SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements

The material properties of cancellous bone from patients with osteoporosis (OP) or osteoarthritis (OA) were determined and compared with normal controls. Samples were selected from defined sites in human femoral heads which are subjected to different loads in vivo. Overall, OP bone had the lowest stiffness and OA the highest, and this same order was reflected in the apparent densities of the bone, with OA being the most dense and OP the least. Normal and OP bone were found to have very similar stiffness–density relationships and composition. However, OA bone differed significantly from normal. The stiffness of OA bone increased more slowly with apparent density and its material density was significantly reduced. These findings were due to an altered composition of the bone in which the mass fraction of mineral is 12% less than normal. There was also greater site variation of both apparent and material density, suggesting an altered sensitivity to applied load. These results support the concept that osteoporosis is a loss of normal bone. They also provide evidence for the hypothesis that osteoarthritis is, at least partly, a bone disease in which proliferation of defective bone results in an increase in bone stiffness.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements

Osteoarthritis (OA) and osteoporosis (OP) are two common, age-related musculoskeletal disorders with considerable morbidity and mortality. Only rarely are both seen in the same patient,(1-4) and clinical experience and epidemiological studies have suggested that there is a negative association between them.(5-7) However, the relationship between the two conditions remains unclear. OP is universally accepted to be a disease of bone, whereas OA is generally considered to be a disease of articular cartilage, the changes in the bone being secondary to the damage to the cartilage. Alternatives to the “cartilage first” hypothesis are an altered hemodynamics (discussed and dismissed by Freeman and Meachim(8)) and the microfracture repair hypothesis of Radin and Paul.(9) They proposed that the primary changes in OA may be in the underlying subchondral bone, and they reasoned that bone stiffness was increased by the healing of microfractures, resulting in overloaded cartilage. Determining whether cartilage or bone is the primary tissue in the pathogenesis of OA is not trivial because cartilage degeneration is generally detected only when it is well advanced, by which time it is too late to determine the early events.

The proximal femur is the area most affected by pathological processes which are responsible for either fractured neck of femur or OA of the hip. Many studies have concentrated on noninvasive measurement of skeletal mass in the hip region,(10-12) but knowledge about the relationship between the mechanical properties of the bone and these measurements is not yet adequate.(13-15) Increasingly, it is evident that bone quality and structure is of at least as much importance as bone mass in determining the risk of fracture in OP.(3,14) However, it is implicit in many of these methods that the bone is essentially normal, merely altered in quantity. This assumption has never been rigorously tested. Though there have been many studies of the mechanical behavior of normal trabecular bone and its dependency on density (summarized by Cowin(16), few have investigated the changes in the properties of bone as a consequence of degenerative and metabolic diseases.(17-20) Density fractionation measurements on powdered bone have suggested that there is hypomineralization of OA bone,(21) but most other studies paid little attention to how the density is determined by the composition of the bone and whether this was normal. This study investigates both the mechanical and material properties of cancellous bone from femoral heads of patients with OA or OP and compares these with normal bone. This will test whether bone from these patients is abnormal or whether it has essentially normal composition and is simply altered in quantity.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements

Sample preparation

Femoral heads were collected from three clinical groups, OA, OP, and normal, and were matched for age and gender. The femoral heads from both patient groups were obtained in theatre from patients undergoing a hip replacement for either a fractured neck of femur attributed to OP or for OA of the hip. Cases with roentgenographic, biochemical, or histological evidence of osteomalacia, multiple myeloma, rheumatoid arthritis, or secondary OP due to corticosteroids were excluded from the OP group. Patients with rheumatoid arthritis, congenital or acquired dysplasia, gout, or avascular necrosis were excluded from the OA group. A control group was collected from hips removed during post mortem examination, and the medical records were examined to exclude disorders affecting bone metabolism. All samples were stored at 4°C in a calcium phosphate–buffered 0.15 M saline solution (containing 0.2 mM CaCl2 · 2H2O, 0.2 mM Na2HPO4, 0.01 mM Na4P2O7 · 10H2O, and 0.4 g of sodium azide/l) because this has been shown to preserve the structure and composition of the bone.(22)

From each femoral head, seven cylindrical cores 9 mm in diameter were removed from the sites shown in Figure 1. These sites were chosen to represent regions subjected to different amounts of loading in vivo; superior being the most highly loaded; posterior, anterior, and medial from the partially loaded region; and inferior being the least loaded.(23,24) The central and lateral regions are not loaded directly but have loads transmitted to them and represent the center of the femoral head and the neck of the femur, respectively. The cores were taken with the cylindrical axis along the preferred orientation of the trabeculae, which, at the surface, is also approximately perpendicular to the surface. The ends of the cancellous bone cores were trimmed parallel by removing the subchondral bone, and the dimensions were measured using a micrometer. The length of the cylinders varied slightly between samples depending on the size of the femoral head; the mean length ± standard deviation was 7.7 ± 1.6 mm. The gross volume of each tissue sample, Vs, was determined from the above dimensions. Care was taken to maintain the bone at all times in a fully hydrated condition.

thumbnail image

Figure FIG. 1. Schematic diagram of the sites on the human femoral head from which samples of cancellous bone were removed.

Download figure to PowerPoint

Mechanical testing

The stiffness of each sample was measured from an unconstrained compression test using an Instron 5564 materials testing machine (Instron, High Wycombe, U.K.). The test was performed at a strain rate of 20%/minute (0.0033/s). The machine was set up to plot the stress-strain curve as the test proceeded, and the test was stopped manually as soon as the slope of the trace, which defines the stiffness, could be seen to be reducing. We assume that a reduction in stiffness as the test proceeds is due to the initial stages of failure of the material and the test is stopped to prevent overt damage to the specimen that would prejudice the subsequent density measurements. After completing the test, the stiffness was plotted as a function of strain and the peak stiffness, the yield strength of the sample, and the energy absorbed to yield were determined. The yield strength was defined by the stress at which the stiffness had reduced by 3% from the maximum stiffness and the energy absorbed to yield by integrating to find the area under the curve up to the defined yield point.

Density and composition measurements

Each sample was cleaned using a water jet to remove all blood, marrow, etc. and defatted using chloroform:ethanol (2:1) in an ultrasonic bath at room temperature for 4 h. It was then washed thoroughly again in a water jet, placed in buffer solution in an ultrasonic bath for 4 h, and finally centrifuged in buffer solution at 175g for 20 minutes to ensure that it was fully hydrated.(25)

Two densities describing the material properties of the bone were measured in this study(25): the density of the trabecular bone material, which we have called the material density, and the density of the cancellous structure itself, called the apparent density. Material density is the mass of the sample divided by the volume of the trabeculae, to determine which requires the use of Archimedes' principle. Apparent density is the mass divided by the gross sample volume.

After cleaning, each sample was weighed immersed in buffer solution to determine the mass mb, centrifuged again at 175g for 20 minutes in a dry centrifuge tube to remove water from the pores and weighed again to determine the mass, ma, which is the mass of the bone material itself. The apparent density, ρa, was found from ma/Vs and the material density, ρm, calculated using Archimedes' principle (ρm = ma ρb/[mamb]) where ρb is the density of the buffer. Finally, the sample was dehydrated at 105°C for 48 h, weighed, ashed in predried crucibles at 600°C for 24 h, and weighed again. Water content is given by the difference between wet and dry masses, mineral content by the final mass after ashing, and organic content by the difference between the dry mass and the ash mass. All are expressed as fractions of the wet mass, ma. Given the nature of the material, it would be surprising if the water content varied significantly from in vivo to its cleaned state equilibrated in a physiological buffered solution. In addition, the salt concentration in the buffer (0.16 molal) is presumed to be similar to that perfusing the bone in vivo and negligible compared with that of the bone itself (∼6 molal) and thus will have no measurable effect on the mineral mass measurements.

Statistical methods

The first step was to compare the overall results in each patient group and test for global differences in the mechanical and compositional properties. Site variation within each group was then investigated further to assess the effects of load bearing on these overall results. However, because these may be inter-related a two-way analysis of variance (ANOVA) was performed. Normality of the distributions was assessed using the Kolmogorov-Smirnov test with the significance level set at 0.05. Where the data were not normally distributed with equal variances, the equivalent nonparametric tests were used. Mean values and associated standard errors are shown for data that are normally distributed, otherwise median values are given. Pairwise multiple comparisons following ANOVA were performed using the Student-Newman-Keuls method or Dunnett's test for comparison with a control. Linear regression was used to determine the correlation between variables.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements

The results for measurements of differences between disease groups will be presented first followed by an analysis of the variation of properties with site. Where there is significant interaction between these parameters, as shown by two-way ANOVA, this will be noted. The median ages and the oldest and youngest of each of the groups are shown in Table 1. The ratio of female to male was not significantly different in any group, and there were no differences in the age distributions between the groups.

Table Table 1. Details of Patient Groups
Thumbnail image of

Figure 2a shows a stress-strain curve measured from a sample of cancellous bone that was tested to failure. The gradient of the curve, found simply as the slope between consecutive points, defines the stiffness, and if this is plotted as a function of strain a curve such as that in Fig. 2b is obtained. This procedure accentuates any noise in the data, and the curve shown was smoothed using a running mean over five consecutive points. The peak modulus was found as the mean of the values shown in bold in which the stiffness varies by less than 3%. The yield point is defined as the stress at the end of this region when the stiffness starts to reduce. The strain axis is the same in Fig. 2a and Fig. 2b to allow easy comparison. The energy absorbed to yield is determined by integrating the stress-strain curve up to the yield point. These curves are typical of those obtained in this study, except that the test was stopped as soon as the stiffness could be seen to be decreasing.

thumbnail image

Figure FIG. 2. (a) Stress-strain curve from a sample of cancellous bone and (b) the stiffness derived from it as the gradient of the curve. The point at which the stiffness starts to decrease defines the yield point, and the energy absorbed to yield is the area under the stress-strain curve to this point, shown shaded.

Download figure to PowerPoint

ANOVA showed that the difference in mean values of stiffness, yield strength, and energy absorbed to yield are greater than would be expected by chance after allowing for the effects of difference in site (p < 0.001). The stiffness and energy absorbed to yield, but not the yield strength, for the OA bone were both different from normal (p < 0.05), whereas the stiffness and yield strength, but not the energy absorbed, were different from normal in the OP group (p < 0.05). The median stiffnesses of all of the cancellous bone samples from the femoral heads of each group are shown in Fig. 3a. As expected, the bone from patients with OA is stiffer (356 MPa) than that from OP patients (247 MPa) with the normal bone intermediate between the two (310 MPa). The yield strengths were found to be in the same order (Fig. 3c) with OA being the strongest (4.3 MPa), then normal bone (3.3 MPa) and OP being the weakest (2.5 MPa). The energy absorbed to yield is shown in Fig. 3e: OA, 31.9 kJ/m3; normal, 21.8 kJ/m3; and OP, 16.3 kJ/m3. The interaction between disease group and site was significant only for stiffness and yield strength with significance values of about 0.03 in both cases. There was no significant interaction between disease group and site for energy absorbed to yield.

thumbnail image

Figure FIG. 3. Average values of mechanical properties for all sites from all patients in each disease group and their decomposition by site to show variation over the femoral head. Cancellous bone cores were removed from the femoral head of patients undergoing a hip replacement for osteoarthritis (OA), or osteoporosis (OP), as shown in Fig. 1, and are compared with equivalent sites from normal controls. (a) Overall stiffness, (b) site variation of stiffness, (c) yield strength, (d) site variation of yield strength, (e) energy absorbed to yield, and (f) site variation of energy absorbed to yield. Data are shown as median values because not all were normally distributed. The extent of the shaded bars shows the 25% and 75% confidence limits, the error bars show the 5% and 95% limits. (* p < 0.05 compared with normal).

Download figure to PowerPoint

The apparent density of the bone in the three groups showed a pattern similar to the stiffness with the OA bone being significantly higher, at 0.71 g/cm3, than normal, 0.47 g/cm3 (p < 0.05), and the OP bone being significantly less at only 0.38 g/cm3 (p < 0.05) (Fig. 4a). However, the material densities showed a very different pattern. The material density of the bone from the OP group (1.86 g/cm3) was not significantly different from that in the normal group (1.89 g/cm3), but the OA bone (1.73 g/cm3) was considerably less dense (p < 0.05). Two-way ANOVA showed that there was no significant interaction between the disease group and the site for either apparent or material densities.

thumbnail image

Figure FIG. 4. (a) Average values of density of cancellous bone from the femoral heads of patients with OA or OP compared with normal controls. (See text for definitions of each of the densities.) These data are subdivided to show the site variation of (b) apparent and (c) material densities over the femoral head. As expected, OP bone has the lowest and OA bone the highest apparent density. However, this order is reversed for the material density, and the largest variation in both properties is found in the OA group. Median values are shown, and the extent of the shaded bars indicates the 25% and 75% confidence limits and the error bars show the 5% and 95% limits. (*p < 0.05 compared with normal).

Download figure to PowerPoint

Analysis of the bone composition in terms of mass fractions of water, organic, and inorganic components provides some basis for these differences (Fig. 5). No significant differences were detected in the composition of the bone from OP and normal groups. However, there was a marked difference in both the water and mineral contents of the OA bone compared with normal (p < 0.001), but no apparent change in the organic mass fraction. Expressing the mineral content as a fraction of dry mass leads to a similar result in which bone from the OA group comprises 61.6 ± 0.4% mineral compared with 65.8 ± 0.7% for the normal group and 64.4 ± 0.4% for the OP group. The OA group is significantly different (p < 0.05) by two-way ANOVA after accounting for site variation, and there was no significant interaction between disease group and site.

thumbnail image

Figure FIG. 5. Composition of normal, OA, and OP cancellous bone from human femoral heads expressed as a fraction of the mass of the sample. OP and normal are very similar, whereas OA bone is significantly different with a reduction in mineral content.

Download figure to PowerPoint

Linear regression was used to assess the relationship between stiffness and density. No correlation was found between stiffness and material density. There was a significant correlation between stiffness, E, and apparent density (ρa) in all three groups (Fig. 6). The regression equations are:

  • equation image

One-way ANOVA and Dunnett's test showed that the gradients of the regression lines for OA and normal were significantly different (p < 0.05), while no difference was found between OP and normal. A best subsets linear regression was also done with each sample coded for the site from which it came by assigning it the value 1 for that site and 0 for all the others. This was to allow for the fact that the samples are not totally independent, because they do not each come from a different femoral head, and to determine the variability due to position within the head. For the normal group, there was an increase in R2 from 0.59 to 0.75 with the inclusion of sites superior, inferior, central, and lateral and a similar increase in R2 in the OP group from 0.44 to 0.59 when superior, central, and lateral sites were taken into account. For the OA group, the increase was not as great, from 0.34 to 0.42, after including superior, inferior, and central sites. Inclusion of further sites made no difference to the correlation coefficient in any of the groups.

thumbnail image

Figure FIG. 6. Linear regression relationships between stiffness and apparent density of cancellous bone. Data are measured from seven independent samples per patient from the sites described in Fig. 1. OP and normal are very similar, whereas OA bone is significantly different. See text for regression equations.

Download figure to PowerPoint

There was a clear site variation of mechanical properties over the femoral head in all three groups. Figures 3b, 3d, and 3f shows that the greatest values of stiffness, yield strength, and energy absorbed to yield were found on the most heavily loaded superior aspect. Slightly lower than this are the partial weight-bearing areas (posterior, anterior, and medial) followed by the central region. The lowest values of stiffness and yield strength were found in the femoral neck for the OP and normal groups but in the inferior region for the OA group. The femoral neck in the OP group also had the lowest value of energy absorbed to yield, whereas this occurred in the inferior region in both of the normal and OA groups. If the average for each area in the patient groups and the corresponding area in the normal group is expressed as a ratio, then the overall increase in these parameters in OA and the decrease in OP is clearly seen (Fig. 7). Shown in this way, two areas immediately appear anomalous: the properties of the neck region in OA are enhanced by a factor of between 2.5 and 4 in contrast to typical values of generally less than two for the other sites, and the inferior region of the OP group has enhanced properties where in all other sites they are reduced compared with normal.

thumbnail image

Figure FIG. 7. For each site (a) the stiffness, (b) the yield strength, and (c) the energy absorbed to yield are shown as a ratio of the corresponding site in the normal group.

Download figure to PowerPoint

The apparent densities (Fig. 4b) in the OP and normal groups reflect the variations in stiffness with the most heavily loaded areas having the greatest density. However, the OA group showed much greater differences in apparent density between sites with the order of the partial load-bearing regions reversed compared with the OP and normal groups. There was only a small variation in the material density in different sites in the OP group and slightly more in the normal group (Fig. 4c), though only the femoral neck (lateral) region of the normal group was significantly more dense (p < 0.05) than the mean of the load-bearing areas. A similar pattern was found in the OA group with the indirectly loaded regions, lateral (p < 0.01) and central (p < 0.02), having an increased density compared with the mean of the directly loaded areas. Once again, there was greater site variation, and this shows the opposite trend to the apparent density, i.e., the sample with highest apparent density had the lowest material density. Expressing the densities as ratios, by dividing the results for each patient group by the corresponding site from the normal group, showed similar patterns for the apparent density to those of the stiffnesses described above (Fig. 8). There was a similar anomalously high value for the inferior region in the OP group and a higher density, though not by such a large factor, in the lateral OA group. The material densities can all be seen clearly to be lower than in the normal group for both the OP and the OA groups. The OA densities were considerably reduced, and the reduction was greatest in the more heavily loaded regions.

thumbnail image

Figure FIG. 8. For each site (a) the apparent and (b) the material densities are shown as a ratio of the corresponding site in the normal group.

Download figure to PowerPoint

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements

This investigation shows that the apparent density and mechanical stiffness of cancellous bone from the femoral head are increased in OA and decreased in OP compared with normal, in agreement with previous studies.(19) However, we also show that whereas OP bone has very similar material density and composition to normal bone, OA bone is considerably less dense and has a reduced mineral content.

The different densities being measured in this experiment provide information on the structure of the bone and the material of which it is made. Apparent density reflects the trabecular bone mass, without bone marrow, per unit volume of the sample of bone. Thus, the greater the volume of the trabeculae, Vt, i.e., the greater their number or the thicker they are, the higher will be the apparent density, assuming the composition of the bone does not change. Alternatively, if the volume occupied by the trabeculae does not alter, it will reflect changes in the material density. Material density is a measure of the bone material of which the trabeculae themselves are comprised.(25) The apparent density, ρa, is related to the material density, ρm, by the relationship

  • equation image(1)

and variation theory then shows that changes in trabecular volume, ΔVt, and material density, Δρm result in changes in apparent density given by

  • equation image(2)

If the change in each parameter is taken to be the difference between the value measured for the disease group and that for the normal group, then from the figures presented it is readily calculated that there is an overall reduction in trabecular volume in the femoral head of the OP group of 18% but an increase of 60% in the OA group. However, the material density of OA bone is lower than normal by 8.5%, whereas the OP bone is very similar to normal. The composition measurements from OP bone also showed no significant difference from normal and taken together these results support the currently accepted concept that OP bone material is normal in quality but reduced in quantity.

Of considerable interest was the result that the fractional inorganic content of OA bone is significantly less than normal while the fractional organic content is unchanged and the fractional water content is increased. However, these results are not easy to interpret. Because the contents are expressed as fractions, a change in one component must lead to apparent changes in the others even if there is no change in absolute amounts of these other components. In this case, if the mineral content is reduced from 54.3% in the normal to 47.2% in OA, and if there were no changes in the absolute amounts of organic and water, then it is readily shown that the organic fraction would apparently increase to 32.7% and the water fraction to 20.1%. That the organic fraction appears not to change while the water fraction increases points to there being actual changes in the absolute amounts of these components, not just a reduction of mineral content. The change in the absolute amount of organic material cannot be determined from this study, but it is certain that whatever the nature of the organic framework it is hypomineralized and there is a relative increase in the amount of water. These results extend those obtained by Grynpas et al,(21) who used a density fractionation technique on finely ground bone to show that OA subchondral bone is hypomineralized. Theirs was a relatively crude method that destroyed the tissue and produced a histogram with binned values of densities; there was no measurement of composition. Mineralization has been shown to have an effect on the packing fraction of collagen fibrils; increased mineralization leads to closer packing of the collagen molecules and a reduced water content.(22,26) Our results are consistent with a defect in the mineralization of the collagen, resulting in a reduced mineralization and a decreased packing fraction. This would result in an increased water content and ultimately a reduction in mechanical stiffness, which will be tested in future experiments.

Further support for this conclusion may be gained from the regression relationship between stiffness and apparent density. Once again, normal and OP bone lie very close together, suggesting little difference between them, though the clustering of OP values at the lower density region is quite marked. In contrast, OA bone is considerably less stiff for a given density, and the data are spread over a considerable range of densities. Theoretical and parametric modeling suggest a relationship between stiffness and apparent density of the form E ∝ ρn, where n = 2 for open cell models and n = 3 for closed cells.(27) Carter and Hayes(28) suggested n = 3 from their experimental data, and Rice et al.(29) comprehensively surveyed the current literature and proposed that the best value for n is 2. We found worse values for R2 using a simple n = 2 or n = 3 relationship and only slightly better values than those quoted if we used a stepwise regression and retained a linear term. Analysis of the residuals from the linear regressions showed them to be normally distributed and to have no correlation with apparent density. This shows that the regression is not being dominated by the higher values of density, especially in the OA group, and that there is no justification for fitting anything other than a straight line to our data. This may be a consequence of the limited range of values for apparent density (0.14–1.4 g/cm3) found in this study, but our results show evidence only for a linear relationship between stiffness and apparent density in these samples of cancellous bone. However, it was not our intention to test these theoretical relationships, and because the experimental design was not optimized for that, we cannot comment on their validity. Linear regression was performed to detect differences between the bone from the different disease groups, and this it does quite clearly.

The smaller number of samples in the normal group arose from the difficulty of obtaining normal tissue post mortem, and there will always be doubt in a study like this as to how “normal” are the normal group members. Because of the requirement for matching by age and gender, the tissue was obtained from an elderly population which must inevitably have contained people with subclinical OA or OP. We excluded any femoral heads with marked tissue degeneration but were not able to collect samples in the same number as can be obtained from the operating theater. It is essential to have some group to which the disease groups can be compared, otherwise it is not possible to determine which of these groups have deviated from normal. Using a clinical definition of disease means that we can define the “normal” group to be clinically normal, though we cannot be sure that there is no overlap in the material properties of the groups. The OP and OA groups are defined clinically by either a fractured neck of femur or the need for a total hip replacement with the classic signs of OA. The self-consistency of the results gives us reasonable confidence that what we have measured are real changes in the bone.

The variation of properties with site, and hence with mechanical loading, are reasonably consistent and similar in the OP and normal groups. The most heavily loaded areas are stiffer and have a greater apparent density than more lightly loaded regions. The apparent density of the central region is comparable with the most heavily loaded superior region, though the stiffness is more comparable with the partial load-bearing regions. The stiffness and the apparent density of the lateral, or femoral neck, region is considerably lower than the direct load-bearing regions in both the normal and the OP group. Curiously, the material density is highest in both of these two regions, whereas it is reasonably constant over the weight-bearing areas. Whether this is reflected in the mechanical properties of individual trabeculae needs to be tested. Compared with these patterns, OA bone again shows a number of differences. There is greater variation between the regions, with the most heavily loaded being still the stiffest and having the greatest apparent density, though the significance, if any, of the reversal in the order within the partial load-bearing regions is not clear. There is also more variation in the material densities, with the most heavily loaded being the least dense and the central and lateral regions again being the most dense. A possible explanation could be that there is a metabolic response from the cells to try to maintain the load-bearing capability of the bone in the face of reduced bone quality and that this results in a greater sensitivity to applied load. The anomalous results for the neck region in the OA group and the inferior femoral head of the OP group, shown by comparison of the ratios of these areas to corresponding areas of the normal group, are not explicable from the data and need further investigation.

Clinical experience and epidemiological evidence that OP and OA, although both very common, are rarely seen together, raises the question of whether there could be common factors between these diseases. OA is generally considered to be a disease of cartilage, but evidence to suggest that the primary problem may lie in the bone is slowly accumulating. OA patients have a better preserved bone mass,(10,30) and studies of iliac crest bone have found changes in the mass, biochemical, and biomechanical properties.(31,32) These results suggest a generalized change in the bone. The hypothesis of microfracture repair(9,33) suggests a possible mechanism for local OA but does not explain why there should be any changes in bone remote from the affected joint. Cartilage is too thin to be effective at attenuating impact loads. Reducing the peak stress produced by an impact by spreading it in time is done mainly by the bone and periarticular soft tissues.(34,35) If the bone becomes stiffer, it may be less able to absorb impact loads, which may in turn lead to increased peak stresses in the cartilage. Conversely, a reduction in stiffness due to bone loss in OP may result in greater force attenuation which will protect the cartilage.(9,33) This hypothesis could explain the inverse correlation between OP and OA in hips.

Though supportive of this hypothesis, our results still do not determine whether the changes observed in OA bone are primary or secondary to degenerative changes in the cartilage, which would lead to increased loading of the bone. Though it is to be expected that the bone would respond to such an increased loading by producing a greater amount of bone, it would also be expected that this bone would be essentially normal and have similar properties to the existing bone. This would also be the case if the increase in the stiffness of the underlying bone was a primary event as a consequence of a repair response to microfractures; the additional bone would still be expected to have a normal composition. In contrast, the results presented here clearly show that the bone is abnormal, with reduced mineral content and material density. An alternative hypothesis is that there is a defect in the bone metabolism leading to the production of bone that is either weaker or less stiff or both. Normal loading will then lead either to microfractures and a repair response and/or to an increased strain being generated in the tissue to which the cells would respond by producing greater quantities of bone. In either case, the newly synthesized bone would still be abnormal. This defective bone is not readily assessed by conventional means, which may be why it appears largely to have been overlooked. Simple measurement of “bone mineral content” or “bone mineral density” (which are neither composition nor density measurements) in vivo by dual-energy X-ray absorptiometry or computer tomography, will merely show an increased quantity of mineral. Equation 2 shows that even no measurable change in apparent density could hide opposing changes in material density and bone volume. Only when the material density or the composition is measured is this defect revealed. Further studies are in progress to determine the exact nature of the compositional changes in OA bone and this may begin to reveal why the cells are producing abnormal bone and suggest new avenues for treatment of patients presenting with early symptoms of OA.

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements

We thank the Medical Research Council for a Senior Fellowship for RMA and the Sir Halley Stewart Trust and the Arthritis and Rheumatism Council for financial support. We are grateful to Marion Campbell for statistical advice.

  • 1
    Pogrund H, Rutemberg M, Makin M, Robin G, Menczel J, Steiberg R 1982 Osteoarthritis of the hip joint and osteoporosis: A radiological study in a random population sample in Jerusalem Clin Orthop 164: 130135.
  • 2
    Weintroub S, Papo J, Ashkenazi M, Tardiman R, Weissman SL, Salama R 1982 Osteoarthritis of the hip and fractures of the proximal end of the femur Acta Orthop Scand 53: 261264.
  • 3
    Dequeker J 1985 The relationships between osteoporosis and osteoarthritis Clin Rheum Dis 11: 271296.
  • 4
    Dequeker J, Johnell O 1995 Osteoarthritis protects against femoral neck fracture: The MEDOS study experience Bone 14: S51S56.
  • 5
    Byers PD, Contepenni CA, Farkas TA 1970 A post mortem study of the hip joint Ann Rheum Dis 29: 1531.
  • 6
    Solomon L, Schnitzler CM, Browett JP 1982 Osteoarthritis of the hip: The patient behind the disease Ann Rheum Dis 41: 118125.
  • 7
    Cooper C, Cook PL, Osmond C, Cawley MID 1991 Osteoarthritis of the hip and osteoporosis of the proximal femur Ann Rheum Dis 50: 540542.
  • 8
    Freeman MAR, Meachim G 1979 Ageing and degeneration. In: FreemanMAR (ed.) Adult Articular Cartilage, 2nd Ed. Pitman Medical, London, U.K., pp. 487543.
  • 9
    Radin EL, Paul IL 1970 Does cartilage compliance reduce skeletal impact loads? The relative force-attenuating properties of articular cartilage, synovial fluid, periarticular soft tissues and bone Arthritis Rheum 13: 139144.
  • 10
    Foss MVL, Byers PD 1972 Bone density, osteoarthrosis of the hip, and fracture of the upper end of the femur Ann Rheum Dis 31: 259264.
  • 11
    Kroger H, Tupperainen M, Honkanen R, Alhava EM, Saarikoski S 1994 Bone mineral density and risk factors for osteoporosis: A population-based study of 1600 perimenopausal women Calcif Tissue Int 55: 17.
  • 12
    Cummings SR, Black DM, Nevitt MC et al. 1993 Bone density at various sites for prediction of hip fractures Lancet 341: 7275.
  • 13
    Alhava EM 1991 Bone density measurements Calcif Tissue Int 49: S21S23.
  • 14
    Cosman F, Schnitzer MB, McCann PD, Parisien MV, Dempster DW, Lindsay R 1992 Relationships between quantitative histological measurements and noninvasive assessments of bone mass Bone 13: 237247.
  • 15
    Deldmas PD 1993 Bone mass measurement: How, where, when and why? Int J Fertil Menopausal Studies 38 (Suppl 2): 7076.
  • 16
    Cowin SC 1989 The mechanical properties of cancellous bone. In: CowinSC (ed.) Bone Mechanics. CRC Press Inc., Boca Raton, FL, U.S.A., pp. 129157.
  • 17
    Lereim P, Goldie I, Dalberg E 1974 Hardness of the subchondral bone of the tibial condyles in the normal state and in osteoarthritis and rheumatoid arthritis Acta Orthop Scand 45: 614627.
  • 18
    Reimann I, Mankin HJ, Trahan CJ 1977 Quantitative histologic analyses of articular cartilage and subchondral bone from osteoarthritic and normal human hips Acta Orthop Scand 48: 6373.
  • 19
    Deligianni DD, Missirlis YF, Tanner KE, Bonfield W 1991 Mechanical behaviour of trabecular bone of the human femoral head in females J Mater Sci Mater Med 2: 168175.
  • 20
    Boyde A, Jones SJ, Aerssens J, Dequeker J 1995 Mineral density quantitation of the human cortical iliac crest by backscattered electron image analysis: Variations with age, sex, and degree of arthritis Bone 16: 619627.
  • 21
    Grynpas MD, Alpert B, Katz I, Lieberman I, Pritzker KPH 1991 Subchondral bone in osteoarthritis Calcif Tissue Int 49: 2026.
  • 22
    Lees S 1988 Sonic velocity and the ultrastructure of mineralised tissues. In: HukinsDWL (ed.) Calcified Tissues. MacMillan, London, U.K., pp. 121152.
  • 23
    Thomas DB, Daniel TS 1983 In vitro contact stress distributions in the natural human hip J Biomech 16: 373384.
  • 24
    Hodge WH, Fijan RS, Carlson KL, Burgess RG, Harris WH, Mann RW 1986 Contact pressures in the human hip joint measured in vivo Proc Natl Acad Sci USA 83: 28792883.
  • 25
    Ashman RB 1989 Experimental techniques. In: CowinSC (ed.) Bone Mechanics. CRC, Boca Raton, FL, U.S.A., pp. 7595.
  • 26
    Lees S 1986 Water content in type I collagen tissues calculated from the generalized packing model Int J Biol Macromol 8: 6672.
  • 27
    Gibson LJ 1985 The mechanical behaviour of cancellous bone J Biomech 18: 317328.
  • 28
    Carter DR, Hayes WC 1977 The compressive behavior of bone as a two-phase porous structure J Bone Joint Surg 59A: 954962.
  • 29
    Rice JC, Cowin SC, Bowman JA 1988 On the dependence of the elasticity and strength of cancellous bone on apparent density J Biomech 21: 155.
  • 30
    Carlsson Å, Nilsson BE, Westlin NE 1979 Bone mass in primary coxarthrosis Acta Orthop Scand 50: 187189.
  • 31
    Gevers G, Dequeker J, Geusens P, Nyssen-Behets C, Dhem A 1989 Physical and histomorphological characteristics of iliac crest bone differ according to the grade of osteoarthritis at the hand Bone 10: 173177.
  • 32
    Gevers G, Dequeker J, Martens M, Van Auderkerke R 1989 Biomechanical characteristics of iliac crest bone in elderly women according to osteoarthritis grade at the hand joints J Rheumatol 16: 660663.
  • 33
    Radin EL, Paul IL, Lowy M 1970 A comparison of the dynamic force transmitting properties of subchondral bone and articular cartilage J Bone Joint Surg 52-A: 444456.
  • 34
    Kempson GE, Weightman B 1979 Load carriage. In: FreemanMAR (ed.) Adult Articular Cartilage, 2nd Ed. Pitman Medical, Tunbridge Wells, U.K., pp. 291331.
  • 35
    Radin EL, Rose RM 1986 The role of subchondral bone in the initiation and progression of cartilage damage Clin Orthop 213: 3440.