Iliac cancellous osteocyte density declines with age, but its relationship to vertebral fracture pathogenesis is unknown. We performed iliac bone biopsy in 44 women with clinical vertebral fracture and 56 healthy women. The fracture patients had 34% fewer osteocytes but no reduction in percent occupied lacunae. Some patients destined to sustain vertebral fracture make cancellous bone with fewer osteocytes.
Introduction: Patient's with vertebral fracture have less bone than appropriate healthy controls, but other factors may contribute to bone fragility. Iliac cancellous osteocyte density declines with age in healthy women; we asked whether this variable differed between fracture patients and healthy controls.
Methods: Two groups of women were assembled. Forty-four (mean age, 66.2 years) had unequivocal evidence of bone fragility manifested as painful nontraumatic vertebral fracture, and 56 (mean age, 62.2 years) were skeletally healthy. All subjects underwent iliac bone biopsy. From archival embedded biopsy cores, new sections were stained with Goldner's trichrome, in which we enumerated osteocyte-occupied lacunae (stained), empty lacunae (unstained), and total lacunae per bone area.
Results: Cancellous osteocyte density was 34% lower in the fracture group than in the controls (p < 0.001); this difference was not a consequence of higher turnover, having less bone, or the small difference in age. The area under the receiver operating characteristic (ROC) curve for discrimination between the groups was >90% for osteocyte density and <75% for bone volume/tissue volume (BV/TV). The disease-related osteocyte deficit was accompanied by a proportionate reduction in empty lacunae and no change in percent occupied lacunae; therefor, it was not the result of premature death. Both superficial bone (<25 μm from the surface) and deep bone (>45 μm from the surface) were affected. In contrast, the age-related deficit is accompanied by an increase in empty lacunae and fall in percent osteocyte-occupied lacunae and occurs only in deep bone, but not in superficial bone.
Conclusions: In some patients destined to sustain spontaneous vertebral compression fracture, iliac cancellous bone is made with fewer osteocytes than normal; the mechanism of osteocyte incorporation into bone needs more detailed study. Osteocyte deficiency could contribute to bone fragility, either by impairing the detection of fatigue microdamage or by reducing canalicular fluid flow. Current practices of defining vertebral fracture based on morphometry alone regardless of symptoms, and diagnosing osteoporosis based on bone densitometry alone regardless of fracture history, should be reexamined.
The age-related increase in the incidence of vertebral compression fracture is commonly attributed to the age-related decline in vertebral bone mass.(1) This relationship is formalized in Consensus Conference definitions of osteoporosis(2) and World Health Organization (WHO) sponsored diagnostic criteria.(3) Loss of vertebral compressive strength is due mainly to loss of bone mass,(4) with some additional contributions from impaired cancellous bone connectivity(5) and low vertebral body size.(6) It is widely believed that patients with vertebral compression fracture are simply those at the lower end of the frequency distributions for these variables and do not differ in the process of bone remodeling or in bone cell function from age-matched healthy control subjects.(7) Nevertheless, it has been suggested several times over the past 40 years that the pathogenesis of vertebral fracture may not be so straightforward.(8–11)
All structural materials that undergo repetitive cyclical loading are susceptible to fatigue microdamage(12); bone is no exception, but is ordinarily able to repair itself.(13) Microdamage repair is carried out by new bone multicellular units (BMU)(14) and requires the participation of osteocytes,(15) the resident cells of bone derived from osteoblasts during the process of bone formation.(9) Undetected metal fatigue can have spectacular consequences,(13,16) which suggests that increased production and/or ineffective detection and repair of microdamage might contribute to bone fragility. Although no conclusive data are available,(17) several findings consistent with this possibility have been reported, including increased osteocyte death in appendicular bone(8) and both low turnover and retarded BMU progression in cortical(9) and cancellous bone.(10) As a result of such observations, Urist et al.(8) and more recently Frost(11) have proposed that physiologic osteoporosis (due only to age-related bone loss) and pathological osteoporosis (due also to retarded microdamage repair) should be distinguished.
The ilium is the only practicable site of bone biopsy; histomorphometry of such samples is a valuable investigative tool and has consistently shown lower values for indices of cancellous and cortical bone mass in patient's with osteoporosis, however defined, than in appropriate control subjects.(18,19) If the concept of pathological osteoporosis is valid, there should also be differences in some nonstructural indices of cellular activity. Conversely, the absence of such differences would tend to invalidate the concept. We have previously reported that, in healthy women, iliac cancellous osteocyte density declines with age in interstitial bone (>45 μm from the bone surface), which is relatively inaccessible to bone remodeling, and consequently older,(20) but not in superficial bone (<25 μm from the surface), which is more accessible to remodeling and consequently younger.(21) In this study, we have compared osteocyte density between white women with nontraumatic vertebral compression fracture and healthy white women of the same age.
MATERIALS AND METHODS
Patients with osteoporotic fracture fulfilled the entry criteria for a controlled trial of sodium fluoride treatment.(22) They were white women, ages 45–75 years, were at least 1 year postmenopause, and had at least one compression fracture with loss of posterior vertebral height, or at least two noncontiguous wedge fractures.(23) The abnormal vertebral body shape was readily apparent without measurement on a lateral spine radiograph, taken for evaluation of back pain that was spontaneous or after no more than trivial trauma. These criteria served to establish the existence of significant vertebral bone fragility and to rule out other causes of vertebral body deformation.(24) No patient had previously received sodium fluoride or was currently taking estrogen or any other treatment. Bone loss because of causes other than age or menopause was excluded. Of 663 potential candidates who were evaluated, 84 were enrolled in the study. Of these, 78 (93%) underwent transiliac bone biopsy before randomization (between August 1981 and December 1987).
Healthy control subjects were recruited in one of two ways.(25) Between 1980 and 1992, institutional employees and their friends and relatives responded to notices in circulars and newsletters. Between 1990 and 1993, members of Health Alliance Plan, the institutional HMO, who volunteered for a prospective study relating baseline indices of bone turnover to rate of bone loss,(26) were approached. Individuals in both groups volunteered for a study of normal bone histology using transiliac biopsy, which was approved by the Institutional Review Board for studies on human subjects at Henry Ford Hospital. All subjects were skeletally healthy according to standard criteria.(19,26) Of the 144 volunteers, 66 were postmenopausal white women.
In each subject, a cylindrical bone biopsy core was obtained(27) and processed, embedded, stained, sectioned, and examined by previously reported methods.(28) All biopsy cores were unfragmented with intact cortices at both ends of the samples. Sufficient histological material suitable for the present study was available from 44 fracture patients (mean age, 66.2 years) and 56 normal subjects (mean age, 62.2 years).
Sections (5 μm thick) were stained with Goldner's trichrome. Using a bright-field light microscope (×20 objective) equipped with a Bioquant System (R&M Biometrics Inc., Nashville, TN, USA), fields were scanned by rows and 10 unbroken regions were sequentially selected for counting the number of osteocyte-occupied lacunae (stained) and empty lacunae (unstained). The sum of both is the number of total lacunae. The lacuna is empty from the time the osteocyte dies and all its remnants are removed until it is filled with mineralized debris and becomes undetectable by standard methods.(24) For each region, the numbers of osteocytes, empty lacunae, and total lacunae were expressed per bone area (/mm2), and percent osteocyte-occupied lacunae calculated as osteocytes/total lacunae × 100. Our method is subject to two opposing errors, which should be random rather than systematic. First, some osteocytes may be lost during sectioning; second, some stainable osteocytes could have recently died by apoptosis.
To examine the effects of distance from the surface as an index of bone age, we measured osteocytes and empty lacunae separately in bone <25 μm from the surface (superficial) and >45 μm from the surface (deep). Based on wall thickness measurements in healthy women,(25) more than 95% of superficial bone was closer to the surface than the cement line, and more than 95% of deep bone was interstitial bone. For this purpose, the distance of each lacuna from the surface was measured, and this value was used for classification of the results.
Data are expressed as mean ± SD. For differences between vertebral fracture patients and normal subjects, we used Student's t-test. For differences of osteocyte density among sections of different age we used one-way ANOVA. The ability of the histological measurements to discriminate between fracture patients and normal subjects was examined by construction of receiver operating characteristic (ROC) curves,(29) using the MEDCALC program (MedCalc Software, Mariakerke, Belgium).
As expected, the patients with vertebral fractures had significantly less cancellous bone, with thinner, fewer, and more widely spaced trabeculae (Table 1). There were 34% fewer osteocytes per bone area in the fracture patients than in the healthy controls, a highly significant difference (Table 2). Empty lacunae and total lacunae were reduced in similar proportion so that percent osteocyte-occupied lacunae did not differ between the two groups. The frequency distribution of osteocyte density in individual subjects was markedly shifted to the left in the fracture patients, with many values below the lower limit in the normal subjects (Fig. 1). A similar shift was observed when the unit of observation was the histological field rather than the subject (data not shown). The CV of osteocyte density was significantly higher in the patients (41.5%) than in the controls (26.0%; p < 0.001).
Table Table 1. Structural Measurement in Healthy Postmenopausal Women and Women With Vertebral Fracture
Table Table 2. Comparison of Mean Values for Osteocyte and Lacunar Density Between Healthy Postmenopausal Women and Women With Vertebral Fracture
The ROC curves for osteocyte and total lacuna density and for bone volume/tissue volume (BV/TV; %) and trabecular number (Tb.N; #/mm) are compared in Fig. 2, and the corresponding values for sensitivity, specificity, and area under the curve are given in Table 3. Both osteocyte variables discriminated much better than the structural variables, a difference that persisted when all 144 biopsy specimens were included, not just the 100 in which osteocyte measurements were made (data not shown).
Table Table 3. ROC Analysis for Selected Histologic Variables
To address the possibility that fewer osteocytes were a consequence of having less bone, we compared the results in 29 subjects from each group who were selected to be matched individually for cancellous bone volume (normal, 17.5 ± 5.3%; fracture, 17.4 ± 5.2%). The absolute and relative differences between the groups for all variables were virtually identical to the complete series (Table 4). There was no correlation between osteocyte density and BV/TV in either group.
Table Table 4. Comparison of Mean Values for Osteocyte and Lacunar Density Between Healthy Postmenopausal Women and Vertebral Fracture Women With Matched Bone Volume
Because of their manner of recruitment, the biopsy-embedded blocks had been in storage for a substantially longer time in one set of normal subjects than in the other, but the values did not differ between them. Furthermore, when the analysis was restricted to the older blocks, the differences between the groups were very similar (Table 5). Consequently, the differences between the normal and fracture groups were not the result of differential changes during sample storage.
Table Table 5. Effect of Section Age
To exclude any contribution from the small difference in age, we calculated for each fracture patient the value for osteocyte density predicted from the exponential regression on age in the normal subjects.(21) The difference between observed and predicted values remained highly significant (p < 0.001; Table 2).
In both fracture patients and controls, there were significantly fewer osteocytes and total lacunae and significantly more empty lacunae in deep bone than in superficial bone (Table 6). In both superficial and deep bone, the fracture patients had fewer osteocytes and total lacunae than the healthy controls. The relative osteocyte deficit in the patients was greater in superficial bone (38%) than in deep bone (25%), although the absolute values were higher. However, this difference was not significant.
Table Table 6. Comparison of Mean Values for Osteocyte and Lacunar Density in Superficial and Deep Bone Between Healthy Postmenopausal Women and Women With Vertebral Fracture
We found a substantial reduction in osteocyte density in iliac cancellous bone in postmenopausal women with osteoporotic vertebral compression fracture; the deficit was not the result of having less bone or differences in age or in duration of sample storage. In healthy women, there is a significant inverse relationship between osteocyte density in superficial bone and bone formation rate (BFR)(30); to explain the osteocyte deficit, cancellous BFR would have to have been 28% higher in the fracture patients, but in an earlier study, we found it to be 35% lower,(31) and in the present study, we found it to be 17% lower (data not shown). Our findings contrast with a report of increased iliac osteocyte density in fracture patients,(32) but the control measurements were from autopsy samples, and deplasticized sections had been stored for about 4 years longer in the fracture patients and could have undergone greater shrinkage. In our study, the areas under the ROC curves were larger for osteocyte and total lacuna density than for any other histological measurement and larger than for spinal DXA measurement, for which values have ranged from 79%(33) to 89%.(34) The deficit was present in both superficial bone (which is younger because more frequently remodeled) and in deep bone (which is older because less frequently remodeled). In this respect, the disease-related osteocyte deficit differed from the age-related deficit in normal women, which was restricted to deep bone and did not occur in superficial bone.(21) A further difference is that the age-related decline in osteocyte density is accompanied by an increase in empty lacunae and a fall in percent osteocyte-occupied lacunae,(21) whereas the deficit in the fracture patients was accompanied by a fall in empty lacunae and no change in percent osteocyte-occupied lacunae.
The existence of a large difference between vertebral fracture patients and healthy controls that is unrelated to bone structure or remodeling suggests that there is more to vertebral fracture pathogenesis than is currently believed; the concept of pathological osteoporosis seems to have some merit. There may indeed be a real disease hidden within the thicket of age-related bone loss. However, to develop this notion further, it is necessary to consider how such a large deficit in osteocytes could have occurred and how it could contribute to bone fragility. An osteocyte deficit could arise if too many died prematurely, but when an osteocyte dies, it leaves an empty lacuna, which is why during normal aging, osteocyte density and empty lacuna density change in opposite directions.(21) When osteocytes and empty lacunae are reduced together in similar proportion, as in the fracture subjects of this study, too few osteocytes must have been present when the bone was formed. This abnormality was evident in more recently formed superficial bone, but was also present in deep bone, some of which was probably made during childhood.(35) Osteoblasts committed to become osteocytes make less bone matrix(36); some may undergo premature apoptosis (37,38), either as a primary defect such as failure of matrix metalloproteinases to activate latent TGF-β,(39) or secondary to faulty burial signals. Clearly, much more needs to be learned about the mechanism of osteocyte incorporation into bone.
We cannot rule out the possibility that osteocyte deficiency is an indirect marker for some subtle change in bone material properties that reduces strength, but it seems reasonable to consider more direct effects. Osteocytes play a key role in the detection of fatigue microdamage(15) so that a 30% deficit in osteocytes could contribute to bone fragility by compromising repair of microdamage and making it more likely that it would accumulate into a macrofracture. A lifelong tendency to delayed fatigue damage repair would not become clinically manifest until sufficient bone had been lost to increase the loads on, and consequent strains within, the remaining bone, which would increase the microdamage burden above some critical level. This would explain why spontaneous vertebral fracture appears to need both less bone and fewer osteocytes. However, there is probably considerable redundancy in the osteocyte network,(38) and whether a 30% deficit would be sufficient to lead to fatigue damage accumulation has not been determined. Furthermore, although microfractures occur in vertebral cancellous bone(40) and are sometimes suggested to be fatigue fractures,(41) there is no evidence that they are caused by fatigue; indeed, identical lesions can be produced by a single episode of experimental compression.(42)
Although the details remain unclear, there is general agreement that osteocytes are needed to maintain fluid flow within the lacunar-canalicular network that is widely and evenly distributed throughout the bone.(38) If, as we have inferred, in some patients destined to sustain vertebral fracture, cancellous bone is made with fewer osteocytes, the canalicular network would be similarly attenuated, although not demonstrable by the methods we used. Conceivably reduced fluid flow could alter the volumetric proportion of water and crystal within the mineral, too little water increasing bone brittleness,(43) and too much water decreasing bone stiffness.(44) Such a mechanism might account for the acute effect of osteocyte death to reduce vertebral compressive strength in mice independent of changes in bone mass or architecture.(45) Evidently much more work is needed to explore these possibilities.
Whatever the mechanism for the osteocyte deficit and its contribution to bone fragility, our results have important implications for practice and research in osteoporosis. Traditionally a vertebral fracture was manifest as an episode of significant back pain, but to facilitate the conduct of clinical trials it has become customary to redefine a fracture as a measurable change in vertebral body shape, regardless of symptoms.(23) All the fractures in our study conformed to the traditional description, and the difference we found might not apply to asymptomatic changes that we prefer to call deformities rather than fractures.(23) All the currently accepted risk factors for vertebral “fracture” can in principle be assessed by noninvasive methods, but our data indicate that about one-half of the patients with a genuine fracture have a risk factor that cannot at present be detected noninvasively. There are evidently limitations in using bone densitometry as the sole diagnostic criterion for “osteoporosis,” because this obscures the possibility that patients with vertebral fracture may differ in important ways from similarly osteopenic subjects without fracture. We believe that clinical investigators should pay more attention to the clinical manifestations of the events that they study, so that research subjects can be classified not just according to their spine X-rays, but also according to their fracture history.
We thank Dr Michael Kleerekoper for his contributions to the definition of vertebral fracture, his insistence on restricting the fluoride trial to patients with genuine fracture, and his support for the importance of bone biopsy in the evaluation of such patients. This study was partly supported by a grant from the National Institutes of Health (DK 43858).