SEARCH

SEARCH BY CITATION

Keywords:

  • Androgens;
  • anogenital distance;
  • prenatal exposure;
  • women

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. References

Objective

Animal models have suggested that anogenital distance (AGD) at birth reflects androgen levels during in utero development and predicts adult AGD. A recent study showed an association between perineal length and androgen levels in men, suggesting that serum testosterone levels in adulthood will depend on factors involved during the fetal period. The aim of this study is to assess the relationship between AGD measures and reproductive hormone levels in women.

Design

Cross-sectional study conducted between February and November 2011.

Setting

University-affiliated fertility clinics.

Population

100 young college students.

Methods

Physical and gynaecological examinations were conducted on university students. All participants provided a blood sample for determination of reproductive hormones and completed an epidemiological questionnaire on lifestyles and gynaecological history. We used multiple linear regression analysis to examine the associations between perineal length measurements [anus-fourchette (AGDAF) and anus-clitoris (AGDAC)] and reproductive hormone levels.

Main outcome measures

Anogenital distance measurements and reproductive hormone levels.

Results

In the multiple linear regression analyses, AGDAF was positively associated with serum testosterone levels. Serum testosterone increased 0.06 ng/ml (95%CI 0.01, 0.10; = 0.02) for each 1-cm increase in AGDAF. None of the measurements was associated with other reproductive hormones.

Conclusions

Anogenital distance may predict normal reproductive development in women, and may be a new tool of potential clinical interest to evaluate ovarian function. Our results suggest that serum testosterone levels in adulthood may depend on factors operating in the prenatal period.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. References

Anogenital distance (AGD) is used as a marker of genital development in animal models and in humans, being longer in males.[1-7] Several experimental studies have shown that AGD at birth reflects androgen levels during in utero development and predicts AGD in adulthood.[8] Basically, AGD has been shown to reflect the amount of androgens to which a fetus is exposed in early development; higher in utero androgen exposure results in longer and more masculine AGD. In males, AGD has been found to have similar values in people of different age groups, suggesting that AGD may provide a measure of genital development and function throughout human adult life.[9]

A number of animal studies have shown that the female reproductive tract is susceptible to virilisation by exogenous androgens, prior to, as well as during, the in utero masculinisation programming window.[10-12] Exposure to androgens in the uterus also results in higher hormones levels later in life in animal models. As shown by Wu et al.[12] prenatally androgenised female rats presented higher levels of androgens postnatally compared with the control groups.

In men, Eisenberg et al.[13] have shown a positive association between AGD and androgen levels in men, and shorter AGD is related to a possible testicular dysfunction. Recently, Mendiola et al.[14] reported significant positive associations between AGD measures and the presence of greater ovarian follicular number in women, suggesting that the prenatal androgenic milieu may exert a long-lasting influence on both AGD and the female reproductive system in humans. However, to the best our knowledge, no study has investigated the relationship between AGD measures and reproductive hormone levels in women, a relationship this study was set up to explore.

Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. References

Study population

The study methods have been described previously.[14] The Murcia Young Women's Study (MYWS) is a cross-sectional study of healthy young university students (18–23 years old), carried out between 9 February and 25 November 2011. Of 124 students who contacted us, 109 met the eligibility criteria for the study and 100 attended the clinical appointment. All participants provided a blood sample and completed an epidemiological questionnaire on lifestyles and gynaecological history. Besides that, at a scheduled clinic visit, women underwent a physical and gynaecological examination, including genital measurements and transvaginal ultrasound. Participants were compensated for their participation (€40 gift card). Written informed consent was obtained from all women. The Research Ethics Committee of the University of Murcia approved this study.

Serum reproductive hormone analysis

One non-fasting blood sample was drawn between 12 and 5 pm on the same day of the clinical appointment and during the early follicular phase (days 1–6). The blood serum was separated by centrifugation, coded and frozen at −80 °C until analysis. All samples were analysed for hormones in the same laboratory at the ‘Virgen de la Arrixaca’ University Hospital (Murcia, Spain). All hormone assessments were carried out simultaneously to reduce intralaboratory variation. Serum levels of follicle-stimulating hormone (FSH), luteinising hormone (LH), prolactin (PRL), testosterone (T) and estradiol (E2) were determined using a time-resolved electrochemiluminiscence inmunoassay (Roche Diagnostic Corporation, Indianapolis, IN, USA). Intra- and interassay coefficients of variation (CVs) were the following: for FSH 2.6 and 3.6%, LH 0.8 and 2.0%, PRL 0.8 and 1.8%, T 2.8 and 3.2%, and E2 2.1 and 2.8%. In this study, assay sensitivities were 0.10 IU/l for FSH and LH, 1.0 μIU/ml for PRL, 0.025 ng/ml for T, and 5.0 pg/ml for E2.

Participants’ physical examination and gynaecological history

A complete gynaecological history was taken for each woman, including history of gynaecological diseases (endometriosis, salpingitis, etc.) (yes/no), previous irregular menstrual cycles (yes/no) and self-reported menstrual cycle length (days).

Body mass index (BMI) was calculated as weight in kilograms divided by squared height in meters. For each woman, two variants of anogenital measurement were taken using a digital calliper: anus–clitoris (AGDAC) and anus–fourchette (AGDAF) (Figure 1).[14] To increase precision, two different examiners measured each AGD three times, resulting in a total of six measures for AGDAF and another six for AGDAC. The mean value of the six measurements of each AGD was used. Neither the examiners nor the support staff had knowledge of the women's reproductive hormone levels at the time of the physical examination. As reported elsewhere,[14] both AGDs (AGDAC and AGDAF) had a normal distribution and were correlated (r = 0.44, < 0.01). Intra- and interexaminer coefficients of variation for both AGD measurements were below 3 and 10%, respectively. We found a positive association between both AGD measures and BMI (< 0.01) and a negative association between AGDs and hormonal contraception (< 0.05).[14]

image

Figure 1. Landmarks for two measurements of AGD: AGDAC, from the anterior clitoral surface to the centre of the anus (point 1 to point 3); and AGDAF, from the posterior fourchette to the centre of the anus (point 2 to point 3). Adapted with permission from Sathyanarayana et al.[5]

Download figure to PowerPoint

Statistical analyses

Descriptive statistics are presented using untransformed data. We used multiple linear regression analyses to identify predictors of each of the two AGD measurements. Covariates initially examined as predictors of AGD measurements were: age, height, weight, BMI (kg/m2), age at menarche, self-reported sexually transmitted diseases (STDs), taking medication (antibiotics or antihistamines; yes/no) and hormonal contraception (yes/no). The covariates that were examined as predictors of serum levels of reproductive hormones were: age (years), BMI (kg/m2), smoking status (current smoker versus not current smoker), season of the year (spring versus summer, fall, winter), day of the menstrual cycle (days 1–6), time of the day when the blood sample was obtained, and hormonal contraception (yes/no). We also used multiple linear regression analysis to examine the associations between AGD measurements and the serum reproductive hormone levels of the young women. When inclusion of a potential covariate resulted in a change in the β coefficient of <10%, the variable was not retained in final models. For these statistical analyses outliers were excluded, women with extreme hormone concentrations (>2 standard deviations of the mean value). There were two women for T, one for LH, two for PRL and four for E2. Statistical analyses were performed with the statistical package IBM SPSS 21.0 (IBM Corporation, Armonk, NY, USA).

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. References

The study population of the MYWS was quite homogeneous. Participants were 18–23 years of age [mean: 20.0, standard deviation (SD): 1.2] with a mean BMI of 21.8 (SD 3.1), predominantly Caucasian (98%) and non-smokers (66%). Almost 40% used hormonal contraception (oral contraceptives or vaginal ring). Of the women, 61% were taking hormonal contraception to regulate menstrual cycles and 39% for contraceptive purposes or other symptoms (e.g. acne). All participants were nulliparous, most of them with a self-reported good or excellent health (92%). None of the women reported diagnosis of endometriosis, salpingitis, STDs, genital lesions or transfusions.

Table 1 summarises the AGD measures and serum reproductive hormone concentrations of the young participants. Significant covariates were identified with regard to serum reproductive hormone levels. We found a statistically significant association between BMI and T levels [β = 0.017; 95%CI 0.008–0.026]. Season of sample collection was significantly related to several reproductive hormones. Samples collected in the fall (versus spring) had lower levels of T [β  =  −0.087; 95%CI −0.156 to −0.017] and FSH [β  =  −1.112; 95%CI −2.159 to −0.066]. Winter (versus spring) presented lower levels of T [β = −0.141; 95%CI −0.263 to −0.019]. Time of the day when the blood sample was obtained was positively associated with serum levels of PRL [β  =  0.604; 95%CI 0.214–0.993], and hormonal contraception was negatively associated with serum levels of LH (β  =  −1.354; 95%CI −2.533 to −0.176), PRL (β  =  −56.252; 95%CI −108.334 to −4.170) and E2 (β  =  −10.470; 95%CI −19.448 to −1.492). For this reason the subsequent statistical analyses were adjusted with these covariates.

Table 1. AGDs and hormonal features of young women participating in the Murcia Young Women's Study (MYWS)
VariablesMean (SD)Median (25–75)
  1. a

    One woman in whom no physical examination was performed (= 99).

  2. AGDAC, anogenital distance from the centre of the anus to the anterior clitoral surface; AGDAF, anogenital distance from the centre of the anus to the posterior fourchette; SD, standard deviation; (25–75), 25th–75th percentile.

Genital measurements a
Anogenital distance (AGDAC) (mm)80.4 (10.5)79.2 (73.2–87.2)
Anogenital distance (AGDAF) (mm)37.7 (6.3)37.2 (33.3–41.8)
Serum hormone levels
FSH (IU/l)5.8 (2.0)6.1 (4.7–6.9)
Testosterone (ng/ml)0.39 (0.82)0.30 (0.20–0.41)
LH (IU/l)5.8 (3.1)5.6 (4.2–6.9)
Prolactin (μIU/ml)256 (154)216 (160–294)
Estradiol (pg/ml)57.7 (110.6)36.5 (27.9–53.2)

As shown in Table 2, AGDAF, but not AGDAC, was positively associated with serum T levels in young women (< 0.05) (Figure 2A,B). For each centimetre increase in AGDAF, serum T increased by 0.06 ng/ml (95%CI 0.01, 0.10; = 0.02). We calculated the estimated increase in serum T levels going from the 25th to the 75th percentile of AGDAF. For the 25th percentile (33.3 mm) the expected serum T level was 0.292 ng/ml, for the 50th percentile 0.337 ng/ml, and for the 75% percentile 0.362 ng/ml (41.8 mm). Thus, based on the best-fitting model, an interquartile increase in AGDAF is associated with an increase in T serum levels that is 20.7% of the median. No associations were found between the remaining reproductive hormones and AGD measurements. In addition, associations between hormone ratios (LH/FSH, E2/T) and AGD measures were assessed, but no significant associations were observed (data not shown). There were only 10 cases of clinically diagnosed PCOS in our study population, therefore any meaningful interpretations with regard to serum T levels or AGD measures are difficult to make in this small subpopulation.

Table 2. Multivariable linear regression models for serum reproductive hormones and AGD measures
Serum reproductive hormonesAGDACAGDAF
UnadjustedAdjustedaUnadjustedAdjusteda
β 95%CIP-value β 95%CIP-value β 95%CIP-value β 95%CIP-value
  1. a

    Controlling for current BMI and hormonal contraception.

  2. b

    Controlling also for season of the year.

  3. c

    Controlling also for time of day when the blood sample was obtained.

FSH−0.008−0.047, 0.0310.70−0.008−0.049, 0.0340.72b0.043−0.022, 0.1080.190.029−0.040, 0.0980.41b
Testosterone0.0030.000, 0.0060.030.002 −0.001, 0.0050.16b0.0070.003, 0.0110.0020.0060.001, 0.0100.02b
LH 0.003−0.050, 0.0570.91−0.019−0.079, 0.0410.530.046−0.043, 0.1360.310.001−0.105, 0.1060.99
Prolactin1.127−1.353, 3.6080.370.563−2.121, 3.2480.68c2.765−1.287, 6.8160.182.138−2.335, 6.6110.35c
Estradiol 0.247−0.167, 0.6610.24−0.016−0.482, 0.4510.950.185−0.519, 0.8880.60−0.210−1.017, 0.5960.61
image

Figure 2. Simple linear regression plot of testosterone levels in young women modeled as a function of (A) AGDAF and (B) AGDAC.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. References

Main findings

In our study, longer AGDAF was significantly associated with higher serum T levels in young women. This is the first study to report an association between AGD measures and levels of reproductive hormones in women. We have previously established a relationship between AGD and ovarian follicular number. But, to the best of our knowledge, the current study represents the first evaluation of the association between AGD measures and serum T in human females.

Strengths and weaknesses

Our population was relatively small and narrow in age and ethnicity. In the absence of bias, a relative small sample size would be relevant for type II error, but not in our study, as significant associations were found. AGD measures were well tolerated by all women with adequate intra- and interexaminer consistency. Besides, hormone concentrations were unknown at the time of genital measurement. Our data were limited by the use of a single serum sample to describe hormone function but there is evidence showing that a single sample can be used to classify women's reproductive hormones.[15] We were unable to measure other potentially important reproductive hormones such as the anti-Müllerian hormone (AMH) or inhibin b, as well as SHBG. Therefore, free androgen concentrations, which may be of relevance and be affected by obesity/insulin resistance, were not assessed. We cannot rule out that type II error might explain the negative findings with regard to other reproductive hormones and AGD. Unfortunately, we measured a 2:4 digit ratio, a marker of androgen exposure, in only 35 of the women and could not draw meaning conclusions about its association with AGD. Getting more information to link 2:4 digit ratios to AGD measures and reproductive outcomes would be extremely useful in future human studies. Lastly, the cross-sectional design of the study curtails the temporality criteria of causality,[16] and therefore prevents causal interpretations of the associations reported here.

We found significant associations with serum T levels only for AGDAF. Other authors have shown similar findings in human males, reporting significant associations between semen quality and the short (equivalent) measurement (anus–scrotum), but not with the long one (anus–penis).[17] This could in part be due to the influence of BMI on adult AGDAC, as BMI influences the size of the fat pad anterior to the pubic symphysis, an area that is included in AGDAC but not in AGDAF.

Two of the five hormones assessed in the current study (T and E2) are synthesised in the ovaries. However, we only found a relationship with AGD measures for T, not for E2. We cannot completely rule out that taking hormonal contraception may have altered those results, although we accounted for that possibility in our multivariate models. Moreover, we carried out a sensitivity analysis to assess whether taking hormonal contraception was a modifier of the association between AGD measures and serum T levels, as well as for all the other reproductive hormones. The associations were similar when we analysed both subpopulations separately, and we therefore concluded that there was not effect modification by hormonal contraception.

Interpretation

Several experimental and observational studies have shown that AGD can be used as a sensitive marker of fetal exposure to different environmental endocrine disruptors, able to alter androgen signaling, and lead to abnormal genital development and reproductive dysfunction.[6, 18-23]

In utero exposure to agents that interfere with the action of androgens in male rodents, induced genital development disorders, involving impaired sperm production, lower testosterone levels and shorter AGD in adults.[3, 24] Prenatal exposure to exogenous androgens in female rodents results in increased ovarian follicular recruitment, higher levels of androgens in adulthood, and longer, more masculine, AGD compared with controls.[12] Thus, longer AGD and a higher level of circulating androgens in adult females may share a common hyperandrogenic environment in utero, suggesting a prenatal origin for ovarian dysfunction.

Two recent studies in human males have correlated AGD to sperm production, showing that men with shortened AGD presented poorer semen quality[17] and were more likely to be infertile.[25] In females, a significant positive association has been found between both AGDs and ovarian follicle number, suggesting common fetal origins for longer AGD and greater follicular recruitment.[14]

Buck Louis et al.[26] have suggested the term ‘ovarian dysgenesis syndrome’. In utero environmental exposures would alter normal development processes, which in turn would be associated with later female reproductive conditions (fecundity impairments, gynaecologic disorders, gravid diseases, etc.).[26]

In our study, longer AGD was associated with high levels of testosterone in adult women. These findings are consistent with previous studies in animal models. Wu et al.[12] suggested that a prenatal androgenic exposition can de-feminise gonadotropin secretions in the female fetus. Therefore, a prenatal hyperandrogenic environment induces an alteration in the normal cyclic hormone secretions of female adult animals. Other authors have suggested that an exposure to excessive in utero androgens increased LH secretions and conferred resistance to the negative feedback of ovarian hormones, producing high levels of T in females.[27-29]

Conclusions

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. References

Our work is the first study showing an association between AGD and androgen concentrations in women. Following recent reports examining adult AGD and ovarian function, AGD might predict normal genital development in women, and could present a new instrument to evaluate ovarian function. It would be advisable to take both AGD measures until a significant body of normative data has been accumulated in both male and female adults. Assuming that AGD at birth predicts adult AGD, our results suggest that serum T levels in adulthood may depend on factors operating in the prenatal period. However, associations between AGD measures and reproductive hormone levels in women are of uncertain clinical significance at this stage. Consequently, further studies should follow to corroborate these findings and assess their potential clinical and public health significance.

Disclosure of interest

The authors declared no conflicts of interest with respect to the authorship and/or publication of this article.

Contribution to authorship

M.R., J.M. and A.M.T.C. designed and initiated the current study. M.R., L.M.A., M.M., J.A.N.V. and J.M. were responsible for collecting and analysing the samples and the interview data. L.M.A. and M.R. coordinated the fieldwork of the study. J.J.L.E., A.C.T. and M.P.M.E. were responsible for statistical analysis. M.P.M.E., J.M., M.M., A.C.T. and A.M.T.C. were responsible for writing the draft version of manuscript. All authors commented on and approved the final manuscript.

Details of ethics approval

The study received ethical approval from The Research Ethics Committee of the University of Murcia. Ref. no 495/2010, 14 May 2010.

Funding

This work was supported by a research contract from Gestión Clínica Avanzada SLU; ‘Ministerio de Ciencia e Innovación, Instituto de Salud Carlos III (FIS) grant no PI10/00985’ and ‘Fundación Séneca, Región de Murcia, Agencia Regional de Ciencia y Tecnología, grant no 08808/PI/08’.

Acknowledgements

The authors gratefully acknowledge Dr C. Millán, L. Sarabia, C. Ruiz, E. Belmonte and all the Quirón Dexeus Murcia clinic staff for their assistance in data collection, and the young women of the study for their participation. We also thank K. J. Ruiz-Ruiz and E. Estrella for their work on database management.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. References