Early menarche and its association with anthropometric and body composition variables in young university students

Este trabajo fue recibido el 28 demayo de 2019. Aceptado con modificaciones: 10 de octubre de 2019. Aceptado para ser publicado: 16 de octubre de 2019. ABSTRACT Objective: Evaluate the association between early menarche and anthropometric/body composition variables in young female university students. Methods: Cross-sectional study involving female university students at a public university in northeastern Brazil. Anthropometric and body composition data were collected, along with a self-report of age at menarche. The conceptual model considered demographic, socioeconomic and lifestyle variables as well as the perception of weight in childhood and at menarche. Results: Among the 162 students analyzed, 62.3% were less than 20 years of age. The frequency of underweight and overweight was 8.6% and 22.3%, respectively. The frequency of above-average body fat and obesity was 32.1% and 14.8%, respectively. Early menarche was reported in 35.2%. The frequency of self-reported excess weight in childhood and at menarche was 31.3% and 25.9%, respectively. A statistically significant association was found between early menarche and height < 1st tertile (< 1.62 m). A tendency toward an association was found between early menarche and both weight, circumference in the risk range and the perception of excess weight in childhood and at menarche. Conclusion: Current excess weight was more associated with excess weight in childhood and at menarche than the occurrence of early menarche.


INTRODUCTION
The relationship between weight and sexual maturity, which is generally evaluated, in girls, by the age at menarche, is quite complex 1 . Early menarche (EM) has been considered a marker for the risk of obesity in adolescence and adulthood. However, girls with EM tend to be overweight prior to puberty. Thus, the occurrence of EM may simply reflect the effect of rapid weight gain in childhood 2 .
Depending on the study, EM is defined from 9 to less than 12 years of age 3,4,5 . According to Gaudineau et al 6 , this variability can be explained by the age of menarche found in different populations. Therefore, there is no consensus on age limits for defining EM. Thus, menarche generally is considered early if it occurs before 12 years of age and late if it occurs at or after 15 years of age 7 .
Most studies conducted in Brazil that have explored the relationship between menarche and nutritional status are limited to the investigation of its effect during the period of adolescence 3,8 . A study involving adult women in the southern region of the country found a protective effect of late menarche for obesity8. Pinheiro et al. 9 found a 67.7% frequency of excess weight among women who had EM (< 13 years), whereas the frequency was 54.0% in the group with a later menarche. This association has also been found in studies conducted in other Brazilian states 10,11,12 . Therefore, the aim of the present study was to investigate whether EM is associated with anthropometric variables and body composition among young university women.

METHODS
A cross-sectional study was conducted involving students studying health majors at a public university in northeast Brazil in 2015. Students aged ≥ 30 years, pregnant women, those who already had children and those who were unable to undergo the anthropometric evaluation due to poor physical condition were excluded. This study received approval from the Human Research Ethics Committee of the Center for Health Sciences of Universidade Federal de Pernambuco (certificate number: 41423215.6.0000.5208).
The sample size was calculated using the Statcalc program of EPI-INFO, version 6.04, considering a 95% significance level (1-α), 80% power (1-β), 1:1 proportion, exposure (obesity) and relative risk of 1.6. A sample of 120 university students was determined. A convenience sample was used of women who agreed to adhere to the study.
The students were evaluated at the beginning of the school year. Age at menarche was defined as the self-reported age at which the first menstrual cycle occurred. Menarche was considered early when it occurred at a younger age than the median of its occurrence among the university students analyzed (12 years).
The anthropometric variables collected were weight, height and waist circumference (WC). Weight was determined using a digital scale (Plena) with a 150 kg capacity and 100 g precision. Height was determined using a portable stadiometer (Ghrum Polar, Switzerland) with 1 mm precision. Both weight and height were measured using the methods proposed by Lohman et al. 13 and served as the basis for the calculation of the body mass index (BMI). For students up to 19 years of age, the BMI was classified using the cutoff points for adolescents recommended by the World Health Organization (WHO) 14 . For those aged 20 years or older, cutoff points for adults were used 15 . WC was measured at the midpoint between the last rib and iliac crest using a non-elastic metric tape. Abdominal obesity was evaluated using both WC and the waist-to-height ratio (WHtR= WC in cm/height in cm). For adolescents, we used the cutoff points for the classification of WC recommended by Taylor et al. 16 For adults, the cutoff points proposed by the WHO were used 17 . For WHtR, we used the cutoff points proposed by Li et al. 18 Body composition was evaluated using bioimpedance analysis. For such, the Maltron BF-906 device (Maltron, United Kingdom) was employed, with a 50-Hz frequency in alternating current and four electrodes. Readings were made with the students lying on a non-conductive surface in the supine position with the limbs bent at 45º and the volunteer free of any accessories or metallic objects. To ensure the accuracy of the readings, participants were instructed to follow the following procedures: four hours of absolute fasting; no rigorous physical exercise in the previous 12 hours; refrain from ingesting alcoholic beverages in the previous 48 hours; no use of medications that affect the hydroelectrolytic balance in the previous seven days; and urinate at least 30 minutes prior to the reading. The cutoff points for above average body weight and obesity were ≥ 24% and ≥ 32%, respectively 13 .
The conceptual model considered demographic, socioeconomic and lifestyle variables as well as perceptions of weight in childhood and at menarche. Socioeconomic status was categorized using the Economic Classification Criteria of Brazil established by the Brazilian Association of Research Companies 19 . Age was stratified into two groups: adolescents (<20 years) and adults (20 years). The lifestyle variables were physical activity level, sedentary behavior, alcohol consumption and smoking.
The short version of the International Physical Activity Questionnaire (IPAQ) was used for the determination of physical activity level 20 . Sedentary behavior was determined by the time spent on activities such as watching television and using a computer, considering two hours per day for each activity was considered excessive 21 . Alcohol consumption was dichotomized as "drinks" or "does not drink" based on intake frequency (number of times per week). Students who reported having smoked one or more days in the previous 30 days were classified as smokers 22 . To identify the occurrence of excess weight in childhood and at menarche, students were asked about their perceptions of their own weight and given four response options: underweight, normal, a little overweight and very overweight 23 .
The statistical analysis was performed with the SPSS version 13.0. The chi-square test was used to determine associations between variables. Continuous variables were tested for the normality of distribution using the Kolmogorov-Smirnov test.
Variables with normal distribution were expressed as mean and standard deviation and those with non-Gaussian distribution were expressed as median and interquartile range. Means were compared using the Student's t-test for unpaired variables and medians were compared using the Mann-Whitney U test. For all analyses, the 5% significance level was adopted, with p-values between 0.06 to 0.10 considered indicative of borderline significance.

RESULTS
A total of 162 university students were evaluated, among whom 35.2% reported the occurrence of menarche prior to 12 years of age (EM). A total of 62.3% of the sample were adolescents and more than half were middle class. Regarding physical activity, 42.6% were classified as insufficiently active or sedentary; 63.9% spent more than two hours a day on the Internet and 18.8% reported spending more than two hours a day watching television ( Table 1).
The anthropometric characteristics and body composition according to the occurrence of EM are displayed in table 2.
In the overall sample, the frequency of excess weight based on the BMI was 22.2% (14.8% prevalence of overweight + 7.4% prevalence of obesity). However, the frequency of the above average percentage of body  fat and obesity measured using bioimpedance analysis was 32.1% and 14.8%, respectively. A total of 31.2% reported excess weight in childhood and 25.9% reported that excess weight continued at menarche. EM was significantly associated with height < 1st tertile (< 1.62 m) and tendencies (borderline p-values) were found between EM and both WC and WHtR in the risk range as well as perceptions of excess weight in childhood and at menarche. No statistically significant differences were found between the students with and without EM regarding the anthropometric variables and body composition, except WHtR, which had a tendency to be higher in the group with EM (Table 3). Analyzing the situation in terms of means and medians (Table 4), no statitically significant differences were found between the studants with and without EM regarding the anthropometric variables and composition, except WHtR, which had a tendency to be higher in the group with EM.
In correlation analysis between age at menarche and the anthropometric variables, inverse correlations were found with BMI, lean mass and WHtR (Figure 1).

DISCUSSION
In recent years, age at menarche has been the focus of attention in the literature due to its possible association with health status in adulthood 24 . In addition, with the influence of economic and social development of populations, there has been a trend toward earlier occurrence in recent decades 25,26 . However, there is no standardization in the literature on what is considered early. The most widely accepted definition is that which considers EM as occurring before 12 years of age 24 . In the present study, EM was considered as occurring prior to the median in the sample (12 years), which is in line with the most cited age in the literature. Despite methodological differences in the comparison of studies, the age that characterizes EM does not fluctuate much in the literature, with mean values ranging from 11.0 to 13.0 years 24,27,28 .
The frequency of EM in the present study was 35.2%. In a study involving adolescents in the city of Porto Alegre (southern Brazil), Spinola et al. 27 found a much higher frequency of EM (73.3), but considered a cutoff point of < 13 years. According to Carvalho et al. 29 , age at menarche seems to continue to diminish throughout the world, but apparently at a much slower pace in recent years when compared to that observed at the beginning of 20th century. However, if this trend continues, menarche will likely occur at nine or ten years of age in the near future, which is in childhood. This could cause a set of problems, such as a reduction in height gain and the occurrence of excess weight and body fat in adulthood, which are aspects investigated in the present study.
Although menarche is an intrinsic biological factor for each individual, the age at which it occurs may be related to environmental factors, eating habits and hormonal characteristics. Therefore, studies indicate the possibility that factors such as excess weight in childhood may contribute to early sexual maturity 27 .
Analyzing the anthropometric characteristics according to the occurrence of EM, an association was found with height < 1st tertile. According to Vitalle et al. 30 , the postmenarche period is characterized by a diminishing in the increase in height and a characteristic increase in weight. Lago et al .31 also found that age at menarche is directly associated with final height in younger generations. The authors found that a later menarche was associated with a taller height, with gains in stature on the order of 0.4 to 0.9 cm for each additional year in the age at menarche. As the majority of girls have reached 95% of their final height at the time of their first menstrual cycle, a later menarche would tend to be associated with a greater final height 31 .
Due to hormonal issues, obesity is believed to exert an influence on maturity in girls, making them menstruate at an earlier age 32 . In a study involving adolescents, Castilho & Nucci 8 found that among girls recruited at private schools, those with excess weight menstruated earlier (11.6 years) than those without excess weight (12.3 years). Results published by Mumby et al. 33 lend support to a causal relationship between Figure 1: Pearson's correlations between age at menarche and body mass index (BMI), lean mass in kg and waist-toheight ratio (WHtR). a high BMI in childhood and a younger age at menarche, suggesting that the greater prevalence of childhood obesity leads to premature sexual maturity.
In the present study, perceptions of excess weight in childhood and at menarche demonstrated a tendency toward an association with early sexual maturity. This association has biological plausibility, as body fat may be a precursor of the onset of puberty due to the synthesis of leptin by adipocytes, thereby accelerating the maturation of the tissue and reproductive function 34 . Moreover, changes in adipose tissue and its distribution during adolescence are strongly influenced by sexual maturity, which, in turn, is marked by human biological variability that originates from the interaction between genetic and environmental factors 35 . There is a growing body of evidence that sexual maturity constitutes a risk factor for a greater percentage of body fat. On the other hand, individuals with a greater percentage of body fat have a greater chance of early maturity 36 .
Pinheiro et al. 9 concluded that the prevalence of excess weight in adulthood was higher among women who reached menarche prior to 13 years of age. Mueller et al. 37 found an association between younger age at menarche and higher BMI at 20 years of age in a study involving more than 8000 women in Brazil. The literature also reports that early puberty onset is an important risk factor for obesity among women who had their first menstrual cycle prior to 12 years of age, with a 59% higher risk of obesity when compared to those who had their menarche after 12 years of age 1 .
In the present sample, we found an association between EM and shorter height as well as a tendency toward an association with past of excess weight and both current WC and WHtR. In other words, our data imply that excess weight in childhood contributed to EM and was perpetuated in the form of abdominal obesity.
In a five-year longitudinal study with girls in the ideal weight range at the time of menarche, Araújo et al. 3 found no statistically significant differences between girls with and without EM in terms of excess weight, abdominal obesity or abnormal serum lipid levels at the end of sexual maturity. Likewise, Must et al. 28 conducted a longitudinal study and found that nutritional status in childhood exerted a greater influence on the occurrence of excess weight in adulthood than EM.
The present data also showed weak, but significant inverse correlations between age at menarche and current BMI and WC. Ibanez et al. 38 found that, independent of BMI, measures that reflect abdominal obesity were significantly more frequent among women who had EM. Mueller et al. 39 also found that EM was associated with visceral and subcutaneous abdominal fat deposits in adulthood.
Few studies in the literature have discussed the association between lean mass and EM. In 2000, Castilho & Barros Filho 40 found that peak weight gain velocity and growth in stature differs considerably among girls who mature early, those who reach menarche at an average age and those who mature later, but menarche nearly always coincides with or occurs a few months after the peak weight gain velocity. Moreover, the authors report that peak gain in muscle mass occurs at menarche in girls and begins to decelerate thereafter.
In the present study, a statistically significant inverse correlation was found between age at menarche and the amount of muscle mass in adulthood. One may therefore conclude that an earlier occurrence of sexual maturity leads to more physiologically mature women (greater lean mass) at the end of adolescence and the onset of adulthood.
The present study has limitations that should be considered. Age at menarche was not collected in real time and this variable may have been subject to recall bias. The perception of excess weight in childhood and at menarche was self-reported by the students and may therefore have been underestimated, especially when one considers that the sample was composed of young women. However, self-reported weight and height have been used in epidemiological studies due to the cost effectiveness and simplification of the fieldwork and have presented acceptable levels of validity, even among individuals with obesity, who may have a greater tendency to underestimate their weight 15 .

CONCLUSION
The findings of the present study reveal that excess weight at the onset of adulthood in young university women may be more associated with excess weight in childhood and at menarche than EM. This study also discusses a finding that is addressed little in the literature, which is the association between EM and muscle mass in early adulthood.