REVIEW
Second to fourth digit length ratio (2D:4D) and adult sex hormone levels: New data and a meta-analytic review

https://doi.org/10.1016/j.psyneuen.2007.01.007Get rights and content

Summary

The relative length of the second (index) to the fourth (ring) finger (2D:4D) is a putative negative correlate of prenatal testosterone (T) exposure. Therefore, 2D:4D (and to a lesser extent Dr−l, the difference between 2D:4D in the right hand and in the left hand) has often been used to study effects of prenatal androgenization on human behavior and cognition. However, evidence suggests that 2D:4D may also be related to levels of circulating sex hormones in adults. This would question the validity of 2D:4D as a means of studying the effects of prenatal sex hormones. Here we present new data from two non-clinical samples (64 women and 102 men) regarding the relationships of 2D:4D and Dr−l with circulating sex hormone levels. We then present a meta-analytic review of all the present evidence regarding this issue. The results suggest that, in the normal population, 2D:4D and Dr−l are not associated with adult sex hormone levels. The findings from this current study add to the growing body of evidence demonstrating that 2D:4D is a suitable tool to study the effects of prenatal androgenization on human behavior and cognition.

Introduction

The ratio of the length of the second (index) to the forth (ring) finger, 2D:4D, is sexually dimorphic in humans. Averaged across samples from various populations, female values were found to be about .25 standard deviations higher than male values (Manning et al., 2000).

Several lines of evidence suggest that fetal sex hormone (especially androgen) levels bring about the sex difference in 2D:4D and also affect intra-sexual variability in this measure. First, the sex difference in 2D:4D is already observable at the end of the first trimester of fetal development (Malas et al., 2006). Second, the development of both, genitals and digits, is controlled by the same genes HoxA and HoxD (Kondo et al., 1997). Third, the sex difference in 2D:4D appears unaffected by puberty as is evidenced by cross-sectional (Manning et al., 1998) and longitudinal (McIntyre et al., 2005; Trivers et al., 2006) data. Fourth, right hand 2D:4D at the age of two years was found to be negatively correlated with the T/estrogen ratio as measured by amniocentesis in the second trimester (Lutchmaya et al., 2004). Fifth, individual 2D:4D values were shown to have high longitudinal stability from age 10 to 14 (Trivers et al., 2006) and some stability from infancy throughout to age 17 (McIntyre et al., 2005). Sixth, children with congenital adrenal hyperplasia, a condition that results in abnormally high T levels during gestation, were found to have lower 2D:4D values than normal controls in two studies (Brown et al., 2002; Ökten et al., 2002). A third study did not find such a difference (Buck et al., 2003), but this study shows some methodological shortcomings (for a thorough critique, see McIntyre, 2006). Seventh, females with male co-twins appear to be exposed to higher T levels in utero and they have lower 2D:4D values than females with female co-twins (van Anders et al., 2006). Eighth, 2D:4D negatively correlates with T sensitivity in androgen receptors as measured by the number of CAG repeats in the androgen receptor gene (Manning et al., 2003). Ninth, autism is believed to result from prenatal hypermasculinization of the brain, and autistic subjects were found to have lower 2D:4D than normal controls (Manning et al., 2001).

For these reasons, 2D:4D may be a valid marker of prenatal T exposure. As prenatal T affects human behavior and cognition and since other ways of studying these effects in humans are laborious and pose various difficulties (Collaer and Hines, 1995; Cohen-Bendahan et al., 2005) 2D:4D has become popular as a means to study the effects of prenatal androgenization in humans, especially regarding sex-linked behaviors and traits (Manning, 2002a). In some areas, findings regarding the relationship between 2D:4D and research variables appear quite diverse. For example, a negative relationship between mental rotation ability and 2D:4D has been found for females but not for males (Burton et al., 2005), for males but nor for females (Sanders et al., 2005), or for neither sex (Alexander, 2006; Coolican and Peters, 2003; Falter et al., 2006). In other areas, research results have been remarkably consistent. For example, all studies focusing at 2D:4D and athletic ability have found negative relationships (Hönekopp et al., 2006b; Manning, 2002a, Manning, 2002b; Pokrywka et al., 2005; Paul et al., 2006). Interestingly, evidence for homologous effects of prenatal steroids as indicated by 2D:4D and of circulating steroids has been found in some fields. For example, not only prenatal T but also circulating T appears to boost sport performance in men (Neave and Wolfson, 2003); and both, low prenatal androgenization as indicated by high 2D:4D values and high levels of circulating estrogen, appear to increase the risk of eating disorders in women (Klump et al., 2006).

In sum, 2D:4D is promising as a marker for studying effects of prenatal androgenization in humans (for a more thorough discussion, see McIntyre, 2006). However, questions remain open. For example, it is unclear why the sex difference in 2D:4D is so low (about .25 standard deviations) whereas the sex difference in prenatal T levels appears to be much higher (d≈1.9; cf. Knickmeyer et al., 2005; van de Beek et al., 2004). This difference in effect sizes suggests that other, yet to be discovered, factors than prenatal T affect 2D:4D and that 2D:4D may not be a very accurate indicator of prenatal T. To clarify the latter point, more longitudinal studies comparing prenatal T with 2D:4D values later in life would be of great value.

Here, we focus on another possible problem of 2D:4D: Potential associations between 2D:4D and levels of circulating sex hormones threaten its validity as a measure of prenatal androgenization as we argue below. Our aim is therefore to clarify the relationship between 2D:4D and circulating sex hormone levels. In order to do so, we present new data regarding associations between 2D:4D and current levels of various sex hormones in women and men. We then review the literature on the relationship between sex hormone levels and 2D:4D, using meta-analytic procedures when appropriate. However, before we turn to the new data and the literature review, we show why potential associations between adult sex hormone levels and 2D:4D would threaten the validity of the latter as a means of studying prenatal androgenization, and we argue why such associations might be expected.

As 2D:4D appears to reflect prenatal T levels, a correlation between 2D:4D and X (the study variable of interest) is usually interpreted as indicating an effect of prenatal T on X (see Fig. 1a). Although this interpretation is plausible, it is not necessarily correct. 2D:4D may not exclusively reflect prenatal androgenization but also adult levels of sex hormones (see below). In this case, an observed relationship between 2D:4D and X may not arise from prenatal T but from activational effects of circulating sex hormones on X, from effects of X on circulating sex hormones, or from bidirectional causation (Fig. 1b). Of course, an association between circulating adult hormones and 2D:4D does not undermine possible prenatal hormonal influences on digit ratios. But in this case, it may be wrong to conclude, on grounds of an observed relationship between the study variable and 2D:4D, that prenatal androgenization affects this variable. It is thus important to clarify the relationship between 2D:4D and adult circulating sex hormone levels, which is our objective.

Is it plausible to assume a correlation between 2D:4D and adult sex hormone titers? A causal effect of adult sex hormones on 2D:4D is unlikely because the onset of puberty does not change the sex difference in 2D:4D (Manning et al., 1998; McIntyre et al., 2005; Trivers et al., 2006). Nonetheless, theoretical considerations make an association plausible. This is, because prenatal exposure to sex hormones, affecting 2D:4D, may be associated with adult circulating sex hormone levels (Fig. 1c). Prenatal T appears to stem from the fetuses and not from their mothers. This is indicated by the fact that amniotic T levels are not correlated with maternal serum T levels (van de Beek et al., 2004). But if prenatal and adult T stem from the same organs their titers may show a positive correlation. This should effect an association between 2D:4D and adult levels of circulating sex hormones (Fig. 1c).

So far, we have focused on a potential association between adult T levels and 2D:4D. For three reasons, however, it is necessary to broaden the scope to other sex hormones. First, 2D:4D may not exclusively reflect prenatal T levels but also prenatal estradiol (Lutchmaya et al., 2004). Second, complex relationships between sex hormones are common. And third, empirical findings suggest an association of 2D:4D with adult FSH and LH levels (see below).

We now turn to the empirical evidence regarding potential associations between 2D:4D and adult sex hormones. Indirect evidence supports the idea of an association between 2D:4D and adult T. Manning et al. (2002) found a negative relationship between parents’ 2D:4D and offspring sex ratio (i.e., proportion of sons). Because parents’ current T levels at the time of conception are positively related to sex ratio, this result suggests a negative association between 2D:4D and adult circulating T levels. Previous direct findings regarding associations between 2D:4D and adult sex hormone levels are listed in detail in Table 3, Table 4. In brief, these findings are incomplete and inconsistent. Relationships between 2D:4D and adult T have been studied best. Conflicting results have been obtained for women (e.g., van Anders and Hampson, 2005, vs. Benderlioglu and Nelson, 2004) and men (e.g., Manning et al., 1998, vs. Bang et al., 2005). Whereas studies on women have focused exclusively on T, studies on men have obtained (partly conflicting) results, which suggest an association of 2D:4D with FSH and LH (Manning et al., 2004, vs. Bang et al. 2005).

We present new data investigating, for the first time, the relationship of 2D:4D with 17-β-estradiol (E2), FSH, LH, and progesterone in women and with E2 in men. Studying E2 is especially important because 2D:4D may be related to prenatal estradiol (Lutchmaya et al., 2004); studying FSH and LH in women is important because of the already mentioned findings in men (Bang et al. 2005; Manning et al., 2004). We also investigated the association between 2D:4D and T, because conflicting findings have been obtained before. We then review all studies that have investigated associations between adult sex hormone levels and 2D:4D, using meta-analytic techniques to summarize multiple results.

Synthesizing multiple results by meta-analyses has several advantages. First, meta-analyses have a higher statistical power than the primary studies they are based upon. Thus, they may reveal effects that tend not to be detected in single studies. Moreover, meta-analyses yield especially precise estimates of the size of such effects. In the present context, this is highly desirable because primary studies have rarely studied large samples. The statistical power of primary studies in this field also suffers from the limited reliability of single hormone measurements. Sex hormone levels tend to vary from day to day, across the day, and they respond to external events (van Anders and Watson, 2006). Thus, multiple hormone measures would be desirable to arrive at precise measurements. Probably for economic reasons, this has not been done in most studies reviewed here (including our own). Second, meta-analyses can resolve conflicts arising from conflicting primary results. Meta-analyses help to understand whether conflicting evidence stems from chance, from differences in statistical power, or from the fact that the studied effect varies across populations (e.g., clinical vs. normal samples) or across particulars of methods (e.g., time of hormone measurement). Conflicting evidence is likely due to particulars of samples or methods if effect sizes vary more strongly across studies than would be expected by chance (Shadish and Haddock, 1994).

Similar to 2D:4D, left hand 2D:4D minus right hand 2D:4D (Dr−l) has also been suggested to be a negative correlate of prenatal androgenization in humans (Manning, 2002a; Manning et al., 2003). As Dr−l is sometimes used as a putative negative measure of prenatal androgenization (e.g., Manning, 2002a) we extend our present analyses to relationships between Dr−l and adult circulating sex hormone levels.

Section snippets

Methods

The original female sample (Sample 1) consisted of 66 women who were recruited via newspaper advertisements, leaflets, and newsgroup postings in Chemnitz, Dresden, and Leipzig (Saxony, Germany). All women gave their informed consent to participating in a larger study on sex hormones and attractiveness. Only women who did not use hormonal contraceptives were eligible to participate. All women were nulliparous and happened to be Caucasian. One woman was dropped from the original sample because

Meta-analytic review

We next review all findings on relationships between 2D:4D and adult levels of circulating sex hormones that are available at present. If more than one study addressed a particular relationship, we integrated the multiple study results using meta-analytic techniques.

General discussion

The length ratio of the second (index) to the fourth (ring) finger (2D:4D), a putative negative correlate of prenatal T, has become a popular variable for studying effects of prenatal androgenization in humans. However, 2D:4D may not exclusively reflect prenatal sex hormones but also adult circulating sex hormone levels. Therefore, observed relations between 2D:4D and variables of interest in adults may not reflect effects of prenatal androgenization, as is commonly assumed, but the effects of

Conflicts of interest statement

None declared.

Acknowledgments

We would like to thank all participants. We also thank Nina Asperger, Astrid Lang, Ute Lausmann, and Anja Miethe for help with the collection of data. We are grateful to Georg Jahn and Udo Rudolph for helpful comments on an earlier draft. This research was supported by Deutsche Forschungsgemeinschaft Grants HO 2506/1-1 and HO 2506/3-1 to J.H. The DFG had no further role in study design; in the collection, analyses and interpretation of the data; in the writing of the report; and in the decision

References (50)

  • S. Lutchmaya et al.

    2nd to 4th digit ratios, fetal testosterone and estradiol

    Early Hum. Dev.

    (2004)
  • M.A. Malas et al.

    Fetal development of the hand, digits and digit ratio (2D:4D)

    Early Hum. Dev.

    (2006)
  • J.T. Manning et al.

    The 2nd:4th digit ratio, sexual dimorphism, population differences, and reproductive success: evidence for sexually antagonistic genes? Evol

    Hum. Behav.

    (2000)
  • J.T. Manning et al.

    2nd to 4th digit ratio and offspring sex ratio

    J. Theor. Biol.

    (2002)
  • J.T. Manning et al.

    The second to fourth digit ratio and variation in the androgen receptor gene

    Evol. Hum. Behav.

    (2003)
  • N. Neave et al.

    Testosterone, territoriality, and the ‘home advantage’

    Physiol. Behav.

    (2003)
  • A. Ökten et al.

    The ratio of second- and fourth-digit lengths and congenital adrenal hyperplasia due to 21-hydroxylase deficiency

    Early Hum. Dev.

    (2002)
  • D.A. Putz et al.

    Sex hormones and finger length. What does 2D:4D indicate?

    Evol. Hum. Behav.

    (2004)
  • G. Sanders et al.

    The ratio of the 2nd to 4th finger length predicts spatial ability in men but not women

    Cortex

    (2005)
  • R. Trivers et al.

    A longitudinal study of digit ratio (2D:4D) and other finger ratios in Jamaican children

    Horm. Behav.

    (2006)
  • S.M. van Anders et al.

    Testing the prenatal androgen hypothesis: measuring digit ratios, sexual orientation, and spatial abilities in adults

    Horm. Behav.

    (2005)
  • C. van de Beek et al.

    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)
  • S.M. van Anders et al.

    Finger-length ratios show evidence of prenatal hormone-transfer between opposite-sex twins

    Horm. Behav.

    (2006)
  • G.M. Alexander

    Associations among gender-linked toy preferences, digit ratio, and spatial ability: evidence from eye-tracking analysis

    Arch. Sex. Behav.

    (2006)
  • A.K. Bang et al.

    A study of finger lengths, semen quality and sex hormones in 360 young men from the general Danish population

    Hum. Reprod.

    (2005)
  • Cited by (277)

    • Sex hormones and choice under risk

      2023, Journal of Economic Psychology
    • The effect of sex hormones on brain development and functioning

      2023, Encyclopedia of Child and Adolescent Health, First Edition
    View all citing articles on Scopus
    View full text