Abnormal Bone Microarchitecture and Evidence of Osteoblast Dysfunction in Premenopausal Women with Idiopathic Osteoporosis

Abstract

Context:

Idiopathic osteoporosis (IOP) in premenopausal women is an uncommon disorder of uncertain pathogenesis in which fragility fractures occur in otherwise healthy women with intact gonadal function. It is unclear whether women with idiopathic low bone mineral density and no history of fragility fractures have osteoporosis.

Objective:

The objective of the study was to elucidate the microarchitectural and remodeling features of premenopausal women with IOP.

Design:

We performed transiliac biopsies after tetracycline labeling in 104 women: 45 with fragility fractures (IOP), 19 with idiopathic low bone mineral density (Z score ≤−2.0) and 40 controls. Biopsies were analyzed by two-dimensional quantitative histomorphometry and three-dimensional microcomputed tomography. Bone stiffness was estimated using finite element analysis.

Results:

Compared with controls, affected women had thinner cortices; fewer, thinner, more widely separated, and heterogeneously distributed trabeculae; reduced stiffness; and lower osteoid width and mean wall width. All parameters were indistinguishable between women with IOP and idiopathic low bone mineral density. Although there were no group differences in dynamic histomorphometric remodeling parameters, serum calciotropic hormones, bone turnover markers, or IGF-I, subjects in the lowest tertile of bone formation rate had significantly lower osteoid and wall width, more severely disrupted microarchitecture, lower stiffness, and higher serum IGF-I than those in the upper two tertiles, suggesting that women with low turnover IOP have osteoblast dysfunction with resistance to IGF-I. Subjects with high bone turnover had significantly higher serum 1,25 dihydroxyvitamin D levels and a nonsignificant trend toward higher serum PTH and urinary calcium excretion.

Conclusions:

These results suggest that the diagnosis of IOP should not require a history of fracture. Women with IOP may have high, normal or low bone turnover; those with low bone turnover have the most marked deficits in microarchitecture and stiffness. These results also suggest that the pathogenesis of idiopathic osteoporosis is heterogeneous and may differ according to remodeling activity.

Idiopathic osteoporosis (IOP) is an uncommon disorder that affects young, otherwise healthy individuals with intact gonadal function and no secondary cause of bone loss (1). The pathogenesis is uncertain and may differ by gender. In men with IOP (2–8), most studies report reduced bone formation on transiliac bone biopsy (5, 6, 9–11), although several find increased (12, 13) or heterogeneous remodeling (14). Male IOP is also commonly associated with low serum levels of IGF-I that correlate directly with histomorphometric parameters of bone formation (6, 9, 10). Thus, in men, the pathogenesis of IOP may be related to impaired osteoblast function, proliferation (15), or recruitment to remodeling sites (16).

There have been relatively few studies of IOP in young women. Therefore, we initiated a comprehensive National Institutes of Health-supported, cross-sectional study of women with IOP. We included women with documented adult, fragility fractures, regardless of their areal bone mineral density (aBMD) measurements (the IOP group) and women with low spine or hip aBMD with no history of adult fragility fractures [the idiopathic low bone mineral density (ILBMD) group] (17). Informed by prior smaller studies, we hypothesized that affected women would have low bone formation by dynamic histomorphometry (18) but normal serum IGF-I compared with controls (19). Because isolated low bone mineral density (BMD) measurements in young women may be related to low peak bone mass or small bone size and may not reflect abnormal bone quality (20), we also hypothesized that women with low aBMD and no history of adult fragility fractures would have better bone microarchitecture and stiffness than those with fractures.

We recently reported the demographic, anthropometric, densitometric, and biochemical results of this study (17). Compared with 40 concurrently recruited, age-matched controls with normal BMD and no fractures, the 64 subjects weighed less and had lower body mass index (BMI). As hypothesized, serum IGF-I did not differ from controls, nor were there differences in serum estradiol or vitamin D metabolites. Although serum PTH levels were within the normal range, they were significantly higher in subjects than controls. In contrast to our expectation of low bone formation, most bone turnover markers (BTM) did not differ between subjects and controls. However, tartrate-resistant acid phosphatase (TRAP)-5b, a resorption marker, was significantly higher in subjects and correlated directly with serum PTH. Except for lower weight, BMI, and later menarche, subjects with low aBMD and those with fractures were clinically and biochemically indistinguishable. By high-resolution peripheral quantitative computed tomography (HRpQCT), subjects had microarchitectural disruption and reduced stiffness of the distal radius and tibia that were also indistinguishable between those with fractures and those with low aBMD (21, 22). Herein we report the results of transiliac bone biopsies in controls and both groups of subjects.

Patients and Methods

Patient population

Premenopausal women, aged 18–48 yr, were recruited at Columbia University Medical Center (New York, NY) and Creighton University (Omaha, NE) by advertisement and self- or physician referral. The ILBMD group included women with low aBMD by dual-energy x-ray absorptiometry (DXA; T score ≤−2.5 or Z score ≤−2.0) at the spine or hip and no history of adult low trauma fracture. The IOP group included women with a documented low-trauma fracture after age 18 yr, regardless of whether aBMD was low. Fractures were ascertained by review of radiographs or reports and categorized as low trauma (equivalent to a fall from a standing height or less) after review by physician panel (E.S., A.C., R.R.R., E.M.S.). Skull and digit fractures were excluded. IOP subjects were evaluated more than 3 months after their most recent fracture. Concurrently recruited controls had normal aBMD by DXA (T score ≥−1.0 or Z score ≥−1.0) and no history of adult low-trauma fractures.

Inclusion and exclusion criteria were previously reported (17, 21). We defined premenopausal status as regular menses off hormonal contraception and early follicular phase FSH levels less than 20 mIU/ml. Secondary causes of osteoporosis were excluded by detailed history and physical and biochemical evaluation in subjects and controls (17): estrogen deficiency, eating disorders associated with amenorrhea, endocrinopathies, celiac or other gastrointestinal diseases, abnormal mineral metabolism, marked hypercalciuria [>300 mg/g creatinine (Cr)], and drug exposures. Women with serum 25-hydroxyvitamin D (25-OHD) levels below 10 ng/ml were excluded. All subjects provided written informed consent. The institutional review boards of both institutions approved these studies.

Laboratory assessments

To exclude secondary causes of osteoporosis, fasting morning blood and a 24-h urine collected during the early follicular phase of the menstrual cycle on the participant's usual diet and supplements were analyzed in a clinical laboratory (Quest Diagnostics, Madison, NJ). Serum and urine were stored at −80 C for batch analyses of serum estradiol, SHBG, 25-OHD, 1,25 dihydroxyvitamin D [1,25(OH)₂D], PTH, N-terminal propeptides of procollagen type 1 (P1NP); bone-specific alkaline phosphatase (BAP), osteocalcin, C-telopeptide (CTX), TRAP5b, IGF-I, and IGF binding protein-3, and 24-h urinary calcium, as previously described (17).

Areal bone mineral density

aBMD was measured by DXA (QDR-4500; Hologic Inc., Walton, MA) at Columbia and Creighton University medical centers as previously described (17).

Transiliac bone biopsy

After double labeling with tetracycline, transiliac biopsy was performed using a Bordier-type trephine with an inner diameter of 7.5 mm (23). The specimens were fixed and dehydrated in ethanol, subjected to microcomputed tomography (μCT), and then embedded in polymethylmethacrylate for quantitative histomorphometry.

Microcomputed tomography

Intact biopsies were scanned by μCT (μCT 40; Scanco Medical AG, Brüttisellen, Switzerland) with the long axis oriented along the rotation axis of the scanner at an isotropic, nominal resolution of 8 μm, as previously described in detail (24, 25). Trabecular indices were determined using a direct three-dimensional (3D) approach (26): bone volume density (BV/TV), trabecular thickness (Tb.Th), trabecular separation (Tb.Sp), and trabecular number (Tb.N). Connectivity density is calculated based on the Euler characteristics of the structure normalized to tissue volume. Higher values represent greater trabecular connectivity. Structure model index quantifies the number of plates and rods by 3D image analysis. For an ideal plate and rod structure, the structure model index value is 0 and 3, respectively (27).

Microfinite element analysis (μFE) of μCT images

Apparent Young's modulus of the biopsies was calculated by μFE of μCT trabecular images as previously described (25). A rectangular volume of interest (640*640*300 voxels) was isolated in the center of the biopsy image; each voxel was converted to an eight-node brick element. A single load case was applied, representing uniaxial compression in the medial-lateral direction. A custom in-house parallel conjugate gradient finite element solver with multilevel preconditioning (28) was used.

Bone histomorphometry

After μCT, biopsy specimens were embedded, sectioned (7 μm for stained and 20 μm for unstained sections) and stained (Goldner trichrome, toluidine blue, and solochrome) according to established procedures (25). Histomorphometry was performed with a digitizing image-analysis system (OsteoMeasure, version 4.00C; OsteoMetrics, Inc., Atlanta, GA). All variables were calculated according to American Society for Bone and Mineral Research recommendations (29). Cancellous BV/TV, Tb.Th, Tb.N, and Tb.Sp were derived from two-dimensional (2D) measurements of total tissue area, cancellous bone area, and perimeter (30). Cortical width was the average distance between periosteal and endocortical surfaces for both cortices. Measurements of structural variables were performed at ×20 magnification.

Statistical analysis

Statistical analyses were performed using SAS software (SAS Institute, Cary, NC). To limit type I error caused by multiple comparisons, we used ANOVA models as an omnibus test to examine overall differences among the three groups (controls, IOP, ILBMD) for each variable. Unless noted, main between-groups comparisons are presented only for variables that showed significant overall differences among the three groups by ANOVA. Between-groups comparisons were conducted using Student's t tests. Pearson correlation coefficients were calculated to test associations between variables. If variables were not normally distributed (Kolmogorov-Smirnov test), Spearman correlations were used. Multivariate linear and logistic regression analyses were used to control for covariates. Because the IOP and ILBMD groups had similar bone volume, microstructure, and remodeling findings, they were combined for correlation analyses. All data are expressed as mean ± sd. Results were considered significant with P < 0.05.

Results

Characteristics of the study population

Forty-five women had low-trauma adult fractures (IOP group) and 19 had low aBMD but no low trauma adult fractures (ILBMD group). The number of fractures ranged from one to 12, including vertebral, rib, hip, pelvic, forearm, humerus, lower leg, ankle, and metatarsal fractures; 25 women had multiple fractures. Mean age at first fracture was 30 ± 9 yr and mean time since most recent fracture was 4 ± 4 yr.

Clinical, densitometric, and biochemical characteristics (Table 1) were recently reported (17). Because inclusion was based on BMD criteria, aBMD was lowest in the ILBMD group, but the IOP group also had substantially lower aBMD than controls. BMD differences remained significant (P < 0.03) after controlling for BMI. Reproductive hormones, indices of mineral metabolism, IGF-I, and most BTM did not differ among groups. Although normal, serum PTH and TRAP5b were significantly higher in both affected groups and were directly associated with each other in all groups (R = 0.35–0.55; P = 0.005–0.03).

Table 1.

Characteristics of the study participants

	Control (n = 40)	IOP (n = 45)	ILBMD (n = 19)
Anthropometric characteristics
Age (yr)	37.3 ± 8.2	37.0 ± 7.7	39.6 ± 6.3
Height (cm)	165.4 ± 7.3	163.8 ± 7.2	162.2 ± 5.7
Weight (kg)	70.7 ± 14.6	63.0 ± 14.9a	57.1 ± 10.2c
BMI (kg/m2)	25.8 ± 4.7	23.4 ± 4.9a	21.6 ± 3.5b
BMD (g/cm2)
Lumbar spine (L1-4)	1.099 ± 0.093	0.875 ± 0.139c	0.793 ± 0.078c,d
Total hip	0.984 ± 0.075	0.792 ± 0.138c	0.736 ± 0.090c
Femoral neck	0.878 ± 0.075	0.677 ± 0.129c	0.615 ± 0.089c
Distal radius (one third)	0.721 ± 0.051	0.692 ± 0.047b	0.679 ± 0.045b
BMD (Z score)
Lumbar spine	0.74 ± 0.88	−1.33 ± 1.28c	−2.06 ± 0.73c,d
Total hip	0.48 ± 0.65	−1.09 ± 1.12c	−1.48 ± 0.69c
Femoral neck	0.39 ± 0.74	−1.31 ± 1.15c	−1.77 ± 0.76c
Distal radius (one third)	0.68 ± 0.89	0.26 ± 0.82a	0.01 ± 0.85b
Indices of mineral metabolism and hormones
Calcium (mg/dl)	9.0 ± 0.3	9.1 ± 0.6	8.9 ± 0.3
25-OHD (ng/ml)	30 ± 13	37 ± 16	38 ± 12
1,25(OH)2D (pg/ml)	49 ± 11	55 ± 19	56 ± 12a
PTH (pg/ml)	22 ± 9	29 ± 11b	30 ± 16a
24-h Urinary calcium (mg/g Cr)	162 ± 68	184 ± 89	211 ± 70
FSH (mIU/ml)	7.2 ± 3.5	7.5 ± 3.7	7.5 ± 2.5
Total estradiol (pg/ml)	39 ± 34	38 ± 29	35 ± 24
Free estradiol (pmol/liter)	1.7 ± 1.4	1.9 ± 3.3	1.4 ± 0.9
IGF-I (ng/ml)	184 ± 55	182 ± 54	172 ± 56
Bone turnover markers
P1NP (μg/liter)	47 ± 18	47 ± 18	52 ± 18
BAP (U/liter)	20 ± 6	22 ± 7	22 ± 5
Osteocalcin (ng/ml)	15 ± 8	16 ± 8	16 ± 6
CTX (ng/ml)	0.287 ± 0.19	0.354 ± 0.19	0.333 ± 0.13
TRAP5b (U/liter)	1.6 ± 1.0	2.2 ± 1.2b	2.3 ± 1.0b
Urine NTx (nMBCE/mMCr)	37 ± 16	44 ± 23	36 ± 19

nMBCE/mMCr, Nanomoles of bone collagen equivalent per liter per millimoles of Cr per liter; NTx, N-telopeptide.

^aP < 0.05 vs. controls.

^bP < 0.01 vs. controls.

^cP < 0.001 vs. controls.

^dP < 0.01 IOP vs. ILBMD.

Bone structure and stiffness

By 2D histomorphometry, cancellous BV/TV and Tb.Th were significantly and comparably lower in the IOP and ILBMD subjects than controls (Table 2). Cortices were thinner than controls (Fig. 1A), by 24% in IOP (P < 0.0001) and 21% in ILBMD (P = 0.07) subjects. Cortical porosity was significantly lower in both affected groups than controls.

Table 2.

Bone microstructure and remodeling measured by quantitative 2D histomorphometry of transiliac crest bone biopsy samples

	Control (n = 40)	IOP (n = 45)	ILBMD (n = 19)
Trabecular (Tb) and cortical microstructure by 2D histomorphometry
BV/TV (%)	21.6 ± 5.4	18.3 ± 4.7b	16.6 ± 3.8c
Tb bone surface/total volume (%)	3.3 ± 0.5	3.2 ± 0.7	3.0 ± 0.5
Tb.N (n/mm)	1.6 ± 0.3	1.6 ± 0.3	1.5 ± 0.2
Tb width (Tb.Wi; μm)	133 ± 24	115 ± 26b	113 ± 23b
Tb.Sp (μm)	621 ± 110	645 ± 175	680 ± 114
Cortical width (μm)	866 ± 212	662 ± 235c	687 ± 376
Cortical porosity (% cortical area)	7.4 ± 3.1	5.9 ± 3.2a	5.4 ± 2.6a
Static and dynamic cancellous bone surface remodeling
MdPm (%)	3.2 ± 2.7	4.1 ± 3.5	3.7 ± 1.7
Osteoid width (# lamellae)	4.1 ± 1.0	3.6 ± 1.0	3.7 ± 1.1
Wall width (μm)	35.0 ± 3.6	33.7 ± 3.8	33.7 ± 4.0
Osteoid perimeter (OPm; %)	5.4 ± 3.8	5.8 ± 4.6	6.2 ± 2.9
MdPm/OPm (%)	62.0 ± 24.9	70.9 ± 27.9	66.1 ± 28.3
Mineral apposition rate (μm/d)	0.64 ± 0.09	0.64 ± 0.09	0.59 ± 0.10
BFR/BS (mm2/mm/yr)	0.008 ± 0.007	0.010 ± 0.009	0.008 ± 0.004
BFR/BV (mm3/mm3/yr)	0.119 ± 0.088	0.173 ± 0.147	0.147 ± 0.075
AjAR (μm/d)	0.492 ± 0.389	0.566 ± 0.458	0.383 ± 0.192
Eroded perimeter (%)	4.3 ± 1.8	4.3 ± 1.9	4.4 ± 1.5
Bone resorption rate (mm2/mm/yr)	0.003 ± 0.003	0.003 ± 0.002	0.002 ± 0.001
Formation period (Yr)	0.263 ± 0.165	0.238 ± 0.273	0.182 ± 0.096
Resorption period (Yr)	0.277 ± 0.264	0.214 ± 0.241	0.140 ± 0.088
Activation frequency (cycle/yr)	0.291 ± 0.237	0.365 ± 0.303	0.312 ± 0.164

P values for between-groups comparisons (by Student's t tests) are presented only for variables that initially showed significant overall differences among the three groups based on ANOVA models. Significant overall differences among the three groups based on ANOVA models were not found for any cancellous surface remodeling variable. For measure of bone formation and resorption rate, the conversion from square millimeters per millimeter per year to square micrometer per micrometer per day is: (square millimeters per millimeter per year × 1000)/365.25 = square micrometer per micrometer per day.

^aP < 0.05 vs. controls.

^bP < 0.01 vs. controls.

^cP < 0.001 vs. controls.

Fig. 1.

Representative images from histomorphometric analyses of cortical width (×40 magnification) (A) and μCT analyses of trabecular bone structure performed on transiliac crest bone biopsy samples (B) from controls, IOP, and ILBMD subjects.

By μCT, IOP and ILBMD subjects had similar microarchitectural deficits compared with controls (Fig. 1B): lower BV/TV, fewer and more widely spaced trabeculae, lower connectivity density, and a less homogeneously distributed trabecular network (Fig. 2, A–G). By μFE, trabecular stiffness was comparably and significantly lower in IOP and ILBMD subjects than controls (Fig. 2H).

Fig. 2.

Trabecular microstructure measured by μCT performed on transiliac crest bone biopsy samples from controls (white), IOP (black), and ILBMD (gray) subjects. *, P < 0.05; **, P < 0.01; ***, P < 0.001 vs. controls; #, P < 0.05 IOP vs. ILBMD.

Older age was associated with lower μCT Tb.N in subjects (r = −0.252; P < 0.05) and tended to be associated with lower μCT BV/TV in controls (r = −0.306; P = 0.06). Higher weight and BMI were associated with lower BV/TV by μCT in subjects (weight: r = −0.337, P = 0.006; BMI: r = −0.248, P = 0.048) and controls (weight: r = −0.338, P = 0.04 and BMI: r = −0.332; P = 0.04).

Bone surface remodeling

Static and dynamic indices of bone surface remodeling were examined on the cancellous (Table 2), intracortical, and endocortical surfaces (data not shown). There were no significant differences in osteoid width, wall width, mineralizing perimeter (MdPm), mineral apposition rate, adjusted apposition rate (AjAR), bone formation rate (BFR/BS), or activation frequency between either group of subjects and controls. In controls, older age was associated with lower cancellous MdPm (r = −0.468; P = 0.002), BFR/BS (r = −0.415; P = 0.008), AjAR (r = −0.548; P = 0.0002), activation frequency ((r = −0.395; P = 0.01), and bone resorption rate (r = −0.455; P = 0.003) and longer formation (r = 0.566; P = 0.0001) and resorption periods (r = 0.417; P = 0.007) at the cancellous surface. In subjects, the only significant association was older age with AjAR (r = −0.252; P = 0.04).

Relationships between BTM, calciotropic hormones, IGF-I and bone remodeling, structure, and stiffness

Because there were few differences between the IOP and ILBMD groups, they were combined in subsequent analyses. Serum BTM correlated positively with most dynamic histomorphometric remodeling parameters in controls and subjects (Table 3). The strongest correlations with BFR/BS were found for serum CTX. Based on the regression results, for every 0.100 ng/ml higher serum CTX, BFR/BS was higher by (estimate ± sem) 0.027 ± 0.002 mm²/mm · yr in controls and 0.026 ± 0.002 mm²/mm · yr in the subjects.

Table 3.

Relationships between serum bone turnover markers and histomorphometric measures of bone turnover at the cancellous surface

Biochemical parameter	Histomorphometric parameter	Controls	Subjects
CTX	Osteoid width	0.052	0.075
	Wall width	0.348	−0.037
	Mineralizing perimeter	0.686c	0.511c
	Mineral apposition rate	0.243	0.202
	BFR/BS	0.622c	0.527c
	Activation frequency	0.581c	0.538c
	Bone resorption rate	0.630c	0.386b
TRAP5b	Osteoid width	0.247	−0.187
	Wall width	0.168	−0.026
	Mineralizing perimeter	0.593c	0.342b
	Mineral apposition rate	0.377a	0.216
	BFR/BS	0.566c	0.367b
	Activation frequency	0.558b	0.396b
	Bone resorption rate	0.366a	0.314a
Osteocalcin	Osteoid width	−0.062	−0.074
	Wall width	0.291	0.217
	Mineralizing perimeter	0.439b	0.430b
	Mineral apposition rate	0.021	0.271a
	BFR/BS	0.374a	0.451b
	Activation frequency	0.311	0.432b
	Bone resorption rate	0.365a	0.397b
BAP	Osteoid width	0.091	−0.086
	Wall width	0.182	0.030
	Mineralizing perimeter	0.340a	0.417b
	Mineral apposition rate	0.213	0.346b
	BFR/BS	0.329a	0.447b
	Activation frequency	0.333a	0.450b
	Bone resorption rate	0.346a	0.304a

^aP < 0.05.

^bP < 0.01.

^cP ≤ 0.0001.

BV/TV by μCT and stiffness by μFE did not correlate with PTH, 25-OHD, or free estradiol in either group (Table 4); both were directly associated with urinary calcium, and μFE was directly associated with 1,25(OH)₂D in subjects but not controls. After adjusting for age-associated declines in IGF-I, BV/TV and stiffness were inversely associated with IGF-I in subjects but not controls. BV/TV and stiffness correlated directly with some BTM in controls but not subjects and with dynamic remodeling parameters in both controls and subjects; higher wall width and BFR/BS were most consistently associated with higher BV/TV and stiffness. Static and dynamic remodeling parameters did not correlate with PTH, 25-OHD, or free estradiol in either group (data not shown).

Table 4.

Correlations with cancellous BV/TV by μCT and stiffness by μFE

	Controls		Subjects
	BV/TV	μFE (E)	BV/TV	μFE (E)
Biochemistries and hormonal measures
PTH	0.105	0.138	0.129	0.113
25-OHD	0.190	0.246	0.112	0.204
1,25(OH)2D	−0.040	−0.055	0.152	0.330a
24-h urinary calcium	−0.135	−0.132	0.319a	0.374b
Free estradiol	−0.083	−0.087	−0.023	−0.082
IGF-I (adjusted for age)	−0.128	−0.125	−0.313a	−0.287a
Bone turnover markers
P1NP	0.300	0.311	0.056	−0.021
BAP	0.212	0.340a	0.225	0.179
Osteocalcin	0.299	0.310	0.158	0.070
CTX	0.392a	0.402a	0.248	0.183
TRAP5b	0.579c	0.521b	0.224	0.188
Cancellous surface remodeling
Osteoid width	0.288	0.256	0.245	0.308a
Wall width	0.394a	0.352a	0.440b	0.357b
Mineralized perimeter	0.411b	0.406a	0.295a	0.250a
Mineral apposition rate	0.268	0.294	0.310a	0.251
BFR/BS	0.418b	0.411a	0.345b	0.302a
Activation frequency	0.367a	0.362a	0.304a	0.262a
Bone resorption rate	0.289	0.310	0.442b	0.370b

^aP < 0.05.

^bP < 0.01.

^cP ≤ 0.0001.

Characteristics of subjects by remodeling status

Osteoid width and wall width, two static parameters that reflect osteoid synthesis by osteoblasts, were significantly lower in subjects than controls at cancellous, endocortical, and intracortical surfaces (Fig. 3, A and B). Hypothesizing that defective bone formation might contribute to the pathogenesis of IOP, we compared subjects according to tertiles of BFR/BS (Table 5). Results were the same if BFR/BV was used (data not shown). Those in the lowest tertile, the low turnover (LT) group, did not differ from the high turnover (HT) group in terms of fractures, age, BMI, or aBMD. The LT group also had significantly lower cancellous osteoid width (Fig. 3C) and both cancellous and endocortical wall width (Fig. 3D) than the HT group. Interestingly, the LT group had substantially lower trabecular BV/TV and stiffness than the upper tertile. In contrast, although the HT group had lower BV/TV and stiffness than controls, the differences were not significant (P = 0.2 and 0.06, respectively); similarly, osteoid and wall width did not differ from controls.

Fig. 3.

The IOP and ILBMD groups were combined for these analyses. A and B, Mean osteoid width (OWi) and wall width (WWi) at the cancellous, intracortical, and endocortical bone surfaces in controls (white) compared with subjects (black). *, P < 0.05 for comparison of controls vs. subjects. C and D, Mean OWi and WWi for subjects in the three tertiles of BFR/BS [low (dark gray), middle (middle gray), high (lightest gray)] are shown as percentage difference from controls at each of the cancellous, intracortical, and endocortical bone surfaces. E and F, Mean BV/TV by μCT and stiffness by μFE for subjects in the three tertiles of BFR/BS are shown as percentage difference from controls. *, P < 0.05, **, P < 0.01, and ***, P ≤ 0.0001, subjects vs. controls; #, P < 0.05 and ##, P < 0.01 for low (LT) compared with high tertile (HT).

Table 5.

Analysis by tertiles of cancellous BFR/BS among affected subjects (means ± sd)

	Controls	Subjects
		Tertile 1 (lowest)	Tertile 2 (middle)	Tertile 3 (highest)
CanBFR/BS (mm2/mm/yr)	0.008 ± 0.007	0.003 ± 0.002	0.008 ± 0.001	0.017 ± 0.010
n	40	21	22	21
Percent with fractures	0	76	68	67
Age (yr)	37.3 ± 8.2	37.7 ± 6.7	38.0 ± 6.7	37.5 ± 8.9
BMI (kg/m2)	25.8 ± 4.7	23.8 ± 5.5	22.9 ± 3.9a	22.0 ± 4.2b
BMD (g/cm2)
Lumbar spine	1.099 ± 0.093	0.893 ± 0.149c	0.843 ± 0.119c	0.815 ± 0.109c
Total hip	0.984 ± 0.075	0.812 ± 0.149c	0.742 ± 0.108c	0.778 ± 0.119c
Femoral neck	0.878 ± 0.075	0.695 ± 0.150c	0.622 ± 0.095c	0.664 ± 0.106c
One third radius	0.721 ± 0.051	0.686 ± 0.051a	0.688 ± 0.050a	0.683 ± 0.053b
μCT BV/TV (%)	23.7 ± 7.7	16.7 ± 4.5b	21.0 ± 6.5	21.3 ± 5.7e
μCT Tb.N (n/mm)	1.8 ± 0.35	1.5 ± 0.16c	1.6 ± 0.24a	1.5 ± 0.22b
μCT Tb.Th (μm)	161 ± 38	154 ± 35	163 ± 36	177 ± 37
μCT Tb.Sp (μm)	624 ± 61	714 ± 76c	678 ± 67b	739 ± 78c
μCT Tb.Sp, sd (μm)	176 ± 19	188 ± 28a	190 ± 23a	210 ± 36c,d
Stiffness by μFE (MPa)	545 ± 350	273 ± 173b	386 ± 202	398 ± 217d
Can OWi (no. lamellae)	4.1 ± 1.0	3.1 ± 1.4b	3.9 ± 0.7	3.8 ± 0.7d
Can WWi (μm)	35.0 ± 3.6	32.6 ± 4.2a	33.2 ± 3.6	35.5 ± 3.3d
PTH (pg/ml)	22.4 ± 9.1	24.3 ± 10.1	33.3 ± 14.7b	30.5 ± 10.0b
Serum calcium (mg/dl)	9.0 ± 0.3	8.9 ± 0.3	9.2 ± 0.8	8.9 ± 0.3
25-OHD (ng/ml)	30 ± 13	35 ± 17	38 ± 14	38 ± 14
1,25(OH)2D (pg/ml)	49 ± 11	49 ± 20	57 ± 16a	60 ± 12b,d
24-h urine Ca (mg/g Cr)	162 ± 68	177 ± 80	171 ± 55	230 ± 101b
Free estradiol (pmol/liter)	1.7 ± 1.4	1.2 ± 1.2	2.4 ± 4.4	1.5 ± 1.1
IGF-I (ng/ml)	184 ± 55	193 ± 46	164 ± 58	179 ± 57
BAP (U/liter)	20.3 ± 6.4	19.5 ± 5.4	21.8 ± 5.7	24.4 ± 6.3a,e
Osteocalcin (ng/ml)	15.4 ± 7.9	11.7 ± 5.4	15.8 ± 6.3	19.8 ± 8.3d
P1NP (μg/liter)	46.6 ± 17.9	41.5 ± 14.1	47.3 ± 17.5	55.4 ± 19.7
CTX (ng/ml)	0.287 ± 0.186	0.255 ± 0.154	0.364 ± 0.114	0.434 ± 0.206d
TRAP5b (U/liter)	1.6 ± 1.0	1.8 ± 1.1	2.3 ± 1.0a	2.8 ± 1.1c,d

Results of between-groups comparisons are shown when overall ANOVA models for between-groups differences were significant. Can, Cancellous; Owi, osteoid width; Wwi, wall width.

^aP < 0.05;

^bP < 0.01;

^cP ≤ 0.0001 in controls vs. tertiles 1, 2 and 3;

^dP < 0.05;

^eP < 0.01; tertile 1 vs. tertile 3.

As expected, serum BTM were lower in the LT than the HT group. Age-adjusted serum IGF-I was significantly higher in the LT group than the middle and upper tertiles combined (195 ± 43 vs. 170 ± 43; P = 0.04). The HT group had significantly higher 1,25(OH)₂D, with no differences in serum 25-OHD or calcium; urinary calcium excretion (P = 0.07) and PTH (P = 0.06) were also higher although the differences were not significant.

Discussion

This is the first prospective study to compare transiliac bone biopsies in premenopausal women with idiopathic osteoporosis and concurrently recruited normal women. Overall, there were significant microarchitectural deficits in affected women: thinner cortices; fewer, thinner, more widely separated, and heterogeneously distributed trabeculae; and lower mechanical competence (stiffness) estimated by μFE of μCT data sets. Contrary to our original hypotheses, bone remodeling, whether assessed by serum BTM or dynamic histomorphometry, was heterogeneous and, on average, did not differ between affected and normal women. Additionally, there were no differences in microarchitecture, stiffness, or remodeling according to whether subjects had fractures. Overall, affected women had lower cancellous osteoid width and endocortical and intracortical wall width than controls, suggesting that osteoblasts were producing less bone matrix per completed osteon. This finding was more pronounced in women in the lowest tertile of BFR/BS, who also had worse microarchitecture and stiffness than those in the highest tertile. They also had significantly higher serum IGF-I than those in the upper two tertiles of BFR/BS, suggesting primary osteoblast dysfunction with resistance to IGF-I. Those in the highest tertile of BFR/BS had higher serum 1,25(OH)₂D and trends toward higher urinary calcium and serum PTH, a pattern consistent with idiopathic hypercalciuria. These results suggest that the pathogenesis of idiopathic osteoporosis is heterogeneous and may differ according to remodeling activity.

The microarchitectural abnormalities are consistent with several other 2D histomorphometry studies of young adults with IOP (3, 6, 7, 9, 14). Khosla et al. (14) reported lower cancellous bone volume and cortical width in young men and women with unexplained fractures. Lower cancellous bone volume and, in some instances, cortical width were seen in men with IOP (3, 6, 7, 9). By μCT, a 3D technique that provides direct volumetric measurements of trabecular structure of the entire biopsy, we detected significant differences in trabecular number and separation that were not apparent by 2D histomorphometry. To estimate stiffness, we used μFE, a computational technique that integrates the density and microarchitectural information in the biopsy and demonstrates excellent agreement with true biomechanical tests of bone specimens (31). The μCT and μFE technologies indicated that decreased trabecular number rather than trabecular thinning underlies reduced bone stiffness in premenopausal women with IOP. These 2D histomorphometry, μCT, and μFE data are consistent with our recent report in a subset of these same women in whom microarchitecture was assessed by HRpQCT of the radius and tibia (21). Despite inherent heterogeneity of different skeletal sites (26), our results suggest that microstructural defects in women with IOP at the iliac crest are generalizable to multiple skeletal sites.

In a previous retrospective study of women with IOP (18), we found significantly lower dynamic parameters of bone formation (MdPm, mineral apposition rate, BFR/BS) compared with age-matched historical controls, results consistent with several studies of men with IOP (6, 11, 15, 32). Although we found no differences in dynamic parameters between subjects and controls in this study, on average, affected women had lower osteoid width and wall width, suggesting that osteoblasts are synthesizing less bone matrix per completed osteon. Moreover, wall width correlated strongly with trabecular microstructure and stiffness, suggesting that osteoblast dysfunction contributed to the microarchitectural disruption. Further analyses revealed that low osteoid width was most pronounced in women in the lowest tertile of BFR/BS. Moreover, the LT group had substantially less favorable microarchitecture (lower BV/TV, Tb.Th, osteoid width, and wall width) and lower stiffness than the HT group. This is unusual because in older women, high rather than low turnover is associated with more bone loss and microarchitectural deterioration (33, 34). One possible explanation is that microarchitectural disruption in the LT group may have resulted from a prior insult now resolved, whereas bone loss may be ongoing in the HT group. It is conceivable that prolonged low bone formation rate, particularly if present during growth, could result in impaired bone structure and reduced stiffness. Longitudinal observation is necessary to determine which of these possibilities is true.

In men with IOP, low serum IGF-I levels that are directly associated with aBMD and histomorphometric parameters of bone formation have been reported (6, 9). In contrast, we found no difference in serum IGF-I levels between subjects and controls (17). However, serum IGF-I was inversely associated with bone microstructure and stiffness in affected women, correlated inversely with cancellous osteoid and wall width in the ILBMD group, and was significantly higher in the lower than the upper two tertiles of BFR/BS. These results suggest that the pathogenesis of low turnover IOP differs in men and women; in men, osteoblast dysfunction may be the result of low IGF-I, whereas in women, osteoblasts may be resistant to IGF-I.

In the HT group, serum 1,25(OH)₂D was significantly higher, and there was a trend toward higher urinary calcium and PTH, with no differences in serum 25-OHD or calcium. This biochemical pattern suggests idiopathic hypercalciuria. Although we excluded women with frank hypercalciuria (>300 mg/g Cr), it is possible that milder degrees of idiopathic hypercalciuria may cause high turnover osteoporosis in young women.

We have argued (35), as has the International Society for Clinical Densitometry (36) and others (37–42), that young, otherwise healthy women should not be diagnosed with osteoporosis on the basis of low aBMD by DXA unless there is a history of fragility fracture or a secondary cause of osteoporosis, such as glucocorticoid exposure. However, both our bone biopsy and HRpQCT data (21) revealed that otherwise healthy women with low aBMD had cortical and trabecular microarchitectural disruption and reduced stiffness comparable to subjects with fractures. Although these data suggest that low aBMD in premenopausal women with IOP represents a presymptomatic phase of the disease and could portend early onset of fractures, they may not be generalizable for several reasons. Although the ILBMD group did not have adult low trauma fractures, 26% reported childhood fractures, 16% reported high-trauma fractures, and 84% had a family history of osteoporosis (17). Because such women might be more likely to volunteer for the study and to have abnormal microarchitecture, our results could represent ascertainment bias. A single low aBMD measurement provides no insight into whether it is due to low peak bone mass or to bone loss that has occurred previously or is ongoing. The ILBMD group weighed significantly less than controls, and small, thin women tend to have lower aBMD (19, 37, 43–45); whether small, thin women also have abnormal bone quality is unknown. Although overall, there were definite microarchitectural defects in the women with ILBMD, they may be heterogeneous in this regard. A larger population-based longitudinal study will be required to determine whether premenopausal women with low aBMD have abnormal bone quality and to quantify their risk for bone loss and/or fractures.

This study has several limitations. Our younger subjects may not have reached peak bone mass. We may have misclassified fractures as low trauma when in fact they did not reflect abnormal fragility. Women may have volunteered for this study because of a family history of osteoporosis, causing ascertainment bias. We cannot determine whether microarchitectural abnormalities resulted from ongoing bone loss or past insults, now resolved. Similarly, dynamic remodeling parameters may not have differed because they represent current conditions rather than past pathophysiological processes. Because we performed μFE analyses on the trabecular compartment, they reflect trabecular rather than whole-bone stiffness. Because cortical thickness was lower in subjects, we may have underestimated differences in stiffness. Finally, assumptions of uniform bone mineralization are incorporated into μFE analyses, which may or may not be appropriate for this cohort.

Our study also has important strengths. It is the largest case-control study of women with IOP to include transiliac crest bone biopsies for the assessment of bone microarchitecture, stiffness and remodeling. Both cases and controls were carefully characterized and common causes of secondary osteoporosis excluded by detailed clinical and biochemical evaluation.

In conclusion, we found significant microarchitectural differences in cortical and trabecular bone and estimated stiffness on transiliac bone biopsies between a group of premenopausal women with unexplained fractures and/or low aBMD and a group of healthy women with normal aBMD. Women with low aBMD who had not fractured were as severely affected as those included on the basis of one or more low-trauma fractures. Some affected women had evidence of osteoblast dysfunction with synthesis of less bone matrix per remodeling site, more profound microstructural defects, and possible resistance to IGF-I. Another group had high bone turnover and a biochemical pattern resembling idiopathic hypercalciuria. These noteworthy findings should serve as the basis for further in vivo and in vitro studies to investigate potential etiologies of IOP in premenopausal women.