Difference in subchondral cancellous bone between postmenopausal women with hip osteoarthritis and osteoporotic fracture: Implication for fatigue microdamage, bone microarchitecture, and biomechanical properties




Osteoarthritis (OA) and osteoporosis (OP) of the hip rarely occur in the same patient. The purpose of this study was to determine whether this difference might be attributable to the different quantity and quality of subchondral cancellous bone in the two conditions.


Subchondral cancellous bone from the femoral head was obtained at the time of hip arthroplasty from 60 postmenopausal women, 30 with OA and 30 with OP. In each group, 10 specimens were subjected to compressive fatigue loading and 20 were left nonloaded. Specimens were examined by compressive mechanical testing, micro–computed tomography scanning, fluorescence microscopy, and nanoindentation techniques.


Both the ultimate stress and the elastic modulus of cancellous bone from OA patients were significantly higher than those of cancellous bone from OP patients (P < 0.05). Compared to cancellous bone from OP patients, the bone volume fraction and trabecular thickness were significantly increased, but bone matrix mineralization was significantly decreased, in cancellous bone from OA patients (P < 0.05 for each comparison). The microcrack density was significantly higher in OP cancellous bone than in OA cancellous bone (P < 0.001), irrespective of fatigue loading. In addition, fatigue loading resulted in a significant increase in microcrack density in both OA and OP cancellous bone (P < 0.001). There was no significant difference in nanoindentation elastic modulus and hardness between cancellous bone from OA and OP patients, as well as between bones with and without fatigue loading.


The difference in mechanical properties between OA and OP cancellous bone is attributed to different bone mass and bone structure. OP cancellous bone is susceptible to fatigue damage due to insufficient structure. However, increased bone volume and plate-like structure provide OA cancellous bone a superior capacity to resist fatigue damage.

Osteoarthritis (OA) and osteoporosis (OP) are two orthopedic disorders that usually result in high rates of morbidity and disability in people over the age of 50 years, particularly in postmenopausal women (1, 2). Although OA and OP commonly occur in the elderly, they rarely coexist in a single patient (1, 3). Healey et al (4) reported that only 4% of OP patients had significant osteoarthritis in the hip. It is well known that OP is associated with bone loss and microstructure deterioration, which can reduce bone strength and thereby increase the risk of fragility fracture (5). In contrast, the pathogenesis of OA remains a subject of controversy (6, 7). In general, OA initially presents with destruction of the articular cartilage (8), which is followed by osteophyte formation, cystic degeneration, and then sclerosis of the subchondral bone (9). This “cartilage first” hypothesis has been challenged, however (6). There is evidence that the cause of primary OA is related to the changes in subchondral bone (10), which includes the subchondral cortical plate and the underlying cancellous bone.

It has been reported that the subchondral cortical plate in the normal femoral head is stiffer than that in the femoral head of patients with OA or OP and so cannot, by itself, explain the opposite changes in articular cartilage that are associated with OA versus OP (11). However, the subchondral cortical plate is supported by the underlying cancellous bone. Changes in subchondral cancellous bone therefore also need to be examined for their influence on the articular cartilage.

In comparison to OP, OA cancellous bone has a significantly increased density and thick, plate-like trabeculae, both of which not only enhance bone strength, but also provide a stiffer structure (12, 13), rendering the overlying cartilage susceptible to more damage because of decreased shock absorption (2, 6). One possible mechanism of OA is that it plays a protective role against the age-related decrease in bone density and deterioration of the microarchitecture of the cancellous bone (1, 12, 14). In addition, the spatial distribution of trabeculae differs between OA and OP (13). Based on these lines of evidence, we postulate that in OP, there is a compliant subchondral cancellous bone that can absorb more impact force, so the fatigue-induced damage resulting from daily activities is more likely to occur in bone than in cartilage, whereas in OA, the situation seems to be the opposite. This may explain why OA and OP do not usually occur in the same patient. Nevertheless, there is no evidence showing the difference in fatigue-induced bone damage between patients with OA and those with OP.

The purpose of this study was to determine the differences in fatigue strength, microarchitecture, mineralization degree, and biomechanical properties between subchondral cancellous bones obtained from the femoral heads of patients with OA and patients with OP.


Study subjects.

Sixty postmenopausal Chinese women who underwent hip arthroplasty for primary osteoarthritis (OA) or for osteoporotic (OP) fracture of the femoral neck were included in the study. The method of patient recruitment has been described previously (10). Briefly, we matched the OA and OP patients as closely as possible in terms of their age (age range 53–86 years in the OA patients and 59–87 years in the OP patients; n = 30 patients per group). All patients were more than 5 years postmenopausal at the time of recruitment for study. Patients who had bone diseases other than OA and OP, and those who took medicines that affect bone metabolism (e.g., calcitonin, denosumab, estrogen, raloxifene, teriparatide, and corticosteroids, etc.) were excluded. All OA patients had grade IV disease according to the Outerbridge classification (15), whereas all OP patients had femoral neck fracture.

Informed consent was obtained from each patient. The study was approved by the Institutional Review Board of Shanghai Xinhua Hospital.

Specimen preparation and grouping.

The specimen of subchondral cancellous bone (15 × 15 mm in cross-section and 10 mm long) was collected from the load-bearing area of the femoral head using a pendulum saw. The specimens were wrapped in saline-soaked gauze and stored at −20°C. Specimens were thawed at room temperature before examination.

Thirty specimens from each group of patients (1 from each OA and OP patient) were used for various examinations. Ten specimens were subjected to compressive fatigue loading, and 20 were not loaded. The specimens that underwent fatigue loading were stained with calcein and embedded in polymethylmethacrylate (PMMA). These specimens were examined using fluorescence microscopy and nanoindentation techniques. For the nonloaded specimens, 10 were used for compressive mechanical testing and the other 10 were used for micro–computed tomography (micro-CT) examination. After micro-CT scanning, these specimens were processed and examined in the same way as those that had been subjected to fatigue loading.

Assessment of fatigue loading.

Fatigue loading analysis was carried out on standard-sized OA and OP cancellous bone specimens under load control, using an Instron 8874 servohydraulic materials testing system. The specimens were subjected to 10,000 cycles of sinusoidal compressive loading at a frequency of 2 Hz, with a minimum load of 20N and a maximum load of 300N.

Evaluation of microdamage.

As described above, the specimens that had been subjected to fatigue loading as well as those that had not were stained with calcein and embedded in PMMA. Briefly, the specimens were dehydrated with graded ethanol solutions (80%, 90%, and 100%) and then immersed in 0.5 mmoles/liter of a calcein–ethanol solution for 2 days at room temperature (16). After in-bulk staining, the specimens were embedded in PMMA and sectioned longitudinally into 100 μm samples.

The bone sections were examined using a Zeiss microscope equipped with a CCD video camera. The microscopic image was imported to an Axioplan 2 image analysis system. Bone area (BAr; mm2), crack number (CrN; /mm2), and crack length (CrLe; μm) were measured using fluorescence (ultraviolet light) microscopy (10× objective). The mean CrLe (μm), the crack numerical density (CrDn; /mm2 [the CrN/BAr]), and the crack surface density (CrSDn; μm/mm2 [total CrLe/BAr]) were computed for the entire bone. The morphology of microcracks was also examined using a Zeiss LSM 510 confocal laser scanning microscope.

Micro-CT imaging.

Twenty nonloaded OA and OP specimens, 10 in each group, were examined using a Micro-CT system (μCT 80, Scanco Medical). Each specimen was scanned continuously at a slice thickness and slice increment of 36 μm. After scanning, a constant volume of interest (VOI) centered over the specimen was selected for analysis of all study samples. Three-dimensional (3-D) images were reconstructed based on the VOI. The bone volume fraction (BV/TV; %), trabecular thickness (TbTh; μm), trabecular number (TbN; /mm2), trabecular spacing (TbSp; μm), connectivity density (ConnD; /mm3), structure model index (SMI), and bone matrix mineralization (tissue mineral density, TMD; mg HA/cm3) were calculated using the software provided with the instrument. SMI is a topological index for estimating the characteristic form in terms of the plates and rods that compose the 3-D structure. This index assumes integer values of 0 and 3 for ideal plates and rods, respectively, whereas for a mixed structure containing both plates and rods, the value lies between 0 and 3 (14).

Mechanical testing.

The mechanical testing method we used has been described elsewhere (17). Briefly, specimens of cancellous bone form OA and OP patients were mechanically tested using an Instron 8874 instrument. Each specimen was compressed in a longitudinal direction between 2 platens at a rate of 2 mm/minute. The test was stopped upon failure of the specimen. The ultimate stress and elastic modulus were obtained from the stress–strain curve.

Nanoindentation testing.

The PMMA-embedded bone blocks used for microdamage evaluation were polished and examined by a nanoindentation technique as described previously (18). A NanoIndenter XP system (MTS Nano Instruments) was used to measure force and displacement during indentation of the polished bone specimen. Four trabeculae in each specimen were randomly selected using an optical microscope, and 5 nanoindentation tests were performed on each trabecula. A Berkovich diamond indenter tip (Ei = 1,141 GPa and νi = 0.07, where Ei represents elastic modulus indentation and νi represents Poisson's ratio indentation) was used for all measurements. After the surface was identified, the indenter was advanced to 500 nm at a speed of 10 nm/second to avoid the effect of bone surface roughness. A typical indentation load-displacement curve included a loading segment, a 10-second holding period at maximum load, an unloading segment, and a 50-second holding period for thermal drift measurement at 10% of minimum load (18). Bone tissue elastic modulus (E) and hardness (H) were calculated from the unloading segment of the load-displacement curve according to the method of Oliver and Pharr (18). Mean values for the elastic modulus and the hardness of cancellous bone were calculated for each specimen.

Statistical analysis.

All data were expressed as the mean ± SD. Student's t-test was used to compare means between groups. Two-way analysis of variance was used to compare the differences in variables between the two disease groups (OA and OP) and the two loading conditions (with and without fatigue loading). All statistical analyses were performed using a commercial software package (SPSS 10.0). P values less than 0.05 were considered significant.


The patients with OP were older (mean ± SD age 72.57 ± 7.82 years) than the patients with OA (mean ± SD age 68.43 ± 8.45 years), but the difference did not reach statistical significance (P = 0.054).

Findings of the microdamage analysis.

The calcein-stained microcracks were visible by fluorescence and confocal microscopy (Figure 1), particularly in bone that had undergone fatigue loading. As shown in Table 1, CrDn and CrSDn were significantly increased in OP cancellous bone as compared to OA cancellous bone (P < 0.001 for each comparison). In addition, fatigue loading resulted in a significant increase in CrDn and CrSDn in both the OA and the OP groups (P < 0.001 for each comparison). The increase in CrDn after fatigue loading was higher in OP (2.52 times) than in OA (1.91 times) samples. There was no significant difference in the mean CrLe between OA and OP samples or between loaded and nonloaded cancellous bone samples. The increase in CrSDn can therefore be attributed to an increase in the number of cracks rather than an increase in the length of the cracks.

Figure 1.

A linear microcrack stained with calcein (arrow) is shown under 3 different imaging techniques: fluorescence microscopy using ultraviolet light (A), confocal microscopy using fluorescent light (B), and confocal microscopy using reflected light (C). Original magnification × 100 in A; × 40 oil lens in B and C.

Table 1. Comparison of microdamage in OA and OP cancellous bone, with and without fatigue loading*
 CrDn,/mm2Mean CrLe, μmCrSDn, μm/mm2
  • *

    Ten samples per group were analyzed in each experiment. Values are the mean ± SD. P values were as follows: for comparisons of crack numerical density (CrDn), P < 0.001 for osteoarthritis (OA) versus osteoporosis (OP), loaded versus unloaded, and interaction; for comparisons of mean crack length (CrLe), P = 0.971 for OA versus OP, P = 0.112 for loaded versus unloaded, and P = 0.496 for interaction; and for comparison of crack surface density (CrSDn), P < 0.01 for OA versus OP and loaded versus unloaded, and P < 0.005 for interaction.

OA patients   
 Unloaded0.178 ± 0.07084.1 ± 18.714.4 ± 5.05
 Loaded0.341 ± 0.103102 ± 26.933.9 ± 12.2
OP patients   
 Unloaded0.321 ± 0.11678.1 ± 18.725.5 ± 8.90
 Loaded0.809 ± 0.22297.2 ± 32.475.2 ± 28.0

Findings of micro-CT imaging.

The results of micro-CT analyses are shown in Table 2. Compared to OA cancellous bone, the BV/TV, TbTh, and ConnD values in OP cancellous bone were significantly decreased (P < 0.05 for each comparison), whereas there was no significant difference in TbN between OA and OP samples. TbSp was significantly higher in OP than in OA cancellous bone (P < 0.05). Additionally, the SMI was significantly higher in OP than in OA samples (P < 0.05). As shown in Figure 2, there were more thin and rod-like trabeculae in OP specimens than in OA specimens. In contrast, the TMD in OA cancellous bone was significantly lower than that in OP cancellous bone (P < 0.01).

Table 2. Comparison of the microarchitecture of cancellous bone from OA and OP patients*
 OA (n = 10)OP (n = 10)P
  • *

    Values are the mean ± SD. OA = osteoarthritis; OP = osteoporosis; BV/TV = bone volume fraction; TbTh = trabecular thickness; TbN = trabecular number; TbSp = trabecular separation; ConnD = connectivity density; SMI = structure model index; TMD = tissue mineral density; HA = hydroxyapatite.

BV/TV, %26.0 ± 6.0219.3 ± 4.360.010
TbTh, μm206 ± 46.7165 ± 25.10.028
TbN,/mm21.27 ± 0.181.16 ± 0.150.132
TbSp, μm593 ± 102709 ± 1260.036
ConnD,/mm33.24 ± 0.812.43 ± 0.640.023
SMI1.03 ± 0.441.47 ± 0.430.039
TMD, mg HA/cm3755 ± 25.7798 ± 29.20.002
Figure 2.

Three-dimensional morphologic features of subchondral cancellous bone samples derived from the femoral heads of postmenopausal women with osteoarthritis (OA) and osteoporosis (OP). Resolution 36 μm.

Results of mechanical testing.

The difference in biomechanical properties between OA and OP subchondral cancellous bone is shown in Figure 3. Both the ultimate stress and elastic modulus of OA bone were significantly higher than those of OP bone (P < 0.05).

Figure 3.

Difference in ultimate stress (A) and elastic modulus (B) of subchondral cancellous bone samples derived from the femoral heads of postmenopausal women with osteoarthritis (OA) and osteoporosis (OP). Values are the mean ± SD of 30 patients per group. ∗ = P < 0.05.

Results of nanoindentation testing.

There were no significant differences in elastic modulus and hardness between OA and OP cancellous bone or between loaded and nonloaded samples. The values for these two variables were very similar among the 4 experimental groups (Table 3).

Table 3. Comparison of the effects of nanoindentation on the elastic modulus and hardness in OA and OP cancellous bone, with and without fatigue loading*
 Elastic modulus, GPaHardness, GPa
  • *

    Ten samples per group were analyzed in each experiment. Values are the mean ± SD. P values were not significant for comparisons of elastic modulus (P = 0.631 for osteoarthritis [OA] versus osteoporosis [OP], P = 0.112 for loaded versus unloaded, and P = 0.845 for interaction) or for comparisons of hardness (P = 0.334 for OA versus OP, P = 0.247 for loaded versus unloaded, and P = 0.794 for interaction).

OA patients  
 Unloaded16.0 ± 1.250.619 ± 0.036
 Loaded16.0 ± 1.900.606 ± 0.050
OP patients  
 Unloaded15.7 ± 1.820.609 ± 0.055
 Loaded15.8 ± 1.370.589 ± 0.033


Microdamage is normally present in bone and can suddenly increase due to the application of excessive cyclical loading (19). The amount of microdamage in the bone matrix can thus reflect the fatigue properties of bone. To the best of our knowledge, this study is the first to determine the difference in fatigue-induced microdamage in subchondral cancellous bone from patients with OA as compared to patients with OP. We assessed linear cracks in subchondral cancellous bone in the femoral heads of OA and OP specimens for two reasons. First, only linear cracks compromise bone strength and trigger bone remodeling (20, 21). Second, linear cracks are more likely to occur in aged bone and under compressive loading (20, 22), both of which were present in our study samples.

Our results showed that both CrDn and CrSDn were significantly higher in OP cancellous bone than in OA cancellous bone, regardless of whether the samples had been subjected to fatigue loading. In general, a small number of microcracks, which are caused by repetitive loading during daily activities, are present in normal bone. The number of microcracks in OA cancellous bone were not significantly increased compared to that in normal cancellous bone (16, 23), suggesting that the fatigue strength of OA bone is not decreased from normal levels. In this context, our data indicate that OP cancellous bone possesses lower fatigue strength than normal.

Our study verified that bone fatigue properties depended on the integrity of the bone structure. It is likely that the lower levels of microdamage in OA cancellous bone are associated with its normal, or even increased, bone mass and trabecular thickness (24). Bone with insufficient structure, as shown in samples from the OP patients, is predisposed to fatigue damage. Both the BV/TV and the SMI are inversely associated with the formation of microcracks in cancellous bone, suggesting that microdamage tend to occur in cancellous bone that has lower bone mass and fewer rod-like trabeculae (25, 26). In addition, the unchanged values for the nanoindentation parameters before and after fatigue testing suggest that fatigue damage may not compromise the bone material properties.

Unlike the CrDn results, this study did not show a significant difference in the mean CrLe in OA versus OP cancellous bone, even after fatigue loading. It suggests that OP cancellous bone is predisposed to microcrack formation rather than propagation. The formation of microcracks depends on the microstructure of the bone and the amount of strain applied to the bone (25, 27, 28). Decreased bone mass and impaired microarchitecture resulting from OP would enhance bone deformation under loading conditions, resulting in the formation of microdamage. Additionally, increased remodeling of OP cancellous bone engenders numerous resorption lacunae on the trabecular surface (29). The resorption lacunae are areas where stress concentrates (30) and are therefore the areas that are susceptible to the formation of microdamage. However, increased bone remodeling would also create more cement lines in bone, and these may act as a barrier to stop the growth of microcracks (31).

In this study, the results showed significantly higher cancellous bone volume, elastic modulus, and strength in femoral head samples from postmenopausal women with OA than from those with OP. There is compelling evidence that there is a positive correlation between bone density and bone stiffness and bone strength (10). Therefore, the increased bone mass in OA patients contributes to greater bone stiffness, and greater bone stiffness would cause subchondral plate sclerosis, which is harmful to the articular cartilage. Yet, the increased bone mass appears to be beneficial in reducing the risk of fragility fracture.

In addition to differences in bone mass, the difference in the mechanical properties of OA and OP cancellous bone is also associated with the trabecular microarchitecture. Micro-CT demonstrated thinner trabeculae in conjunction with poorer connectivity in OP cancellous bone than in OA cancellous bone. Moreover, the SMI suggested that there were more rod-like trabeculae in OP cancellous bone than in OA cancellous bone. The rod-like trabeculae are able to absorb higher-energy impact because they are more likely to yield under load. In contrast, the thick, plate-like trabeculae are less flexible, so they can bear a higher load with less deformation.

Such a large structural difference between OA and OP subchondral cancellous bones is likely to be attributable to differences in bone remodeling. In postmenopausal OP patients, bone remodeling is considerably accelerated and is associated with increased osteoclastic bone resorption and normal or decreased osteoblastic bone formation, thereby resulting in bone loss (32). It has been reported, however, that OA may arrest the postmenopause- and age-related destruction of cancellous bone (1, 12, 14, 33). This is attributed to decreased bone remodeling in OA (1, 33). In addition, a large number of studies indicate that the function and metabolism of OA osteoblasts are different from those of OP osteoblasts (34, 35). OP osteoblasts showed reduced anabolic function (34), whereas OA osteoblasts favored bone formation (35). The altered osteoblast phenotype may also increase the level of type I collagen in OA cancellous bone (36). The consequence of these biologic changes is an increase in the volume and strength of OA bone.

Nanoindentation has been used to characterize the elastic modulus and hardness of cortical and trabecular bone (37). The parameters determined by nanoindentation are related to bone tissue stiffness (38). However, this method measures bone intrinsic material properties at microscale levels (38), which does not always correlate with the stiffness of the entire bone sample (37). It has been suggested that bone tissue stiffness depends heavily on its degree of mineralization (38, 39). Burket et al (40) found that in bone samples from female baboons, the indentation modulus and hardness were most influenced by the degree of mineralization of the tissue. The ratio of mineral to matrix alone accounted for 78% of the variation in the indentation modulus and 70% of the variation in hardness (40). In this study, we did not find significant differences in nanoindentation elastic modulus and hardness between OA and OP subchondral cancellous bone, although the degree of mineralization of OP bone was moderately higher (5.8%) than that of OA bone. It has been reported that there were no significant differences in nanomechanical properties or the degree of bone mineralization between postmenopausal women with and without osteoporosis (18, 41), suggesting that osteoporosis induced by estrogen depletion may not change bone mineralization and bone material properties. Accordingly, the reduction in bone strength in postmenopausal OP is related to impaired bone structure rather than degraded bone material.

Coats et al (42) measured OA and OP subchondral cancellous bones using microindentation techniques and found a moderate decrease (7%) in the hardness of OA cancellous bone. However, the difference in hardness between OA and OP cancellous bones was not significant in the current study using nanoindentation techniques. The conflicting findings may be related to differences in sample processing and indentation testing conditions (e.g., maximum load, load/unload rate, load holding time, sites of bone collection, and indentation depth and area) (42–44). Taken together, our findings and the results from other studies (39, 42, 45) lead us to postulate that in OA cancellous bone, the degree of mineralization is slightly decreased as compared with normal or OP cancellous bone, but the change in bone hardness is uncertain.

It has been hypothesized that increased cancellous bone volume in OA is the result of decreased mineralization (46). The lower level of mineralized OA bone would be subject to higher strain under loading, stimulating osteogenesis to increase the bone volume. However, the available evidence suggests that decreased mineralization of OA bone plays little role in increasing the bone volume (46). We found that compared to OP, OA femoral heads had a 34.7% increase in cancellous bone volume (BV/TV) at the expense of a 5.8% decrease in bone mineralization (TMD). Cox et al (46) recently reported that in advanced OA, a 69% increase in cancellous bone volume was associated with 6% decrease in bone mineralization as compared to bone without OA. A simulation model determining the effect of mechanoregulated bone adaptation in response to different degrees of mineralization showed that a 6% decrease in bone mineralization could result in only a 9% increase in bone volume (46), which is remarkably lower than 69%.

Accordingly, the increased cancellous bone volume in OA is likely to be associated with other factors. Bone marrow lesions, characterized by increased water, blood, or other fluid inside the bone, often occur in patients with hip and knee OA and are associated with increased subchondral cancellous bone volume and decreased bone mineralization (47–49). Bone marrow lesions also correlate highly with intraosseous hypertension due to poor venous drainage (47). Venous stasis increases intramedullary pressure and, consequently, the interstitial fluid flow in bone, which in turn, stimulates osteogenesis (50). We think that the increased bone formation in OA is the cause of the moderately decreased bone mineralization due to the accumulation of less-mineralized new bone.

In conclusion, our findings from biomechanical testing indicate that subchondral cancellous bone from the femoral head of patients with OP possesses significantly lower stiffness and strength than that from patients with OA. The different trabecular morphology but similar nanoindentation properties suggest that the biomechanical differences between OA and OP cancellous bone are attributable to the bone mass and microstructure rather than to the bone materials. The insufficient structure and decreased stiffness make OP cancellous bone susceptible to fatigue damage, which is present in the form of microcracks. Viewed from another angle, the compliant OP subchondral cancellous bone would absorb more impact force resulting from activities of daily living, thus diminishing the damage to the overlying articular cartilage. In contrast, the highly increased bone volume and strong plate-like structure in OA cancellous bone provide superior capacity for resisting fatigue damage. The subchondral plate sclerosis, however, would increase the likelihood of damage to the articular cartilage.


All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Qiu had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study conception and design. Li, Dai, Jiang, Qiu.

Acquisition of data. Li, Jiang.

Analysis and interpretation of data. Li, Qiu, Dai.