjects involved in the study. All school authorities and parents received information about the study and on all requirements for children's voluntary participation. Written consent forms were provided to parents and tutors, and only students with written consent participated.
Data Availability Statement: Not applicable due to the small sample.

Conflicts of Interest:
The authors declare no conflict of interest. The community and school authorities in all participant communities had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript; or in the decision to publish the results.