Biomechanical implications of degenerative joint disease in the apophyseal joints of human thoracic and lumbar vertebrae.код для вставкиСкачать
AMERICAN JOURNAL OF PHYSICAL ANTHROPOLOGY 136:318–326 (2008) Biomechanical Implications of Degenerative Joint Disease in the Apophyseal Joints of Human Thoracic and Lumbar Vertebrae Kate Robson Brown,1* Phill Pollintine,2 and Mike A. Adams2 1 2 Department of Archaeology and Anthropology, University of Bristol, Bristol BS8 1UU, UK Department of Anatomy, University of Bristol, Bristol BS2 8EJ, UK KEY WORDS apophyseal joint; cartilage loss; joint disease; loading; biomechanics ABSTRACT An experimental technique for quantifying load-sharing in cadaveric spines is used to test the hypothesis that degenerative changes in human apophyseal joints are directly related to high levels of compressive load-bearing by these joints. About 36 cadaveric thoraco-lumbar motion segments aged 64–92 years were subjected to a compressive load of 1.5 kN. The distribution of compressive stress was measured in the intervertebral discs using a miniature pressure transducer, and stress measurements were summed over area to give the compressive force resisted by the disc. This was subtracted from the applied 1.5 kN to indicate compressive load-bearing by the apophyseal joints. The cartilage of each apophyseal joint surface was then graded for degree of degeneration. After maceration, each joint surface was scored for degenerative joint disease (DJD) affecting the bone. Results demonstrated that the apophyseal joints resisted 5–96% (mean 45%) of the applied compressive force. A signiﬁcant positive correlation was demonstrated between age and cartilage degeneration, age and DJD bone score, apophyseal joint load-bearing and bone score, and cartilage score and load-bearing. The latter correlation was strongest for load-bearing above 50%. Ordinal regression showed that the variables describing bone DJD (marginal osteophytes, pitting, bony contour change, and eburnation) were signiﬁcantly correlated with degree of cartilage degeneration. It is concluded that in elderly individuals apophyseal joint load-bearing above a threshold of 50% is associated with severe degenerative changes in cartilage and bone, and that markers of DJD observed palaeopathologically may be used as predictors of such loading in life. Am J Phys Anthropol 136:318–326, 2008. V 2008 Wiley-Liss, Inc. The extent to which degenerative joint disease (DJD) in bone may be seen as an indicator of mechanical or occupational stress has long been a focus of debate within osteoarchaeology (e.g., Jurmain, 1991; Waldron, 1991; Knüsel et al., 1997; Sofaer Derevenski, 2000). While some studies of paleopathology continue to use DJD as a predictor of speciﬁc activity patterns (Gerszten et al., 2001), most urge caution (Jurmain, 1990, 1991; Bridges, 1991, 1994; Lovell, 1994; Knüsel et al., 1997). Clinical studies have failed to demonstrate a simple relationship between DJD and speciﬁc patterns of movement or activity (Lane et al., 1986; Videman et al., 1990, and recently reviewed by Lequesne et al., 1997 and by Conaghan, 2002). Certain aspects of DJD appear to be related to age (Rogers et al., 1987; Jurmain, 1991; Waldron, 1991, 1992; Knüsel et al., 1997), whereas others appear to be related to sex, genetic inheritance, or body weight (Merbs, 1983; Waldron, 1994; Maat et al., 1995; Sofaer Derevenski, 2000; Meulenbelt et al., 2006). DJD is probably best thought of as resulting from a combination of ‘‘systemic’’ risk factors (which can lead to degenerative changes to many joints within an individual) and localized factors which may be more closely related to the mechanical loading experienced at a particular joint (Kellgren and Lawrence, 1958). An additional complication is that skeletal tissues are able to strengthen in response to mechanical loading which does not damage them, so that moderate loading appears to be better for the joints than either low or high loading (Videman et al., 1990). Not surprisingly, this complexity has tended to discourage consideration of the biomechanical implica- tions of DJD, even where the presence and pattern of such pathology is well documented (e.g., Weber and Czarnetzki, 2002; Weber and Spring, 2004). The vertebral column is one region of the skeleton that has become a focus of interest for paleopathological study into DJD. Comparative studies between species have suggested that modern humans suffer vertebral osteophytosis and osteoarthritis more commonly and more severely than any other living primate, which might support the view that bipedal posture and locomotion generate localized mechanical factors affecting the spine that differ from those of other primate species (Jurmain, 1990, 2000). Detailed study of the morphology and distribution of degenerative changes in the human spine suggest that complex interacting factors are involved. For example, postmortem studies indicate that traumatic ﬂexion or torsion injury to the lumbar spine C 2008 V WILEY-LISS, INC. C Grant sponsors: BBSRC and University of Bristol Research Fellowship. *Correspondence to: Kate Robson Brown, Department of Archaeology and Anthropology, University of Bristol, Bristol BS8 1UU. E-mail: email@example.com Received 25 July 2007; accepted 28 December 2007 DOI 10.1002/ajpa.20814 Published online 6 March 2008 in Wiley InterScience (www.interscience.wiley.com). VERTEBRAL FACET JOINT DEGENERATION can damage the articular surfaces of the apophyseal joints, possibly precipitating DJD (Adams and Hutton, 1981; Twomey et al., 1989), and the pattern of cartilage damage may be indicative of the type of mechanical overload involved (Swanepoel et al., 1995). Palaeopathological studies have suggested that DJD occurs in the apophyseal joints of the cervical spine, which is subjected to relatively wide range of motion (Jurmain, 1977; Knüsel et al., 1997) and also in the lower lumbar spine, which is heavily loaded (Lewin, 1964; Jurmain, 1977; Bridges, 1994; Knüsel et al., 1997; Sofaer Derevenski, 2000; Haefeli et al., 2006). It is not clear if DJD reﬂects excessive movements, or loading, or both. The situation can be further complicated by the adjacent intervertebral discs. Disc degeneration and narrowing increase compressive load-bearing by the apophyseal joints (Dunlop et al., 1984; Adams et al., 2006), especially when specimens are loaded to simulate a lordotic posture (Dunlop et al., 1984), and clinical studies conﬁrm that lumbar disc degeneration increases the risk of DJD in the apophyseal joints (Butler et al., 1990). Evidence from postmortem studies has led to speculation that patterns of cartilage wear on the apophyseal joint surfaces may reﬂect focal loading of these surfaces in certain individuals and postures (Swanepoel et al., 1995; Tischer et al., 2006). However, there is little ﬁrm evidence to support this suggestion. Given this complex background, it is perhaps not surprising that DJD in the apophyseal joints is of uncertain biomechanical relevance to palaeopathology. There has recently been a call for experimentally tested and validated biomechanical models to enable skeletal function to be inferred from skeletal morphology (Pearson and Lieberman, 2004; Van der Merwe et al., 2006). The present study responds to this call by combining experimental biomechanics and quantitative osteoarchaeology in order to investigate the mechanical implications of degenerative changes in human apophyseal joints. Direct quantiﬁcation of the compressive load transmitted by the apophyseal joints is difﬁcult, even in cadaveric specimens (el-Bohy et al., 1989), but it can be achieved indirectly, by calculating the compressive force transmitted by the adjacent intervertebral disc (Pollintine et al., 2004a,b; Adams et al., 2006). Here this technique will be used to compare load-bearing in the apophyseal joints of elderly cadavers with direct evidence of DJD in the same joints, assessed from morphological changes in cartilage and bone. MATERIALS AND METHODS Twenty-two cadaveric thoraco-lumbar spines aged 64– 92 years (mean 77.4 years, STD 8.5 years) were obtained from post-mortem rooms and stored at 2208C until required for testing. None of the individuals had a history of traumatic injury to the spine, or suffered any long-term disease that might inﬂuence disc degeneration or bone strength. Spines were subsequently thawed and dissected into 36 ‘‘motion segments.’’ The motion segment, consisting of two whole vertebrae and the intervening soft tissues (disc and ligaments) is the smallest repeating unit of the spine that preserves the intervertebral joints intact, and as such has been used extensively for in-vitro modeling of the mechanical behavior of the spine. The following ligaments were preserved: the anterior and posterior longitudinal ligaments, the supraspinous and interspinous ligaments, the ligamentum ﬂa- 319 Fig. 1. Apparatus for measuring distributions of compressive ‘‘stress’’ in loaded cadaveric intervertebral discs. Different postures could be simulated by adjusting the height of the rear roller. Compressive stress was measured by means of a pressure transducer side-mounted in a 1.3-mm diameter needle. vum and the capsule ligaments of the left and right apophyseal joints. Overall within the study, all levels between T9-T10 and L5-S1 were represented, but it was not possible to obtain identical segments for each individual. The total number and range of segments included in the study were constrained by availability and practicality. Each motion segment was radiographed prior to testing as a record of bone density and distribution and disc height (Adams et al., 1986). Each motion segment was secured in two cups of dental stone and loaded on a computer-controlled hydraulic materials testing machine (Dartec Ltd., Stourbridge, UK) as shown in Figure 1. The compressive force on the spine is conventionally deﬁned as that force which acts perpendicular to the mid-plane of the disc (the plane equidistant from its endplates). By keeping this plane horizontal, we ensured that only compression was applied. Each specimen underwent 2 h of ‘‘creep,’’ during which a constant compressive force of 1.5 kN was applied. This was used to reduce disc height and water content by an amount similar to that which occurs in life during the ﬁrst few hours of each day in response to moderate physical activity (McMillan et al., 1996). To obtain the compressive force transmitted by the apophyseal joints, it was ﬁrst necessary to calculate the compressive load transmitted by the intervertebral disc and adjoining vertebral bodies. As a ﬁrst step, the distribution of compressive stress within the disc was measured by stress proﬁlometry, as described in detail elsewhere (Adams et al., 1996; Pollintine et al., 2004a,b. Brieﬂy, each motion segment was subjected to a static compressive load of 1.5 kN to represent the combined effects of muscle tension and superincumbent body weight during light manual labor (Nachemson, 1981; Sato et al., 1999). During the loading period of 20 s, the distribution of vertically acting compressive stress (force per unit area) was measured inside the disc by pulling a miniature strain-gauged pressure transducer (Gaeltec, Dunvegen, Scotland) through the disc along its mid-sagittal diameter (see Fig. 1). Stress proﬁles were obtained with the motion segment positioned in 28 of extension to simulate the lordotic erect standing posture (Adams et al., 1988). The compressive stress distribution was then integrated over crosssectional area to obtain the compressive force acting on the intervertebral disc. Subtracting this force from the applied 1.5 kN then indicates the compressive force transmitted by the neural arch (effectively, the apophyseal joints: Adams and Hutton, 1980). American Journal of Physical Anthropology 320 K.R. BROWN ET AL. Validation tests have shown that this technique predicts apophyseal joint load-bearing with errors of 2–8% (Pollintine et al., 2004a). Following the completion of the biomechanical testing phase of the project, the two vertebrae of each motion segment were separated. The joint capsules were opened dorsally and the inferior and superior articular processes were carefully separated. In turn, the left and right apophyseal joint surfaces (LAP, RAP) were stained with Indian ink to reveal the degree of cartilage damage (Meachim, 1972; Tischer et al., 2006). The presence and severity of damage was recorded on a scale of 0–3, where Grade 0 signiﬁed no damage, Grade 1 indicated moderate damage (minor or peripheral pits or ﬁssures on an otherwise normally shiny surface), Grade 2 indicated severe damage (many ﬁssures, severe pitting, thin and rough surfaced cartilage remaining), and Grade 3 indicated complete cartilage loss. After the cartilage had been scored, the remaining soft tissue was removed by maceration. When completely dry, the apophyseal joint articular surfaces were scored again for pathological bone lesions. The following lesions were recorded: distinct osteophytes, joint surface contour or shape change, pitting, and eburnation. These changes were each scored for presence and severity on a scale of 0–3, where 0 signiﬁed no lesions or bone change, 1 indicated moderate lesions or change, 2 indicated severe lesions or change, and 3 indicated that very severe change was observed. This grading system was based on previously published descriptions and scoring schemes (Sager, 1969; Resnick and Niwayama, 1988; Jurmain, 1990; Knüsel et al., 1997). Osteophytosis (OP) was recorded as a proportion of the articular surface width, based on the methodology of Knüsel et al. (1997). Articular surface width was measured using callipers, and the proportion of that width presented by OP informed the score. When OP was smaller than half the width and less than 50% of the circumference of the margin affected, a 1 was scored. If in excess of these values, but lower than 75% of the width and margin the lesion was recorded as a 2. A third category was included to account for osteophytes larger than 75% of the joint surface width or affecting more than 75% of the joint margin. Pitting was recorded as 1 when less than half the articular surface was affected, as 2 when between 50 and 75% was affected, and as 3 when more was affected. Similarly eburnation was recorded as 1 when a small area of less than half the articular surface was affected, as 2 when between 50 and 75% was affected, and as 3 when more was affected. Lesions were recorded independently for each apophyseal joint, and to create an overall measure of DJD for each, the four bone scores were summed to give an overall score out of 12. There were no missing values. To assess the strength of the relationships between neural arch loading, age, cartilage scores, and bone scores Spearman’s rank correlation was used. This method is particularly useful for datasets which may not meet the assumptions about normality and homoscedasticity required by linear regression correlation (SPSS, Chicago, IL, version 12). Unless otherwise stated, each segment was treated as an independent data point. In the examination of the key relationship between apophyseal joint load-bearing and summed bone score a loess curve was employed to present a locally weighted regression (Cleveland, 1979). To investigate the effect of sex on summed bone score, a Mann-Whitney U test was employed. The Mann-Whitney American Journal of Physical Anthropology U tests the null hypothesis that two samples are drawn from the same population; it is a nonparametric alternative to the independent-samples t test. This rank sum test is used when the normality assumption is questionable or when data are ordinal. One of the disadvantages of nonparametric tests is their lower power compared to their parametric equivalents when the assumptions underlying the test are met, but the advantages are that the test is not affected by outliers and is more suitable for small datasets (Howell, 1996). To test the hypothesis that summed bone score medians from the left and right apophyseal joints are equal a Wilcoxon matched pairs, signed ranks test was employed. This test is a nonparametric counterpart of the paired-samples t test. To investigate the utility of the four constituent bone score observations (marginal osteophyte, pitting, bony contour change, eburnation) as predictors of cartilage score, an ordinal logistic regression model was constructed. The ordinal logistic model is an extension of the general linear model to ordinal categorical data. Cartilage score was taken as the dependent variable and the reference category was Score 3. Finally, the distribution of apophyseal joint degeneration across the spine was investigated by plotting the mean bone score against spinal level. Mixed ANOVAs were used to examine the interacting inﬂuences of within-subject factors (spinal level) and between-subject factors (age, sex) on apophyseal joint load-bearing, and DJD degeneration scores. RESULTS Load-bearing results and DJD scores are summarized in Table 1. Results of the loading experiments demonstrate that the apophyseal joints transmitted a measurable amount of the compressive force applied. The proportion averaged 45% and ranged from 5 to 96% of the total compressive force (Table 1). The mean and median summed bone score for the female group (mean 5 4.09, SD 5 3.47, median 5 3.93) was slightly higher than that of the male group (mean 5 3.43, SD 5 2.70, median 5 3.0). The Mann-Whitney U test showed this difference not to be signiﬁcant: U 5 619.5; exact P [ 0.05 (two-tailed) (Table 2). The Wilcoxon matched-pairs, signed ranks test showed that the difference between the median summed bone score for left (mean 5 3.56, SD 5 3.11, median 5 3.0) and right (mean 5 3.92, SD 5 3.04, median 5 3.0) apophyseal joints was not signiﬁcant: exact P [ 0.05 (two-tailed). A positive correlation was found between apophyseal joint load-bearing and summed bone score (P \ 0.01) (Table 3 and Fig. 2). A marginally less strong correlation was found between apophyseal joint load-bearing and cartilage score (P \ 0.01). Load-bearing was not found to correlate with age (Table 3). However, a positive correlation was found between age and cartilage score (P \ 0.05), and a stronger correlation was found between age and summed bone score (P \ 0.01) (Table 3). Cartilage score and bone score were also strongly positively correlated (P \ 0.01) (Table 3). The relationship between degree of cartilage loss and the four constituents of observed bone change is presented as a matrix of correlations in Table 4. All four variables describing bone changes (marginal osteophytes, pitting, bony contour change, eburnation) are positively 321 VERTEBRAL FACET JOINT DEGENERATION TABLE 1. Summarized results Specimen Age Sex Motion seg 1 64 64 64 64 67 67 71 71 71 71 71 71 76 76 76 76 76 76 85 85 85 85 87 87 87 87 88 88 92 92 92 92 65 65 67 67 f f f f f f f f f f f f f f f f f f f f f f f f f f f f f f f f m m m m L2/L3 L2/L3 T12/L1 T12/L1 T12/L1 T12/L1 L2/L3 L2/L3 T12/L1 T12/L1 T10/T11 T10/T11 L2/L3 L2/L3 T12/L1 T12/L1 T12/L1 T12/L1 T11/T12 T11/T12 L2/L3 L2/L3 L2/L3 L2/L3 L4/L5 L4/L5 L4/L5 L4/L5 L3/L4 L3/L4 T12/L1 T12/L1 L2/L3 L2/L3 T11/T12 T11/T12 2 3 4 5 6 7 8 9 10 11 Joint Na load Cart score Bone score LAP RAP LAP RAP LAP RAP LAP RAP LAP RAP LAP RAP LAP RAP LAP RAP LAP RAP LAP RAP LAP RAP LAP RAP LAP RAP LAP RAP LAP RAP LAP RAP LAP RAP LAP RAP 41 41 33 33 65 65 5 5 65 65 85 85 40 40 49 49 30 30 48 48 64 64 96 96 96 96 70 70 86 86 36 36 50 50 54 54 1 1 1 1 1 1 1 2 0 1 2 3 1 2 1 0 1 1 1 1 3 3 2 3 3 3 2 2 1 2 1 2 1 1 2 2 2 2 0 0 1 2 2 3 1 1 3 6 4 6 0 0 2 2 3 4 9 10 8 8 11 11 9 8 5 5 1 2 2 4 1 1 Specimen 13 12 16 12 14 15 17 18 19 20 21 22 Age Sex Motion seg Joint Na load Cart score Bone score 67 67 72 72 72 72 72 72 72 72 72 72 72 72 74 74 79 79 81 81 81 81 81 81 82 82 82 82 89 89 90 90 90 90 90 90 m m m m m m m m m m m m m m m m m m m m m m m m m m m m m m m m m m m m L4/L5 L4/L5 L1/L2 L1/L2 T12/L1 T12/L1 T12/L1 T12/L1 T9/T10 T9/T10 L3/L4 L3/L4 L5/S1 L5/S1 T10/T11 T10/T11 L3/L4 L3/L4 L1/L2 L1/L2 L4/L5 L4/L5 T11/T12 T11/T12 T11/T12 T11/T12 L1/L2 L1/L2 T11/T12 T11/T12 L1/L2 L1/L2 L5/L6 L5/L6 T11/T12 T11/T12 LAP RAP LAP RAP LAP RAP LAP RAP LAP RAP LAP RAP LAP RAP LAP RAP LAP RAP LAP RAP LAP RAP LAP RAP LAP RAP LAP RAP LAP RAP LAP RAP LAP RAP LAP RAP 9 9 12 12 81 81 10 10 37 37 34 34 13 13 81 81 15 15 28 28 33 33 50 50 74 74 27 27 71 71 10 10 20 20 10 10 2 2 1 2 1 2 1 1 0 1 1 2 1 0 3 2 2 2 2 3 1 1 3 3 2 3 1 2 2 2 2 2 1 1 0 1 4 4 2 3 6 6 0 1 0 1 2 3 0 0 10 10 6 4 5 3 4 4 6 7 8 9 3 3 3 2 3 3 2 2 0 1 Abbreviation: Motion seg, motion segment; T, thoracic vertebra; L, lumbar vertebra; LAP, left apophyseal joint; RAP, right apophyseal joint; Na load, % load transmitted by the neural arch; Cart score, cartilage score; Bone score, Summed DJD bone score. and signiﬁcantly correlated with degree of cartilage loss (P \ 0.01). The ordinal logistic regression method was used to model the relationship between cartilage score and these four observed bone scores, whilst allowing for age effects. The predictor (marginal osteophytes, pitting, bony contour change, eburnation) was entered into the model with cartilage score as the dependent variable, and with the reference category as Score 3. A value of P \ 0.05 (two sided) was considered signiﬁcant (Norus̆is, 2005). The results of this analysis (Table 5) show that for the category of marginal osteophytes, all scores are signiﬁcantly related to cartilage score. Specimens graded low for marginal osteophytes are less likely to have high scores for cartilage loss than specimens with high marginal osteophyte scores. For the category of pitting, Scores 0 and 1 are signiﬁcantly related to cartilage score, although pitting Score 2 is not. So here, specimens graded low (0 or 1) for pitting are less likely to have high scores for cartilage loss than specimens a pitting Score of 3, but pitting Score 2 is not a good predictor. For the category of bony contour change, Scores 0 and 1 are strongly signiﬁcantly related to cartilage score (P \ 0.01) while Score 2 is lightly less strongly related (P \ 0.05). For the category of eburnation, Score 0 is signiﬁcantly related to cartilage score (P \ 0.01), and Score 1 is slightly less strongly related (P \ 0.05). In conclusion, TABLE 2. Mann–Whitney U test for sexual differentiation in summed bone score Mann-Whitney U Wilcoxon W Z Exact Sig. (two-tailed) 619.500 1399.500 20.273 0.788 low bone scores (0, 1) of all categories are good predictors of low cartilage score, while high bone scores are more likely to predict high cartilage scores. Finally, to investigate any patterning of apophyseal joint degeneration with spinal level within the whole data set (all represented levels and all individuals), the summed bone score was plotted against spinal level (see Fig. 3). Here, the bone score represents the total score for each parameter of bone degeneration, averaged for left, right, and upper and lower articular processes. Each box shows the median, quartiles (25–75%), and extreme values (range). Mixed ANOVAs were used to examine the interacting inﬂuences of within-subject factors (spinal level) and between-subject factors (age, sex) on apophyseal joint load-bearing, and DJD degeneration scores. The withinsubject analysis showed that level was not associated with neural arch loading (P 5 0.1) or apophyseal joint American Journal of Physical Anthropology 322 K.R. BROWN ET AL. TABLE 3. Spearman’s rank correlations between age, neural arch load-bearing, cartilage score, and summed bone score Spearman’s rho Age Naload Cartscore Bonescore a b Correlation coefﬁcient Sig. (two-tailed) Correlation coefﬁcient Sig. (two-tailed) Correlation coefﬁcient Sig. (two-tailed) Correlation coefﬁcient Sig. (two-tailed) Age Naload Cartscore Bonescore – 0.112 0.347 0.252a 0.033 0.342b 0.003 . 0.349b 0.003 0.481b 0.000 . 0.737b 0.000 . Correlation is signiﬁcant at the 0.05 level (two-tailed). Correlation is signiﬁcant at the 0.01 level (two-tailed). Fig. 2. Scatterplot showing the % apophyseal joint loadbearing against summed bone score, with the loess line marked. degeneration (P 5 0.9). The between subject analysis revealed that gender and age had no signiﬁcant effect on neural arch loading and apophyseal joint degeneration (P [ 0.1). DISCUSSION Apophyseal joints evidently play a major, but variable, role in resisting compressive forces acting on the human thoracolumbar spine. On average, they resisted 45% of this force, but in particular specimens it could be as much as 96% or as little as 5% (Table 1). This wide variation in levels of compressive load may be explained by the degree of disc degeneration. In a previous study that included the specimens described here, it was shown that disc degeneration transfers compressive load bearing from the anterior vertebral body to the neural arch in upright postures, but that this effect does not depend on spinal level or sex (Adams et al., 2006). In the present study, the results of the mixed ANOVA examining within-subject factors suggest that apophyseal joint loadbearing and degeneration were not systematically inﬂuenced by the spinal level. However, distribution of apophyseal joint degeneration with spinal level across the whole data set appears to reveal a pattern of peaks and troughs (see Fig. 3), which may mask any systematic trend within a single individual. The peaks of severity across the whole group lie between T10-T11 and L2-L5. This ﬁnding is consistent with previous work (Nathan, 1962; Sager 1969; Bridges, 1994; Roberts and Manchester, 1995; Knüsel et al.,1997). American Journal of Physical Anthropology The ‘‘compressive force’’ that may underlie this ﬁnding is deﬁned as the force acting down the long axis of the spine, perpendicular to the mid-plane of the intervertebral disc at each spinal level. Compression dominates spinal loading during vigorous activities, and arises primarily from tension in the paraspinal muscles (Dolan et al., 1994) which tend to follow the curvature of the spine. Even in simple standing postures, the need to stabilize the upright spine leads to high antagonistic activity of trunk muscles (Cholewicki et al., 1997), and considerable spinal compression (Sato et al., 1999). Shear loading of the spine acts in a direction perpendicular to the compressive force (i.e. parallel to the mid-plane of the disc) and is resisted mainly by the articular surfaces of the apophyseal joints (Cyron and Hutton, 1980). Shear loading arises mainly from superincumbent body weight, and so tends to be greatest in the lower lumbar spine, where the discs are most inclined to the vertical. This may explain the relatively large size of the lower lumbar apophyseal joints (Davis, 1961). Lumbar apophyseal joints also resist axial rotation (Adams and Hutton, 1981), and they couple axial rotation with lateral bending in a posture-dependent manner (Cholewicki et al., 1996). Nevertheless, it is compressive loading that acts on the apophyseal joints for most of the time in upright posture, and which is applied to all spinal levels, and this is why compressive loading is the focus of the present study. Previous studies have suggested that mechanical loading may underlie the development of DJD in the lower spine (e.g. Jurmain, 1977; Putz, 1985; Swanepoel et al., 1995; Knüsel et al., 1997; Sofaer Derevenski, 2000), and the results presented here provide more detailed evidence which helps to explain how this occurs, at least in relatively elderly individuals (see Fig. 2). The loess curve in this ﬁgure presents a locally weighted regression, and at any particular point on the x axis it is determined only by the points in that vicinity (Cleveland, 1979). Figure 2 suggests that specimens in which less than 50% of the experimental compressive load was resisted by the apophyseal joints did not show a strong relationship between increasing load and increasing degeneration of the apophyseal joints. However, where the apophyseal joints resisted more than 50% of the compressive load, bone changes in the apophyseal joints were much more severe. This may imply that there is a compressive force threshold under which the apophyseal joints can accommodate the force transmission, but over which degeneration of cartilage and (especially) bone becomes more likely. There is experimental evidence that disc degeneration and narrowing can cause the apophyseal joints to transmit compressive load by extra-articular impingement between the bony tips of the inferior articular proc- 323 VERTEBRAL FACET JOINT DEGENERATION TABLE 4. Spearman’s rank correlations between age, cartilage score, and the four bone score categories Spearman’s rho Age Cartscore Marost Pitting Bonycont Ebur a Correlation coefﬁcient Sig. (two-tailed) Correlation coefﬁcient Sig. (two-tailed) Correlation coefﬁcient Sig. (two-tailed) Correlation coefﬁcient Sig. (two-tailed) Correlation coefﬁcient Sig. (two-tailed) Correlation coefﬁcient Sig. (two-tailed) Age Cartscore Marost Pitting Bonycont – 0.285a 0.000 0.224a 0.004 0.405a 0.000 0.287a 0.000 0.238a 0.002 – 0.636a 0.000 0.606a 0.000 0.610a 0.000 0.580a 0.000 0.595a 0.000 0.783a 0.000 0.475a 0.000 0.532a 0.000 0.478a 0.000 0.565a 0.000 Ebur Correlation is signiﬁcant at the 0.01 level (two-tailed). TABLE 5. Ordinal regression analyses with cartilage score as the dependent variable and marginal osteophytes, pitting, bony contour change, and eburnation scores as the independent factors, with Score 3 as the reference Variable Estimate Std Error Sig. Marost 0 Marost 1 Marost 2 Pitting 0 Pitting 1 Pitting 2 Bonycont 0 Bonycont 1 Bonycont 2 Ebur 0 Ebur 0 Ebur 0 27.433 25.562 24.387 24.571 22.811 20.194 25.505 23.588 21.706 23.468 21.459 0.616 1.176 1.143 1.132 1.009 0.990 0.997 0.845 0.729 0.724 0.709 0.725 0.864 0.000 0.000 0.000 0.000 0.005 0.845 0.000 0.000 0.180 0.000 0.044 0.476 95% conﬁdence interval 29.748 27.802 26.606 26.549 24.751 22.149 27.162 25.018 23.125 24.857 22.880 1.078 to to to to to to to to to to to to 25.137 23.321 22.168 22.593 20.871 1.760 23.848 22.158 20.287 22.079 20.037 2.310 esses and the laminae below (Dunlop et al., 1984). Such load-transmission would largely by-pass the articular surfaces, and may explain why high neural arch loadbearing in the present study was marginally more closely associated with degenerative changes in the bone than cartilage. Another factor involved in the development of DJD is age. Previous studies using a large number of specimens dispersed over a wide age range (19– 92years) showed that the compressive force resisted by the apophyseal joints increased signiﬁcantly with age (P \ 0.005) (Pollintine et al., 2004a). However, there was a large variation in the force resisted by the apophyseal joints at any particular age (for example, at 79 years the force resisted varied between 20 and 65%). The lack of correlation between apophyseal joint loading and age found in the present study may be due to the narrower age range, combined with this variation in loading at any particular age. However, Table 3 shows that while age and load-bearing by the apophyseal joints are not signiﬁcantly correlated, markers of joint degeneration such as cartilage loss and bone change clearly are. It is possible that, with advancing years, senescent chondrocytes are less likely to activate so that cartilage damage is more likely to accumulate (Martin and Buckwalter, 2003), subsequently causing lesions and other changes in the underlying bone. Three categories of bone change— pitting, the presence of marginal osteophytes, and eburnation—appear to be particularly good predictors of this cartilage loss. Fig. 3. Distribution of apophyseal joint degeneration with spinal level for the whole data set. Each box shows the median, quartiles (25–75%), and extreme values (range). The results of the present study probably depend on the manner in which mechanical loading was applied to the cadaveric specimens. Each motion segment was creep-loaded to reduce disc water content and height by an amount similar to the normal diurnal variation seen in life (Botsford et al., 1994), and this height loss is known to increase compressive load-bearing by the apophyseal joints (Dunlop et al., 1984) Also, specimens were tested in 28 of extension to simulate the erect standing posture in which the ‘‘normal’’ lordosis (i.e. the lordosis of an excised unloaded cadaveric lumbar spine) is slightly increased (Adams et al., 1988). In moderately ﬂexed postures, the apophyseal joints resist very little compressive force, even when the discs are narrowed (Adams et al., 2006). In this way, the relationship between DJD and spinal loading can be seen to depend on posture, and the time of day. Load-sharing in the spine is sensitive to small variations in posture and disc height: for example, just 28 of backwards bending of a motion segment increases stress concentrations within the intervertebral discs by 16% (Adams et al., 2000), reduces pressure in the nucleus pulposus by 10% (Adams et al., 1994), and approximately doubles load-bearing by the apophyseal joints (Adams and Hutton, 1980). American Journal of Physical Anthropology 324 K.R. BROWN ET AL. The results of this present experiment suggest that DJD in the apophyseal joints of the elderly human thoracolumbar spine can be strongly suggestive of biomechanical environment in life. Even the observation of mild DJD in this region is a good indicator of cartilage loss on the articular surface, and therefore joint space narrowing, and increased interfacet forces. Observation of severe bone change in the spine of an archaeological skeleton of similar age at death to those included in this study may suggest that the apophyseal joints were bearing much of the compressive load on the vertebra, probably in a lordotic posture. This situation is known to arise when intervertebral disc degeneration is severe, and it is associated with a reduction in bone density in the anterior vertebral body, and high bone density in the apophyseal joints relative to the vertebral body (Swanepoel et al., 1995; Fugiwara et al., 1999; Gries et al., 2000; Adams et al., 2006; Tischer et al., 2006). Other factors, such as variations in apophyseal joint morphology within and between spines may also be important. Together, these results indicate a change in vertebral function and show how bone (in particular) responds to a changing mechanical environment. This study has demonstrated how pathological and biomechanical data can be combined to reveal new insights into a skeletal condition of interest to palaeopathology. This approach is relatively uncommon, as comparative cadaveric or clinical studies involving collaborative efforts across discipline boundaries may be difﬁcult to establish. It is also the case that for some aspects of palaeopathological study clinical comparison may not hold the key to archaeological interpretation (Roberts et al., 1998). For studies of DJD, however, such an approach provides an essential grounding (Rogers et al., 1987; Dieppe et al., 1997; Pearson and Lieberman, 2004; Roberts et al., 2006). Not only is the relationship between cartilage and bone change exposed, but also the experimental framework permits testing of speciﬁc biomechanical hypotheses related to human movement, posture, and activity widely discussed in more general terms (e.g. Knüsel et al., 1997; Plochoki, 2002; Robson Brown et al., 2002; Van der Merwe et al., 2006; Mays, 2006). This may be of signiﬁcance to the study of human remains from archaeological contexts in a number of ways. It provides support for the palaeopathological investigation of sex or age related differences in the participation of activities that involve signiﬁcant compressive loading on the vertebrae (e.g. Merbs, 1983; Van der Merwe et al., 2006), both within populations and between them (Bennike, 1985; Kramar et al., 1990; Bridges, 1994; Knüsel et al., 1997). Similarly, there may also be implications for the assessment of vertebral pathology in extinct hominin populations. For example, degenerative changes have been observed on some of the Hadar australopithecine vertebrae, particularly in the thoracolumbar region (Robinson, 1972; Cook et al., 1983). The relatively large size of the neural arch in australopithecines, coupled with the patterning of DJD described, could be an indication that the apophyseal joints played a signiﬁcant role in resisting compressive forces acting on the australopithecine thoracolumbar spine, which in turn suggests intervertebral disc degeneration or lordotic posture, or both (Sanders, 1998; Whitcome et al., 2007). This possibility requires further investigation, but it is encouraging that while it may not be possible to isolate a particular behavior or activity that most stressed peoples’ bodies in life (Jurmain, 1990), the American Journal of Physical Anthropology comparative, experimental approach presented here can nevertheless yield insights into human biomechanics in archaeological populations. CONCLUSIONS Previous work has shown that human apophyseal joints have a variable load-bearing function, which increases following degeneration and narrowing of the intervertebral discs, and in lordotic postures. The present study shows that in elderly individuals, when loadbearing exceeds 50% of the compressive force acting on the spine, then degenerative changes are to be expected in the apophyseal joints, particularly in the subchondral bone. This suggests that load transmission is largely extra-articular, between the tips of the inferior articular processes and the laminae. The variables describing bone change (pitting, marginal osteophytes, bony contour change and eburnation) are positively and signiﬁcantly correlated with degree of cartilage loss, and therefore may be seen as good palaeopathological predictors of such soft tissue degeneration in elderly individuals. ACKNOWLEDGMENTS Many thanks to Chris Knüsel and the anonymous reviewers for their thoughtful comments on previous drafts of this article. LITERATURE CITED Adams MA, Hutton WC. 1980. The effect of posture on the role of the apophyseal joints in resisting intervertebral compressive forces. J Bone Joint Surg (Br) 62:358–362. Adams MA, Hutton WC. 1981. The relevance of torsion to the mechanical derangement of the lumbar spine. Spine 6:241– 248. Adams MA, Dolan P, Hutton WC. 1986. The stages of disc degeneration as revealed by discograms. J Bone Joint Surg (Br) 68:36–41. Adams MA, Dolan P, Hutton WC. 1988. The lumbar spine in backward bending. Spine 13:1019–1026. Adams MA, May S, Freeman BJ, Morrison HP, Dolan P. 2000. Effects of backward bending of lumbar intervertebral discs. Relevance to physical therapy treatments for low back pain. Spine 25:41–47. Adams MA, McNally DS, Chinn H, Dolan P. 1994. Posture and the compressive strength of the lumbar spine. Clin Biomech 9:5–14. Adams MA, McNally DS, Dolan P. 1996. ‘‘Stress’’ distributions inside intervertebral discs. The effects of age and degeneration. J Bone Joint Surg 78:965–972. Adams MA, Pollintine P, Tobias JH, Wakley GK, Dolan P. 2006. Intervertebral disc degeneration can predispose to anterior vertebral fractures in the thoracolumbar spine. J Bone Miner Res 21:1409–1416. Bennike P. 1985. Paleopathology of Danish skeletons. A comparative study of demography, disease and injury. Copenhagen: Akademisk Forlag. Botsford DJ, Esses SI, Ogilvie-Harris DJ. 1994. In vivo diurnal variation in intervertebral disc volume and morphology. Spine 19:95–940. Bridges PS. 1991. Degenerative joint disease in hunter-gatherers and agriculturalists of the southeastern United States. Am J Phys Anthropol 85:379–391. Bridges PS. 1994. Vertebral arthritis and physical activities in the prehistoric southeastern United States. Am J Phys Anthropol 93:83–93. Butler D, Traﬁmov JH, Andersson GB, McNeill TW, Huckman MS. 1990. Discs degenerate before facts. Spine 15:111–113. VERTEBRAL FACET JOINT DEGENERATION Cholewicki J, Crisco JJ III, Oxland TR, Yamamoto I, Panjabi MM. 1996. Effects of posture and structure on three-dimensional coupled rotations in the lumbar spine. A biomechanical analysis. Spine 21:2421–2428. Cholewicki J, Panjabi MM, Khachatryan A. 1997. Stabilizing function of trunk ﬂexor-extensor muscles around a neutral spine posture. Spine 22:2207–2212. Cleveland WS. 1979. Robust locally weighted regression and smoothing scatterplots. J Am Stat Assoc 74:829–836. Conaghan PG. 2002. Update on osteoarthritis part 1: current concepts and the relation to exercise. Br J Sports Med 36: 330–333. Cook DC, Buikstra JE, DeRousseau CJ, Johanson DC. 1983. Vertebral pathology in the Afar australopithecines. Am J Phys Anthropol 60:83–102. Cyron BM, Hutton WC. 1980. Articular tropism and stability of the lumbar spine. Spine 5:168–172. Davis PR. 1961. Human lower lumbar vertebrae; some mechanical and osteological considerations. J Anat 95:337–344. Dieppe PA, Cushnagan J, Shepstone L. 1997. The Bristol ‘‘OA500’’ Study: progression of osteoarthritis (OA) over 3 years and the relationship between clinical and radiographic changes at the knee joint. Osteoarthritis Cartilage 5:87–97. Dolan P, Earley M, Adams MA. 1994. Bending and compressive stresses acting on the lumbar spine during lifting activities. J Biomech 27:1237–1248. Dunlop RB, Adams MA, Hutton WC. 1984. Disc space narrowing and the lumbar facet joints. J Bone Joint Surg (Br) 66:706–710. el-Bohy AA, Yang KH, King AI. 1989. Experimental veriﬁcation of facet load transmission by direct measurement of facet lamina contact pressure. J Biomech 22:931–941. Fugiwara A, Tamai K, Yamato M. 1999. The relationship between facet joint osteoarthritis and disc degeneration of the lumbar spine: an MRI study. Eur Spine J 8:396–401. Gerszten PC, Gerszten E, Allison MJ. 2001. Diseases of the spine in South American mummies. Neurosurgery 48:208– 213. Gries NC, Berlemann U, Moore RJ, Vernon-Roberts B. 2000. Early histological changes in lower lumbar discs and facet joints and their correlation. Eur Spine J 9:23–29. Haefeli MF, Kalberer F, Saegesser D, Nerlich AG, Boos N, Paesold G. 2006. The course of macroscopic degeneration in the human lumbar intervertebral disc. Spine 31:1522–1531. Howell DC. 1996. Statistical methods for psychology, 3rd ed. Belmont California: Duxbury Press. Jurmain RD. 1977. Stress and the etiology of osteoarthritis. Am J Phys Anthropol 46:353–366. Jurmain RD. 1990. Paleoepidemiology of a central California prehistoric population from CA-ALA-329: II. Degenerative disease. Am J Phys Anthropol 83:83–94. Jurmain RD. 1991. Degenerative changes in peripheral joints as indicators of mechanical stress: opportunities and limitations. Int J Osteoarchaeol 1:247–252. Jurmain RD. 2000. Degenerative joint disease in African great apes: an evolutionary perspective. J Hum Evol 39:185–203. Kellgren JH, Lawrence JS. 1958. Osteo-arthrosis and disk degeneration in an urban population. Ann Rheum Dis 17: 388–397. Knüsel CJ, Göggel S, Lucy D. 1997. Comparative degenerative joint disease of the vertebral column in the medieval monastic cemetery of the Gilbertine Priory and St. Andrew, Fishergate, York, England. Am J Phys Anthropol 103:481–495. Kramar C, Lagier R, Baud CA. 1990. Rheumatic diseases in Neolithic and medieval populations of western Switzerland. Zeitschrift f€ vr Rheumatol 49:338–345. Lane NE, Bloch DA, Jones HH, Marshall WH, Wood PD, Fries JF. 1986. Long-distance running, bone density and osteoarthritis. J Am Med Assoc 255:1147–1151. Lequesne MG, Dang N, Lane NE. 1997. Sport practice and osteoarthritis of the limbs. Osteoarthritis Cartilage 5:75– 86. Lewin T. 1964. Osteoarthritis in lumbar synovial joints. A morphologic study. Acta Orthop Scand Suppl 73:1–112. 325 Lovell NC. 1994. Spinal arthritis and physical stress at Bronze Age Harappa. Am J Phys Anthropol 93:149–164. Maat GJR, Mastwijk RW, van der Velde E. 1995. Skeletal distribution of degenerative changes in vertebral osteophytosis, vertebral osteoarthritis and DISH. Int J Osteoarchaeol 5:289–298. Martin JA, Buckwalter JA. 2003. The role of chondrocyte senescence in the pathogenesis of osteoarthritis and in limiting cartilage repair. J Bone Joint Surg Am 85A (Suppl 2):106–110. Mays S. 2006. Spondylolysis, spondylolisthesis, and lumbo-sacral morphology in a medieval English skeletal population. Am J Phys Anthropol 131:352–362. McMillan DW, McNally DS, Garbutt G, Adams MA. 1996. Effect of sustained loading on the water content of intervertebral discs: implications for disc metabolism. Ann Rheum Dis 55:880–887. Meachim G. 1972. Light microscopy of Indian ink preparations of ﬁbrillated cartilage. Ann Rheum Dis 31:457–464. Merbs CF. 1983. Patterns of activity-induced pathology in a Canadian Inuit population. Ottawa: Archaeological Survey of Canada. National Museum of Man, Mercury Series 119. Meulenbelt I, Kloppenburg M, Kroon HM, Houwing-Duistermaat JJ, Garnero P, Hellio Le Graverand MP, Degroot J, Slagboom PE. 2006. Urinary CTX-II levels are associated with radiographic subtypes of osteoarthritis in hip, knee, hand, and facet joints in subject with familial osteoarthritis at multiple sites: the GARP study. Ann Rheum Dis 65:360–365. Nachemson AL. 1981. Disc pressure measurements. Spine 6:93– 97. Nathan H. 1962. Osteophytes of the vertebral column. J Bone Joint Surg 44:243–268. Norus̆is, MJ. 2005. SPSS 13.0 guide to data analysis. Englewood Cliffs: Prentice Hall. Pearson OM, Lieberman DE. 2004. The aging of Wolff ’s Law: ontogeny and responses to mechanical loading in cortical bone. Am J Phys Anthropol 125:63–99. Plochoki JH. 2002. Directional bilateral asymmetry in human sacral morphology. Int J Osteoarchaeol 12:349–355. Pollintine P, Przybyla AS, Dolan P, Adams MA. 2004a. Neural arch load-bearing in old and degenerated spines. J Biomech 37:197–204. Pollintine P, Dolan P, Tobias JH, Adams MA. 2004b. Intervertebral disc degeneration can lead to ‘‘stress-shielding’’ of the anterior vertebral body: a cause of osteoporotic vertebral fracture? Spine 29:774–782. Putz R. 1985. The functional morphology of the superior articular processes of the lumbar vertebrae. J Anat 143:181–187. Resnick D, Niwayama G. 1988. Diagnosis of bone and joint disorders, 2nd ed. Washington DC: WB Saunders. Roberts CA, Manchester KM. 1995. The archaeology of disease. Gloucester: Cornell University Press. Roberts CA, Boylston A, Buckley L, Chamberlain AC, Murphy EM. 1998. Rib lesions and tuberculosis: the paleopathological evidence. Tubercle Lung Dis 79:55–60. Roberts AM, Robson Brown K, Musgrave JH, Leslie I. 2006. A case of bilateral scapholunate advanced collapse in a RomanoBritish skeleton from Ancaster. Int J Osteoarchaeol 16:208– 220. Robinson JT. 1972. Early hominid posture and locomotion. Chicago: University of Chicago Press. Robson Brown KA, Davies EN, McNally DS. 2002. The angular distribution of vertebral trabeculae in modern humans, chimpanzees and the Kebara 2 Neanderthal. J Hum Evol 43:189– 205. Rogers J, Waldron T, Dieppe P, Watt I. 1987. Arthropathies in palaeopathology: the basis of classiﬁcation according to most probably cause. J Archaeol Sci 14:179–193. Sager P. 1969. Spondylosis cervicalis. Copenhagen: Munksgaard. Sanders WJ. 1998. Comparative morphometric study of the australopithecine vertebral series Stw-H8/H41. J Hum Evol 34:249–302. Sato K, Kikuchi S, Yonezawa T. 1999. In vivo intradiscal pressure measurement in healthy individuals and in patients with ongoing back problems. Spine 24:2468–2474. American Journal of Physical Anthropology 326 K.R. BROWN ET AL. Sofaer Derevenski JR. 2000. Sex differences in activity-related osseous change in the spine and the gendered division of labor at Ensay and Wharram Percy, UK. Am J Phys Anthropol 111:333–354. Swanepoel MW, Adams LM, Smeathers JE. 1995. Human lumbar apophyseal joint damage and intervertebral disc degeneration. Ann Rheum Dis 54:182–188. Tischer T, Aktas T, Milz S, Putz RV. 2006. Detailed pathological changes of human lumbar facet joints L1-L5 in elderly individuals. Eur Spine J 15:308–315. Twomey LT, Taylor JR, Taylor MM. 1989. Unsuspected damage to lumbar zygapophyseal (facet) joints after motor-vehicle accidents. Med J Aust 151:210–212, 215–217. Van der Merwe AE, I scan MY, L’Abbè EN. 2006. The pattern of vertebral osteophyte development in a South African population. Int J Osteoarchaeol 16:459–464. Videman T, Nurmine M, Troup JD. 1990. 1990 Volvo Award in clinical sciences. Lumbar spinal pathology in cadaveric mate- American Journal of Physical Anthropology rial in relation to history of back pain, occupation, and physical loading. Spine 15:728–740. Waldron T. 1991. The prevalence of, and the relationship between, some spinal diseases in a human skeletal population from London. Int J Osteoarchaeol 1:103–110. Waldron T. 1992. Osteoarthritis in a Black Death cemetery in London. Int J Osteoarchaeol 2:235–240. Waldron T. 1994. Counting the dead: the epidemiology of skeletal populations. Chichester: Wiley. Weber J, Czarnetzki A. 2002. Paleopathology of the lumbar spine in the early medieval period. Z Orthop Grenzgeb 140: 637–643. Weber J, Spring A. 2004. Degenerative changes of the cervical facet joints in the medial and lateral atlantoaxial joint—a paleopathological study. Z Orthop Grenzgeb 142:184–187. Whitcome KK, Shapiro LJ, Lieberman DE. 2007. Fetal load and the evolution of lumbar lordosis in bipedal hominins. Nature 450:1075–1080.