Effects of raloxifene and alendronate on bone turnover as assessed by procollagen type I N-terminal propeptide
- First Online:
- Cite this article as:
- Eastell, R., Rogers, A., Ni, X. et al. Osteoporos Int (2011) 22: 1927. doi:10.1007/s00198-010-1380-5
- 190 Views
Raloxifene decreases PINP into the lower half of the premenopausal reference interval; alendronate decreases PINP more, with approximately 60% of alendronate-treated women having PINP concentrations below the lower limit of the premenopausal reference interval.
The purpose of this study was to evaluate the distribution of serum procollagen type I N-terminal propeptide (PINP) concentrations in women with postmenopausal osteoporosis prior to treatment and after treatment with either raloxifene or alendronate for 12 or more months.
Included were data from 1,323 postmenopausal women aged 45 to 87 years, collected at baseline or after treatment with either alendronate 10 mg/day or raloxifene 60 mg/day. These patients had participated in one of four clinical trials in which intact PINP was measured by radioimmunoassay (Orion Diagnostica). A premenopausal reference interval from 16.0 to 75.8 μg/L was determined from 68 premenopausal, non-pregnant women.
Most postmenopausal osteoporotic patients prior to treatment had PINP values in the upper half of the premenopausal reference interval at baseline (70%). After ≥12 months of therapy, most patients who received raloxifene had PINP concentrations in the lower half of the premenopausal reference interval (58%), whereas among those who received alendronate, around 60% of patients had PINP concentrations below the lower limit of the premenopausal reference interval.
PINP may be useful for assessing differences in bone turnover response to different types of anti-resorptive therapy.
Bone remodeling is the continuous physiologic process by which old or damaged bone is broken down and replaced by new bone to maintain mechanical strength .
Before menopause, when there are high levels of endogenous estrogen production, bone resorption and formation are said to be balanced . At menopause, endogenous estrogen levels are reduced and women have an increase in bone turnover, with an excess of bone resorption versus formation . The level of bone turnover optimal for long-term skeletal health is unknown; however, we do know that excessive bone turnover with a negative balance of formation to resorption is unhealthy for the skeleton. It may lead to decreases in bone mineral density (BMD) and bone quality . Conversely, excessively low bone turnover, such as that which occurs in osteopetrosis, also results in a high risk of fracture. This is due to an increase in bone age (as old bone is not replaced), an accumulation of damage, and excessive mineralization of the skeleton .
Anti-resorptive drugs increase bone strength and reduce fractures by rapidly decreasing bone turnover, resulting in a decrease in resorption site number and depth with an associated decrease in the risk of perforation . Anti-resorptive therapies result in decreases in bone formation markers as well as bone resorption markers, as these processes are coupled; thus, treatment response can be detected by the lowering of bone formation markers as well as bone resorption markers. In clinical practice, the change in bone turnover markers from pretreatment baseline is used to assess therapeutic response . A natural way to describe the concentration of bone turnover markers in women after anti-resorptive treatment is in relation to levels found in healthy premenopausal women, as healthy premenopausal women are thought to have a level of turnover associated with long-term skeletal health and a stable bone mass. With this approach, levels can be described in qualitative terms (high, upper half of normal, lower half of normal, and low). A T-score approach and a least significant change approach have also been described [6, 7].
The bone formation marker PINP has the potential to be a particularly useful bone turnover marker . It is released into the circulation during the posttranslational cleavage of propeptides of the collagen precursor, procollagen type I . The assays for PINP recognize different species; some assays recognize only the trimeric form (“intact PINP”) and some also recognize monomeric fragments (“total PINP”). Procollagen type I N-terminal propeptide has been included in a number of reference interval studies, is a stable analyte with little effect of temperature or hemolysis, and is not sensitive to feeding or circadian rhythm . Measurement variability for PINP is low compared to other markers, and the favorable signal-to-noise ratio of treatment response and the relatively early response compared to changes in bone density make PINP an appropriate marker to compare treatment responses between therapies .
Alendronate and raloxifene are both licensed anti-resorptive therapies for the prevention and treatment of postmenopausal osteoporosis . Alendronate is incorporated into bone matrix and, when ingested by osteoclasts, inhibits the cholesterol biosynthesis pathway enzyme, farnesyl diphosphate synthase, resulting in impaired geranylgeranylation (attachment of the lipid to regulatory proteins) and osteoclast inactivation . Raloxifene binds to estrogen receptors and appears to act as an estrogen agonist in bone . Procollagen type I N-terminal propeptide decreases during treatment with both raloxifene and alendronate, and reaches its nadir after 6 to 12 months of treatment .
The aim of this study was to compare the distribution of PINP in women with postmenopausal osteoporosis after 12 or more months of therapy with either raloxifene or alendronate.
Studies, patients, and treatments
Included were Eli Lilly and Company clinical trials of patients treated with alendronate or raloxifene in which intact PINP was evaluated. In addition, a group of premenopausal women for the calculation of a PINP reference interval was included. Serum intact PINP was measured by radioimmunoassay (Orion Diagnostica, Espoo, Finland) from serum samples collected from patients in these studies. All studies were reviewed and approved by the participating sites' ethical review boards prior to enrolling any patient. The studies were conducted in accordance with the 1964 Declaration of Helsinki and in keeping with Good Clinical Practice. All participating patients provided written informed consent prior to undergoing any study procedure or receiving any study therapy.
Study 1 Premenopausal
Sixty-eight premenopausal, non-pregnant women aged 31 to 55 years were previously recruited and a sample of serum taken for the assessment of PINP concentration [16, 17]. These women were not taking calcium, vitamin D, or any medications known to affect bone turnover. Reference intervals were calculated after log10 transformation of data and obtained from the mean ± 1.96 multiplied by the standard deviation (SD). The reference interval was 16.0 to 75.8 μg/L in premenopausal women, with a geometric mean of 34.9 μg/L.
Study 2 MORE
Details of the Multiple Outcomes of Raloxifene Evaluations (MORE) trial have been previously published . Briefly, this phase III study included a total of 7,705 women aged 31 to 80 years, from 25 countries, who had been postmenopausal for ≥2 years and who met World Health Organization criteria for having osteoporosis. Participants were randomly assigned to receive either raloxifene 60 mg/day, raloxifene 120 mg/day, or placebo. All participating women received supplemental calcium and cholecalciferol. Procollagen type I N-terminal propeptide was assessed at baseline and at 12 months in subsets of patients receiving placebo (n = 365), raloxifene 60 mg/day (n = 347), and raloxifene 120 mg/day (n = 254) . For this analysis, data from the placebo and raloxifene 60 mg/day (the approved dose) groups at baseline and at 12 months and the raloxifene 120 mg/day group at baseline were included.
Study 3 FACT
Detailed methods of the Forteo-Alendronate Comparator Trial (FACT) have been previously published . This 18-month randomized, parallel, double-blind phase IV study included postmenopausal women aged 45 to 85 years who were postmenopausal for ≥5 years, with hip or spine T-score between −2.5 and −4.0, inclusive. Patients were randomly assigned to once daily doses of teriparatide 20 μg or alendronate sodium 10 mg. Markers of bone turnover were assessed in all women, including 100 who were randomly assigned to alendronate and 100 to teriparatide. Procollagen type I N-terminal propeptide was assessed at about the same time of day at baseline and at 1, 3, 6, and 12 months. For this analysis, data from the alendronate group at all time points were included. We also included the baseline data from the teriparatide group to provide further PINP data from untreated women.
Study 4 AAA
Detailed methods of the Anabolic After Antiresorptive (AAA) study have been published . This phase III study included 59 postmenopausal women aged 60 to 87 years, with BMD T-scores ≤−2.5 prior to study entry and ≤−2.0 at study enrollment. Women in this study had previously received either alendronate or raloxifene therapy for at least 18 months. Only baseline samples (i.e., after 18 months of either alendronate or raloxifene) obtained from fasting patients were included in this analysis. Patients were then treated with daily subcutaneous injections of teriparatide 20 μg for 18 months.
Study 5 GHAH
Details of this study have also been published . Briefly, this was a phase III global, multicenter, double-blind, parallel randomized trial of 146 postmenopausal women with osteoporosis. Women were treated with either teriparatide 20 μg/day or alendronate sodium 10 mg oral capsule per day. All women received once daily oral supplementation with calcium (1,000 mg) and vitamin D (400 to 1,200 IU). Serum samples were collected and stored at −20°C at Covance Central Laboratories and later shipped to SUPREME SA (Liege, Belgium) for testing. All PINP samples were processed as one batch. PINP results from this trial were available at baseline and after 12 months of study drug. For this analysis, data from the alendronate group at both time points and data from the teriparatide group at baseline only were included.
The reference interval was calculated by exponentiating the interval obtained through mean ± 1.96 SD on a log10 scale using the data from Study 1 Premenopausal. Jitter plots were constructed to show PINP concentrations in individual patients relative to the lower and upper limits of the premenopausal reference interval and the geometric mean at baseline in all groups and post-baseline for patients treated with placebo, alendronate 10 mg/day, and raloxifene 60 mg/day. Density estimation  is the construction of the underlying probability distribution of observed data. Probability density curves offer a convenient way to present observed data. A Kernel Density Estimation (KDE) non-parametric method was used because it does not require rigid assumptions about the underlying distribution and allows a more flexible fit to the data. KDE plots for PINP concentration were constructed for Study 1 Premenopausal, at baseline and after 12 months of raloxifene or alendronate treatment for Studies 2 MORE, 3 FACT, and 5 GHAH, and at baseline from study 4 AAA (at which time patients had taken raloxifene or alendronate for ≥18 months). We drew the density curves for PINP concentration on both raw and log10 scales. The Shapiro–Wilk test was used to test for normality. We used the reference interval and geometric mean of PINP to categorize patients into low, lower half of the premenopausal reference interval, upper half of normal, and high concentrations for PINP. Analyses were performed using SAS version 9.1.3 (SAS Institute, Cary, NC, USA).
Study 3 FACT showed that during alendronate treatment, a shift to within the lower half of the premenopausal reference interval was evident after 1 month, and some patients had bone turnover below the lower limit of the premenopausal reference interval after 3 months. After 6 to 12 months of alendronate therapy, more than 50% of patients had bone turnover below the lower limit of the premenopausal reference interval.
Study 4 AAA showed a difference in distributions between patients treated for ≥18 months with either alendronate or raloxifene. Most patients in the alendronate group were below the lower limit of the premenopausal reference interval, and in contrast most patients in the raloxifene group were distributed throughout the premenopausal reference interval.
In Study 5 GHAH, most patients were in the upper half of the premenopausal reference interval at baseline; after 12 months of alendronate, most patients were below the lower limit of the premenopausal reference interval.
The number (percent) of postmenopausal women with osteoporosis at baseline in each bone turnover category defined by the lower limit, geometric mean, and upper limit of the premenopausal reference interval
Study 2 MORE
Study 3 FACT
Study 5 GHAH
The number (percent) of patients after at least 12 months of treatment with raloxifene or alendronate in each bone turnover category defined by the lower limit, geometric mean, and upper limit of the premenopausal reference interval
Study 1 Premenopausal
Study 2 MORE
Study 4 AAA
Study 3 FACT
Study 4 AAA
Study 5 GHAH
During raloxifene treatment, 3% of patients had PINP concentrations below the lower limit of the premenopausal reference interval when treated for ≥12 months, 37% to 60% had PINP in the lower half of the premenopausal reference interval, and 33% to 57% were in the upper half of the premenopausal reference interval. Only 3% to 4% of patients had PINP concentrations above the upper limit of the premenopausal reference interval. Of the 374 total patients treated with ≥12 months of raloxifene, 216 patients (58%) had PINP concentrations in the lower half of the premenopausal reference interval.
In this analysis we report that, prior to treatment, most postmenopausal women with osteoporosis had PINP concentrations in the upper half of the premenopausal reference interval. After treatment with alendronate, most women had PINP concentrations below the lower limit of the premenopausal reference interval; after treatment with raloxifene, most women had PINP concentrations within the lower half of the premenopausal reference interval (weighing the evidence from the 344 raloxifene-treated women in Study 2 MORE higher than the evidence from the 30 raloxifene-pretreated women in Study 4 AAA).
There are limited previous data regarding PINP as a measure for response to alendronate or raloxifene therapy. However, the findings from this study confirm and extend those of two previous studies that included markers of bone turnover as secondary endpoints [24, 25]. In one study , women with postmenopausal osteoporosis, who were treated with alendronate 10 mg/day for a mean of 43 months, were randomly assigned to double-blind raloxifene or placebo, or to continuation of open-label alendronate. Patients who continued on alendronate did not have an increase in their serum PINP, while those who switched to raloxifene had a significant increase in serum PINP to within the lower half of the premenopausal reference interval, and those who switched to placebo had a significant increase in serum PINP to the upper half of the premenopausal reference interval. The percentage of patients whose maximum post-baseline serum PINP value was below the lower limit of the premenopausal reference interval was significantly greater in the alendronate group compared with the raloxifene or placebo groups. In addition, the percentage of patients whose maximum PINP was above the upper limit of the premenopausal reference interval was higher in the placebo group compared with the raloxifene or alendronate group.
In another study , postmenopausal women with a BMD T-score <−2 were randomly assigned to alendronate, raloxifene, alendronate plus raloxifene, or placebo. While PINP was not assessed in this study, women who received alendronate had lower concentrations of other markers of bone turnover than those who received raloxifene. The percentage of women with after-treatment urinary C-telopeptide corrected for creatinine (CTX/Cr) below the lower limit of the premenopausal reference interval (52.02 μg/mmol) was significantly greater with alendronate (52.6%) compared with placebo (3.8%) or raloxifene (13.9%), and was similar to that observed in the combination (56.2%) group. However, only one woman in the combination therapy group had a post-baseline urinary N-telopeptide corrected for creatinine (NTX/Cr) value below the lower limit of the premenopausal reference interval .
Because premenopausal women are thought to have long-term skeletal health, a possible goal of anti-resorptive therapy is to reduce bone turnover markers to the lower half of the reference interval for healthy young premenopausal women. However, in postmenopausal women with osteoporosis treated with raloxifene, this level of bone turnover has been demonstrated to reduce the risk for vertebral but not for non-vertebral fractures . Bone turnover below the lower limit of the premenopausal reference interval may be associated with lower risk of fracture in 3- to 4-year fracture studies . In a meta-analysis of studies that examined the relationship between change in bone turnover and incidence of non-vertebral fracture, Hochberg et al. demonstrated a reduced risk of non-vertebral fracture in those studies with a decrease in bone turnover markers >50% . However, there have been reports which suggest that long-term suppression of bone turnover may be associated with an increased risk of fracture, osteonecrosis of the jaw and atypical femoral fractures [28, 29].
It is well known that the processes of bone resorption and formation are closely linked; a process known as “coupling”. Cells of the osteoblast lineage regulate the activity of osteoclasts by the release of various local signaling factors . Osteoclast activity in turn stimulates the differentiation and activity of osteoblasts. Therapies which reduce bone resorption (osteoclast activity) thus also indirectly reduce bone formation. Bone formation markers, such as PINP, are on the whole less variable than resorption markers  and are therefore potentially useful candidates for the measurement of treatment response in patients treated with anti-resorptive drugs.
We have shown that the majority of patients receiving alendronate therapy have bone turnover below the lower limit of the premenopausal reference range. In this analysis, we were not able to assess the full clinical significance of these findings. Further prospective studies that examine the change in systemic bone turnover markers in comparison with histological changes in bone and also to fracture incidence would be useful additions to the literature.
A strength of this study is that findings were consistent across a number of trials. Additionally, we recently published a reference interval for intact PINP from another group of healthy premenopausal women, which was similar to the one we used in this study; in that study, the reference interval calculated from 145 healthy premenopausal women aged 35 to 45 years was 17.4–61.6 μg/L (geometric mean = 32.7) .
The studies included in this analysis were small in some cases or else included markers in only subsets of patients, precluding a direct linkage of the marker findings to fracture outcomes. Only one marker was measured consistently in all the studies, and so the studies are limited to the consideration of one bone turnover marker. In addition, the studies were not planned to compare the distributions of markers by treatment group and as such the samples were not all run at the same time in the same laboratory, as would be good practice.
We conclude that raloxifene and alendronate result in different levels of suppression of bone turnover. Raloxifene decreases PINP into the lower half of the premenopausal reference interval, while alendronate decreases PINP more, with about 60% of treated women below the premenopausal reference interval. Procollagen type I N-terminal propeptide is a stable and sensitive marker of bone turnover which may be used to differentiate between the effects of different types of anti-resorptive therapy.
We acknowledge the assistance of statistician Richard Jacques, the School of Health and Related Research, University of Sheffield, in the preparation of this manuscript. This manuscript was reviewed by the editorial board of the NIHR Sheffield Bone Biomedical Research Unit. The views expressed in this publication are those of the authors and not necessarily those of the NHS, the NIHR, or the Department of Health. The authors would like to acknowledge Dr. Jody Arsenault, Lilly USA, LLC, for assistance with editing and formatting.
Conflicts of interest
Dr. Eastell has received consulting or advisory board fees from Amgen, Novartis, Procter & Gamble, Servier, Ono, and GlaxoSmithKline (less than $10,000 each), lecture fees from Eli Lilly (less than $10,000), and grant support from AstraZeneca, Procter & Gamble, and Novartis (less than $10,000 each). Dr. Rogers has no conflicts of interest to report. Drs. Ni and Krege are employees of Lilly USA, LLC and own stock or stock options in Eli Lilly. The authors had full control of all primary data and agree to allow the journal to review their data if requested.