Skip to main content
Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

The association of prenatal amniotic sex hormones and digit ratio (2D:4D) in children aged 5 to 70 months: A longitudinal study

  • Luisa Ernsten ,

    Roles Conceptualization, Data curation, Formal analysis, Methodology, Project administration, Visualization, Writing – original draft

    Luisa.Ernsten@hhu.de

    Affiliation Department of Experimental Psychology, Heinrich-Heine-University, Düsseldorf, Germany

  • Lisa M. Körner,

    Roles Conceptualization, Methodology, Writing – review & editing

    Affiliations Department of Experimental Psychology, Heinrich-Heine-University, Düsseldorf, Germany, Department of General Pediatrics, Neonatology and Pediatric Cardiology, University Children’s Hospital, Medical Faculty, Heinrich-Heine-University, Düsseldorf, Germany

  • Marie Luisa Schaper,

    Roles Conceptualization, Formal analysis, Writing – review & editing

    Affiliation Department of Experimental Psychology, Heinrich-Heine-University, Düsseldorf, Germany

  • Judith Lawrenz,

    Roles Conceptualization, Writing – review & editing

    Affiliation Department of Experimental Psychology, Heinrich-Heine-University, Düsseldorf, Germany

  • Gareth Richards,

    Roles Conceptualization, Methodology, Writing – review & editing

    Affiliation School of Psychology, Faculty of Medical Sciences, Newcastle University, Newcastle upon Tyne, United Kingdom

  • Martin Heil,

    Roles Conceptualization, Investigation, Methodology, Resources, Supervision, Writing – review & editing

    Affiliation Department of Experimental Psychology, Heinrich-Heine-University, Düsseldorf, Germany

  • Nora K. Schaal

    Roles Conceptualization, Formal analysis, Methodology, Supervision, Writing – review & editing

    Affiliation Department of Experimental Psychology, Heinrich-Heine-University, Düsseldorf, Germany

Abstract

The sex difference of the 2D:4D digit ratio (female > male)–a proposed marker for prenatal testosterone exposure—is well established. Studies suggest it already exists in utero and is of moderate effect size in adulthood. However, evidence for the claim that 2D:4D reflects prenatal androgen action is limited, and the sex difference may exhibit lability during childhood. In the present study, 244 mothers were recruited in the course of an amniocentesis examination (performed between gestational weeks 14 and 18). Prenatal testosterone (T) and estradiol (E) levels were determined from amniotic fluid for boys and girls. The majority (97.4%, n = 114) of available female T levels (n = 117) were found below the level of quantification. Therefore, only male amniotic fluid data (n = 117) could be included for the analysis of associations between amniotic sex hormones (T levels and T to E ratio (T/E)) and 2D:4D. The families were then invited to each of the five consecutive follow-ups (ages: 5, 9, 20, 40, and 70 months) where children’s 2D:4D was measured for both hands. The alternative marker D[r-l] reflects the directional asymmetry of 2D:4D (right subtracted by left 2D:4D) and was subsequently calculated as an additional measure for prenatal T exposure. No significant correlations between amniotic T or the T/E ratio (measured between week 14 and 18 of gestation) with 2D:4D respectively D[r-l] were observed for any time point. There was a significant sex difference (females > males) and a significant age effect with moderate correlations of 2D:4D between time points. 2D:4D increased between 20 and 40 months and between 40 and 70 months of age. The findings raise questions regarding the applicability of 2D:4D as a marker for prenatal androgen action and are discussed in terms of the reliability of obtained digit ratio data as well as in terms of the developmental timing of amniocentesis.

Introduction

Since Manning, Scutt, Wilson, and Lewis-Jones [1] suggested the sexually differentiated ratio between the second and fourth digit length (2D:4D or digit ratio) as a marker for prenatal testosterone (T) exposure, a vast number of studies have investigated relationships between 2D:4D and different human behaviors and traits (for reviews see [26]). The fascination stems from the assumption that prenatal T has organizing effects on the brain and in turn affects sex-specific behavior later in life [7]. Considering the difficulties inherent with measuring prenatal sex hormones, a convenient and easily obtainable proxy such as 2D:4D is particularly appealing. However, questions remain regarding its validity.

The sex difference in 2D:4D (female > male) is well established [8] and studies of aborted fetuses suggest that it already exists in utero [9, 10]. People characterized by atypical prenatal hormonal environments have also been reported to exhibit differences in 2D:4D. Notably, individuals with complete androgen insensitivity syndrome (CAIS; 46,XY karyotype, but female phenotype) exhibit feminized 2D:4D [11, 12]. However, the effect sizes reported for the difference between healthy controls and females with CAIS were small. Further, the variance in 2D:4D for these samples does not appear to differ from that of controls, which questions the premise that it is prenatal testosterone acting on the developing tissues that explains the effect [13]. Individuals affected by congenital adrenal hyperplasia (CAH), who are exposed to high levels of androgens during the prenatal period, show masculinized 2D:4D compared to females without CAH [1416]. However, evidence is inconsistent and the effects sizes are relatively small [1417]. A recent meta-analysis [16] showed only small to medium sized differences in 2D:4D between CAH patients and healthy controls. Additionally, it was noted that individual studies have typically utilized small sample sizes, and that there has been relatively little research published in this area during the past decade. Other studies have reported that men with Klinefelter syndrome (47,XXY karyotype), i.e. phenotypical males with an additional X chromosome, have a more feminized 2D:4D than unaffected controls [18, 19].

2D:4D has repeatedly been shown to correlate with various behavioral measures [20], developmental conditions [21, 22], and even various types of cancer [23, 24]. However, the validity of 2D:4D as a marker for prenatal testosterone exposure has been questioned [20, 25, 26]. Nonetheless, according to a PubMed search using the keywords “2D:4D” and “digit ratio” there were more than 100 newly published studies from 2020 till today that have investigated associations between 2D:4D and various behaviors, traits, disorders, or diseases, like muscular strength and fitness [5, 27], parental income [28], sexual preferences [29], concentration of steroids like cortisol and vitamin D [30], thyroid disease [31], and migraine in adults [32].

In order to investigate whether 2D:4D is a valid proxy for prenatal androgen levels, studies using a direct measurement of hormonal exposure are required. However, it is challenging to obtain such data in human studies due to obvious ethical considerations. Some researchers have measured prenatal sex hormone exposure from amniotic fluid obtained during the course of medically necessary amniocenteses [7, 33, 34]. The advantage of this is the timing of amniocentesis, which is normally conducted between gestational weeks 14 and 20, a timeframe in which sex differences in amniotic T and fetal serum T are highest [3538]. Amniocentesis examinations are conducted in cases of suspected chromosomal anomalies or due to other risk factors (e.g., higher maternal age) during the second trimester of pregnancy. As this procedure is associated with a small risk of miscarriage as a consequence of invasive sampling [39, 40], non-invasive first trimester screening techniques have recently replaced many amniocentesis examinations [41]. It has therefore become less feasible to collect amniotic fluid samples, with only a few studies relating such data to subsequent phenotypic outcomes [42].

To date there are only three studies investigating associations between amniotic T and 2D:4D. Lutchmaya, Baron-Cohen, Raggatt, Knickmeyer, and Manning [43] published a prominent study showing that 2D:4D of the right hand of 2-year-old children was negatively correlated with the ratio of amniotic T to amniotic estradiol (E), referred to as the T/E ratio, measured during the second trimester of pregnancy. The sample consisted of 29 mother-infant dyads (18 male infants). Male infants exhibited descriptively lower 2D:4D than female infants, but this sex difference was not statistically significant. No significant association emerged between 2D:4D and T or E. The authors suggested the T/E ratio has an influence on fetal growth, developmental conditions like autism, protection against early onset of breast cancer in women or myocardial infarcts in men, as well as athletic abilities in men. This study is frequently cited as evidence for 2D:4D being a valid marker for prenatal sex hormone exposure, despite its methodological limitations. Notably, the sample size was quite small and correlational analyses were not conducted separately for girls and boys, although sex was controlled for as a covariate in multiple regression analysis [43]. Furthermore, and critically, a replication of the study did not find the proposed negative correlation between prenatal T and E measured in amniotic fluid as well as the T/E ratio between weeks 15 and 20 of gestation and 2D:4D measured after 4.5 years in a sample of 66 children and their mothers [44]. Richards Browne, and Constantinescu [44] found, in contrast to the assumed correlations based on Lutchmaya et al. [43], no associations between T as well as the T/E ratio and 2D:4D for either girls or boys. The authors also examined the difference between right and left 2D:4D (D[r-l]). This variable is discussed as an additional measure, with lower values thought to indicate a higher intrauterine androgen exposure [4547]. However, this variable also showed no association with prenatal androgens [44]. The third amniocentesis study from Ventura, Gomes, Pita, Neto, and Taylor found the expected negative relationship between amniotic T and 2D:4D measured a few days after birth [48]. However, this was present only for the left hand and only in newborn girls (significant correlation for left 2D:4D, statistical trend for right 2D:4D [48]).

Another issue in 2D:4D research is limited evidence of temporal stability, especially in young cohorts. Based on the assumption that 2D:4D is (at least partially) determined by prenatal androgens one can assume that the temporal stability should be high, especially before the onset of puberty. However, a review of relevant studies revealed limited evidence for the temporal stability of 2D:4D, and, moreover, the sex difference in prepubertal cohorts [49]. Longitudinal studies have shown an increase of 2D:4D with age, although with moderate to high correlations between the time points [20, 5052]. Across these studies, the sex difference in 2D:4D was in general small and did not always reach statistical significance, although the direction (female > male) was always the same [50, 51]. A longitudinal study of 0-2-year-olds (age: 2 weeks, 12 months, 24 months) reported a significant sex effect for 2D:4D measured two weeks after birth as well as a decrease in 2D:4D in the first year and an increase in the second year of life [53]. Interestingly, at 12 and 24 months there were no significant sex differences in 2D:4D. However, cross-sectional studies revealed a sex difference as well as an increase in 2D:4D for children between the ages of 2 and 5 years [54] and in a cohort aged between 5 and 17 years [55].

Manning and Fink [56] have promoted the assumption that 2D:4D remains independent of skeletal growth, with an earlier cross-sectional study [1] reporting a stable sex difference and no significant age effects between cohorts of 2-25-year-old subjects. However, the majority of longitudinal as well as cross-sectional studies do not indicate the temporal stability of 2D:4D in younger cohorts (especially under the age of two), which may be affected by skeletal and overall growth. Studies with older cohorts reveal greater and more stable sex differences (specifically adult and adolescent samples [8]), in a timeframe where the finger lengths remain more or less the same [57].

It becomes apparent that the available literature does not allow precise conclusions on the hypothesized validity and reliability of 2D:4D as a marker of prenatal androgen action. Nonetheless, 2D:4D remains a popular tool used by researchers attempting to investigate the effects of prenatal androgen action. For these reasons, the current study examines amniocentesis data of pregnant women and their offspring’s 2D:4D in a longitudinal design. This approach provides an opportunity to test whether 2D:4D is a valid proxy for prenatal T levels, and further to test its temporal stability within a young cohort. Amniotic fluid was obtained between gestational weeks 14 and 18 in a sample of 244 pregnant women in Germany. The overall levels of amniotic T and E were determined, and the T/E ratio was calculated for further analysis. Unfortunately, data were only available for the male subsample as female T levels were under the limit of detection or quantification. Children’s 2D:4D on both hands was measured at 5, 9, 20, 40, and 70 months of age. With respect to the hypothesis that 2D:4D acts as a marker for prenatal androgen exposure, negative correlations between amniotic T as well as the T/E ratio and 2D:4D were predicted for the male subsample at each time point. As other studies also analyzed a possible association between amniotic T and T/E ratio with D[r-l], this additional variable was included. We also predicted that 2D:4D would show moderate to high correlations between the measurement time points, as well as a stable sex difference (females > males for all time points). Further, based on the available literature, an increase of 2D:4D with age was expected.

Methods

Participants

Two-hundred forty-four mothers who underwent amniocentesis at a practice of gynecologists and human genetics (Praenatal.de) in Düsseldorf, Germany, were recruited between 2010 and 2012. The children were born between January 2011 and February 2013 and were invited to the Department of Experimental Psychology at the University of Düsseldorf, Germany, on five occasions. More specifically, participants took part in follow-up research at the ages of 5 months (T1), 9 months (T2), 20 months (T3), 40 months (T4), and 70 months (T5) as part of another study that also obtained behavioral data published elsewhere [58, 59]. Although every initially recruited family was invited to participate, not every family took part in all consecutive measurement time points resulting in different sample sizes (for mean ages, sample sizes and sex distributions, see Table 1). Mothers were between 22 and 48 years old (M = 38.13 years, SD = 3.50 years) when they gave birth. All families were white. Of the 244 initially recruited mothers (n = 123 male fetuses and n = 121 female fetuses), n = 231 were used for data analysis, as for n = 2 female samples there was no hormonal data as well as no hand scan for any of the measurement time points, and for n = 5 male and n = 6 female samples there were no corresponding hand scans for any of the measurement time points, leaving n = 118 male samples and n = 113 female samples. Of the 244 amniocentesis examinations, n = 1 male and n = 4 female amniotic samples were missing, however, as hand scans were available, these were used for data analysis. Further, the majority of female fetuses T values (97.4%) were below the limit of detection (0.02 ng/ml) or between the limit of detection and the limit of quantification (0.05 ng/ml) and therefore, the female sample was not included for the respective data analysis including hormonal data.

thumbnail
Table 1. Number and age (in months) of participants at T1-T5.

https://doi.org/10.1371/journal.pone.0282253.t001

For recruitment, mothers gave their written informed consent for the use of the data obtained through amniocentesis as well as for re-contacting for the following measurements. At every following measurement (T1-T5) the parents again gave written informed consent for participation. The longitudinal study was approved by the local Ethics Committee of the Science Faculty of the University of Düsseldorf, Germany.

Materials

Prenatal hormone concentration measurements.

Total prenatal T and E levels (in ng/ml) were determined in amniotic fluid from amniocentesis samples of mothers recruited from Praenatal.de (Düsseldorf, Germany). The amniocentesis examinations were carried out between weeks 14 and 18 of gestation (M = 14.78, SD = 0.84). Amniotic fluid samples were assayed with ultra-performance liquid chromatography and tandem mass spectrometry (described elsewhere [58, 59]). For data analysis, n = 117 male samples could be included.

As discussed by Lutchmaya et al. [43], the ratio between amniotic T and E (amniotic T/E) may have a similar impact on early fetal and postnatal development as amniotic T alone. Therefore, we included this ratio in our analysis. Amniotic T/E was calculated as follows:

2D:4D.

Both hands of the children were scanned using a FIJUTSI fi-60F image scanner at every time point (T1-T5), as hand scans yield both larger sex differences [8] and a higher measurement precision [60, 61] compared to direct measurements. The freeware program Autometric [62] was used to measure the ratio between the second and the fourth digit length (2D:4D). Each digit was measured as the length of the midpoint of the ventral proximal crease to the tip in pixels using a 100-dpi monitor (100 pixels = 2.54 cm). Scans in which the fingertips or the ventral creases were not distinct were excluded as well as scans that could not be used for calculation of 2D:4D (e.g. only the length of the second finger was measurable but not of the fourth finger). Two raters, each blind to the sex of the children, measured all hand scans (max. 10 per child). Intraclass correlations indicated high inter-rater reliabilities (all ICCs > .90). The measurements of the two raters were averaged to increase reliability. 2D:4D was calculated as follows: D[r-l].

Additionally, the measure D[r-l] defined as the difference between right and left 2D:4D was computed:

Statistical analyses

A multilevel linear regression model [63, 64] using the R packages lme4 and lmerTest [6567] with participants as random effects was used to examine the association between amniotic T as well as amniotic T/E and 2D:4D in boys. Age (months since birth, centralized to the participant mean), amniotic T level (logarithmized to achieve normal distribution, centralized to the grand mean) and week of pregnancy the amniocentesis took place (centralized to the grand mean) were entered as predictors of 2D:4D in the first model; age, amniotic T/E level (logarithmized to achieve normal distribution, centralized to the grand mean) and week of pregnancy the amniocentesis took place were entered as predictors in the second model.

The same multilevel linear regression models [63, 64] with participants as random effects were used to examine the association between amniotic T as well as amniotic T/E and D[r-l] in boys. Multilevel linear regression models with participants as random effects were also used to test whether age, sex, and hand were associated with 2D:4D. The advantage of this analysis is that it accounts for participant heterogeneity as well as missing data (distribution of participants for each measurement time point can be seen in Table 1). Age (months since birth, centralized to the participant mean), sex (first coded as boys = 0, girls = 1, then centralized to the grand mean), and hand (first coded as left = 0, right = 1, then centralized to the participant mean) were entered as predictors of 2D:4D. Knickmeyer, Woolson, Hamer, Konneker, and Gilmore [53] also used mixed models in their longitudinal study of 2D:4D. To follow-up on the association between age and 2D:4D, four additional multilevel linear regression models comparing the influence of age between T1 and T2, T2 and T3, T3 and T4, and T4 and T5 were calculated. Lastly, Pearson correlations of 2D:4D between the different time points were conducted and interpreted as follows: small correlation r ≥ .10, moderate correlation r ≥ .30, and large correlation r ≥ .50 [68].

Results

2D:4D and amniotic T and T/E

Boys had mean T levels of M = 0.099 ng/ml (SD = 0.065 ng/ml, range: 0.014–0.339 ng/ml, see Fig 1 for distribution) and mean E levels of M = 0.146 ng/ml (SD = 0.080 ng/ml, range: 0.047–0.496 ng/ml). There were no significant associations between amniotic T or amniotic T/E and 2D:4D, and no interactions with any other factor in boys (controlling for age at measurement of 2D:4D and week of pregnancy of amniocentesis) as indicated by two multilevel linear regression models (see Tables 2 and 3).

thumbnail
Table 2. Multilevel linear model with age, amniotic T and week of pregnancy as predictors, and boys’ 2D:4D as outcome.

https://doi.org/10.1371/journal.pone.0282253.t002

thumbnail
Table 3. Multilevel linear model with age, amniotic T/E and week of pregnancy as predictors, and boys’ 2D:4D as outcome.

https://doi.org/10.1371/journal.pone.0282253.t003

D[r-l] and amniotic T and T/E

There were no significant associations between amniotic T or amniotic T/E and D[r-l], and no interactions with any other factor in boys (controlling for age at measurement of D[r-l] and week of pregnancy of amniocentesis) as indicated by two multilevel linear regression models (see Tables 4 and 5).

thumbnail
Table 4. Multilevel linear model with age, amniotic T and week of pregnancy as predictors, and boys’ D[r-l] as outcome.

https://doi.org/10.1371/journal.pone.0282253.t004

thumbnail
Table 5. Multilevel linear model with age, amniotic T/E and week of pregnancy as predictors, and boys’ D[r-l] as outcome.

https://doi.org/10.1371/journal.pone.0282253.t005

Temporal stability of 2D:4D

Girls had larger 2D:4D compared to boys at every measurement time point. On a descriptive level, boys’ 2D:4D decreased from T1 to T2 and increased from T2 to T3, T3 to T4 and T4 to T5; in girls, right 2D:4D increased whereas left 2D:4D decreased from T1 to T2, 2D:4D of both hands decreased from T2 to T3, and showed an increase from T3 to T4 and T4 to T5 (see Table 6 and Fig 2). The multilevel linear regression model with age, sex, and hand as predictors of 2D:4D (see Table 7) showed a significant main effect of sex. Moreover, there was a significant main effect of age indicating that 2D:4D increased with age. The interaction between sex and age was marginally significant (p = .050) and the three-way interaction between sex, age, and hand was significant (p = .036). Follow-up analyses on these interactions revealed that a) the effects of age and sex were significant in both hands (all p < .001), b) in the right hand, the effect of age was stronger in boys compared to girls (p = .006), and c) in the left hand, the effect of age was of equal size for boys and girls (p = .761). There was no main effect of hand, no interaction between sex and hand, and no interaction between age and hand. The additional multilevel linear regression models to follow up on the effect of age showed that age was marginally associated with 2D:4D between T1 and T2, but not between T2 and T3. There were significant age effects between T3 and T4, as well as between T4 and T5 (see Table 8). Table 9 indicates that the sex effect remained stable for each measurement time point separately. There was no main effect for hand and no significant interaction between hand and sex for each measurement point separately (see Table 9).

thumbnail
Fig 2. Means and Standard Deviations of (a) Right and (b) Left 2D:4D for every Time Point in Girls and Boys.

https://doi.org/10.1371/journal.pone.0282253.g002

thumbnail
Table 6. Sample sizes, means and standard deviations for right and left 2D:4D at T1-T5.

https://doi.org/10.1371/journal.pone.0282253.t006

thumbnail
Table 7. Multilevel linear model with sex, age and hand as predictors, and 2D:4D as outcome.

https://doi.org/10.1371/journal.pone.0282253.t007

thumbnail
Table 8. Multilevel linear model with age as predictor and 2D:4D as outcome.

https://doi.org/10.1371/journal.pone.0282253.t008

thumbnail
Table 9. Multilevel linear model with hand and sex as predictors and 2D:4D as outcome for each measurement time point.

https://doi.org/10.1371/journal.pone.0282253.t009

Regarding the reliability of 2D:4D between measurement time points, Pearson correlations revealed significant positive correlations for the right hand between T1 and T2, r = .42, p = .001, between T2 and T3, r = .29, p = .037, between T3 and T4, r = .50, p < .001, and between T4 and T5, r = .66, p < .001. For the left hand, significant positive correlations could be found between T1 and T2, r = .47, p < .001, between T2 and T3, r = .30, p = .023, between T3 and T4, r = .39, p = .001, and between T4 and T5, r = .61, p < .001.

Discussion

The current study used a longitudinal design and had two main aims: (1) to determine whether 2D:4D is associated with prenatal androgen (T and T/E ratio) concentrations measured from amniotic fluid obtained between 14 and 18 weeks of gestation, and (2) to examine the stability of sex differences in 2D:4D at 5, 9, 20, 40, and 70 months of age. There was no significant correlation between 2D:4D respectively D[r-l] and either amniotic T or the T/E ratio in boys for any time point. For girls, amniotic T and T/E ratio could not be determined and therefore no analysis was performed. At each time point, a consistent sex effect was revealed with girls showing larger 2D:4D ratios than boys. Age was associated with 2D:4D, with post-hoc-analyses revealing a marginally significant decrease in 2D:4D from 5 to 9 months of age and statistically significant increases in 2D:4D from 20 to 40 and from 40 to 70 months of age. Significant positive correlations for 2D:4D between different time points suggest a moderate level of stability during infancy and childhood. However, when taken together, the results of this study indicate that 2D:4D exhibits considerable lability during infancy and also question the validity of 2D:4D as a marker of prenatal androgen exposure.

The current state of research does not allow precise conclusions about the hypothesis that 2D:4D is a marker of prenatal androgen action, with heterogeneity in study designs and differing results providing an inconsistent picture. In our study, we did not find any associations between 2D:4D respectively D[r-l] and either amniotic T or T/E in boys for any of the postnatal time points. However, the absence of female amniocentesis data is especially unfavorable. This is because it precluded the opportunity to try to replicate the finding of Ventura et al. [48] of a significant negative association between T levels and left 2D:4D in girls (but not boys). Ventura et al. [48] hypothesized that because boys are already exposed to higher prenatal T levels than girls, increased T above the average might not have a similar impact on male compared to female 2D:4D, and so may potentially account for the non-significant correlation in boys. On the other hand, Lutchmaya et al. [43] found a statistically significant negative association between the T/E ratio and 2D:4D that remained even when controlling for sex. The attempted replication of Lutchmaya et al. [43] by Richards et al. [44] did not find any of the hypothesized correlations for either sex. This was despite the higher statistical power associated with there being more than twice as many mother-child-dyads as in the original study. These differences in findings may stem from a bias concerning the reliability of data obtained through amniocentesis. Notably, amniocentesis samples reflect only a small time window during prenatal development. Lutchmaya et al. [43] reported that the examinations took place during the second trimester of pregnancy without specifying the mean and range of gestational weeks. Other studies in this area examined amniocentesis samples collected between 15 and 22 weeks [44], 16 and 24 weeks (M = 17.2 [48]), and 14 and 18 weeks (M = 14.78; current study). This may account for the differing results, as literature suggests considerable fluctuations of T during the second trimester as well as a peak around week 17 [35, 36]. A more recent study of Kuijper, Ket, Caanen, and Lambalk [69] reports that T levels seem to increase between weeks 14 and 22, with no clear peak for boys and a peak around week 18 for girls. Due to risks and ethical considerations, repeated sampling of amniotic fluid in humans is not possible. Nevertheless, amniocentesis data may allow the most precise information available to assess the hormonal intrauterine environment of a developing human fetus [33]. As Hollier, Keelan, Hickey, Maybery, and Whitehouse [70] discussed, to date there is no “gold standard” to examine hormone levels to which the fetus is exposed during pregnancy. Other studies have revealed inconsistent and mostly non-significant associations between children’s 2D:4D and alternative measures like maternal plasma or cord blood samples to assess T levels in pregnant women [7174].

Another point to consider is that prenatal androgens are unlikely to be the only determining factor for 2D:4D as the sex difference in prenatal androgen levels is much larger than the sex difference observed for 2D:4D [11, 75]. Furthermore, studies of individuals with CAIS show differences in 2D:4D despite T having no physiologic effect [11, 12]. However, Zheng and Cohn [76] showed that active signaling and density of androgen and estrogen receptors differ in the second and fourth digit, which may lead to the sexually differentiated pattern of 2D:4D (female > male) in a mouse model. The sex difference in 2D:4D in mice was present as early as the 17th day of embryonic life [76]. A direct manipulation of the intrauterine environment of mice showed that an activation or blocking of the androgen receptor resulted in changes of the total paw and digit length as well as 2D:4D [77]. That is, an activation of the androgen receptor led to an increase of 2D:4D in males, whereas a blocking of the androgen receptor led to a decrease of 2D:4D in females [77]. This finding is also in line with Swift-Gallant, Di Rita, Coome, and Monks [78] who found an increase in 2D:4D in mice with global androgen receptor overexpression compared to wildtype mice and mice with neural-specific androgen receptor overexpression regardless of sex. Other studies administered T directly as an intramuscular injection into the dam and examined 2D:4D in the offspring. Talarovičová, Kršková, and Blažeková [79] found a significant effect of T administration on the length of the second and fourth digit as well as on 2D:4D. Administration of T, compared to sesame oil as a control, led to a lengthening of the fourth digit, a shortening of the second digit, and a smaller 2D:4D. Contrary, another study found shortened second and fourth digits in the offspring after T (respectively T combined with flutamide) administration compared to a control group that was supplemented with olive oil only, however, no effect of T administration on 2D:4D [80]. These findings indeed suggest an influence of androgens on digit lengths and 2D:4D but also reveal the complex interplay of hormones, chromosomes, and gonads to determine the sex and sex-specific characteristics of organisms and the influence on digit growth.

Despite the current study revealing stable sex differences with girls exhibiting larger 2D:4D compared to boys, 2D:4D appears unable to reliably discriminate between the sexes as the magnitude of 2D:4D in girls and boys reveals a large overlap. This could be shown in various other studies reporting the expected pattern of 2D:4D, however, without significant sex differences and/or only small effect sizes [5053, 81]. Furthermore, girls show considerable variability in 2D:4D yet relative homogeneity in amniotic T levels [36, 44]; this was also true for the current study, as 97.4% of girls’ T levels were under the limit of detection or quantification. Another observation that questions the validity of 2D:4D is its association with age in young cohorts, an effect reported not only in the current study but also in various others within the literature [5155, 81, 82]. The significant effect of age indicates that 2D:4D continues to develop during childhood. Importantly, the reliability of 2D:4D at this time is also questionable, as the current study observed only moderately sized correlations between time points. Likewise, the association between age and 2D:4D makes the results of Lutchmaya et al. [43] and Richards et al. [44] difficult to compare, as 2D:4D was measured at 2 and 4.5 years of age respectively.

Notably, research does not support reliable sex differences of 2D:4D in younger cohorts, as noted in another study of our group [49], and it appears that sex differences become larger and more stable during adolescence and adulthood [8, 20, 5052]. Prepubertal girls and boys show nearly the same concentrations of fluctuating sex hormones [83]. There is only a short period of time during the first 6 months after birth in which specifically T levels increase in boys, whereas the levels do not differ to a greater extent between girls and boys until the onset of puberty [84, 85]. A comparison of characteristics that can be associated with the influence of prenatal sex hormones should be highly reliable between these two time points—birth and puberty. However, studies failed to replicate a stable sex difference in young cohorts [56, 86], questioning the assumption that prenatal amniotic T levels explain a significant proportion of variance in 2D:4D. The measurement of digit lengths in young cohorts compared to adult cohorts may face other methodological issues that challenge the reliability and validity of the measurement itself. Young infants show a much higher proportion of body fat tissue compared to adults [87], which may specifically impair indirect measurement techniques like hand scans, as hand creases tend to shift more in between different measurements when pressed onto the scanner glass. Furthermore, very young children and infants may be less likely to cooperate in the measurement procedure like adults normally do. To control for this, another more direct measurement technique may be a better approach to examine 2D:4D in young cohorts and, compared to indirect measurement techniques, would also yield greater effect sizes [88]. In fact, McIntyre, Ellison, Lieberman, Demerath, and Towne [51] and McIntyre, Cohn, and Ellison [52] used radiographs in their cohorts aged 1 month up to 18 years. However, they found only low to moderate differences in 2D:4D between males and females with more stable differences emerging with age. The lack of comparability between different studies in terms of measurement time point and measurement technique impairs the interpretation of different results specifically in prepubertal cohorts.

Lastly, there are some limitation of the current study that should be addressed. It has already been discussed that the timing of the amniocentesis examination may account for different results between studies. Further, there is a relative large interval between human limb development and amniocentesis examinations as the human limb development starts much earlier, approximately during week 4 of gestation and by the end of week 8 of gestation the upper and lower limbs are already mostly developed [89, 90], compared to amniocenteses which are usually performed around week 15+0 of gestation [91]. Therefore, prenatal hormone levels obtained through amniotic sampling may not fully reflect the intrauterine environment during limb development. Additionally, amniotic fluid only reflects the hormonal environment the fetus is subjected to, not the actual hormonal concentrations of the fetus [69]. Nevertheless, amniotic fluid sampling may be the best method available to obtain information about fetal hormonal environment, due to obvious ethical considerations. Also the 2D:4D digit ratio is facing important influencing factors like the number and order of siblings and their sex (however, with contradictory results, see Králík, Hupková, Zeman, Hložek, Hlaváček, Slováčková, et al. [92] and Saino, Leoni, and Romano [93]). Further, as it is hypothesized that 2D:4D differs between different ethnic groups [94, 95], the current results are not generalizable as they only refer to the German and solely white population.

The motivation to obtain reliable and valid information about the intrauterine hormonal environment by simply measuring digit lengths is understandable considering the difficulties associated with obtaining more direct measurements. The literature does provide some evidence for 2D:4D as it exhibits a moderate sized sex difference as well as associations with other sexually differentiated characteristics. However, the evidence that 2D:4D reflects prenatal androgen action is more limited and a comparison of available studies is difficult. Therefore, 2D:4D does not seem to be a feasible proxy for intrauterine androgen exposure. Although the sex difference in 2D:4D shows a good replicability, especially in adult samples, the hypothetical fundament of its origins is not sufficiently examined. If 2D:4D shall be used as a marker for prenatal androgen action it is necessary to examine the developmental basis. Further, instead of conducting novel research on associations between 2D:4D and various other characteristics, research should focus on replications with highly comparable study designs and on testing the underlying hypothesis that 2D:4D is a valid marker of intrauterine androgen action.

Supporting information

S1 Fig. Data inclusion and exclusion chart.

https://doi.org/10.1371/journal.pone.0282253.s001

(TIF)

References

  1. 1. Manning J, Scutt D, Wilson J, Lewis-Jones D. The Ratio of 2nd to 4th Digit Length: A Predictor of Sperm Numbers and Concentrations of Testosterone, Luteinizing Hormone and Oestrogen. Hum Reprod. 1998;13:3000–4. pmid:9853845
  2. 2. Breedlove SM. Minireview: Organizational hypothesis: instances of the fingerpost. Endocrinology. 2010;151(9):4116–22. pmid:20631003
  3. 3. Valla J, Ceci SJ. Can Sex Differences in Science Be Tied to the Long Reach of Prenatal Hormones? Brain Organization Theory, Digit Ratio (2D/4D), and Sex Differences in Preferences and Cognition. Perspect Psychol Sci. 2011;6(2):134–6. pmid:22164187
  4. 4. Wang Y, Wang HL, Li YH, Zhu FL, Li SJ, Ni H. Using 2D: 4D digit ratios to determine motor skills in children. European review for medical and pharmacological sciences. 2016;20(5):806–9. pmid:27010133
  5. 5. Pasanen BE, Tomkinson JM, Dufner TJ, Park CW, Fitzgerald JS, Tomkinson GR. The relationship between digit ratio (2D:4D) and muscular fitness: A systematic review and meta-analysis. Am J Hum Biol. 2021:e23657. pmid:34331730
  6. 6. Siegmann EM, Müller T, Dziadeck I, Mühle C, Lenz B, Kornhuber J. Digit ratio (2D:4D) and transgender identity: new original data and a meta-analysis. Scientific reports. 2020;10(1):19326. pmid:33168880
  7. 7. Cohen-Bendahan CC, van de Beek C, Berenbaum SA. Prenatal sex hormone effects on child and adult sex-typed behavior: methods and findings. Neurosci Biobehav Rev. 2005;29(2):353–84. pmid:15811504
  8. 8. Hönekopp J, Watson S. Meta-analysis of digit ratio 2D:4D shows greater sex difference in the right hand. Am J Hum Biol. 2010;22(5):619–30. pmid:20737609
  9. 9. Galis F, Ten Broek CM, Van Dongen S, Wijnaendts LC. Sexual dimorphism in the prenatal digit ratio (2D:4D). Arch Sex Behav. 2010;39(1):57–62. pmid:19301112
  10. 10. Malas MA, Dogan S, Evcil EH, Desdicioglu K. Fetal development of the hand, digits and digit ratio (2D:4D). Early Hum Dev. 2006;82(7):469–75. pmid:16473482
  11. 11. Berenbaum SA, Bryk KK, Nowak N, Quigley CA, Moffat S. Fingers as a marker of prenatal androgen exposure. Endocrinology. 2009;150(11):5119–24. pmid:19819951
  12. 12. van Hemmen J, Cohen-Kettenis PT, Steensma TD, Veltman DJ, Bakker J. Do sex differences in CEOAEs and 2D:4D ratios reflect androgen exposure? A study in women with complete androgen insensitivity syndrome. Biol Sex Differ. 2017;8:11. pmid:28413602
  13. 13. Wallen K. Does Finger Fat Produce Sex Differences in Second to Fourth Digit Ratios? Endocrinology. 2009;150(11):4819–22. pmid:19846613
  14. 14. Brown WM, Hines M, Fane BA, Breedlove SM. Masculinized finger length patterns in human males and females with congenital adrenal hyperplasia. Horm Behav. 2002;42(4):380–6. pmid:12488105
  15. 15. Ökten A, Kalyoncu M, Yariş N. The ratio of second- and fourth-digit lengths and congenital adrenal hyperplasia due to 21-hydroxylase deficiency. Early Hum Dev. 2002;70(1–2):47–54. pmid:12441204
  16. 16. Richards G, Browne WV, Aydin E, Constantinescu M, Nave G, Kim MS, et al. Digit ratio (2D:4D) and congenital adrenal hyperplasia (CAH): Systematic literature review and meta-analysis. Horm Behav. 2020;126:104867. pmid:32998030
  17. 17. Buck JJ, Williams RM, Hughes IA, Acerini CL. In-utero androgen exposure and 2nd to 4th digit length radio. Hum Reprod. 2003;18(5):976–9.
  18. 18. Chang S, Skakkebæk A, Trolle C, Bojesen A, Hertz JM, Cohen A, et al. Anthropometry in Klinefelter Syndrome—Multifactorial Influences Due to CAG Length, Testosterone Treatment and Possibly Intrauterine Hypogonadism. The Journal of Clinical Endocrinology & Metabolism. 2015;100(3):E508–E17. pmid:25514102
  19. 19. Manning JT, Kilduff LP, Trivers R. Digit ratio (2D:4D) in Klinefelter’s syndrome. Andrology. 2013;1(1):94–9. pmid:23258636
  20. 20. Wong WI, Hines M. Interpreting digit ratio (2D:4D)-behavior correlations: 2D:4D sex difference, stability, and behavioral correlates and their replicability in young children. Horm Behav. 2016;78:86–94. pmid:26542674
  21. 21. Hönekopp J. Digit ratio 2D:4D in relation to autism spectrum disorders, empathizing, and systemizing: a quantitative review. Autism research: official journal of the International Society for Autism Research. 2012;5(4):221–30. pmid:22674640
  22. 22. Fink B, Manning JT, Williams JHG, Podmore-Nappin C. The 2nd to 4th digit ratio and developmental psychopathology in school-aged children. Pers Individ Dif. 2007;42(2):369–79.
  23. 23. Bunevicius A. The Association of Digit Ratio (2D : 4D) with Cancer: A Systematic Review and Meta-Analysis. Disease markers. 2018;2018:7698193.
  24. 24. Fonseca CAD, Martelli DRB, Maia CMF, Dias VO, Carvalho AA, Júnior HM. Digital biomarker 2D:4D as a predictor of cancer: A systematic review. Early Hum Dev. 2022;164:105521. pmid:34922146
  25. 25. Smoliga JM, Fogaca LK, Siplon JS, Goldburt AA, Jakobs F. Giving science the finger-is the second-to-fourth digit ratio (2D:4D) a biomarker of good luck? A cross sectional study. BMJ (Clinical research ed). 2021;375:e067849. pmid:34911738
  26. 26. Richards G. What is the evidence for a link between digit ratio (2D:4D) and direct measures of prenatal sex hormones? Early Hum Dev. 2017;113:71–2. pmid:28807670
  27. 27. Pruszkowska-Przybylska P, Sitek A, Rosset I, Sobalska-Kwapis M, Słomka M, Strapagiel D, et al. Association of the 2D:4D digit ratio with body composition among the Polish children aged 6–13 years. Early Hum Dev. 2018;124:26–32.
  28. 28. Manning JT, Fink B, Mason L, Trivers R. Parental income inequality and children’s digit ratio (2D:4D): a ’Trivers-Willard’ effect on prenatal androgenization? Journal of biosocial science. 2022;54(1):154–62. pmid:33557976
  29. 29. Swift-Gallant A, Di Rita V, Major CA, Breedlove CJ, Jordan CL, Breedlove SM. Differences in digit ratios between gay men who prefer receptive versus insertive sex roles indicate a role for prenatal androgen. Scientific reports. 2021;11(1):8102. pmid:33854100
  30. 30. Pruszkowska-Przybylska P, Sitek A, Rosset I, Sobalska-Kwapis M, Słomka M, Strapagiel D, et al. Associations between second to fourth digit ratio, cortisol, vitamin D, and body composition among Polish children. Scientific reports. 2021;11(1):7029. pmid:33782473
  31. 31. Pruszkowska-Przybylska P, Kobus M, Iljin A, Wiktorska JA, Żądzińska E, Sitek A. Thyroid diseases and second to fourth digit ratio in Polish adults. Scientific reports. 2021;11(1):18979. pmid:34556783
  32. 32. Kobus M, Sitek A, Antoszewski B, Rożniecki J, Pełka J, Żądzińska E. Prenatal oestrogen-testosterone balance as a risk factor of migraine in adults. The journal of headache and pain. 2021;22(1):119. pmid:34620097
  33. 33. van de Beek C, Thijssen JH, Cohen-Kettenis PT, van Goozen SH, Buitelaar JK. Relationships between sex hormones assessed in amniotic fluid, and maternal and umbilical cord serum: what is the best source of information to investigate the effects of fetal hormonal exposure? Horm Behav. 2004;46(5):663–9. pmid:15555509
  34. 34. Knickmeyer RC, Baron-Cohen S. Fetal testosterone and sex differences. Early Hum Dev. 2006;82(12):755–60. pmid:17084045
  35. 35. Judd HL, Robinson JD, Young PE, Jones OW. Amniotic fluid testosterone levels in midpregnancy. Obstetrics and gynecology. 1976;48(6):690–2. pmid:999757
  36. 36. Robinson JD, Judd HL, Young PE, Jones OW, Yen SS. Amniotic fluid androgens and estrogens in midgestation. The Journal of clinical endocrinology and metabolism. 1977;45(4):755–61. pmid:144143
  37. 37. Abramovich DR. Human sexual differentiation—in utero influences. The Journal of obstetrics and gynaecology of the British Commonwealth. 1974;81(6):448–53. pmid:4407079
  38. 38. Finegan JA, Bartleman B, Wong PY. A window for the study of prenatal sex hormone influences on postnatal development. The Journal of genetic psychology. 1989;150(1):101–12. pmid:2496195
  39. 39. Di Mascio D, Khalil A, Rizzo G, Buca D, Liberati M, Martellucci CA, et al. Risk of fetal loss following amniocentesis or chorionic villus sampling in twin pregnancy: systematic review and meta-analysis. Ultrasound in Obstetrics & Gynecology. 2020;56(5):647–55. pmid:32632979
  40. 40. Akolekar R, Beta J, Picciarelli G, Ogilvie C, D’Antonio F. Procedure-related risk of miscarriage following amniocentesis and chorionic villus sampling: a systematic review and meta-analysis. Ultrasound in Obstetrics & Gynecology. 2015;45(1):16–26. pmid:25042845
  41. 41. Nicolaides KH. First-trimester screening for chromosomal abnormalities. Seminars in perinatology. 2005;29(4):190–4. pmid:16104667
  42. 42. Constantinescu M, Hines M. Relating Prenatal Testosterone Exposure to Postnatal Behavior in Typically Developing Children: Methods and Findings. Child Dev Perspect. 2012:n/a-n/a.
  43. 43. Lutchmaya S, Baron-Cohen S, Raggatt P, Knickmeyer R, Manning JT. 2nd to 4th digit ratios, fetal testosterone and estradiol. Early Hum Dev. 2004;77(1–2):23–8. pmid:15113628
  44. 44. Richards G, Browne W, Constantinescu M. Digit ratio (2D:4D) and amniotic testosterone and estradiol: An attempted replication of Lutchmaya et al. (2004). J Dev Orig Health Dis. 2020.
  45. 45. Manning JT. Digit Ratio: A Pointer to Fertility, Behavior, and Health. New Brunswick, New Jersey, London: Rutgers University Press; 2002.
  46. 46. Manning JT, Fink B, Mason L, Kasielska-Trojan A, Trivers R. The effects of sex, nation, ethnicity, age and self-reported pubertal development on participant-measured right-left 2D:4D (Dr-l) in the BBC internet study. Journal of biosocial science. 2022:1–13. pmid:35088686
  47. 47. Manning J, Kilduff L, Cook C, Crewther B, Fink B. Digit Ratio (2D:4D): A Biomarker for Prenatal Sex Steroids and Adult Sex Steroids in Challenge Situations. Frontiers in endocrinology. 2014;5. pmid:24523714
  48. 48. Ventura T, Gomes MC, Pita A, Neto MT, Taylor A. Digit ratio (2D:4D) in newborns: influences of prenatal testosterone and maternal environment. Early Hum Dev. 2013;89(2):107–12. pmid:23017880
  49. 49. Ernsten L, Körner LM, Heil M, Richards G, Schaal NK. Investigating the reliability and sex differences of digit lengths, ratios, and hand measures in infants. Scientific reports. 2021;11(1):10998. pmid:34040007
  50. 50. Trivers R, Manning J, Jacobson A. A longitudinal study of digit ratio (2D:4D) and other finger ratios in Jamaican children. Horm Behav. 2006;49:150–6. pmid:16040033
  51. 51. McIntyre MH, Ellison PT, Lieberman DE, Demerath E, Towne B. The development of sex differences in digital formula from infancy in the Fels Longitudinal Study. Proc R Soc Lond B Biol Sci. 2005;272(1571):1473–9. pmid:16011922
  52. 52. McIntyre MH, Cohn BA, Ellison PT. Sex dimorphism in digital formulae of children. Am J Phys Anthropol. 2006;129(1):143–50. pmid:16224778
  53. 53. Knickmeyer RC, Woolson S, Hamer RM, Konneker T, Gilmore JH. 2D:4D ratios in the first 2 years of life: Stability and relation to testosterone exposure and sensitivity. Horm Behav. 2011;60(3):256–63. pmid:21664359
  54. 54. Williams JHG, Greenhalgh KD, Manning JT. Second to fourth finger ratio and possible precursors of developmental psychopathology in preschool children. Early Hum Dev. 2003;72(1):57–65. pmid:12706312
  55. 55. Richards G, Bellin W, Davies W. Familial digit ratio (2D:4D) associations in a general population sample from Wales. Early Hum Dev. 2017;112:14–9. pmid:28668648
  56. 56. Manning JT, Fink B. Sexual dimorphism in the ontogeny of second (2D) and fourth (4D) digit lengths, and digit ratio (2D:4D). Am J Hum Biol. 2018;30(4):e23138. pmid:30133905
  57. 57. Matzdorff I. Metrische und allometrische Wachstumsuntersuchungen an der menschlichen Hand. Z Morphol Anthropol. 1967;59(2):158–84.
  58. 58. Körner LM, Pause BM, Meinlschmidt G, Tegethoff M, Frohlich S, Kozlowski P, et al. Prenatal testosterone exposure is associated with delay of gratification and attention problems/overactive behavior in 3-year-old boys. Psychoneuroendocrinology. 2019;104:49–54. pmid:30802710
  59. 59. Körner LM, Schaper ML, Pause BM, Heil M. Parent-Reports of Sex-Typed Play Preference in Preschool Children: Relationships to 2D:4D Digit Ratio and Older Siblings’ Sex. Arch Sex Behav. 2020;49(7):2715–24. pmid:32222854
  60. 60. Kemper CJ, Schwerdtfeger A. Comparing indirect methods of digit ratio (2D:4D) measurement. Am J Hum Biol. 2009;21(2):188–91. pmid:18988284
  61. 61. Mikac U, Buško V, Sommer W, Hildebrandt A. Analysis of different sources of measurement error in determining second-to-fourth digit ratio, a potential indicator of perinatal sex hormones exposure. Annu Rev Psychol. 2016;23:39–49.
  62. 62. DeBruine LM. AutoMetric software for measurement of 2D:4D ratios. www.facelab.org/debruine/Programs/autometric.php2004.
  63. 63. Krull J, MacKinnon D. Multilevel Modeling of Individual and Group Level Mediated Effects. Multivariate Behav Res. 2001;36:249–77. pmid:26822111
  64. 64. Kenny DA, Korchmaros JD, Bolger N. Lower level mediation in multilevel models. Psychol Methods. 2003;8(2):115–28. pmid:12924810
  65. 65. Bates D, Mächler M, Bolker B, Walker S. Fitting Linear Mixed-Effects Models Using lme4. J Stat Softw. 2015;67(1):1–48.
  66. 66. Kuznetsova A, Brockhoff PB, Christensen RHB. lmerTest: Tests for random and fixed effects for linear mixed effect models [Software]. http://cran.r-project.org/package=lmerTest 2014.
  67. 67. R Core Team. R: A language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing; 2019.
  68. 68. Cohen J. Statistical Power Analysis for the Behavioral Sciences. 2 ed: Lawrence Erlbaum Associates; 1988.
  69. 69. Kuijper EAM, Ket JCF, Caanen MR, Lambalk CB. Reproductive hormone concentrations in pregnancy and neonates: a systematic review. Reprod Biomed Online. 2013;27(1):33–63. pmid:23669015
  70. 70. Hollier LP, Keelan JA, Hickey M, Maybery MT, Whitehouse AJO. Measurement of Androgen and Estrogen Concentrations in Cord Blood: Accuracy, Biological Interpretation, and Applications to Understanding Human Behavioral Development. Frontiers in endocrinology. 2014;5(64). pmid:24829559
  71. 71. Hickey M, Doherty DA, Hart R, Norman RJ, Mattes E, Atkinson HC, et al. Maternal and umbilical cord androgen concentrations do not predict digit ratio (2D:4D) in girls: A prospective cohort study. Psychoneuroendocrinology. 2010;35(8):1235–44. pmid:20299156
  72. 72. Barrett E, Thurston SW, Harrington D, Bush NR, Sathyanarayana S, Nguyen R, et al. Digit ratio, a proposed marker of the prenatal hormone environment, is not associated with prenatal sex steroids, anogenital distance, or gender-typed play behavior in preschool age children. J Dev Orig Health Dis. 2020:1–10.
  73. 73. Richards G, Gomes M, Ventura T. Testosterone measured from amniotic fluid and maternal plasma shows no significant association with directional asymmetry in newborn digit ratio (2D:4D). J Dev Orig Health Dis. 2019;10(3):362–7. pmid:30376903
  74. 74. Richards G, Aydin E, Tsompanidis A, Padaigaitė E, Austin T, Allison C, et al. Digit ratio (2D:4D) and maternal testosterone-to-estradiol ratio measured in early pregnancy. Scientific reports. 2022;12(1):13586. pmid:35945232
  75. 75. Swift-Gallant A, Johnson BA, Di Rita V, Breedlove SM. Through a glass, darkly: Human digit ratios reflect prenatal androgens, imperfectly. Horm Behav. 2020;120:104686. pmid:32014464
  76. 76. Zheng Z, Cohn MJ. Developmental basis of sexually dimorphic digit ratios. Proc Natl Acad Sci U S A. 2011;108(39):16289–94. pmid:21896736
  77. 77. Huber SE, Lenz B, Kornhuber J, Müller CP. Prenatal androgen-receptor activity has organizational morphological effects in mice. PloS one. 2017;12(11):e0188752. pmid:29176856
  78. 78. Swift-Gallant A, Di Rita V, Coome LA, Monks DA. Overexpression of Androgen Receptors Masculinizes 2D:4D Digit Ratios in Mice. Arch Sex Behav. 2021. pmid:34625883
  79. 79. Talarovičová A, Kršková L, Blažeková J. Testosterone enhancement during pregnancy influences the 2D:4D ratio and open field motor activity of rat siblings in adulthood. Horm Behav. 2009;55(1):235–9. pmid:19022257
  80. 80. Suchonova M, Borbelyova V, Renczes E, Konecna B, Vlkova B, Hodosy J, et al. Does the 2nd and 4th digit ratio reflect prenatal androgen exposure? Bratislavske lekarske listy. 2019;120(9):703–10. pmid:31475559
  81. 81. McIntyre MH. The use of digit ratios as markers for perinatal androgen action. Reproductive biology and endocrinology: RB&E. 2006;4:10. pmid:16504142
  82. 82. Manning JT. Sex Differences and Age Changes in Digit Ratios: Implications for the Use of Digit Ratios in Medicine and Biology. In: Preedy VR, editor. Handbook of Anthropometry: Physical Measures of Human Form in Health and Disease. New York, NY: Springer New York; 2012. p. 841–51.
  83. 83. Berenbaum SA, Beltz AM. Sexual differentiation of human behavior: effects of prenatal and pubertal organizational hormones. Front Neuroendocrinol. 2011;32(2):183–200. pmid:21397624
  84. 84. Auyeung B, Lombardo MV, Baron-Cohen S. Prenatal and postnatal hormone effects on the human brain and cognition. Pflugers Arch. 2013;465(5):557–71. pmid:23588379
  85. 85. Lamminmaki A, Hines M, Kuiri-Hanninen T, Kilpelainen L, Dunkel L, Sankilampi U. Testosterone measured in infancy predicts subsequent sex-typed behavior in boys and in girls. Horm Behav. 2012;61(4):611–6. pmid:22373494
  86. 86. Manning JT. Digit ratio (2D:4D), sex differences, allometry, and finger length of 12-30-year olds: evidence from the British Broadcasting Corporation (BBC) Internet study. Am J Hum Biol. 2010;22(5):604–8. pmid:20737606
  87. 87. Kuzawa CW. Adipose tissue in human infancy and childhood: An evolutionary perspective. Am J Phys Anthropol. 1998;107(S27):177–209. pmid:9881526
  88. 88. Ribeiro E, Neave N, Morais RN, Manning JT. Direct Versus Indirect Measurement of Digit Ratio (2D:4D): A Critical Review of the Literature and New Data. Evolutionary psychology: an international journal of evolutionary approaches to psychology and behavior. 2016;14(1):1474704916632536.
  89. 89. Burton L, Carachi R. The Limbs. In: Carachi R, Helmi S, Doss E, editors. Clinical Embryology: Springer Cham; 2019. p. 449–61.
  90. 90. Doss SH, Sneddon SF. General Embryology. In: Carachi R, Helmi S, Doss E, editors. Clinical Embryology: Springer Cham; 2019. p. 11–30.
  91. 91. Navaratnam K, Alfirevic Z, Obstetricians tRCo, Gynaecologists. Amniocentesis and chorionic villus sampling. BJOG: An International Journal of Obstetrics & Gynaecology. 2022;129(1):e1–e15.
  92. 92. Králík M, Hupková A, Zeman T, Hložek M, Hlaváček L, Slováčková M, et al. Family effects on the digit ratio (2D:4D): The role of the interbirth interval. American Journal of Human Biology. 2019;31(4):e23260. pmid:31183942
  93. 93. Saino N, Leoni B, Romano M. Human digit ratios depend on birth order and sex of older siblings and predict maternal fecundity. Behavioral Ecology and Sociobiology. 2006;60(1):34–45.
  94. 94. Manning JT, Churchill AJG, Peters M. The Effects of Sex, Ethnicity, and Sexual Orientation on Self-Measured Digit Ratio (2D:4D). Archives of sexual behavior. 2007;36(2):223–33. pmid:17373585
  95. 95. Butovskaya M, Burkova V, Apalkova Y, Dronova D, Rostovtseva V, Karelin D, et al. Sex, population origin, age and average digit length as predictors of digit ratio in three large world populations. Scientific reports. 2021;11(1):8157. pmid:33854119