Predicting Stature From Body Trunk Dimensions: A Regression Analysis Approach

Abstract

It is not uncommon in forensic medicine to receive only the body trunk for postmortem examination. The study intended to derive linear regression formulae for estimating stature with two parameters of the body trunk: the inter-acromial length and the distance from the suprasternal notch to the pubic symphysis, and to correlate them with the known stature. The cross-sectional study was conducted on subjects more than 18 years of age belonging to the North-West region of India. Pearson’s correlation test was applied to assess whether there was a relationship between the parameters measured and stature. Gender-specific linear regression models for estimating stature were established. The paired t-test was employed to compare the known and estimated stature. Pearson’s correlation in the combined gender group showed good correlations between parameters measured and stature. The paired t-test between known and estimated stature was statistically not significant, with a p value of >.05. The present study has established a definite correlation between stature and the parameters measured. Stature can be estimated from these parameters in situations where only the body trunk is available for postmortem examination and subsequent identification.

Keywords

Stature identification anthropology inter-acromial length suprasternal notch pubic symphysis

Introduction

Stature is a decisive parameter for identifying individuals, particularly in cases where bodies are unidentified due to decomposition or mass disasters such as floods, earthquakes, tsunamis, landslides, aviation or train accidents and terrorist activities.¹ It is not uncommon that autopsy surgeons receive only body trunks for postmortem examination. It is the duty of stakeholders and forensic experts conducting medicolegal autopsies to establish the identity of an unknown dead individual, ensuring cremation as per religious customs and practices, while keeping the dignity of the dead. Race, sex, age and stature are the ‘big four’ in the Bertillon identification system, and hence stature estimation plays a vital role in identifying both living and deceased individuals.^1–3 Forensic anthropology is a branch that employs data obtained from body parts and skeletons to assist in legal cases. It focuses on identifying individuals, whether living, deceased or reduced to skeletal remains, using measurement techniques known as anthropometry.^4–6 Most of the researchers have conducted studies on the estimation of stature from subcutaneous bone length or skeletal remains, percutaneous measurement of various body parts, including arm, forearm, leg, foot and so on,^{7, 8} the regression analysis has been the most commonly used method.⁹ Considering India’s diverse racial, climatic, dietary and ethnic variations, findings from one population may not apply to another,^{9, 10} and thus, there is a need to estimate stature based on body trunk dimensions across different regions. The present study employs two body trunk parameters: inter-acromial length (IAL) and the distance from the suprasternal notch to the pubic symphysis. The objective is to derive linear regression formulae for stature estimation using these two parameters and correlate them with known stature. This study uniquely focuses on torso measurements for stature estimation in the Northwest Indian population, unlike prior studies focusing primarily on limb measurements.

Material and Methods

This cross-sectional study was conducted at a tertiary care centre in Southern Haryana from January 2020 to June 2020, comprising medical students of age 18 and above, including all ages above 18 years, belonging to the North-West region of India. However, the study included medical students aged above 18 years, and no explicit upper age limit was predefined. The age of participants in our sample ranged up to 28 years. Prior approval was obtained from the Institutional Ethical Committee vide letter no SHKM/IEC/2019/66 dated 19 April 2019. The study was conducted following the ethical principles outlined in the Declaration of Helsinki. Participation of students was voluntary, and informed consent was taken. Subjects with evident physical abnormalities were excluded. The sample size of 240 subjects was determined based on previous similar studies, with an expected moderate correlation (r ≈ 0.4), power of 80% and alpha of 0.05. Using G*Power software, a minimum sample size of 84 was determined as adequate for correlation studies. Our study sample size (n = 240) exceeds this, thus ensuring sufficient power. The following measurements were obtained and recorded in the proforma.

Stature

After making sure that the individuals were not wearing any kind of footwear or headwear, the stature was measured on a stadiometer. They were instructed to stand upright on the stadiometer with their heads in the Frankfurt Horizontal Plane, hands hanging at their sides and feet parallel or slightly divergent. The stadiometer’s movable horizontal headpiece was gently placed on the scalp, and height was recorded in centimetres.^{11, 12}

Inter-acromial Length

The subjects were made to sit erect with their arms hanging freely by the sides, ensuring the shoulders were neither too far nor too forward. The lateral border of the acromial process was located on each shoulder. The blunt end of the spreading calliper was placed on a bony ridge on one side, and then the spreading arm of the calliper was placed on the skin directly next to the lateral border of the acromial process on the opposite side.

Reasonable pressure was applied to compress the soft tissues over the acromial process while precautions were taken not to hurt the volunteers. Each measurement was taken twice, first with the spreading calliper and then with a self-retracting measuring tape.^{1, 4, 13, 14} The spreading calliper had a precision of 0.1 cm, and the metallic tape had an accuracy of ±0.05 cm.

Distance from Suprasternal Notch to Pubic Symphysis

After locating the suprasternal notch and upper level of pubic symphysis percutaneously, the distance from the suprasternal notch to the pubic symphysis was measured with a metallic tape.^{15, 16} All measurements were recorded in centimetres to the nearest two decimal places. To minimise diurnal variations, measurements were performed daily between 2:00 pm and 3:00 pm. The measurements of females were obtained by the female observer, while those of male subjects were recorded by the male observer. To eliminate intra-observer error, the same observer recorded the measurements each day. A pilot study was conducted to evaluate data reliability and reproducibility. The intraclass correlation coefficient (ICC) ranged between 0.92 and 0.96, indicating excellent inter- and intra-observer reliability.

Statistical Evaluation

IBM SPSS Statistics Version 23.0 for Windows (Armonk, NY: IBM Corp) was utilised for statistical analysis. Pearson’s correlation test was applied to determine the relationship between measured parameters and stature. A p value of .05 was considered statistically significant. Gender-specific linear regression models were developed for stature estimation. The coefficient of correlation (R), coefficient of determination (R²) and standard error of estimation (SEE) were calculated. Using regression analysis, the equations for multiple regressions were generated based on various parameter combinations. The paired t-test was applied to compare the known and estimated stature.

Results

A total of 240 students participated in the study, comprising 120 males and 120 females. The mean ages for males and females were 21.30 and 20.55 years, respectively. The descriptive statistics for age, stature, IAL and the distance from suprasternal notch to pubic symphysis (SP) are depicted in Table 1. Pearson’s correlation in the combined gender group showed good correlations between SP, IAL parameters and stature, with SP exhibiting the highest correlation in males and IAL exhibiting the highest correlation among females. Pearson’s correlation between SP, IAL parameters and stature in males, females and combined gender groups is depicted in Table 2. A summary of linear regressions in males, females and combined genders is depicted in Table 3. By employing the linear regression equation, stature can be approximated by using the regression: y (stature) = b (constant) + a (regression coefficient of the independent parameter) × SP/IAL. The SP showed the lowest SEE in males as compared to IAL, while IAL showed the lowest SEE among females (Table 3). The regressions obtained by the multiple regression analysis for both parameters by using various combinations in males, females and the combined gender group have been depicted in Table 4. In the male group, the multiple regressions showed the lowest SEE (4.593). The descriptive statistics of mean values of known stature and estimated stature in males and females are depicted in Table 5. The paired t-test was applied between known and estimated stature, which was statistically not significant with a p value of >.05 and the same is represented in Table 6.

Table 1.

Descriptive Statistics of Age, Stature, Suprasternal Notch to Pubic Symphysis (SP) and Inter-acromial Length (IAL).

Parameter	Gender	Mean	SD	Min	Max
Age	Male (120)	21.3	1.63	18	28
	Female (120)	20.55	1.89	18	27
	Combined (240)	20.93	1.8	18	28
Stature	Male	174.43	5.89	162	189
	Female	161.36	5.85	148	175
	Combined	167.89	8.79	148	189
SP	Male	54.05	5.25	20	65
	Female	48.16	4.35	38	58
	Combined	51.1	5.64	20	65
IAL	Male	44.06	3.83	19	49
	Female	38.61	4.22	16.7	46.5
	Combined	41.34	4.86	16.7	49

Table 2.

Pearson’s Correlation Analysis Between Body Trunk Parameters and Stature.

Parameters	Males		Females		Combined
Parameters	R	p Value	R	p Value	R	p Value
SP	0.479	<.001	0.239	.009	0.599	<.001
IAL	0.395	<.001	0.265	.003	0.599	<.001

Note: Correlation is significant at the level 0.01.

Table 3.

Linear Regression Equations for Stature (cm) Estimation in Males, Females and Combined Gender.

Male
Parameter	Equation	R	R ²	Adjusted R²	SEE
SP	145.371 + 0.538 SP	0.479	0.229	0.223	5.19736
IAL	147.645 + 0.608 IAL	0.395	0.156	0.149	5.4393
Female
Parameter	Equation	R	R ²	Adjusted R²	SEE
SP	145.878 + 0.322 SP	0.239	0.057	0.049	5.7065
IAL	147.144 + 0.368 IAL	0.265	0.070	0.063	5.6661
Combined Gender
Parameter	Equation	R	R ²	Adjusted R²	SEE
SP	120.219 + 0.933 SP	0.599	0.359	0.356	7.04758
IAL	123.132 + 1.083 IAL	0.599	0.359	0.356	7.0498

Table 4.

Multiple Regression Equations for Stature (cm) Estimation in Males, Females, and Combined Gender.

Male
Equation	R	R ²	Adjusted R²	SEE
139.774 + 0.423 SP + 0.267 IAL	0.499	0.249	0.236	5.15229
Female
Equation	R	R ²	Adjusted R²	SEE
128.021 + 0.364 SP + 0.409 IAL	0.378	0.143	0.128	5.46357
Combined Gender
Equation	R	R ²	Adjusted R²	SEE
105.083 + 0.634 SP + 0.735 IAL	0.698	0.488	0.484	6.31378

Table 5.

Descriptive Statistics of Mean Values of Known Stature and Estimated Stature in Males and Females.

Parameter	Gender	N	Mean	SD	Min	Max
Known stature	Male	120	174.43	5.89	162.00	189.00
	Female	120	161.36	5.85	148.00	175.00
	Combined gender	240	167.89	8.79	148.00	189.00
Estimated stature	Male	120	174.40	2.94	153.31	180.22
	Female	120	161.34	2.21	155.24	166.08
	Combined gender	240	167.86	6.13	131.73	181.94

Table 6.

Paired t-test Between Known and Estimated Height.

Parameter	Gender	Mean	Std Error Mean	t-value	Two-sided Significance
Estimated stature	Male	0.024	0.466	0.051	0.959
	Female	0.021	0.495	0.042	0.966
	Combined Gender	0.030	0.406	0.073	0.942

Discussion

Stature is an important component of the biological profile of an individual and is influenced by several factors, such as racial or ethnic background, hereditary characteristics, climatic conditions and nutritional status. In forensic anthropology, stature constitutes one of the classical ‘big four’ parameters used in identification, along with age, sex and race. Estimating stature plays an important role in identification during a medicolegal autopsy, especially in cases involving unknown bodies, mutilated bodies, advanced decomposition or mass disaster. Anthropometric measurements derived from various body parts have long been used in forensic anthropology to estimate the stature of an individual in cases where the complete body is not available. The current study was conducted in the North-West Indian population to derive the regression equations for stature estimation using IAL and the distance from suprasternal notch to pubic symphysis (SP). In the present study, males exhibited higher mean values for stature, IAL and SP compared to females, which is in concurrence with other studies.^{2, 17} The age of puberty being 2 years later in males gives additional time for growth; Males are generally taller than females, and hence it explains this difference. In the present study, a positive correlation was observed between distance from the suprasternal notch to pubic symphysis (SP) and stature in males, females and a combined gender group (Table 2). Among the parameters measured, SP demonstrated higher correlation with stature in males, while IAL showed relatively stronger correlations in females. These findings indicate that trunk dimensions can serve as a reliable predictor of stature. The comparative analysis of the correlation of IAL with stature, along with studies on different populations, is depicted in Table 7. It was observed from the results of various studies that IAL exhibits a moderately strong correlation with stature like in studies done by Ozalen et al.,² Patel et al.,⁴ Jasuja et al.¹⁷ and Koulapur et al.¹⁸ Our results for correlation of stature with IAL are consistent with studies conducted by Ozalen et al.,² Patel et al.,⁴ Jasuja et al.¹⁷ and Koulapur et al.¹⁸ Similarly, distance from suprasternal notch to pubic symphysis depicted moderately strong positive correlation with the stature in the current study as well as in the study by Bandyopadhyay and Dutta.¹⁶ The similarities between the findings of the present study and the previous research suggest that trunk measurements may provide valuable information for the estimation of stature in forensic contexts.

The regression equations were derived for the estimation of stature from IAL and distance from the suprasternal notch to pubic symphysis (SP) in both males and females, as well as in the combined groups (Tables 3 and 4). The accuracy of regression equations in anthropometric studies is commonly evaluated by the SEE, where the lower SEE indicates better predictive accuracy. In the present study, the SEE was lowest for SP in males and IAL in females (Table 3). However, the difference between the SEE for both parameters is minimal. In the study conducted by Gambhir et al.,¹³ the standard error from the linear regression equation for IAL was ± 6 and ± 5 for males and females, respectively. In another study by Koulapur et al.¹⁸ the standard error from the linear regression equation for IAL was ± 6 and ± 4 for males and females, respectively. The standard errors came out to be ± 8 cm for males and 181 + 5 cm for females in a study done by Momonchand and Devi.¹⁹ The standard errors came out to be ± 6 cm in males and ± 5 cm in females in the study conducted by Patel et al.⁴ The SEE value for the linear regression equation, and stature was 2.2395 and 1.8632 in the study conducted by Ozaslan et al.² on Turkish population.

The mean difference between the estimated stature and known stature for males, females and in the combined gender group in the current study is depicted in Table 5 and was found to be similar. No statistically significant difference (p value >.05) was observed between the known stature and estimated stature by using a paired t-test (Table 6). Therefore, the derived equations can be employed for the estimation of stature. No statistically significant difference was observed between the known stature and estimated stature based on the IAL by paired t-test in other studies as well.^{1, 13, 18}

Study Limitations

Population migration, colonisation and genetic admixture have an impact on local communities; as a result, these aspects were overlooked while classifying study participants. The current study’s sample size was not very large. Therefore, additional research with a bigger sample size that accurately depicts the specific population after accounting for confounding variables like migration, and so on, could be done.

Table 7.

Comparison of Stature and IAL of the Present Study with Various Studies.

Study (Citation)	Population	Parameter	M	F	CG
Present Study	North-west Indian population	n	120	120	240
		Mean IAL	44.06	38.61	41.34
		Mean Stature	174.43	161.36	167.89
		r	0.395	0.265	0.599
Bharath and Babu¹⁴	Tamil Population	n	30	30	60
		Mean IAL	45.33	39.43
		Mean Stature	174.16	158.4
		r	0.282	–0.235
Nayak and Sheikh¹	Hyderabad Population	n	97	97	194
		Mean IAL	46.029	41.63
		Mean Stature	169.83	156.26
		r	0.827	0.7	0.8163
Patel et al.⁴	Gujrati Population	n	150	150	300
		Mean IAL
		Mean Stature
		r	0.31	0.23	0.59
Gambhir et al.¹³	Tamil Population	n	100	100	200
		Mean IAL	45	40
		Mean Stature	168.25	156.5
		r	0.38	0.45	0.61
Ozaslan et al.²	Turkish Population	n	216	121	337
		Mean IAL	38.6	34.9
		Mean Stature
		r	0.42	0.26

Conclusion

The present study has established a definite correlation between stature and two parameters of the body trunk, namely IAL and distance from the suprasternal notch to the pubic symphysis. Out of these two parameters studied, the distance from the suprasternal notch to pubic symphysis showed the highest degree of correlation with stature in males, and IAL showed the highest correlation in females. The regression equations formulated in this study provide valid and reliable stature estimations. The known stature and estimated stature from the derived regression equations were found to be similar, with no statistically significant difference observed. Therefore, it is concluded that stature can be estimated from these parameters of the body trunk in situations where only the body trunk is available for postmortem examination and subsequent identification.

Footnotes

Declaration of Conflict of Interests

The authors declared no potential conflicts of interest with respect to the research, authorship and/or publication of this article.

Ethical Approval

Prior approval was obtained from the Institutional Ethical Committee (SHKM/IEC/2019/66 dated 19 April 2019).

Funding

The authors received no financial support for the research, authorship and/or publication of this article.

Informed Consent

Written informed consent was obtained from all participants prior to their inclusion in the study.

ORCID iDs

Praveen Dixit

Hitesh Chawla

Ashish Tyagi

References

Nayak

and Sheikh

NA.

Regression equation analysis on inter-acromial length for stature estimation in South India population. Indian J Forensic Med Pathol 2019; 12(1): 13–17.

Ozaslan

, Karadayi

, Kolusayin

, . Stature estimation from bi-acromial and bi-iliocristal measurements. Rom J Leg Med 2011; 19(3): 171–176.

Sume

BW.

Estimation of body height from percutaneous length of tibia in Debre Markos University students, North West Ethiopia. Egypt J Forensic Sci 2019; 9: 1–8.

Patel

, Bastia

, Kumar

, . Estimation of adult human stature from measurements of inter-acromial length in Gujarati population of India. J Indian Acad Forensic Med 2015; 37(4): 365–368.

Gupta

, Kumari

, Palimar

, . Stature estimation using interacromial length in South and North Indian population. Ann Bioanthropol 2015; 3(2): 63–65.

Krishan

Anthropometry in forensic medicine and forensic science: ‘forensic anthropometry’. Internet J Forensic Sci 2007; 2(1): 95–97.

Joshi

, Patel

and Dongre

AV.

Regression equation of height on ulnar length. Indian J Med Res 1964; 52: 1088–1091.

Ozaslan

, Işcan

, Ozaslan

, . Estimation of stature from body parts. Forensic Sci Int 2003; 132(1): 40–45.

Verma

, Krishan

, Rani

, . Stature estimation in forensic examinations using regression analysis: A likelihood ratio perspective. Forensic Sci Int Rep 2020; 2: 100069.

10.

Siddiqui

and Shah

MA.

Estimation of stature from long bones of Punjabis. Indian J Med Res 1944; 32: 104–108.

11.

Shakya

, Mishra

and Pandey

Estimation of stature from upper arm length. Int J Health Sci Res 2021; 11(5): 23–29.

12.

Kamal

and Yadav

PK.

Estimation of stature from different anthropometric measurements in Kori population of North India. Egypt J Forensic Sci 2016; 6(4): 468–477.

13.

Gambhir

, Arun

PKS

, Sanjeev

, . Estimation of stature from percutaneous inter-acromial length in Tamil population. Int J Progr Res Sci Eng 2022; 3(6): 150–152.

14.

Bharath

and Babu

KY.

Estimation of stature using measurement of inter-acromial distance. Eur J Mol Clin Med 2022; 9(8): 263–270.

15.

Selvamurugan

, Muralidharan

, Tamilmani

, . A study in correlation of stature assessment using practising anthropometric parameters like forearm length, sternal notch to upper border of pubic symphysis and tibial length in local population by using bodies subjected to autopsy. Indian J Forensic Med Toxicol 2019; 13(3): 60–62.

16.

Bandyopadhyay

and Dutta

SK.

Estimation of stature based on arm span, vertex of head to upper level of symphysis pubis distance and suprasternal notch to upper level of symphysis pubis distance in adult Bengali population: An autopsy based study. Int J Sci Res 2019; 8(2): 1–3.

17.

Jasuja

, Singh

and Jain

Estimation of stature from foot and shoe measurements by multiplication factors: A revised attempt. Forensic Sci Int 1991; 50(2): 203–215.

18.

Koulapur

, Agarwal

and Mestri

SC.

Estimation of stature by anthropometric measurements of inter-acromial length. J Indian Acad Forensic Med 2010; 32(2): 101–103.

19.

Momonchand

and Devi

TM.

Determination of stature from the interacromial length. J Forensic Med Toxicol 1999; 16(1): 72–73.