Effect of the interaction between high altitude and socioeconomic factors on birth weight in a large sample from South America.код для вставкиСкачать
AMERICAN JOURNAL OF PHYSICAL ANTHROPOLOGY 129:305–310 (2006) Effect of the Interaction Between High Altitude and Socioeconomic Factors on Birth Weight in a Large Sample From South America Jorge S. López Camelo,1,2* Hebe Campaña,1 Rita Santos,1 and Fernando A. Poletta1,2 1 Latin-American Collaborative Study of Congenital Malformations (ECLAMC), Instituto Multidisciplinario de Biologia Celular (IMBICE), 1900 La Plata, Argentina 2 Latin-American Collaborative Study of Congenital Malformations (ECLAMC), Dirección de Investigación, Centro de Educación Médica e Investigaciones Clı́nica (CEMIC), 1431 Buenos Aires, Argentina KEY WORDS altitude; birth weight; interaction risk; socioeconomic level ABSTRACT Several studies of South American populations showed that altitude is associated with low birth weight and intrauterine growth retardation. Although some of them analyzed the altitude-birth weight association, only a few assessed the effect of the interaction between altitude and socioeconomic factors on birth weight. The purpose of this research is to evaluate such effects on birth weight, after adjustment for confounding factors. This observational descriptive study includes a sample of 37,022 live-born infants without congenital anomalies, weighing 500 g, from 75 South American maternity hospitals, during 1982–1999. Of the total sample, 1,187 infants were born in two South American cities located at more than 2,000-m altitude: La Paz, According to a report by UNICEF (1994), 12% of live-born infants have low birth weight (LBW, less than 2,500 g), and one third suffer from some type of brain damage. Furthermore, epidemiological studies revealed that low birth weight signiﬁcantly contributes to perinatal mortality (Morrison and Olsen, 1985; Wilcox, 1993). Ten percent of LBW infants are stillborn, while the stillbirth rate among infants weighing more than 2,500 g is 1:154. Social factors such as maternal malnutrition (Haas et al., 1980; Beall, 1981a; DelgadoRodrı́guez et al., 1998; Dutra et al., 1993) or biological factors such as low maternal age (Fraser et al., 1995; Dutra et al., 1993) are known to decrease birth weight. Several studies on South American populations showed that altitude is associated with low birth weight and reduced growth rate during childhood. Based on a review of the literature from the last 35 years, at least 70 publications reported data on the weight of infants born at high altitude (Moore, 2003). When compared with neonates born at sea level, those from higher altitudes (>2,000 m) showed a 2–3-fold increase in LBW rate, mainly related to a higher incidence of intrauterine growth retardation (IUGR) (Haas et al., 1977; Beall, 1981a; Yip, 1987; Dutra et al., 1993; Jensen and Moore, 1997; Mortola et al., 2000; Keyes et al., 2003). However, demographic differences between lowland and highland populations, such as racial admixture, income, and availability of social security, among others, interfere when correlating altitude and birth weight. Hypertension during pregnancy or preeclampsia (Arias and Tomich, 1982; Keyes et al., 2003), primiparity (Sanjose and Roman, 1991), low maternal age (Fraser et al., 1995) and social level (Wilcox et al., 1995), lack of prenatal care C 2005 V WILEY-LISS, INC. Bolivia (N ¼ 974 at 3,600 m) and Bogota, Colombia (N ¼ 274 at 2,600 m). Among the seven risk factors analyzed, altitude was the main predictor of birth weight (except for gestational age). After adjustment for the other risk factors, birth weight at cities located above 2,000 m showed a decrease of approximately 200 g. When comparing highest and lowest socioeconomic levels, birth weight also showed differences for levels of altitude analyzed (lowland, <2,000 m; middle land, 2,600 m; and highland, 3,600 m). Interaction between both factors showed no effect. High altitude seems to act independent of socioeconomic status in explaining birth weight reduction. Am J Phys Anthropol 129:305–310, 2006. C 2005 V Wiley-Liss, Inc. (Kramer, 1987), low weight gain during pregnancy (Abrams et al., 1989), and maternal smoking (Wilcox, 1993), are well-known predictors of LBW that could interact with high altitude. Few studies (Jensen and Moore, 1997; Mortola et al., 2000; Giussani et al, 2001) reported that socioeconomic factors might relatively contribute to LBW in comparison with altitude effects. However, limited access to specialized medical care and poverty are signiﬁcant factors inducing LBW and mortality rates. Therefore, it would be interesting to quantify this contribution and to evaluate the possible effect of the interaction between altitude and economic status on birth weight. The aims of this work are: 1) to assess the effect of altitude on birth weight, adjusting for other risk factors, and 2) to evaluate the joint effect of altitude and socioeconomic factors on birth weight. Grant sponsor: Agencia Nacional de Promoción Cientı́ﬁca y Tecnológica; Grant sponsor: Consejo Nacional de Investigaciones Cientı́ﬁcas y Técnicas; Grant sponsor: Comisión de Investigaciones Cientı́ﬁcas de la Provincia de Buenos Aires. *Correspondence to: Dr. J. López-Camelo, Instituto Multidisciplinario de Biologia Celular (IMBICE), Casilla de Correo 403, CP 1900 La Plata, Argentina. E-mail: firstname.lastname@example.org Received 16 June 2004; accepted 17 December 2004. DOI 10.1002/ajpa.20274 Published online 1 December 2005 in Wiley InterScience (www.interscience.wiley.com). 306 J.S. LÓPEZ CAMELO ET AL. MATERIALS AND METHODS Sample deﬁnition and size This observational descriptive study was performed with material obtained by the Estudio Colaborativo Latinoamericano de Malformaciones Congénitas (ECLAMC) (Castilla and López-Camelo, 1990) network involving 75 maternity hospitals from eight South American countries: Argentina (32), Bolivia (2), Brazil (16), Chile (9), Colombia (3), Peru (1), Uruguay (6), and Venezuela (6). During 1982–1999, 1,444,646 consecutive live- and stillborn infants weighing 500 g were examined for congenital malformations. For every malformed infant, the next liveborn nonmalformed baby of the same sex and from the same hospital was selected as control. Of these controls, 37,022 were included in the study; 1,187 were born in two South American cities located above 2,000 m altitude: La Paz, Bolivia (3,600 m, N ¼ 913) and Bogota, Colombia (2,600 m, N ¼ 274). Birth weight was registered for each infant with a gestational age of 24–43 weeks and weighing 500 g. Gestational age was calculated by last menstrual date. The hospital of birth was used as proxy for maternal residence during pregnancy. According to these data, three altitude levels were deﬁned: lowland (below 2,000 m), middle land (2,600 m), and highland (3,600 m). The following risk factors were analyzed: gestational age (date of last menstrual period, in days), sex of newborn, maternal age, parity, and Native or Black ethnic origin. The ECLAMC operational manual includes eight ethnic categories: Latin European, non-Latin European (Anglo-Saxon), Jewish, Arabian, Native, Black, and Oriental. The newborn ethnic deﬁnition depends on phenotypic characteristics, grandparents’ place of birth, and other features considered relevant by the pediatrician after questioning the mother. Native ethnicity is deﬁned as the group with highest probability of Amerindian and Latin European admixture (ancestors born in Latin America). A newborn can have up to eight types of ethnic ancestry. In this work, the Native category includes only Native ethnic ancestry, excluding any other ethnic admixture; the Black category includes Black phenotype of the baby, regardless of the ethnic background of ancestors. Socioeconomic level: deﬁnition and score Four variables were considered when assigning socioeconomic level (SEL): maternal and paternal education (eight levels), paternal occupation (eight levels), and health system of the hospital taking care of the mother and her baby. The eight levels considered for maternal and paternal education were: 1, illiterate; 2, less than 7 years of education; 3, 7 years of education; 4, less than 12 years of education; 5, 12 years of education; 6, less than 18 years of education; and 9, 18 years of education. Paternal occupational levels were: 1, housework; 2, unqualiﬁed worker; 3, independent worker; 4, boss, chief, or owner; 5, unemployed; 6, qualiﬁed worker; 7, clerk; and 8, professional or executive. Hospital health systems considered were either public or private. The eight levels of maternal and paternal education were combined into three levels with the following scores: low (score ¼ 2, less than 5 years of education), medium (score ¼ 1, from 6–8 years of education), and high (score ¼ 0, more than 10 years of education). Paternal occupational levels were also combined into three levels, with the following scores: low (score ¼ 2, including unemployed, unqualiﬁed worker, and independent worker), medium (score ¼ 1, including boss, chief, or owner), and high (score ¼ 0, including qualiﬁed worker, clerk, and professional or executive). According to the health system of the hospital, score 1 was assigned to the public system, and score 0 to a private system. Adding partial scores, a ﬁnal score was determined for each case. Socioeconomic level scores ranged from 0 (high) to 7 (low). Statistics In order to compare variables among the three altitude levels, a one-way ANOVA test (for continuous variables) and chi-square test (for categorical variables) were applied. A multiple lineal regression analysis was performed to assess the independent effects of altitude and socioeconomic level on birth weight, adjusted for other confounding factors. The dependent variable was birth weight (in grams); independent variables were altitude (three altitude levels, and two dummy variables), socioeconomic level (seven levels), Native ancestry (yes/no), Black ancestry (yes/no), sex of newborn (male/female), maternal age (in years), and gravidity (number of pregnancies). In order to assess the joint effect of altitude and ethnic origin and socioeconomic factors on birth weight, four multiplicative terms (highland * SEL, middle land * SEL, Native ancestry * SEL, and Black ancestry * SEL) were included in the regression analysis. RESULTS Altitude as a risk factor for low birth weight Table 1a shows the distribution of birth weight, gestational age, maternal age, parity, socioeconomic level, sex, and ethnic origin for the three altitude levels. Signiﬁcant differences were detected for maternal age, parity, socioeconomic level, and ethnic origin in the three altitude levels (Table 1a,b). These differences were observed among the three levels of altitude for maternal age and socioeconomic level, and between lowland and the other two altitude categories for birth weight and parity (Table 1b). Nonsigniﬁcant differences were observed for gestational age and sex. Birth weight distribution of highland and middle land populations shifted to the left when compared with the lowland population. Birth weight mean for the lowland population (3,221.9 6 556.0 g) was approximately 200 g lower than for highlanders (3,068.4 6 494.1 g) and 100 g lower than for the middle land. No differences were observed among lowland (273.8 6 18.5 days), middle land (275.1 6 18.4 days), and highland (273.6 6 19.3 days) gestational age means. After adjustment for confounding risk factors, birth weight mean differences among lowland, middle land, and highland infants were similar to nonadjusted birth weight mean differences. Effect of the variables on birth weight Table 2 shows regression coefﬁcients and levels of signiﬁcance of each risk factor on birth weight. After adjusting for the remaining risk factors, gestational age was the main birth weight predictor, followed by altitude, sex, parity, maternal age, socioeconomic level, and Native ancestry. All these variables showed a signiﬁcant regression coefﬁcient. Black ancestry did not show any effect on birth weight. 307 ALTITUDE, SOCIOECONOMIC FACTORS, AND BIRTH WEIGHT TABLE 1a. Distribution of risk factors for birthweight in lowland, middle-land, and highland populations Risk factor Lowland (N ¼ 35835) Means 6 SD Middle land (N ¼ 274) Means 6 SD Highland (N ¼ 913) Means 6 SD Birth weight Gestational age Maternal age Parity SEL Sex (male) Native ancestry Black ancestry 3221.8 6 556.0 273.8 6 18.5 25.5 6 6.4 2.8 6 2.0 3.7 6 1.9 51.3% 70.8% 16.2% 3133.7 6 511.4 275.1 6 18.4 29.8 6 5.4 2.3 6 1.4 0.8 6 1.0 51.7% 95.2% 4.4% 3068.4 6 494.1 273.6 6 19.4 24.1 6 5.9 2.2 6 1.7 4.0 6 1.5 51.6% 98.5% 0.1% Test P F F F F v2 v2 v2 ¼ ¼ ¼ ¼ ¼ ¼ ¼ ¼ 37.27 0.74 85.48 46.50 335.9 0.38 411.5 201.2 P <0.001*** 0.476, NS <0.001*** <0.001*** <0.001*** 0.825, NS <0.001*** <0.001*** TABLE 1b. A posteriori multiple comparison test (Bonferroni test)1 Risk factor Individual comparison Birth weight 1) 2) 3) 1) 2) 3) 1) 2) 3) 1) 2) 3) Maternal age Parity SEL Lowland vs. Lowland vs. Middle land Lowland vs. Lowland vs. Middle land Lowland vs. Lowland vs. Middle land Lowland vs. Lowland vs. Middle land middle land highland vs. highland middle land highland vs. highland middle land highland vs. highland middle land highland vs. highland DM P 88.1 153.4 65.2 4.33 1.39 5.73 0.47 0.60 0.14 2.91 0.33 3.25 0.026* <0.001*** 0.262, NS <0.001*** <0.001*** <0.001*** <0.001*** <0.001*** 0.985, NS <0.001*** <0.001*** <0.001*** 1 DM; difference of magnitude between means. *, P < 0.05; **, P < 0.01; ***, P < 0.001. TABLE 2. Multiple lineal regression on birth weight for risk factors Risk factor b1 t2 P3 Altitude 2,600 m Altitude 3,600 m Socioeconomic level Native ancestry Black ancestry Male sex Maternal age Parity Gestational age SEL - altitude 2600 m interaction SEL - altitude 3600 m interaction SEL - Black ancestry interaction SEL - Native ancestry interaction 163.2 187.8 11.9 61.6 34.4 123.9 3.5 20.0 12.4 34.6 4.3 3.9 3.3 4.4 1.7 23.9 6.7 12.0 88.9 1.2 <0.001*** <0.001*** 0.001*** <0.001*** 0.099, NS <0.001,*** <0.001*** <0.001*** <0.001*** 0.216, NS 12.2 1.1 0.270, NS 3.5 0.6 0.496, NS 11.8 2.9 0.003** 1 b, estimated regression coefﬁcient. t, Student’s t-test. 3 NS, not signiﬁcant. *, P < 0.05; **, P < 0.01; ***, P < 0.001. 2 The only statistically signiﬁcant interaction observed was Native ancestry-socioeconomic level. Effect of interaction between altitude and socioeconomic level on birth weight The regression coefﬁcient for altitude-socioeconomic level interaction on birth weight was not signiﬁcant (Table 2). Birth weight means were adjusted for all covariates used in the regression model. Adjusted birth weight mean, standard deviation, and sample size of each SEL for alti- tude levels analyzed are shown in Table 3. Adjusted birth weight mean differences among the three levels of altitude were similar for each socioeconomic level. The parallel lines of birth weight means by socioeconomic levels shown in Figure 1 conﬁrm the lack of interaction effect. DISCUSSION Maternal age, parity, SEL, and Native ancestry showed statistically signiﬁcant differences among populations for the altitude levels analyzed. Some of these differences, such as for Native ancestry, were expected. The others could mainly be explained by sample size differences among groups, as well as by the fact that data were obtained through a hospital-based registry. Altitude’s effect on birth weight In agreement with Zamudio et al. (1993) and DelgadoRodrı́guez et al. (1998), when compared with the lowland population, the birth weight distribution of highlanders showed a shift to the left. Highland birth weight means were approximately 200 g lower than those of the lowland population, before and after adjustment for other variables. Nonsigniﬁcant differences in gestational age were observed among lowland, middle land, and highland populations. We considered three altitude levels in order to assess the effect of lowland (above 2,000 m), middle land (2,600 m), and highland (3,600 m) on birth weight. Mortola et al. (2000) proposed that at high altitude, birth weight decreases after reaching a critical barometric pressure (BP). Those authors showed that a threshold for a hypoxic effect exists at an altitude of approximately 2,000 m, corresponding to a BP of 590 mm. Except for the obvious relationship between birth weight and gestational age, altitude was the main predic- 308 J.S. LÓPEZ CAMELO ET AL. TABLE 3. Altitude-socioeconomic level interaction1 Socioeconomic level Altitude 0 1 2 3 4 5 6 7 Lowland >2,000 m M 3,304.0 SD 182.8 n 3,018.0 3,272.1 221.6 2,516.0 3,256.5 229.5 3,471.0 3,238.4 237.8 5,565.0 3,216.8 243.0 8,030.0 3,192.9 251.5 6,800.0 3,175.3 274.9 4,134.0 3,154.3 280.7 2,300.0 Middle land 2,600 m M 3,142.8 SD 242.1 n 150.0 3,150.8 221.8 61.0 3,147.6 170.7 38.0 3,018.4 221.3 20.0 3,302.3 225.0 3.0 2,568.6 597.2 2.0 Highland 3,600 m M SD n 3,140.3 240.7 64.0 3,087.2 271.0 84.0 3,110.0 225.1 123.0 3,070.3 230.1 303.0 3,043.3 262.6 189.0 3,015.7 301.6 105.0 3,033.9 264.5 45.0 Birth weight means were adjusted for all covariates used in regression model.1 M, means; SD, standard deviation; n, sample size. in the causal pathway. Although the association is wellknown, this observation is relevant for validating the socioeconomic level indicators used in the present study. Effect of interaction between altitude and socioeconomic level on birth weight Fig. 1. Adjusted birth weight means by socioeconomic level in lowland (<2,000 m), middle land (2,600 m), and highland (3,600 m) populations. Due to small sample size of SELs 3–5 for middle land, they were included in SEL 2. tor of birth weight variations: its effect was even more signiﬁcant than that of sex or any other risk factor known for low birth weight. The effect of altitude on birth weight was reported by several authors (Haas et al., 1977; Beall, 1981a; Mortola et al., 2000; Keyes et al., 2003; Moore, 2003), and results suggest that hypoxia is by far the most important parameter responsible for low birth weight at high altitude (Ballew and Haas, 1986; Yip et al., 1988; Delgado-Rodrı́guez et al., 1998). In agreement with others (Yip, 1987; Mortola et al., 2000; Giussani et al., 2001), no gestational age differences were observed, suggesting that altitude only leads to IUGR and not to prematurity. Reduced fetal O2 supply (maternal, placental, or fetal causes) was associated with IUGR. Although the mechanisms by which hypoxia retards fetal growth are unclear, recent studies were carried out regarding two possible contributing factors: an increased incidence of preeclampsia (Keyes et al., 2003), and reduced maternal glucose concentrations (Krampl et al., 2001). Socioeconomic level’s effect on birth weight A signiﬁcant association was observed between low SEL and reduced birth weight at any of the altitude levels studied. Low SEL is obviously associated with low birth weight, and factors such as maternal nutrition (Frisancho et al., 1977; Haas et al., 1980; Beall, 1981b) and maternal smoking (Wilcox, 1993), among others, could be involved Although birth weight seems to depend on the availability of both oxygen and glucose (Krampl et al., 2001; Keyes et al., 2003), the extent to which altitude affects birth weight, independent of other risk factors, is still unknown. Maternal plasma glucose fasting that is lower at high altitude than at sea level in the presence of similar insulin secretion (Krampl et al., 2001) may partly explain lower birth weight at high altitude. Since transplacental transfer of glucose is directly proportional to maternal blood glucose (Economides and Nicolaides, 1989), decreased maternal nutrition could also affect the availability of glucose for the fetus. At this point, it would be interesting to determine whether altitude and decreased maternal nutrition affect birth weight through the same biological mechanism or if they act through different mechanisms. A possible interaction effect would support the hypothesis that the actions unchained by both factors converge in a purely biological mechanism (e.g., low availability of glucose for the fetus). Haas et al. (1980) suggested that fetal growth retardation at high altitude correlates with the duration of highaltitude residence, independent of maternal nutrition. Jensen and Moore (1997) analyzed the effect of the interaction between altitude and 17 risk factors on birth weight. Only two of these factors (nulliparity and premature rupture of membranes) were associated with reduced birth weight at high altitude, but the signiﬁcance of their effect was low. Those authors concluded that high altitude acts independently of other risk factors in reducing birth weight. A study carried out in a relatively large population of North American children of low socioeconomic status (Yip et al., 1988) showed a dose-response relationship between reduction of height/age and weight/height scores at increasing altitude. Those authors proposed that two components seem to contribute to growth restriction when related to altitude: an indirect effect mediated by birth weight reducing the effect of altitude, and a direct effect of altitude, independent of birth weight. In a study performed at lowland and highland hospitals on patients with low and high incomes from two Bolivian cities, Giussani ALTITUDE, SOCIOECONOMIC FACTORS, AND BIRTH WEIGHT et al. (2001) did not ﬁnd evidence of an altitude-socioeconomic level interaction. Those authors suggested that regardless of maternal economic status, high altitude is associated with low weight and body shape alteration at birth. Our results from a large sample of South American populations agree with these observations. No evidence for altitude and socioeconomic level interaction was observed, which suggests that both factors could be acting independently of birth weight. Strengths This large South American sample (37,122 births) covers a wide range of altitude levels, from sea level up to the highest urban settlements in the world, such as La Paz, Bolivia (3,600 m, 913 births) and Bogota, Colombia (2,600 m, 274 births). Low maternal/paternal education, low paternal occupational level, and health system of the hospital were used as proxies for low socioeconomic level, and the indicator applied was able to detect birth weight differences among socioeconomic levels, independent of altitude. Therefore, socioeconomic level indicators showed acute sensibility for the indirect measurement of maternal economic and nutrition status. Limitations Although highland and lowland populations differed in several demographic characteristics, socioeconomic levels as well as the other variables included in this surrogate were registered in a similar way. According to the ECLAMC operational manual, ancestry refers to the ethnic admixture in the child’s ancestors and not to the child’s race. This concept applies to Latin American populations in which the wide variety of racial admixtures makes it difﬁcult to establish a newborn’s race. However, newborn ancestry can be determined with reasonable reliability. South American racial admixture is di-hybrid, basically formed by Natives and Caucasians in countries such as Argentina, Chile, Bolivia, and Peru, and tri-hybrid, formed by Natives, Caucasians, and Blacks as in Uruguay, Brazil, Ecuador, Colombia, and Venezuela (Carlson, 1937; Parra et al., 1998; Vieira et al., 2002). Selection bias may occur if migration exists between highland and lowland populations. Using data from the Maryland Birth Defects Reporting and Information System, Khoury et al. (1988) evaluated the potential bias introduced by migration. Twenty percent of 295 cases with one or more of the 12 birth defects analyzed reported residence changes during gestation, and 7.5% had migrated from their original county. When place of residence at moment of birth of the child is taken as a proxy for residence at conception, when the putative exposure might have occurred, the classiﬁcation error increases with the mobility rate and frequency of exposure. Thus, this error tends to minimize the association expected. Although not evaluated in the present study, migration values between lowland and highland are probably lower than in Khoury et al. (1988), and therefore a low bias is expected. Healthcare, diet, or maternal smoking were not evaluated. Although weight for gestational age was not measured directly, nonsigniﬁcant differences in gestational age were observed between lowland and highland populations. Therefore, differences in birth weight cannot be attributed to differences in gestational age, and we can conclude that 309 LBW at high altitude was the result of IUGR and not prematurity. CONCLUSIONS This study, performed on a large sample from South American populations, is in agreement with results published in literature, in the sense that high altitude seems to act independently of socioeconomic status in explaining birth weight reduction. LITERATURE CITED Abrams B, Newman V, Key T, Parker J. 1989. Maternal weight gain and preterm infants. Obstet Gynecol 74:577–583. Arias F, Tomich P. 1982. Etiology and outcome of low weight and preterm infants. Obstet Gynecol 60:270–281. Ballew C, Haas JD. 1986. Hematologic evidence of fetal hypoxia among newborn infants at high altitude in Bolivia. Am J Obstet Gynecol 155:166–169. Beall CM. 1981a. Optimal birthweights in Peruvian populations at high and low altitudes. Am J Phys Anthropol 56:209–216. Beall CM. 1981b. Growth in a population of Tibetan origin at high altitude. Ann Hum Biol 8:31–38. Carlson FA. 1937. Geography of Latin America. New York: Prentice Hall, Inc. Castilla EE, López-Camelo JS. 1990. The surveillance of birth defects in South America: I—the search for time clusters: epidemics. Adv Mutagen Res 2:191–210. Delgado-Rodrı́guez M, Pérez-Iglesias R, Gómez-Olmedo M, Bueno-Cavanillas A, Gálvez-Vargas R. 1998. Risk factors for low birth weight: results from a case-control study in southern Spain. Am J Phys Anthropol 105:419–424. Dutra MG, López-Camelo JS, Castilla EE. 1993. Prenatal growth in South-American populations I: a study of major determinants. Acta Med Auxol 25:97–105. Economides DL, Nicolaides KH. 1989. Blood glucose and oxygen tension levels in small-for-gestational-age fetuses. Am J Obstet Gynecol 160:385–389. Fraser AM, Brockert JE, Ward RH. 1995. Association of young maternal age with adverse reproductive outcomes. N Engl J Med 332:1113–1117. Frisancho AR, Klayman JE, Matos J. 1977. Inﬂuence of maternal nutritional status on prenatal growth of a Peruvian urban population. Am J Phys Anthropol 46:265–274. Giussani DA, Phillips S, Anstee S, Barker DJP. 2001. Effects of altitude versus economic status on birth weight and body shape at birth. Pediatr Res 49:490–494. Haas JD, Baker PT, Hunt EE. 1977. The effects of high altitude on body size and composition of the newborn infant in southern Peru. Hum Biol 49:611–628. Haas JD, Frongillo EA Jr, Stepick CD, Beard JL, Hurtado L. 1980. Altitude, ethnic and sex differences in birth weight and length in Bolivia. Hum Biol 52:459–477. Jensen GM, Moore LG. 1997. The effect of high altitude and other risk factors on birth weight: independent or interactive effects? Am J Public Health 87:1003–1007. Keyes LE, Armaza JF, Niermeyer S, Vargas E, Young DA, Moore LG. 2003. Intrauterine growth restriction, preeclampsia, and intrauterine mortality at high altitude in Bolivia. Pediatr Res 54:20–25. Khoury MS, Stewart W, Weinstein A, Panny S, Lindsay P, Eisenberg M. 1988. Residential mobility during pregnancy: implications for environmental teratogenesis. J Clin Epidemiol 41:15–20. Kramer M. 1987. Determinants of low birth weight: methodological assessment and meta-analysis. Bull WHO 65:663–757. Krampl E, Kametas AN, Nowotny P, Roden M, Nicolaides KH. 2001. Glucose metabolism in pregnancy at high altitude. Diabetes Care 24:817–822. Moore LG. 2003. Fetal growth restriction and maternal oxygen transport during high altitude pregnancy. High Alt Med Biol 4:141–156. 310 J.S. LÓPEZ CAMELO ET AL. Morrison MB, Olsen J. 1985. Weight-speciﬁc stillbirths and associated causes of death: an analysis of 765 stillbirths. Am J Obstet Gynecol 152:975. Mortola JP, Frappell PB, Agüero L, Armstrong K. 2000. Birth weight and altitude: a study in Peruvian communities. J Pediatr 136:324–329. Parra EJ, Marcini A, Akey J, Martinson J, Batzer MA, Cooper R, Forrester T, Allison DB, Deka R, Ferrell RE, Shriver MD. 1998. Estimating African admixture proportions by use of population-speciﬁc alleles. Am J Hum Genet 63:1839–1851. Sanjose S, Roman B. 1991 Low birthweight, preterm, and small for gestational age babies in Scotland, 1981–1984. J Epidemiol Community Health 45:207–210. UNICEF. 1994. Estado mundial de la infancia: Foro de las Naciones Unidas para la Infancia. Barcelona: UNICEF. Vieira AR, Karras JC, Orioli IM, Castilla EE, Murray JC. 2002. Genetic origins in a South American clefting population. Clin Genet 62:458–463. Wilcox AJ. 1993. Birth weight and perinatal mortality: the effect of maternal smoking. Am J Epidemiol 137:1098–1104. Wilcox MA, Smith SJ, Johnson IR, Maynard PV, Chilvers CED. 1995. The effect of social deprivation on birthweight, excluding physiological and phatological effects. Br J Obstet Gynaecol 79:571–574. Yip R. 1987. Altitude and birth weight. J Pediatr 111:869–876. Yip R, Binkin NJ, Trowbridge LF. 1988. Altitude and childhood growth. J Pediatr 113:486–489. Zamudio S, Droma T, Norkyel K, Acharya G, Zamudio J, Niermeyer S, Moore L. 1993. Protection from intrauterine growth retardation in Tibetans at high altitude. Am J Phys Anthropol 91:215–224.