Teriparatide Reduces Bone Microdamage Accumulation in Postmenopausal Women Previously Treated With Alendronate


  • Harald Dobnig,

    Corresponding author
    1. Department of Internal Medicine, Division of Endocrinology and Nuclear Medicine, Medical University of Graz, Graz, Austria
    • Department of Internal Medicine, Division of Endocrinology, Medical University of Graz, Auenbruggerplatz 15, A-8036 Graz, Austria
    Search for more papers by this author
  • Jan J Stepan,

    1. Institute of Rheumatology, Charles University, Faculty of Medicine, Prague, Czech Republic
    Search for more papers by this author
  • David B Burr,

    1. Department of Anatomy and Cell Biology and Department of Orthopaedic Surgery, Indiana University School of Medicine, Indianapolis, Indiana, USA
    Search for more papers by this author
    • Dr. Burr is a member of the Scientific Advisory Board for Eli Lilly and Co. He also serves as a consultant to The Alliance for Better Bone Health. He has research support from Eli Lilly and Co., The Alliance for Better Bone Health, and Amgen. Drs. Sipos and Pavo are shareholders and fulltime employees of Eli Lilly and Co. Dr. Petto is a full-time employee of Eli Lilly and Co. All other authors state that they have no conflicts of interests.

  • Jiliang Li,

    1. Department of Biology, Indiana University-Purdue University, Indianapolis, Indiana, USA
    Search for more papers by this author
  • Dana Michalská,

    1. Department of Internal Medicine 3, Charles University, Faculty of Medicine, Prague, Czech Republic
    Search for more papers by this author
  • Adrien Sipos,

    1. Lilly Research Laboratories, Indianapolis, Indiana, USA
    Search for more papers by this author
  • Helmut Petto,

    1. Lilly Research Laboratories, Indianapolis, Indiana, USA
    Search for more papers by this author
  • Astrid Fahrleitner-Pammer,

    1. Department of Internal Medicine, Division of Endocrinology and Nuclear Medicine, Medical University of Graz, Graz, Austria
    Search for more papers by this author
  • Imre Pavo

    1. Lilly Research Laboratories, Indianapolis, Indiana, USA
    Search for more papers by this author

  • These data were previously presented in part in the form of an abstract and an oral presentation at the 29th Annual Meeting of the American Society of Bone and Mineral Research, Honolulu, HI, USA, September 16–19, 2007.

  • Published online on May 18, 2009;


Suppression of bone turnover by bisphosphonates is associated with increased bone microdamage accumulation in animal models. Our objective was to study the effects of teriparatide treatment on changes in microdamage accumulation at the iliac crest in previously treatment-naïve patients or in those switched from alendronate to teriparatide. Sixty-six postmenopausal women with osteoporosis (mean age, 68.0 yr; and mean BMD T-score of −2.8 at lumbar spine and −1.7 at total hip; 62% with prevalent fractures) entered this prospective, nonrandomized study and started with 24-mo 20 μg/d subcutaneous teriparatide treatment in monotherapy: 38 patients stopped previous alendronate treatment (10 mg/d or 70 mg/wk for a mean duration of 63.6 mo) and switched to teriparatide, whereas 28 were previously treatment naïve. Thirty-one paired biopsies with two intact cortices were collected and analyzed for microstructure and microdamage accumulation at baseline and after 24 mo of teriparatide administration. After 24 mo of teriparatide treatment, crack density (Cr.Dn), crack surface density (Cr.S.Dn), and crack length (Cr.Le) were decreased in previously alendronate-treated patients, whereas only Cr.Le was reduced in former treatment-naïve patients. Patients with lower initial femoral neck BMD also showed a higher reduction of microdamage accumulation. Better bone microarchitecture correlated positively, whereas bone turnover markers and age did not correlate with reduced microdamage accumulation on teriparatide. In conclusion, teriparatide reduces microdamage accumulation in the iliac crest of patients previously treated with alendronate. There is insufficient evidence to suggest that age or bone turnover would be associated with this change.


Mechanical loading of the bone, aging, and increased bone fragility are associated with accumulation of skeletal microdamage.(1–4) Microdamage formation occurs in response to mechanical loads and is removed by locally induced remodeling. The level of microdamage accumulation is a consequence of the balance between microdamage formation and its repair.(1–7) Therefore, conditions that either increase microdamage formation or decrease its repair can result in elevated microdamage accumulation. In contrast, decreased microdamage formation and/or increased removal of microdamage would result in less existing microdamage.

Bisphosphonates, like alendronate, are efficacious for reducing fractures because they suppress bone remodeling and subsequently increase bone strength.(8–10) Based on animal studies, one consequence of reduced bone remodeling by bisphosphonates is the increase of microdamage accumulation.(11–15) This occurs with alendronate and risedronate in a dose-dependent manner starting with doses comparable to those used for the treatment of postmenopausal osteoporosis.(15) The implications of increased microdamage on fracture risk with bisphosphonates are uncertain as animal studies also showed increases in vertebral bone strength and stiffness, even though toughness was reduced.(11,15) In humans, few data exist on bisphosphonate treatment and microdamage accumulation.(16,17) Overall, bisphosphonate treatment does not seem to significantly increase microdamage accumulation.(16,17) On the other hand, alendronate-treated patients with low BMD had a higher accumulation of microdamage in the iliac crest than those with higher BMD.(16) This difference was not present in untreated patients. Moreover, after adjustment for age, prevalent fractures, femoral neck BMD, activation frequency, and investigational center, microdamage density was significantly increased in transiliac crest biopsies of women who had been treated with alendronate for an average of 5 yr.(16)

Teriparatide is a bone-forming agent used for the treatment of osteoporosis.(18) Teriparatide significantly increases bone turnover, BMD, and bone volume and improves both trabecular and cortical microarchitecture through effects on bone remodeling that are essentially opposite from those of bisphosphonates.(19–21) In contrast to the latter, the more frequent initiation of bone remodeling units by teriparatide would be expected to result in increased removal of microdamage. It is unknown how the lower mean tissue age under teriparatide treatment could influence microdamage accumulation. At present, neither animal nor clinical data exist on microdamage accumulation after teriparatide treatment.

Our goal was to study the effect of increased bone turnover and improved bone structure on microdamage accumulation in patients with osteoporosis who started with teriparatide as first treatment and in those who switched from long-term alendronate to teriparatide treatment.


In two clinical centers (Center I: Department of Internal Medicine 3, Charles University, Faculty of Medicine, Prague, Czech Republic; Center II: Department of Internal Medicine, Medical University of Graz, Graz, Austria), 66 white postmenopausal women with osteoporosis were enrolled (Fig. 1). Thirty-eight had been treated previously with alendronate sodium (10 mg/d or 70 mg/wk), calcium (1000 mg daily), and vitamin D (800 IU daily) for a mean duration of 63.6 mo. Twenty-eight patients were treatment naïve. In alendronate-treated patients, date of therapy initiation, compliance, and persistence on drug were assessed by interviewing patients. Date of therapy initiation and persistence were also confirmed by review of medical records. Information about filling the prescriptions at least 70% in the last 12 mo before study start was verified specifically. Ambulatory postmenopausal women who at some point were diagnosed with osteoporosis at either total hip or lumbar spine (T-score ≤ 2.5) and who were at least 55 yr of age were included. Patients were excluded if they had a history of any secondary causes of osteoporosis, history of malignant neoplasm in the prior 5 yr, history of nephrolithiasis or urolithiasis in the prior 2 yr, abnormal thyroid function, active liver disease, significantly impaired renal function, treatment with androgens or other anabolic steroids, treatment with vitamin D > 50,000 IU/wk or active vitamin D analogs, and treatment with calcitonin, fluoride, estrogens, progestins, estrogen analogs, estrogen agonists, estrogen antagonists, selective estrogen receptor modulators, tibolone, systemic corticosteroids, and bisphosphonates other than alendronate in the prior 36 mo. History of excessive consumption of alcohol or abuse of drugs was also an exclusion criterion. The actual serum concentrations of the markers were not inclusion or exclusion criteria.

Figure Figure 1.

Study design.

Patients signed informed consent to the treatment and investigation protocol, which was approved by the Institutional Review Board for Research Involving Human Subjects, at each participating center.

At baseline, we obtained information on demographics, health history, and medication use. History of clinical fractures and spine X-rays were obtained in all patients. All enrolled patients were supplemented with a daily intake of 1000 mg elemental calcium and 400–800 IU vitamin D3. All patients self-administered a once-daily subcutaneous injection of 20 μg teriparatide for 24 mo.

Biochemical measurements

Biochemical markers of bone formation (the serum concentration of intact amino terminal propeptide of type I procollagen [PINP]) and resorption (serum concentration of type 1 collagen cross-linked C-telopeptide [CTX]) were assessed at baseline and at 1, 3, 6, 12, and 24 mo or at early discontinuation. Blood specimens were collected in the morning after an overnight fast. The serum concentration of PINP was assessed by radioimmunoassay (Procollagen Intact PINP; Orion Diagnostica). The assay is not sensitive to the small-molecular-weight degradation products of the propeptide. The within-run imprecision was <5%, and between-run imprecision was <7% at concentrations between 20 and 90 μg/liter. The serum concentration of CTX was assessed using electrochemiluminescence-based immunoanalysis (the Elecsys 1010 Analyzer; Roche Diagnostics). The within-run imprecision of the CTX was <5% for samples >500 ng/liter, <7% for samples between 200 and 500 ng/liter, and <10% for very low CTX concentrations samples. The between-run imprecision results were <7% for samples >500 ng/liter and <9% for samples between 200 and 500 ng/liter. The detection limit was <10 ng/liter. Bone-specific alkaline phosphatase (bone ALP) was determined using an immunoradiometric assay (Tandem-R Ostase; Hybritech, San Diego, CA, USA). The intra- and interassay coefficients of variation were 4.2–6.8% and 7.4–7.9%, respectively.

BMD measurements

BMD at lumbar spine, total femur, femoral neck, distal forearm, and total body was determined using the QDR 4500A bone densitometer (Hologic, Waltham, MA, USA) at baseline and after 6, 12, and 24 mo of teriparatide treatment. Normative values provided by Hologic (NHANES III normative values for the proximal femur) were used for the determination of T-scores (comparison with a sex-matched young normal reference population). The short-term in vivo precision error for the lumbar spine (L1−L4), total femur, femoral neck, distal forearm, and total body was 0.7%, 0.9%, 1.9%, 0.9%, and 1.5%, respectively. Daily scanning of a phantom showed absence of a machine drift during the study. No cross-calibration between the two centers was performed.


Double-tetracycline labeling was initiated 22–24 days before biopsy surgery. Transiliac crest bone biopsies were obtained by manual drill using a 7.5-mm trephine (Medical Innovations International, Rochester, MN, USA). Biopsies were obtained from patients at baseline, before receiving teriparatide treatment, and at study end (24 mo). Biopsy specimens were stored and transported in 70% ethanol to the central bone histomorphometry laboratory for assessment and analysis of biopsies and interpretation of results (D.B. Burr, Indianapolis, IN, USA). The bone specimens were prepared and analyzed in the same laboratory (D.B.B.) with the same plastic embedment processes and materials. Any biopsy that consisted of two well-defined cortical tables and that did not have a single complete fracture extending between the cortices was considered complete and suitable for analysis.

On arrival, biopsies were stained en bloc under vacuum (15–20 mmHg) in 1% basic fuchsin and increasing concentrations of ETOH (80%, 95%, and 100%; one change each after 2.5 h). After staining, bones were rinsed in 100% ETOH followed by 100% methylmethacrylate for 4 h under vacuum and embedded in methylmethacrylate with 3% dibutyl phthalate. For analysis of microdamage, three sections of 80 μm thickness were cut from each biopsy using a diamond wire saw (Histosaw; DDK). All fields at least 0.5 mm from any cortical wall or the ends of the biopsy were assessed; the same fields were used for microdamage assessment as for subsequent histomorphometry. Stained microcracks were identified by their shape, depth of field, and permeation of stain into the crack walls. This technique can discriminate artifactual damage caused by histological preparation processes from those that occurred in vivo but cannot distinguish in vivo cracks from those generated by the process of biopsy removal. The number of linear microcracks was counted and normalized to bone area (Cr.Dn = Cr.N/bone area, #/mm2). Crack lengths (Cr.Le, μm) were measured and mean Cr.Le was used to calculate crack surface density (Cr.S.Dn, μm/ mm2), which is the product of Cr.Dn and mean Cr.Le. Validation of the techniques and the rationale for measuring microdamage from three sections was presented in an earlier paper.(16)

Four sections of 5-μm thickness were cut from each block using a Reichert-Jung 2050 microtome (Magee Scientific) and stained with McNeal's tetrachrome for static 2D histomorphometry. The protocol included predefined static histomorphometric parameters, bone volume fraction (BV/TV), trabecular thickness (Tb.Th, μm), trabecular number (Tb.N, mm−1), and trabecular separation (Tb.Sp, μm). Dynamic parameters included activation frequency (Ac.f, cycles/yr), mineralizing surface (MS/BS, %), bone formation rate (BFR/BV, %/yr), and mineral apposition rate (MAR, μm/d). All measurements were carried out by one histomorphometrist blinded to group affiliation using a semiautomatic digitizing system (Bioquant OSTEO 7.20.10; Bioquant Image Analysis) attached to a Nikon Optiphot 2 microscope. All measured variables were expressed in accordance with histomorphometric criteria recommended by the subcommittee on bone histomorphometry of the American Society for Bone and Mineral Research.(22)

Statistical analyses

The study was designed to enroll 60 patients, with ∼30 patients from each pretreatment group, to account for those biopsies that were not suitable for analysis. Calculation of the sample size to detect differences in active mineralizing surface between the two groups was based on results by Chavassieux et al.,(23) indicating that the mean active mineralizing surface in treatment-naïve patients is 0.06 with an SD of 0.035 and that 2–3 yr of alendronate therapy reduces the active mineralization surface by 96–98%. Assuming a common SD of 0.035, at least 90% power existed to detect a difference of 0.06 in active mineralizing surface at baseline between the long-term alendronate-treated and treatment-naïve patients with 20 evaluable biopsies in each treatment group.

All patients enrolled were included in the analysis data set. For each analysis of baseline, endpoint and change from baseline all available and valid values of the analysis data set were included. No imputation of missing values was performed.

Patients' characteristics were summarized at baseline for the two study groups (treatment-naïve and alendronate-treated groups). For categorical data, percentages were calculated, and for continuous variables, arithmetic means and SDs were calculated. For approximately normally distributed continuous variables, the difference between the study groups was tested by two-tailed t-tests. Log-transformation was used for non-normal data and the Wilcoxon test if the normalization failed. For categorical data, the χ2 test was used. Microdamage indices were summarized at baseline and endpoint. The differences between treatment groups were tested by two-tailed Wilcoxon tests and the change from baseline by signed rank tests.

For testing the relationships between microdamage indices and age, Spearman rank correlation coefficients were calculated at baseline and endpoint by study group and for all patients. Partial correlation coefficients were used to age-adjust the Spearman rank correlations between crack surface density and microarchitectural indices. The adjusted correlations were calculated at baseline and at endpoint by study group and for all patients and for the changes from baseline only for all patients.

Similar Spearman rank-correlations were calculated between microdamage indices and biochemical markers and in-between microdamage indices and BMD variables. Also, this analysis was performed at baseline, at endpoint, and for the changes from baseline.

For crack density and separately for crack surface density, linear regression analysis with stepwise variable selection was performed (selection criteria p < 0.05), using the percentage changes from baseline as dependent variables and pretreatment, investigational sites, prevalent fracture status, baseline femoral neck BMD, and baseline activation frequency as independent variables.

For all test results, the statistical level of significance was set to 5%. This explorative analysis was performed posthoc to support the results of the primary and secondary study objectives. No multiplicity adjustments were performed.


Sixty-six patients were eligible for histomorphometric analysis at baseline (treatment naïve, n = 28; alendronate pretreated, n = 38). Forty-five patients had valid histomorphometric measurements both at baseline and endpoint (treatment naïve, n = 16; alendronate pretreated, n = 29). Finally, biopsies of 31 patients could be evaluated for microdamage accumulation parameters, both at baseline and endpoint (treatment naïve, n = 13; alendronate pretreated, n = 18; Table 1; Fig. 2)

Table Table 1.. Patient Characteristics With Paired Biopsy Values at Baseline (N = 31)
original image
Figure Figure 2.

Patient flow.

At baseline, compared with treatment-naïve patients, the lumbar spine BMD results were higher and the biochemical markers of bone turnover and some dynamic histomorphometric indices (BFR/BV, MS/BS) were significantly lower in the alendronate pretreated group. There was insufficient evidence for any further differences between parameters, including structural histomorphometric indices of the two groups.

At baseline, no difference in microdamage accumulation between the two treatment groups could be observed. In patients pretreated with alendronate, there was a significant correlation between age and microdamage accumulation (Cr.Dn and Cr.S.Dn) that was not present in treatment-naïve patients (Table 2). After adjusting for age, trabecular area inversely correlated with elevated microdamage accumulation represented by Cr.S.Dn (Table 3).

Table Table 2.. Correlation Between Age and Microdamage Accumulation at Baseline and After 24 mo of Teriparatide Treatment
original image
Table Table 3.. Age-Adjusted Correlations Between Crack Surface Density and Various 2D and 3D Microarchitectural Indices at Baseline and After 24 mo of Teriparatide Treatment
original image

After 24 mo of teriparatide treatment, Cr.Dn, Cr.S.Dn, and Cr.Le were significantly decreased in previously alendronate-treated patients, whereas only Cr.Le was significantly reduced in former treatment-naïve patients (Fig. 3).

Figure Figure 3.

Changes in microdamage accumulation in patients treated with teriparatide *Change from baseline within group. **Change from baseline between groups. All data presented as means with SE.

Among baseline characteristics (pretreatment, investigational site, prevalent fracture status, femoral neck BMD, and activation frequency), pretreatment and femoral neck BMD interacted significantly with change in microdamage accumulation by teriparatide. Those patients with lower initial femoral neck BMD (p = 0.0003) and on previous alendronate therapy (p = 0.042) showed a significantly higher reduction of microdamage accumulation (Cr.S.Dn).

At endpoint, after 24 mo of teriparatide treatment, no interaction between age and microdamage accumulation could be shown. The correlations between age and microdamage accumulation present after alendronate treatment were not observed after subsequent teriparatide treatment (Table 2).

After 24 mo of treatment, when adjusted for age, numerous microstructural indices in both groups correlated with microdamage accumulation (Table 3; Fig. 4).

Figure Figure 4.

Stained microcracks in paired iliac crest biopsies from a patient previously treated with alendronate.

We could not show any significant correlation between microdamage parameters and biochemical or histomorphometric parameters indicative of bone turnover at baseline or endpoint. Likewise, the baseline-endpoint changes of microdamage parameters did not correlate with the respective changes in dynamic histomorphometry parameters. Moreover, there was no correlation between endpoint total hip or femoral neck BMD and microdamage accumulation.


We found that 24 mo of teriparatide treatment reduced microdamage accumulation in osteoporotic patients who had been treated before this with alendronate. There was no change in microdamage accumulation in those patients who were not treated with any osteoporosis medication before starting with teriparatide administration.

There are several factors that could account for this. The differential effect of teriparatide on microdamage accumulation in the two patient populations may be caused by the baseline amount of damage accumulation. In our previous analysis of this study population at baseline, crack density was higher in alendronate-treated patients compared with treatment-naïve patients after adjustment for potential confounders (clinical site, age, history of prevalent fractures, femoral neck, BMD, and activation frequency). In that study, patients who had been treated for an average of 5.3 yr with alendronate had 30% more microdamage than those who had not been treated for osteoporosis, even though treatment-naïve patients had significantly lower BMD at the lumbar spine.(16) This higher baseline accumulation of damage in the alendronate pretreatment group may have allowed a greater window to show the reduction in microdamage accumulation when these patients were placed on teriparatide.

A recent study has also investigated microdamage accumulation in 50 patients on different bisphosphonate treatment regimens and has documented a 2.6 times higher mean crack density compared with a smaller pool of control samples (cadaveric specimens from 12 significantly elderly individuals).(17) This increase, although not statistically significant, is numerically comparable to the 2.5- to 3.5-fold statistically significant increase of crack density in canine vertebrae associated with the use of clinically relevant doses of alendronate or risedronate.(15) In agreement with the conclusion of the human study by Chapurlat et al.,(17) we suggest that, because of the large individual variation of microdamage accumulation in the human iliac crest, the statistical power of both studies is not sufficient to finally conclude whether long-term alendronate use is associated with increased microdamage accumulation.

Bone remodeling is considered to be targeted toward the repair of fatigue microdamage.(24) Consequently, an excessive reduction of bone turnover may result in inadequate microdamage repair. While reducing global bone turnover, alendronate as a potent anticatabolic agent also suppresses targeted remodeling.(25) Indeed, in preclinical studies, where long-term alendronate use was associated with increased microdamage accumulation, the magnitude of bone turnover reduction correlated in a nonlinear fashion with this accumulation.(11–15,22)

In contrast, however, we could not find any significant relationship between markers of bone turnover, determined either by biochemical or histomorphometric parameters, and microdamage accumulation in this study. Our findings suggest that the degree of bone turnover suppression achieved showed no evidence of detrimental effect on microcrack repair and consequent microdamage accumulation in treatment-naïve or alendronate-treated osteoporotic patients. It may be that, in most patients, the suppressed bone turnover during bisphosphonate therapy is sufficient to maintain an adequate level of microdamage repair. This is also supported by the observations in monkey vertebrae where microdamage accumulation did not occur when bone turnover was suppressed by 50% or less with estrogen or raloxifene.(26)

Recent preclinical data also showed that longer alendronate treatment does not cause additional microdamage accumulation above that produced by 1 yr of treatment.(27) It was hypothesized that either microdamage can be controlled at a new equilibrium level when turnover is reduced by 70%, and/or that there is a reduced formation of microdamage development caused by lower trabecular strains associated with the increased bone volume in alendronate-treated animals.

The absence of a relationship between turnover and the reduction of microdamage accumulation after teriparatide administration in the pretreated group, however, could also be indicative of the timing between the increased bone remodeling that must have occurred after the initiation of teriparatide therapy and the measurement of bone microdamage. Based on biomarkers, it is well known that the primary anabolic effect of teriparatide occurs between 1 and 18 mo after initiation of treatment.(28–35) Because the biopsies were not examined until 24 mo after the initiation of teriparatide therapy, the rate of bone remodeling had likely reduced, and thus no effect of turnover could be shown in our study conditions. Although no correlation was found between biomarkers of turnover in the first 12 mo and microdamage accumulation at 24 mo, the variability of both biomarker and microdamage measurements, and their measurement at different time points would make any strong correlation between them unlikely.

The significant reduction in mean crack length, in both pretreated and treatment-naïve groups, raises the possibility that an increase in remodeling may affect the change in damage parameters. However, recent data have also shown that bone matrix from patients treated with teriparatide exhibits signs of collagen properties suggesting a reduction of mean tissue age,(36) whereas the presence of aging collagen and the increased glycation of older bone is associated with decreased fracture load(37) and decreased bone toughness.(38) Physicochemical properties of younger bone matrix on teriparatide could be responsible for decreased crack length in both patient groups.

One result of the teriparatide anabolic effect was a change in bone microarchitecture, indicated among others by a 73% increase in BV/TV after 24 mo of teriparatide treatment.(39) There was a significant inverse correlation of BV/TV and Tb.N with the overall microdamage burden (assessed by Cr.S.Dn), indicating that an improvement in the integrity of the cancellous bone network is associated with decreased microdamage accumulation. These effects of changing architecture on microdamage were observed in both treatment groups. Our results are in agreement with recent findings in cancellous vertebral bone obtained from 23 randomly selected donors 54–93 yr of age(39) and with an earlier cross-sectional study in which microcrack density inversely correlated with microarchitectural parameters.(3) The fact that this improved microarchitecture is associated with reduced microdamage accumulation after teriparatide treatment can not answer the important question whether reduced microdamage accumulation influences the mechanical properties of bone.

At baseline, we found a correlation between age and microdamage accumulation in alendronate-treated patients, but when adjusted to femoral neck BMD, the correlation between age and microdamage accumulation was not statistically significant. Among selected confounders including age, femoral neck BMD was the only measure shown to have a significant interaction with microdamage accumulation in alendronate-treated patients.(16)

No correlation was present in osteoporotic patients free from medication at baseline, nor did we find any correlations between age and microdamage parameters in any group after 24 mo of treatment with teriparatide. Aging generally is associated with increased microdamage accumulation in cortical bone.(2,5,40) There is less clarity about whether microdamage accumulates with age in cancellous bone. In an earlier study, no age-dependent accumulation in cancellous bone of the vertebrae was found.(3) In contrast, others have shown that microdamage increases with age in the human proximal femur and vertebral cancellous bone.(6,39) Our study population is, however, markedly different from the other study populations. Other studies collected samples either from routine serial autopsies of men and women,(3,6,39) of different races,(3) and in the absence of complete medical history or from patients with osteoarthrosis.(6) In contrast, we studied a well-characterized, clinically relevant patient population with postmenopausal osteoporosis.

One important limitation of the study is the lack of an untreated control group after alendronate treatment. The response to teriparatide in our study could theoretically also be caused by the cessation of alendronate treatment alone rather than being a specific teriparatide effect. We believe that the likelihood of the cessation of alendronate treatment being responsible for the reduction of microdamage accumulation is small. It is well known that stopping alendronate only gradually and slowly increases bone turnover that often remains markedly suppressed in the first 1–2 yr after cessation of treatment.(9) Switching from alendronate to teriparatide, on the other hand, results in almost immediate increases in bone turnover, translating into bone anabolic effects. In addition to these observations, this was shown previously.(31,41) Thus, changes that are associated with a reduction in microdamage accumulation after cessation of alendronate treatment are most likely driven by the marked effect that teriparatide has on bone turnover and/or structural improvements.

The hypothesis of this study was that reduced bone turnover by alendronate slows down microdamage repair and may lead to microdamage accumulation. Administration of teriparatide after stopping alendronate could reduce microdamage accumulation by the substantial increase seen in bone turnover. Indeed, teriparatide reduced microdamage accumulation after alendronate. We failed, however, to show a meaningful correlation between microdamage reduction and the observed increase in bone turnover. On the other hand, our data suggest that the improvement in bone structure by teriparatide itself could be associated with the reduction in microdamage. At least at the human iliac crest, improved microstructure could be more relevant in the process of microdamage accumulation than potential faster repair. This finding is supported by numerous observations, because deterioration in microarchitecture and/or the process of aging,(42) two fundamental and closely interrelated components of increased fragility, are associated with increased microdamage accumulation in humans. It is possible that under clinical conditions, microdamage accumulation is influenced to a greater degree by bone microarchitecture than by the process of bone turnover.

We conclude that teriparatide reduces microdamage accumulation in the iliac crest of patients previously treated with alendronate and reduces crack length regardless of prior treatment. In addition, we conclude that an intact microarchitecture is essential for maintaining microdamage accumulation at physiologically normal levels in osteoporotic patients.


Funding was provided by Eli Lilly and Company.