Fetal anogenital distance using ultrasound

Abstract Objective This study measured anogenital distance (AGD) during late second/early third trimester of pregnancy to confirm previous findings that AGD can be measured noninvasively in the fetus using ultrasound and further showed differences in reference ranges between populations. Method Two hundred ten singleton pregnancies were recruited at the Rosie Hospital, Cambridge, UK. A 2D ultrasound was performed between 26 and 30 weeks of pregnancy. AGD was measured from the centre of the anus to the base of the scrotum in males and to the posterior convergence of the fourchette in females. Results A significant difference in AGD between males and females (P < .0001) was found, replicating previous results with a significant correlation between estimated fetal weight (EFW) and AGD in males only (P = .006). A comparison of AGD using reference data from an Israeli sample (n = 118) and our UK sample (n = 208) showed a significant difference (P < .0001) in both males and females, after controlling for gestational age (GA). Conclusion Our results confirm that AGD measurement in utero using ultrasound is feasible. In addition, there are strong sex differences, consistent with previous suggestions that AGD is influenced by prenatal androgen exposure. AGD lengths differ between the UK and Israel; therefore, population‐specific normative values may be required for accurate clinical assessments.

found to further influence AGD 13 in humans. This supports the theory that prenatal androgen exposure acts as the driver of AGD length. In adults, AGD correlates with circulating serum testosterone, 16 as well as with the aromatisation ratio (of circulating testosterone to oestradiol). 17 This suggests that AGD reflects the different aspects of the masculinisation pathways in development, and the relative balance between testosterone and oestrogens, rather than testosterone levels alone. It was thus suggested that AGD may be a suitable proxy to estimate prenatal androgen exposure.
Measuring AGD in 2D ultrasound scans has been shown to be feasible and reliable. 1,12,18 This suggests that this measure has the potential to aid in the early identifications or understanding of pathogenesis of genitalia development, 12 PCOS, 19 and anorectal 4,20 and male genitalia 18 malformation. However, there may be populationspecific differences in this measurement, with ethnicity or other regional factors affecting AGD centile charts and reference ranges. 1 This study aims to further establish the feasibility of measuring AGD during the late second/early third trimester of pregnancy, for the first time in a UK sample, in particular to test for sex differences in AGD, and to assess population-specific reference ranges, by comparing UK measures with those collected in Israel.

| Participants
Two hundred nineteen healthy fetuses, 104 male and 115 female, were recruited prospectively in the Rosie Maternity Hospital in Cambridge, UK. The same inclusion and exclusion criteria were used to those used in the Chaim Sheba Medical Centre study. 1 Eligibility inclusion criteria for the study were as follows: pregnant women who were willing to have an additional ultrasound scan between 26 and 30 weeks of gestation (average GA: 28 weeks, SD = 1.25), with (a) little/no consumption of alcohol during pregnancy, (b) no smoking or recreational drug use during pregnancy, (c) a singleton appropriate-for-gestational age fetus, (d) the absence of any major fetal anomalies, and (e) fetus is not considered to have intrauterine growth restriction (IUGR) or be large for gestational age (LGA). Eligibility criterion for inclusion of the data in the final analysis was the birth of a clinically healthy baby. To observe difference in reference ranges, normal modelled AGD charts created from an Israeli population 1  LGA or developed any abnormalities since the last routine scan. AGD measurements were taken on the tangential section of the fetal perineum, where the anal sphincter would first be observed. AGD was measured from the centre of the anus to the base of the scrotum in males and to the posterior convergence of the fourchette in females using electronic callipers, following the same procedure used in previous research measuring fetal AGD 1,12 ( Figures 1 and 2). For this measurement to be taken, the fetus' legs must be apart to accurately visualise the scrotum (in males) and fourchette (in females). If the legs were not separated when this measure was attempted, the mother was given cold water or asked to walk around for several minutes. In instances where the fetus was breech in presentation, the examination bed was tilted to move the fetus out of the pelvis and remove shadowing whilst the measurement was being taken.
Failure to obtain any AGD measurement was only seen in 4.1% (nine fetuses, three male and six female; fetal positioning, four breech, four cephalic, and one not available) of the total sample (see Table 1 for population details). Because of the low failure rate of obtaining this measure, more invasive techniques such as transvaginal ultrasound and external cephalic version were not considered.
Sex assignment was performed by observation of sonographic landmarks such as labial lines or fetal scrotum, in line with the established gold standard, 12,21 and was further confirmed at birth. As there was slight variability in AGD measurement due to placement of the fetus, What is already known about this topic?
• AGD is a sexually dimorphic measure.
• Previous research has linked AGD to prenatal androgen excess.
• This measure could assist with early identification of a range of neurodevelopmental, endocrine outcomes and the early diagnosis of conditions characterised by steroidogenic excesses or deficits.
What does this study add?
• This study confirms that, using ultrasound, AGD can be successfully measured in utero.
• We identify the need for population-specific normative charts if this measure is to be used clinically and of diagnostic value. several measurements (range 1-3 freeze frame measurements) were taken and averaged for interobserver reliability purposes.

| Analysis
SPSS statistical version 25 was used. Data were examined using a t test and a paired-samples t test. To test for population differences (see Table 2 for raw AGD measurements of both populations), statistical analysis was performed using bilinear interpolation of the observed data from the UK set against an average fetal AGD biometry chart using the normal modelled AGD charts created from an Israeli population. 1 This used the gestation and centiles to provide average expected AGD from a different population sample that were then gestation and centile matched to this UK population sample and analysed using a paired-samples t test. As several freeze frame measurements were taken for AGD, intraobserver variability was assessed. The difference between the raw measures is presented in a Bland-Altman  Associations of AGD with maternal and fetal characteristics were assessed with univariate Pearson linear regression (Table 3) separately for each sex, to investigate potential covariates. The reported maternal body mass index (BMI) was calculated on weight and height measurements, as reported by the participants at the time of this additional scan. Nominally significant variables were then consecutively introduced to multivariate linear regression models, according to their level of significance, with AGD as the dependent variable and nominally significant characteristics as the independent variables (Table 4).

| RESULTS
The study included 210 fetuses (101 male and 109 female). Nine fetuses (three male and six female) were excluded from the analysis as adequate AGD measures could not be obtained because of fetal orientation during the ultrasound. There was no significant association between any of the specific fetal positions (eg, breech; see Table 1 for frequencies) and whether an adequate AGD measure could be obtained (Pearson chi-squared test: χ 2 = 6.278, There was a significant difference in AGD between males (range: to test for a correlation between EFW and AGD. There was a partial significant correlation between EFW and AGD (r = .193, P = .005); however, when the analysis was split by sex (females, r = .135, P = .162; males, r = .272, P = .006), this correlation only remained significant in males.
Variables that were associated with AGD (Table 3) were included in multivariate linear regression models. The maximum variance that could be explained by all predictors was 64.5% (adjusted R 2 : .645, P < .0001), with fetal sex having the most pronounced effect on AGD, across all models (Table 4).  and five females from the UK sample were not included in this analysis as centile measurements from their ultrasound were not obtained. See Table 5 for means and SD of the two populations ( Figure 6).

| DISCUSSION
This study supports previous findings that AGD can be reliably measured in utero during the second and third trimesters of pregnancy.
The minimal variability in this measurement between independent raters suggests that this can be measured reliably by both sonographers and researchers. In addition, we replicated previous fetal   Neonatal and animal research shows a slight correlation between weight and AGD measurement. 11, 25 We found a similar significant correlation to EFW and birth weight (Table 3) in males but not in female fetuses. This could be due to sexual dimorphism in the regulation of growth, 26 or it may be because the range of AGD is greater defined in males, making it easier to detect the correlation. In addition, we noted a modest association of maternal BMI at the time of the ultrasound scan (mean GA: 28 weeks) with AGD in males. This may be attributed to fetal weight as well, since the effect was not detected after combining maternal BMI with measures of fetal weight (Table 4).   Previous research has suggested potential differences in AGD means and ranges between ethnic backgrounds. 1 The influence of biological and environmental factors on prenatal growth has been well documented. [27][28][29] We found significant variation between our UK sample and gestational growth charts from a sample in Israel. This demonstrates that similar to fetal biometry charts (such as femur length and head circumference), there is population-specific variability in AGD. The observed variation in AGD length between our UK sample and Israel samples could be related to differences in fetal positioning during the ultrasound. As the same anthropometric protocol was implemented (ie, initial identification of the anal sphincter) to guide the measurement and use of the same ultrasound plane, this should result in little to no variability in the measurement taken between the two sites. Alternatively, differences in genetic predisposition to congenital adrenal hyperplasia in Israel could account for the observed variation in measured distances. 30 Causes of this variability would need to be investigated, but given its existence, population-specific charts need to be created if this measure is to be used clinically or for research purposes.

| CONCLUSIONS
Overall, we show significant prenatal sexual dimorphism and demonstrate the ability to successfully measure AGD in utero. These results support the potential for this measure to be utilised in the prediction of later outcomes. We also suggest the need for large scale population-specific normative charts if this measure is to be used to inform research or used clinically. In order to fully assess the utility of prenatal AGD as a clinical measure, longitudinal research from prenatal to adult life is needed. We will continue to follow this cohort longitudinally and assess and compare the feasibly and utility of measuring AGD length at prenatal and infant stages of development.

CONFLICT OF INTEREST
The authors have no potential conflicts of interest to disclose.

FINANCIAL DISCLOSURE
All authors have indicated that they have no financial relationships relevant to this article to disclose.

DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.