Skeletal Histomorphometry in Subjects on Teriparatide or Zoledronic Acid Therapy (SHOTZ) Study: A Randomized Controlled Trial

Abstract

Context:

Recent studies on the mechanism of action (MOA) of bone-active drugs have rekindled interest in how to present and interpret dynamic histomorphometric parameters of bone remodeling.

Objective:

We compared the effects of an established anabolic agent, teriparatide (TPTD), with those of a prototypical antiresorptive agent, zoledronic acid (ZOL).

Design:

This was a 12-month, randomized, double-blind, active-comparator controlled, cross-sectional biopsy study.

Setting:

The study was conducted at 12 U.S. and Canadian centers.

Subjects:

Healthy postmenopausal women with osteoporosis participated in the study.

Interventions:

Subjects received TPTD 20 μg once daily by sc injection (n = 34) or ZOL 5 mg by iv infusion at baseline (n = 35).

Main Outcome Measures:

The primary end point was mineralizing surface/bone surface (MS/BS), a dynamic measure of bone formation, at month 6. A standard panel of dynamic and static histomorphometric indices was also assessed. When specimens with missing labels were encountered, several methods were used to calculate mineral apposition rate (MAR). Serum markers of bone turnover were also measured.

Results:

Among 58 subjects with evaluable biopsies (TPTD = 28; ZOL = 30), MS/BS was significantly higher in the TPTD group (median: 5.60 vs. 0.16%, P < 0.001). Other bone formation indices, including MAR, were also higher in the TPTD group (P < 0.05). TPTD significantly increased procollagen type 1 N-terminal propeptide (PINP) at months 1, 3, 6, and 12 and carboxyterminal cross-linking telopeptide of collagen type 1 (CTX) from months 3 to 12. ZOL significantly decreased PINP and CTX below baseline at all time points.

Conclusions:

TPTD and ZOL possess fundamentally different mechanisms of action with opposite effects on bone formation based on this analysis of both histomorphometric data and serum markers of bone formation and resorption. An important mechanistic difference was a substantially higher MS/BS in the TPTD group. Overall, these results define the dynamic histomorphometric characteristics of anabolic activity relative to antiresorptive activity after treatment with these two drugs.

Bone histomorphometry has allowed researchers to study bone remodeling in humans at the tissue level (1). Using a standard panel of tetracycline-based parameters of bone formation, researchers are able to evaluate biopsy samples to determine whether a bone-active drug is antiresorptive (anticatabolic), or anabolic. Both classes of drug are used to treat women with postmenopausal osteoporosis.

Antiresorptive agents increase bone mineral density by inhibiting osteoclast-mediated bone resorption, which slows bone turnover, closes the remodeling space, and allows for more complete secondary mineralization (2). Anabolic agents increase bone mass by increasing bone turnover, preferentially stimulating formation over resorption, thereby creating a positive bone balance within each bone remodeling unit (3).

Measurement and calculation of several dynamic parameters of bone formation, such as mineralizing surface/bone surface (MS/BS), mineral apposition rate (MAR), and bone formation rate (BFR), play a key role in defining the mechanism of action (MOA) of bone-active drugs. However, there has been considerable variation in the inclusion and exclusion of samples for the calculation of group means, especially for MAR (4). In particular, there has been a lack of uniformity in the presentation of data from studies in which a substantial proportion of the biopsies lack labels or display single labels only. This occurs frequently in the setting of low turnover after treatment with potent antiresorptive agents, such as bisphosphonates (5, 6) or denosumab (7).

Two methodological studies attempted to address the problem of how to interpret MS/BS and MAR in biopsies lacking double labels or when only single labels are present. Foldes et al. (8) proposed that, in the absence of labels, MAR be treated as a missing value and MS/BS be assigned a value of zero and be included in the calculation of group mean values. For biopsies with single labels only, they proposed that MAR be assigned a value of 0.3 μm/d, the lowest value measured in their study. In 1999, Hauge et al. (9) cautioned that censoring missing labels may bias the results by the preferential inclusion of subjects with higher turnover. Instead, these authors recommended that, in the absence of labels, both MAR and MS/BS be assigned a value of zero but that these subjects be reported separately and excluded from the calculation of group means. When only single labels are present, Hauge et al. recommended that MAR be assigned a lower limit of 0.1 μm/d.

The recommendations from these studies have not been widely adopted. For example, in studies of two potent antiresorptive agents, risedronate (5) and zoledronic acid (ZOL) (6), biopsies with missing labels, single labels only, or too few double labels to accurately measure MAR were excluded from the calculation of mean MAR values. Many more biopsies (36–48%) were excluded from the bisphosphonate- than from the placebo-treated groups (8–10%). In the ZOL study (6), MAR was significantly higher in the ZOL group than in the placebo group, leading the authors to speculate that ZOL may possess anabolic activity, although the authors correctly raised the caveat that this result could be due to the type of selection bias noted above. The interpretation of this finding has been a subject of debate (10, 11).

The purpose of the present study was to compare the histomorphometric profile of an established anabolic agent, teriparatide (TPTD), with that of a prototypical antiresorptive agent, ZOL, using a standard panel of static and dynamic indices. We also sought to explore the effect of several methods of handling data from biopsies with absent or only single labels on group mean values for MAR.

Materials and Methods

Study population

Healthy, ambulatory postmenopausal women with osteoporosis aged 55–89 yr inclusive were enrolled based on the following criteria: 1) bone mineral density T-score of −2.5 or less at the lumbar spine, total hip, or femoral neck or 2) bone mineral density T-score of −1.5 or less at the lumbar spine, total hip, or femoral neck and fragility fractures (vertebral or nonvertebral). Subjects were required to have normal serum calcium, PTH, and alkaline phosphatase laboratory values and serum 25-hydroxyvitamin D levels of 10 ng/ml or greater at screening. All subjects received supplementation of vitamin D3 800-1200 IU/d or equivalent and calcium 1000 mg/d throughout the study.

Exclusion criteria included prior or current use of TPTD, PTH, or a PTH analog; use of any iv bisphosphonate at any time, or an oral bisphosphonate without a specified time off before study entry (i.e. 6 months off if oral bisphosphonates were used for more than 2 wk but 2 months or less, 1 yr off if used for more than 2 months but 1 yr or less, 2 yr off if used for more than 1 yr); treatment with other bone-active agents for 3–6 months or less of study entry; treatment with systemic glucocorticoids (≥5 mg/d prednisone or equivalent) for more than 30 d within 12 months of study entry; any active or suspected disease known to affect bone metabolism; significantly impaired hepatic or renal function; or any medical conditions that could potentially increase the risk of an adverse event (AE) due to the bone biopsy procedure.

Study design

This was a phase 4, multicenter, randomized, stratified, double-blind, double-dummy, active comparator-controlled study of 12 months' duration. The primary end point was MS/BS in the bone biopsy at month 6. MAR was chosen as an important secondary endpoint. After the biopsy, the study became open label for 6 additional months. The study was approved by the ethics committee at each participating center and conducted in accordance with the Declaration of Helsinki. Written informed consent was obtained from each participant. During the screening period, nonfasting serum procollagen type 1 N-terminal propeptide (P1NP) levels were measured for stratification purposes to achieve between-group comparability for baseline bone turnover. Subjects were assigned to one of three strata, based on baseline P1NP levels of less than 30 μg/liter, 30 μg/liter or greater and 50 μg/liter or less, or greater than 50 μg/liter and then were randomized to either treatment group. During the screening period, subjects were instructed in the use of the delivery device to self-administer the injectable study drug.

Subjects were assigned in a 1:1 ratio as determined by computer-generated random sequence to receive standard doses of TPTD (20 μg administered once daily as a sc injection) or ZOL (5 mg administered at baseline by iv infusion over at least 15 min). All subjects received one active treatment and the alternate placebo.

After blood was collected for laboratory assessments, the first dose of TPTD or matching placebo was administered by the subject into the abdomen or thigh while at the study site. Blinded study personnel administered the iv at the study site. Teriparatide was provided by Eli Lilly and Co. as a prefilled, fixed-dose delivery device. Teriparatide placebo was supplied in an identical delivery device. Zoledronic acid was obtained commercially as prefilled bottles of 100 ml ready-to-infuse aqueous solution. The active and placebo ZOL (saline) were administered to subjects in identical iv bags to preserve the blinding.

Biopsy procedure

At month 3, 250-mg tetracycline capsules were dispensed, and subjects were provided instruction for administration before the bone biopsy procedure. In addition, just before each labeling period, the study site personnel contacted subjects to review the dosing schedule for tetracycline administration to maximize compliance with the dosing schedule. Furthermore, at the time of histomorphometric assessment, biopsies devoid of cancellous tetracycline label were assessed for the presence of label in other bone envelopes to confirm tetracycline administration. Tetracycline labeling began 25 d before biopsy and followed a 3:14:3:5 d schedule (12). A prebiopsy assessment was performed approximately 2 wk before the bone biopsy procedure and included vital signs and laboratory and clinical evaluations to assess any emerging risk for undergoing bone biopsy. Medically cleared subjects underwent trans-iliac bone biopsy at month 6 (5–8 d after the last dose of tetracycline). Biopsies were performed using a Rochester needle or similar large bore (6–8 mm) manual trephine system. All biopsies were analyzed and interpreted by a central bone histomorphometry laboratory that was blinded to the assigned treatment.

Assessments

Preparation and histomorphometric analysis of bone biopsies were performed as previously described (13). Methyl methacrylate-embedded blocks were sectioned with a Reichert Polycut heavy duty microtome. Sections were collected when about one third of the block was trimmed or until a full face was obtained. Seven- and 20-μm-thick sections were collected at three levels of the block with each level approximately 100 μm apart. For each of the three levels, two adjacent 7-μm sections were stained with Goldner's trichrome and toluidine blue for assessment of static parameters, and one adjacent 20-μm section was collected and left unstained for assessment of dynamic parameters. All cancellous bone within the three sections was analyzed. The demarcation between cancellous bone surface and endocortical bone surface was defined following the rules described by Duncan (14) and American Society of Bone and Mineral Research guidelines (1). Depending on the length of each biopsy core, the total cancellous tissue area measured in each subject ranged from 34.7 to 228.6 mm² (mean 86.6 mm²), the trabecular area ranged from 2.2 to 29.3 mm² (mean 13.0 mm²), and the trabecular perimeter ranged from 45.0 to 665.6 mm (mean 227.8 mm).

For this study, we chose MS/BS (expressed as percent of trabecular surface) as the primary endpoint; MS/BS was calculated with the numerator being the sum of the double-labeled surface plus half of the single-labeled surface. For biopsies missing both double and single labels in the region of cancellous bone analyzed, the measured value for MS/BS was zero and these values were included in the calculation of median group values. Secondary end points included MAR and other standard dynamic and static histomorphometric indices.

MAR was calculated using four different methods (Table 1): MAR 1 was derived by measuring samples with double labels only; to obtain MAR 2, we measured samples with double labels and assigned an imputed value of 0.3 μm/d to samples with single labels only (8) and zero to samples with no labels; to obtain MAR 3, we measured samples with double labels, treated samples with single labels only as missing values, and assigned a value of zero for samples with no labels; and to obtain MAR 4, we measured samples with double labels, assigned an imputed value of 0.3 μm/d to samples with single labels only, and treated samples with no labels as missing values. Parameters calculated using MAR, such as BFR/bone surface (BS), activation frequency, etc., were derived using MAR 1. The length of the individual double labels was calculated with the total double label perimeter divided by the total number of the individual double labels. All other parameters were measured, calculated, and expressed according to the recommendations of the American Society of Bone and Mineral Research Nomenclature Committee (1).

Table 1.

Four methods for calculating MAR

End point	Specimens with double label	Specimens with single labels only	Specimens with no labels
MAR 1	Include	Exclude	Exclude
MAR 2	Include	Include (0.3 μm/d)a	Include (0 μm/d)b
MAR 3	Include	Exclude	Include (0 μm/d)b
MAR 4	Include	Include (0.3 μm/d)a	Exclude

^a

MAR is assigned this value for these specimens and they are included in the whole group for calculation of MAR.

^b

MAR is assigned this value for these specimens and they are included in the whole group for calculation of MAR.

Fasting biochemical markers of bone turnover were measured in serum at baseline and months 1, 3, 6, and 12 or at discontinuation. These markers included P1NP (a measure of skeletal bone formation), serum carboxyterminal cross-linking telopeptide of collagen type 1 (CTX; a measure of bone resorption), and serum osteocalcin (a measure of osteoblast function).

Safety evaluations included vital signs and hematology and clinical chemistry measurements, drawn 12–16 h after the dose of injectable study drug, and prebiopsy assessments. Serum calcium was measured at prespecified visits; elevated serum calcium was defined as a serum calcium concentration that exceeded 10.6 mg/dl. In addition, treatment-emergent adverse events (TEAE) and serious adverse events (SAE) were assessed at the time originally reported and subsequently throughout the study. TEAE were defined as any untoward medical occurrence that either occurred or worsened at any time after treatment baseline and that did not necessarily have a causal relationship with treatment. SAE were defined as any AE that resulted in death or an initial or prolonged hospitalization, was life threatening, caused persistent or significant disability/incapacity, caused congenital anomaly or birth defect, required intervention, or was considered significant by the investigator for any other reason. Treatment compliance was assessed at each postbaseline visit after the first dose of TPTD by quantifying the study drugs returned.

Sample size

We anticipated an absolute mean difference of 6.73% in MS/BS between the two groups and sd of 7.83% for the TPTD group and of 2.16% for the ZOL group based on the published literature (6, 15). Therefore, 18 completers per group with evaluable biopsies (with sufficient cancellous bone for accurate measurement) would provide 90% power to detect a significant difference in MS/BS between the two treatment groups at month 6 with a two-sided alpha level of 0.05.

Statistical methods

Efficacy and safety analyses were performed by a modified intent-to-treat principle for all subjects who received one or more doses of the study drug. All tests of treatment effects were conducted with a two-sided alpha level of 0.05 unless otherwise stated. Because there is only one primary analysis, no adjustment was made for multiple comparisons.

Reasons for discontinuation and dropout rates by visit and during the study were compared between the groups using Fisher's exact test. Comparisons of baseline characteristics between the groups were made by t tests for continuous variables and Fisher's exact tests for categorical variables. Differences in the histomorphometric parameters and biochemical markers of bone turnover between the TPTD and ZOL groups were analyzed using the Wilcoxon rank-sum test. The study also explored the correlations between biochemical markers of bone turnover and remodeling assessed by dynamic histomorphometry in cancellous tissue. Spearman's rank correlation test was used to assess the strength of the correlations. Comparisons of TEAE between the groups were conducted using the Fisher's exact test.

Results

Baseline demographics and characteristics

A total of 69 subjects were randomized to study treatment. Eleven subjects discontinued from the study before the 12-month end point: seven discontinued due to AE, two decided to withdraw, one was lost to follow-up, and one was discontinued due to a protocol violation. Three of those who discontinued were subjects receiving TPTD who were noncompliant with study drug at a single visit. Two of these three subjects discontinued before the biopsy; the third subject was included in the histomorphometric analyses. Treatment groups were balanced with regard to age and other baseline characteristics, and the proportion of subjects within each P1NP stratum was not significantly different between the two treatment groups (Table 2).

Table 2.

Baseline demographics and characteristics

Variable	Teriparatide (n = 34)	Zoledronic acid (n = 35)
Age (yr) (mean ± sd)	65 ± 7	64 ± 6
Race (n, %)
Black or African-American	1 (2.9)	1 (2.9)
White	33 (97.1)	34 (97.1)
Weight (kg) (mean ± sd)	61 ± 9	65 ± 12
Previous fracture (n, %)	14 (41.2)	17 (48.6)
Previous osteoporosis therapy (n, %)	17 (50)	21 (60)
T-scores (mean ± sd)
Lumbar spine	−2.79 ± 0.8	−2.93 ± 0.8
Femoral neck	−2.25 ± 0.7	−2.29 ± 0.7
Total hip	−1.78 ± 0.8	−1.88 ± 0.9
Bone turnover markers (mean ± sd)
P1NP (μg/liter)	49.18 ± 18.3	52.88 ± 19.4
CTX (ng/ml)	0.43 ± 0.2	0.42 ± 0.2
OC (μg/liter)	25.05 ± 8.4	24.92 ± 8.5
P1NP strata (n, %)
<30 μg/liter	5 (15.2)	2 (6.1)
≥30 and ≤50 μg/liter	15 (45.5)	14 (42.4)
>50 μg/liter	13 (39.4)	17 (51.5)

OC, Osteocalcin.

Bone histomorphometric results

Fifty-nine subjects completed 6 months of treatment: 58 contributed evaluable biopsy specimens (TPTD, n = 28; ZOL, n = 30), whereas one biopsy sample in the ZOL group was fractured and could not be analyzed. All biopsy samples in the TPTD group had cancellous double labels for analysis (Fig. 1A). In the ZOL group, the cancellous envelope had 16 samples with double labels, two with single labels only, and 12 (40%) with no label (P < 0.001 for between-group comparisons; Fig. 1A). Five of the 12 samples in the ZOL group with no labels in the cancellous compartment did have label in the endocortical or periosteal envelope; seven samples were devoid of label in all three compartments. The percentages of double or single labels per bone surface were significantly higher in the TPTD compared with the ZOL group (Table 3). Furthermore, the median length of each double-labeled surface was significantly greater in the TPTD group (TPTD = 0.35 mm; ZOL = 0.24 mm, P = 0.002).

Fig. 1.

Distribution of tetracycline labels (A) and MS/BS (B) in the TPTD and ZOL acid groups: values are medians (interquartile range). Representative photomicrograph of tetracycline labels in cancellous bone from the ZOL (MS/BS = 0.20%) (C) and TPTD (MS/BS = 5.38%) (D) groups.

Table 3.

Dynamic and static indices of bone formation and resorption assessed in cancellous bone at month 6, median (interquartile range)

Dynamic parameters	Teriparatide, n		Zoledronic acid, n		Wilcoxon rank-sum P Value
Mineralizing surface/bone surface (MS/BS, %)	28	5.60 (3.00, 11.75)	30	0.16 (0.00, 0.58)	<0.001
Activation frequency (Ac.f, per year)a	28	0.37 (0.17, 0.82)	16	0.03 (0.01, 0.04)	<0.001
Bone formation rate/bone surface (BFR/BS, mm3/mm2/yr)a	28	0.0116 (0.0051, 0.0265)	16	0.0009 (0.0003, 0.0012)	<0.001
Bone formation rate/bone volume (BFR/BV, % per year)a	28	0.21 (0.10, 0.43)	16	0.01 (0.00, 0.03)	<0.001
Mineral apposition rate (MAR, μm/d)a	28	0.56 (0.48, 0.62)	16	0.50 (0.40, 0.52)	0.031
Adjusted apposition rate (Aj.AR, μm/d)a	28	0.34 (0.15, 0.51)	16	0.05 (0.02, 0.18)	<0.001
Mineralization lag time (Mlt, d)a	28	13.63 (11.22, 29.22)	16	75.72 (26.84, 193.78)	<0.001
Osteoid maturation time (Omt, d)a	28	9.99 (8.39, 10.74)	16	9.05 (7.83, 10.76)	0.502
Total formation period (Tt.FP, d)a	28	0.24 (0.16, 0.50)	16	1.46 (0.44, 3.10)	<0.001
Active formation period (a.FP, d)a	28	0.15 (0.14, 0.17)	16	0.16 (0.15, 0.19)	0.154
Double-label surface/bone surface (dLS/BS, %)	28	4.13 (2.14, 8.88)	30	0.07 (0.00, 0.30)	<0.001
Single-label surface/bone surface (sLS/BS, %)	28	3.19 (1.58, 4.86)	30	0.02 (0.00, 0.38)	<0.001
Double-label length (mm)	28	0.35 (0.28, 0.38)	16	0.24 (0.19, 0.30)	0.002
Static parameters, bone formation
Osteoid volume/bone volume (OV/BV, %)	28	1.32 (0.81, 2.13)	30	0.24 (0.08, 0.49)	<0.001
Osteoid surface/bone surface (OS/BS, %)	28	11.34 (6.58, 16.52)	30	2.51 (1.27, 4.58)	<0.001
Osteoid thickness (OsTh, μm)	28	4.92 (4.29, 6.68)	30	3.77 (3.51, 4.22)	<0.001
Wall thickness (WTh, μm)	28	31.29 (28.80, 33.26)	30	28.63 (27.16, 30.43)	0.014
Static parameters, bone resorption
Eroded surface/bone surface (ES/BS, %)	28	4.59 (3.14, 6.01)	30	2.71 (1.73, 3.21)	<0.001

^a

MAR and parameters derived from MAR were calculated using measurements from subjects with double labels, i.e. MAR 1 (n = 28 vs. 16). The parameters not affected by the calculation of MAR were calculated from all evaluable biopsies (n = 28 vs. 30).

Median values for MS/BS were significantly higher in the TPTD group than in the ZOL group (5.60 vs. 0.16%, P < 0.001) (Fig. 1B) at month 6. This difference is apparent in photomicrographs of representative biopsies from subjects treated with either ZOL (Fig. 1C) or TPTD (Fig. 1D). Results for MAR, calculated four different ways as described in Materials and Methods, are shown in Fig. 2. The median value for MAR was significantly greater with TPTD than with ZOL, regardless of how MAR was calculated; the difference was most evident in MAR 2 and 3 in which MAR was assigned a value of zero in the 12 ZOL samples that were devoid of labels. Significantly fewer subjects with double labels were available for analysis in the ZOL group than in the TPTD group (53 vs. 100%; P < 0.001; Fig. 1A). The median values for most of the other dynamic indices were significantly higher in the TPTD group (Table 3). Static indices of bone formation and resorption were significantly higher with TPTD than with ZOL.

Fig. 2.

MAR in the TPTD and ZOL groups: values are medians (interquartile range). MAR was calculated in four different ways. MAR 1, Double labels only. MAR 2, Double labels, single labels: MAR = 0.3 μm/d, no labels: MAR = zero. MAR 3, Double labels, single labels: MAR = missing, no labels: MAR = zero. MAR 4, Double labels, single labels: MAR = 0.3 μm/d, no labels: MAR = missing.

Serum bone turnover markers

The different effects of the two drugs on bone formation (P1NP) and resorption (CTX) markers were evident after 1 month, and the differences were significant (P < 0.001) between groups at months 1, 3, 6, and 12 (Fig. 3, A and B). Treatment with TPTD produced significant increases in P1NP above baseline at all time points measured. However, CTX did not increase significantly above baseline for subjects receiving TPTD until month 3. Treatment with ZOL decreased P1NP and CTX significantly below baseline by month 1, and these values remained low through month 12. Similar results were seen with OC (data not shown). The levels of P1NP at month 6 were strongly correlated with MS/BS at the same time point (Fig. 3C).

Fig. 3.

Percentage change from baseline to month 12 (mean ± se) in serum P1NP and CTX for ZOL (A) and TPTD (B) (P < 0.001 for all between-group comparisons for change from baseline). C, Correlation of P1NP and MS/BS at month 6.

The Spearman correlation analysis was performed with and without the data point corresponding to a MS/BS of 50.5% and a P1NP of 710 μg/liter at 6 months. In each analysis, the data were strongly correlated (r = 0.85 with the outlier and r = 0.84 without the outlier) and highly significant (P < 0.0001 for both comparisons). After assessing this subject's laboratory data for baseline and posttreatment serum markers, the investigator determined that the clinical data suggested a robust response to TPTD, that the subject required no further follow-up, that she should not be discontinued from the study, and that her data were appropriate to include in the analysis.

Safety

The incidence of TEAE was similar in each group at month 12. In the overall study population, 88% of subjects experienced one or more TEAE. The most commonly occurring events in the TPTD group were arthralgia (23.5%); nasopharyngitis (17.6%); procedural pain, muscle spasms, and pain in extremity (each 14.7%); postprocedural hematoma and hematoma (each 11.8%); and back pain, headache, and nausea (each 8.8%). The most commonly occurring events in the ZOL group were arthralgia, back pain, and headache (each 20%); cough (17.1%); nasopharyngitis, nausea, pain, and postprocedural hematoma (each 11.4%); and chills, hypertension, muscle spasms, dizziness, and vomiting (each 8.6%). Most TEAE were of mild or moderate severity; there were no significant differences between groups in the incidence of TEAE that were classified as severe. Five subjects in the TPTD group and two in the ZOL group discontinued the study because of TEAE.

Four subjects experienced an SAE before the 12-month end point: one subject receiving TPTD (noncardiac chest pain) and three receiving ZOL (increase in size of preexisting liver hemangioma, ophthalmic aneurysm, and asthmatic bronchitis); none of these SAE was considered to be related to study drug. No subject withdrew as the result of an SAE. There were no deaths in this study.

Vital signs were within normal limits at baseline for all randomized subjects and did not change in any clinically meaningful way in either group over the 12-month study. Elevated laboratory values for serum calcium were observed in seven subjects in the TPTD group, although clinical symptoms of hypercalcemia were not observed. Four subjects had an elevated serum calcium value at baseline or 1 month after treatment that was normal at retest; one subject had an elevated serum calcium value at month 12; and one subject had elevated serum calcium values at month 1 and month 12. The latter two subjects completed the study at month 12, and study drug was discontinued. One additional subject had persistent but mild hypercalcemia as observed by laboratory assessment; this subject discontinued from the study after about 3 months because of a TEAE (arthralgia). There were no statistically significant or clinically relevant differences in mean laboratory values between treatment groups.

Discussion

In this study, TPTD and ZOL were shown to possess fundamentally different mechanisms of action with opposite effects on bone formation based on data evaluating a standard panel of dynamic and static histomorphometric bone indices and serum markers of bone turnover. The opposite effects on bone formation described here are also consistent with previously published data exploring the effect of these two drug classes on bone when compared with placebo. Stimulation of osteoblast activity with TPTD, as evidenced by significantly higher indices of bone formation, is consistent with previous histomorphometric studies evaluating TPTD compared with placebo (16, 17) and confirms the anabolic MOA of this drug. Inhibition of osteoclast activity by ZOL reported here, as suggested by significantly lower bone resorption and bone formation indices, is characteristic of skeletal effects previously reported for bisphosphonates vs. placebo (18, 19) and confirms the antiresorptive activity of ZOL.

The ability of a standard panel of histomorphometric indices to differentiate the MOA of these two drugs is illustrated graphically in Fig. 4. The plots show that there is very little overlap between the effect of the two drugs on the dynamic indices, with the clearest separation in the dynamic parameters of MS/BS, BFR/BS, and total formation period (Fig. 4A). However, there is overlap in MAR, which indicates that MAR alone is not a good parameter to distinguish anabolic from antiresorptive activities of drugs. Furthermore, the static parameters alone (Fig. 4B) are not useful in this regard because they differ only in magnitude rather than in direction.

Fig. 4.

Spider plots of dynamic (A) and static histomorphometric (B) parameters in the TPTD and ZOL groups: values are medians based on MAR 1 analyses.

The results of the current trial are consistent with the findings of Arlot et al. (20), which was the first study to provide a direct comparison of TPTD with a bisphosphonate (alendronate) in 23 women with postmenopausal osteoporosis. In that study, MS/BS for TPTD was 8.1% compared with 0.22% for alendronate at month 6 (P < 0.001); in our study, MS/BS for TPTD was 5.60% compared with 0.16% for ZOL at month 6 (P < 0.001). Similar results were found across the panel of histormorphometric indices, assessed by Arlot et al. (20), including MAR, which was significantly higher for subjects receiving TPTD compared with those receiving alendronate (0.77 vs. 0.49 μm/d; P < 0.01). It should be noted, however, that Arlot et al. did not specify how MAR was calculated in the setting of single or missing labels.

The findings for serum bone turnover markers in the current study further support the mechanistic differences observed by histomorphometry and are consistent with data from an earlier study by McClung et al. (21), which demonstrated the distinct MOA of TPTD compared with alendronate based on serum P1NP and urinary N-telopeptide corrected for creatinine, another marker of bone resorption. In the current trial, we also demonstrated that serum P1NP was strongly correlated with MS/BS at 6 months.

The significant number of samples in the ZOL group that had no label (n = 12) and, therefore, where MS/BS = zero, as well as those samples with a very low value for MS/BS, were largely responsible for driving differences in bone formation rate between groups. MAR 1 through MAR 4 were all significantly lower in the ZOL group than in the TPTD group, but the difference was most evident when MAR was assigned a value of zero in the 12 samples that showed absent labels (MAR 2 and MAR 3). This suggests that the most accurate way to assess MAR may be to measure it only in samples with double labels, i.e. MAR 1 in this study. In the current study, when MAR was measured only in samples with double labels, it was still significantly higher in the TPTD group than in the ZOL group. In the absence of a placebo arm, we cannot determine whether this is due to a stimulation of MAR in the TPTD group or inhibition of MAR in the ZOL group. Alternately, this difference could be due to a sampling issue as there were many fewer double labels in the ZOL group. However, considering the large differences in MS/BS between these two drugs, small differences in MAR, even if statistically significant, would have minimal impact on BFR/BS. This emphasizes why MS/BS, rather than MAR, is a superior index to differentiate between anabolic and antiresorptive activity and why MAR should not be considered separately in making such a distinction. This is particularly relevant for potent antiresorptive agents because the small number of double labels makes interpretation of MAR problematic.

Recently Recker et al. (4) proposed guidelines to improve the consistency of reporting data from histomorphometric studies. For MS/BS, the authors recommend reporting the measured value for biopsies with double and/or single labels and assigning a value of zero to MS/BS when no label is found. For MAR, the authors recommend reporting the measured value when the double label is present with adequate sampling, i.e. MAR 1 in the present study (Tables 1 and 3). In samples with only single or no label, they recommend that MAR be counted as missing or assigned a value of 0.3 μm/d, provided that the value for MAR is reported with and without the samples in which an imputed value was used. The authors acknowledge that imputation may underestimate MAR, whereas counting single or no label cases as missing will likely overestimate MAR. This issue is apparent in the current study in which point estimates are higher for MAR 1 (single and no label cases counted as missing) and MAR 4 (single labels imputed, no labels counted as missing) than for MAR 2 or MAR 3, when biopsies with no labels are assigned a value of zero irrespective of whether an imputed value of 0.3 μm/d is assigned to the two subjects with single labels only in the ZOL group. These findings underscore the potential pitfalls of interpreting the value of MAR in the setting of low bone turnover and confirm the importance of explicitly stating the distribution of single and/or double labels across the samples, the percent labeled surface, and the method for calculating MAR. As Recker et al. (4) noted, in the absence of a mineralization defect, MAR reflects the rate of matrix production in individual remodeling units, and measurement of MAR alone is not sufficient to define anabolism.

The primary strengths of this study are 2-fold: 1) it used an active-comparator design to provide a head-to-head comparison of TPTD and ZOL across a standard panel of histomorphometric indices and was well powered to detect significant differences between groups, and 2) it compared different approaches to assessing MAR in the setting of biopsies with missing or only single labels. Regardless of how MAR was calculated in the study, it was significantly lower in women receiving ZOL than in those receiving TPTD.

There also were several limitations to this study: there was no placebo group, and the cross-sectional design of the study did not allow us to assess histomorphometric changes with respect to baseline. We also did not perform an extended search for labels as has recently been done in some studies (6, 7); however, we did sample the biopsies thoroughly by analyzing the entire cancellous compartment at three levels in all biopsies in a systematic and blinded fashion. Finally, the duration of the study was too short to evaluate potential differences between the effects of the two agents on bone structure and matrix properties.

Conclusions

The results of this study define the dynamic histomorphometric characteristics of an anabolic drug, TPTD, vs. an antiresorptive drug, ZOL, and, when interpreted with the results of analyses of serum markers of bone turnover, underscore that TPTD and ZOL possess fundamentally distinct mechanisms of action with opposite effects on bone formation. Furthermore, this study demonstrates the importance of reporting a full panel of histomorphometric parameters, including MS/BS and BFR/BS, and the need to state clearly the number of samples with double labels, single labels, or no labels, as well as the method used to calculate MAR, particularly in the setting of low turnover. This will continue to be an important requirement in defining the mechanisms of action of new drugs under development, especially because some may be formation-sparing or mixed antiresorptive/anabolic agents.

Acknowledgments

We acknowledge Eileen R. Gallagher, a full-time employee of PharmaNet/i3, a part of the inVentiv Health Co., for her assistance in preparing this manuscript.

This work was sponsored by Lilly USA, LLC.

Disclosure Summary: D.W.D. has received research grants from Eli Lilly; is a consultant and on the advisory board for Eli Lilly, Amgen, and Merck; is on the speakers' bureau at Eli Lilly, Amgen, Novartis, Procter & Gamble, and Merck; and has received honoraria from Eli Lilly, Amgen, Novartis, Procter & Gamble, and Merck. H.Z. and R.R.R. have nothing to disclose. J.P.B. is on the advisory council for and has received research grants from Amgen, Eli Lilly, Merck, Novartis, and Warner Chilcott; and has received honoraria from Amgen, Eli Lilly, and Novartis. M.A.B. has served as a speaker for Eli Lilly. C.P.R. is a consultant for and has received advisory fees from Takeda, Dramatic Health, Novartis, and Eli Lilly and lecture fees from Amgen, Novartis, Publicis Meetings, and Warner Chilcott. D.L.K. is on the advisory council and speakers' board for and has received honoraria from Eli Lilly and has received research grants from Amgen, Glaxo-Smith Kline, Johnson & Johnson, Eli Lilly, Novartis, and Roche. E.M.L. has received research grants from Amgen, Eli Lilly, Novartis, Merck, Warner Chilcott, and Glaxo-Smith Kline; is on the advisory board for Amgen, Eli Lilly, Novartis, and Merck; and is on the speakers' board for and has received honoraria from Amgen, Eli Lilly, Novartis, and Warner Chilcott. D.A.H. is on the advisory council for and has received honoraria from Eli Lilly and Novartis. D.S.R. is on the speakers' bureau for and has received research grants from Eli Lilly. P.D.M. has received research grants from and served on the advisory council for Eli Lilly and Novartis; has received honoraria from Novartis; and has provided expert testimony for Novartis. G.C.W. is on the speakers' bureau for and has received research grants from Eli Lilly and Amgen. R.L. has received research grants from, has been a consultant for, and has served on the advisory council for Eli Lilly; has been a speaker for Amgen and Novartis; and has received honoraria from Eli Lilly and Amgen. N.B. has a financial relationship with Eli Lilly, Amgen, Merck, and Tarsa Therapeutics; and is a consultant and is on the advisory council for Eli Lilly and Merck. X.W., V.A.R., and K.A.T. are employees of Eli Lilly, and B.J. is employed by Lilly Canada, Inc.; each owns stock in Eli Lilly.

Abbreviations

 
• AE

  Adverse event

 
• BFR

  bone formation rate

 
• BS

  bone surface

 
• CTX

  carboxyterminal cross-linking telopeptide of collagen type 1

 
• MAR

  mineral apposition rate

 
• MOA

  mechanism of action

 
• MS/BS

  mineralizing surface/BS

 
• P1NP

  procollagen type 1 N-terminal propeptide

 
• SAE

  serious AE

 
• TEAE

  treatment-emergent AE

 
• TPTD

  teriparatide

 
• ZOL

  zoledronic acid.