Skeletal evidence of osteoarthritis: a palaeopathological

Annals of the Rheumatic Diseases 1995; 54: 443-450

REVIEW

Skeletal evidence of osteoarthritis:a palaeopathological perspective

Robert D Jurmain, Lynn Kilgore

Osteoarthritis (OA), or degenerative jointdisease, is the most common form of jointpathology, and it has an extensive history.Indeed, it has been observed in a variety offossil animals, and has been reported in thefossil remains of dinosaurs.'Among preindustrial human groups, as

documented by skeletal remains, osteoarthriticchanges are quite frequently the most commonpathological lesion seen. Such degenerativechanges have been described for archaeologicalsamples varying widely in time and location.2"In many cases,"'5 frequencies of involvementhave also been calculated and presented.

Skeletal evidence from archaeological sitesthus has potential to expand considerably therange of epidemiological data concerning OA.Human populations long since extinct can besampled. Current epidemiological data arelimited almost entirely to clinical samplesdrawn from urban contexts, mostly from NorthAmerica and Europe. Earlier populations,ranging from arctic hunters, to early Romano-British agriculturists, to Mediaeval Nubians(and, obviously, many others) may have ledlifestyles very different from those character-istic of contemporary groups. Accordingly,skeletal analysis provides a 'window' into avastly broader sample of human groups thanafforded solely by clinical samples. If varyinglevels of mechanical loading are important inthe aetiology of OA, as suggested by the 'stresshypothesis','6 earlier populations could beexpected to exhibit definitive patterns of jointinvolvement. Such patterns should be reflectedin differential involvement both within andamong joints; moreover, patterns of bilateralasymmetries, sex based variation, and perhaps,most crucially, clearly demarcated frequenciesof involvement among prehistoric populationsshould be manifest.

Available materialsIn recognition of the fact the osteoarthriticprocess begins in articular cartilage, analysis ofthe 'hard' tissues can provide another windownot commonly available to the clinician. Withjoint surfaces free of overlying tissue, subtledegenerative changes involving bone are easilyseen. Presumably slight arthritic involvementincludes small marginal lipping (osteophytes)or small erosive changes to the joint surfaceitself. In modern contexts, such minor boneinvolvement is not usually recognised as it does

not present radiographically; occasionally,following surgical replacement, a few macer-ated specimens have become available,'7 18thereby giving researchers a rare glimpse ofunderlying bony tissue.

In addition, materials derived at postmortemare a most useful supplement and help bridgethe gulf between clinical data and fully macer-ated remains. Several detailed studies have, infact, investigated the frequency of OA inpostmortem collections.'9-21 Moreover, in a fewcases large numbers of macerated specimenshave carefully been obtained from dissectingroom samples and have been curated aspermanent collections (most notably, the ToddCollection at the Cleveland Museum ofNatural History and the Terry Collection at theUS National Museum of Natural History).These collections have proven a rich source ofinformation, as the skeletons are virtuallyintact, sex and age (approximated) wererecorded at the time of dissection, and formany individuals cause of death (and someclinical history) is also noted. Building uponthis rich data source, skeletal biologists2 9 20have used dissecting room skeletal collectionsto refine their descriptions of osseous degener-ative lesions, and to make epidemiologicalcomparisons with other groups.For archaeologically derived specimens

there are tens of thousands of examples ofremodelled joint surfaces. Moreover, whenskeletal materials are well preserved, numerousregions can be systematically observed, in-cluding the large peripheral joints, the vertebralcolumn (body surfaces in addition toapophyseal articulations), and the numeroussmall articulations of the hands and feet. Inthis way, a pattern of involvement withinindividuals and, more generally, within popu-lations can be established.

Finally, in addition to a palaeoepidemiologyof degenerative arthritis in human groups, suchcomparative perspectives can be expanded toinclude non-human animals. Recent work withnon-human primates,22 23 especially the studyof joint involvement in great apes,2426 hasopened the potential for an even broaderperspective concerning the aetiology of osteo-arthritis. By studying animals with differingpostural/locomotory adaptations and corre-lating these factors with the pattern of degener-ative lesions, we should be able to gain a fullerunderstanding of some of the mechanicalfactors influencing the aetiology of OA.

Department ofAnthropology,San Jose StateUniversity,San Jose,CA 95192, USAR D JurmainDepartment ofAnthropology,Colorado StateUniversity,Fort Collins,CO, USAL KilgoreCorrespondence to:R D Jurmain.Accepted for publication13 October 1994

443

on April 10, 2022 by guest. P

rotected by copyright.http://ard.bm

j.com/

Ann R

heum D

is: first published as 10.1136/ard.54.6.443 on 1 June 1995. Dow

nloaded from

http://ard.bmj.com/

44urmain, Kilgore

Methodological approachesDegenerative lesions of joints are most

commonly evaluated in skeletal materials usingordinal scaling (most frequently, as none/slight; moderate; severe). Criteria usuallyinclude both marginal changes (osteophyticdevelopment) and articular surface changes(pitting, ebumation, or both). In fact, many

researchers27 28 argue that the most reliablediagnostic criteria for osteological determi-nation is a combination of marginal changesaccompanied by articular surface alterations.Additionally, in some studies20 periarticularareas of tendinous insertions (for example thegreater and lesser tubercules of the humerus)have also been evaluated.

Finally, evaluation of radiographic alter-ations, most importantly subchondral re-

modelling, has also been attempted; however,few systematic results have emerged here, so

Figure 1 Stages ofosteoarthritis as shown in the knee.A: Slight involvement with small marginal osteophyte.B: Moderate involvement with larger marginal osteophyte.C: Severe involvement, showing large osteophyte andeburnated, grooved articular surface.

that, with few exceptions,28 radiographic studyof skeletal OA has not yet become routine.Figure 1 shows an example of the scoring ofthestages of OA of the knee.

PERIPHERAL INVOLVEMENT

Evaluation of the large peripheral joints(particularly the shoulder, elbow, hip, andknee) by osteologists has been undertakenmore commonly and systematically than forother regions of the body. Most commonly, thedegree of OA involvement has been reportedas a single value for an entire joint (for examplethe shoulder). This observed level of involve-ment is derived from assessment of theindividual articulations (for example glenoidof scapula, head of humerus) and usuallyreported as the greater (more severe)score.'3 14 30

In a few, more detailed, investigations,individual areas within joints have been scoredseparately.9 11"31 For example, in a study ofdissecting room samples of Black and WhiteAmericans (Terry Collection) compared withtwo archaeological samples,29 numerous specificareas within joints were evaluated: sevenvariables for the shoulder; 16 for the elbow; 10for the hip; and 17 variables in the knee. Foreach side of the body, therefore, a total of 50variables were evaluated, yielding a maximumof 100 variables per individual. As an example,the table shows a list of these regionalindicators of skeletal OA for the elbow joint.

Scoring the osteoarthritic changes in such aspecific fashion allows a more comprehensiveinterpretation of arthritic involvement withinparticular joints. For example, do marginalchanges occur independently of articular sur-face degeneration? Do functional componentswithin joints (for example patello-femoral vfemoro-tibial) act somewhat independently?Does age interact differentially as a factorwithin various joint segments?

Another advantage of using multiple indi-cators per joint (as opposed to a single averagevalue) is that the values can be more easilymanipulated statistically. Ordinal measures,such as those used almost universally forskeletal assessment of OA, are not parametric.Thus most of the more robust types ofstatistical analyses cannot, strictly, be applied.However, as the number of individual variables

Specific areas scoredfor degenerative joint involvement inthe elbow

Variable Ordinalsconngrange

Distal humerus, trochlea, medial margin 0-2Distal humerus, trochlear ridge 0-2Distal humerus, capitulum, lateral margin 0-2Distal humerus, olecranon fossa 0-2Distal humerus, coronoid fossa 0-2Distal humerus, trochlea, articular surface 0-3Distal humerus, capitulum, articular surface 0-4Proximal ulna, coronoid process, margin 0-2Proximal ulna, olecranon process, margin 0-3Proximal ulna, radial facet, margin 0-1Proximal ulna, coronoid process, articular surface 0-2Proximal ulna, olecranon process, articular surface 0-2Proximal ulna, radial facet, articular surface 0-2Proximal radius, head, superior articular surface 0-4Proximal radius, head, inferior margin 0-2Proximal radius, head, lateral articular surface 0-2

444



j.com/

Ann R

heum D


nloaded from

http://ard.bmj.com/

Skeletal evidence of osteoarthritis: a palaeopathological perspective

increases (giving a much wider dispersion ofvalues), the total joint score approaches theconditions of parametric normality. Accord-ingly, analyses such as analysis of variancecomparing degree of degenerative jointinvolvement with bone density32 can befacilitated.

VERTEBRAL INVOLVEMENT

Within the vertebral column the fibrocarti-lagenous joints of adjoining vertebral bodiesare usually scored exclusively on the basis ofmarginal osteophytes (and the lesion is thuscalled vertebral osteophytosis or spondylosis).Figure 2 shows comparative stages of vertebralosteophytosis. Although these fibrocarti-lagenous joints are not synovial, the processesof degenerative arthritis that occur here aresuperficially similar to those seen in truesyiovial joints. Owing to these similarities ingross involvement, in some palaeopathologicalreports there is confusion relating toterminology. In this paper we follow theclinical consensus and reserve the term 'osteo-arthritis' for synovial joint involvement, whilevertebral body involvement is referred toseparately as 'vertebral osteophytosis.'Most commonly, in evaluating vertebral

osteophytosis the entire body suface is givena single score,13 33 but in some studies themargin was subdivided into more specificregions (anterior, anterolateral, posterior).One researcher34 has attempted to calculate

metrically the degree of osteophyte develop-ment.For apophyseal OA, the nature of degener-

ative changes tends to be very subtle, mani-fested as a slight thickening around thejoint margin or sharp edged pitting of thearticular surface. Until now, few systematicattempts have been made to standardise themethodology.

Limitations ofthe osteological approachAGE DETERMINATIONThe lack of precision in determining age atdeath for human skeletal remains imposes amajor limitation on osteological analyses ofOA. As this disease is clearly age correlated,comparisons among different groups (orbetween males and females within groups)cannot be accomplished without accuratedetermination of the age structure of therespective populations. Here lies the difficulty.Even though attempts have been made tocharacterise systematically the progressive agechanges in human skeletons,35-37 for adultskeletons the deteminations are still only roughapproximations. The most reliable age indi-cator involves progressive remodelling of thepubic symphysis, but as a well documentedcontemporary sample from the Los AngelesCounty Coroner's Office has shown,37 the ageestimates from the pubic symphysis can beestablished only within broad age brackets(± 10 years, or more); and, even more

Figure 2 Stages of vertebral osteophytosis. A: Slight. B: Moderate. C: Severe. D: Ankylosis.

445



j.com/

Ann R

heum D


nloaded from

http://ard.bmj.com/

46urmain, Kilgore

disturbing, as age increases, the estimatesbecome even less precise.

Other age approximations of older adults,such as calibration of cranial suture closure,have not proved as accurate as changes at theos pubis, and thus are not routinely used.However, some recent innovations, includingevaluation of remodelling of the auricularsurface of the ilium38 or sternal ends of theribs39 have proved useful, at least for someskeletal groups.One approach that has sought to control for

the lack of precise age control in archaeologicalmaterials is to utilise contemporary maceratedsamples. As mentioned above, such dissectingroom samples as the Todd and Terrycollections have been regularly used by skeletalbiologists, including studies of degenerativejoint disease.9 40

CLINICAL CORRELATIONAnother obvious limitation in skeletal researchon archaeologically derived materials is the lackof clinical records for any of the individuals.Without information regarding symptoma-tology the assumption has been that at leastsome of the osteological changes observed(osteophytes or articular surface remodelling)relate to pain or other symptoms during life.For OA of the peripheral joints, the moderateand severe changes regularly scored by osteo-logists probably do relate to a degree of im-paired function.4' Some, although as yet smallscale, efforts have begun to assemble materialsto correlate clinical records with hard tissuealterations. These collections, derived as surgi-cally removed macerated specimens,'7 18probably provide the best means to address thisconstraint.

STANDARDISATION OF SCORINGAnother problem that must still be resolved byskeletal biologists doing research on degener-ative joint disease is the current lack of system-atic scoring criteria. While all researchers usesome ordinal scaling system and look atbasically equivalent bony changes (osteo-phytes, porosity, eburnation of surface), littlehas been agreed upon in terms of standardis-ation of technique. Certainly, some greaterdegree of standardisation is desirable. At present,researchers are reasonably confident that, atleast for moderate and severe changes (whichare easily recognisable), reasonably compar-able criteria are generally being applied.

HISTORICAL DOCUMENTATIONAnother unavoidable, yet major, difficulty thatarises with archaeological material is the poorquality of general ethnographic informationfor the groups represented. Certainly, ethno-historical data can be of some use-forexample, activities of Inuit arctic hunters" ormediaeval Nubian agriculturists.'3 However,even here, the degree of precision relating tospecific types of behaviours practised by thesepeoples is very superficial.

Even more limiting, we have almost no ideaof the actual intensity of the significantmechanical loadings involved, or theirduration, in the activities of early populations.Some modern data suggest that duration ofmechanical stress may be crucial-forexample, the observation that degenerativeelbow involvement appears only after three to10 years of use of pneumatic tools.42

BEHAVIOURAL INTERPRETATIONS

From comparisons among various ancientpopulations of the frequency and severity ofosteoarthritic changes, skeletal biologists haveroutinely sought to draw conclusions relatingto presumed behavioural factors. However,given the only very superficial nature of thebehavioural data which are available, suchconclusions must be regarded as extremelytenuous. Indeed, extreme caution must beused in evaluating much of the osteologicalliterature on OA, including early publicationsby the authors.9 10 These investigators beginwith the assumption that culturally patternedmechanical stress leads to the onset of OA;moreover, the differences in frequency andseverity of involvement within groups (males vfemales) are explained by such differences inbehaviour, as are most of the differencesbetween groups.

Following along these logical lines, theseinvestigators then peruse the limited ethno-historical documentation in order to isolatethose presumed activities that initiated thedegenerative disease (such as rowing amongthe Inuit or use of the hoe in Nubians). It isclear that such reasoning is largely circular, butit nevertheless continues to persist in theliterature.43 Certainly, on the basis of currently-available data, the conclusions are virtuallyuntestable.The wisdom that extreme caution should be

exercised in not too quickly making func-tionally based explanations ofOA is attested toin the clinical literature. Some, largelyanecdotal, data have suggested some links withactivity, for example in the ankles of footballplayers44 45 or ballet dancers,45 the elbow ofsome pneumatic tool users,46 or spinal diseasein coal miners and longshoremen,46 47 yet otherstudies have found no such conclusiveevidence relating strenuous activity to anyincrease in OA in the lower limb of physicaleducation teachers,48 the elbow of otherpneumatic tool users,49 or in the leg ofparachutists.48Given the quite mixed results of contem-

porary epidemiological attempts to establish afirm correlation between specific activities andincreased incidence of OA, any such attemptsextended to the much less well documentedarchaeological past represent an extremelyhazardous intellectual venture.

Analyses ofdegenerative disease inhumn skeletonsAnthropologists, anatomists, and others havelong been interested in the patterning of

446



j.com/

Ann R

heum D


nloaded from

http://ard.bmj.com/


degenerative disease as seen in fully maceratedor partially mummified human remains. Asearly as 1911, Ruffer and Rietti50 describedseveral examples of arthritic disease in ancientEgyptians. For many years much of the focuswas centred on vertebral lesions, especiallyvertebral osteophytosis.2 3 33 In the past fewdecades, however, attention has also beenshifted to describing the distribution of lesionsin the large peripheral joints.8 9 1113 14 Inaddition, some researchers have also reportedon the temporomandibular joint. 13 30 51 Todate, however (with few exceptions30 33 52) littlesystematic analysis of the small joints of thehands and feet has been accomplished.

Osteoarthritis ofthe peripheral skeletonA basic approach applied to the interpretationof skeletal series borrows much from contem-porary epidemiological methodology. Such anapproach, when utilised for archaeologicallyderived materials, is usually termed 'palaeo-epidemiology.' As in epidemiological surveys,the most useful data are derived from stratifiedsamples-that is, those well controlled for bothage and sex. Figure 3, for example, shows thefrequencies of elbow OA in three archaeo-logical populations (Alaskan Inuit, PuebloIndians, and Central California Indians). Ascan be seen, Inuit have the earliest onset andby far the greatest incidence of involvement.The extreme pattern displayed by Inuit, whichis particularly characterised by erosion andultimate destruction of the radiohumeralcomponent, is quite unique.10 In general, inmany (but not all) prehistoric populations, theincidence of elbow OA is greater than forcontemporary groups; in contrast, thefrequency of hip involvement is much lower.The explanation for the differences in hip

involvement in prehistoric groups comparedwith recent ones may lie in the relative agedistributions. Hip disease is strongly correlatedwith advancing age. As prehistoric populationsexperienced a considerably shorter lifeexpectancy, they would thus be expected to

L

L

R

31 -40

RL R

41+

Age (years)

Figure 3 Incidence of osteoarthritis of the left (L) and right (R) elbows as shown in threearchaeological populations: Ala-329from Central California (Ol); Pecos Pueblofrom theSouthwestern US ( ); andAlaskan Inuit (0 ). Note the earlier onset and muchincreased incidence among the Inuit. For each population, * represents severe disease andthe remainder ofany column represents moderate disease.

show less hip involvement than is typically seentoday. Moreover, as discussed below, theelbow may be more vulnerable to mechanicalloading than the hip (which appears to be moreprone to systematic factors). It thus becomesvery tempting to try to correlate such patternsof OA with specific activities. For example,among Inuit highly intensified impulse loadingassociated with hunting (rowing, harpooning,etc.) might explain their unusually frequentand severe expression of elbow OA. Asirresistible as such behavioural hypotheses maybe, it must be recognised that anthropologistsusually lack the precise data to make anythingbut very general (and usually untestable)behaviour based hypotheses.What is clear, however, is that the pattern of

degenerative disease within the elbow conformswell with a functional-mechanical aetiology ofOA. Those areas presumably most subjected tomechanical loading show the greatest degree ofvariation. In addition, age appears to playconsiderably less of a role than is true for theshoulder. Inuit-as shown in figure 3-have avery high degree of elbow disease: more thanany other archaeological population sampledto date. Inuit are known ethnohistorically tohave experienced unusually high levels ofmechanical loading associated with arctichunting activities.53 We might thus reasonablyinfer that the frequent and severe levels ofelbow OA result directly from this severemechanical stress. However, the aetiologicalpicture, as explicated by skeletal data, is notthis unambiguous.

Mediaeval Sudanese Nubian groups are alsothought to have engaged in a rigorous life style.However, they have considerably less elbowOA than is seen among Inuit but, conversely,show proportionately greater knee involve-ment.'3 Indeed, in a general overview of thepalaeoepidemiology of OA in the Americas,Bridges54 could find little evidence ofconsistent patterns. Some populations showgreater degrees of involvement in the knee(analogous to contemporary patterns), whileothers show patterns more similar to that ofgreater elbow involvement in Inuit. Onegeneral pattern does emerge from the archaeo-logical data-the shoulder and hip are usuallyless severely involved than the knee or elbow.Given the overall lack of systematic

patterning in OA involvement among pre-historic groups, it would be unwise to drawunduly broad aetiological conclusions fromthese data. Certainly, as discussed above, evenmore hazardous are attempts to relate specificbehaviours in prehistory to the onset ofarthritic disease.

Degenerative joint disease ofthe spineNumerous osteological studies have investi-gated the incidence of degenerative jointdisease of the spine, particularly the manifes-tation referred to as vertebral osteophytosis orspondylosis of the fibrocartilagenous joints ofvertebral body surfaces.'3 30 33 34 55 56 Lessattention has been directed at involvement ofthe true synovial articulations of the dorsal

0

G1)Vc

-a

._

c

21 -30

447



j.com/

Ann R

heum D


nloaded from

http://ard.bmj.com/

_7urmain, Kilgore448

apophyseal facets. I13 30 3356 From these studiesit appears that the pattern of involve-ment of spinal arthropathy, particularlyvertebral osteophytosis, is more uniformbetween populations than that of peripheralinvolvement.

Vertebral osteophytosis is most common inthose motion segments located at the peaks ofthe cervical and thoracic curves, in addition tothroughout the lumbar area (fig 4). Thispattern of involvement has been reported fornumerous populations, including skeletalcollections representing various populations ofunspecified origin housed at the University ofWitwatersrand and Cambridge University,41lUnited States WVhites and Blacks,44 ancientEgyptians,"5 Mediaeval Eastern Europeans,34Sadlermiut Inuit,33 Sudanese Nubians, l 3Southeast US Indians,'7- and Central CaliforniaIndians.3Although arthritis of the dorsal vertebral

facets has been reported upon less frequentlythan spondylosis, similarities of expression alsoemerge from between group comparisons.These similarities include peaks of involvementin mid cervical and upper to mid thoracicsegments and a gradual increase in incidencebeginning in the lower thoracic region(TIO-11) and continuing throughout thelumbar area. I13 33 34 41 Rather than attemptingto explain spinal arthropathies as the result ofspecific activities, a more valid approach is toexamine the pattern of lesions in terms ofwhere they occur and to understand thegeneral types of movement most likely to occurin particular motion segments. 13 5' Thus whilelumbar involvement probably results largelyfrom compression caused bV weight bearing,lesions in the cervical area are most likelycaused by movement of all types.

It is evident from existing studies thatperipheral articulations show a wider range ofvariation in incidence than do spinal joints.Although there are a number of limitationsinvolved in interpopulation comparisons, this

pattern does emerge. We suggest that theobserved variation, especially that exhibitedby the elbow and knee, may reflect differencesin culturally based activities. However, thesimilarities in incidence, particularly thoseobserved in the spine, occur as a function ofspecies specific biomechanical parametersimposed by weight bearing and locomotion inan upright posture.

Osteoarthritis in non-human primatesComparing the distribution of osteoarthritis invaried human populations provides a moreinclusive perspective than concentratingstrictly on contemporary groups. Even broaderin its potential implications is the comparativestudy of non-human animals, most especiallythe non-human primates.

Degenerative lesions, as they are distributedin animals with different locomotory adap-tations, can give insight into mechanical factorsin the aetiopathogenesis of OA. Such aperspective has only recently been put tosystematic use.21 2 One recent study59surveyed 267 prosimians and 1250 anthropoidskeletons and found very low frequencies ofOA. Among captive animals, the rates ofinvolvement were quite low (4 8% ofprosimians and 3-7 of anthropoids) and evenlower in free ranging animals (0-8% amongprosimians and 0-9%/ of anthropoids).Moreover, there was a difference in pattern:elbow and hip lesions were seen only in captiveanimals, but knee involvement was greater infree ranging individuals. Similar results wereseen in our own study (in preparation) andanother 24 of a much smaller (but very welldocumented) series of free ranging chimpan-zees from Gombe National Park. In this group,vertebral osteophytosis was also investigated,and no involvement whatsoever was seen in483 vertebral body surfaces examined. Furtherreinforcing the total lack of vertebral osteo-phytosis was the observation that osteophytes

80

70

60-

::g 50 -'

a)

c: 40

-30

20

1 0 A.4i

rl 1U- 7

-iC2- C3- C4- C5- C6- C7- Tl- T2- T3- T4- T5- T6- T7- T8- T9- T10-Tll-Tl2- Ll- L2- L3- L4- L5-C3 C4 C5 C6 C7 Tl T2 T3 T4 T5 T6 T7 T8 T9 T10 Tll T12 Ll L2 L3 L4 L5 S1

Figure 4 Incidence of vertebral osteophytosis ( 0) involvingo vertebral bodies compared with i'ncidence of osteoarthriti's (O)involving the dorsalfacet (apophyseal) joints (data from Mediaeval Nubian sample).



j.com/

Ann R

heum D


nloaded from

http://ard.bmj.com/


did not occur even in individuals known to bequite old (in one case, an older female >40years of age).Another comprehensive study of three great

ape species (free ranging chimpanzees, gorillas,and orangutans)23 also found low rates ofdegenerative involvement in the peripheralskeleton and spine. The one exception wasobserved in mountain gorillas, in which theincidence of vertebral osteophytosis approached20%. Finally, in a comparative pongid samplealso from free ranging contexts, we foundless than 1% vertebral involvement (>2500surfaces examined) (paper in preparation). Aswith the study noted above, gorillas (lowlandin this case) had a greater frequency ofvertebral osteophytosis (3-6% ofvertebral bodysurfaces involved, and affecting close to 35%of individuals).The explanation for the higher degree of

involvement among gorillas compared withchimpanzees is thought to relate to the muchgreater body weight of the former specieS,23-up to four to five times that of chimpanzees.The differences in frequency of involvement incaptive versus free ranging primates noted byRothschild and Woods59 is also intriguing.Quite possibly, the variable patterns may beexplained by differences in substrate surface oractivity patterns.Of even greater aetiological interest is

comparison of the non-human primate datawith the human pattern. In all cases, both forthe spine and peripheral joints, humansconsistently have markedly greater rates ofdegenerative joint involvement. For the spine,especially, the increased loading concomitantwith bipedal locomotion may well provide theaetiological explanation. Certainly, vertebralosteophytosis is a ubiquitous human condition,seen in all cultural groups, and it is nearlypervasive after age 40. Its almost completeabsence in chimpanzees, and comparativerarity, even in the gorilla, argue for somefundamentally altered mechanical demands inthe human species.Always a concern, however, is the need to

control for the influence of age. In most of thenon-human primate studies noted above, thematerials were obtained as 'wild shot'collections, and age was only very roughlyestimated from often quite misleadingobservations made in the field ('young adult','old adult'). With no good criteria for agingadult skeletons of non-human primates, thislack of age control persists as a majorlimitation.Another possible confounding issue is the

pattern of degenerative joint disease seen inOld World monkeys. Osteological study ofbaboons60 and radiographic survey ofmacaques23 have indicated a high rate ofvertebral osteophytosis in these animals. Asboth are quadrupeds, it is difficult to explainwhy monkeys would have more spinalinvolvement than great apes (which aremodified, larger quadrupeds). One speculationis that increased flexion and extension in thelumbar spine of monkeys may be an importantmechanical factor.

The study of arthritic patterns in closelyrelated animals (including humans) who differin locomotory adaptations is one that holdsconsiderable promise. However, the pro-visional data now available suggest we wouldbe too optimistic if we expect comfortablysimple patterns to emerge from this work.

Suggested directions for future studiesThere are numerous areas on which osteo-logists can and should focus serious attentionin future studies of OA. Critical to comparativestudies is the well recognised need forstandardisation of scoring techniques. In thepast, investigators have used a variety ofmethods for scoring OA, and there exists ageneral consensus regarding moderate andsevere involvement. However, the diversity ofscoring techniques frequently makes betweenpopulation comparisons difficult; therefore dis-cussion must, of necessity, remain superficial.The need for standardisation has frequentlybeen the topic of informal discussion amongpalaeopathologists, and attempts aimed atimproving the situation have been made.6'Another recognised need is increased access

to surgically derived, macerated speicmens ofarthritic joints. Such specimens, accompaniedby detailed patient histories (age, occupation,symptomology) would be invaluable tools inhelping osteologists assess the aetiopatho-genesis of OA and the impact of this disease inarchaeological populations. Obviously, osteo-logists must rely upon clinical colleagues toprovide such materials with patient consent.Perhaps, with increased awareness of this needwithin the medical community, such speci-mens will be collected more systematically.Non-human studies also would be tremen-

dously aided by greater availability ofmacerated skeletons of captive animals, par-ticularly non-human primates. In the UnitedStates, remains of zoo animals are mostcommonly donated to universities fordissection and the skeletons are almost neveravailable for systematic study. If zoo derivedspecimens were made available, along withinformation pertaining to age at death, habitatdesign, substrate materials, and activitypatterns, osteologists could begin to delineatethose factors important to the development ofOA in non-human primates. This ability wouldbe greatly enhanced through comparisons withremains of free ranging primates. The smallsample of 14 chimpanzees from Dr JaneGoodall's study group in Tanzania has beeninvaluable because of existing, often detailed,life history information and known or approxi-mate age at death of individuals. Plans are nowunder way for the collection of chimpanzee andother primate remains from groups understudy in the Kibale National Park, Uganda.Field researchers are increasingly aware of theimportance of osteological analysis of theiranimals, and certainly there will be greaterattention to the collection of bodies in thefuture. Unfortunately, at least in the case ofchimpanzees, animals sometimes simply dis-appear and their fate remains unknown.

449



j.com/

Ann R

heum D


nloaded from

http://ard.bmj.com/

40urmain, Kilgore

Osteoarthritis is one of the most common

ailments of modem humans. This disease,apparently, was also quite common inantiquity. By expanding the horizons of thedatabase ofOA in time and space, we also maybe able to expand the horizons of our under-standing of this often crippling disease.

1 Rothschild B. Radiologic assessment of osteoarthritis indinosaurs. Ann Carnegie Mus 1990; 59: 295-301.

2 Stewart T D. The rate of development of vertebralosteophytosis in skeletal age identification. Leech 1958;28: 144-51.

3 Chapman F H. Incidence of arthritis in a prehistoric MiddleMississippian Indian population. Indiana Acad Sci Proc1962; 72: 59-62.

4 Anderson J E. The people of Fairty: an osteological analysisof an Iroquois ossuary. National Museum of CanadaBulletin, 193, Contributions to Anthropology, 1961-1962.Toronto: National Museum, 1963.

5 Angel J L. Early skeletons from Tranquility, California.Smithsonian ContribAnthropol 1966; 2: No 1.

6 Ortner D J. Description and classification of degenerativebone changes in the distal joint surfaces of the humerus.Am Phys Anthropol 1968; 28: 139-55.

7 Rogers J, Watt I, Dieppe P. Arthritis in Saxon and medievalskeletons. BMJ 1981; 283: 1668-70.

8 Bennike P. Paleopathology of Danish skeletons. Copenhagen:Akademisk Forlag, 1985.

9 Jurmain R D. Stress and the etiology of osteoarthritis. AmIPhysAnthropol 1977; 46: 353-66.

10 Jurmain R D. Paleoepimemiology of degenerative jointdisease. Med Coil Virginia Q 1978; 14: 45-56.

11 Jurmain R D. The pattern of involvement of appendiculardegenerative joint disease. Am Phys Anthropol 1980; 53:143-50.

12 Larsen C S. The anthropology of St Catherine's Island 3.Prehistoric human biological adaptation. Am Mus NatHist Anthropol Papers 1982; 57: 159-270.

13 Kilgore L. Degenerative joint disease in a medieval Nubianpopulation [Thesis]. Boulder, CO: University ofColorado, 1984.

14 Bridges P S. Degenerative joint disease in hunter-gatherersand agriculturists from the southeastern United States.AmJrPhysAnthropol 1991; 85: 379-91.

15 Lovell N. Spinal arthritis and physical stress at Bronze AgeHarappa. Am Phys Anthropol 1994; 93: 149-64.

16 Cobb S. The frequency of the rheumatic diseases. Cambridge:Harvard University Press, 1971.

17 Rothschild B M, Woods R J, Ortel W. Rheumatoid arthritis"in the buff:" erosive arthritis in defleshed bones. AmPhysAnthropol 1990; 82: 441-9.

18 Leisen J C, Duncan H, Riddle J M. Rheumatoid erosivearthropathy as seen in macerated (dry) bone specimens.In: Ortner D J, Aufderheide A C, eds. Humanpaleopathology. Current synthesis and future options.Washington DC: Smithsonian Institution Press, 1991;211-5.

19 Beitzke H. Ueber de sogen. Arthritis deformans atrophica.ZKlin Med 1912; 74: 215-29.

20 Heine J. Uber die Arthritis deformans. Virchows Arch PatholAnat 1926; 260: 521-663.

21 Chung E B. Aging in human joints. I. Articular cartilage.3'NatlMedAssoc 1966; 58: 87.

22 DeRousseau C J. Osteoarthritis in rhesus monkeys and gibbons:A locomotor model of joint degeneration. Contributions toPrimatology, Vol 25. Basel: Karger, 1988.

23 Lovell N C. An evolutionary framework for assessing injuryand illness in nonhuman primates. Yearbook PhysAnthropol 1991; 34: 117-55.

24 Lovell N C. Patterns of injury and illness in great apes. Askeletal analysis. Washington DC: Smithsonian InstitutionPress, 1990.

25 Rothschild B M, Woods R D. Reactive erosive arthritis. AmPrPtimatol 1991; 25: 49-56.

26 Jurmain R D. Trauma, degenerative disease, and otherpathologies among the Gombe chimpanzees. Am PhysAnthropol 1989; 80: 229-37.

27 Bourke J B. A review of the paleopathology of the arthriticdiseases. In: Brothwell D L, Sandison A T, eds. Diseasesin antiquity. Springfield IL: Charles C Thomas, 1967;352-70.

28 Rogers J, Waldron T, Dieppe P, Watt I. Arthropathies inpaleopathology: The basis of classification according tomost probable cause. JArchaeol Sci 1987; 14: 179-93.

29 Jurmain R D. The distribution of degenerative joint diseasein skeletal populations (Thesis]. Cambridge: HarvardUniversity, 1975.

30 Jurmain R D. Paleoepidemiology of a Central Califomiaprehistoric population from CA-Ala-329. II. Degener-ative disease. AmJ Phys Anthropol 1990; 83: 83-94.

31 Pierce L C. A comparison of the pattem of involvement ofdegenerative joint disease between an agricultural and anon-agricultural skeletal series [Thesis]. Knoxville:University ofTennessee, 1987.

32 Burr D B, Martin R B, Schaffier M B, Jurmain R D,Harner E J, Radin E L. Osteoarthrosis: Sex-specificrelationship to osteoporosis. Am Jf Phys Anthropol 1983;61: 299-303.

33 Merbs C F. Patterns of activity-induced pathology in aCanadian Inuit population. Archaeol Surv Canada 1983;Paper 119.

34 Swedborg I. Degenerative changes of the human spine.Stockholm: Osteological Research Laboratory, Universityof Stockholm, 1974.

35 Todd T W. Age changes in the pubic bone. I: The maleWhite pubis. Am_J Phys Anthropol 1920; 3: 285-334.

36 Bass W M. Human osteology. A laboratory and field manual,3rd edn. Columbia, MO: Missouri ArchaeologicalSociety, 1987.

37 Katz D, Suchey J. Age determination of the male os pubis.AmJPhysAnthropol 1986; 69: 427-35.

38 Lovejoy C 0, Meindl R S, Pryzbeck T R, Mensforth R P.Chronological metamorphosis of the auricular surface ofthe ilium: A new method for the determination of adultskeletal age at death. Am J Phys Anthropol 1985; 68:15-28.

39 Iscan M Y, Loth S R, Wright R K. Metamorphosis at thestemal rib end: A new method to estimate age at deathin White males. AmJPhysAnthropol 1984; 65: 147-56.

40 Nathan H. Osteophytes of the vertebral column: Ananatomical study of their development according to age,race, and sex with consideration as to their etiology andsignificance. JfBoneJoint SurgAm 1962; 44: 243-68.

41 Collins D H. The pathology of articular and spinal diseases.London: E Arnold, 1949.

42 Fischer A. Rheumatismus als Berufskrankheit. ActaRheumatologica 1932; 4: 24-8.

43 Molleson T. The eloquent bones of Abu Hureyra. Sci Am1994; 271: 70-5.

44 Solonen K A. The joints of the lower extremities of footballplayers. Ann Chir Gynaecol 1966; 55: 176-80.

45 Brodelius A. Osteoarthritis of the talar joints in footballersand ballet dancers. Acta Orthop Scand 1961; 30: 309-14.

46 Lawrence J S. Rheumatism in coal miners. Part III:Occupational factors. BrJ Ind Med 1955; 12: 249-61.

47 Lawrence J S, de Graff R, Laine V A. Degenerative jointdisease in random samples and occupational groups. In:Jeffrey M R, Ball J, eds. The epidemiology of chronicrheumatism, Vol 1. Oxford: Blackwell, 1963; 98-119.

48 Mankin H, Brandt K D, Shulman L E. Workshop onetiopathogenesis of osteoarthritis. Proceedings andrecommendations.7 Rheumatol 1986; 13: 1130-60.

49 Roche L, Maitrepairre J, Lejune E, Mermet J. Les atteintesdu membre superieur chez ouvriers travaillant au marteaupneumatique. Arch Mal Prof 1961; 22: 57-61.

50 Ruffer M A, Rietti A. On osseous lesions in ancientEgyptians.JPatholBacteriol 1911; 16: 439-65.

51 Hodges D C. Temporomandibular joint osteoarthritis in aBritish skeletal population. AmJfPhys Anthropol 1991; 85:367-77.

52 Pfeiffer S. The skeletal biology ofArchaic populations of theGreat Lakes region. Archaeol Surv Canada 1977; Paper64.

53 Nelson R K. Hunters of the northern ice. Chicago: Universityof Chicago Press, 1969.

54 Bridges P S. Prehistoric arthritis in the Americas. Annu RevAnthropol 1992; 21: 67-91.

55 Bourke J B. Trauma and degenerative diseases in ancientEgypt and Nubia.JHum Evol 1969; I: 225-32.

56 Shore L R. On osteoarthritis in the dorsal intervertebraljoints. BrJSurg 1935; 22: 833-49.

57 Bridges P S. Vertebral arthritis and physical activities in theprehistoric southeastem United States. Am _J PhysAnthropol 1994; 93: 83-93.

58 White A A, Panjabi M M. Clinical biomechanics of the spine.Philadelphia: J B Lippincott, 1978.

59 Rothschild B M, Woods R J. Osteoarthritis, calciumpyrophosphate deposition disease, and osseous infectionin Old World primates. Am J Phys Anthropol 1992; 87:341-7.

60 Bramblett C A. Pathology of the Darajani baboon. Am J7PhysAnthropol 1967; 26: 331-40.

61 Waldron T, Rogers J. Interobserver variation in codingosteoarthritis in human skeletal remains. Int J Osteo-archaeol 1991; 1: 49-56.

450



j.com/

Ann R

heum D


nloaded from

http://ard.bmj.com/

Documents

Skeletal evidence of osteoarthritis: a palaeopathological