Drs Marcus, Wang, Satterwhite, and Mitlak are full-time employees of Eli Lilly & Co
In a recent study of women with postmenopausal osteoporosis, treatment with teriparatide for a median of 19 months increased bone mineral density and decreased the risk of vertebral and nonvertebral fractures. Using the same cohort, the current study evaluated the relationship between these therapeutic effects and the patient's baseline age, vertebral bone mineral density, and prevalent vertebral fractures. In women over 65 years of age, treatment resulted in a greater increase in vertebral bone mineral density than in younger women (treatment-by-age interaction, p = 0.037), but baseline age had no effect on the relative risk reduction for vertebral fractures (treatment-by-age interaction, p = 0.558). In women receiving placebo (with calcium and vitamin D), there was an inverse relationship between baseline vertebral bone mineral density and vertebral fracture risk. When compared across bone mineral density tertiles, the effects of teriparatide on the relative risk for developing new vertebral fractures and increase in vertebral bone mineral density did not differ significantly (p = 0.817 and p = 0.615, respectively). Teriparatide treatment significantly decreased vertebral fracture risk in patients with a vertebral bone mineral density T score of less than −3.3 or a score between −2.1 and −3.3 (p < 0.001 and p = 0.027, respectively) and showed a trend toward reduced fracture risk in the group with a T score greater than −2.1 (p = 0.115). Placebo-treated women with two or more prevalent vertebral fractures had a significantly greater risk of developing new vertebral fractures than women with zero or one prevalent vertebral fracture (p < 0.001). When compared within prevalent vertebral fracture subgroups, the effects of teriparatide on the relative risk for developing new vertebral fractures were similar. The results of this study indicate that teriparatide offers clinical benefit to patients across a broad range of age and disease severity.
OSTEOPOROSIS IS currently managed with a range of therapies that primarily decrease bone resorption.(1) Soon patients with osteoporosis may also benefit from bone formation therapy.
Daily subcutaneous administration of teriparatide (rDNA origin) injection [rhPTH(1–34)] has been shown to increase bone mass and strength in several species.(2) Studies in humans have also confirmed that teriparatide increases bone mass and have suggested that this agent will also reduce the risk of vertebral fracture.(3, 4) In a large multinational study of women with osteoporosis, treatment with teriparatide (20 μg/day) for a median of 19 months was shown to increase spinal and femoral neck bone mineral density (BMD) and total body bone mineral content (BMC) and to reduce the risk of vertebral and nonvertebral osteoporotic fractures by 65% and 53%, respectively.(5) Importantly, teriparatide administration leads to improvement in skeletal microarchitecture in both animals(6) and humans.(7) Thus, teriparatide seems to have clinical potential as a bone-forming agent for the treatment of osteoporosis.
Several clinical trials have suggested that the response to treatment with antiresorptive agents may depend on the subjects' characteristics, such as baseline BMD and prevalent fractures.(8, 9) The current analyses were therefore conducted to evaluate the relationship between these same baseline characteristics of participants in the teriparatide fracture prevention study(5) and their subsequent BMD and fracture response to teriparatide.
MATERIALS AND METHODS
A brief summary of the clinical trial will be presented here; a full report has been published by Neer et al.(5) A total of 1637 ambulatory postmenopausal women, 42–86 years of age, were enrolled in the study. They were all at least 5 years beyond menopause and had a minimum of either one moderate or two mild atraumatic vertebral fractures and a minimum of seven evaluable nonfractured vertebrae on baseline spinal radiographs. Women with fewer than two moderate fractures were also included if their hip or vertebral BMD values were less than 1 SD below the mean value for normal premenopausal white women (i.e., T score < −1.0). Women with diseases related to bone or calcium metabolism, urolithiasis within the preceding 5 years, alcohol or drug abuse, impaired hepatic function, or a serum creatinine concentration ≥177 μM were excluded from the study. Women who had taken drugs affecting bone metabolism within the past 2–24 months were also excluded. An ethics committee at each participating center approved the study, and all participants gave written informed consent.
Each woman received daily supplements of calcium (1000 mg) and vitamin D (400–1200 IU) and daily injections of placebo for 2 weeks before randomization. Women were subsequently divided into the following treatment groups: placebo (n = 544), teriparatide 20 μg (n = 541), and teriparatide 40 μg (n = 552). All women were on a regimen of daily, self-administered subcutaneous injections. All participants received calcium and vitamin D throughout the trial.
During the course of this trial, results of a long-term rodent teriparatide carcinogenicity study revealed the dose-related occurrence of skeletal proliferative lesions, including osteosarcoma. Accordingly, Eli Lilly and Company suspended all clinical trials with teriparatide until the meaning of this finding could be thoroughly evaluated. All participants were called back to their individual study site for repeat assessments of BMD and spine radiographs. Rather than achieving the planned 36 months of teriparatide administration for all subjects, the median duration of drug exposure was 19 months with no subject exposed to drug for more than 25 months. Thus, results reported here reflect changes from baseline to study endpoint(s). Additional information regarding the rat osteosarcoma finding has been reported.(5,10,11)
Assessment of drug efficacy
The number of women experiencing new vertebral fractures during the course of the trial was determined by evaluating anteroposterior and lateral radiographs of the thoracic and lumbar spine of each participant at baseline and at the end of the study. Individual vertebrae were graded by the semiquantitative method of Genant et al.(12) as normal (i.e., normal height) or as mildly, moderately, or severely deformed (representing decreases in the anterior, middle, or posterior vertebral heights of 20–25%, 26–40%, or >40%, respectively). Vertebrae exhibiting scoliosis, fusion, or other anomalies were excluded from the analysis. A new vertebral fracture was reported if a normal vertebra became deformed.
The effect of teriparatide on the incidence of nonvertebral fractures also was reported by Neer et al.(5) As the number of nonvertebral fractures was not sufficient to support an analysis stratified by baseline characteristics, they will not be considered here.
Lumbar spine BMD was assessed periodically by dual-energy X-ray absorptiometry using Hologic (Hologic Corp., Bedford, MA, USA), Norland (Norland Corp., Ft. Atkinson, WI, USA), and Lunar equipment (Lunar Corp., Madison, WI, USA). Spine BMD values were converted to standardized units (expressed as milligrams per square centimeter) to eliminate differences attributable to densitometer manufacturer.(13) The percent change in BMD from baseline to endpoint(s) after therapy was calculated for each patient. Vertebral fracture sites and regions of severe scoliosis were excluded from lumbar spine BMD measurements. Serial measurements of a spine phantom at each center were used to assess the consistency of measurements between centers and to adjust for minor changes in densitometer performance.(14)
For the continuous response variables of BMD changes, subgroup analyses were performed using analysis of variance (ANOVA). Effects in the model included country, treatment, country-by-treatment interaction, subgroup, and treatment-by-subgroup interaction. Type III sum of squares was used to assess the treatment-by-subgroup interaction. Within each level of the subgroup, treatment difference was assessed using an ANOVA with country and treatment in the model.
For the binary response variable of new vertebral fractures, subgroup analyses were performed using logistic regression. Effects in the logistic regression model included treatment, subgroup, and treatment-by-subgroup interaction. The treatment-by-subgroup interaction was examined using a likelihood ratio test. Treatment effect within each level of the subgroup was assessed using a likelihood ratio test based on a logistic regression model with only the term of treatment.
A typical subgroup based on tertiles has approximately 180 patients per treatment group. For the observed vertebral fracture incidence rates of 14% versus 5% (placebo vs. teriparatide 20 μg) in Neer et al.,(5) each subgroup has approximately 80% power to detect such treatment group differences. When analyzing changes in lumbar spine BMD, the power is well beyond 99%.
In the study, 1637 women were randomly assigned to one of three treatment groups: teriparatide 20 μg, teriparatide 40 μg, or placebo. These groups were well matched for age, vertebral BMD, and prevalent vertebral fractures, as well as for other characteristics that could affect the risk of fracture (Table 1).
Table Table 1.. Baseline Demographic and Clinical Characteristics of Study Population
Vertebral BMD response to teriparatide
Teriparatide treatment caused rapid increases in vertebral BMD. Statistically significant increases in vertebral BMD were observed after 3 months of treatment (Fig. 1). Compared with placebo, vertebral BMD had increased 10% and 14% at study endpoint(s) (median of 19 months) in women assigned to treatment with teriparatide 20 μg and 40 μg, respectively. This increase in BMD was associated with a 65% and 69% reduction, respectively, in the incidence of new vertebral fractures.(5)
Baseline age, change in vertebral BMD, and risk of vertebral fracture
To evaluate the relationship between baseline age and response to treatment, women were divided into three age subgroups: <65 years, 65 to <75 years, and ≥75 years. The mean age was 69 ± 7 years. In women assigned to treatment with placebo, there was a small age-related increase in mean vertebral BMD during the study (Fig. 2A). Compared with placebo, treatment with teriparatide significantly increased vertebral BMD, regardless of age, although the percentage increase was greater in those in the older subgroups (treatment-by-age interaction, p = 0.037). This statistically significant interaction could be attributed to the fact that the study is overpowered with respect to analysis of BMD. However, we note that even in the placebo group the change in spine BMD from baseline was higher in older women. This finding may reflect the fact that older women may have been relatively more deficient in calcium and vitamin D on entry into the trial.
A greater proportion of women assigned to placebo had an incident vertebral fracture in the older subgroups (Fig. 2B). Treatment with teriparatide was associated with a similar reduction in the relative risk of fracture in each subgroup of age (treatment-by-age interaction, p = 0.558).
Baseline vertebral BMD, change in vertebral BMD, and risk of vertebral fracture
The relationship between baseline vertebral BMD and response to treatment was evaluated by dividing the women into three equal BMD subgroups (i.e., tertiles). This resulted in assignment of women with vertebral BMD T scores < −3.3 to the low BMD group, those with T scores between −3.3 and −2.1 to the middle group, and those with T scores > −2.1 to the high group. The mean lumbar vertebral BMD T score was −2.6 ± 1.4 (SD).
In the subgroups of women assigned to placebo, vertebral BMD increased by a mean of 1–2%, regardless of baseline BMD (Fig. 3A). Vertebral BMD increased significantly and in a dose-related fashion in women assigned to treatment with teriparatide. The percentage increase in vertebral BMD was inversely related to the baseline vertebral BMD.
Because percentage change in BMD is dependent on the baseline BMD, actual change in BMD was also examined in the subgroups. In contrast to the inverse relationship between baseline BMD and percentage change in BMD after teriparatide therapy, women had similar absolute increases in vertebral BMD regardless of baseline BMD subgroup (Fig. 3A; treatment-by-bone mineral density interaction, p = 0.615).
In the placebo group, the relative risk for developing new vertebral fractures was inversely related to baseline vertebral BMD. New vertebral fractures occurred in 22.1%, 11.0%, and 7.9% of women in the low, medium, and high vertebral BMD tertiles, respectively (Fig. 3B; p < 0.001).
When compared within subgroups of baseline BMD, the effect of treatment with teriparatide on the relative risk for developing new vertebral fractures did not differ significantly (Fig. 3B; treatment-by-baseline BMD T score interaction, p = 0.817).
When the response was examined in women with a vertebral BMD T score of < −2.5 or ≥ −2.5, the corresponding percentage of women with fractures was 19.2%, 6.2%, and 4.0%, respectively, for the placebo, teriparatide 20 μg, and teriparatide 40 μg groups with scores < −2.5, and it was 7.5%, 3.4%, and 4.0%, respectively, for the placebo, teriparatide 20 μg, and teriparatide 40 μg groups with scores ≥ −2.5. A similar response by subgroups (treatment-by-baseline BMD interaction, p = 0.203) was observed, although treatment p values were p < 0.001 for the group with a T score of < −2.5 and p = 0.130 for the group with a T score of ≥ −2.5.
Because baseline vertebral BMD is inversely correlated with body weight, the influence of baseline body weight on BMD absolute change was also examined. Although treatment-by-weight interaction was statistically significant, the magnitude of treatment response across weight groupings was not clinically different. The absolute changes in vertebral BMD for the low, medium, and high body weight subgroups, respectively, were 0.00, 0.01, and 0.01 g/cm2 for the placebo group; 0.07, 0.08, and 0.07 g/cm2 for the teriparatide 20 μg group; and 0.12, 0.10, and 0.10 g/cm2 for the teriparatide 40 μg group.
Impact of prevalent vertebral fractures on the incidence of new vertebral fracture
The relationship between the number of prevalent vertebral fractures and response to treatment was evaluated. The mean number of fractures was 2.3 ± 1.8. Ten percent of patients were enrolled in the study but were reclassified as having no vertebral fracture at endpoint(s); 90% had at least one fracture and 60% had two or more fractures. The percentage increase in BMD was similar, regardless of whether women had no vertebral fractures, one fracture, or two or more fractures.
Women in the placebo group with at least two prevalent vertebral fractures had a significantly greater risk for developing new vertebral fractures (19.9% of patients) than patients with zero or one prevalent vertebral fracture (4–6.8% of patients; p < 0.001; Fig. 4). Teriparatide reduced the percentage of patients who developed new vertebral fractures, regardless of the number of prevalent fractures (Fig. 4).
Comparisons within vertebral fracture prevalence subgroups showed that a decrease in the relative risk for developing new vertebral fractures was similar for each dose of teriparatide (treatment-by-prevalent vertebral fracture interaction, p = 0.649).
This extended analysis from the pivotal trial of teriparatide in postmenopausal osteoporotic women show substantial independence of the beneficial skeletal effects of teriparatide from the variables of age, baseline vertebral BMD, and prevalent fractures. Although women with the lowest vertebral BMD on study entry exhibited a greater percentage increase in vertebral BMD after teriparatide therapy than those with higher initial vertebral BMD, absolute changes in vertebral BMD, expressed as grams per square centimeter, were the same. These observations suggest that baseline vertebral BMD values did not influence the ability of teriparatide to increase vertebral BMD.
The risk reduction for vertebral fractures in women treated with teriparatide was approximately the same for each vertebral BMD subgroup, regardless of the fact that patients in the lower BMD tertile had a greater risk for experiencing a vertebral fracture. However, fracture protection for women in the highest BMD tertile did not achieve statistical significance, reflecting the relatively low event rate in this subgroup. Although increases in vertebral BMD over time were greater in older patients assigned to placebo, age had only a modest effect on the degree to which teriparatide enhanced vertebral BMD, and had no effect on the reduction of vertebral fracture risk. This is especially important because more than 50% of all women over the age of 75 years are estimated to have osteoporosis, have a higher risk of fracture caused also by their advanced age, and should therefore benefit from the bone-forming effects of teriparatide.(15) Alendronate has also been shown to reduce vertebral fractures to a similar degree in osteoporotic patients in diverse age groups (i.e., <65 years of age and ≥65 years of age).(16)
In the current study, patients with two or more prevalent vertebral fractures were at higher risk for developing new vertebral fractures than those with fewer than two prevalent vertebral fractures. However, teriparatide decreased the risk of new vertebral fractures regardless of the prevalent fracture status of these patients before treatment.
We note that women treated with placebo experienced gains in BMD during the course of this trial, albeit to a degree that was significantly lower than that associated with teriparatide. There are two plausible explanations for this observation. These women may have shown a beneficial response to the supplemental calcium and vitamin D that was provided to all participants. Alternatively, an increased BMD over time in older populations could reflect the gradual accumulation of degenerative changes in the spine.
Teriparatide represents a potentially important therapeutic agent for women with postmenopausal osteoporosis, showing significant anabolic effects on bone as reflected by increased vertebral BMD and a decrease in new vertebral fracture risk. The results of the present studies indicate that teriparatide offers clinical benefit to patients across a broad range of age and disease severity.
The authors thank Mary Ellen Perron for preparation of the figures and assistance with the manuscript and the study's chief investigators.(5) This study was funded by Eli Lilly and Company. The study drug was provided by Eli Lilly and Company, Indianapolis, Indiana.