- Email: [email protected]
The far interpupillary distance. A gender-specific variation with advancing age
Ophthal. Physiol. Opt. Vol. 19, No. 4, pp. 317±326, 1999 # 1999 The College of Optometrists. Published by Elsevier Science Ltd All rights reserved. Printed in Great Britain 0275-5408/99 $20.00 + 0.00
PII: S0275-5408(98)00082-9
The far interpupillary distance. A genderspecific variation with advancing age Jonathan S. Pointer Optometric Research, 4A Market Square, Higham Ferrers, Northants NN10 8BP, UK Summary A knowledge of the magnitude of the far interpupillary distance (FIPD) in relation to a specific population is of clinical, practical and theoretical interest. A FIPD database is presented here, comprising material collated from the spectacle dispensing records of n = 1800 subjects seen in routine optometric practice. All measurements were taken by the author on healthy Caucasian (white, Northern European) males and females. The data were equi-partitioned either across three age bands (16±25, 26±40, 41±65 years: mixed refractive types, total n = 900) or between the three distance refractive types (emmetropia, hypermetropia, myopia: all subjects aged between 41±65 years, total n = 900). A consistent gender difference (male>female) was confirmed throughout this material, irrespective of age group; refractive type, however, had no influence on the magnitude of this facial parameter. Summary results of this anthropometric survey are presented in tabular form, facilitating reference by ophthalmic and dental clinicians and by the designers of binocular optical instruments. There was also revealed evidence of a gender-specific pattern of change in the FIPD variable with advancing age. An approximately 3% increase in the magnitude of the human FIPD from the mid-teens to later middle age was attained in males by early middle age, being little altered thereafter: in contrast females continued to record an increase in this facial parameter into later middle age. An explanation for this hitherto unremarked feature of human facial anthropometry might be sought in the gender-specific changes post-puberty of the cranial skeletal anatomy and in the soft tissues of the orbital region. # 1999 The College of Optometrists. Published by Elsevier Science Ltd. All rights reserved
Introduction
the eye's optic axis rather than the line of sight which is approximately centred on the pupillary aperture. This decentration is occasioned by the infero-temporal oset of the fovea, itself a feature of possible teleological signi®cance given that it minimises chromatic aberration in the human refractive system. Consequently the optic axis lies temporal to the visual axis: the two axes intersect at the (single) nodal point of the reduced eye, the angle alpha between them amounting to ca 5 deg, being less in myopic and greater in hypermetropic and infants' eyes (Emsley, 1952). Practical determination of this angular displacement using a point source of light and an arc perimeter, for example, results in the measurement of angle kappa (Landolt: Emsley, 1952); this is the angle formed between the visual axis and the pupillary line (an approximation of the optic axis, sited perpendicular to the cornea). Although the pupil centre is oset slightly nasally in relation to the corneal centre, the dierence
The far interpupillary distance (FIPD) is the facial measurement (mm) in the horizontal plane between the geometric centres of the pupillary apertures of a pair of eyes, which latter are focused at optical in®nity. This measurement is typically taken across the upper bridge of a subject's nose, in the approximate spectacle (frontal) plane (Sasieni, 1975). To minimise the introduction of undesirable optical aberration and prismatic eects when wearing spectacles or when using an optical instrument, the line of sight of each of an individual's pair of eyes should ideally pass through the optical centre of each of a pair of ophthalmic lenses. But as has been appreciated at least since the time of Helmholtz (Emsley, 1952) it is Received: 3 August 1998 Revised form: 24 November 1998
317
318
Ophthal. Physiol. Opt. 1999 19: No 4
in magnitude between the two angular measurements (alpha and kappa) is clinically insigni®cant. Further, given the practical ease with which the pupillary centres (or rather, the relative linear displacement between complementary points on the two eyes) can be estimated, ophthalmic lenses are invariably aligned with the ``optic'' (pupillary) axes rather than the ``visual'' axes of an individual: whence the IPD. The magnitude of the FIPD is of basic and clinical signi®cance. For a given population type, anthropometric surveys can establish descriptive statistics for this measurement: the FIPD then becomes a useful diagnostic tool for developmental anatomists, geneticists and cranio-facial surgeons (e.g., Roy, 1985), a practical aid in restorative facial surgery and in the sphere of prosthodontics (e.g., Cesario and Latta, 1984), a necessary parameter in the design of binocular instruments and equipment (Harvey, 1982) and, not least, in the allied commercial (frame design and lens blank manufacture) and clinical (ophthalmic dispensing) ®elds of optometry. Osuobeni and Al-Musa (1993) have tabulated IPD values from several published sources, facilitating comparisons between gender, age band and ethnic group. Sasieni (1975) quotes average values for the FIPD in the adult (assumed Caucasian/European) subject as follows: males = 63±64 mm (variation 58±72), females = 60±61 mm (variation 57±65). A new FIPD database is presented here, comprising material generated in routine optometric practice and compiled in a balanced format. Statistical analysis of this material permits speci®c conclusions to be drawn with regard to the distribution and characteristics of the FIPD in a contemporary Caucasian population. In addition certain anatomical considerations underlying age-related variations in this parameter are explored.
Method and materials Viktorin's method The FIPD was measured routinely by the author, using a ruler calibrated in mm, after the method of Viktorin (see Sasieni, 1975). Despite the several objections to the absolute measurement accuracy of this technique (summarised by Anderson, 1954) its facility has ensured that it has remained perhaps the most widely taught and practised interocular measurement procedure: it is also the method most frequently cited in published investigations involving the assessment of this facial parameter (Osuobeni and Al-Musa, 1993). All of the FIPD material collated herein was obtained in the following manner. The subject was seated opposite to, on the same level as and within arm's length of the practitioner in a well-illuminated
ophthalmic dispensing area. A ¯at calibrated (mm) gauge was laid across the subject's nose and lightly supported by the practitioner's left hand. The subject was instructed to keep both eyes open and to ®xate the practitioner's open left eye; using this eye the practitioner aligned the ``zero'' mark on the gauge with the temporal limbus of the subject's right eye. Keeping the ruler still, the practitioner then closed his left eye and opened his right, at the same time instructing the subject to move ®xation across to the practitioner's open right eye. The position of the nasal limbus of the patient's left eye could then be read on the ruler: given that approximately straight and parallel lines of sight were maintained throughout this procedure, this calibration value was taken as the FIPD measurement (mm). Of course this approach ignores the error associated with angle alpha (see Introduction), and also any inaccuracies introduced by parallax arising from a marked dierence between the subject's and practitioner's IPD and inequalities in the vertical plane due to stature or sitting position. However, in practice these theoretical inaccuracies are small (Sasieni, 1975). Subject groups FIPD values (mm), as measured by the author on n = 900 male and n = 900 female Caucasian (white, Northern European) subjects, were taken at random but in accordance with certain criteria (see below) from the spectacle dispensing records held at the author's optometric practice. Data compilation was such that equal subject numbers were assembled across the various categories: all material required for this report (i.e., subject gender, age group and FIPD value) was entirely non-attributable in analysis and patient con®dentiality was not compromised. All subjects were between the ages of 16 and 65 years at the time of the spectacle dispensing. The choice of these speci®c age limits was with the intention of minimising mensuration uncertainties associated with, on the one hand childhood physical growth changes (a source of confounding error encountered in an IPD study undertaken by Osuobeni and Faden, 1993) and on the other, the increased laxity of the soft tissues around the orbit of the elderly subject (Fledelius and Stubgaard, 1986; Whitnall, 1921). In addition, given the possibility of systematic variation in the dimensions of the globe of the eye (cf. myopic vs hypermetropic axial length dierences, and the in¯uence of degree of myopia upon the amount of exophthalmos: Quant and Woo, 1992) distance refractive classi®cation was also recorded. Two data groups were assembled: Group A comprised three age bands (16±25, 26±40, 41±65 years) of unspeci®ed (i.e., mixed)
Far interpupillary distance: J. S. Pointer distance refractive type with n = 150 male and n = 150 female FIPD values for each age band (i.e., total n = 900): Group B comprised material from subjects aged 41±65 years equally partitioned between the three refractive types (emmetropoia, hypermetropia, myopia) with n = 150 male and n = 150 female FIPD values for each refractive classi®cation (i.e., total n = 900). Data selection criteria were not extensive. All subjects were in good health, recorded an acuity of 6/6 or better and had no binocular vision anomalies; distance refractive corrections >2 6.00 DS and > ÿ 2.00 DC barred the subject's data from inclusion in the sample, as did anisometropia of the spherical elements >2.00 D. In accord with a previous publication (Pointer, 1995b), emmetropia was de®ned as the refractive range ÿ0.25 D to +0.75 D spherical equivalent refraction (SER: sphere power plus one-half of any cylinder power, at the spectacle plane); hypermetropia included individuals whose spectacle correction was r + 0.875 D SER; and myopia covered individuals whose spectacle correction was rÿ 0.375 D SER. All facial measurements had been taken by the author as a prerequisite to the dispensing of a complete pair of (distance vision) spectacles: no ``special requirement'' or occupational dispensings were included in the data sets. Data from each subject were only entered once, so the results represent material from n = 1800 separate individuals.
Results Descriptive statistics1 for the two groups of FIPD data are given in Table 1A, B. The magnitude, both of the range and of the three measures of central tendency associated with the distribution of the FIPD variable, is seen to be consistently greater in male than in female subjects. A degree of variation with age group (Table 1A)Ðbut not with refractive type (Table 1B)Ðis also evident. The small values of the coecients of skewness and kurtosis are indicative of the FIPD variable being normally distributed in each of these data groupings. Probability plots (not shown here) typically indicated normality for at least two standard deviations (SD) either side of the mean value, with cumulative frequency distributions yielding an ogive. Data of male and female subjects in the 16±25 and 26±40 years age groups display a small positive skew (Table 1A), with the exception of a very small negative skew in the material relating to 16±25 year old 1 Data management and analysis handled by STATISTICA/Mac software (Release 3, 1992: StatSoft Inc., Tulsa, Oklahoma 74104, USA) on Apple Macintosh LCII computer hardware running System 7.
319
malesÐwhich data also have the highest SD. Taken together these two statistical features suggest a wider variation of FIPD measurements (including a few low values) in adolescent males as compared to adolescent females and to adults of either gender. Certainly the range of FIPD values is greatest for this particular subject group (16 mm as compared to 13±15 mm for all of the other gender/age groups: Table 1A). Consequently these speci®c observations probably re¯ect the earlier physical maturation of the human female as compared to the male teenager, a feature also recorded by Fledelius and Stubgaard (1986). Both genders show a small negative skew for the oldest age band represented here (41±65 years), indicating that this age grouping contains a few low FIPD values. This tendency to a small negative skew largely persists in the data of the three refractive types (Table 1B), this material also having been collated from subjects in the 41±65 years age band: this uniformity is taken as being indicative of the internal consistency of the new FIPD database presented here.
Analysis Student's t-test for independent samples was run on all available male and female age and refractive type group pairings: the results are summarised in Table 2A and B. Irrespective of age group or refractive type, the FIPD of males is always statistically signi®cantly dierent (larger) compared to that of females. With regard to refractive type (Table 2B), a genderbased dierence in the mean FIPD is greatest between hypermetropic males and females (3.34 mm) and least between myopes (2.50 mm), with emmetropes falling between the two (2.82 mm). However, cross-pairing of FIPD values between refractive types fails to attain statistical signi®cance within either gender. The picture emerging from the results of the pairwise analyses of the age-grouped data is more complex (Table 2A). The largest gender-based dierence in mean FIPD values lies within those individuals aged 26±40 years (3.56 mm): the adolescents (16±25 years) and the oldest individuals (41±65 years) both display smaller gender-based dierences (2.75 and 2.97 mm, respectively). Within each gender the pattern of transition across age groups diers. In males the mean FIPD of the 16±25 years group is statistically signi®cantly dierent (smaller) compared to that of both the 26±40 and the 41±65 years groupings (by 1.58 and 1.85 mm, respectively), but which latter are not statistically signi®cantly dierent to one another (0.27 mm). By contrast, in females the mean FIPD of the 26±40 years group is statistically signi®cantly dierent to the mean value of this parameter in both 16±25 year olds
320
Ophthal. Physiol. Opt. 1999 19: No 4
Table 1. The far interpupillary distance (FIPD): magnitude (*mm) and summary statistics 1A. Variation with age group and gender (mixed distance refractive types) Range* Percentile* Age group (years)
Gender
n
Min
Max
5th
95th
Mean 2 1st SD*
Median*
Mode*
Skewness
Kurtosis
16±25
Male Female Male Female Male Female
150 150 150 150 150 150
56 54 58 55 58 55
72 67 73 68 73 70
58.06 57.20 60.14 57.33 60.37 58.00
69.58 65.58 70.32 66.58 70.00 66.41
63.65 2 3.38 60.90 2 2.53 65.23 2 3.14 61.67 2 2.72 65.50 2 2.77 62.53 2 2.76
64 61 65 61 66 63
66 60 66 61 66 61
ÿ0.01 +0.23 +0.29 +0.12 ÿ0.32 ÿ0.17
ÿ0.06 +0.01 ÿ0.13 ÿ0.22 +0.18 ÿ0.28
26±40 41±65
1B. Variation with refractive type and gender (age group: 41±65 years) Range* Percentile* Refractive type
Gender
n
Min
Max
5th
95th
Mean 2 1st SD*
Median*
Mode*
Skewness
Kurtosis
Myopia
Male Female Male Female Male Female Male Female
150 150 150 150 150 150 450 450
59 57 59 56 58 55 58 55
74 70 72 71 73 70 74 71
60.78 58.14 59.87 58.66 60.69 58.55 60.58 58.52
70.37 67.00 70.00 66.67 70.15 66.46 70.14 66.75
65.35 2 2.80 62.85 2 2.72 65.33 2 2.95 62.51 2 2.75 65.75 2 2.94 62.41 2 2.70 65.48 2 2.90 62.59 2 2.73
65 63 66 63 66 62 66 63
65 61 66 61 66 60 65 61
+0.03 ÿ0.04 ÿ0.24 ÿ0.26 ÿ0.32 +0.09 ÿ0.10 ÿ0.10
+0.28 ÿ0.05 ÿ0.04 ÿ0.09 ÿ0.28 ÿ0.06 ÿ0.22 ÿ0.26
Emmetopia Hyperopia Combined
(0.77 mm) and 41±65 year olds (0.86 mm): however, like males, the dierence between the mean FIPD of 16±25 and 41±65 year old individuals (1.63 mm) is statistically signi®cant.
Discussion I. The in¯uence of gender on the FIPD To the present author's knowledge there has been only one comparable investigation into the variation
Table 2. The FIPD: results of a series of Student's t-tests for independent sample comparisons 2A. Influence of age group and gender
Difference between means Absolute (mm)
%
Student's t-value
p-level
Male 16±25 vs Female 16±25 Male 26±40 vs Female 26±40 Male 41±65 vs Female 41±65 Male 16±25 vs Male 26±40 Male 16±25 vs Male 41±65 Male 26±40 vs Male 41±65 Female 16±25 vs Female 26±40 Female 16±25 vs Female 41±65 Female 26±40 vs Female 41±65
+2.75 +3.56 +2.97 ÿ1.58 ÿ1.85 ÿ0.27 ÿ0.77 ÿ1.63 ÿ0.86
+4.52 +5.77 +4.75 ÿ2.42 ÿ2.82 ÿ0.41 ÿ1.25 ÿ2.61 ÿ1.37
+7.98 +10.52 +9.30 ÿ4.19 ÿ5.17 ÿ0.78 ÿ2.52 ÿ5.34 ÿ2.74
0.0000001 0.0000001 0.0000001 0.00005 0.0000005 0.5 0.01 0.0000005 0.005
2B. Influence of refractive type and gender Male MY vs Female MY Male EM vs Female EM Male HY vs Female HY Male MY vs Male EM Male MY vs Male HY Male EM vs Male HY Female MY vs Female EM Female MY vs Female HY Female EM vs Female HY Male ALL vs Female ALL
+2.50 +2.82 +3.34 +0.02 ÿ0.40 ÿ0.42 +0.34 +0.44 +0.10 +2.89
+3.98 +4.51 +5.35 +0.03 ÿ0.61 ÿ0.64 +0.54 +0.70 +0.16 +4.62
+7.81 +8.56 +10.26 +0.04 ÿ1.23 ÿ1.23 +1.08 +1.43 +0.34 +15.38
0.0000001 0.0000001 0.0000001 0.9 0.5 0.5 0.5 0.5 0.9 0.0000001
Inter group comparison
Far interpupillary distance: J. S. Pointer of the FIPD measurement in Caucasian subjects: as an adjunct to a study of changes in eye position during life, Fledelius and Stubgaard (1986) measured the FIPD in healthy male and female subjects between the ages of 5 and 80 years. The larger subject numbers in this present study corroborate the conclusions of the earlier Danish investigation and contribute an additional ®nding. Speci®cally, it is substantiated (i) that the magnitude of the FIPD is always greater in males than females of the same age group, and (ii) that the absolute value of the FIPD increases with age, at least until the third decade of life (for a further discussion of this feature see below: Discussion II). The mean FIPD values obtained for male and female subjects in these two studies across broad age bands spanning childhood, adolescence and adulthood are presented chronologically in Figure 1. The age-related increase in the absolute value of the FIPD is evident for either gender at least into adulthood; for both series of data the plotted male and female mean values are statistically signi®cantly dierent at p = 0.01 or greater. This consistent gender dierence is demonstrated in Figure 2 where, for equal numbers of male and female subjects (n = 450 each) aged 41±65 years, cumulative frequency ogives of the FIPD values are plotted. It is evident that in age-matched subjects the
321
male and female FIPD parameter diers only by a simple linear transposition along the x-axis: male values are consistently ca 3 mm larger than those of females of the same age group, irrespective of the absolute magnitude of the FIPD variable. The in¯uence of gender on the magnitude of the human interpupillary separation is entirely predictable. On the basis of the FIPD data presented herein, female values are typically ca 95% of the values recorded for similarly-aged males. This result accords with established gender dierences in human anatomical and skeletal dimensions as given in post-mortem and forensic tables, for example. Indeed such a genderbased physical (size) dierence has recently been implicated (Pointer, 1995a) in an explanation for the clinical observation that a presbyopic female typically requires a positive reading addition of marginally greater magnitude than an age-matched male subjectÐessentially because a woman's arms are shorter, producing a slightly closer reading position than a man. An additional ®nding presented here is that an individual's refractive typeÐemmetropia, hypermetropia or myopiaÐhas no in¯uence on the magnitude of the FIPD in mature adult Caucasian subjects. Evidently for low to moderate degrees of refractive error clinical measurement of eye position in the frontal (horizontal)
Figure 1. Variation of the FIPD with age group and gender. The published data of Fledelius and Stubgaard (1986: Table 4, p. 484) are identified by the square symbols, the new data of Pointer (refer to Table 1A) are represented by the circular symbols. In each case the solid black symbols indicate male and the unfilled symbols female mean values, with error bars indicating, respectively, + or ÿ 1st SD. Note the discontinuity midway along the x-axis, occasioned by the different age ranges studied in the two reports. Each mean value plotted from the Danish study was obtained from either ca 50 subjects (5±10 years) or ca 70 subjects (11±19 years): all other values are each the mean of measurements upon 150 subjects. At each age group the gender difference is statistically significant at p = 0.01 or greater.
322
Ophthal. Physiol. Opt. 1999 19: No 4
Figure 2. The gender difference in FIPD. Cumulative frequency/percentage ogives for male (solid black symbols: n = 450) and female (unfilled symbols: n = 450) FIPD values (combined data for all three refractive types, age group 41±65 years: refer to Table 1B). The normal distribution of the FIPD material for either gender is demonstrated, with the two ogives being separated by a simple linear displacement of ca 3 scale units (mm) along the abscissa, male values being the greater.
planeÐunlike the assessment of orbital protrusion taken orthogonal to this (Quant and Woo, 1992)Ðis not in¯uenced by the axial length or size of the globe within the bony orbit. A reference table of human FIPD values It is the intention that the FIPD information presented here relating to a broad age range of male and female Caucasian subjects will be a useful reference source. The information is relevant not only within the ®eld of ophthalmic optics but also to clinicians in the dental and prosthetic/restorative surgical specialities. In addition, the inclusion of such details as the magnitude of the 5th and 95th percentile values in Table 1A and B may be of assistance to designers of binocular optical instruments and vision systems (Harvey, 1982): such equipment, whether intended for clinical, industrial or defence application, must have an adequate adjustment range appropriate to the speci®c group of personnel who will be using that equipment.
Discussion II. A gender-speci®c variation of the FIPD with age Studies which have quanti®ed the FIPD variable in children and young subjects over a period of time have
reported an increase in magnitude with increasing years (Kaye and Obstfeld, 1989; Osuobeni and Faden, 1993). Many investigators appear to have made the implicit assumption that this parameter attains a plateau at some point between the second and third decades of life (see Osuobeni and Faden, 1993). Certainly in a longitudinal biometric study of a group of Caucasian male subjects Bruckner et al. (1987) could ®nd no increase in FIPD beyond 30 years of age, with the majority of subjects attaining a stable value in their late teens or early twenties. Fledelius and Stubgaard (1986) reported a consistent gender dierence (male>female) not only in the absolute value of the FIPD in age-matched Caucasian subject groups but also in the age at which the FIPD attained the ``adult'' level. In line with observations on general bodily growth and maturation, human female subjects attained the ``adult'' FIPD dimension around 14±16 years of age (61.8 mm) and males at around 17± 19 years (65.8 mm). However, further study of the tabulated FIPD values for male and female subject groups of increasing age in the publication of Fledelius and Stubgaard (1986: Table 3, p. 484) reveals an interesting subsidiary gender dierence. Beyond the late teenage years female subjects in particular recorded an increase in the FIPD. This increase was slight in the 31±55 years age band (1.8% rise, to 62.9 mm) but
Far interpupillary distance: J. S. Pointer
323
Table 3. The FIPD: variation with age group and gender (data from Table 1A) Age group (years)
Gender
n
Mean 2 1st SD (mm)
16±25
Male Female Male Female Male Female
150 150 150 150 150 150
63.65 2 3.38 60.90 2 2.53 65.23 2 3.14 61.67 2 2.72 65.50 2 2.77 62.53 2 2.76
26±40 41±65
more marked in the subject group aged >55 years (a further 3.5% rise, to 65.1 mm). Male values increased only slightly in the 31±55 years age range from the level of FIPD reached in the late teens (1.0% rise, to 66.4 mm) and actually reduced (0.6% drop, to 66.0 mm) within the >55 years age grouping. The authors made no speci®c comment regarding this particular gender dierence in their data, aside from remarking that age-related soft tissue changes could account for slight alterations in orbital/facial measurements in elderly subjects. A study restricted to females of Arab origin (Osuobeni and Faden, 1993) also reported an increase (statistically signi®cant, pR 0.005) in the FIPD with advancing age. Between 7±15 and 16±25 years there was a 4.7% rise (57.55 to 60.27 mm), and between 16± 25 and 26±40 years a further 1.1% increase (to 60.90 mm). Unfortunately individuals beyond forty years of age were not studied. A subsequent study of the FIPD of Saudi males of Arab origin (Osuobeni and Al-Musa, 1993) used dierent age groupings and does not provide a satisfactory comparison with the earlier female Arab material. Nevertheless, taking these male data in isolation, a substantial increase in the FIPD was recorded between childhood and young adulthood but registered only a small change thereafter. The database of FIPD values established here (refer to Tables 1A and 3) was obtained from a subject group comparable in pro®le to that utilised by Fledelius and Stubgaard (1986): i.e., balanced numbers of age-matched healthy Caucasian males and females. However, the subject age range was intentionally restricted to the ®ve decades between 16±65 years and comprised many more subjects than in the earlier Danish study. For either gender there was revealed an approximately 3% increase in the mean FIPD value between adolescence (16±25 years) and later middleage (41±65 years): female = 2.7%, male = 2.9%. However, as also suggested in the data of Fledelius and Stubgaard (1986) a dierent pattern of change could be detected between genders. In the female data the increase was continuous, from 16±25 to 26±40
Increase from preceding age group Absolute (mm)
%
p-level
Ð Ð +1.58 +0.77 +0.27 +0.86
Ð Ð +2.48 +1.26 +0.41 +1.39
Ð Ð 0.00005 0.01 0.5 0.005
years (a 1.3% rise) and from 26±40 to 41±65 years (a further 1.4% rise). In the male data the increase was substantially accomplished by early middle age (a 2.5% rise between 16±25 and 26±40 years) and increased only very slightly thereafter (a 0.4% rise to 41±65 years). Consequently, it would appear that the FIPD parameter in human females, in contrast to human males, continues to increase not merely post-puberty but throughout adult life. This seems to be a feature which has not been commented upon by previous investigators in this area of human facial anthropometry. Whether this is a gender-speci®c issue which can be related to developments in the underlying cranial skeletal anatomy or to soft (adipose) tissue changes in the orbital areaÐor a combination of bothÐremains unclear, and necessitates further specialist investigation. However a brief review of certain aspects of human facial developmental anatomy might be useful in this regard. Age and gender changes in the human facial skeleton The cranial bony anatomy must be the primary determinant of the FIPD, with subject age and gender in¯uencing the magnitude of this parameter at any given time of life. It is recognised (Whitnall, 1921; Wol, 1968) that in the infant the inter-orbital linear skeletal dimension is small (especially relative to the size of the globe, this often contributing to the appearance of pseudo-squint) but increases with the development of the nasal cavity, the frontal and ethmoidal sinuses and with the eruption of the teeth. The secondary ossi®cation changes beyond puberty establish a gender dierence in the human facial skeleton. The female retains the ``rounded'' skull appearance of the infant, with sharp-edged and almost circular orbits (width approximately equal to height of orbital entrance). This is quite dierent to the male development in the brow region, where increasing prominence adjacent to the supra-orbital margin combined with an increase in width of the facial bones produces a horizontally-rectangular orbital entrance.
324
Ophthal. Physiol. Opt. 1999 19: No 4
The foremost publication on the anatomy of the human orbit (Whitnall, 1921) considers that the full dimensions of the bony orbit itself are only attained at puberty. The possibility of subsequent growth or variation during adulthood is not mentioned, aside from the likelihood of changes around the orbital margin in the elderly individual as a consequence of bone atrophy. Two skeletal orbit-related dimensions in the horizontal frontal plane which may possibly be associated with, or have an in¯uence on, the magnitude of the FIPD are the inter- and extra-orbital widths. To the present author's knowledge there have not been published reference tables of human skeletal mensuration detailing gender- or age-grouped variations in either of these linear dimensions. Certain material as is available is presented here in Table 4 and Figure 3.
Any discussion of the in¯uence of the inter-orbital width (lowest pair of curves in Figure 3) upon the magnitude of the FIPD in living subjects is limited by the restricted age range of the published data (Hansman, 1966; Table 1A and B, pp. 90±91). A parallel and steady increase is evident for either gender (absolute value male>female) from infancy through to puberty (ca 15 years of age). Over the subsequent decade of early adult life there is evidence of the attainment of a plateau in the magnitude of this parameter in female subjects, whereas in males the inter-orbital width continues to increase. This gender dierence is undoubtedly a direct re¯ection of the secondary ossi®cation changes in the orbital and brow regions postpuberty, as previously described. The rather more extensive data (Fledelius and Stubgaard, 1986: Tables 3 and 4, p. 484) describing the
Table 4. Orbital-separation mensuration: certain published dimensions of human facial anatomy Parameter Subject (Method) Extra-orbital width* Skeletal (calipers?) Living
(Hertel)
Living
(Hertel)
Inter-orbital width** Skeletal (calipers?) Living
(X-ray)
Dimension (mm)
Gender
Age (years)
Nationality
Source
80.8 99.7 96.0 97.5 94.4 88.32 3.8 85.62 3.8 94.12 4.0 92.22 3.7 98.52 3.6 93.52 2.9 99.12 3.3 94.22 3.9 97.82 2.9 93.22 3.2 95.62 3.4 94.12 4.8
Ð Male Female Male Female Male Female Male Female Male Female Male Female Male Female Male Female
10±17 20±67 23±67 Adult
Ð Ð Ð Caucasian
Emmert (1880): cited by Whitnall (1921) Knudtzon (1949)
5±10
Danish
Fledelius and Stubgaard (1986)
19.5±30.7 18.5±29.5 18.22 *** 17.84 20.73 20.07 23.57 23.13 26.50 25.24 27.78 25.87 28.43 25.83
Male Female Male Female Male Female Male Female Male Female Male Female Male Female
Adult
German
Gunther (1933): cited by Freihofer (1980)
2
N. American
Hansman (1966)
11±19 20±24 25±30 31±55 >55
5 10 15 20 24
*Horizontal distance between the lateral margins of the orbits.**Horizontal distance between the medial margins of the orbits.***50th percentile value.ÐIndicates information not specified
Far interpupillary distance: J. S. Pointer
325
Figure 3. The variation with age and gender of three human facial dimensions in the horizontal frontal plane of living subjects. The uppermost pair of curves (diamond symbols) indicate the mean extra-orbital width (Fledelius and Stubgaard, 1986), and the lowest truncated pair of curves (triangular symbols) the inter-orbital width (50th percentile values: Hansman, 1966): refer to Table 4. The two sets of data across the centre of the figure indicate the FIPD data of Fledelius and Stubgaard (1986: square symbols linked by continuous line) and Pointer (refer to Table 3; circular symbols linked by dotted line). For all three parameters the solid black symbols indicate male and the unfilled symbols female mean values, with error bars indicating + or ÿ 1st SD. Each FIPD value plotted for the Danish study (square symbols) represents the mean result obtained from between 17±36 subjects: the other six values (three pairs of circular symbols) are each the mean of measurements upon 150 subjects with, for each of these three age groups, male values being statistically significantly greater than female values. There is a marked increase in the FIPD for both genders during childhood and adolescence, but this trend becomes much reduced during adulthood and into later middle-age, especially in male subjects (see text).
variation of extra-orbital width in living subjects (highest pair of curves in Figure 3) also shows a parallel increase in this parameter for both genders (absolute value male>female) up to the mid-teens. Subsequently females display a much reduced rise to the mid-twenties, and no substantial variation thereafter. In male
subjects the adolescent increase is maintained through to the end of the second decade of life: thereafter the extra-orbital width stabilizes and then appears to decline through middle age, eventually attaining a value only slightly greater than that of similarly-aged female subjects.
326
Ophthal. Physiol. Opt. 1999 19: No 4
From the foregoing description of the longitudinal trends in two skeletal orbit-related dimensions which might be expected to have an in¯uence on the magnitude of the FIPD, it will be appreciated that for neither gender do they adequately describe the variation in the FIPD parameter throughout adult life as recorded here (see family of curves across the centre of Figure 3). Consequently, on the basis of the available evidence, determination of the dependence of the FIPD upon speci®c skeletal mensuration details remains unresolved. The inference must be that features other than solely the facial bony anatomy in¯uence the FIPD parameter. A possible candidate in this regard could be an age/gender dierence in volume of the orbital soft tissue. Unfortunately, however, even less well documented than age-related changes in the cranial bony structures are growth and developmental alterations in the soft tissues of the human orbital regionÐincluding, for example, changes arising from senile laxity of muscle tone and reduced adipose tissue volume within and around the orbit. This paucity of information is no doubt a re¯ection of the clinical applications that such quantitative material would have, even for the diagnostician or the restorative facial surgeon (Freihofer, 1980). Until such time as further gender-segregated and age-grouped mensuration details relating to the human facial skeleton are forthcoming, the variation in the magnitude of the FIPD described here for Caucasian subjects remains the only (indirect) evidence of a gender-speci®c pattern of anatomical development in the area of the orbits in the young and mature human adult.
References Anderson, A. L. (1954). Accurate clinical means of measuring intervisual axis distance. Arch. Ophthalmol. 52, 349±352. Bruckner, R., Batschelet, E. and Hugenschmidt, F. (1987). The Basel longitudinal study on aging (1955±1978). Ophthalmogerontological research results. Doc. Ophthalmol. 64, 235±310.
Cesario, V. A. and Latta, G. H. (1984). Relationship between the mesiodistal width of the maxillary central incisor and interpupillary distance. J. Prosthet. Dent. 52, 641±643. Emsley, H. H. (1952). Visual Optics. Vol. 1: Optics of Vision, 5th edn, Butterworths, London. Fledelius, H. C. and Stubgaard, M. (1986). Changes in eye position during growth and adult life as based on exophthalmometry, interpupillary distance, and orbital distance measurements. Acta Ophthalmol. (Kbh) 64, 481± 486. Freihofer, H. P. M. (1980). Inner intercanthal and interorbital distances. J. Max.-Fac. Surg. 8, 324±326. Hansman, C. F. (1966). Growth of interorbital distance and skull thickness as observed in roentgenographic measurements. Radiol. 86, 87±96. Harvey, R. S. (1982). Some statistics of interpupillary distance. Optician 184, 29(No. 4766: 12th November). Kaye, J. and Obstfeld, H. (1989). Anthropometry for children's spectacle frames. Ophthal. Physiol. Opt. 9, 293± 298. Knudtzon, K. (1949). On exophthalmometry. The result of 724 measurements with Hertel's exophthalmometer on normal adult individuals. Acta Psychiat. Neurol. Scand. 24, 523±537. Osuobeni, E. P. and Al-Musa, K. A. (1993). Gender dierences in interpupillary distance among Arabs. Optom. Vis. Sci. 70, 1027±1030. Osuobeni, E. P. and Faden, F. K. (1993). Interpupillary distance of females of Arab origin. Optom. Vis. Sci. 70, 244±247. Pointer, J. S. (1995a). The presbyopic add. II Age-related trend and a gender dierence. Ophthal. Physiol. Opt. 15, 241±248. Pointer, J. S. (1995b). The presbyopic add. III. In¯uence of the distance refractive type. Ophthal. Physiol. Opt. 15, 249± 253. Quant, J. R. and Woo, G. C. (1992). Normal values of eye position in the Chinese population of Hong Kong. Optom. Vis. Sci. 69, 152±158. Roy, F. H. (1985). Ocular Syndromes and Systemic Diseases, Grune & Stratton, Philadelphia. Sasieni, L. S. (1975). The Principles and Practice of Optical Dispensing and Fitting, 3rd edn, Butterworths, London. Whitnall, S. E. (1921). The Anatomy of the Human Orbit and Accessory Organs of Vision, Henry Frowde Hodder & Stoughton, London. Wol, E. (1968). Anatomy of the Eye and Orbit, 6th edn.; revised by Last R.J. Lewis, London.
PII: S0275-5408(98)00082-9
The far interpupillary distance. A genderspecific variation with advancing age Jonathan S. Pointer Optometric Research, 4A Market Square, Higham Ferrers, Northants NN10 8BP, UK Summary A knowledge of the magnitude of the far interpupillary distance (FIPD) in relation to a specific population is of clinical, practical and theoretical interest. A FIPD database is presented here, comprising material collated from the spectacle dispensing records of n = 1800 subjects seen in routine optometric practice. All measurements were taken by the author on healthy Caucasian (white, Northern European) males and females. The data were equi-partitioned either across three age bands (16±25, 26±40, 41±65 years: mixed refractive types, total n = 900) or between the three distance refractive types (emmetropia, hypermetropia, myopia: all subjects aged between 41±65 years, total n = 900). A consistent gender difference (male>female) was confirmed throughout this material, irrespective of age group; refractive type, however, had no influence on the magnitude of this facial parameter. Summary results of this anthropometric survey are presented in tabular form, facilitating reference by ophthalmic and dental clinicians and by the designers of binocular optical instruments. There was also revealed evidence of a gender-specific pattern of change in the FIPD variable with advancing age. An approximately 3% increase in the magnitude of the human FIPD from the mid-teens to later middle age was attained in males by early middle age, being little altered thereafter: in contrast females continued to record an increase in this facial parameter into later middle age. An explanation for this hitherto unremarked feature of human facial anthropometry might be sought in the gender-specific changes post-puberty of the cranial skeletal anatomy and in the soft tissues of the orbital region. # 1999 The College of Optometrists. Published by Elsevier Science Ltd. All rights reserved
Introduction
the eye's optic axis rather than the line of sight which is approximately centred on the pupillary aperture. This decentration is occasioned by the infero-temporal oset of the fovea, itself a feature of possible teleological signi®cance given that it minimises chromatic aberration in the human refractive system. Consequently the optic axis lies temporal to the visual axis: the two axes intersect at the (single) nodal point of the reduced eye, the angle alpha between them amounting to ca 5 deg, being less in myopic and greater in hypermetropic and infants' eyes (Emsley, 1952). Practical determination of this angular displacement using a point source of light and an arc perimeter, for example, results in the measurement of angle kappa (Landolt: Emsley, 1952); this is the angle formed between the visual axis and the pupillary line (an approximation of the optic axis, sited perpendicular to the cornea). Although the pupil centre is oset slightly nasally in relation to the corneal centre, the dierence
The far interpupillary distance (FIPD) is the facial measurement (mm) in the horizontal plane between the geometric centres of the pupillary apertures of a pair of eyes, which latter are focused at optical in®nity. This measurement is typically taken across the upper bridge of a subject's nose, in the approximate spectacle (frontal) plane (Sasieni, 1975). To minimise the introduction of undesirable optical aberration and prismatic eects when wearing spectacles or when using an optical instrument, the line of sight of each of an individual's pair of eyes should ideally pass through the optical centre of each of a pair of ophthalmic lenses. But as has been appreciated at least since the time of Helmholtz (Emsley, 1952) it is Received: 3 August 1998 Revised form: 24 November 1998
317
318
Ophthal. Physiol. Opt. 1999 19: No 4
in magnitude between the two angular measurements (alpha and kappa) is clinically insigni®cant. Further, given the practical ease with which the pupillary centres (or rather, the relative linear displacement between complementary points on the two eyes) can be estimated, ophthalmic lenses are invariably aligned with the ``optic'' (pupillary) axes rather than the ``visual'' axes of an individual: whence the IPD. The magnitude of the FIPD is of basic and clinical signi®cance. For a given population type, anthropometric surveys can establish descriptive statistics for this measurement: the FIPD then becomes a useful diagnostic tool for developmental anatomists, geneticists and cranio-facial surgeons (e.g., Roy, 1985), a practical aid in restorative facial surgery and in the sphere of prosthodontics (e.g., Cesario and Latta, 1984), a necessary parameter in the design of binocular instruments and equipment (Harvey, 1982) and, not least, in the allied commercial (frame design and lens blank manufacture) and clinical (ophthalmic dispensing) ®elds of optometry. Osuobeni and Al-Musa (1993) have tabulated IPD values from several published sources, facilitating comparisons between gender, age band and ethnic group. Sasieni (1975) quotes average values for the FIPD in the adult (assumed Caucasian/European) subject as follows: males = 63±64 mm (variation 58±72), females = 60±61 mm (variation 57±65). A new FIPD database is presented here, comprising material generated in routine optometric practice and compiled in a balanced format. Statistical analysis of this material permits speci®c conclusions to be drawn with regard to the distribution and characteristics of the FIPD in a contemporary Caucasian population. In addition certain anatomical considerations underlying age-related variations in this parameter are explored.
Method and materials Viktorin's method The FIPD was measured routinely by the author, using a ruler calibrated in mm, after the method of Viktorin (see Sasieni, 1975). Despite the several objections to the absolute measurement accuracy of this technique (summarised by Anderson, 1954) its facility has ensured that it has remained perhaps the most widely taught and practised interocular measurement procedure: it is also the method most frequently cited in published investigations involving the assessment of this facial parameter (Osuobeni and Al-Musa, 1993). All of the FIPD material collated herein was obtained in the following manner. The subject was seated opposite to, on the same level as and within arm's length of the practitioner in a well-illuminated
ophthalmic dispensing area. A ¯at calibrated (mm) gauge was laid across the subject's nose and lightly supported by the practitioner's left hand. The subject was instructed to keep both eyes open and to ®xate the practitioner's open left eye; using this eye the practitioner aligned the ``zero'' mark on the gauge with the temporal limbus of the subject's right eye. Keeping the ruler still, the practitioner then closed his left eye and opened his right, at the same time instructing the subject to move ®xation across to the practitioner's open right eye. The position of the nasal limbus of the patient's left eye could then be read on the ruler: given that approximately straight and parallel lines of sight were maintained throughout this procedure, this calibration value was taken as the FIPD measurement (mm). Of course this approach ignores the error associated with angle alpha (see Introduction), and also any inaccuracies introduced by parallax arising from a marked dierence between the subject's and practitioner's IPD and inequalities in the vertical plane due to stature or sitting position. However, in practice these theoretical inaccuracies are small (Sasieni, 1975). Subject groups FIPD values (mm), as measured by the author on n = 900 male and n = 900 female Caucasian (white, Northern European) subjects, were taken at random but in accordance with certain criteria (see below) from the spectacle dispensing records held at the author's optometric practice. Data compilation was such that equal subject numbers were assembled across the various categories: all material required for this report (i.e., subject gender, age group and FIPD value) was entirely non-attributable in analysis and patient con®dentiality was not compromised. All subjects were between the ages of 16 and 65 years at the time of the spectacle dispensing. The choice of these speci®c age limits was with the intention of minimising mensuration uncertainties associated with, on the one hand childhood physical growth changes (a source of confounding error encountered in an IPD study undertaken by Osuobeni and Faden, 1993) and on the other, the increased laxity of the soft tissues around the orbit of the elderly subject (Fledelius and Stubgaard, 1986; Whitnall, 1921). In addition, given the possibility of systematic variation in the dimensions of the globe of the eye (cf. myopic vs hypermetropic axial length dierences, and the in¯uence of degree of myopia upon the amount of exophthalmos: Quant and Woo, 1992) distance refractive classi®cation was also recorded. Two data groups were assembled: Group A comprised three age bands (16±25, 26±40, 41±65 years) of unspeci®ed (i.e., mixed)
Far interpupillary distance: J. S. Pointer distance refractive type with n = 150 male and n = 150 female FIPD values for each age band (i.e., total n = 900): Group B comprised material from subjects aged 41±65 years equally partitioned between the three refractive types (emmetropoia, hypermetropia, myopia) with n = 150 male and n = 150 female FIPD values for each refractive classi®cation (i.e., total n = 900). Data selection criteria were not extensive. All subjects were in good health, recorded an acuity of 6/6 or better and had no binocular vision anomalies; distance refractive corrections >2 6.00 DS and > ÿ 2.00 DC barred the subject's data from inclusion in the sample, as did anisometropia of the spherical elements >2.00 D. In accord with a previous publication (Pointer, 1995b), emmetropia was de®ned as the refractive range ÿ0.25 D to +0.75 D spherical equivalent refraction (SER: sphere power plus one-half of any cylinder power, at the spectacle plane); hypermetropia included individuals whose spectacle correction was r + 0.875 D SER; and myopia covered individuals whose spectacle correction was rÿ 0.375 D SER. All facial measurements had been taken by the author as a prerequisite to the dispensing of a complete pair of (distance vision) spectacles: no ``special requirement'' or occupational dispensings were included in the data sets. Data from each subject were only entered once, so the results represent material from n = 1800 separate individuals.
Results Descriptive statistics1 for the two groups of FIPD data are given in Table 1A, B. The magnitude, both of the range and of the three measures of central tendency associated with the distribution of the FIPD variable, is seen to be consistently greater in male than in female subjects. A degree of variation with age group (Table 1A)Ðbut not with refractive type (Table 1B)Ðis also evident. The small values of the coecients of skewness and kurtosis are indicative of the FIPD variable being normally distributed in each of these data groupings. Probability plots (not shown here) typically indicated normality for at least two standard deviations (SD) either side of the mean value, with cumulative frequency distributions yielding an ogive. Data of male and female subjects in the 16±25 and 26±40 years age groups display a small positive skew (Table 1A), with the exception of a very small negative skew in the material relating to 16±25 year old 1 Data management and analysis handled by STATISTICA/Mac software (Release 3, 1992: StatSoft Inc., Tulsa, Oklahoma 74104, USA) on Apple Macintosh LCII computer hardware running System 7.
319
malesÐwhich data also have the highest SD. Taken together these two statistical features suggest a wider variation of FIPD measurements (including a few low values) in adolescent males as compared to adolescent females and to adults of either gender. Certainly the range of FIPD values is greatest for this particular subject group (16 mm as compared to 13±15 mm for all of the other gender/age groups: Table 1A). Consequently these speci®c observations probably re¯ect the earlier physical maturation of the human female as compared to the male teenager, a feature also recorded by Fledelius and Stubgaard (1986). Both genders show a small negative skew for the oldest age band represented here (41±65 years), indicating that this age grouping contains a few low FIPD values. This tendency to a small negative skew largely persists in the data of the three refractive types (Table 1B), this material also having been collated from subjects in the 41±65 years age band: this uniformity is taken as being indicative of the internal consistency of the new FIPD database presented here.
Analysis Student's t-test for independent samples was run on all available male and female age and refractive type group pairings: the results are summarised in Table 2A and B. Irrespective of age group or refractive type, the FIPD of males is always statistically signi®cantly dierent (larger) compared to that of females. With regard to refractive type (Table 2B), a genderbased dierence in the mean FIPD is greatest between hypermetropic males and females (3.34 mm) and least between myopes (2.50 mm), with emmetropes falling between the two (2.82 mm). However, cross-pairing of FIPD values between refractive types fails to attain statistical signi®cance within either gender. The picture emerging from the results of the pairwise analyses of the age-grouped data is more complex (Table 2A). The largest gender-based dierence in mean FIPD values lies within those individuals aged 26±40 years (3.56 mm): the adolescents (16±25 years) and the oldest individuals (41±65 years) both display smaller gender-based dierences (2.75 and 2.97 mm, respectively). Within each gender the pattern of transition across age groups diers. In males the mean FIPD of the 16±25 years group is statistically signi®cantly dierent (smaller) compared to that of both the 26±40 and the 41±65 years groupings (by 1.58 and 1.85 mm, respectively), but which latter are not statistically signi®cantly dierent to one another (0.27 mm). By contrast, in females the mean FIPD of the 26±40 years group is statistically signi®cantly dierent to the mean value of this parameter in both 16±25 year olds
320
Ophthal. Physiol. Opt. 1999 19: No 4
Table 1. The far interpupillary distance (FIPD): magnitude (*mm) and summary statistics 1A. Variation with age group and gender (mixed distance refractive types) Range* Percentile* Age group (years)
Gender
n
Min
Max
5th
95th
Mean 2 1st SD*
Median*
Mode*
Skewness
Kurtosis
16±25
Male Female Male Female Male Female
150 150 150 150 150 150
56 54 58 55 58 55
72 67 73 68 73 70
58.06 57.20 60.14 57.33 60.37 58.00
69.58 65.58 70.32 66.58 70.00 66.41
63.65 2 3.38 60.90 2 2.53 65.23 2 3.14 61.67 2 2.72 65.50 2 2.77 62.53 2 2.76
64 61 65 61 66 63
66 60 66 61 66 61
ÿ0.01 +0.23 +0.29 +0.12 ÿ0.32 ÿ0.17
ÿ0.06 +0.01 ÿ0.13 ÿ0.22 +0.18 ÿ0.28
26±40 41±65
1B. Variation with refractive type and gender (age group: 41±65 years) Range* Percentile* Refractive type
Gender
n
Min
Max
5th
95th
Mean 2 1st SD*
Median*
Mode*
Skewness
Kurtosis
Myopia
Male Female Male Female Male Female Male Female
150 150 150 150 150 150 450 450
59 57 59 56 58 55 58 55
74 70 72 71 73 70 74 71
60.78 58.14 59.87 58.66 60.69 58.55 60.58 58.52
70.37 67.00 70.00 66.67 70.15 66.46 70.14 66.75
65.35 2 2.80 62.85 2 2.72 65.33 2 2.95 62.51 2 2.75 65.75 2 2.94 62.41 2 2.70 65.48 2 2.90 62.59 2 2.73
65 63 66 63 66 62 66 63
65 61 66 61 66 60 65 61
+0.03 ÿ0.04 ÿ0.24 ÿ0.26 ÿ0.32 +0.09 ÿ0.10 ÿ0.10
+0.28 ÿ0.05 ÿ0.04 ÿ0.09 ÿ0.28 ÿ0.06 ÿ0.22 ÿ0.26
Emmetopia Hyperopia Combined
(0.77 mm) and 41±65 year olds (0.86 mm): however, like males, the dierence between the mean FIPD of 16±25 and 41±65 year old individuals (1.63 mm) is statistically signi®cant.
Discussion I. The in¯uence of gender on the FIPD To the present author's knowledge there has been only one comparable investigation into the variation
Table 2. The FIPD: results of a series of Student's t-tests for independent sample comparisons 2A. Influence of age group and gender
Difference between means Absolute (mm)
%
Student's t-value
p-level
Male 16±25 vs Female 16±25 Male 26±40 vs Female 26±40 Male 41±65 vs Female 41±65 Male 16±25 vs Male 26±40 Male 16±25 vs Male 41±65 Male 26±40 vs Male 41±65 Female 16±25 vs Female 26±40 Female 16±25 vs Female 41±65 Female 26±40 vs Female 41±65
+2.75 +3.56 +2.97 ÿ1.58 ÿ1.85 ÿ0.27 ÿ0.77 ÿ1.63 ÿ0.86
+4.52 +5.77 +4.75 ÿ2.42 ÿ2.82 ÿ0.41 ÿ1.25 ÿ2.61 ÿ1.37
+7.98 +10.52 +9.30 ÿ4.19 ÿ5.17 ÿ0.78 ÿ2.52 ÿ5.34 ÿ2.74
0.0000001 0.0000001 0.0000001 0.00005 0.0000005 0.5 0.01 0.0000005 0.005
2B. Influence of refractive type and gender Male MY vs Female MY Male EM vs Female EM Male HY vs Female HY Male MY vs Male EM Male MY vs Male HY Male EM vs Male HY Female MY vs Female EM Female MY vs Female HY Female EM vs Female HY Male ALL vs Female ALL
+2.50 +2.82 +3.34 +0.02 ÿ0.40 ÿ0.42 +0.34 +0.44 +0.10 +2.89
+3.98 +4.51 +5.35 +0.03 ÿ0.61 ÿ0.64 +0.54 +0.70 +0.16 +4.62
+7.81 +8.56 +10.26 +0.04 ÿ1.23 ÿ1.23 +1.08 +1.43 +0.34 +15.38
0.0000001 0.0000001 0.0000001 0.9 0.5 0.5 0.5 0.5 0.9 0.0000001
Inter group comparison
Far interpupillary distance: J. S. Pointer of the FIPD measurement in Caucasian subjects: as an adjunct to a study of changes in eye position during life, Fledelius and Stubgaard (1986) measured the FIPD in healthy male and female subjects between the ages of 5 and 80 years. The larger subject numbers in this present study corroborate the conclusions of the earlier Danish investigation and contribute an additional ®nding. Speci®cally, it is substantiated (i) that the magnitude of the FIPD is always greater in males than females of the same age group, and (ii) that the absolute value of the FIPD increases with age, at least until the third decade of life (for a further discussion of this feature see below: Discussion II). The mean FIPD values obtained for male and female subjects in these two studies across broad age bands spanning childhood, adolescence and adulthood are presented chronologically in Figure 1. The age-related increase in the absolute value of the FIPD is evident for either gender at least into adulthood; for both series of data the plotted male and female mean values are statistically signi®cantly dierent at p = 0.01 or greater. This consistent gender dierence is demonstrated in Figure 2 where, for equal numbers of male and female subjects (n = 450 each) aged 41±65 years, cumulative frequency ogives of the FIPD values are plotted. It is evident that in age-matched subjects the
321
male and female FIPD parameter diers only by a simple linear transposition along the x-axis: male values are consistently ca 3 mm larger than those of females of the same age group, irrespective of the absolute magnitude of the FIPD variable. The in¯uence of gender on the magnitude of the human interpupillary separation is entirely predictable. On the basis of the FIPD data presented herein, female values are typically ca 95% of the values recorded for similarly-aged males. This result accords with established gender dierences in human anatomical and skeletal dimensions as given in post-mortem and forensic tables, for example. Indeed such a genderbased physical (size) dierence has recently been implicated (Pointer, 1995a) in an explanation for the clinical observation that a presbyopic female typically requires a positive reading addition of marginally greater magnitude than an age-matched male subjectÐessentially because a woman's arms are shorter, producing a slightly closer reading position than a man. An additional ®nding presented here is that an individual's refractive typeÐemmetropia, hypermetropia or myopiaÐhas no in¯uence on the magnitude of the FIPD in mature adult Caucasian subjects. Evidently for low to moderate degrees of refractive error clinical measurement of eye position in the frontal (horizontal)
Figure 1. Variation of the FIPD with age group and gender. The published data of Fledelius and Stubgaard (1986: Table 4, p. 484) are identified by the square symbols, the new data of Pointer (refer to Table 1A) are represented by the circular symbols. In each case the solid black symbols indicate male and the unfilled symbols female mean values, with error bars indicating, respectively, + or ÿ 1st SD. Note the discontinuity midway along the x-axis, occasioned by the different age ranges studied in the two reports. Each mean value plotted from the Danish study was obtained from either ca 50 subjects (5±10 years) or ca 70 subjects (11±19 years): all other values are each the mean of measurements upon 150 subjects. At each age group the gender difference is statistically significant at p = 0.01 or greater.
322
Ophthal. Physiol. Opt. 1999 19: No 4
Figure 2. The gender difference in FIPD. Cumulative frequency/percentage ogives for male (solid black symbols: n = 450) and female (unfilled symbols: n = 450) FIPD values (combined data for all three refractive types, age group 41±65 years: refer to Table 1B). The normal distribution of the FIPD material for either gender is demonstrated, with the two ogives being separated by a simple linear displacement of ca 3 scale units (mm) along the abscissa, male values being the greater.
planeÐunlike the assessment of orbital protrusion taken orthogonal to this (Quant and Woo, 1992)Ðis not in¯uenced by the axial length or size of the globe within the bony orbit. A reference table of human FIPD values It is the intention that the FIPD information presented here relating to a broad age range of male and female Caucasian subjects will be a useful reference source. The information is relevant not only within the ®eld of ophthalmic optics but also to clinicians in the dental and prosthetic/restorative surgical specialities. In addition, the inclusion of such details as the magnitude of the 5th and 95th percentile values in Table 1A and B may be of assistance to designers of binocular optical instruments and vision systems (Harvey, 1982): such equipment, whether intended for clinical, industrial or defence application, must have an adequate adjustment range appropriate to the speci®c group of personnel who will be using that equipment.
Discussion II. A gender-speci®c variation of the FIPD with age Studies which have quanti®ed the FIPD variable in children and young subjects over a period of time have
reported an increase in magnitude with increasing years (Kaye and Obstfeld, 1989; Osuobeni and Faden, 1993). Many investigators appear to have made the implicit assumption that this parameter attains a plateau at some point between the second and third decades of life (see Osuobeni and Faden, 1993). Certainly in a longitudinal biometric study of a group of Caucasian male subjects Bruckner et al. (1987) could ®nd no increase in FIPD beyond 30 years of age, with the majority of subjects attaining a stable value in their late teens or early twenties. Fledelius and Stubgaard (1986) reported a consistent gender dierence (male>female) not only in the absolute value of the FIPD in age-matched Caucasian subject groups but also in the age at which the FIPD attained the ``adult'' level. In line with observations on general bodily growth and maturation, human female subjects attained the ``adult'' FIPD dimension around 14±16 years of age (61.8 mm) and males at around 17± 19 years (65.8 mm). However, further study of the tabulated FIPD values for male and female subject groups of increasing age in the publication of Fledelius and Stubgaard (1986: Table 3, p. 484) reveals an interesting subsidiary gender dierence. Beyond the late teenage years female subjects in particular recorded an increase in the FIPD. This increase was slight in the 31±55 years age band (1.8% rise, to 62.9 mm) but
Far interpupillary distance: J. S. Pointer
323
Table 3. The FIPD: variation with age group and gender (data from Table 1A) Age group (years)
Gender
n
Mean 2 1st SD (mm)
16±25
Male Female Male Female Male Female
150 150 150 150 150 150
63.65 2 3.38 60.90 2 2.53 65.23 2 3.14 61.67 2 2.72 65.50 2 2.77 62.53 2 2.76
26±40 41±65
more marked in the subject group aged >55 years (a further 3.5% rise, to 65.1 mm). Male values increased only slightly in the 31±55 years age range from the level of FIPD reached in the late teens (1.0% rise, to 66.4 mm) and actually reduced (0.6% drop, to 66.0 mm) within the >55 years age grouping. The authors made no speci®c comment regarding this particular gender dierence in their data, aside from remarking that age-related soft tissue changes could account for slight alterations in orbital/facial measurements in elderly subjects. A study restricted to females of Arab origin (Osuobeni and Faden, 1993) also reported an increase (statistically signi®cant, pR 0.005) in the FIPD with advancing age. Between 7±15 and 16±25 years there was a 4.7% rise (57.55 to 60.27 mm), and between 16± 25 and 26±40 years a further 1.1% increase (to 60.90 mm). Unfortunately individuals beyond forty years of age were not studied. A subsequent study of the FIPD of Saudi males of Arab origin (Osuobeni and Al-Musa, 1993) used dierent age groupings and does not provide a satisfactory comparison with the earlier female Arab material. Nevertheless, taking these male data in isolation, a substantial increase in the FIPD was recorded between childhood and young adulthood but registered only a small change thereafter. The database of FIPD values established here (refer to Tables 1A and 3) was obtained from a subject group comparable in pro®le to that utilised by Fledelius and Stubgaard (1986): i.e., balanced numbers of age-matched healthy Caucasian males and females. However, the subject age range was intentionally restricted to the ®ve decades between 16±65 years and comprised many more subjects than in the earlier Danish study. For either gender there was revealed an approximately 3% increase in the mean FIPD value between adolescence (16±25 years) and later middleage (41±65 years): female = 2.7%, male = 2.9%. However, as also suggested in the data of Fledelius and Stubgaard (1986) a dierent pattern of change could be detected between genders. In the female data the increase was continuous, from 16±25 to 26±40
Increase from preceding age group Absolute (mm)
%
p-level
Ð Ð +1.58 +0.77 +0.27 +0.86
Ð Ð +2.48 +1.26 +0.41 +1.39
Ð Ð 0.00005 0.01 0.5 0.005
years (a 1.3% rise) and from 26±40 to 41±65 years (a further 1.4% rise). In the male data the increase was substantially accomplished by early middle age (a 2.5% rise between 16±25 and 26±40 years) and increased only very slightly thereafter (a 0.4% rise to 41±65 years). Consequently, it would appear that the FIPD parameter in human females, in contrast to human males, continues to increase not merely post-puberty but throughout adult life. This seems to be a feature which has not been commented upon by previous investigators in this area of human facial anthropometry. Whether this is a gender-speci®c issue which can be related to developments in the underlying cranial skeletal anatomy or to soft (adipose) tissue changes in the orbital areaÐor a combination of bothÐremains unclear, and necessitates further specialist investigation. However a brief review of certain aspects of human facial developmental anatomy might be useful in this regard. Age and gender changes in the human facial skeleton The cranial bony anatomy must be the primary determinant of the FIPD, with subject age and gender in¯uencing the magnitude of this parameter at any given time of life. It is recognised (Whitnall, 1921; Wol, 1968) that in the infant the inter-orbital linear skeletal dimension is small (especially relative to the size of the globe, this often contributing to the appearance of pseudo-squint) but increases with the development of the nasal cavity, the frontal and ethmoidal sinuses and with the eruption of the teeth. The secondary ossi®cation changes beyond puberty establish a gender dierence in the human facial skeleton. The female retains the ``rounded'' skull appearance of the infant, with sharp-edged and almost circular orbits (width approximately equal to height of orbital entrance). This is quite dierent to the male development in the brow region, where increasing prominence adjacent to the supra-orbital margin combined with an increase in width of the facial bones produces a horizontally-rectangular orbital entrance.
324
Ophthal. Physiol. Opt. 1999 19: No 4
The foremost publication on the anatomy of the human orbit (Whitnall, 1921) considers that the full dimensions of the bony orbit itself are only attained at puberty. The possibility of subsequent growth or variation during adulthood is not mentioned, aside from the likelihood of changes around the orbital margin in the elderly individual as a consequence of bone atrophy. Two skeletal orbit-related dimensions in the horizontal frontal plane which may possibly be associated with, or have an in¯uence on, the magnitude of the FIPD are the inter- and extra-orbital widths. To the present author's knowledge there have not been published reference tables of human skeletal mensuration detailing gender- or age-grouped variations in either of these linear dimensions. Certain material as is available is presented here in Table 4 and Figure 3.
Any discussion of the in¯uence of the inter-orbital width (lowest pair of curves in Figure 3) upon the magnitude of the FIPD in living subjects is limited by the restricted age range of the published data (Hansman, 1966; Table 1A and B, pp. 90±91). A parallel and steady increase is evident for either gender (absolute value male>female) from infancy through to puberty (ca 15 years of age). Over the subsequent decade of early adult life there is evidence of the attainment of a plateau in the magnitude of this parameter in female subjects, whereas in males the inter-orbital width continues to increase. This gender dierence is undoubtedly a direct re¯ection of the secondary ossi®cation changes in the orbital and brow regions postpuberty, as previously described. The rather more extensive data (Fledelius and Stubgaard, 1986: Tables 3 and 4, p. 484) describing the
Table 4. Orbital-separation mensuration: certain published dimensions of human facial anatomy Parameter Subject (Method) Extra-orbital width* Skeletal (calipers?) Living
(Hertel)
Living
(Hertel)
Inter-orbital width** Skeletal (calipers?) Living
(X-ray)
Dimension (mm)
Gender
Age (years)
Nationality
Source
80.8 99.7 96.0 97.5 94.4 88.32 3.8 85.62 3.8 94.12 4.0 92.22 3.7 98.52 3.6 93.52 2.9 99.12 3.3 94.22 3.9 97.82 2.9 93.22 3.2 95.62 3.4 94.12 4.8
Ð Male Female Male Female Male Female Male Female Male Female Male Female Male Female Male Female
10±17 20±67 23±67 Adult
Ð Ð Ð Caucasian
Emmert (1880): cited by Whitnall (1921) Knudtzon (1949)
5±10
Danish
Fledelius and Stubgaard (1986)
19.5±30.7 18.5±29.5 18.22 *** 17.84 20.73 20.07 23.57 23.13 26.50 25.24 27.78 25.87 28.43 25.83
Male Female Male Female Male Female Male Female Male Female Male Female Male Female
Adult
German
Gunther (1933): cited by Freihofer (1980)
2
N. American
Hansman (1966)
11±19 20±24 25±30 31±55 >55
5 10 15 20 24
*Horizontal distance between the lateral margins of the orbits.**Horizontal distance between the medial margins of the orbits.***50th percentile value.ÐIndicates information not specified
Far interpupillary distance: J. S. Pointer
325
Figure 3. The variation with age and gender of three human facial dimensions in the horizontal frontal plane of living subjects. The uppermost pair of curves (diamond symbols) indicate the mean extra-orbital width (Fledelius and Stubgaard, 1986), and the lowest truncated pair of curves (triangular symbols) the inter-orbital width (50th percentile values: Hansman, 1966): refer to Table 4. The two sets of data across the centre of the figure indicate the FIPD data of Fledelius and Stubgaard (1986: square symbols linked by continuous line) and Pointer (refer to Table 3; circular symbols linked by dotted line). For all three parameters the solid black symbols indicate male and the unfilled symbols female mean values, with error bars indicating + or ÿ 1st SD. Each FIPD value plotted for the Danish study (square symbols) represents the mean result obtained from between 17±36 subjects: the other six values (three pairs of circular symbols) are each the mean of measurements upon 150 subjects with, for each of these three age groups, male values being statistically significantly greater than female values. There is a marked increase in the FIPD for both genders during childhood and adolescence, but this trend becomes much reduced during adulthood and into later middle-age, especially in male subjects (see text).
variation of extra-orbital width in living subjects (highest pair of curves in Figure 3) also shows a parallel increase in this parameter for both genders (absolute value male>female) up to the mid-teens. Subsequently females display a much reduced rise to the mid-twenties, and no substantial variation thereafter. In male
subjects the adolescent increase is maintained through to the end of the second decade of life: thereafter the extra-orbital width stabilizes and then appears to decline through middle age, eventually attaining a value only slightly greater than that of similarly-aged female subjects.
326
Ophthal. Physiol. Opt. 1999 19: No 4
From the foregoing description of the longitudinal trends in two skeletal orbit-related dimensions which might be expected to have an in¯uence on the magnitude of the FIPD, it will be appreciated that for neither gender do they adequately describe the variation in the FIPD parameter throughout adult life as recorded here (see family of curves across the centre of Figure 3). Consequently, on the basis of the available evidence, determination of the dependence of the FIPD upon speci®c skeletal mensuration details remains unresolved. The inference must be that features other than solely the facial bony anatomy in¯uence the FIPD parameter. A possible candidate in this regard could be an age/gender dierence in volume of the orbital soft tissue. Unfortunately, however, even less well documented than age-related changes in the cranial bony structures are growth and developmental alterations in the soft tissues of the human orbital regionÐincluding, for example, changes arising from senile laxity of muscle tone and reduced adipose tissue volume within and around the orbit. This paucity of information is no doubt a re¯ection of the clinical applications that such quantitative material would have, even for the diagnostician or the restorative facial surgeon (Freihofer, 1980). Until such time as further gender-segregated and age-grouped mensuration details relating to the human facial skeleton are forthcoming, the variation in the magnitude of the FIPD described here for Caucasian subjects remains the only (indirect) evidence of a gender-speci®c pattern of anatomical development in the area of the orbits in the young and mature human adult.
References Anderson, A. L. (1954). Accurate clinical means of measuring intervisual axis distance. Arch. Ophthalmol. 52, 349±352. Bruckner, R., Batschelet, E. and Hugenschmidt, F. (1987). The Basel longitudinal study on aging (1955±1978). Ophthalmogerontological research results. Doc. Ophthalmol. 64, 235±310.
Cesario, V. A. and Latta, G. H. (1984). Relationship between the mesiodistal width of the maxillary central incisor and interpupillary distance. J. Prosthet. Dent. 52, 641±643. Emsley, H. H. (1952). Visual Optics. Vol. 1: Optics of Vision, 5th edn, Butterworths, London. Fledelius, H. C. and Stubgaard, M. (1986). Changes in eye position during growth and adult life as based on exophthalmometry, interpupillary distance, and orbital distance measurements. Acta Ophthalmol. (Kbh) 64, 481± 486. Freihofer, H. P. M. (1980). Inner intercanthal and interorbital distances. J. Max.-Fac. Surg. 8, 324±326. Hansman, C. F. (1966). Growth of interorbital distance and skull thickness as observed in roentgenographic measurements. Radiol. 86, 87±96. Harvey, R. S. (1982). Some statistics of interpupillary distance. Optician 184, 29(No. 4766: 12th November). Kaye, J. and Obstfeld, H. (1989). Anthropometry for children's spectacle frames. Ophthal. Physiol. Opt. 9, 293± 298. Knudtzon, K. (1949). On exophthalmometry. The result of 724 measurements with Hertel's exophthalmometer on normal adult individuals. Acta Psychiat. Neurol. Scand. 24, 523±537. Osuobeni, E. P. and Al-Musa, K. A. (1993). Gender dierences in interpupillary distance among Arabs. Optom. Vis. Sci. 70, 1027±1030. Osuobeni, E. P. and Faden, F. K. (1993). Interpupillary distance of females of Arab origin. Optom. Vis. Sci. 70, 244±247. Pointer, J. S. (1995a). The presbyopic add. II Age-related trend and a gender dierence. Ophthal. Physiol. Opt. 15, 241±248. Pointer, J. S. (1995b). The presbyopic add. III. In¯uence of the distance refractive type. Ophthal. Physiol. Opt. 15, 249± 253. Quant, J. R. and Woo, G. C. (1992). Normal values of eye position in the Chinese population of Hong Kong. Optom. Vis. Sci. 69, 152±158. Roy, F. H. (1985). Ocular Syndromes and Systemic Diseases, Grune & Stratton, Philadelphia. Sasieni, L. S. (1975). The Principles and Practice of Optical Dispensing and Fitting, 3rd edn, Butterworths, London. Whitnall, S. E. (1921). The Anatomy of the Human Orbit and Accessory Organs of Vision, Henry Frowde Hodder & Stoughton, London. Wol, E. (1968). Anatomy of the Eye and Orbit, 6th edn.; revised by Last R.J. Lewis, London.