Twin births in Yucatan, Mexico during 2008–2020: trends in maternal sociodemographic factors and differences in birth weight and length according to sex of co-twin

Abstract

Background: Over the years, there has been a noticeable increase in the incidence of multiple pregnancies, due in part to the increased use of assisted reproductive technologies in recent years. The increase in twin pregnancies constitutes a public health challenge due to the adverse outcomes sometimes they entail. Twin pregnancies inherently carry a higher risk of complications, and one of the critical associated factors is the risk of low birth weight. Twin birth weight discordance among different populations can be attributed to differences in non-shared environmental influences. The presence of two or more fetuses in the uterus may lead to an unequal distribution of nutritional and oxygen resources, increasing the likelihood that at least one of the twins will experience insufficient fetal development. Other factors, such as ethnicity, genetics, sociodemographic characteristics, gestational age, parity, and chorion type, have also been related to the birth weight discordance in twin pregnancies. However, it is unclear to what extent the associations between these factors can explain the differences in birth and length size. The frequency of twin births varies among populations and over time, so it is crucial to gain a deeper understanding of the factors contributing to the rise in the twinning rate. The official records of twins continue to make significant contributions to our understanding of the causes of individual differences, and new twin registries are still being created. The recent availability of data from developing countries allows the analysis of trends in regions with sociodemographic and reproductive profiles. Obtaining a more comprehensive understanding of the epidemiology, as well as the related morbidity and mortality, is clinically crucial. Objectives: The aim of the study was to describe the trends of twin births in Yucatan, Mexico during 2008–2021, analyze their association with maternal sociodemographic factors, and compare birth outcomes between types of twin pairs: female-female, male-male, and female-male. Methods: A retrospective cohort analysis was conducted using data from the Mexican Ministry of Health, encompassing information on all births registered in the State of Yucatán, Mexico, from 2008 through 2021. The data was obtained from both public and private hospitals. The variables, including date of birth, sex, gestational age, birth weight and length of newborns, mother´s date of birth, educational level, and number of previous live offspring, were extracted from each dataset. Multiple births (three or more) were excluded from the study. We graphically analyzed the rates (per 1000 births) and percentages of twins according to maternal age, education level, and parity during the study period. A multiple logistic regression model was used to analyze the association between maternal sociodemographic factors and the occurrence of twin births. Comparisons of the gestational age and birth weight and length between types of twin pairs were performed using analysis of variance. Results: A total of 478,118 live births, including 1.4% twins (accounting 6,766 twin births), were analyzed. The rates increased from 11.21 during 2008–2011 to 13.34 during 2012–2017 and reached 20.08 in 2019. The percentages increased in women aged ≥30 years and those with higher educational levels. Older maternal age (coefficient = 0.03; OR = 1.03, per each year), greater education level (coefficient = 0.55 and OR= 1.74 for medium and coefficient = 1.05; OR = 2.57 for high level, compared with no education) and higher parity (coefficient= 0.26; OR = 1.30 per each previous offspring) increased the odds for having twins. Twins’ male-male showed a slightly increased of preterm birth than a co-twin female. Opposite-sex twins showed measurable but small increases in birth weight and length compared with same-sex twins. Conclusion: The rate of twins in Yucatan increased substantially during 2008–2020 in specific sociodemographic groups. Opposite-sex twins were slightly larger than same-sex twins at birth.

Keywords:

• Birth weight
• multiple birth offspring
• sociodemographic factors
• maternal age
• developing countries

Introduction

The study of twins provides relevant insights into reproductive biology, maternal-fetal and neonatal medicine, genetics, and certain aspects of dynamic demography. From an epidemiological perspective, twins show greater morbidity and mortality at birth than newborns resulting from singleton pregnancies [1]. Twin pregnancies are a significant risk factor for preterm birth, low birth weight, and other consequences [2–4]. Studies have shown that variability in the size of twins at birth is influenced by the sex of the co-twin owing to the complex processes occurring during gestation [5–7].

Birth weight depend on environmental and genetic/epigenetic factors [8,9]. Twin pregnancies are associated with a lower birth weight and size than singleton pregnancies. Both low birth weight and small birth size are associated with an increased likelihood of developing serious diseases in adulthood.

Twin rates have increased globally over the last 3–4 decades. A study that analyzed the distribution of and changes in twin birth rates in 112 countries between 1980–2015 found that 74(66%) of the countries showed an increase in the number of twin births of approximately 10% [10]. The rise in twinning rates may be related to delayed childbearing age and the increased use of infertility treatments, including assisted reproductive technologies (ART) [11]. Socioeconomic differences in access to ART, along with the change in the age of women giving birth over the last few decades, may indicate that the increase in twin births does not occur homogeneously in the sociodemographic groups of the population. Studies have shown that higher levels of maternal education and income, and better occupational positions are associated with higher rates of twinning [12–14].

Until recently, data on twin births were available only for European countries and the United States. The recent availability of data from low- and middle-income countries has made it possible to analyze trends in regions with demographic, sociocultural, and reproductive profiles. This study was conducted in the state of Yucatan in southeastern Mexico. In 2020, Yucatan had 2,321,000 inhabitants and had the seventh highest population growth rate in the country (7/32) [15] because of an interstate migration process that began at the end of the twentieth century. Yucatan has experienced a significant reduction in its birth rate, a phenomenon that was exacerbated in 2019–2020 during the coronavirus disease pandemic [16]. The reduction in birth rate has been more pronounced in women between 20 and 29 years of age, with a low educational level, and no previous children. However, twin births have been poorly studied not only in Yucatan, but also in the entire country, even in the context of increasing availability and access to assisted reproductive techniques.

This study has three aims: (1) to describe the trends of twin births in Yucatan during 2008–2021, including whether potential changes in the number of twin births vary according to maternal age, education level, and the number of previous offspring; (2) to analyze the association between maternal sociodemographic factors and the occurrence of twin births; and (3) to compare birth outcomes between types of twin pairs: female-female, male-male, and female-male.

Methods

The data used in this study were obtained from annual datasets produced by the National Birth Information System (SINAC) of the Mexican Ministry of Health. The databases include information on pregnancy, newborn, and sociodemographic characteristics of the mother, all obtained in public and private hospitals by trained health personnel using standardized techniques. The data came from births registered in approximately 30 public and private hospitals, and a minimal proportion of births (<1%) occurred in households with the help of traditional midwives. All twin births included in this study occurred in hospitals. For this study, we used datasets corresponding to the births of singletons and twins registered in Yucatan between January 1, 2008–December 31, 2021. While the number of multiple products (three or more) showed slight increases during the period, given the small number of cases (n = 50), multiple products were excluded from the study. The final dataset comprised 478,118 live births (98.6% singletons and 1.4% twins). Over 97% of the mothers included in this study resided in Yucatan; the remaining percentage corresponded to women from neighboring states, and their inclusion in the analyses did not modify the trends.

We obtained date of birth, sex, gestational age, birth weight and length of newborns, mother’s date of birth, education level, and number of previous live offspring. We calculated the maternal age in years at the date of childbirth and grouped the mothers into the following age groups: 1:<20, 2:20–29, 3:30–39, and 4:>40 years; both the continuous years variable and the categorical age groups were used in the analyses. We grouped mothers into the following school categories: 1.Below Elementary; 2.Low: elementary, or junior high school; 3. Medium: high school; and 4.High: technical career/university. These categories correspond to the levels of the classification commonly used in Mexico. The number of previous live offspring variables were used in the analyses both as a numerical and categorical (1.None, 2.One; 3.Two, 4.Three or more). Gestational age was calculated in weeks according to the date of the last menstrual bleeding. Birth weight(g) and length(cm) were measured during the first hour after birth following mandatory protocols for health personnel (NOM-008-SSA2-1993, 1994). We calculated the percentages of preterm births (<37 weeks of gestation) and infants with low birth weight (<2500 g).

Statistical analysis

First, the percentages and annual rates of twin births (per 1000 births) were calculated to analyze the trends of these events during the study period. We then calculated and graphically analyzed the percentage of twin births according to maternal age, education, and number of previous live births categories to determine whether the potential changes in the number of twins differed according to sociodemographic characteristics. Second, we adjusted the multiple logistic regression model (reporting regression coefficients and odds ratios [OR]) to analyze the association between maternal sociodemographic factors and the occurrence of twin births (0:singletons vs. 1:twins). Maternal age (years) and the number of previous births were introduced into the model as numerical variables, and education level as a categorical variable. We tested interactions between predictors using the variables in the scale in which they were entered into the model. Third, we analyzed the differences in gestational age, birth weight and length, and percentages of preterm births and infants with low birth weight according to the type of twin pair: female-female(F-F), male-male(M-M), and female-male(F-M). We applied one-way analysis of variance for gestational age and two-way analysis of variance for birth weight, including gestational age as a covariate in the model. Bonferroni adjustments for multiple comparisons were used when analysis of variance was statistically significant. Chi-square tests were used to analyze differences in the proportions of preterm births and low-birth weight infants. Two-tailed tests were considered statistically significant at p< 0.05.

Ethical considerations

Retrospective, secondary, anonymized datasets were used under permission from the Ministry of Health of the State of Yucatan for the present study. The health authorities exempted the study from ethical review.

Results

During 2008–2021, a total of 6,766 (14.15 per 1000 live births) twin births were recorded, of which 39% were F-F, 34% were M-M, and 27% were female-male. The rates remained stable during 2008–2011, with an average of 11.21 cases (Figure 1). During 2012–2017, average rate was 13.34 cases, that is, an increase of 20% from the previous period. In 2019, the rate reached 20.08 cases, which was ∼50% and ∼80% higher than the rate in 2012–2017 and 2008–2011, respectively. In 2020, the rate decreased to 17.68 cases, but it increased to 18.78 during 2021.

Figure 1. Trend of annual rates of twin births in Yucatan, Mexico, during 2008–2021.

On average, mothers of twins were 2.6 years older than mothers of singletons; the proportions of women aged 30–39 and ≥40 years were greater in mothers of twins, and the proportions of women younger than 20 years and between 20 and 29 years were greater in mothers of singletons (Table 1). The proportion of women with a high level of education was greater among mothers of twins and the mothers of twins had one previous live birth more than the mothers of singletons. The percentage of twin births in women aged 20–29 years decreased throughout the study period, and the percentages in women aged 30–39 and ≥40 years increased from 2016 onwards (Figure 2). There were substantial decreases in the percentage of twin births in women with none and low level of education from 2012 and 2018, respectively, and there were notable increases in women with medium and high educational levels throughout the period studied (Figure 3). There was no significant change in the percentage of twin births according to the number of previous offspring during the study period (Figure 4).

Figure 2. Trends in the percentages of twin births occurring in Yucatan by categories of maternal age. Sample sizes: <20: 648, 20–29: 3,236, 30–39: 2,516, ≥40: 352.

Figure 3. Trends in the percentages of twin births occurring in Yucatan by categories of maternal education. Sample sizes: None: 612, Low: 3,023, Medium: 2,286, High: 312.

Figure 4. Trends in the percentages of twin births occurring in Yucatan by categories of maternal parity. Sample sizes: None: 1,153, One: 2,197, Two: 1,705, Three or more: 1,699.

Table 1. Differences in sociodemographic characteristics between mothers of singletons and twins born in Yucatan, Mexico during 2008–2020.

Variable	Singletons (n = 471,352)	Twins (n = 6,766)	Statistic	p Value
Maternal age (years)	25.71 (SD = 6.24)	28.29 (SD = 6.74)	t= −33.70	p < 0.001
Maternal age categories
<20	82,747 (18%)	648 (10%)	X^2(3)=110	p < 0.001
20–29	258,491 (55%)	3,236 (48%)
30–39	120,206 (26%)	2,516 (37%)
>40	8,680 (2%)	352 (5%)
Education
None	39,410 (9%)	525 (8%)	X^2(3)=455	p < 0.001
Low	254,296 (55%)	3,231 (49%)
Medium	96,231 (21%)	1,247 (19%)
High	70,159 (15%)	1,645 (25%)
Number of previous births	Median= 2 (IR = 1–3)	Median= 3 (IR= 2–4)	Z=-41.52	p < 0.001
Number of previous births categories
None	177,818 (38%)	1,153 (17%)	X^2(3)=180	p < 0.001
One	148,030 (31%)	2,197 (33%)
Two	85,191 (18%)	1,705 (25%)
Three or more	59,211 (13)	1,699 (25%)

T-Student test was applied to analyze differences in maternal age (years) between mothers of singletons and twins.

Chi-square was used to analyze differences in the proportions of categories of maternal age, education, and number of previous offspring.

Wilcoxon rank sum test was applied to compare medians of number of previous offspring between mothers of singletons and twins.

IR = interquartile range.

The multiple logistic regression models showed that maternal age, educational level, and parity were associated with the occurrence of twin births (Table 2). Specifically, the risk of twin births increased 3% (coefficient = 0.03[CI95%:0.02–0.03], OR = 1.03[CI95%:1.02–1.04]) for each year increase in maternal age and 30% (coefficient = 0.26[CI95%:0.24–0.27], OR = 1.30[CI95%:1.28–1.32] for each previous offspring. Low (coefficient = 0.41[CI 95% 0.31, 0.51], OR = 1.51[CI95%:1.37–1.67], medium (coefficient = 0.55[CI95%:0.44:0.66], OR = 1.74[CI95%:1.56–1.94] and high (coefficient = 1.05[CI95%:0.94–1.16], OR = 2.87[CI95%:2.57–3.20] levels of education increased the odds of having twins compared with those cases with none level of education.

Table 2. Logistic regression model of maternal sociodemographic factors for twin births in Yucatan, Mexico.

Variable	Coefficient (CI 95%)	OR (CI 95%)	SE	p Value
Main effects
Maternal age (years)	0.03 (0.02, 0.03)	1.03 (1.02, 1.04)	0.01	<0.001
Maternal schooling
None (reference)
Low: primary/secondary	0.41 (0.31, 0.51)	1.51 (1.37, 1.67)	0.07	<0.001
Medium: high school	0.55 (0.44, 0.66)	1.74 (1.56, 1.94)	0.09	<0.001
High: technical career/university	1.05 (0.94, 1.16)	2.87 (2.57, 3.20)	0.16	<0.001
Number of previous offspring	0.26 (0.24, 0.27)	1.30 (1.28, 1.32)	0.01	<0.001
Interaction terms
Age (numerical) X Schooling (categories)
None (reference)	–	–	–	–
Low: primary/secondary	0.02 (0.01, 0.03)	1.02 (1.01, 1.03)	0.006	0.005
Medium: high school	0.05 (0.03, 0.06)	1.05 (1.03, 1.07)	0.008	<0.001
High: technical career/university	0.10 (0.08, 0.12)	1.10 (1.08, 1.12)	0.008	<0.001
Age (numerical) X Number of previous offspring (numerical)	−0.02 (-0.03, −0.01)	0.98 (0.97, 0.98)	0.001	<0.001
Number of previous offspring (numerical) X Schooling (categories)
None (reference)	–	–	–	–
Low: primary/secondary	0.11 (0.08, 0.15)	1.12 (1.08, 1.17)	0.02	<0.001
Medium: high school	0.28 (0.22, 0.33)	1.32 (1.25, 1.40)	0.03	<0.001
High: technical career/university	0.22 (0.17, 0.28)	1.25 (1.18, 1.31)	0.03	<0.001

Analytical sample size of the model: 464,583.

All interactions between predictors were significant. The effect of maternal age on the risk for twin births increased as the level of maternal education was higher (low level: OR = 1.02[CI95%:1.01–1.03]; medium level: OR = 1.05[CI95%:1.03–1.07]; high level: OR = 1.10[CI95%:1.08–1.12]). The effect of maternal age on the risk of twin births decreased as the number of previous offspring increased (OR = 0.98[CI95% 0.97–0.98]. Finally, the effect of number of previous offspring on the risk for twin births was greater in higher levels of maternal education (low level: OR = 1.12[CI 95%:1.08–1.17], medium level: OR = 1.32[CI95%:1.25–1.40] and high level: OR = 1.25[CI95%:1.18–1.31].

The mean birth weight and length of the twins in the overall sample were 2,176 g(SD = 531) and 44.8 cm(SD = 4.1), respectively. Opposite-sex twins had a greater gestational age and birth weight and length than same-sex twins (Table 3). Multiple comparisons tests showed that the gestational age in F-M twins was 0.3 weeks longer than that in same-sex twins, and they were 109 and 115 g heavier and 0.8 and 1 cm larger than M-M and F-F twins, respectively. No significant differences in birth outcomes were found between M-M and F-F twins. The mean birth weight and length of males in opposite-sex twins were greater than those of male twins in M-M pairs (2300 g[SD = 499] vs. 2141 g[SD = 542], p < 0.001; 45.7 cm[SD = 4.5] vs. 44.6 cm[SD = 4.2], p < 0.001) and females in F-M pairs showed the same trend as females in F-F pairs (2206 g[SD = 489] vs. 2135 g[SD = 520], p < 0.001; 45.1 cm [SD = 3.6] vs. 44.4 cm[SD = 4.2], p < 0.001) (Figure 5). M-M twins showed a slightly higher prevalence of preterm births than F-F and F-M twins, while opposite-sex twins had the lowest prevalence of low birth weight (Figure 6).

Figure 5. Means of birthweight of twins born in Yucatan during 2008–2020 according to sex of twin pairs. Male-Male twins n = 2,274, Female-Female twins n = 2,572, Female-Male twins n = 1,758.

Figure 6. Means of length at birth of twins born in Yucatan during 2008–2020 according to sex of twin pairs. Male-Male twins n = 2,274, Female-Female twins n = 2,572, Female-Male twins n = 1,758.

Table 3. Differences in birth outcomes according to sex of twin pairs.

Variable	Male-Male twins (n = 2,274)	Female-Female twins (n = 2,572)	Female-Male twins (n = 1,758)	Differences
Gestational age (weeks)	35.6 (SD = 2.6)	35.7 (SD = 2.5)	36.0 (SD = 2.5)	F = 7.78, p < 0.001
Prematurity (<37 weeks)	60%	57%	55%	X^2(2)=6.02, p < 0.05
Birthweight (g)	2141 (SD = 542)	2135 (SD = 520)	2250 (SD = 492)	F = 27.83, p < 0.001
Low birthweight (<2500 g)	74%	75%	66%	X^2(2)=43.29, p < 0.001
Length (cm)	44.6 (SD = 4.1)	44.4 (SD = 4.2)	45.4 (SD = 3.6)	F = 39.02, p < 0.001

One-way ANOVA was applied to analyze differences in gestational age.

Two-way ANOVAs were applied to analyze differences in birthweight and length, and models were adjusted for gestational age.

Chi-square was used to analyze differences in the proportions of prematurity and low birthweight.

Discussion

The study of twin pregnancies is of great medical interest because of the higher frequency of maternal, fetal, and perinatal complications compared to that in singleton pregnancies [17].

The rate of twin births in Yucatan in 2019 (20.08 per 1000 live births) was higher than that in high-income countries such as Canada (15.5), Australia (14.7), France (16.1), Spain (17.9), Germany (18.4), Denmark (15.3), and the United States (16.3) in the same year (Human Multiple Births Database 2023, French Institute for Demographic Studies), but lower than that in African countries such as Ghana, Senegal, Togo, and Tunes, where the rates have been historically high. Studies on trends in twin birth rates in Mexico are scarce. A global study by Monden et al. [10] showed that the rates in Mexico increased from 9.5 cases in 1980–1985 to 11.4 cases in 2010–2015, representing an increase of 20% over three decades. The average rate in Yucatan during 2010–2015 was 12.3, which indicates that recent rates in Yucatan were than the national reference.

The increase in twin births in Yucatan may be influenced, to some extent, by the increased use of ART. ART increases the odds of twin births depending on the type of treatment, up to 24% of successful in vitro fertilization procedures result in twin gestation [18]. However, data on the number of births resulting from these procedures are not available in Mexico, and legal regulation is practically nonexistent. Women on ART tend to be older, more educated, and less likely to be nulliparous [19]. During the period analyzed, an increase in the frequency of twin pregnancies was noted, highlighting an upward trend, especially in women >40 years of age and with higher levels of education.

We found interactive effects of maternal age with level of maternal education, which suggest the postponement of motherhood as a potential reason for increased twining rates in Yucatan, as has been proposed in populations from high income nations [20]. Studies have shown that advanced maternal age (>35 years) increase the occurrence of twin conception and births [20]. In a study of birth weight in twins comparing discordant with non-discordant newborns, it was found that the discordant growth group comprised a higher proportion of mothers aged > 30 years (monochorionic twin pregnancy: p = 0.044; dichorionic twin pregnancy: p = 0.027). In that case, monochorionic twins were proportionally more frequent among women aged >30 years when compared to women below 30; findings revealed serum placental growth factor levels were lower in the discordant group, therefore, it is plausible that age related reduction of serum placental growth factor levels influences discordant twin birthweight, affecting mainly women at older age [21]. The natural increase in follicle-stimulating hormone levels in the follicular phase, may increase the risk of twinning through multiple follicle growth and subsequent multiple ovulations in hereditary and non-hereditary twinning [22,23]. As suggested by Zhang and colleagues in a study with 5082 participants on frozen-thawed embryo transfer, women with normal menstrual cycles exhibited twin pregnancies in 12.94% among women with hMG-induced cycle rate, expecting high ovarian response [24].

The overall birth rate in Yucatan has been decreasing since 2008 (14.6 births per 1000 inhabitants in 2008 to 11.48 births per 1000 inhabitants in 2021 [16]; win pregnancy is associated with a lower birth weight and size than singleton pregnancy. Both low birth weight and small birth size are associated with an increased likelihood of developing serious diseases in adulthood, such as cardiovascular disease, obesity, and type 2 diabetes. In addition, in M-H twin pregnancies, males were larger and had higher birth weights, which is consistent with previously reported findings [25].

In the present study, the odds of twining increased as the level of maternal education increased, in a population-level study that analyzed 17 million births from 1972 to 2013 (2.7% twins) in 72 developed and developing countries found that taller, not underweight, not having smoked before pregnancy, and having no morbidities prior to conception, including obesity, diabetes, hypertension, asthma, and kidney disease, were more likely to give birth to twins [26]; therefore, a link between education and healthier pre-pregnancy habits may at least partially explain this phenomenon.

Our analysis showed that opposite-sex twins were heavier and longer than same-sex twins. Specifically, we found that males in opposite-sex twins were 159 g heavier and 1.1 cm longer than males in twins of the same sex. Meanwhile, females who had a male co-twin were 71 g heavier and 0.7 cm longer than females in twins of the same sex. The magnitude of these differences seems to be small, as they represent a decrease in birth weight of approximately 7% in males and 3.5% in females in relation to the average birth weight of twins in this sample (2176 g). However, the trends found in our study are like those of an earlier study conducted in Yucatan [27] and to those shown in several studies with larger samples [5–7]. For example, using data from 191,838 twin births recorded in 214 nationally representative household surveys in 72 countries between 1990 and 2016, Alemu et al. [7] found that having a male co-twin was associated with lower birth weight for boys and higher birth weight for girls. Boys with a female co-twin were 70 g heavier than boys with a male co-twin, a difference smaller than that found in our study.

It has been proposed that the greater birth size of female twins in opposite-sex pairs than that of females in same-sex pairs could be due to testosterone transfer from a male to a female co-twin during prenatal development, that is, the so-called twin testosterone transfer hypothesis. Possibly, the larger birth size of males in opposite-sex twins than that of males in same-sex twins could be because males in M-F twin pairs are more successful in the competition for resources in the presence of a female co-twin [28].

Studies on monozygotic twins have suggested differences in DNA methylation patterns in the promoter and/or coding regions of genes involved in cellular metabolism and biosynthesis, thus regulating development. In twins with birth weight differences, variations in methylation patterns are found in regions of chromosome 1 where TYW3 and CRYZ are located [29] and participate in metabolism. Other described mechanisms that could be related to the difference in birth weight include variations in the promoter region of the glucokinase gene, which encodes a key enzyme in glucose metabolism, and variations in mitochondrial DNA, which contains genes essential for cellular energy production [30]. To determine the genetic and epigenetic modifications involved in the decrease in birth weight and height in women with m-h twin pregnancies, more detailed research is needed at the genetic and molecular levels.

Limitations

One of the most evident limitations in our study is the retrospective nature that would not allow us to truly identify temporal relation of studied outcomes, a second important limitation is the absence of information on the monochoriocity or dichorionicity of fetuses, given the fact that discordant weight frequency may vary. Another important limitation is that we can only speculate in feasible aspects regarding age related discordant twin weight incidence, as we cannot identify specific conditions that may have influenced the late pregnancy.

Conclusion

In conclusion, the rate of twin births in Yucatan, Mexico, increased substantially during 2008–2020. The largest increase in the rate of twins was observed in women over 30 years of age and in those with a high level of education. Opposite-sex twins had a greater birth size than same-sex twins. Differences in birth weight and length were greater between males in opposite-sex twins and males in same-sex twins than between females in opposite-sex twins and females in same-sex twins.

Acknowledgements

The authors thank the Subsystem of Births Information (SINAC) and the Ministry of Health of the State of Yucatan for providing us with access to the dataset.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability

http://www.dgis.salud.gob.mx/contenidos/basesdedatos/da_nacimientos_gobmx.html.

Additional information

Funding

The author(s) reported there is no funding associated with the work featured in this article.