femur subtrochanteric shape and ancestry assessment in ...schepartz2015). as such,...

1
References Femur Subtrochanteric Shape and Ancestry Assessment in Modern Japanese and Thai Individuals* Sean D. Tallman, Department of Anthropology, University of Tennessee, Knoxville, TN, [email protected] Introduction The paucity of cranial remains presents a significant problem to bioarchaeologists and forensic anthropologists in the determination of ancestry. While postcranial ancestry methods generally lack the accuracy attained from cranial methods, morphoscopic and morphometric approaches have been applied to the proximal, diaphyseal, and distal femur with varying success. Notably, femur subtrochanteric size and shape have been employed to differentiate between broad ancestral groups using anterior-posterior (A-P) and medial-lateral (M-L) sectioning lines (Gilbert and Gill 1990; Gill and Rhine 1990), in addition to the platymeric index (PI; Wescott 2005). The PI is calculated by dividing the A-P by M-L subtrochanteric diameter and multiplying by 100, with individuals exhibiting platymeric (medio-laterally broad; PI ≤ 84.9), eurymeric (rounded; PI 85.0 - 99.9), or stenomeric (antero- posteriorly broad; PI ≥ 100.0) subtrochanteric regions (Fig. 1). While the traditional sectioning line positions and PI thresholds are population dependent, they achieve success rates in excess of 80% when differentiating between Native Americans (generally platymeric) and pooled Afro- and Euro- Americans (generally eurymeric). Though it is often assumed that East Asian groups also exhibit similarly platymeric subtrochanteric morphologies, this assumption is based mainly on data derived from Native American archaeological assemblages. However, recent work has demonstrated that some East Asian groups are less platymeric than Native Americans (Tallman and Winburn 2015), while some Eastern European populations exhibit unexpectedly high rates of platymeria (McIlvaine and Schepartz 2015). As such, population-specific subtrochanteric ancestry methods should be used. Figure 1. Cross sections of platymeric (a), eurymeric (b), and stenomeric (c) femora (Tallman and Winburn 2015). Several methodological issues and inherent assumptions have been identified with the use of the proximal femur in ancestry assessment. In particular, the A-P and M-L subtrochanteric measurements generally exhibit high intra- and inter- observer error rates (Adams and Byrd 2002). Further, subtrochanteric methods assume: that there is minimal sexual dimorphism within populations; that populations are temporally and geographically homogenous; and that population differences in subtrochanteric form are attributable to genetic variation rather than biomechanical stress (Wescott and Srikanta 2008). While Wescott and Srikanta found that dimorphism, intrapopulation variance, and biomechanical stresses do affect subtrochanteric form, these factors do not significantly impede discriminating between Native American and Afro-/Euro- American individuals. The current study further explores femur subtrochanteric size and shape variability and its use in forensic ancestry assessment by testing Wescott’s (2005), Gill and Rhine’s (1990), and Tallman and Winburn’s (2015) femur subtrochanteric methods and associated assumptions on modern Japanese and Thai individuals aged 17-96 years. The Japanese sample consists of 287 known individuals from the greater Tokyo region who died during the late 19th/early 20th centuries (Chiba University) and those who died during the 1960s/1970s (Jikei University). The Thai sample is composed of 149 known individuals from northern Thailand who died in very recent decades (Khon Kaen University). A Euro-American sample of 77 males identified by the Department of Defense was used to help test the discriminatory power of subtrochanteric size and shape. To examine intra- and inter- observer error rates, 45 Chiba femora were re-measured; to examine the effects of body size, femur maximum length (FML) and maximum femoral head diameter (MHD) were measured for a subsample of 136 Thai individuals. Intraclass correlation coefficients, discriminant functions, descriptive statistics, t-tests, and ANOVAs were calculated in the SPSS statistical software package. Materials and Methods Sample N Mean PI SD PI Range Platymeric Eurymeric Stenomeric Japanese females 70 80.21 1,2 6.45 57.44-99.22 T: 80.0% 4 A: 90.0% T: 20.0% A: 10.0% T: 0% A: 0% Japanese males 217 82.16 1,3 7.43 55.65-103.72 T: 62.7% 4 A: 79.7% T: 35.9% A: 20.3% T: 1.4% A: 0% Thai females 45 79.71 1 7.14 64.29-93.55 T: 75.6% 4 A: 95.6% T: 24.4% A: 4.4% T: 0% A: 0% Thai males 104 83.51 1 7.13 64.71-110.71 T: 60.6% 4 A: 82.7% T: 36.5% A: 15.4% T: 2.9% A: 1.9% Results Table 3. Descriptive statistics for Japanese and Thai with traditional (T) and adjusted (A) PI definitions*. *T: Platymeric (PI ≤ 84.9), Eurymeric (85.0 ≤ PI ≤ 99.9), Stenomeric (PI ≥ 100.0) (Wescott 2005); A: Platymeric (PI ≤ 88.9), Eurymeric (89.0 ≤ PI ≤ 103.9), Stenomeric (PI ≥ 104.0) (Tallman and Winburn 2015). 1 Japanese females differ from Japanese males (p=0.038); Thai females differ from Thai males (p=0.004). 2 Japanese females who died during the early 19th/20th century (n=35) differ (more platymeric) from those who died during the 1960s/1970s (n=35; p=0.001). 3 Japanese males who died during the early 19th/20th century (n=134) differ (more platymeric) from those who died during the 1960s/1970s (n=83; p=0.001). 4 The Japanese and Thai females and males are less platymeric than Wescott’s (2005) Native American females (85%) and males (79%) using traditional PI thresholds. Test A-P M-L Intraobserver .989 .963 Interobserver .981 .957 Table 1. Intraclass correlation coefficients for 45 re-measured Chiba femora. Test Sample Sum of squares df Mean square F Sig. Between groups Japanese 623.170 7 89.024 1.716 .105 Thai 538.511 6 89.752 1.698 .126 Within groups Japanese 14471.572 279 51.869 Thai 7505.363 142 52.855 Total Japanese 15094.743 286 Thai 8043.874 148 Table 4. ANOVA: Age effects on PI for 8 Japanese groups and 7 Thai groups. Test Meas. Sum of squares df Mean square F Sig. Between groups FML 773.479 7 110.497 2.035 .055 MHD 544.093 3 181.364 3.335 .021* Within groups FML 6948.492 128 54.285 MDH 7177.878 132 54.378 Total FML 7721.971 135 MDH 7721.971 135 Table 5. ANOVA: FML and MHD effects on PI for 136 Japanese individuals. Japanese and Thai individuals display considerable variability in subtrochanteric size and shape. Subtrochanteric ancestry assessment methods derived from Native American data exhibit reduced discriminatory power when applied to these two modern Asian groups, therefore underscoring the importance for the ongoing development and testing of population-specific methods. As such, the amended sectioning lines and PI thresholds should be used over those derived from Native Americans when presented with remains potentially originating from East or Southeast Asia. Intra- and inter- observer variation introduces 1.2-4.3% error to measurements. Additionally, sex, time period, and MHD affect PI, while FML and age do not. This suggests that subtrochanteric form is not entirely under genetic control, but is also under the influence of intrapopulation variation, biomechanical stresses, secular change, and/or body size/proportions, potentially complicating its use in ancestry assessment. Despite these complications, this study demonstrates that population-specific PI thresholds and A-P and M-L sectioning lines are useful and appropriate in discriminating modern Japanese and Thai individuals from some non-Asian groups when cautiously applied. Discussion and Conclusions *Bonferonni test shows significant difference (.024) in mean PIs between MHD ≤ 40 mm and MHD 47 - 50 mm. Group N Mean PI PI AP and ML Correct Incorrect Correct Incorrect Japanese males 77 81.72 68.8% 31.2% 80.5% 19.5% Euro-American males 77 91.36 58.4% 41.6% 62.3% 37.7% Table 2. Cross-validated discriminant function classification rates for Japanese and Euro-American males. Adams BJ, Byrd JE. 2002. Interobserver variation of selected postcranial skeletal measurements. Journal of Forensic Sciences 47:1193-1202. Gilbert R, Gill GW. 1990. A metric technique for identifying American Indian femora. In: Gill GW, Rhine S, editors. Skeletal attribution of race: methods for forensic anthropology. Maxwell Museum of Anthropology Anthropological Papers No. 4. Albuquerque: University of New Mexico. p. 97-99. Gill GW, Rhine S. Appendix A. A metric technique for identifying American Indian femora, by Gilbert R, Gill GW. In: Gill GW, Rhine S, editors. Skeletal attribution of race: methods for forensic anthropology. Maxwell Museum of Anthropology Anthropological Papers No. 4. Albuquerque: University of New Mexico. p. 99. McIlvaine BK, Schepartz LA. 2015. Femoral subtrochanteric shape in Albania: implications for use in forensic applications. HOMO-Journal of Comparative Biology 66:79-89. Tallman SD, Winburn AP. 2015. Forensic applicability of femur subtrochanteric shape to ancestry assessment in Thai and White American males. Journal of Forensic Sciences (July 2015; in press). Wescott DJ. 2005. Population variation in femur subtrochanteric shape. Journal of Forensic Sciences 50:286-293. Wescott DJ, Srikanta D. 2008. Assessing ancestry using femur subtrochanteric shape revisited: testing the assumptions of the Gilbert and Gill method. HOMO-Journal of Comparative Human Biology 59:347-363. *Portions of this research were supported by the National Science Foundation (grant #1414742), the Japanese Society for the Promotion of Science, and the Joint POW/MIA Accounting Command’s Central Identification Laboratory. Acknowledgements: Allysha P. Winburn, Carrie A. Brown, Dr. Yoshiharu Matsuno, Dr. Yoshikatsu Negishi, Dr. Panya Tuamsuk, Kathryn E. Kulhavy Figure 2. Scatterplot of Japanese, Thai, and Euro-American A-P and M-L measurements with traditional (T) Gill and Rhine (1990) and amended (A) Tallman and Winburn (2015) sectioning lines. (T) (A)

Upload: others

Post on 04-Jul-2020

2 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Femur Subtrochanteric Shape and Ancestry Assessment in ...Schepartz2015). As such, population-specific subtrochanteric ancestry methods should be used. Figure 1. Cross sections of

References

FemurSubtrochantericShapeandAncestryAssessmentinModernJapaneseandThaiIndividuals*

SeanD.Tallman,DepartmentofAnthropology,UniversityofTennessee,Knoxville,TN,[email protected]

Introduction

Thepaucityofcranialremainspresentsasignificantproblemtobioarchaeologists andforensicanthropologistsinthedeterminationofancestry.Whilepostcranialancestrymethodsgenerallylacktheaccuracyattainedfromcranialmethods,morphoscopicandmorphometricapproacheshavebeenappliedtotheproximal,diaphyseal,anddistalfemurwithvaryingsuccess.Notably,femursubtrochantericsizeandshapehavebeenemployedtodifferentiatebetweenbroadancestralgroupsusinganterior-posterior(A-P)andmedial-lateral(M-L)sectioninglines(GilbertandGill1990;GillandRhine1990),inadditiontotheplatymericindex(PI;Wescott2005).ThePIiscalculatedbydividingtheA-PbyM-Lsubtrochantericdiameterandmultiplyingby100,withindividualsexhibitingplatymeric(medio-laterallybroad;PI≤84.9),eurymeric(rounded;PI85.0- 99.9),orstenomeric(antero-posteriorlybroad;PI≥100.0)subtrochantericregions(Fig.1).WhilethetraditionalsectioninglinepositionsandPIthresholdsarepopulationdependent,theyachievesuccessratesinexcessof80%whendifferentiatingbetweenNativeAmericans(generallyplatymeric)andpooledAfro- andEuro-Americans(generallyeurymeric).ThoughitisoftenassumedthatEastAsiangroupsalsoexhibitsimilarlyplatymericsubtrochantericmorphologies,thisassumptionisbasedmainlyondataderivedfromNativeAmericanarchaeologicalassemblages.However,recentworkhasdemonstratedthatsomeEastAsiangroupsarelessplatymericthanNativeAmericans(TallmanandWinburn2015),whilesomeEasternEuropeanpopulationsexhibitunexpectedlyhighratesofplatymeria(McIlvaine andSchepartz 2015).Assuch,population-specificsubtrochantericancestrymethodsshouldbeused.

Figure1.Crosssectionsofplatymeric(a),eurymeric(b),andstenomeric(c)femora(TallmanandWinburn2015).

Severalmethodologicalissuesandinherentassumptionshavebeenidentifiedwiththeuseoftheproximalfemurinancestryassessment.Inparticular,theA-PandM-Lsubtrochantericmeasurementsgenerallyexhibithighintra- andinter- observererrorrates(AdamsandByrd2002).Further,subtrochantericmethodsassume:thatthereisminimalsexualdimorphismwithinpopulations;thatpopulationsaretemporallyandgeographicallyhomogenous;andthatpopulationdifferencesinsubtrochantericformareattributabletogeneticvariationratherthanbiomechanicalstress(WescottandSrikanta 2008).WhileWescottandSrikanta foundthatdimorphism,intrapopulation variance,andbiomechanicalstressesdoaffectsubtrochantericform,thesefactorsdonotsignificantlyimpedediscriminatingbetweenNativeAmericanandAfro-/Euro- Americanindividuals.

ThecurrentstudyfurtherexploresfemursubtrochantericsizeandshapevariabilityanditsuseinforensicancestryassessmentbytestingWescott’s (2005),GillandRhine’s(1990),andTallmanandWinburn’s (2015)femursubtrochantericmethodsandassociatedassumptionsonmodernJapaneseandThaiindividualsaged17-96years.

TheJapanesesampleconsistsof 287 knownindividualsfromthegreaterTokyoregionwhodiedduringthelate19th/early20thcenturies(ChibaUniversity)andthosewhodiedduringthe1960s/1970s(JikeiUniversity).TheThaisampleiscomposedof149 knownindividualsfromnorthernThailandwhodiedinveryrecentdecades(KhonKaenUniversity).AEuro-Americansampleof77malesidentifiedbytheDepartmentofDefensewasusedtohelptestthediscriminatorypowerofsubtrochantericsizeandshape.Toexamineintra- andinter- observererrorrates,45Chibafemorawerere-measured;toexaminetheeffectsofbodysize,femurmaximumlength(FML)andmaximumfemoralheaddiameter(MHD)weremeasuredforasubsampleof136Thaiindividuals.Intraclasscorrelationcoefficients,discriminantfunctions,descriptivestatistics,t-tests,andANOVAswerecalculatedintheSPSSstatisticalsoftwarepackage.

MaterialsandMethods

Sample N MeanPI SD PIRange Platymeric Eurymeric Stenomeric

Japanesefemales 70 80.211,2 6.45 57.44-99.22 T:80.0%4

A: 90.0%T:20.0%A:10.0%

T: 0%A:0%

Japanesemales 217 82.161,3 7.43 55.65-103.72 T: 62.7%4

A:79.7%T:35.9%A:20.3%

T:1.4%A:0%

Thai females 45 79.711 7.14 64.29-93.55 T:75.6%4

A: 95.6%T:24.4%A: 4.4%

T:0%A: 0%

Thai males 104 83.511 7.13 64.71-110.71 T: 60.6%4

A:82.7%T:36.5%A:15.4%

T:2.9%A: 1.9%

Results

Table3. DescriptivestatisticsforJapaneseandThaiwithtraditional(T)andadjusted(A)PIdefinitions*.

*T:Platymeric(PI≤84.9),Eurymeric(85.0≤PI≤99.9),Stenomeric(PI≥100.0) (Wescott2005);A:Platymeric(PI≤88.9),Eurymeric(89.0≤PI≤103.9),Stenomeric(PI≥104.0)(TallmanandWinburn2015).1Japanese femalesdifferfromJapanesemales (p=0.038);Thaifemales differfromThaimales (p=0.004).2 Japanesefemaleswhodiedduringtheearly19th/20thcentury(n=35)differ(moreplatymeric)fromthosewhodiedduringthe1960s/1970s(n=35;p=0.001).

3 Japanesemaleswhodiedduringtheearly19th/20thcentury(n=134)differ(moreplatymeric)fromthosewhodiedduringthe1960s/1970s(n=83;p=0.001).

4TheJapaneseandThaifemalesandmalesarelessplatymericthanWescott’s (2005)NativeAmericanfemales(85%)andmales(79%)usingtraditionalPIthresholds.

Test A-P M-L

Intraobserver .989 .963

Interobserver .981 .957

Table1. Intraclass correlationcoefficientsfor45re-measuredChibafemora.

Test Sample Sum ofsquares df Mean

square F Sig.

Between groupsJapanese 623.170 7 89.024 1.716 .105Thai 538.511 6 89.752 1.698 .126

WithingroupsJapanese 14471.572 279 51.869Thai 7505.363 142 52.855

TotalJapanese 15094.743 286Thai 8043.874 148

Table4. ANOVA:AgeeffectsonPIfor8Japanesegroupsand7Thaigroups.

Test Meas. Sum ofsquares df Mean

square F Sig.

BetweengroupsFML 773.479 7 110.497 2.035 .055MHD 544.093 3 181.364 3.335 .021*

WithingroupsFML 6948.492 128 54.285MDH 7177.878 132 54.378

TotalFML 7721.971 135MDH 7721.971 135

Table5. ANOVA:FMLandMHDeffectsonPIfor136Japaneseindividuals.

JapaneseandThaiindividualsdisplayconsiderablevariabilityinsubtrochantericsizeandshape.SubtrochantericancestryassessmentmethodsderivedfromNativeAmericandataexhibitreduceddiscriminatorypowerwhenappliedtothesetwomodernAsiangroups,thereforeunderscoringtheimportancefortheongoingdevelopmentandtestingofpopulation-specificmethods.Assuch,theamendedsectioninglinesandPIthresholdsshouldbeusedoverthosederivedfromNativeAmericanswhenpresentedwithremainspotentiallyoriginatingfromEastorSoutheastAsia.

Intra- andinter- observervariationintroduces1.2-4.3%errortomeasurements.Additionally,sex,timeperiod,andMHDaffectPI,whileFMLandagedonot.Thissuggeststhatsubtrochantericformisnotentirelyundergeneticcontrol,butisalsoundertheinfluenceofintrapopulation variation,biomechanicalstresses,secularchange,and/orbodysize/proportions,potentiallycomplicatingitsuseinancestryassessment.Despitethesecomplications,thisstudydemonstratesthatpopulation-specificPIthresholdsandA-PandM-LsectioninglinesareusefulandappropriateindiscriminatingmodernJapaneseandThaiindividualsfromsomenon-Asiangroupswhencautiouslyapplied.

DiscussionandConclusions

*Bonferonni testshowssignificantdifference(.024)inmeanPIsbetweenMHD≤40mmandMHD47- 50mm.

Group N MeanPIPI AP andML

Correct Incorrect Correct Incorrect

Japanese males 77 81.72 68.8% 31.2% 80.5% 19.5%

Euro-Americanmales 77 91.36 58.4% 41.6% 62.3% 37.7%

Table2. Cross-validateddiscriminantfunctionclassificationratesforJapaneseandEuro-Americanmales.

AdamsBJ,ByrdJE.2002.Interobserver variationofselectedpostcranialskeletalmeasurements.JournalofForensicSciences47:1193-1202.

GilbertR,GillGW.1990.AmetrictechniqueforidentifyingAmericanIndianfemora.In:GillGW,RhineS,editors.Skeletalattributionofrace:methodsforforensicanthropology.MaxwellMuseumofAnthropologyAnthropologicalPapersNo.4.Albuquerque:UniversityofNewMexico.p.97-99.

GillGW,RhineS.AppendixA.AmetrictechniqueforidentifyingAmericanIndianfemora,byGilbertR,GillGW.In:GillGW,RhineS,editors.Skeletalattributionofrace:methodsforforensicanthropology.MaxwellMuseumofAnthropologyAnthropologicalPapersNo.4.Albuquerque:UniversityofNewMexico.p.99.

McIlvaine BK,Schepartz LA.2015.FemoralsubtrochantericshapeinAlbania:implicationsforuseinforensicapplications.HOMO-JournalofComparativeBiology66:79-89.

TallmanSD,WinburnAP.2015.ForensicapplicabilityoffemursubtrochantericshapetoancestryassessmentinThaiandWhiteAmericanmales.JournalofForensicSciences(July2015;inpress).

WescottDJ.2005.Populationvariationinfemursubtrochantericshape.JournalofForensicSciences50:286-293.

WescottDJ,Srikanta D.2008.Assessingancestryusingfemursubtrochantericshaperevisited:testingtheassumptionsoftheGilbertandGillmethod.HOMO-JournalofComparativeHumanBiology59:347-363.

*PortionsofthisresearchweresupportedbytheNationalScienceFoundation(grant#1414742),theJapaneseSocietyforthePromotionofScience,andtheJointPOW/MIAAccountingCommand’sCentralIdentificationLaboratory.

Acknowledgements:AllyshaP.Winburn,CarrieA.Brown,Dr.YoshiharuMatsuno,Dr.Yoshikatsu Negishi,Dr.Panya Tuamsuk,KathrynE.Kulhavy

Figure2. ScatterplotofJapanese,Thai,andEuro-AmericanA-PandM-Lmeasurementswithtraditional(T)GillandRhine(1990)andamended(A)TallmanandWinburn(2015)sectioninglines.

(T) (A)