- Email: [email protected]
Is the length of the sternum reliable for estimating adult stature? A pilot study using fresh sterna and a test of two methods using dry sterna
Forensic Science International 220 (2012) 292.e1–292.e4
Contents lists available at SciVerse ScienceDirect
Forensic Science International journal homepage: www.elsevier.com/locate/forsciint
Forensic anthropology population data
Is the length of the sternum reliable for estimating adult stature? A pilot study using fresh sterna and a test of two methods using dry sterna Luı´sa Marinho a, Dina Almeida b, Agostinho Santos b,c, Hugo F.V. Cardoso c,d,* a
Departamento de Cieˆncias da Vida – FCTUC, Apartado, 3046 3001-401 Coimbra, Portugal Instituto Nacional de Medicina Legal, I.P. – Norte, Jardim Carrilho Videira, 4050-167 Porto, Portugal c Universidade do Porto, Faculdade de Medicina, Jardim Carrilho Videira, 4050-167 Porto, Portugal d Universidade de Lisboa, Museu Nacional de Histo´ria Natural & Centro de Biologia Ambiental, Rua da Escola Polite´cnica, 58 1269-102 Lisboa, Portugal b
A R T I C L E I N F O
A B S T R A C T
Article history: Received 26 July 2011 Received in revised form 27 January 2012 Accepted 15 February 2012 Available online 14 March 2012
Stature estimation is one of the four attributes of the biological profile obtained from human skeletal remains. The length of the long bones has been consistently used to estimate stature from regression equations, but these may be useless when dealing with fresh or decomposed mutilated remains. Until recently, there was no consistent assessment of the reliability of measurements of the sternum for stature estimation. The purpose of this paper is to test previously developed regression formulae for stature based on measurements of the dry sternum and to assess the reliability of measurements of the fresh sternum in estimating stature. The formulae developed by Menezes et al. and Singh et al. were applied to a sample of 5 known stature skeletons from the identified human skeletal collection curated at the National Museum of Natural History, in Lisbon, Portugal. Testing of these formulae showed that estimated stature confidence intervals do not allow discrimination between individuals with similar stature. The length of the fresh sternum was measured on a sample of 45 male individuals autopsied at the National Institute of Legal Medicine – North Delegation (Porto, Portugal). Cadaver length was regressed on sternum length and a simple linear regression formula was obtained. The regression model provided a 95% confidence interval of 13.32 cm and a correlation coefficient of only 0.329. Compared to other studies, regression formulae based on the length of the sternum provided considerably larger standard errors than that based on long bone lengths. These results suggest that the length of the sternum has limited forensic value and relatively low reliability in estimating stature from mutilated human skeletal remains, either skeletonized or fresh. ß 2012 Elsevier Ireland Ltd. All rights reserved.
Keywords: Forensic anthropology Identification Stature estimation Sternum
1. Introduction Forensic anthropology plays a crucial role in the medico-legal investigation of unidentified skeletonized or decomposed human remains. The determination of sex, age, stature and ancestry of the decedent from the skeleton forms the basis of the so-called biological profile [1,2]. The first attempts in estimating stature from human remains involved the re-articulation of the whole skeleton and measuring the total length of several anatomical elements [3,4]. Although highly accurate, this anatomical approach can be particularly problematic whenever the skeleton is incomplete or badly preserved. Regression equations, based on the proportional relationship between long bone length and stature, has provided important alternatives to the anatomical method [5– 10]. The regression approach has become very popular because it
* Corresponding author at: Universidade do Porto, Faculdade de Medicina, Jardim Carrilho Videira, 4050-167 Porto, Portugal. Tel.: +351 222073850. E-mail address: [email protected] (Hugo F.V. Cardoso). 0379-0738/$ – see front matter ß 2012 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.forsciint.2012.02.012
has proved to be very useful in most circumstances, allowing the expert to estimate stature from the size of only one bone and providing reasonably good accuracy. Developments in stature estimation methods have continued and exploring the relationship between stature and the size of various skeletal elements is useful because in certain scenarios, such as mass disasters or crimes involving dismemberment of the victim, the long bones may not be available or complete. In these cases the expert requires alternative solutions for stature estimation [11]. Several studies have assessed the relationship between stature and the size of the cranium [12,13], scapula [14], vertebrae [15,16] or sternum [17–19]. The sternum may be useful in stature estimation of skeletonized remains, but in particular of fresh or partially decomposed mutilated human remains, because it is easily removed and dissected from the rib cage or torso, even if considerable soft tissue is present. Recently, Menezes et al. [18], as well as Singh et al. [19], have assessed the usefulness of the length of the macerated sternum in stature estimation of two adult Indian populations and concluded that sternum length correlates relatively well with stature and can be used in the estimation of
L. Marinho et al. / Forensic Science International 220 (2012) 292.e1–292.e4
292.e2
stature, when long bones are not available for that purpose. Prior to Menezes et al. and Singh et al., only Dwight [17], in the late 19th century, had discussed the potential use of the sternum for estimation of stature. This study has two different but closely intertwined goals. The first is to test previously developed stature estimation formulae from sterna measurements. This was carried out by testing the Menezes et al.’s [18] and Singh et al.’s [19] formulae in a small sample of individuals of known stature from the identified skeletal collection curated at the National Museum of Natural History, in Lisbon, Portugal [20]. Although stature estimation methods tend to be population-specific, testing these methods on a Portuguese sample is important to understand the cross-population applicability of the methods and whether methods are specific to a certain population or specific to populations with similar body proportions. The second goal was to further explore the association between the size of the sternum and stature, which can be particularly useful in stature estimation of mutilated human remains. To explore this association we have used an autopsy sample where the length of the fresh sternum and stature were obtained from Portuguese cadavers admitted to the Northern Delegation of the National Institute of Legal Medicine, in Porto, Portugal. 2. Materials and methods This study is based on two sets of data. The first was used to test the formulae developed by Menezes et al. [18] and Singh et al. [19] for stature estimation from the length of the sternum, and consists of 5 macerated dry sterna obtained from skeletons of known stature in the identified skeletal collection curated at the National Museum of Natural History, in Lisbon, Portugal [20]. The maximum length of the sternum was measured according to the technique described by Ashley [21], using a spreading caliper. The 5 individuals had been subjected to autopsy and their cadaver length was obtained from the archives of the National Institute of Legal Medicine – South Delegation, in Lisbon. Only 5 individuals could be utilized in this study, because from a list of over 40 individuals with known cadaver length, only these preserved a complete sternum. Cadaver length for each of these individuals was then compared with the stature estimated from the Menezes et al.’s [18] and Singh et al.’s [19] formulae. The second set of data was used to develop regression equations from the fresh sternum. Cadaver and sternum length was obtained from 45 male individuals submitted to autopsy at the National Institute of Legal Medicine – North Delegation, in Porto (Portugal), between February and August 2010. The ages of the individuals vary between 18 and 90 years, with a mean (SD) of 52.89 (15.61) years. Each individual is identified with a registration number which was used throughout this research, thus ensuring absolute confidentiality of personal data. The cause of death was not considered and the only exclusion criterion was the mutilation of the body affecting the measurement of cadaver stature and of the sternum as a consequence of disease, trauma or surgery. The sternum was measured during autopsy after the removal of the chest plate and using a spreading caliper. The distance between the inferior point of the jugular notch and the inferior point of the xiphoid process was recorded. The main concern with this measurement was to assure that both extremities of the spreading caliper were directly over bone, by removing as much soft tissue as possible and exposing the underlying bone surface (Fig. 1). The sternum was not preserved beyond the autopsy and was subsequently returned to the body. Cadaver length is routinely measured prior to autopsy using a sliding caliper, as the maximum length between the vertex of the head and heel of the foot, with the cadaver laying on its back and in full extension. Data were analyzed using SPSS Inc. [22] version 17.0 to assess the association between stature and sternum length by simple regression analysis. Both variables were found to be normally distributed and their variances homogeneous. Rather than comparing the point stature estimate and cadaver length
Fig. 1. Measurement of the length of the sternum using a spreading caliper, as the distance between the inferior point of the jugular notch and the inferior point of the xiphoid process.
we wished to assess whether cadaver length is included in the 95% confidence interval of estimated stature. A test of the reliability of the data was carried out by estimating the amount of inter-observer error on a small subset (n = 9) of the sample, where sterna had been measured by two different researchers. Intra-observer error could not be assessed because the sternum was not preserved after autopsy and, consequently, not enough time could have elapsed for the sterna to be re-measured by the same researcher. Inter-observer error was estimated by calculating the intraclass correlation coefficient using SPSS Inc. [22] version 17.0 as well.
3. Results The results from the testing of Menezes et al.’s [18] and Singh et al.’s [19] formulae are shown in Table 1. The confidence intervals encompasses cadaver length in all individuals using both methods, and neither the formulae of Menezes et al.’s [18] or that of Singh et al.’s [19] seem to consistently overestimate or underestimate cadaver length. The fact that stature in the test sample does not show great variation, and that individuals cluster together at two sternum lengths (162–164 mm and 172–173 mm), demonstrates that stature estimated from the sternum cannot be reliably used as exclusion criteria.
Table 1 Test of Menezes et al.’s [18] and Singh et al.’s [19] stature estimation formulae. Sternum length (cm)
13.9 12.6 15.8 14.1 15.9
Cadaver length (cm)
162 164 164 174 173
Menezes et al. [18]
Singh et al. [19]
Estimated stature (cm)
C.I. (95%)
Estimated stature (cm)
C.I. (95%)
165 161 172 166 172
154.0–176.1 150.0–173.1 161.0–184.1 155.0–178.1 161.1–184.1
164 160 171 165 171
149.1–178.9 145.1–174.9 156.1–185.9 150.1–179.9 156.1–185.9
L. Marinho et al. / Forensic Science International 220 (2012) 292.e1–292.e4
In our autopsy sample, cadaver length ranged from 156 cm to 184 cm, with a mean (SD) of 167.9 (6.90) cm, and mean sternum length (SD) was 203.9 (14.21) mm, ranging from 171 mm to 235 mm. The intra-class correlation coefficient (ICC = 0.994, 95%CI = 0.975–0.999) shows an almost perfect agreement between the measurements of two researchers. When regressing cadaver length (CL) on sternum length (SL), the following simple linear regression formula was obtained: CL = 135.322 + 0.160 SL 13.32 (95% C.I.). The standard error of the estimate (SEE) obtained is 6.59 and the correlation coefficient (R) is 0.329. Only 11% of the variation in the sample is explained by this regression model (R2 = 0.11). 4. Discussion Since Dwight’s study [17], in the late 19th century, the sternum has received very little attention as a means to estimate stature from human skeletal remains. This study, together with that of Menezes et al.’s [18] and Singh et al.’s [19], has filled an important gap. Whereas the first two focused on dry bone observations, in this study we wished to provide an alternative means to the estimation of stature with measurements of fresh sterna, namely when maceration cannot be carried out, due to time constraints for example, or when a fresh or decomposed mutilated cadaver is being identified. Overall, our results show that the sternum is not particularly useful for stature estimation. Not only is the correlation between sternum length and stature low, but also the standard error of the estimate is relatively large, resulting in confidence intervals for stature which are also large and of limited use to discriminate between individuals of similar stature. A simple comparison with the results of other studies, where stature has been regressed on long bone lengths and other bone measurements, shows that the standard error of the estimate (as well as the confidence interval) in stature estimation formulae using the sternum is among the highest (Table 2). Studies shown in this table were chosen solely to illustrate the variation in standard errors obtained from stature regression formulae. In this table, the rank of increasing error for each regression formula is shown. Regression errors for the length of the sternum are almost twice of that for the length of the long bones and are similar to measurement of the fleshed hand and foot, and measurements of the sacrum and coccyx.
292.e3
Although in this study the standard error of the estimate obtained is smaller than that of Singh et al.’s [19], Menezes et al.’s [18] shows the smallest error. On the other hand, the correlation coefficient obtained by Menezes et al. [18] and Singh et al. [19] is greater than that of this study. Nonetheless, confidence intervals are still large and the correlation is always under 0.7. This indicates that for the same stature, individuals vary considerably in sternum length. The fact that we have used cadaver length and measured the fresh sternum is unlikely to have biased our results significantly, because any errors tend to be systematic. Cadaver length tends to be greater than true living stature and the length of the fresh sternum may also tend to be greater than that of the macerated sternum, due to the presence of any soft tissue and lack of shrinkage. However, this will not affect the relationship between cadaver length and the length of the sternum. In addition, our inter-observer test shows a high reliability for the data. Unfortunately, we could not preserve the sterna and assess the amount of shrinkage with maceration. Consequently, the formula obtained in this study cannot be applied to dry macerated sterna. An important consideration to make when analyzing and applying different stature estimation methods is that they are not universally applicable and several studies have demonstrated that a model developed from a specific population may not give as reliable estimates when applied to another [20,25]. This results mostly from differences in body proportions and differences in proportions between stature and bone size, that result from differences in environmental conditions during growth that affect these proportions. For example, stunted children and adults have proportionally shorter legs when compared to their normal counterparts [28]. As a consequence, individuals from developed nations may differ in body proportions relative to stature when compared to individuals from the developing countries. Unfortunately, due to our small sample we could not assertively conclude whether the Menezes et al.’s [18] and Singh et al.’s [19] formulae can accurately be applied to the Portuguese population. Our results suggest that these formulae may not tend to overestimate or underestimate true stature. However, the Lisbon collection sample of macerated sterna is comprised of individuals who lived and died in the late 19th and early 20th centuries in Portugal. In the late
Table 2 Stature estimation error (SEE – standard error of the estimate and CI – confidence intervals) reported in other studies and compared to that of the sternum in this study and other. Formulae in each study is ranked by increasing estimation error. Rank
SEE
CI (95%)
Trotter and Gleser [6]
Femur Humerus
1 7
3.27 4.05
6.41 7.94
Mendonc¸a [9]
Femur Humerus
4 8
3.50 4.25
6.96 8.44
Ross and Konigsberg [23]
Humerus Femur Tibia
6 2 3
4.03 3.30 3.39
7.90 6.47 6.64
Duyar and Pelin [24] Pelin et al. [16] Sanli et al. [25]
Tibia Sacrum (e.g.) Hand (length) Foot (length)
5 18 9 10
3.94 6.59 4.27 4.30
7.82 13.32 8.50 8.56
Duyar et al. [26]
Ulna Tibia
15 5
5.59 3.94
11.07 7.80
Celbis and Agritmis [10]
Radio Ulna
12 13
4.70 4.80
9.35 9.55
Nagesh and Kumar [15] Krishan and Sharma [27]
Vertebral column Hand (length, e.g.) Foot (length, e.g.)
11 14 17
4.38 5.22 6.19
8.73 10.34 12.26
Menezes et al. [18] Singh et al. [19] Present study
Sternum (length) Sternum (combined length) Sternum (length)
16 19 18
5.64 7.24 6.59
11.05 14.19 13.32
Bone
292.e4
L. Marinho et al. / Forensic Science International 220 (2012) 292.e1–292.e4
20th century, the country has experienced dramatic shifts and improvements in living conditions, so that modern Portuguese are much taller than their earlier counterparts [29,30]. Compared to modern Indians, represented in the Menezes et al. and Singh et al. samples [18,19], the Lisbon collection individuals may compare favorably, and hence the lack of significant errors in stature estimation using their formulae. Given that regression equations may not be population-specific per se, but specific to populations who share similar body proportions, it would be interesting to know whether a similar result will be obtained using samples of other bones from this collection on previously derived equations for Indian population groups. When applying regression formulae to estimate stature it is also important to consider what is the ‘‘stature’’ that is being determined. Most regression models are based on cadaver length or living standing height, but in most forensic scenarios, the antemortem information available about stature is usually reported stature, obtained from driver’s license, national citizenship identification documents or from relatives [31,32]. Although cadaver length, standing height and reported height are related, cadaver length tends to be greater than standing height, and possibly reported stature is the greatest of the three [31,33,34]. Consequently, it is crucial to distinguish between living, cadaver or forensic stature [35–37]. 5. Conclusion The study does not provide support for the use of the length of the sternum as a reliable estimator of stature in mutilated human skeletal remains, either skeletonized or fleshed. This is contrary to previous studies, which do not seem to stress enough the limitations of using sternum measurements for that purpose. Although if a quick estimate is required and maceration cannot be carried out, as the sternum is easily removed and dissected from a relatively fresh or badly decomposed body, then this may work as an alternative for statute estimation. However, particular attention should be given in the exclusion of cases based on the stature data, as there seems to be considerable variation in sternum length for the same stature. Acknowledgments We would like to thank Dr. Jorge Soares, former director of the Southern Delegation of the National Institute of Legal Medicine in Lisbon, for authorizing the access to the autopsy records of individuals in the Lisbon collection. Thank you to the anonymous reviewers for their remarks, which helped improving the manuscript. References [1] E. Cunha, C. Cattaneo, Forensic pathology and forensic anthropology: the state of art, in: A. Schmitt, E. Cunha, J. Pinheiro (Eds.), Forensic Anthropology and Medicine, Humana Press, Totowa, New Jersey, 2006, pp. 39–53. [2] L. Scheuer, Application of osteology to forensic medicine, Clin. Anat. 15 (4) (2002) 297–312 (review). [3] G. Fully, Une nouvelle me´thode de de´termination de la taille, Ann. Med. Legale 35 (1956) 266–273. [4] M.H. Raxter, B.M. Auerbach, C.B. Ruff, Revision of the fully technique for estimating statures, Am. J. Phys. Anthropol. 130 (3) (2006) 374–384.
[5] M. Trotter, G.C. Gleser, A re-evaluation of estimation of stature based on measurements of stature taken during life and of long bones after death, Am. J. Phys. Anthropol. 16 (1) (1958) 79–123. [6] M. Trotter, G.C. Gleser, Estimation of stature from long bones of American Whites and Negroes, Am. J. Phys. Anthropol. 10 (4) (1952) 463–514. [7] B.C. Didia, E.C. Nduka, O. Adele, Stature estimation formulae for Nigerians, J. Forensic Sci. 54 (1) (2009) 20–21. [8] A. Ozaslan, M.Y. Is¸can, I. Ozaslan, H. Tug˘cu, S. Koc¸, Estimation of stature from body parts, Forensic Sci. Int. 132 (1) (2003) 40–45. [9] M.C. Mendonc¸a, Estimation of height from the length of long bones in a Portuguese adult population, Am. J. Phys. Anthropol. 112 (1) (2000) 39–48. [10] O. Celbis, H. Agritmis, Estimation of stature and determination of sex from radial and ulnar bone lengths in a Turkish corpse sample, Forensic Sci. Int. 158 (2–3) (2006) 135–139. [11] C. Cattaneo, Forensic anthropology: developments of a classical discipline in the new millennium, Forensic Sci. Int. 165 (2–3) (2007) 185–193 (review). [12] S. Kalia, S.K. Shetty, K. Patil, Mahima, Stature estimation using odontometry and skull anthropometry, Indian J. Dent. Res. 19 (2) (2008) 150–154. [13] K. Krishan, R. Kumar, Determination of stature from cephalo-facial dimensions in a North Indian population, J. Leg. Med. 9 (2007) 128–133. [14] R.M. Burke, Can we estimate stature from the scapula? A test considering sex and ancestry, Thesis, BSc, University of Idaho (2005). [15] K.R. Nagesh, G.P. Kumar, Estimation of stature from vertebral column length in South Indians, Leg. Med. (Tokyo) 8 (5) (2006) 269–272. [16] C. Pelin, I. Duyar, E.M. Kayahan, R. Zag˘yapan, A.M. Ag˘ildere, A. Erar, Body height estimation based on dimensions of sacral and coccygeal vertebrae, J. Forensic Sci. 50 (2) (2005) 294–297. [17] T. Dwight, The sternum as an index of sex, age and height, J. Anat. 15 (1881) 327– 330. [18] R.G. Menezes, et al., Stature estimation from the length of the sternum in South Indian males: a preliminary study, J. Forensic Leg. Med. 16 (2009) 441–443. [19] J. Singh, et al., Skeletal height estimation from regression analysis of sterna lengths in a Northwest Indian population of Chandigarh region: a postmortem study, Forensic Sci. Int. 206 (2011) 211.e1–211.e8. [20] H.F.V. Cardoso, Brief communication: the collection of identified human skeletons housed at the Bocage Museum (National Museum of Natural History), Lisbon, Portugal, Am. J. Phys. Anthropol. 129 (2006) 173–176. [21] G.T. Ashley, The relationship between the pattern of ossification and the definitive shape of the mesosternum in man, J. Anat. 90 (1956) 87–105. [22] SPSS Inc., SPSS Version 17. 0 for Windows User’s Guide, SPSS Inc., Chicago, 2008. [23] A.H. Ross, L.W. Konigsberg, New formulae for estimating stature in the Balkans, J. Forensic Sci. 47 (1) (2002) 165–167. [24] I. Duyar, C. Pelin, Body height estimation based on tibia length in different stature groups, Am. J. Phys. Anthropol. 122 (1) (2003) 23–27. [25] S.G. Sanli, E.D. Kizilkanat, N. Boyan, E.T. Ozsahin, M.G. Bozkir, R. Soames, et al., Stature estimation based on hand length and foot length, Clin. Anat. 18 (8) (2005) 589–596. [26] I. Duyar, C. Pelin, R. Zagyapan, A new method of stature estimation for forensic anthropological application, Anthropol. Sci. 114 (2006) 23–27. [27] K. Krishan, A. Sharma, Estimation of stature from dimensions of hands and feet in a North Indian population, J. Forensic Leg. Med. 14 (6) (2007) 327–332. [28] B. Bogin, L. Rios, Rapid morphological change in living humans: implications for modern human origins, Comp. Biochem. Physiol. A: Mol. Integr. Physiol. 136 (1) (2003) 71–84. [29] C. Padez, Secular trend in stature in the Portuguese population (1904–2000), Ann. Hum. Biol. 30 (2003) 262–278. [30] H.F.V. Cardoso, Secular changes in body height and weight of Portuguese boys over one century, Am. J. Hum. Biol. 20 (2008) 270–277. [31] E. Giles, D.L. Hutchinson, Stature- and age-related bias in self-reported stature, J. Forensic Sci. 36 (1991) 765–780. [32] S. Ousley, Should we estimate biological or forensic stature? J. Forensic Sci. 40 (5) (1995) 768–773. [33] J.H. Himes, A.F. Roche, Reported versus measured adult statures, Am. J. Phys. Anthropol. 58 (1982) 335–341. [34] L.W. Konigsberg, S.M. Hens, L.M. Jantz, W.L. Jungers, Stature estimation and calibration: Bayesian and maximum likelihood perspectives in physical anthropology, Am. J. Phys. Anthropol. 27 (27) (1998) 65–92. [35] B.J. Adams, N.P. Herrmann, Estimation of living stature from selected anthropometric (soft tissue) measurements: applications for forensic anthropology, J. Forensic Sci. 54 (4) (2009) 753–760. [36] E. Giles, Corrections for age in estimating older adults’ stature from long bones, J. Forensic Sci. 36 (3) (1991) 898–901. [37] P. Willey, T. Falsetti, Inaccuracy of height information on driver’s licenses, J. Forensic Sci. 36 (3) (1991) 813–819.
Contents lists available at SciVerse ScienceDirect
Forensic Science International journal homepage: www.elsevier.com/locate/forsciint
Forensic anthropology population data
Is the length of the sternum reliable for estimating adult stature? A pilot study using fresh sterna and a test of two methods using dry sterna Luı´sa Marinho a, Dina Almeida b, Agostinho Santos b,c, Hugo F.V. Cardoso c,d,* a
Departamento de Cieˆncias da Vida – FCTUC, Apartado, 3046 3001-401 Coimbra, Portugal Instituto Nacional de Medicina Legal, I.P. – Norte, Jardim Carrilho Videira, 4050-167 Porto, Portugal c Universidade do Porto, Faculdade de Medicina, Jardim Carrilho Videira, 4050-167 Porto, Portugal d Universidade de Lisboa, Museu Nacional de Histo´ria Natural & Centro de Biologia Ambiental, Rua da Escola Polite´cnica, 58 1269-102 Lisboa, Portugal b
A R T I C L E I N F O
A B S T R A C T
Article history: Received 26 July 2011 Received in revised form 27 January 2012 Accepted 15 February 2012 Available online 14 March 2012
Stature estimation is one of the four attributes of the biological profile obtained from human skeletal remains. The length of the long bones has been consistently used to estimate stature from regression equations, but these may be useless when dealing with fresh or decomposed mutilated remains. Until recently, there was no consistent assessment of the reliability of measurements of the sternum for stature estimation. The purpose of this paper is to test previously developed regression formulae for stature based on measurements of the dry sternum and to assess the reliability of measurements of the fresh sternum in estimating stature. The formulae developed by Menezes et al. and Singh et al. were applied to a sample of 5 known stature skeletons from the identified human skeletal collection curated at the National Museum of Natural History, in Lisbon, Portugal. Testing of these formulae showed that estimated stature confidence intervals do not allow discrimination between individuals with similar stature. The length of the fresh sternum was measured on a sample of 45 male individuals autopsied at the National Institute of Legal Medicine – North Delegation (Porto, Portugal). Cadaver length was regressed on sternum length and a simple linear regression formula was obtained. The regression model provided a 95% confidence interval of 13.32 cm and a correlation coefficient of only 0.329. Compared to other studies, regression formulae based on the length of the sternum provided considerably larger standard errors than that based on long bone lengths. These results suggest that the length of the sternum has limited forensic value and relatively low reliability in estimating stature from mutilated human skeletal remains, either skeletonized or fresh. ß 2012 Elsevier Ireland Ltd. All rights reserved.
Keywords: Forensic anthropology Identification Stature estimation Sternum
1. Introduction Forensic anthropology plays a crucial role in the medico-legal investigation of unidentified skeletonized or decomposed human remains. The determination of sex, age, stature and ancestry of the decedent from the skeleton forms the basis of the so-called biological profile [1,2]. The first attempts in estimating stature from human remains involved the re-articulation of the whole skeleton and measuring the total length of several anatomical elements [3,4]. Although highly accurate, this anatomical approach can be particularly problematic whenever the skeleton is incomplete or badly preserved. Regression equations, based on the proportional relationship between long bone length and stature, has provided important alternatives to the anatomical method [5– 10]. The regression approach has become very popular because it
* Corresponding author at: Universidade do Porto, Faculdade de Medicina, Jardim Carrilho Videira, 4050-167 Porto, Portugal. Tel.: +351 222073850. E-mail address: [email protected] (Hugo F.V. Cardoso). 0379-0738/$ – see front matter ß 2012 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.forsciint.2012.02.012
has proved to be very useful in most circumstances, allowing the expert to estimate stature from the size of only one bone and providing reasonably good accuracy. Developments in stature estimation methods have continued and exploring the relationship between stature and the size of various skeletal elements is useful because in certain scenarios, such as mass disasters or crimes involving dismemberment of the victim, the long bones may not be available or complete. In these cases the expert requires alternative solutions for stature estimation [11]. Several studies have assessed the relationship between stature and the size of the cranium [12,13], scapula [14], vertebrae [15,16] or sternum [17–19]. The sternum may be useful in stature estimation of skeletonized remains, but in particular of fresh or partially decomposed mutilated human remains, because it is easily removed and dissected from the rib cage or torso, even if considerable soft tissue is present. Recently, Menezes et al. [18], as well as Singh et al. [19], have assessed the usefulness of the length of the macerated sternum in stature estimation of two adult Indian populations and concluded that sternum length correlates relatively well with stature and can be used in the estimation of
L. Marinho et al. / Forensic Science International 220 (2012) 292.e1–292.e4
292.e2
stature, when long bones are not available for that purpose. Prior to Menezes et al. and Singh et al., only Dwight [17], in the late 19th century, had discussed the potential use of the sternum for estimation of stature. This study has two different but closely intertwined goals. The first is to test previously developed stature estimation formulae from sterna measurements. This was carried out by testing the Menezes et al.’s [18] and Singh et al.’s [19] formulae in a small sample of individuals of known stature from the identified skeletal collection curated at the National Museum of Natural History, in Lisbon, Portugal [20]. Although stature estimation methods tend to be population-specific, testing these methods on a Portuguese sample is important to understand the cross-population applicability of the methods and whether methods are specific to a certain population or specific to populations with similar body proportions. The second goal was to further explore the association between the size of the sternum and stature, which can be particularly useful in stature estimation of mutilated human remains. To explore this association we have used an autopsy sample where the length of the fresh sternum and stature were obtained from Portuguese cadavers admitted to the Northern Delegation of the National Institute of Legal Medicine, in Porto, Portugal. 2. Materials and methods This study is based on two sets of data. The first was used to test the formulae developed by Menezes et al. [18] and Singh et al. [19] for stature estimation from the length of the sternum, and consists of 5 macerated dry sterna obtained from skeletons of known stature in the identified skeletal collection curated at the National Museum of Natural History, in Lisbon, Portugal [20]. The maximum length of the sternum was measured according to the technique described by Ashley [21], using a spreading caliper. The 5 individuals had been subjected to autopsy and their cadaver length was obtained from the archives of the National Institute of Legal Medicine – South Delegation, in Lisbon. Only 5 individuals could be utilized in this study, because from a list of over 40 individuals with known cadaver length, only these preserved a complete sternum. Cadaver length for each of these individuals was then compared with the stature estimated from the Menezes et al.’s [18] and Singh et al.’s [19] formulae. The second set of data was used to develop regression equations from the fresh sternum. Cadaver and sternum length was obtained from 45 male individuals submitted to autopsy at the National Institute of Legal Medicine – North Delegation, in Porto (Portugal), between February and August 2010. The ages of the individuals vary between 18 and 90 years, with a mean (SD) of 52.89 (15.61) years. Each individual is identified with a registration number which was used throughout this research, thus ensuring absolute confidentiality of personal data. The cause of death was not considered and the only exclusion criterion was the mutilation of the body affecting the measurement of cadaver stature and of the sternum as a consequence of disease, trauma or surgery. The sternum was measured during autopsy after the removal of the chest plate and using a spreading caliper. The distance between the inferior point of the jugular notch and the inferior point of the xiphoid process was recorded. The main concern with this measurement was to assure that both extremities of the spreading caliper were directly over bone, by removing as much soft tissue as possible and exposing the underlying bone surface (Fig. 1). The sternum was not preserved beyond the autopsy and was subsequently returned to the body. Cadaver length is routinely measured prior to autopsy using a sliding caliper, as the maximum length between the vertex of the head and heel of the foot, with the cadaver laying on its back and in full extension. Data were analyzed using SPSS Inc. [22] version 17.0 to assess the association between stature and sternum length by simple regression analysis. Both variables were found to be normally distributed and their variances homogeneous. Rather than comparing the point stature estimate and cadaver length
Fig. 1. Measurement of the length of the sternum using a spreading caliper, as the distance between the inferior point of the jugular notch and the inferior point of the xiphoid process.
we wished to assess whether cadaver length is included in the 95% confidence interval of estimated stature. A test of the reliability of the data was carried out by estimating the amount of inter-observer error on a small subset (n = 9) of the sample, where sterna had been measured by two different researchers. Intra-observer error could not be assessed because the sternum was not preserved after autopsy and, consequently, not enough time could have elapsed for the sterna to be re-measured by the same researcher. Inter-observer error was estimated by calculating the intraclass correlation coefficient using SPSS Inc. [22] version 17.0 as well.
3. Results The results from the testing of Menezes et al.’s [18] and Singh et al.’s [19] formulae are shown in Table 1. The confidence intervals encompasses cadaver length in all individuals using both methods, and neither the formulae of Menezes et al.’s [18] or that of Singh et al.’s [19] seem to consistently overestimate or underestimate cadaver length. The fact that stature in the test sample does not show great variation, and that individuals cluster together at two sternum lengths (162–164 mm and 172–173 mm), demonstrates that stature estimated from the sternum cannot be reliably used as exclusion criteria.
Table 1 Test of Menezes et al.’s [18] and Singh et al.’s [19] stature estimation formulae. Sternum length (cm)
13.9 12.6 15.8 14.1 15.9
Cadaver length (cm)
162 164 164 174 173
Menezes et al. [18]
Singh et al. [19]
Estimated stature (cm)
C.I. (95%)
Estimated stature (cm)
C.I. (95%)
165 161 172 166 172
154.0–176.1 150.0–173.1 161.0–184.1 155.0–178.1 161.1–184.1
164 160 171 165 171
149.1–178.9 145.1–174.9 156.1–185.9 150.1–179.9 156.1–185.9
L. Marinho et al. / Forensic Science International 220 (2012) 292.e1–292.e4
In our autopsy sample, cadaver length ranged from 156 cm to 184 cm, with a mean (SD) of 167.9 (6.90) cm, and mean sternum length (SD) was 203.9 (14.21) mm, ranging from 171 mm to 235 mm. The intra-class correlation coefficient (ICC = 0.994, 95%CI = 0.975–0.999) shows an almost perfect agreement between the measurements of two researchers. When regressing cadaver length (CL) on sternum length (SL), the following simple linear regression formula was obtained: CL = 135.322 + 0.160 SL 13.32 (95% C.I.). The standard error of the estimate (SEE) obtained is 6.59 and the correlation coefficient (R) is 0.329. Only 11% of the variation in the sample is explained by this regression model (R2 = 0.11). 4. Discussion Since Dwight’s study [17], in the late 19th century, the sternum has received very little attention as a means to estimate stature from human skeletal remains. This study, together with that of Menezes et al.’s [18] and Singh et al.’s [19], has filled an important gap. Whereas the first two focused on dry bone observations, in this study we wished to provide an alternative means to the estimation of stature with measurements of fresh sterna, namely when maceration cannot be carried out, due to time constraints for example, or when a fresh or decomposed mutilated cadaver is being identified. Overall, our results show that the sternum is not particularly useful for stature estimation. Not only is the correlation between sternum length and stature low, but also the standard error of the estimate is relatively large, resulting in confidence intervals for stature which are also large and of limited use to discriminate between individuals of similar stature. A simple comparison with the results of other studies, where stature has been regressed on long bone lengths and other bone measurements, shows that the standard error of the estimate (as well as the confidence interval) in stature estimation formulae using the sternum is among the highest (Table 2). Studies shown in this table were chosen solely to illustrate the variation in standard errors obtained from stature regression formulae. In this table, the rank of increasing error for each regression formula is shown. Regression errors for the length of the sternum are almost twice of that for the length of the long bones and are similar to measurement of the fleshed hand and foot, and measurements of the sacrum and coccyx.
292.e3
Although in this study the standard error of the estimate obtained is smaller than that of Singh et al.’s [19], Menezes et al.’s [18] shows the smallest error. On the other hand, the correlation coefficient obtained by Menezes et al. [18] and Singh et al. [19] is greater than that of this study. Nonetheless, confidence intervals are still large and the correlation is always under 0.7. This indicates that for the same stature, individuals vary considerably in sternum length. The fact that we have used cadaver length and measured the fresh sternum is unlikely to have biased our results significantly, because any errors tend to be systematic. Cadaver length tends to be greater than true living stature and the length of the fresh sternum may also tend to be greater than that of the macerated sternum, due to the presence of any soft tissue and lack of shrinkage. However, this will not affect the relationship between cadaver length and the length of the sternum. In addition, our inter-observer test shows a high reliability for the data. Unfortunately, we could not preserve the sterna and assess the amount of shrinkage with maceration. Consequently, the formula obtained in this study cannot be applied to dry macerated sterna. An important consideration to make when analyzing and applying different stature estimation methods is that they are not universally applicable and several studies have demonstrated that a model developed from a specific population may not give as reliable estimates when applied to another [20,25]. This results mostly from differences in body proportions and differences in proportions between stature and bone size, that result from differences in environmental conditions during growth that affect these proportions. For example, stunted children and adults have proportionally shorter legs when compared to their normal counterparts [28]. As a consequence, individuals from developed nations may differ in body proportions relative to stature when compared to individuals from the developing countries. Unfortunately, due to our small sample we could not assertively conclude whether the Menezes et al.’s [18] and Singh et al.’s [19] formulae can accurately be applied to the Portuguese population. Our results suggest that these formulae may not tend to overestimate or underestimate true stature. However, the Lisbon collection sample of macerated sterna is comprised of individuals who lived and died in the late 19th and early 20th centuries in Portugal. In the late
Table 2 Stature estimation error (SEE – standard error of the estimate and CI – confidence intervals) reported in other studies and compared to that of the sternum in this study and other. Formulae in each study is ranked by increasing estimation error. Rank
SEE
CI (95%)
Trotter and Gleser [6]
Femur Humerus
1 7
3.27 4.05
6.41 7.94
Mendonc¸a [9]
Femur Humerus
4 8
3.50 4.25
6.96 8.44
Ross and Konigsberg [23]
Humerus Femur Tibia
6 2 3
4.03 3.30 3.39
7.90 6.47 6.64
Duyar and Pelin [24] Pelin et al. [16] Sanli et al. [25]
Tibia Sacrum (e.g.) Hand (length) Foot (length)
5 18 9 10
3.94 6.59 4.27 4.30
7.82 13.32 8.50 8.56
Duyar et al. [26]
Ulna Tibia
15 5
5.59 3.94
11.07 7.80
Celbis and Agritmis [10]
Radio Ulna
12 13
4.70 4.80
9.35 9.55
Nagesh and Kumar [15] Krishan and Sharma [27]
Vertebral column Hand (length, e.g.) Foot (length, e.g.)
11 14 17
4.38 5.22 6.19
8.73 10.34 12.26
Menezes et al. [18] Singh et al. [19] Present study
Sternum (length) Sternum (combined length) Sternum (length)
16 19 18
5.64 7.24 6.59
11.05 14.19 13.32
Bone
292.e4
L. Marinho et al. / Forensic Science International 220 (2012) 292.e1–292.e4
20th century, the country has experienced dramatic shifts and improvements in living conditions, so that modern Portuguese are much taller than their earlier counterparts [29,30]. Compared to modern Indians, represented in the Menezes et al. and Singh et al. samples [18,19], the Lisbon collection individuals may compare favorably, and hence the lack of significant errors in stature estimation using their formulae. Given that regression equations may not be population-specific per se, but specific to populations who share similar body proportions, it would be interesting to know whether a similar result will be obtained using samples of other bones from this collection on previously derived equations for Indian population groups. When applying regression formulae to estimate stature it is also important to consider what is the ‘‘stature’’ that is being determined. Most regression models are based on cadaver length or living standing height, but in most forensic scenarios, the antemortem information available about stature is usually reported stature, obtained from driver’s license, national citizenship identification documents or from relatives [31,32]. Although cadaver length, standing height and reported height are related, cadaver length tends to be greater than standing height, and possibly reported stature is the greatest of the three [31,33,34]. Consequently, it is crucial to distinguish between living, cadaver or forensic stature [35–37]. 5. Conclusion The study does not provide support for the use of the length of the sternum as a reliable estimator of stature in mutilated human skeletal remains, either skeletonized or fleshed. This is contrary to previous studies, which do not seem to stress enough the limitations of using sternum measurements for that purpose. Although if a quick estimate is required and maceration cannot be carried out, as the sternum is easily removed and dissected from a relatively fresh or badly decomposed body, then this may work as an alternative for statute estimation. However, particular attention should be given in the exclusion of cases based on the stature data, as there seems to be considerable variation in sternum length for the same stature. Acknowledgments We would like to thank Dr. Jorge Soares, former director of the Southern Delegation of the National Institute of Legal Medicine in Lisbon, for authorizing the access to the autopsy records of individuals in the Lisbon collection. Thank you to the anonymous reviewers for their remarks, which helped improving the manuscript. References [1] E. Cunha, C. Cattaneo, Forensic pathology and forensic anthropology: the state of art, in: A. Schmitt, E. Cunha, J. Pinheiro (Eds.), Forensic Anthropology and Medicine, Humana Press, Totowa, New Jersey, 2006, pp. 39–53. [2] L. Scheuer, Application of osteology to forensic medicine, Clin. Anat. 15 (4) (2002) 297–312 (review). [3] G. Fully, Une nouvelle me´thode de de´termination de la taille, Ann. Med. Legale 35 (1956) 266–273. [4] M.H. Raxter, B.M. Auerbach, C.B. Ruff, Revision of the fully technique for estimating statures, Am. J. Phys. Anthropol. 130 (3) (2006) 374–384.
[5] M. Trotter, G.C. Gleser, A re-evaluation of estimation of stature based on measurements of stature taken during life and of long bones after death, Am. J. Phys. Anthropol. 16 (1) (1958) 79–123. [6] M. Trotter, G.C. Gleser, Estimation of stature from long bones of American Whites and Negroes, Am. J. Phys. Anthropol. 10 (4) (1952) 463–514. [7] B.C. Didia, E.C. Nduka, O. Adele, Stature estimation formulae for Nigerians, J. Forensic Sci. 54 (1) (2009) 20–21. [8] A. Ozaslan, M.Y. Is¸can, I. Ozaslan, H. Tug˘cu, S. Koc¸, Estimation of stature from body parts, Forensic Sci. Int. 132 (1) (2003) 40–45. [9] M.C. Mendonc¸a, Estimation of height from the length of long bones in a Portuguese adult population, Am. J. Phys. Anthropol. 112 (1) (2000) 39–48. [10] O. Celbis, H. Agritmis, Estimation of stature and determination of sex from radial and ulnar bone lengths in a Turkish corpse sample, Forensic Sci. Int. 158 (2–3) (2006) 135–139. [11] C. Cattaneo, Forensic anthropology: developments of a classical discipline in the new millennium, Forensic Sci. Int. 165 (2–3) (2007) 185–193 (review). [12] S. Kalia, S.K. Shetty, K. Patil, Mahima, Stature estimation using odontometry and skull anthropometry, Indian J. Dent. Res. 19 (2) (2008) 150–154. [13] K. Krishan, R. Kumar, Determination of stature from cephalo-facial dimensions in a North Indian population, J. Leg. Med. 9 (2007) 128–133. [14] R.M. Burke, Can we estimate stature from the scapula? A test considering sex and ancestry, Thesis, BSc, University of Idaho (2005). [15] K.R. Nagesh, G.P. Kumar, Estimation of stature from vertebral column length in South Indians, Leg. Med. (Tokyo) 8 (5) (2006) 269–272. [16] C. Pelin, I. Duyar, E.M. Kayahan, R. Zag˘yapan, A.M. Ag˘ildere, A. Erar, Body height estimation based on dimensions of sacral and coccygeal vertebrae, J. Forensic Sci. 50 (2) (2005) 294–297. [17] T. Dwight, The sternum as an index of sex, age and height, J. Anat. 15 (1881) 327– 330. [18] R.G. Menezes, et al., Stature estimation from the length of the sternum in South Indian males: a preliminary study, J. Forensic Leg. Med. 16 (2009) 441–443. [19] J. Singh, et al., Skeletal height estimation from regression analysis of sterna lengths in a Northwest Indian population of Chandigarh region: a postmortem study, Forensic Sci. Int. 206 (2011) 211.e1–211.e8. [20] H.F.V. Cardoso, Brief communication: the collection of identified human skeletons housed at the Bocage Museum (National Museum of Natural History), Lisbon, Portugal, Am. J. Phys. Anthropol. 129 (2006) 173–176. [21] G.T. Ashley, The relationship between the pattern of ossification and the definitive shape of the mesosternum in man, J. Anat. 90 (1956) 87–105. [22] SPSS Inc., SPSS Version 17. 0 for Windows User’s Guide, SPSS Inc., Chicago, 2008. [23] A.H. Ross, L.W. Konigsberg, New formulae for estimating stature in the Balkans, J. Forensic Sci. 47 (1) (2002) 165–167. [24] I. Duyar, C. Pelin, Body height estimation based on tibia length in different stature groups, Am. J. Phys. Anthropol. 122 (1) (2003) 23–27. [25] S.G. Sanli, E.D. Kizilkanat, N. Boyan, E.T. Ozsahin, M.G. Bozkir, R. Soames, et al., Stature estimation based on hand length and foot length, Clin. Anat. 18 (8) (2005) 589–596. [26] I. Duyar, C. Pelin, R. Zagyapan, A new method of stature estimation for forensic anthropological application, Anthropol. Sci. 114 (2006) 23–27. [27] K. Krishan, A. Sharma, Estimation of stature from dimensions of hands and feet in a North Indian population, J. Forensic Leg. Med. 14 (6) (2007) 327–332. [28] B. Bogin, L. Rios, Rapid morphological change in living humans: implications for modern human origins, Comp. Biochem. Physiol. A: Mol. Integr. Physiol. 136 (1) (2003) 71–84. [29] C. Padez, Secular trend in stature in the Portuguese population (1904–2000), Ann. Hum. Biol. 30 (2003) 262–278. [30] H.F.V. Cardoso, Secular changes in body height and weight of Portuguese boys over one century, Am. J. Hum. Biol. 20 (2008) 270–277. [31] E. Giles, D.L. Hutchinson, Stature- and age-related bias in self-reported stature, J. Forensic Sci. 36 (1991) 765–780. [32] S. Ousley, Should we estimate biological or forensic stature? J. Forensic Sci. 40 (5) (1995) 768–773. [33] J.H. Himes, A.F. Roche, Reported versus measured adult statures, Am. J. Phys. Anthropol. 58 (1982) 335–341. [34] L.W. Konigsberg, S.M. Hens, L.M. Jantz, W.L. Jungers, Stature estimation and calibration: Bayesian and maximum likelihood perspectives in physical anthropology, Am. J. Phys. Anthropol. 27 (27) (1998) 65–92. [35] B.J. Adams, N.P. Herrmann, Estimation of living stature from selected anthropometric (soft tissue) measurements: applications for forensic anthropology, J. Forensic Sci. 54 (4) (2009) 753–760. [36] E. Giles, Corrections for age in estimating older adults’ stature from long bones, J. Forensic Sci. 36 (3) (1991) 898–901. [37] P. Willey, T. Falsetti, Inaccuracy of height information on driver’s licenses, J. Forensic Sci. 36 (3) (1991) 813–819.