European Journal of Forensic Sciences

Original Research
www.ejfs.co.uk
DOI: 10.5455/ejfs.191463

Prediction of stature based on upper

limb measurements among Egyptian
population
Abeer Abdel-wahab Ibrahim Sharaf El-Din1, Shereen Mohamed
Sobhy Elkholy1, Mohamed Ibrahim Yousef2

1
Depatrment of ABSTRACT
Forensic Medicine and
Background: Identification of living individual or dead is a challenging task for forensic experts and physical
Clinical Toxicology,
Faculty of Medicine,
anthropologists worldwide. At the same time, prediction of stature is one of the key parameters in the identity
Benha University, of victims of mass disasters or fatal assaults, especially in incomplete or isolated extremities, mutilated
Egypt, 2Department human remains, and highly decomposed bodies. Purpose: The aim of the current study was to assess the
of Radiology, Faculty correlation between some upper limb measurements and personal stature and to set regression equations
of Medicine, Benha for stature estimation in an adult Egyptian population. Subjects and Methods: To construct the equations;
University, Egypt nine anthropometric measurements were recorded for 500 healthy adult Egyptian Arab volunteers comprising
250 males and 250 females with ages ranging from 25 to 40 years, the stature, arm length, forearm length,
Address for correspondence:
hand length, hand breadth, and four fingers (index, middle, ring, and little fingers) lengths were measured.
Abeer Abdel-wahab Ibrahim
Sharaf El-Din, Department Results: A significant positive correlation between stature and all measurements were observed in both
of Forensic Medicine sexes. Regression models are constructed to estimate stature on the basis of all measurements for males
and Clinical Toxicology, and females. The accuracy of the regression equations was verified by comparing the estimated stature with
Faculty of Medicine,
the actual measured stature. Conclusion: This study provides new forensic standards for stature estimation
Benha University, Egypt.
E-mail: abeer2539801@ from the upper limb measurements of the adult Egyptian population.
yahoo.com/ abir.
sharafeldein@fmed.bu.edu.
eg

Received: June 22, 2015

Accepted: October 05, 2015 KEY WORDS: Forensic sciences, forensic anthropology, Egyptian population, personal identification,
Published: November 09, 2016 stature estimation, upper limb

INTRODUCTION feet [14], and hand dimensions [15]. It is well established

Identification of living or dead one is the main goal of forensic
investigations [1]. It is required in recently dead persons,
decomposing, mutilated bodies, and fragmentary remains [2]. In
such situation, estimation if stature becomes equally important
along with other parameters as age, sex, and race in personal
identity [3]. This process can narrow down the pool of victims
needing to be matched, allowing for more definitive markers,
such as DNA, to be later used for the confirmation of final
identifications [4].
When the bodies are dismembered or mutilated in wars, mass
disasters, and crimes; a practical alternative is to develop new
standards that utilize different parts of the skeleton, e.g.,
the sternum [5], metacarpal bones [6], calcaneus bone [7],
and maximum skull length [8]. Furthermore, many previous
studies have shown the correlation of stature with body
appendages [9-11]. Others have tried to estimate the stature
using different body parts as lower limb [12], arm span [13],
that there are inter-racial and inter-geographical differences in
measurements and their correlation with stature. Therefore, a
formula for one population does not necessarily yield reliable
results for another due to inherent population variations that
may be attributed to genetic and environmental factors as
climate, nutrition, and lifestyle. Therefore, regression formula
for stature estimation should be population-specific [16,17].
The two basic methods of estimating living stature from long
bones and body parts are the anatomical method and the
mathematical method. The anatomical method is considered
to provide the best approximation of stature, but its main
drawbacks are that it requires a complete skeleton [18,19]. In
contrast, the main advantage of the mathematical method is
that it requires a single bone or body part, but its predictive
ability is less accurate because of wide variability in population
body proportions. Within the mathematical method, regression
analysis is considered to be a better and more reliable tool than
multiplication factor analysis [20,21].

Eur J Forensic Sci ● Apr-Jun 2016 ● Vol 3 ● Issue 2 1

Sharaf El-Din, et al.: Prediction of stature based on upper limb measurements
In circumstances where only mutilated arm and/or forearm
portions are available for a deceased person, it becomes quite
challenging to formulate the proper and adequate regression
models. Hence, this study was done to assess the relationship
between personal stature and some upper limb measurements
among a group of living adult Egyptian males and females
and to develop predictive equations to estimate stature of an
individual from some upper limb measurements.
SUBJECTS AND METHODS
This study was conducted on 500 healthy adult Arab Egyptian
volunteers, comprising 250 males and 250 females, from the
companions of cases referred to Radiological Department,
Benha University Hospitals, Benha Faculty of Medicine, Egypt.
The volunteers were recruited from Benha city, Qalyubia
Governorate, Egypt, and surrounding vicinities. The ages of
the subjects ranged from 25 to 40 years. The lower age limit
was 25 years to be sure of complete development and attaining
maximum growth and maturity, as previous studies have stated
that although stature at 18-year-old is generally accepted as
being adult level stature, the median age for attaining full stature
in males is 21.2-year-old, with growth continuing in 10% of
males until they are 23.5-year-old [22] while the upper limit is
taken as 40 years since growth regression starts thereafter [23].
Left-handed subjects were excluded from the study to avoid any
variation on two sides caused by limb dominance. Therefore,
right-handed subjects were included in the study. The subjects
with any disease, injury, fracture, physical deformity, amputation,
or history of any surgical procedures of the upper limb were
excluded from the study. Informed consents were obtained from
each individual. The data collection was conducted at fixed
time, between 9:00 am and 11:00 am, to avoid the influence
of diurnal variation as it affects the standards generated and
equations developed for the estimation of stature [24].
Measurement from both sides were obtained from randomly
selected 10 males and 10 females and tested for a significant
difference. No statistical significant difference was observed
between both sides. Therefore, all measurements were obtained
from the left side in all cases. The measurements were done by
the same observer and with the same instrument, to avoid any
technical and/or inter-observer error. The current study was done
according to standard ethics drawn by the Benha University
ethical committee for human experimentation. Using standard
anthropometric instruments and techniques, the following
anthropometric measurements were measured:
• Stature: Height of the subject was measured in standing
position. The subject was instructed to stand barefooted
on the board of a standard standiometer with both feet
in close contact with each other, trunk braced along the
vertical board, and the head adjusted in Frankfurt plane
(the lateral palpebral commissure and the upper border of
the external auditory meatus). The measurement was taken
in centimeters, by bringing the horizontal sliding bar to the
vertex [25].
• Arm length (AL): Was measured as the distance between the
acromion (the most lateral point on the end of the acromial
2 Eur J Forensic Sci
process of the shoulder blade) to the most distal point on the
capitulum of humerus. This is done by holding the forearm
at right angles to the upper arm, when the capitulum of the
humerus can be palpated very easily [25].
• Forearm length (FL): Was measured as the direct distance
between the most proximal point of the olecranon process
and the styloid process while the elbow was flexed to
90 degrees and the fingers were extended in the direction
of the long axis of the forearm [26].
• Hand length (HL): Was measured with a digital sliding
caliper as the direct linear distance between the distal wrist
crease and the distal end of the most anterior projecting
point, i.e. tip of the middle finger. The subjects were asked
to place their hands supine on a flat hard horizontal surface
with fingers extended and adducted, following which the
hand length was measured [11].
• Hand breadth (HB): Was measured with a digital sliding
caliper as the distance between the most lateral point on the
head of the second metacarpal bone and the most medial
point on the head of the fifth metacarpal bone [25].
• Fingers length: Was measured with a digital sliding caliper on
the ventral surface of the left hand from the most proximal
crease of the finger to the tip of the index finger, middle
finger, ring finger, and little finger in the hand using a
digital sliding caliper. Thumb was excluded because reliable
measurement landmarks are difficult to establish [27].
Statistical Analysis
Before data collection, assessment of precision and reliability
of the stature and upper limb measurement were done in
accordance with the study of Ahmed [4], where the stature
and upper limb dimensions of ten subjects were measured on
three different evaluation days, with two days between the re-
measurements. The absolute technical error of measurement
(TEM) was initially calculated using the following equation.
TEM = ∑D 2
2N
Where D represents the difference between each two successive
measurements, and N represents the number of individuals.
Then, the relative technical error of measurement (rTEM) was
calculated from each two successive sets of measurements by
dividing the TEM for a given variable by the grand mean of that
variable and multiplying the result by 100. In addition to these
measurements, the coefficient of reliability was calculated. The
intra-observer measurement error and reliability were calculated
to be within the acceptable standards for all measurements
(coefficient of reliability > 0.9 and rTEM < 5%).
The data were computed and analyzed using SPSS (Statistical
Package for Social Sciences, version 16) computer software. Data
were presented as means, standard deviations, and ranges to
summarize the anthropometric measurements. An independent
Student’s t-test was used to test the differences between males
and females for all variables. Pearson correlation coefficients
were calculated to measure the strength of correlation between
● Apr-Jun 2016 ● Vol 3 ● Issue 2
 Sharaf El-Din, et al.: Prediction of stature based on upper limb measurements

stature and the upper limb measurements. Sex-specific simple different left upper limb measurements among Egyptian
and multiple linear regression equations for stature estimation population. Among males, it is observed that the highest value
were developed using the upper limb measurements. of r is exhibited by the arm length (AL) (0.844), whereas for
the index finger, the r value is the least, (0.309). At the same
Simple linear regression equations for estimation of stature were time, the arm length (AL) exhibits the overall highest value
formed for each of the measured parameter as: Y (stature) = of correlation (r = 0.801) with stature while ring finger length
a (constant) + b (regression coefficient of the independent (RFL) exhibits the least correlation (r = 0.383) with stature in
variable) X (individual variable) ± SEE (standard error of the females. All measurements presented statistically significant
estimate). The multiple regression models were derived as correlation coefficients with stature. Thus, stature is positively
Y (stature) = a (constant) + b1 (regression coefficient of the and significantly related to various upper limb measurements.
first variable) X1 (first variable) + b2 (regression coefficient of
the second variable) X2 (second variable) + -- bn (regression Simple Linear Regression
coefficient for the nth variables) Xn (nth variable) ± SEE.
A paired-samples t-test was used to compare the differences Tables 3 and 4 depict the simple linear regression equations for
between the actual and estimated means using different estimation of stature from left upper limb measurements for
equations. A P < 0.05 was considered significant. The accuracy males and females, respectively, in the Egyptian population.
of these equations was validated by determination coefficients Regression equations have been computed separately, for both
(r2 for bivariate analysis and R2 for multivariate analysis) and sexes and for each measurement of the upper limb. The tables
standard error of estimate (SEE). also exhibit standard error of estimate (SEE) and determination
coefficient (r2 for bivariate analysis and R2 for multivariate
RESULTS analysis). The standard error of estimate predicts the deviation
of estimated stature from the actual stature and a low value
Descriptive statistics indicates greater reliability and higher accuracy in the estimated
stature. In males, arm length (AL) provides the best estimate
The total of 500 healthy adult Egyptian Arab volunteers of stature as it exhibits the overall highest value of correlation
(250 males and 250 females) were studied for estimation of
stature by upper limb measurements, i.e. arm length (AL), Table 2: Correlation between stature and left upper limb
forearm length (FL), hand length (HL), hand breadth (HB), measurements
and fingers length. Table 1 shows descriptive statistics of stature Variable Males (n=250) Females (n=250)
in males and females. It is evident that the stature in males is (cm) r P value r P value
ranging from 162 to 185.5 cm with the mean of 174.59 cm and AL 0.844 <0.001 0.801 <0.001
standard deviation of 4.13 cm while, the stature in females is FL 0.643 <0.001 0.450 0.001
varied from 153 to 170 cm with the mean of 163.02 cm and HL 0.443 0.001 0.526 <0.001
standard deviation of 3.79 cm. As expected, the mean values HB 0.488 <0.001 0.406 0.003
IFL 0.309 0.029 0.434 0.002
of stature and left upper limb measurements for males were MFL 0.645 <0.001 0.534 <0.001
significantly higher than females. RFL 0.337 0.017 0.383 0.006
LFL 0.349 0.013 0.405 0.004
Table 2 displays the correlation coefficient values (r) of AL: Arm length, FL: Forearm length, HL: Hand length, HB: Hand
breadth, IFL: Index finger length, MFL: Middle finger length, RFL: Ring
finger length, LFL: Little finger length, r: Correlation coefficient, level of
Table 1: Descriptive statistics for ranges, means and SD of significance was set at P<0.05
left upper limb measurements along with stature in males and
females (all measurements are in cm) Table 3: Simple linear regression equations for stature
Variable Males (n=250) Females (n=250) P value estimation from left upper limb measurements in males
(cm) Range Mean±SD Range Mean±SD Variable Equation Y=a+bX ±SEE r2 t-statistic P value
Min. Max. Min. Max. (cm)
S 162 185.5 174.59±4.13 153 170 163.02±3.79 <0.001 AL S=108.06+1.966x (AL) 2.24 0.712 10.9 <0.001
AL 31 38 33.84±1.77 27.5 35 31.14±1.84 <0.001 FL S=126.07+1.661x (FL) 3.20 0.414 5.82 <0.001
FL 26 31 29.20±1.60 24 30 26.87±1.54 <0.001 HL S=142.72+1.543x (HL) 3.74 0.196 3.42 0.001
HL 19 25 20.66±1.18 16.8 20.8 18.50±0.86 <0.001 HB S=138.75+3.474x (HB) 3.65 0.238 3.87 <0.001
HB 9.5 11.2 10.31±0.58 7.7 10.2 8.90±0.53 <0.001 IFL S=156.61+2.296x (IFL) 3.97 0.096 2.25 0.029
IFL 7.1 9.5 7.83±0.55 6.1 8.2 7.21±0.46 <0.001 MFL S=128.52+5.301x (MFL) 3.19 0.416 5.85 <0.001
MFL 8 9.5 8.69±0.50 6.8 9 7.96±0.57 <0.001 RFL S=151.24+2.922x (RFL) 3.93 0.114 2.48 0.017
RFL 6.8 9 7.99±0.47 6.3 8.8 7.33±0.51 <0.001 LFL S=149.52+3.777x (LFL) 3.91 0.121 2.58 0.013
LFL 6 7.5 6.64±0.38 5.1 6.9 6.05±0.42 <0.001
S: Stature, AL: Arm length, FL: Forearm length, HL: Hand length,
S: Stature, AL: Arm length, FL: Forearm length, HL: Hand length, HB: Hand breadth, IFL: Index finger length, MFL: Middle finger length,
HB: Hand breadth, IFL: Index finger length, MFL: Middle finger length, RFL: Ring finger length, LFL: little finger length, SEE: Standard error
RFL: Ring finger length, LFL: Little finger length, Student’s t-test of estimate, r2: Coefficient of determination, Y: Stature, a: Constant,
is used to calculate P value; level of significance was set at P<0.05, b: Regression coefficient, X: Length of the predictor variable, Level of
SD: Standard deviation significance was set at P<0.05

Eur J Forensic Sci ● Apr-Jun 2016 ● Vol 3 ● Issue 2 3

Sharaf El-Din, et al.: Prediction of stature based on upper limb measurements

(r= 0.844) and least value of standard error of estimate DISCUSSION

(SEE = ± 2.24). However, index finger length which exhibits the
highest error of estimate (3.97) and lowest value of r (0.309). In Estimation of stature is one of the important parameter
females, arm length also shows the highest value of correlation in medico-legal and forensic examinations when highly
(r = 0.801) and least value of standard error of estimate (SEE = decomposed and mutilated dead bodies with fragmentary
± 2.30) while, ring finger length shows the highest error of remains are brought for postmortem examination. In this
estimate (±3.54) and lowest value of correlation (r =0.383). So, situation, it becomes difficult to identify the deceased [3]; hence,
index finger length in males and ring finger length in females there is a need to more studies on the estimation of stature from
should be used only in the absence of the other measurements various body parts in different population groups [20]. Wilson
of the upper limb. The regression coefficients were found to be et al. [28] stated that it is essential for stature estimation to
statistically significant in all derived equations. The predictive use not only equations based on forensic statures but also
value, or coefficient of the determination, was greatest for the equations based on modern samples. Therefore, this study
reconstruction of stature from arm length in both sexes. was designed to investigate the relationship between some
upper limb measurements and stature in Egyptian population
Multiple Regression Analysis and to develop predictive equations to estimate stature of
an individual from some upper limb measurements. Eight
Multiple regression equations were formulated for males and left upper limb measurements and stature of adult Egyptian
females [Table 5]. The multiple regression equations revealed population volunteers were taken; the prediction of stature was
a lower standard error of estimate for males (± 2.39) and for derived through simple linear regression and multiple regression
females (± 2.45) as compared with the developed estimate analysis for males and females separately. The results of the
models utilizing different variables for males and females except present study demonstrated that the stature and recorded upper
that of arm length which is (± 2.24) for males and (± 2.30) limb measurements of males are significantly larger than those
for females. of females (P < 0.001). These results agree with the results of
other studies that have observed the same findings in different
Comparison between Actual and Estimated Stature populations [4,15,29-31]. This finding can be attributed to
the fact that males are genetically taller than females, and the
Comparisons of the actual stature and the estimated stature puberty occurs in females approximately 2 years earlier than in
from the upper limb measurements were performed utilizing males, and the earlier fusion of the epiphysis results in less time
simple and multiple regression equations [Tables 6 and 7]. for its growth. In addition, it is known that there is a relation
In both sexes, the mean value estimates were close to the actual between Y chromosome and stature [11,32].
stature, and there were no statistically significant differences
between them when utilizing different equations. The recorded upper arm lengths in adult Egyptian population
in the current study were longer than Sudanese [4] and shorter
Table 4: Simple linear regression equations for stature than those of the studied Iranian [30], but closer to those of
estimation from left upper limb measurements in females Bulgarians subjects [33] for males and females. At the same
time, the forearm lengths were greater than those of Turkish
Variable Equation Y=a+bX ±SEE r2 t-statistic P value
subjects [34] and Iranian subjects [30] and close to those of
AL S=111.6+1.65x (AL) 2.30 0.642 9.27 <0.001 studied Sudanese subjects [4]. Furthermore, the hand length
FL S=133.3+1.11x (FL) 3.43 0.203 3.49 0.001
HL S=120.34+2.31x (HL) 3.26 0.277 4.29 <0.001 and breadth of Egyptian subjects in the current study were found
HB S=137.42+2.875x (HB) 3.51 0.165 3.07 0.003 to be longer than those of Sudanese [4], West Australian [31],
IFL S=137.34+3.563x (IFL) 3.46 0.188 3.33 0.002 Turkish [34], Iranian [30], Mauritian [15], and Rajput Indian
MFL S=134.73+3.552x (MFL) 3.25 0.285 4.37 <0.001 population [10].
RFL S=142.1+2.852x (RFL) 3.54 0.147 2.88 0.006
LFL S=141.24+3.601x (LFL) 3.51 0.164 3.07 0.004
When lengths of index, middle and ring fingers in the left
S: Stature, AL: Arm length, FL: Forearm length, HL: Hand length, hand in this work were compared to other studies, it was found
HB: Hand breadth, IFL: Index finger length, MFL: Middle finger length, that their lengths were close to those of the West Australian
RFL: Ring finger length, LFL: Little finger length, SEE: Standard error
of estimate, r2: Coefficient of determination, Y: Stature, a: Constant, subjects [31] but longer than the Iranian populations. On the
b: Regression coefficient, X: Length of the predictor variable, level of other hand, the length of the little finger in the present work
significance was set at P<0.05 was shorter than the Iranian population [30].

Table 5: Multiple linear regression equations for the stature from left upper limb measurements in males and females
Sex Equation ±SEE r2 P value
Males (n=250) S=113+2.02x (AL)-0.189x (FL)+0.06x (HL)-1.38x (HB)-0.106x (IFL)+ 2.39 0.719 <0.001
1.78x (MFL)+1.23x (RFL)-1.92x (LFL)
Females (n=250) S=109.96+1.57 x (AL)-0.17x (FL)+0.125x (HL)+ 2.45 0.652 <0.001
0.43x (HB)+0.57x (IFL)+0.49x (MFL)-1.08x (RFL)+0.38x (LFL)
S: Stature, AL: Arm length, FL: Forearm length, HL: Hand length, HB: Hand breadth, IFL: Index finger length, MFL: Middle finger length, RFL: Ring
finger length, LFL: Little finger length, SEE: Standard error of estimate, r2: Coefficient of determination, Level of significance was set at P<0.05

4 Eur J Forensic Sci ● Apr-Jun 2016 ● Vol 3 ● Issue 2

 Sharaf El-Din, et al.: Prediction of stature based on upper limb measurements

Table 6: Actual stature versus stature estimated from coefficient for upper arm length [34], and Sudanese males had
simple linear regression analysis in males and females a higher correlation coefficient for ulnar length [4].
(all measurements are in cm)
Variable Mean ±SD Min Max. Paired P value In the current study, the correlation coefficients of hand breadth
t-test and stature in males were greater than hand length while, in
Estimated females, the correlation coefficients of hand length and stature
stature (males) were greater than hand breadth. This finding indicates that hand
AL 174.62 3.46 169.0 182.8 0.085 0.932 length is more reliable than hand breadth for estimating stature
FL 174.66 2.60 169.3 177.6 0.141 0.888 among Egyptian females, which is in agreement with previous
HL 174.61 1.82 172.04 181.8 0.035 0.972
HB 174.63 1.99 171.8 177.7 0.067 0.947
results recorded for studied Sudanese [4], Upper Egyptian [32],
IFL 174.64 1.25 172.9 178.4 0.078 0.938 Iranian [30], Turkish [34], Indian [10,11], and West Australian
MFL 174.64 2.62 170.9 178.9 0.101 0.920 subjects [35].
RFL 174.65 1.30 171.1 177.5 0.100 0.921
LFL 174.55 1.48 172.2 177.8 0.092 0.927 Regarding the correlation between stature and fingers lengths in
Actual stature 174.59 4.13 162 185.5 -
Estimated
this study, middle finger length and stature was stronger in males
stature (females) (r = 0.645) than females (r = 0.534) while index finger length and
AL 163.05 3.06 157 169.4 0.082 0.935 stature was stronger in females than males (r = 0.434 for females; r
FL 163.08 1.67 159.8 166.5 0.105 0.917 =0.309 for males), ring finger (r =0.383 for females; r = 0.337 for
HL 163.07 1.97 159.1 168.3 0.081 0.936 males), and little finger (r=0.405 for females; r = 0.349 for males).
HB 163.08 1.52 159.6 166.7 0.104 0.918
So, the middle finger had the highest correlation coefficient
IFL 163.07 1.63 159.07 166.6 0.076 0.939
MFL 163.09 1.97 158.9 166.7 0.148 0.883 with stature in both sexes in this study. This is in agreement
RFL 163.07 1.42 160.07 167.2 0.092 0.927 with Iranian subjects [30] as the relation between middle
LFL 162.98 1.49 159.6 166.09 0.093 0.926 finger length and stature was stronger in males (r =0.674) than
Actual stature 163.02 3.79 153 170 - females (r = 0.644). At the same time, Sen et al. [27] revealed
S: Stature, AL: Arm length, FL: Forearm length, HL: Hand length, a statistically significant relationship in both genders between
HB: Hand breadth, IFL: Index finger length, MFL: Middle finger length, finger lengths and stature and concluded that the middle finger
RFL: Ring finger length, LFL: Little finger length, level of significance will give the exact height of the individual among all fingers.
was set at P<0.05, SD: Standard deviation On the other hand, Krishan et al. [36] estimated stature from
index and ring finger length in a North Indian population and
Table 7: Actual stature versus stature estimated from observed that stature can be estimated from these finger lengths
multiple linear regression analysis in males and females with a reasonable accuracy. This suggests that there is a genetic
(all measurements are in cm) difference between males and females and hence formula for one
Estimated stature using Mean actual Paired t-test P value sex cannot be applied for the other while estimating the stature.
multiple regression stature of
equations for males males
Standard error of estimate (SEE) is a good parameter to show
174.61±4.13 174.59±4.13 0.047 0.962 the relation between real value and estimated value and to assess
Estimated stature using Mean actual Paired t-test P value the accuracy of the simple regression equations using single
multiple regression stature of parameter, the SEE was the lowest for arm length measurements
equations for females females
in both sexes (SEE = ± 2.24 for males; SEE = ± 2.30 for
163.048±3.02 163.02±3.79 0.063 0.950 females). So, the arm length was found to be more predictive
Level of significance was set at P<0.05 than other upper limb measurements. This is in agreement
with several previous studies had found that all proximal limb
In the current study, the correlation coefficient between the bones are better predictors than distal limb bones [30,37].
stature and the upper limb measurements was found to be On the other hand, a study on Turkish males revealed that
highly significant and positively correlated, indicating that it is forearm length was found to be a better predictor than upper
possible to develop regression equations for the estimation of arm length [34]. The results obtained in this study for arm
stature from various upper limb measurements. The correlation length are more predictive than that found by Ahmed [4]
coefficient in this work was greatest between stature and arm (SEE = ±4.40 cm for males; 4.48 cm for females) and Ozaslan
length (r = 0.844 for males; r = 0.801 for females), which et al. [34] (SEE = ± 4.88 cm for males; 5.93 cm for females).
indicates that the upper arm length is a good predictor for
stature in both sexes. It was greater than that reported for Regarding hand measurements, the hand length presented
Sudanese [4], Iranian [30], and Turkish subjects [34]. lower standard error of estimate in Egyptian females
(SEE =± 3.26 cm) than hand breadth, while hand breadth
The present study revealed that males having higher correlation presented the lower standard error of estimate in Egyptian
coefficients for upper arm length, forearm length, hand breadth, males (SEE = ±3.65 cm) than hand length. This finding is
and middle finger length than females. On the other hand, different from other studies utilizing hand measurements that
females having higher correlation coefficients for hand length, showed comparable standard errors of estimates as hand length
index, ring, and little finger length than males. On the contrary presented the lowest standard error of estimate, such as those of
to these findings, the Turkish females had a higher correlation Ahmed, [4] (SEE = ± 5.01 cm for males; 4.53 cm for females),

Eur J Forensic Sci ● Apr-Jun 2016 ● Vol 3 ● Issue 2 5

Sharaf El-Din, et al.: Prediction of stature based on upper limb measurements
Krishan and Sharma, [10] (SEE = ± 5.17 cm for males; 3.82 cm
for females), Agnihotri et al. [15] (SEE = ± 4.96 cm for males;
4.23 cm for females), Ishak et al. [31] (SEE = ± 4.74 cm
for males; 5.10 cm for females), Habib and Kamal, [29]
(SEE = ± 5.48 cm for males; 4.54 cm for females), and Ozaslan
et al. [34] (SEE = ±5.62 cm for males; 6.06 cm for females). The
variation in observations can be attributed to the population
and ethnic differences between the Egyptian population and
the other populations in earlier studies.
Regarding the multiple regression equations developed for
the upper limb dimensions, this study demonstrated a higher
degree of prediction accuracy, as indicated by a lower SEE
(± 2.39-2.45 cm for males and females, respectively) and higher
R2 (0.665- 0.584 for males and females, respectively) as compared
with the developed estimate models utilizing different variables
separately in simple linear regression for males and females
except that of arm length which is ± 2.24 for males and ± 2.30
for females indicated that the multiple regression equations are
better indicators of stature estimation. Krishan and Sharma [10],
Habib and Kamal [29], and Abdel-Malek et al. [32] demonstrated
increases in prediction accuracy when using multiple variables. At
the same time, Rastogi et al. [11] suggested that when predicting
stature, multiple regression equations give more accurate results.
Furthermore, Ahmed [4] stated that when hand measurements
were combined to develop a multiple regression models, the
calculated SEE was ± 4.45 cm in males and 5 cm in females,
which is comparable to the developed simple linear models for
hand length (± 5.1 cm for males and 4.53 in females) and hand
breadth (± 5.85 cm in males and 5.19 cm in females) [4]. On
the other hand, Ishak et al. [31] demonstrated an equivalent
degree of prediction accuracy (± 4.74 cm) in multiple regression
models compared to the simple linear models.
CONCLUSION
Estimation of stature from limbs or dismembered body parts is of
great importance to forensic experts and anthropologists. In the
present research, there is a strong relationship between stature
and upper limb measurements. Eight regression equations
derived can be used accurately to estimate the stature from
upper limb measurements in the adult Egyptian population. No
significant difference between measured and estimated stature
from upper limb measurements were found. The obtained
equations are specific to the adult Egyptian population;
therefore, application of these by the other populations
might result in incorrect results. It is hoped that the equation
formulae would help in forensic identification when only some
remains of the body are found. In addition to the medico-legal
implications of this study; it may be useful for the clinician in
plastic and reconstructive surgery, anatomists, archeologists, and
anthropologist. Further researches on the Egyptian population
with different age groups are recommended.
REFERENCES
1. Kanchan T, Menezes RG, Moudgil R, Kaur R, Kotian MS, GargRK.
Stature estimation from foot dimensions. Forensic Sci Int
2008;179:241.e1-5.
6 Eur J Forensic Sci
2. Kanchan T, Krishan K. Anthropometry of hand in sex determination
of dismembered remains - A review of literature. J Forensic Leg Med
2011;18:14-7.
3. Krishan K. Estimation of stature from cephalo-facial anthropometry
in North Indian population. Forensic Sci Int 2008;181:52.e1-6.
4. Ahmed AA. Estimation of stature from the upper limb measurements
of Sudanese adults. Forensic Sci Int 2013;228:178.e1-7.
5. Menezes RG, Kanchan T, Kumar GP, Rao PP, Lobo SW, Uysal S, et al.
Stature estimation from the length of the sternum in South Indian
males: A preliminary study. J Forensic Leg Med 2009;16:441-3.
6. Meadows L, Jantz RL. Estimation of stature from metacarpal lengths.
J Forensic Sci 1992;37:147-54.
7. Bidmos M, Asala S. Calcaneal measurement in estimation of stature
of South African blacks. Am J Phys Anthropol 2005;126:335-42.
8. Patil KR, Mody RN. Determination of sex by discriminant function
analysis and stature by regression analysis: A lateral cephalometric
study. Forensic Sci Int 2005;147:175-80.
9. Sanli SG, Kizilkanat ED, Boyan N, Ozsahin ET, Bozkir MG, Soames R,
et al. Stature estimation based on hand length and foot length. Clin
Anat 2005;18:589-96.
10. Krishan K, Sharma A. Estimation of stature from dimensions of
hands and feet in a North Indian population. J Forensic Leg Med
2007;14:327-32.
11. Rastogi P, Nagesh KR, Yoganarasimha K. Estimation of stature
from hand dimensions of north and south Indians. Leg Med (Tokyo)
2008;10:185-9.
12. Nor FM, Abdullah N, Mustapa AM, Qi Wen L, Faisal NA, Ahmad
Nazari DA. Estimation of stature by using lower limb dimensions in
the Malaysian population. J Forensic Leg Med 2013;20:947-52.
13. Laila SZ, Begum JA, Ferdousi R, Parveen S, Husain MS, Holy SZ, et al.
Anthropometric measurements of the arm span and their correlation
with the stature of Bangladeshi adult Muslim females. Mymensingh
Med J 2010;19:561-4.
14. Krishan K, Kanchan T, Passi N. Estimation of stature from the foot
and its segments in a sub-adult female population of North India.
J Foot Ankle Res 2011;4:24.
15. Agnihotri AK, Agnihotri S, Jeebun N, Googoolye K. Prediction of
stature using hand dimensions. J Forensic Leg Med 2008;15:479-82.
16. Hunnargi SA, Menezes RG, Kanchan T, Lobo SW, Uysal S, Herekar NG,
et al. Sternal index: Is it a reliable indicator of sex in the Maharashtrian
population of India? J Forensic Leg Med 2009;16:56-8.
17. Kanchan T, Menezes RG, Moudgil R, Kaur R, Kotian MS, Garg RK.
Stature estimation from foot length using universal regression
formula in a North Indian population. J Forensic Sci 2010;55:163-6.
18. Lundy JK. The mathematical versus anatomical methods of stature
estimate from long bones. Am J Forensic Med Pathol 1985;6:73-6.
19. Maijanen H. Testing anatomical methods for stature estimation
on individuals from the W. M. Bass donated skeletal collection.
J Forensic Sci 2009;54:746-52.
20. Krishan K, Kanchan T, Sharma A. Multiplication factor versus
regression analysis in stature estimation from hand and foot
dimensions. J Forensic Leg Med 2012;19:211-4.
21. Iscan MY. Forensic anthropology of sex and body size. Forensic Sci
Int 2005;147:107-12.
22. Roche AF, Davila GH. Late adolescent growth in stature. Pediatrics
1972;50:874-80.
23. Trotter M, Gleser G. The effect of ageing on stature. Am J Phys
Anthropol 1951;9:311-24.
24. Krishan K, Kanchan T, Menezes RG, Ghosh A. Forensic anthropology
casework essential methodological considerations in stature
estimation. J Forensic Nurs 2012;8:45-50.
25. Vallois HV. Anthropometric techniques. Curr Anthropol 1965;6:127-44.
26. Martin AD, Carter JE, Hendy KC, Malina RM. Segment lengths.
In: Lohman TG, Roche AF, Martorell R, editors. Anthropometric
Standardization Reference Manual. Champaign, IL: Human Kinetics;
1988. p. 9-26.
27. Sen J, Kanchan T, Ghosh A, Mondal N, Krishan K. Estimation of
stature from lengths of index and ring fingers in a North-eastern
Indian population. J Forensic Leg Med 2014;22:10-5.
28. Wilson RJ, Herrmann NP, Jantz LM. Evaluation of stature estimation
from the database for forensic anthropology. J Forensic Sci
2010;55:684-9.
29. Habib SR, Kamal NN. Stature estimation from hand and phalanges
lengths of Egyptians. J Forensic Leg Med 2010;17:156-60.
● Apr-Jun 2016 ● Vol 3 ● Issue 2
 Sharaf El-Din, et al.: Prediction of stature based on upper limb measurements

30. Akhlaghi M, Hajibeygi M, Zamani N, Moradi B. Estimation of stature
from upper limb anthropometry in Iranian population. J Forensic Leg
Med 2012;19:280-4.
31. Ishak NI, Hemy N, Franklin D. Estimation of stature from hand and
handprint dimensions in a Western Australian population. Forensic
Sci Int 2012;216:199.e1-7.
32. Abdel-Malek AK, Ahmed AM, el-Sharkawi SA, el-Hamid NA. Prediction
of stature from hand measurements. Forensic Sci Int 1990;46:181-7.
33. Radoinova D, Tenekedjiev K, Yordanov Y. Stature estimation from
long bone lengths in Bulgarians. Homo 2002;52:221-32.
34. Ozaslan A, Koç S, Ozaslan I, Tugcu H. Estimation of stature from
upper extremity. Mil Med 2006;171:288-91.
35. Reed JC, Algee-Hewitt BF. Comments on “Estimation of stature from
cranial sutures in a South Indian male population” by P. P. J. Rao et al.
Int J Legal Med 2011;125:469-71.
36. Krishan K, Kanchan T, Asha N. Estimation of stature from index and
ring finger length in a North Indian adolescent population. J Forensic
Leg Med 2012;19:285-90.
37. Vercellotti G, Agnew AM, Justus HM, Sciulli PW. Stature estimation
in an early medieval (XI-XII c.) Polish population: Testing the accuracy
of regression equations in a bioarcheological sample. Am J Phys
Anthropol 2009;140:135-42.
Source of Support: Nil, Conflict of Interest: None declared.

Eur J Forensic Sci ● Apr-Jun 2016 ● Vol 3 ● Issue 2 7