Radiographic assessment of the index to ring finger ratio (2D:4D) in adults

Abstract

The smaller index to ring finger (2D:4D) ratio has been considered as a ‘male finger pattern’ and is associated with sporting ability and a number of conditions. However, the ratio may vary according to what is measured, the hand selected and the method used. This study aimed to determine: (1) which bones (phalanges, metacarpals or both) account for variation in the 2D:4D ratio; (2) whether the ratio shows right–left symmetry or relates to hand dominance; and (3) the correlation between visual classification and measured determinations of the ratio based on radiographs. Hand radiographs obtained as part of a large osteoarthritis genetic study were examined. Each hand was classified visually into three types according to the relative length of the index and ring finger: Type 1 (index longer than ring), Type 2 (index = ring) and Type 3 (index shorter than ring). For both index and ring fingers we measured (1) from base of proximal to tip of distal phalanx and (2) metacarpal length. Reproducibility of the classification and measurements were examined using kappa and intraclass correlation coefficient; symmetry between left and right hands was examined using Bland and Altman's agreement analysis; and correlation between visual classification and 2D:4D ratio data was analysed using the anova linearity test. Data were obtained from 3172 radiographs (1636 men, 1536 women; mean age 67 ± 7.9 years, range 45–86 years). Prevalence of Type 3 hand was 61% in men and 37% in women (P < 0.001). Men had smaller 2D:4D ratios than women for phalanges (0.908 versus 0.922, P < 0.01), metacarpals (1.152 versus 1.157, P < 0.01) and the sum of phalanges plus metacarpals (1.005 versus 1.015, P < 0.01). The mean difference between right and left was –0.001 (95% limit of agreement –0.035, 0.032) for the phalangeal ratio and 0.003 (95% limit of agreement –0.051 to 0.057) for the metacarpal ratio. The 2D:4D ratio did not associate with handedness or age. There was a linear trend between the visual classification of hand type and the 2D:4D ratio data (P < 0.001). More technical difficulties (due to positioning, finger trauma, osteoarthritis) were encountered with the phalangeal ratio and visual categorization than with the metacarpal ratio: the latter could be measured in 98.7% of the study population. We concluded that measured 2D:4D ratios and visual categorization can be derived from hand radiographs. The phalanges and metacarpals both contribute to the variation in 2D:4D ratio with smaller ratios observed in men than in women. The ratio is symmetrical with only very small differences between right and left hands. Visual classification may be a useful simple tool for future epidemiological studies but is more prone to bias from positioning than direct measurement. If radiographs are used for this purpose, we recommend the metacarpal ratio with measurement of a single index hand or an average of both as it is least affected by bias from malpositioning, trauma or common joint disease.

Introduction

Men have been found to have a smaller ratio between index and ring finger length (the 2D:4D ratio) than women (Buck et al. 2003; Manning et al. 2005, 2007; McIntyre et al. 2006; Saino et al. 2006; Trivers et al. 2006). The 2D:4D ratio is thought to be determined by testosterone exposure during early intra-uterine life (Lutchmaya et al. 2004). It has been examined in relation to a number of physiological processes and psychological aspects in the last decade (Putz et al. 2004). More recently, there has been increasing interest in its apparent relationship with sporting prowess (Manning & Taylor, 2001b; Paul et al. 2006b) and diverse health conditions (Manning & Bundred, 2000, 2001a; James 2005; Vehmas et al. 2006).

However, despite considerable interest in this ratio several basic questions have yet to be answered. First, we do not know which elements of the finger (phalanges, metacarpals or both) contribute to the variation of this ratio. Previous radiographic studies have measured phalangeal 2D:4D ratios but have not examined the possible contribution from metacarpals or whether the 2D:4D metacarpal ratio is also smaller in men than in women. Secondly, ratio differences between the right and the left hand have been identified in some studies (Manning et al. 1998, 2004), but not others (Paul et al. 2006b). Such differences could relate to handedness and to relative speed of hand performance (Manning & Bundred, 2001a). However, apart from two reports from the same study population of over 400 female twin pairs (Paul et al. 2006a,b) and more recently an Internet survey of self-reported directly measured digits (Manning et al. 2007), this has been examined only in small samples. Therefore, whether the ratio is symmetrical between right and left hands in men and women still needs to be addressed, especially as this might guide selection of an index hand for this measure in epidemiological studies. Research into the 2D:4D ratio has a long history (Peters et al. 2002), using a variety of measurement techniques, including (1) direct finger length measurement from hands (Manning et al. 1998; Manning & Taylor, 2001b; Peters et al. 2002); (2) indirect finger length measurement from photocopies/scans (Lutchmaya et al. 2004; Trivers et al. 2006) and radiographs of hands (Vehmas et al. 2006; Paul et al. 2006a); and (3) distal extent measurement of the fingertips (Peters et al. 2002). However, we have found no work exploring the merits of a simple classification of hand types based on visual inspection of the relative length of index and ring fingers.

We therefore undertook this radiographic study in a large cohort of individuals who had hand radiographs undertaken as part of a gene–environmental study of osteoarthritis (OA) of the knee and hip. The objectives of this study were to: (1) examine the contributions of phalanges and metacarpals to the 2D:4D ratio; (2) determine the agreement between right and left 2D:4D ratios; and (3) examine the feasibility, validity and reproducibility of a visual classification based on hand radiographs.

Methods

Participants

Participants were men and women who underwent hand, knee and hip radiographs as part of a gene–environment interaction study of large-joint OA undertaken in Nottingham during 2001–2005 – the Genetics of Osteoarthritis and Lifestyle (GOAL) study. All were Caucasian (exclusion criteria included being of Asian or African descent). Participants with knee or hip OA were recruited from the joint replacement waiting lists and registers in Nottingham. The non-OA control participants were recruited from lists of people undergoing intravenous urography who had no radiographic evidence of large-joint OA on pelvis and subsequent knee radiographs. Full approval to assemble this cohort and to examine radiographs for research purposes was obtained from the Nottingham Research Ethics Committee.

Hand radiographs

Separate radiographs were taken of right and left hands. The participant was seated adjacent to the X-ray table with the forearm and hand flat and prone on the table with no lateral angulation at the wrist. The hand was centred on the cassette with fingers slightly spread apart but flat. The X-ray beam was centred on the third metacarpophalangeal joint. Images were obtained using a small focal point and a detail cassette. Exposures and distances were: 48 kV; 3.2 mA.s; 90 cm source to image distance (ffd). Each film was scanned and written to CD using the Hipax 4.2 X-Ray Image Processor. CD images were read using the Hipax Private Health Disk Image Viewer that enables straight-line measurements to an accuracy of 0.01 mm.

Radiographic measurements

The following measurements were made on the index and ring fingers of both hands: (1) from the mid-point of the base of the proximal phalanx to the mid-point of the tip of the distal phalanx; and (2) from the mid-point of the base to the mid-point of the tip of the metacarpal.

All measurements were made by one observer and entered directly into a Microsoft Access database. If no measurement was possible the reason for this was recorded.

Visual classification

Each hand was classified according to whether the index finger was longer (Type 1), equal to (Type 2) or shorter than the ring finger (Type 3), by visual comparison of the soft tissue outline of the finger ends on the radiograph (Fig. 1). Because of common difficulty in confidently assigning hands to one of these three categories in our pilot work, each category was divided into ‘definite’ or ‘probable’ according to the confidence of the observer to assign to a single category. If no category could be assigned the likely reason for this was recorded.

Fig. 1.

Visual classification of radiographs of hand types.

Quality of data

The data were checked for outliers (data more than three standard deviations above or below the mean). Nineteen data points were inspected and re-measured; six were correct, 13 had been incorrectly entered onto the database.

Statistical analyses

Reproducibility was examined on a random sample of 30 subjects generated from computer-generated random numbers. The investigator was blinded to subject identity. The radiographic combined phalangeal lengths and metacarpal lengths (mm) for index and ring fingers, and the visual classification of each hand were measured at baseline and repeated at the middle and end of the study. The intraclass correlation coefficient (ICC) and weighted kappa were used to determine the reproducibility of the radiographic measurements and of the visual classifications, respectively. Radiographic measurements and ratio data were assessed for normality and compared using analysis of variation (anova) between men and women. Mean and standard deviation were calculated for each measure after a normality test. Comparison between men and women was undertaken using an independent t-test. The standard mean difference [Cohen's d or effect size, ES (Cohen, 1988)] between men and women and 95% confidence interval (CI) were calculated to allow cross-measurement (e.g. the phalangeal ratio versus the metacarpal ratio) or cross-study comparisons. Symmetry between right and left was examined using Bland and Altman's method for agreement (Bland & Altman, 1986). The mean difference and 95% limit of agreement were calculated under the null hypothesis of no difference between right and left (mean difference = 0) and possible variation of the agreement. Hand dominance was analysed using anova, and covariates (gender and hand) were adjusted by covariance analysis. Pearson's correlation coefficient (r) was calculated for the relationship between the 2D:4D data and age. Prevalence of Type 3 or possible Type 3 hand was calculated and χ² test was used for group comparison. Symmetry of visual classification between right and left hands were cross-tabulated and tested using a contingency coefficient as they were nominal categories. The relationship between the radiographic measures and the visual classification was examined using a linearity test.

Results

Of 3172 subjects, 1636 were men and 1536 were women. Age ranged from 45 to 86 years with a mean of 66.6 (SD 7.9) with no significant difference between genders. Men were significantly taller and heavier than women and had a significantly smaller body mass index (BMI) (Table 1). The pattern of handedness was similar between genders.

Table 1.

Characteristics of the study participants

Characteristic	Men	Women	Total
No. of participants	1636	1536	3172
Age (years)	66.6 ± 7.8	66.7 ± 7.9	66.6 ± 7.9
Height (cm)	172.5 ± 6.8	159.6 ± 6.3**	166.3 ± 9.2
Weight (kg)	86.3 ± 15.0	75.5 ± 15.6**	81.0 ± 16.2
BMI	28.9 ± 4.6	29.6 ± 5.9**	29.3 ± 5.3
Hand dominance (%)
Right	1467 (90%)	1404 (91%)	2871 (90%)
Left	138 (8%)	109 (7%)	247 (8%)
Ambidextrous	31 (2%)	23 (2%)	54 (2%)

Mean ± SD unless otherwise stated.

SD, standard deviation; BMI, body mass index (kg m⁻²); * P ≤ 0.05;

^**P ≤ 0.01.

Of the 6344 hands (3172 pairs) assessed, the phalangeal ratio could be measured in 5336 (84.1%) whereas the metacarpal ratio was measurable in 6260 (98.7%). With the visual classification, 4525 (71.3%) could be attributed a definite category, 1341 (21.1%) were attributed a probable category, but 478 (7.5%) could not be classified. Problems encountered both with the measured phalangeal ratio and the visual classification included marked radial or ulnar deviation secondary to interphalangeal arthritis, marked flexion deformity/attitude, or injury and partial loss of finger length. In addition, malpositioning of hands (ulnar or radial deviation at wrists or metacarpophalangeal joints) or splaying of fingers compromised visual classification but not phalangeal measures. The predominant factor that interfered with metacarpal ratio measurement was shortening or deformity relating to previous trauma or arthritis at either end of the metacarpal, but this was much less prevalent than at phalanges.

2D:4D length ratio

The ICC for radiographic reproducibility of finger length measures ranged from 0.91 to 1. Men had lower 2D:4D phalangeal, metacarpal and combined ratios than women (Table 2). The ES between men and women for the right phalangeal ratio was –0.64 (95% CI –0.71, –0.56) and for the left was –0.55 (95% CI –0.63, –0.48). The ES between men and women for the right metacarpal ratio was –0.17 (95% CI –0.24, –0.10) and for the left was –0.12 (95% CI –0.19, –0.05). The mean difference between right and left metacarpal ratios was 0.003 and the 95% limit of agreement ranged from –0.051 to 0.057, suggesting a good agreement between right and left. Similar results were observed for the mean differences between right and left phalanx ratios (0.001, 95% limit of agreement –0.035, 0.032) and the ratio of the combined metacarpal and phalangeal lengths (0.0004, 95% limit of agreement –0.028, 0.029).

Table 2.

Comparison of 2D:4D ratios between men and women

	Right		Left		Average
2D:4D ratio	Men	Women	Men	Women	Men	Women
Phalangeal
Mean	0.906	0.921	0.909	0.922	0.908	0.922
SD	0.024	0.023**	0.024	0.023**	0.022	0.021**
n	1433	1312	1476	1363	1374	1267
Effect size	–0.64 (–0.71, –0.56)		–0.55 (–0.63, –0.48)		–0.65 (–0.73, –0.57)
Metacarpal
Mean	1.153	1.159	1.151	1.155	1.152	1.157
SD	±0.035	±0.036**	±0.033	±0.034**	±0.031	±0.033**
n	1610	1522	1616	1523	1599	1514
Effect size	–0.17 (–0.24, –0.10)		–0.12 (–0.19, –0.05)		–0.16 (–0.23, –0.09)
Combined
Mean	1.004	1.015	1.005	1.014	1.005	1.015
SD	±0.021	±0.021**	±0.020	±0.020**	±0.018	±0.019**
n	1426	1305	1465	1356	1361	1255
Effect size	–0.52 (–0.60, –0.45)		–0.45 (–0.53, –0.38)		–0.54 (–0.62, –0.46)

^**P ≤ 0.01 between men and women.

SD, standard deviation.

N, number of subjects.

The 2D:4D ratios were similar for dominant hands, non-dominant hands and ambidextrous hands regardless of the bone measured (Table 3). Age was not correlated with the 2D:4D phalangeal (r = –0.012, P = 0.551) or metacarpal measures (r = –0.002, P = 0.913).

Table 3.

2D:4D ratios and hand dominance adjusted by gender and hand

	2D:4D ratio, mean (95% confidence interval)
Hand dominance	Phalangeal	Metacarpal	Combined
Dominant	0.913 (0.911, 0.914)	1.155 (1.152, 1.157)	1.008 (1.007, 1.010)
Non-dominant	0.915 (0.913, 0.916)	1.154 (1.152, 1.156)	1.010 (1.008, 1.011)
Ambidextrous	0.911 (0.906, 0.916)	1.155 (1.148, 1.161)	1.008 (1.004, 1.012)
P	0.058	0.868	0.295

Visual classification

The weighted kappa for reproducibility of the visual classification was 0.66–0.78. Prevalence of definite or probable Type 3 (i.e. longer ring finger) on either hand was 60.88% in men and 37.29% in women (P < 0.001); men were nearly three times as likely to have a Type 3 hand than women (OR = 2.62, 95% CI 2.27, 3.02).

Visual classification was associated with the measured 2D:4D length ratio (Table 4).

Table 4.

Association between visual classification and the measured 2D:4D length ratios

	2D:4D length ratio, mean (SD)
Visual classification	Phalanges	Metacarpals	Combined
Right
Type 1	0.934 (0.021)	1.171 (0.035)	1.026 (0.017)
Type 2	0.916 (0.021)	1.158 (0.034)	1.012 (0.016)
Type 3	0.900 (0.022)	1.145 (0.033)	0.997 (0.018)
Plinearity	< 0.001	< 0.001	< 0.001
Left
Type 1	0.933 (0.020)	1.170 (0.034)	1.026 (0.017)
Type 2	0.918 (0.020)	1.154 (0.031)	1.012 (0.016)
Type 3	0.903 (0.023)	1.142 (0.032)	0.997 (0.017)
Plinearity	< 0.001	< 0.001	< 0.001

The contingency coefficient between right and left was 0.75 (P < 0.01), supporting the symmetry between right and left seen with the direct measurements. However, 45 individuals showed clear discordance in visual classification (i.e. one hand Type 1, the other Type 3). Detailed examination of these films showed that this was always accounted for by mechanical factors such as: lateral deviation at metacarpophalangeal joints; radial deviation at the wrist; attrition of the trapezium with proximal subluxation of the index metacarpal in association with thumb-base OA; or OA of the index metacarpophalangeal joint. However, the measured 2D:4D ratios in these 45 visually discordant pairs (left vs. right) were almost identical for phalangeal (0.917 vs. 0.918, P = 0.83), metacarpal (1.160 vs. 1.167, P = 0.21) and combined measures (1.013 vs. 1.016, P = 0.35).

Discussion

This study confirms that the 2D:4D finger ratio is measurable on radiographs and differentiates between male and female hands (P < 0.01), in accordance with the majority of published studies (Buck et al. 2003; McIntyre et al. 2006; Manning et al. 2007). To our knowledge, this is the first study to investigate metacarpal length separately and it is of interest that the 2D;4D metacarpal ratio also differentiates between male and female hands with the male mean ratio being significantly less than the female (1.152 compared with 1.157, P < 0.01).

One obvious advantage of the metacarpal ratio over the phalangeal ratio in our study was that it is less affected by finger problems such as interphalangeal joint OA and trauma and could thus be applied to almost the entire study population (99% versus 84% for the phalangeal ratio). This practical problem of phalangeal ratio measurement has been noted in one previous study (Vehmas et al. 2006) in which the investigators concluded that shorter index finger measures were often attributable to involvement by OA. Because our study population was assembled primarily to investigate gene–environmental interaction in knee and hip OA, many of the participants were middle-aged and elderly and had a higher than expected prevalence of hand OA. In younger populations with less frequent hand OA and trauma this advantage of the metacarpal ratio over the phalangeal ratio may be diminished.

Overall symmetry of the ratios was seen for right and left hands with no clear effect of hand dominance. Although there were some significant differences between the right and left measurements in this large sample, the differences were extremely small for both phalangeal (–0.001) and metacarpal (0.003) ratios and were not systematic, presumably reflecting the lack of absolute symmetry for all elements of the human body when very precise measurements are undertaken. This very close symmetry of 2D:4D is in accord with one other radiographic study that focused only on women (Paul et al. 2006a).

The effect sizes between men and women for the phalangeal ratio were larger than those recently reported by Manning et al. (2007) (right ratio d = 0.17, left ratio d = 0.13; current study right ratio d = 0.64, left ratio d = 0.55) but were comparable with other published work (e.g. Manning et al. 2005; Trivers et al. 2006). As discussed before, this may be because the subjects of this study were primarily selected for an OA case control study. About two-thirds of our subjects had hip and/or knee OA and all were aged between 45 to 85 years, both of which are associated with history of finger injury or interphalangeal nodes. These result in a greater likelihood of finger deformity, affecting the precision of the phalangeal measurement and the resultant ratio. However, the ES for the metacarpal from this study was similar to that obtained from direct finger measurement from Manning's Internet study. This may be because the metacarpal measure in our specific population is less biased than the phalangeal measure and therefore closer to the true measure of the 2D:4D ratio (although self-measurement is likely to be associated with a higher rate of error than is usual, the very large numbers of participants and error checks will have balanced this to a large extent).

In contrast to Manning's study, however, the sex dimorphism was almost identical for right and left hands in the present study, for both the phalangeal and the metacarpal measures. The symmetry is supported by another radiographic study (Paul et al. 2006b). Whether this is due to the radiographic measure or sampling variation (e.g. age difference between studies) remains unknown. Further radiographic evidence from a normal population is needed.

To our knowledge, this is the first study to attempt a simple visual classification for the 2D:4D relative length ratio, albeit indirectly from a radiograph. We have demonstrated good reproducibility (kappa 0.66–0.78) for this classification system and good criteria validity as compared with the measured 2D:4D length ratio – a proxy ‘gold standard’ (Table 4). Furthermore, results obtained with the visual classification showed more men than women with Type 3 hand (the male pattern 2D:4D finger type) and supported the symmetry between right and left hands demonstrated by the measured ratios (contingency r = 0.75). However, care must be taken when using this classification as it is prone to bias due to hand positioning (ulnar/radial deviation at wrists and metacarpophalangeal joints) and splaying of fingers, as well as to the problems that can bias the measured phalangeal ratio (finger trauma, interphalangeal arthropathy). These problems may even cause two discordant classifications in the right and left hands of an individual, contributing to the difficulty of assigning a definite or probable finger type to 7.5% of our study population. Greater attention to correct positioning of wrists, metacarpophalangeal joints and fingers at the time of radiography might improve the performance of this measure.

There are several caveats to this study. First, being assembled for a case-control study that focuses primarily on genetic and environmental risk factors for hip and knee OA, the study population was non-random and comprised predominantly older subjects (mean age 67 years), all Caucasian. Therefore, the generalizability of the estimates, for example the prevalence of Type 3 hand, to other populations cannot be assumed. Secondly, the combined ratio was generated from adding the phalanx to the metacarpal length rather than direct measure from metacarpal base to finger tip. The contribution from cartilage and soft tissues at the metacarpophalangeal joint was therefore excluded.

In conclusion, the phalanges and metacarpals both contribute to variation in the 2D:4D ratio. The phalangeal and metacarpal ratios both differentiate between men and women with the ratio being smaller in men. The metacarpal ratio is less affected by hand trauma, interphalangeal OA and malpositioning and thus has the advantage of being applicable to more individuals, especially in an older population. As the 2D:4D ratio is symmetric and there is no sex dimorphism between right and left hands, a single index right-hand measure, rather than a mean of the two, may be used. Although prone to positioning bias in addition to finger trauma and OA, visual classification of the 2D:4D ratio is very simple and quick to undertake and merits further study, including clinical and self-reported classification as well as radiographic assessment.