AMERICAN JOURNAL OF HUMAN BIOLOGY 27:494–500 (2015)
Original Research Article
Intergenerational Influences on the Growth of Maya Children: The Effect of Living Conditions Experienced by Mothers and Maternal Grandmothers During Their Childhood ^ VARELA-SILVA2 HUGO AZCORRA,1* FEDERICO DICKINSON,1 BARRY BOGIN,2 LUIS RODRIGUEZ,3 AND MARIA INES Departamento de Ecologıa Humana, Centro de Investigaci on y de Estudios Avanzados del Instituto Politecnico Nacional (Cinvestav)-Unidad Merida, 97310 Merida, Yucat an, Mexico 2 Centre for Global Health and Human Development, School of Sports, Exercise and Health Sciences, Loughborough University, Loughborough LE11 3TU, United Kingdom 3 Facultad de Matem aticas, Universidad Aut onoma de Yucat an, Merida, Yucat an, Mexico
1
Objectives: To test the hypothesis that living conditions experienced by maternal grandmothers (F1 generation) and mothers (F2 generation) during their childhood are related to height and leg length (LL: height 2 sitting height) of their 6-to-8 year old children (F3 generation). Methods: From September 2011 to June 2012 we obtained height and LL, and calculated z-score values of these measurements for 109 triads (F1, F2, F3) who are Maya living in Merida, Yucatan, Mexico. Multiple regression models were adjusted to examine the relation of anthropometric and intergenerational socioeconomic parameters of F1 (house index and family size during childhood) and F2 (paternal job loss during childhood) with the z-score values of height and LL of F3. Results: Children’s height and LL were positively associated with maternal height and LL. This association was relatively stronger in LL. Better categories of grand-maternal house index were significantly associated with higher values of height and LL in grandchildren. Grand-maternal family size was positively related with LL, but not with height. Conclusions: Our findings partially support the hypothesis that living conditions experienced by recent maternal ancestors (F1 and F2) during their growth period influence the growth of descendants (F3). Results suggest that LL is more sensitive to intergenerational influences than is total height and that the transition from a traditional rural lifestyle to urban conditions results in new exposures for risk in human physical growth. Am. J. Hum. Biol. 27:494–500, C 2015 Wiley Periodicals, Inc. V 2015. INTRODUCTION The process of growth is a complex phenomenon strongly influenced by the social, economic and political conditions surrounding the group (Bogin and Loucky, 1997; Steckel, 2012; Stinson, 2012). It has been suggested that the current biological status of a group can also be explained by ecological conditions experienced by recent ancestors during their growth period. In 1986, Irving Emanuel defined the intergenerational influences hypothesis (IIH) as “those factors, conditions, exposures, and environments experienced by one generation that relate to the health, growth, and development of the next generation” (Emanuel, 1986, p 27). Subsequent theoretical contributions suggest that growth trajectories of children are also influenced by factors that trace the nutritional history of matrilineal ancestors, including their nutrition during prenatal stage and first years of postnatal life (Kuzawa, 2005). Most of the accumulated evidence of intergenerational influences on growth is on the prenatal stage (Alwasel et al., 2011, 2013; Emanuel, 1986, 1992, 1999, 2004; Hypponen et al., 2004; Jasienska, 2009; Klebanoff and Yip, 1987; Kuzawa, 2005; Martin et al., 2004; Ounsted et al., 1986). The intergenerational effects on postnatal growth have been less studied. Several studies provide evidence for an association between parental and offspring height. Correlation coefficients show that associations between mothers and offspring are stronger than associations between fathers and offspring in all ages (Alberman, 1991; Ghosh and Malik, 2007; Gray et al., C 2015 Wiley Periodicals, Inc. V
2012; Hypponen et al., 2004). However, parent-offspring correlations reflect both the effect of genes and shared environmental factors such as diet and life style (Towne et al., 2012). The Maya are one of the largest and most important ethnic groups in Mesoamerica (Thompson, 1973). Merida, the capital city of the state of Yucatan, in Mexico, is the home of a great number of Maya people who live in poor neighborhoods (Lizama, 2012). Historically, the Maya from Yucatan have lived under very unfavorable socioeconomic conditions (Bracamonte, 2007), and continue to do so (Azcorra et al., 2013; Varela-Silva et al., 2009, 2012). This has negatively influenced the biosocial status of the Maya. Previous studies show that Maya from Yucatan exhibit high rates of undernutrition (mainly short stature) during the growing period (Azcorra et al., 2009; Kelley, 1991; Siniarska and Wolanski, 1999; Wolanski et al., 1993; Varela-Silva et al., 2009, 2012), and in adulthood. Despite overall improvements of social and economic conditions since the end of the colonial period in Mexico (in 1821), research indicates the absence of a positive secular change in stature for the Maya people (McCollough, 1982; Siniarska and Wolanski, 1999). Contract grant sponsor: The National Council of Science and Technology of Mexico; Contract grant number: 168047. *Correspondence to: Hugo Azcorra. E-mail: [email protected] Received 17 June 2014; Revision received 18 November 2014; Accepted 10 December 2014 DOI: 10.1002/ajhb.22675 Published online 9 January 2015 in Wiley Online Library (wileyonlinelibrary.com).
495
INTERGENERATIONAL INFLUENCES ON THE GROWTH OF MAYA CHILDREN
In a previous article we described the nutritional status of a sample of 109 triads of Maya children, their mothers, and maternal grandmothers from Merida, Mexico (Azcorra et al., 2013). We confirmed the hypothesis that leg length-relative-to-stature is a more sensitive indicator of nutrition and health status than is total height or sitting height alone. In that study the analysis was based on the examination of correlations between grandmothers and children and mothers and children. In the present study we extend the analysis adding a bio-cultural model to test the IIH and ascertain the influence of the biosocial background of urban Maya grandmothers (F1 generation) and mothers (F2 generation) on the linear growth and nutritional status of their children (F3 generation). This study contributes to the body of knowledge about the intergenerational influences on postnatal growth in groups with a long adverse socioeconomic history. METHODS From September 2011 to June 2012 a cross-sectional study was undertaken on a sample of 109 triads of urban Maya children (6-to-8 year old), their mothers and their maternal grandmothers from the city of Merida. The number of triads needed for the study was obtained through a power analysis calculation. Maya surnames were used as a proxy for Maya ancestry; the requirement for inclusion in the sample was that children, mothers and grandmothers had to have at least a maternal Maya surname. Schools (n 5 20) where recruitment was conducted were randomly selected from the south of Merida, where Maya language speakers and people with the lowest level of income are concentrated. To allow for maximum geographic and socioeconomic variability in the sample, we selected no more than 10 children from each school. More details of sample size calculation and recruitment process of participants are described by Azcorra et al. (2013). Measurements We only focused, in this study, on height and sitting height of the participants. Leg length (LL) was calculated by subtracting sitting height from height. Height and LL were transformed to z-score values using the Comprehensive References published by Frisancho (2008). Particularly in the case of mothers and grandmothers, z-score values were useful to accommodate the effect of loss of height with age. We focused on the intergenerational effects on the height and LL of children because these measurements provide information about the nutritional history of individuals (Bogin and Keep, 1999; Bogin and Varela-Silva, 2010), and are related to the risk for adult mortality (Barker et al., 1990; Leon et al., 1995), and health outcomes at adulthood such as obesity, insulin resistance, type 2 diabetes, and heart diseases (Asao et al., 2006; Forsen et al., 2000; Lopez-Alvarenga et al., 2003). The biosocial background of the mothers and grandmothers was analyzed through (1) current anthropometric characteristics: height and LL (z-score values) and (2) socioeconomic indicators experienced during childhood. Children were measured at schools and adult women at homes, all of them by trained personnel. Anthropometric measurements were taken following standardized methods (Lohman et al., 1988).
Home visits were used to administer socioeconomic questionnaires to mothers and grandmothers. Data on socioeconomic conditions experienced during maternal and grand-maternal childhood were obtained through these visits. The questionnaire was based on our field experience and ethnographic knowledge of the local context. Mothers and grandmothers were interviewed separately. The assistance of the mothers was requested when grandmothers did not understand the questions because they did not speak Spanish fluently. Interviews indicated that for grandmothers there were two sensitive indicators of their childhood living conditions: (1) the materials of construction of the house, and (2) family size. Three categories of grand-maternal house type were defined: (1) perishable materials (dirt or cement floor, cardboard or metal sheets for walls and ceilings), (2) traditional Maya rural type (dirt floor, palm leaf ceilings and wattle and daub walls), and (3) durable materials (cement floor and cement blocks in walls and ceilings). Family size was recorded as the total number of people living in the home. Maternal house type was almost uniformly of durable materials, therefore it was not possible to use as a variable. Instead, our interviews revealed that for mothers, the important variable of their childhood living conditions was their father’s employment stability. The rationale for the use of this variable is that around 57% of the mothers’ parents migrated from rural areas to the city of Merida in order to get a job. But, the loss of paid employment among men was common because of the nature of the jobs and to their low qualifications. The most common occupations were masons and laborers in the construction industry. The more frequent the father’s job loss, the generally lower the childhood living conditions of the mother. Ethical concerns The research was approved by the Bioethics Committee for the Study of Human Beings of the Center for Research and Advanced Studies of the National Polytechnic Institute of Merida, Mexico and by the Loughborough University Advisory Ethical Committee. The mothers and grandmothers signed consent forms for themselves and mothers signed on behalf of their children. All children provided us with verbal consent to be measured. Those grandmothers who could not sign their name were asked to provide their fingerprints as proof of consent. For ethical reasons, adult women were measured only by women of the research team. Statistical analysis Intergenerational influences of grandmothers and mothers on children’s growth status were assessed through multiple regression models. Models were adjusted for the outcome variables of z-score values of height and LL. Outcome variables were normally distributed. For each analysis, model 1 included only the maternal stature or LL (as z-score values) as predictors. In model 2 the variable paternal job loss during childhood (0 5 No, 1 5 Yes) was added. In model 3, three grandmaternal intergenerational predictors were added: (1) short stature (0 5 No, 1 5 Yes), (2) house type (for purposes of modelling used as two dummy variables: 0 5 perishable materials vs. 1 5 Maya type; 0 5 perishable materials vs. 1 5 durable materials), and (3) family size American Journal of Human Biology
496
H. AZCORRA ET AL. TABLE 1. Intergenerational effects on children z-score values for height Model 1
Maternal z-score of height Paternal job loss during maternal childhood Grandmaternal short height Grandmaternal home index (reference 5 Perishable materials (cardboard & metal) -Household Maya typea -Durable material (cement) Grandmaternal family size during childhood Constant R2 adjusted
Model 2
Model 3
B (SE)
P
B (SE)
P
B (SE)
P
0.419 (0.093)
Original Research Article
Intergenerational Influences on the Growth of Maya Children: The Effect of Living Conditions Experienced by Mothers and Maternal Grandmothers During Their Childhood ^ VARELA-SILVA2 HUGO AZCORRA,1* FEDERICO DICKINSON,1 BARRY BOGIN,2 LUIS RODRIGUEZ,3 AND MARIA INES Departamento de Ecologıa Humana, Centro de Investigaci on y de Estudios Avanzados del Instituto Politecnico Nacional (Cinvestav)-Unidad Merida, 97310 Merida, Yucat an, Mexico 2 Centre for Global Health and Human Development, School of Sports, Exercise and Health Sciences, Loughborough University, Loughborough LE11 3TU, United Kingdom 3 Facultad de Matem aticas, Universidad Aut onoma de Yucat an, Merida, Yucat an, Mexico
1
Objectives: To test the hypothesis that living conditions experienced by maternal grandmothers (F1 generation) and mothers (F2 generation) during their childhood are related to height and leg length (LL: height 2 sitting height) of their 6-to-8 year old children (F3 generation). Methods: From September 2011 to June 2012 we obtained height and LL, and calculated z-score values of these measurements for 109 triads (F1, F2, F3) who are Maya living in Merida, Yucatan, Mexico. Multiple regression models were adjusted to examine the relation of anthropometric and intergenerational socioeconomic parameters of F1 (house index and family size during childhood) and F2 (paternal job loss during childhood) with the z-score values of height and LL of F3. Results: Children’s height and LL were positively associated with maternal height and LL. This association was relatively stronger in LL. Better categories of grand-maternal house index were significantly associated with higher values of height and LL in grandchildren. Grand-maternal family size was positively related with LL, but not with height. Conclusions: Our findings partially support the hypothesis that living conditions experienced by recent maternal ancestors (F1 and F2) during their growth period influence the growth of descendants (F3). Results suggest that LL is more sensitive to intergenerational influences than is total height and that the transition from a traditional rural lifestyle to urban conditions results in new exposures for risk in human physical growth. Am. J. Hum. Biol. 27:494–500, C 2015 Wiley Periodicals, Inc. V 2015. INTRODUCTION The process of growth is a complex phenomenon strongly influenced by the social, economic and political conditions surrounding the group (Bogin and Loucky, 1997; Steckel, 2012; Stinson, 2012). It has been suggested that the current biological status of a group can also be explained by ecological conditions experienced by recent ancestors during their growth period. In 1986, Irving Emanuel defined the intergenerational influences hypothesis (IIH) as “those factors, conditions, exposures, and environments experienced by one generation that relate to the health, growth, and development of the next generation” (Emanuel, 1986, p 27). Subsequent theoretical contributions suggest that growth trajectories of children are also influenced by factors that trace the nutritional history of matrilineal ancestors, including their nutrition during prenatal stage and first years of postnatal life (Kuzawa, 2005). Most of the accumulated evidence of intergenerational influences on growth is on the prenatal stage (Alwasel et al., 2011, 2013; Emanuel, 1986, 1992, 1999, 2004; Hypponen et al., 2004; Jasienska, 2009; Klebanoff and Yip, 1987; Kuzawa, 2005; Martin et al., 2004; Ounsted et al., 1986). The intergenerational effects on postnatal growth have been less studied. Several studies provide evidence for an association between parental and offspring height. Correlation coefficients show that associations between mothers and offspring are stronger than associations between fathers and offspring in all ages (Alberman, 1991; Ghosh and Malik, 2007; Gray et al., C 2015 Wiley Periodicals, Inc. V
2012; Hypponen et al., 2004). However, parent-offspring correlations reflect both the effect of genes and shared environmental factors such as diet and life style (Towne et al., 2012). The Maya are one of the largest and most important ethnic groups in Mesoamerica (Thompson, 1973). Merida, the capital city of the state of Yucatan, in Mexico, is the home of a great number of Maya people who live in poor neighborhoods (Lizama, 2012). Historically, the Maya from Yucatan have lived under very unfavorable socioeconomic conditions (Bracamonte, 2007), and continue to do so (Azcorra et al., 2013; Varela-Silva et al., 2009, 2012). This has negatively influenced the biosocial status of the Maya. Previous studies show that Maya from Yucatan exhibit high rates of undernutrition (mainly short stature) during the growing period (Azcorra et al., 2009; Kelley, 1991; Siniarska and Wolanski, 1999; Wolanski et al., 1993; Varela-Silva et al., 2009, 2012), and in adulthood. Despite overall improvements of social and economic conditions since the end of the colonial period in Mexico (in 1821), research indicates the absence of a positive secular change in stature for the Maya people (McCollough, 1982; Siniarska and Wolanski, 1999). Contract grant sponsor: The National Council of Science and Technology of Mexico; Contract grant number: 168047. *Correspondence to: Hugo Azcorra. E-mail: [email protected] Received 17 June 2014; Revision received 18 November 2014; Accepted 10 December 2014 DOI: 10.1002/ajhb.22675 Published online 9 January 2015 in Wiley Online Library (wileyonlinelibrary.com).
495
INTERGENERATIONAL INFLUENCES ON THE GROWTH OF MAYA CHILDREN
In a previous article we described the nutritional status of a sample of 109 triads of Maya children, their mothers, and maternal grandmothers from Merida, Mexico (Azcorra et al., 2013). We confirmed the hypothesis that leg length-relative-to-stature is a more sensitive indicator of nutrition and health status than is total height or sitting height alone. In that study the analysis was based on the examination of correlations between grandmothers and children and mothers and children. In the present study we extend the analysis adding a bio-cultural model to test the IIH and ascertain the influence of the biosocial background of urban Maya grandmothers (F1 generation) and mothers (F2 generation) on the linear growth and nutritional status of their children (F3 generation). This study contributes to the body of knowledge about the intergenerational influences on postnatal growth in groups with a long adverse socioeconomic history. METHODS From September 2011 to June 2012 a cross-sectional study was undertaken on a sample of 109 triads of urban Maya children (6-to-8 year old), their mothers and their maternal grandmothers from the city of Merida. The number of triads needed for the study was obtained through a power analysis calculation. Maya surnames were used as a proxy for Maya ancestry; the requirement for inclusion in the sample was that children, mothers and grandmothers had to have at least a maternal Maya surname. Schools (n 5 20) where recruitment was conducted were randomly selected from the south of Merida, where Maya language speakers and people with the lowest level of income are concentrated. To allow for maximum geographic and socioeconomic variability in the sample, we selected no more than 10 children from each school. More details of sample size calculation and recruitment process of participants are described by Azcorra et al. (2013). Measurements We only focused, in this study, on height and sitting height of the participants. Leg length (LL) was calculated by subtracting sitting height from height. Height and LL were transformed to z-score values using the Comprehensive References published by Frisancho (2008). Particularly in the case of mothers and grandmothers, z-score values were useful to accommodate the effect of loss of height with age. We focused on the intergenerational effects on the height and LL of children because these measurements provide information about the nutritional history of individuals (Bogin and Keep, 1999; Bogin and Varela-Silva, 2010), and are related to the risk for adult mortality (Barker et al., 1990; Leon et al., 1995), and health outcomes at adulthood such as obesity, insulin resistance, type 2 diabetes, and heart diseases (Asao et al., 2006; Forsen et al., 2000; Lopez-Alvarenga et al., 2003). The biosocial background of the mothers and grandmothers was analyzed through (1) current anthropometric characteristics: height and LL (z-score values) and (2) socioeconomic indicators experienced during childhood. Children were measured at schools and adult women at homes, all of them by trained personnel. Anthropometric measurements were taken following standardized methods (Lohman et al., 1988).
Home visits were used to administer socioeconomic questionnaires to mothers and grandmothers. Data on socioeconomic conditions experienced during maternal and grand-maternal childhood were obtained through these visits. The questionnaire was based on our field experience and ethnographic knowledge of the local context. Mothers and grandmothers were interviewed separately. The assistance of the mothers was requested when grandmothers did not understand the questions because they did not speak Spanish fluently. Interviews indicated that for grandmothers there were two sensitive indicators of their childhood living conditions: (1) the materials of construction of the house, and (2) family size. Three categories of grand-maternal house type were defined: (1) perishable materials (dirt or cement floor, cardboard or metal sheets for walls and ceilings), (2) traditional Maya rural type (dirt floor, palm leaf ceilings and wattle and daub walls), and (3) durable materials (cement floor and cement blocks in walls and ceilings). Family size was recorded as the total number of people living in the home. Maternal house type was almost uniformly of durable materials, therefore it was not possible to use as a variable. Instead, our interviews revealed that for mothers, the important variable of their childhood living conditions was their father’s employment stability. The rationale for the use of this variable is that around 57% of the mothers’ parents migrated from rural areas to the city of Merida in order to get a job. But, the loss of paid employment among men was common because of the nature of the jobs and to their low qualifications. The most common occupations were masons and laborers in the construction industry. The more frequent the father’s job loss, the generally lower the childhood living conditions of the mother. Ethical concerns The research was approved by the Bioethics Committee for the Study of Human Beings of the Center for Research and Advanced Studies of the National Polytechnic Institute of Merida, Mexico and by the Loughborough University Advisory Ethical Committee. The mothers and grandmothers signed consent forms for themselves and mothers signed on behalf of their children. All children provided us with verbal consent to be measured. Those grandmothers who could not sign their name were asked to provide their fingerprints as proof of consent. For ethical reasons, adult women were measured only by women of the research team. Statistical analysis Intergenerational influences of grandmothers and mothers on children’s growth status were assessed through multiple regression models. Models were adjusted for the outcome variables of z-score values of height and LL. Outcome variables were normally distributed. For each analysis, model 1 included only the maternal stature or LL (as z-score values) as predictors. In model 2 the variable paternal job loss during childhood (0 5 No, 1 5 Yes) was added. In model 3, three grandmaternal intergenerational predictors were added: (1) short stature (0 5 No, 1 5 Yes), (2) house type (for purposes of modelling used as two dummy variables: 0 5 perishable materials vs. 1 5 Maya type; 0 5 perishable materials vs. 1 5 durable materials), and (3) family size American Journal of Human Biology
496
H. AZCORRA ET AL. TABLE 1. Intergenerational effects on children z-score values for height Model 1
Maternal z-score of height Paternal job loss during maternal childhood Grandmaternal short height Grandmaternal home index (reference 5 Perishable materials (cardboard & metal) -Household Maya typea -Durable material (cement) Grandmaternal family size during childhood Constant R2 adjusted
Model 2
Model 3
B (SE)
P
B (SE)
P
B (SE)
P
0.419 (0.093)
