Rheumatology 2002;41:375–380 Sacral and iliac articular cartilage thickness and cellularity: relationship to subchondral bone end-plate thickness and cancellous bone density G. J. McLauchlan* and D. L. Gardner1 Department of Orthopaedic Surgery, Princess Margaret Rose Hospital, Frogston Road West, Edinburgh EH10 7ED and1Department of Pathology, University Medical School, Teviot Place, Edinburgh EH8 9AG, UK Abstract Objectives. To measure the thickness and cellularity of adult human sacral and iliac articular cartilages and the thickness and density of the subchondral bones. Methods. The right sacroiliac joints of 15 adult patients were examined post-mortem. HOME (Highly Optimized Microscope Environment) microscopy was used to measure articular cartilage and subchondral bone end-plate thickness. Conventional morphometric techniques were employed to estimate cartilage cellularity and cancellous bone density. Results. Sacral articular cartilage was thicker than iliac (1.81 vs 0.80 mm, P<0.001). Iliac cartilage cell density in all zones was higher than sacral. The overall mean was 31.19 3 1023 vs 23.23 3 1023umm3, P<0.001. Superficial zones contained more cells than middle and deep zones but there were large differences between the cell numbers of the middle and deep zones of both sacral and iliac cartilages. Iliac subchondral bone end-plates were thicker than sacral (0.36 vs 0.23 mm, P<0.001). The thickness of these plates was related inversely to that of the overlying articular cartilages. Iliac subchondral cancellous bone was twice as dense as sacral (22.07 vs 12.05%, P<0.001), a ratio recognized anteriorly, centrally and posteriorly. Conclusions. Adult human sacral cartilage is thick and of low cell density. It rests upon a thin bone end-plate supported by porous, cancellous bone. Iliac cartilage and bone display the converse proportions. The identification of these variables may assist understanding of normal sacroiliac joint function and the interpretation of tissue changes in the spondylarthropathies. KEY WORDS: Sacroiliac joint, Cartilage, Bone, Morphometry, HOME microscopy, Laser scanning microscopy. They have been reported on several occasions w5–9x. Conspicuous differences between the thicknesses of the sacral and iliac cartilages have been recorded w5–11x. By contrast, there have been relatively few microscopic studies of normal SI joints w12x. As one consequence, records of the cellularity of the sacral and iliac articular cartilages w13x and of the thickness and density of sub- chondral sacral and iliac bones are rare. The present studies were made to remedy this deficiency. Some of the results have been presented in preliminary form w14, 15x. Disorders of the non-ligamentous, synovial components of the sacroiliac (SI) joints characterize the spondyl- arthropathies and are pathognomonic of ankylosing spondylitis (AS) w1–3x. Definition of these defects, essential if the spondylarthropathies are to be fully understood, is proceeding systematically w4x. However, progress has been handicapped, not only by the rarity of material from early cases and by the inaccessibility of the SI joints, but also by the paucity of measurements of the size and cellularity of the components of the normal joints. Measurements of the circumference, area and dia- meter of the bearing surfaces of normal human sacro- iliac articulations can be made readily by the naked eye. Materials and methods Right sacroiliac joints were collected at autopsy from 15 adult patients (mean age 51.8"14.7 yr, range 23–83) dying in hospital with no gross evidence of osteo- cartilaginous disease. There were eight males (mean age 54.5 yr) and seven females (mean age 48.7 yr). Submitted 14 May 2001; revised version accepted 10 October 2001. Correspondence to: D. L. Gardner. *Present address: Orthopaedic and Rheumatology Directorate, Chorley and South Ribble District General Hospital, Preston Road, Chorley PR7 1PP, UK. ? 2002 British Society for Rheumatology 375
G. J. McLauchlan and D. L. Gardner 376 Dissection The upper part of each left femur was divided trans- versely. The right innominate bone was transected 30 mm lateral to the right sacroiliac joint and the pelvis was excised by cutting through the L3–L4 interface. The vertebral column was divided sagittally. Four to 11 slab blocks were cut at right angles to this plane of section, the number varying with the size of the individual. The blocks were fixed in buffered formalin at pH 7.0 before decalcification in EDTA (ethylene- diamine tetraacetic acid). From each block, semiserial sections were prepared and stained differentially with haematoxylin and eosin, toluidine blue, picro-Sirius red, phosphotungstic aciduhaematoxylin, Martius scarlet blue and alkaline Congo red. The mean anteroposterior length of the 15 SI joints was 23.40"11.72 mm. Bone thickness Sacral and iliac bone end-plate thickness was measured at the 11 points selected for assessing cartilage thickness. Sites were chosen where marrow cavities lay closest to the deepest aspects of the articular cartilage. The areas delineated by the tide-line and by the bone margin were included in the counting procedure. A total of 1452 bone thickness measurements was recorded and expressed in millimetres. Bone density A Weibel eyepiece graticule engraved with 42 lines w16x was used within a Leitz Orthoplan microscope. The graticule was positioned so that its inner edge lay immediately beyond the outermost margin of the sub- chondral bone end-plate. Care was taken to avoid dense, cortical bone. In each of 69 sections, linear intercept counts were made of sacral and iliac bone. Fields encompassed the anterior, middle and posterior parts of the subarticular, cancellous bone. A total of 414 fields was therefore counted. The results were expressed as the mean percentage of the fields occupied by bone trabeculae. Cartilage thickness The thickness of the sacral and iliac articular cartilages was measured with a Zeiss Highly Optimised Micro- scope Environment (HOME) system in 66 sections. Cartilage thickness (the distance from the bearing surface to the chondro-osseous junction) was recorded at 11 equidistant points along the anteroposterior axis of the sacral and iliac cartilages, giving a total of 1452 measurements.Integrationofthemeasurementsatpoints 1, 2 and 3; at 5, 6 and 7; and at 9, 10 and 11 allowed direct comparison with those of cartilage thickness. Statistics There was an asymmetrical, non-normal distribution of data (Fig. 1). The Mann–Whitney procedure was there- fore used, within Minitab v. 12, to test the significance of differences between samples. In searching for the possible influences of age, sex and location, regression analysis was employed to take account of unequal replication and missing values. Statistical significance was accepted at the 5% level. Section thickness Section thickness was measured with a Zeiss LSM laser confocal scanning microscope and 340 water- immersion objective lens. The precision of the micro- scope measurements was tested against a Mitutoyo metal block standard. Measurements were made in the z-axis mode, using an argon ion laser. Brightness (blue) and contrast (red) were balanced. Observer error was assessed by repetitive measurements at single sites. From each case, slides representing five different staining techniques were taken. A total of 750 determinations was made. The mean thickness of 75 sections was 11.002"4.38 mm (SD). Cell numbers The numbers of cells, nuclei and empty chondrocyte lacunae were recorded by conventional light microscope and eyepiece graticule in 62 sections of sacral and iliac cartilage. The number of empty lacunae (on average 0.3 per field) had no influence on any analysis and was discarded. Differences between nuclear and cellular counts were minimal. Only counts of whole cell numbers were therefore considered. Sections were divided into three equal superficial, middle and deep cartilage zones and subdivided into nine equal anteroposterior segments. Cell counts were made in each of these areas. From these 27 areas, 3348 measurements were therefore obtained. The results were adjusted for section thickness and expressed as cells 3 1023umm3. FIG. 1. Chondrocyte counts as cells 3 1023umm3. (a) Sacral cartilage zones. 1, superficial; 2, middle; 3, deep. (b) Iliac cartilage zones. 1, superficial; 2, middle; 3, deep.
Morphometry of sacroiliac cartilage and bone 377 cartilage (1.92 mm) to be thicker than male (1.71 mm) but this difference, and that between male and female iliac cartilage thicknesses, was not significant. Results Cartilage thickness Sacral cartilage was thicker than iliac (Table 1; Fig. 2). The thickness of sacral cartilage was related inversely to that of sacral bone (P=0.005) and the thickness of iliac cartilage was related inversely to that of iliac bone (P=0.017). There was a trend for female sacral Chondrocyte numbers Iliac cartilage was more cellular than sacral (Table 1; Fig. 3). Cell numbers of superficial zones of both sacral and iliac cartilage were greater than those of the FIG. 2. Iliac and sacral cartilage and bone. Relatively thin iliac cartilage (left) is rich in perpendicular and oblique collagen fibre bundles. Joint space is seen (centre). Thicker sacral cartilage (right) closely resembles hyaline cartilage of limb joints. Relatively thick iliac bone is separated from bone marrow by artefactual space, the result of tissue shrinkage at fixation. Scale bar=0.5 mm. FIG. 3. Iliac and sacral midzone chondrocytes. Relatively cellular iliac cartilage (left) contrasts with less cellular sacral cartilage (right). Many iliac cells are assembled in groups, within chondrons. Cartilage matrix is interspersed perpendicularly with swathes of collagen fibre bundles that extend from the bearing surface to the chondro-osseous margin. Sacral chondrocytes, arranged singly, in pairs or larger groups, are widely separated by a homogeneous matrix, richer in proteoglycan than iliac cartilage. Scale bar=30 mm.
G. J. McLauchlan and D. L. Gardner 378 TABLE 1. Cartilage thickness, numbers of chondrocytes, bone end-plate thickness and subchondral bone density of sacroiliac joints Significance of difference 95% confidence interval Sacral Iliac Cartilage thickness (mm; mean" S.D.) Total Male Female Chondrocyte number (cells 3 1023umm3; mean" S.D.) Total Superficial Middle Deep Bone end-plate thickness (mm" S.D.) Total Male Female Mean trabecular bone volume (%" S.D.) Total Anterior Middle Posterior 1.81"0.56 1.71"0.52 1.92"0.58 0.80"0.21 0.81"0.20 0.79"0.22 0.90, 21.09 0.85, 0.98 1.00, 1.14 P<0.001 P<0.001 P<0.001 23.23"12.56 33.05"14.09 18.30"7.90 18.55"8.36 31.19"17.71 36.08"19.51 27.72"15.26 29.74"17.04 25.14, 27.70 23.85, 0.001 28.98, 26.41 210.27, 27.70 P<0.001 P<0.013 P<0.001 P<0.001 0.23"0.14 0.23"0.14 0.24"0.26 0.36"0.25 0.37"0.27 0.35"0.29 20.12, 20.07 20.11, 20.07 20.11, 20.06 P<0.001 P<0.001 P<0.001 12.05"6.96 13.36"6.97 11.22"6.31 11.45"7.44 22.07"13.65 23.46"13.57 21.39"12.45 21.36"14.90 29.52, 25.47 213.09, 24.76 211.90, 24.76 29.52, 23.09 P<0.001 P<0.001 P<0.001 P<0.001 corresponding middle and deep zones. The cellularity of each of the three iliac cartilage zones (superficial, middle and deep) was greater than that of the corresponding sacral zone. Cell density declined from the superficial to the deep zone. However, the degree of this difference varied. There was a smaller reduction in cellularity from the surface downwards in iliac cartilage than in sacral cartilage. There was no demonstrable relationship between age or the anteroposterior or superoinferior position of the sample within the joint and the cellularity of sacral or iliac cartilage. sacral cartilage. This difference is largely accounted for by the distinctive populations of the middle and deep zones. The high cell densities of the superficial zones of both cartilages are closely similar. There are further characteristics. The subchondral iliac bone end-plate is 50% thicker than the subchondral sacral bone end-plate. Moreover, the density of the cancellous bone subjacent to the iliac cartilage is nearly twice as great as that of the cancellous bone subjacent to the sacral cartilage. Understanding of the structure and function of the human SI joints has progressed slowly, over many years. Meckel w17x first described the articulations in 1816. Although their diarthrodial nature was claimed by Luschka in 1854 w18x, their synovial character remained contentious until modern times. The unusual shape and size of the paired joints was recognized more than a century ago w19x. Early workers recorded the most obvious anatomical features of the joints: their oblique position, the boomerang shape of the load-bearing sur- faces, and the distinct appearances and unequal thick- ness of the sacral and iliac bearing surfaces. Inequalities of sacral and iliac cartilage thickness were detected early in life and the contrasting dimensions of embryonic and infantile sacral and iliac cartilages were established w8, 12x. Age-related structural divergences persisted and increased. In children and young adults, the sacral cartilage was 4–6 mm in thickness; it was greater anteriorly than posteriorly w20x. By contrast, the iliac cartilage was no more than 2 mm in thickness. This ratio continued into the second decade w21x, during which sacral and iliac measurements of 2–3 and 1.5 mm w22x or 1–4 and 0.5–2 mm w20x were described, respect- ively. By 60–69 yr of age, these figures became 2–3 and <1.0 mm respectively, while in those aged more than 70 yr they were 1.5 and 0.5 mm w22x. The thicker sacral cartilage closely resembles that of the large limb joints w23x. The thinner iliac cartilage is quite different. It is less rich in proteoglycan and is Bone end-plate thickness End-plate iliac bone was thicker than end-plate sacral bone (Table 1). Male iliac bone end-plates were thicker than female iliac end-plates. There was no demonstrable relationship between end-plate bone thickness and age or between the anteroposterior or superoinferior posi- tion of the sample within the joint and this variable. An inverse relationship between bone end-plate thick- ness and overlying cartilage thickness was observed consistently. Bone density Subchondral, iliac cancellous bone density was greater than sacral (Table 1). This relationship was confirmed at each of the anterior, central and posterior parts of the joint. However, differences between the three parts of the individual sacral and iliac bones did not reach acceptable levels of significance. Discussion The evidence presented in this paper confirms that human sacral articular cartilage is twice as thick as iliac. Overall, the relatively thin iliac cartilage contains many more cells per unit volume than the relatively thick
Morphometry of sacroiliac cartilage and bone 379 and that of a 1-mm slice of knee joint ;300 mm, the resolution of a 1-mm slice of hip joint may be no more than ;1500 mm. This constraint contrasts with the much higher resolution of the light microscope lenses used in the present study, in which, in theory, the physical limit is that of the wavelength of the light source. This limit is ;300 nm, the diameter of a staphylococcus. To a small but encouraging extent, it has been shown that the horizons of conventional MRI can be extended but, so far, only in the laboratory. Xia et al. w31x have demonstrated that, by scaling down the receiver coil of a Bruker AMX 300 NMR (nuclear magnetic resonance) microscope equipped with a 7-teslau89-mm vertical-bore superconducting magnet and microimaging accessory, they can obtain a pixel resolution of 10 mm. Xia et al. describe the technique variously as NMR microscopy, microscopic MRI or mMRI. Using this technique, they have made an in vitro analysis of canine humeral cartilage. Comparing their mMRI results with those of polarized light microscopy, they have shown that articular cartilage zones are statistically equivalent to cartilage collagen fibre orientation assessed by polarized light microscopy. To what extent the possibilities raised by this investigation can be applied clinically is a matter for further research. The SI joints play a crucial role in the pathogenesis of the spondyloarthropathies, in particular of AS, and it is to be hoped that the methods of morphometry may contribute to further understanding of the micro- scopic properties of the joint components. In an invest- igation of the structural changes in AS w4x, active, peripheral and central cartilage destruction and severe basal cartilage injury were recognized with greater frequency in affected patients than in control subjects. Although observed and graded subjectively, the car- tilage changes in the diseased tissues could not be measured adequately. There were two reasons. Only three patients with AS had a disease duration of less than 3 yr and many other abnormalities, including in particular chondroid fusion, made attempts at cartilage morphometry technically difficult. Full understanding of the structure and function of the normal SI joints is still incomplete. For example, there is only a single account of the chemical composi- tion of the sacral and iliac cartilages w10x. The sacral and iliac moieties are parts of a complex shock-absorbing system that protects the mammalian brain and spinal cord against diurnal static and impact forces. The con- gruent nature of the articulation, with each opposed bearing surface wholly in contact with the other, suggests strongly that forces derived from a sacral direction are transmitted equally and uniformly to the iliac cartilage surfaces. The evidence presented in this report highlights the paradoxical nature of SI joint structure: if the forces applied to both bearing surfaces are identical, why is their structure so different? One explanation for this enigma may lie in postulating differential responses by the sacrum and ilium. The open, porous, bone of the former contrasts with the immensely strong, dense bone interspersed with dense, tangential and perpendicular collagen bundles that often extend to the bone end- plate. The differential features of male and female joints have been documented w24x, as have those of some racial and ethnic groups w25x. Generally, however, evidence of anatomical distinctions between male and female SI joints is still inconclusive. In the present study, the influence of sex on cartilage thickness could not be substantiated, possibly because the numbers of cases studied was small. Because of incomplete data, it has not been possible to apply correction factors for body mass or stature. Conventional radiology has been used extensively in the clinical analysis of diseased SI joints w26, 27x but reveals limited structural detail and then only long after the onset of symptoms. The advent of non-invasive magnetic resonance imaging (MRI) w28, 29x has offered new scope for early diagnosis and has made it increas- ingly possible to correlate structure and function. Braun et al. w30x have employed dynamic MRI with fast imaging. The technique offers a clear advantage over CT scanning in the evaluation of cartilage disease, although the precise measurement of joint space and cartilage thickness has proved difficult. In a MRI study of 114 children with no clinical evidence of sacroiliitis, Bollow et al. w28x have shown that the width of the adolescent SI joint is, on average, 4 mm, a figure closely similar to that obtained in adult joints. In this investigation, the sacral cartilage is reported as approximately 3 mm in thickness, the iliac cartilage approximately 1 mm. Braun et al. w30x are careful to emphasize that, in the clinic, confirmatory morphological evidence of phenom- ena observed by MRI is not yet available; there is a lack of direct correlation between histological and scanning results and no gold standard w28x. There are several reasons for this deficiency. First, very few controlled microscopic analyses of sacroiliac joint structure have been made w4x. Secondly, it is difficult to obtain the necessary post-mortem material. The SI joints cannot easily be examined directly in the living patient. Much of the evidence rests upon autopsy enquiries. Thirdly, although many needle biopsy investigations of SI joint synovia have been made, sacroiliac joint surgery is uncommon and cartilaginous tissue is rarely sampled. Fourthly, it is unethical to undertake contrast-enhanced MRI studies of the sacroiliac joints in entirely normal individuals or even in patients with unrelated diseases, such as cancer w28x. A serious problem in the use of MRI for the meas- urement of SI joint cartilage thickness remains the limited resolution of MRI scanners. In the majority of cases, a 3 mm slice is required to yield a resolution of 1–1.5 mm. The resolution of scanners varies but is determined principally by the filling factor of the detector coil wL. Hall, personal communicationx. This factor is high in relation to small joints, such as those of a finger, and low in relation to large joints, such as those of the hip and pelvis. Whereas the best plane resolution in a 1-mm slice of distal interphalangeal joint is ;75 mm
G. J. McLauchlan and D. L. Gardner 380 joint. Part I. Clinical anatomical aspects. Spine 1990; 15:130–2. 12. Kampen WU, Tillmann B. Age-related changes in the articular cartilage of human sacroiliac joint. Anat Embryol 1998;198:505–13. 13. Gardner DL, Brereton JD, Hollinshead MS. Com- parative thickness, cell density and histochemistry of sacral and iliac cartilages in man. Br J Rheumatol 1985;24:90. 14. Gardner DL, McLauchlan GJ. Trials and tribulations of the sacroiliac joints. Clin Rheumatol 1998;17:439. 15. McLauchlan GJ, Gardner DL. Sacral and iliac cartilage and bone: thickness, proportions and cell numbers. Ann Rheum Dis 1999;(Suppl. Proc XIVth EULAR Congress):129. 16. Aherne WA, Dunnill MS. Edward Arnold, 1982:33–59. 17. Meckel JF. Handbuch der menschlichen Anatomie, Vol. II. Halle: Buchhandlung der 1816:354. 18. Luschka H von. Die Schambeinfuge des Menschen. Virchows Arch Pathol Anat Physiol 1854;7:299–324. 19. Meyer GH. Der Mechanismus der Symphysis sacro-iliaca. Arch Anat Physiol 1878;I:1–19. 20. Carter ME, Loewi G. Anatomical changes in normal sacro-iliac joints during childhood and comparison with the changes in Still’s disease. Ann Rheum Dis 1962; 21:121–34. 21. Ishimine T. Histopathological study of the aging process in the human sacroiliac joint. Nippon Seikeigeka Gakkai Zasshi 1989;63:1074–84. 22. Macdonald GR, Hunt TE. Sacroiliac joints. Observations on the gross and histological changes in the various age groups. Can Med Assoc J 1952;66:157–63. 23. Stockwell RA. Biology of cartilage cells. Cambridge: Cambridge University Press, 1979:75–8. 24. Salsabili N, Valojerdy MR, Hogg DA. Variations in thickness of articular cartilage in the human sacroiliac joint. Clin Anat 1995;8:388–90. 25. Stewart TD. Pathologic changes in aging sacroiliac joints. A study of dissecting-room skeletons. Clin Orthop 1984; 183:188–96. 26. Horwitz T, Smith RM. An anatomical, pathological and roentgenological study of the intervertebral joints of the lumbar spine and the sacroiliac joints. AJR Am J Roentgenol 1940;43:173–86. 27. Cohen AS, McNeill JM, Calkins E, Sharp JT, Schubart A. The ‘normal’ sacroiliac joint. Analysis of 88 sacroiliac roentgenograms. Clin Orthop 1967;100:559–63. 28. Bollow M, Braun J, Kanneburg J et al. Normal morphology of sacroiliac joints in children: magnetic resonance studies related to age and sex. Skeletal Radiol 1997;26:697–704. 29. Braun J, Bollow M, Sieper J. Radiologic diagnosis and pathology of the spondyloarthropathies. Rheum Dis Clin North Am 1998;24:697–735. 30. Braun J, Bollow M, Eggens U, Ko ¨nig H, Distler A, Sieper J. Use of dynamic magnetic resonance imaging with fast imaging in the detection of early and advanced sacroiliitis in spondylarthropathy Rheum 1994;37:1039–45. 31. Xia Y, Moody JB, Burton-Wurster Quantitative in situ correlation between microscopic MRI and polarized light microscopy studies of articular cartilage. Osteoarthritis Cart 2001;9:393–406. of the latter. The deformable sacrum can be viewed as a cushion in a sacroiliac shock-absorbing system, transmitting loads but not absorbing them. The non- deformable (elastic) ilium, by contrast, is responsible for absorbing the forces imposed by the weight of the upper half of the body in all upright andsemi-upright positions, and for transmitting these forces to the legs and feet. In conclusion, investigations of the SI joints should now be extended by taking advantage of continued advances in the techniques of imaging and microscopy. By these means, and by the investigation, in parallel, of non-human mammals, it should prove possible to ana- lyse the components of the SI joints in greater detail and to extend to their understanding new micromechanical, immunohistochemical, cytogenetic and molecular bio- logical techniques. The correlation of the biological properties of the extracellular matrix components of the SI joints with the selective distribution and nature of individual cells is likely to permit a fuller under- standing of the role of these structures in modulating compressive and shear forces at the junction between the vertebrae and the pelvis and in explaining their disorganization in diseases such as AS. Investigations to fulfil these aims are in progress. Morphometry. London: Hallischen Waisenhauser, Kreuzdarmbeinfuge und die Acknowledgements The Mr W. Adams of the Department of Public Health Sciences, University of acknowledged. expert statistical assistance and advice of Edinburgh, is gratefully References 1. Calin A, Taurog JD. The spondylarthritides. Oxford: Oxford University Press, 1999. 2. Franc ¸ois RJ. Some pathological features of anky- losing spondylitis as revealed by microradiography and tetracycline labelling. Clin Rheumatol 1982;1:23–9. 3. Moll JMH. Ankylosing spondylitis. Edinburgh: Churchill Livingstone, 1980. 4. Franc ¸ois RJ, Gardner DL, Degrave EJ, Bywaters EGL. Histopathologic evidence that sacroiliitis in ankylosing spondylitis is not merely enthesitis. Arthritis Rheum 2000; 43:2011–24. 5. Albee FH. A study of the anatomy and the clinical importance of the sacroiliac joint. J Am Med Assoc 1909; 53:1273–6. 6. Brooke R. The sacro-iliac joint. J Anat 1924;58:299–305. 7. Sashin D. A critical analysis of the anatomy and the pathologic changes of the sacro-iliac joints. J Bone Joint Surg Am 1930;12:891–910. 8. Schunke GB. The anatomy and development of the sacro- iliac joint in man. Anat Rec 1938;72:313–31. 9. Weisl H. The articular surfaces of the sacro-iliac joint and their relations to the movements of the sacrum. Acta Anat 1951;22:1–14. 10. Paquin JD, Rest M van der, Marie PJ et al. Biochemical and morphologic studies of cartilage from the adult human sacroiliac joint. Arthritis Rheum 1983;26:887–95. 11. Vleeming A, Stoeckart R, Volkers ACW, Snijders CJ. Relation between form and function in the sacroiliac patients. Arthritis N, Lust G.