The effect of prior bisphosphonate therapy on the subsequent BMD and bone turnover response to strontium ranelate

Abstract

Strontium ranelate is an effective treatment for osteoporosis in treatment-naive women. In the United Kingdom, bisphosphonates are often used first line. Prior bisphosphonate use may blunt the bone mineral density (BMD) response to strontium ranelate by reducing strontium uptake into the bone. Sixty bisphosphonate-naive women and 60 women discontinuing bisphosphonates were recruited. All women commenced strontium ranelate and calcium/vitamin D. BMD and bone turnover markers were recorded for 12 months. After 12 months, the bisphosphonate-naive group's BMD increased by 5.6% (p < .001) at the spine, 3.4% (p < .001) at the total hip, and 4.0% (p < .001) at the heel. By comparison, the prior bisphosphonate group had a 2.1% (p = .002) increase at the spine but no change at the hip or heel. At all time points, BMD was significantly greater in the bisphosphonate-naive group. In the prior bisphosphonate group, there was no significant change in BMD during the first 6 months at the spine, but between months 6 and 12 there was a parallel gain in BMD (0.027 versus 0.020 g/cm², p = .40). The baseline difference in bone markers was no longer significant by 3 months for bone-specific alkaline phosphatase (BSAP) and 6 months for procollagen type 1 amino-terminal propeptide (P1NP) and carboxy-terminal cross-linking telopeptide of type I collagen (CTX). More women in the prior bisphosphonate group suffered a vertebral fracture (2 versus 8 women, p = .047). After bisphosphonates, bone turnover remains suppressed for up to 6 months, with blunting of the BMD response to strontium ranelate during this time. After 6 months, BMD increases in the spine but not at the hip or heel. © 2010 American Society for Bone and Mineral Research.

Introduction

Strontium ranelate is an effective treatment for osteoporosis that reduces the incidence of both vertebral and nonvertebral fractures.1, 2 After absorption from the gastrointestinal tract, strontium, the active component, is incorporated into bone, resulting in an increase in bone formation and a reduction in bone resorption.3, 4 This is thought to lead to a positive remodeling balance at the level of the basic multicellular unit and an overall gain in bone tissue with each remodeling cycle. Studies using micro-computed tomography (µCT) demonstrate that compared with placebo treatment, women treated with strontium ranelate have an increased trabecular number, an improved trabecular structural model index, a reduction in trabecular separation, and an increase in cortical thickness.4 Compression testing confirms that strontium ranelate therapy leads to an improvement in the mechanical properties of bone and an overall increase in bone strength.5

To date, all clinical studies assessing the effects of strontium ranelate have involved treatment-naive women. Strontium ranelate therefore is licensed as a first-line treatment for osteoporosis, although in clinical practice many women are not treatment-naive and are already receiving bisphosphonate therapy. Furthermore, for newly diagnosed women, bisphosphonates are recommended as the first-line therapy for osteoporosis in the United Kingdom unless contraindicated.6 Women on bisphosphonates may be switched over to strontium ranelate if they develop adverse effects from treatment, such as esophagitis, or have a poor treatment response.

Bisphosphonates are potent suppressors of bone turnover, and their actions persist long after discontinuation of therapy.7, 8 The continued inhibition of bone turnover, leading to reduced new bone formation, even after the discontinuation of bisphosphonates provides two theoretical reasons why prior bisphosphonate therapy may inhibit the subsequent bone mineral density (BMD) response to strontium ranelate. First, since strontium is deposited predominantly in newly formed bone,9, 10 prior bisphosphonate exposure may inhibit the incorporation of strontium into the hydroxyapatite crystals. Second, alendronate has been reported to blunt the anabolic properties of teriparatide,11, 12 and if strontium ranelate has anabolic effects, then prior bisphosphonate exposure also may lead to a similar blunting of these bone-forming properties.

The inhibition of strontium uptake into the bone, leading to less X-ray attenuation and/or reduced bone formation, would be expected to result in a reduction in the BMD response to strontium ranelate. This study aims to investigate the effects of prior bisphosphonate exposure on the treatment response to strontium ranelate.

Methods

Subjects

We prospectively recruited women attending for either an outpatient appointment or bone densitometry assessment at the Centre for Metabolic Bone Disease in Hull (UK). Two groups of women were recruited: bisphosphonate-naive women and women treated with an oral bisphosphonate for more than 1 year and who had stopped treatment within the last month owing to an inadequate response or an adverse side effect. An inadequate response to oral bisphosphonates was defined as a reduction in spine BMD greater than the least significant detectable change of our bone densitometers (2.7%) despite the patient reporting compliance with treatment. All women were aged 50 to 80 years and had either a T-score of less than –2.5 at the hip/spine or a T-score of less than –2.0 at either site and one other risk factor for fracture (e.g., previous osteoporotic fracture, maternal hip fracture, previous steroid use, or body mass index < 19). Women were excluded if they had had prior treatment with strontium ranelate or teriparatide, were unable to give informed consent, had impaired mobility resulting in difficulty undergoing dual-energy X-ray absorptiometry (DXA), or had a lumbar spine that could not be evaluated by DXA. Women also were excluded if they had current or likely future steroid use or medical conditions associated with bone loss, including renal disease (creatinine clearance < 30 mL/min), active malignancy, osteomalacia, hyperparathyroidism, and malabsorption syndromes.

Eligible women were enrolled after providing written informed consent. At their first study visit, the women underwent a full medical history and physical examination. BMD was measured at the spine (L2–4) and hip (total hip) by DXA (Lunar Prodigy, GE Lunar, Madison, WI, USA). Heel (right os calcis) BMD also was measured (Lunar PIXI, GE Lunar, Madison, WI, USA). Blood was collected for bone turnover markers between 9 and 11 a.m. after an overnight fast and was transported to the laboratory within an hour for separation and freezing. The bone turnover markers assessed were procollagen type 1 amino terminal propeptide (P1NP; Elecsys 2010, Roche Diagnostics, Indianapolis, IN, USA), bone-specific alkaline phosphatase (BSAP) (Metra BAP, Quidel Corp., San Diego, CA, USA), and carboxy-terminal cross-linking telopeptide of type I collagen (CTX) (β-Crosslaps, Elecsys 2010, Roche Diagnostics).

Intervention and follow-up

All subjects received treatment with strontium ranelate 2 g once a day at bedtime (2 hours after food) and 1.2 g calcium and 800 IU vitamin D daily (Adcal D₃). The women were followed up for 1 year with visits at 3, 6, and 12 months between 9 and 11 a.m. At each visit, details regarding compliance (based on returned medication), adverse side effects, concomitant medication, and incident fractures were recorded. All women fasted overnight prior to each visit, and blood for bone turnover markers was collected at the same time each visit. Axial and heel DXA was repeated at the 6 and 12 month visits. A DXA-based vertebral fracture assessment (VFA) was performed with the patient in the left lateral position at baseline and the 12 month visit to identify incident vertebral fractures.

The primary endpoint was change in axial BMD after 12 months. The average lumbar spine BMD was used for analysis, but if there was a prevalent fracture at baseline or an incident fracture during the study in one of these vertebrae, then the fractured vertebra was excluded from the analysis. At the hip, total hip BMD was used for the analysis because this region of interest demonstrates the greatest increase in hip BMD in response to strontium ranelate1, 2 and is the recommended region of interest for assessing treatment response at the hip.13 The secondary endpoints were change in heel BMD, change in bone turnover markers (i.e., P1NP, CTX, and BSAP), and fracture incidence.

Data analysis

The sample size was determined using data from previous studies of strontium ranelate that demonstrated a mean annual increase in lumbar spine BMD of 7.3% (0.0512 g/cm²) with a standard deviation of 4.9% (0.0343 g/cm²).14 Allowing for a 10% withdrawal rate, it was calculated that 120 women were needed to detect a 30% difference in BMD gain between the two groups with a power of 90%.

The study was analyzed on a per-protocol basis rather than as an intention-to-treat analysis. This was done so that the results would accurately reflect the changes in BMD in women who successfully switch from a bisphosphonate to strontium ranelate. To check the validity of using a per-protocol analysis, an intention-to-treat analysis also was performed that included all women who consented for the study, with the last recorded value carried forward to replace missing data.

Initially, the Kolmogorov-Smirnov test for normality was used to assess the distribution of the data. Baseline characteristics then were analyzed using either a two-sample t test or Mann-Whitney U test depending on the distribution of the data. A Fisher exact test was used for categorical data. The absolute change in BMD at the spine (L2–4), hip (total hip), and heel (right os calcis) after 6 and 12 months of therapy was compared between the groups using a multivariate ANOVA. Within each group, a repeated-measures ANOVA was used to assess the change in BMD from baseline. For bone turnover markers, a multivariate ANOVA was used to examine the difference between the two groups at each visit of the four visits. A repeated-measures ANOVA was used to assess the overall change from baseline of each bone marker during the course of the study. The significance level chosen was .05. The program package used was SPSS for Windows (Version 14.0, SPSS, Inc., Chicago, IL, USA).

Study approval and funding

Ethical approval was obtained from the Hull and East Riding Local Research Ethics Committee. Clinical trial authorization was obtained from the Medicines and Healthcare Products Regulatory Agency, UK (EudraCT Number 2005-003138-16). Servier Laboratories provided the strontium ranelate and a grant to fund the study. ProStrakan provided Adcal D₃ as the calcium supplement. The study design; the data collection, analysis, and interpretation; and manuscript preparation all were carried out by the authors independent of Servier. All authors had full access to the data and were involved in manuscript preparation.

Results

Subjects and baseline demographics

In total, 120 Caucasian women were recruited into the study: 60 women who were currently taking a bisphosphonate (prior bisphosphonate group) and 60 who had no prior bisphosphonate use (bisphosphonate-naive group). Prior to the first follow-up visit at 3 months, eight women discontinued from the prior bisphosphonate group and four discontinued in the bisphosphonate-naive group. These women had no outcome data, leaving 108 women who made up the study population (52 and 56 women in each group, respectively). A further three women withdrew between the 6 and 12 month visits, so overall, 105 of the 108 women in the study population completed the full year.

The prior bisphosphonate group was older (66.9 versus 62.5 years, p = .001) and had a lower baseline BMD at the spine (0.801 versus 0.836 g/cm², p = .03) than the bisphosphonate-naive group. BMD was similar between the groups for total hip and heel BMD. In the prior bisphosphonate group, bone turnover markers were significantly lower than in the bisphosphonate-naive group, consistent with recent antiresorptive therapy. There were no other significant differences between the groups at baseline. Full baseline demographics are demonstrated in Table 1.

Table 1. Baseline Characteristics

	Bisphosphonate-naive (n = 56)	Prior bisphosphonate (n = 52)	p Value
a Significant difference between groups. Numbers are n (%) or mean (SD). VFA = vertebral fracture assessment.
Age (years)	62.5 (6.8)	66.9 (6.8)	.001a
Menopause age (years)	47.3 (6.1)	46.8 (6.0)	.68
Positive family history (n)	15 (26.8%)	16 (30.8%)	.68
Prior steroid use (n)	5 (8.9%)	4 (7.7%)	1.00
Current smoking (n)	7 (12.5%)	4 (7.7%)	.53
Alcohol (U/week)	5.4 (9.0)	3.5 (4.7)	.63
BMI (kg/m^2)	24.9 (3.4)	24.3 (3.5)	.37
Vertebral fracture on VFA (n)	10 (17.9%)	15 (28.8%)	.25
Prior nonvertebral fracture (n)	26 (46.4%)	24 (46.2%)	1.00
Vitamin D (nmol/L)	72.9 (31.1)	71.9 (26.6)	.86
Parathyroid hormone (µg/L)	31.4 (10.4)	32.3 (11.4)	.66
BMD spine (g/cm^2)	0.836 (0.08)	0.801 (0.09)	.03a
BMD total hip (g/cm^2)	0.780 (0.11)	0.751 (0.11)	.18
BMD heel (g/cm^2)	0.391 (0.07)	0.369 (0.09)	.15
P1NP (µg/L)	54.42 (18.9)	29.64 (13.9)	<.001a
CTX (µg/L)	0.38 (0.14)	0.18 (0.11)	<.001a
BSAP (U/L)	23.16 (6.7)	17.52 (6.6)	<.001a

In the prior bisphosphonate group, the mean (SD) duration of bisphosphonate use was 64.3 (38.5) months. One woman discontinued her bisphosphonate 3 weeks before commencing strontium ranelate; all the remaining women in the prior bisphosphonate groups switched immediately from bisphosphonate to strontium ranelate. Details of prior bisphosphonate usage are contained in Table 2.

Table 2. Bisphosphonate Usage Before Switching to Strontium Ranelate in the Prior Bisphosphonate Group

	Immediately prior		Ever used
	n	%	n	%
Alendronate	27	51.9	39	75.0
Risedronate	24	46.2	29	55.8
Ibandronate	1	1.9	1	1.9
Didronel	0	0.0	15	28.8
Pamidronate	0	0.0	1	1.9

Over the 12 month period, the mean level of compliance with strontium ranelate was 95.6% in the bisphosphonate-naive group and 95.0% in the prior bisphosphonate group.

Change in spine BMD with strontium ranelate

After 6 months of therapy, BMD at the spine had increased by 0.020 g/cm² (2.4%, p = .001) in the bisphosphonate-naive group, whereas there was no change in BMD in the prior bisphosphonate group (–0.003 g/cm², p = .65). After 12 months, BMD at the spine had increased significantly by 0.047 g/cm² (5.6%, p < .001) in the bisphosphonate-naive group and by 0.017 g/cm² (2.1%, p = .002) in the prior bisphosphonate group. These changes are demonstrated in Fig. 1A. The increase in BMD was significantly greater in the bisphosphonate-naive group than in the prior bisphosphonate group at both 6 months (difference 0.023 g/cm², p = .005) and 12 months (difference 0.030 g/cm², p = .003). After adjusting for the baseline differences in age and BMD, the bisphosphonate-naive group maintained a greater gain in spine BMD at both 6 months (difference 0.028 g/cm², p = .002) and 12 months (difference 0.036 g/cm², p = .001). In contrast to the lack of change in BMD in the prior bisphosphonate group during the first 6 months of the study, between months 6 and 12 there was a similar gain in BMD at the spine in each group (0.027 versus 0.020 g/cm², p = .40).

Figure 1. Mean (± SE) change from baseline in BMD at the lumber spine (A), total hip (B), and heel (C) after 6 and 12 months of treatment with strontium ranelate in bisphosphonate-naive women (broken line) and in women with prior bisphosphonate exposure (solid line). a = significant increase from baseline.

Change in total hip BMD with strontium ranelate

In the bisphosphonate-naive group, total hip BMD had increased by 0.014 g/cm² (1.9%, p < .001) at 6 months and by 0.027 g/cm² (3.4%, p < .001) at 12 months. In the prior bisphosphonate group, there was no significant change in total hip BMD during the 12 months [0.006 g/cm² (0.8% increase), p = .096]. These changes are demonstrated in Fig. 1B. The increase in total hip BMD was significantly greater in the bisphosphonate-naive group at 6 months (difference 0.013 g/cm², p < .001) and 12 months (difference 0.021 g/cm², p < .001). After adjusting for the baseline differences in age and BMD, the difference in total hip BMD between the two groups remained significant at both 6 months (difference 0.014 g/cm², p < .001) and 12 months (difference 0.020 g/cm², p < .001).

Change in heel BMD with strontium ranelate

In the bisphosphonate-naive group, heel BMD had increased by 0.011 g/cm² (2.9%, p = .002) after 6 months and by 0.016 g/cm² (4.0%, p < .001) after 12 months. In the prior bisphosphonate group, there was no change in heel BMD over the 12 months [0.001 g/cm² (0.3% increase), p = .93]. These changes are demonstrated in Fig. 1C. The increase in BMD was significantly greater in the bisphosphonate-naive group at 6 months (difference 0.011 g/cm², p = .013) and 12 months (difference 0.015 g/cm², p = .012). After adjusting for the baseline differences in age and BMD, the bisphosphonate-naive group maintained a greater gain in heel BMD at both 6 months (difference 0.013 g/cm², p = .006) and 12 months (difference 0.015 g/cm², p = .010).

Change in bone turnover markers with strontium ranelate

At baseline, all bone markers were significantly lower in the prior bisphosphonate group, consistent with recent antiresorptive therapy. In the prior bisphosphonate group, 12 months after switching to strontium ranelate, there was a significant increase of 55.1%, 61.0%, and 46.3% in P1NP, CTX, and BSAP, respectively. There was no longer a significant difference in bone turnover markers between the two groups by 3 months for BSAP and 6 months for P1NP and CTX. However, with CTX, the difference between the two groups remained borderline at 12 months compared with P1NP and BSAP, where the difference was negligible. The change in bone markers at each visit is provided in Table 3.

Table 3. Changes in CTX, P1NP, and BSAP in Response to Strontium Ranelate in Women With and Without Prior Bisphosphonate Exposure

		Month				12 month change p value
		0	3	6	12
BP = bisphosphonate.
CTX (µg/L)	Prior BP	0.18	0.26	0.29	0.29	<.001
	BP-naive	0.38	0.32	0.34	0.35	.004
Between-group p value		<.001	.016	.069	.046
P1NP (µg/L)	Prior BP	29.64	38.00	41.54	45.98	<.001
	BP-naive	54.42	49.70	48.09	47.75	.011
Between-group p value		<.001	.001	.065	.603
BSAP (U/L)	Prior BP	17.52	21.06	23.86	25.64	<.001
	BP-naive	23.16	24.14	23.81	24.45	.498
Between-group p value		<.001	.055	.976	.465

In the bisphosphonate-naive group, after 12 months of treatment with strontium ranelate, there was a significant decrease of 8.1% in CTX and 12.2% in P1NP (Fig. 2). There was an increase of 5.6% in BSAP, although this was not significant.

Figure 2. Mean (± SE) percentage change from baseline in P1NP (), CTX (), and BSAP () in response to strontium ranelate in treatment-naive women.

Fracture incidence during therapy with strontium ranelate

During the year, there were significantly more women who suffered an incident vertebral fracture in the prior bisphosphonate group (2 versus 8 women, p = .047). During the study, one woman suffered a wrist fracture in the bisphosphonate-naive group, whereas four women reported nonvertebral fractures in the prior bisphosphonate group (two rib fractures, wrist, and humerus). All fractures were confirmed on X-ray or VFA, with the exception of rib fractures. Fracture incidence is summarized in Table 4.

Table 4. Fractures Occurring During 12 Months of Treatment With Strontium Ranelate

	Number of women suffering a fracture		p Value
	Bisphosphonate-naive	Prior bisphosphonate
Any fracture	3	12	0.014
Vertebral fractures	2	8	0.047
Nonvertebral fractures	1	4	0.194

Adverse events and subject withdrawal from the study

There were a total of six serious adverse events during the study (four prior bisphosphonate, two bisphosphonate-naive), none of which were felt to be likely to be related to the study medication or lead to withdrawal from the study. For analysis, adverse events were divided into gastrointestinal (GI; predominantly nausea, altered bowel habit, and bloating), central nervous system (CNS; predominantly headache and lethargy), musculoskeletal (arthralgia or leg cramps), and skin (itching or rashes). Only adverse events considered by the investigators to be probably or definitely related to the study medication were counted. Of the 12 women who withdrew from the study prior to visit 3, the reasons for withdrawal were GI (n = 4), CNS (n = 3), musculoskeletal (n = 3), skin (n = 1), and one woman requested to change back to weekly therapy with bisphosphonates. Of the 108 women in the study population, the numbers of reported adverse events were GI 26 (24.1%), CNS 3 (2.8%), musculoskeletal 1 (0.9%), and skin 1 (0.9%). Most of these (23/31, 74.1%) occurred in the first 3 months and settled without withdrawing medication. The three women who withdrew from the study population between months 6 and 12 did so because of lost contact (n = 1), disliked of the taste of strontium ranelate (n = 1), and dyspepsia (n = 1).

Intention-to-treat analysis

All 120 women who were recruited into the study were included in the intention-to-treat analysis. In the bisphosphonate-naive group, BMD increased significantly after 1 year at the spine (0.043 g/cm², 5.2%, p < .001), hip (0.025 g/cm², 3.2%, p < .001), and heel (0.015 g/cm², 3.8%, p < .001). In the prior bisphosphonate group, the BMD increased significantly after 1 year at the spine (0.015 g/cm², 1.8%, p = .002) but not the hip (0.005 g/cm², 0.7%, p = .10) or heel (0.001 g/cm², 0.3%, p = .93). The increase in BMD was significantly greater in the bisphosphonate-naive group at all sites. After 1 year, P1NP, CTX, and BSAP had increased significantly in the prior bisphosphonate group by 47.5%, 52.0%, and 39.3%, respectively. In the bisphosphonate-naive group after 1 year, there was a significant reduction in P1NP (–11.6%, p = .011) and CTX (–7.5%, p = .004), whereas the increase in BSAP was not significant (+5.2%, p = .50).

Discussion

This is the first study to investigate the BMD response to switching osteoporosis therapy from a bisphosphonate to strontium ranelate. In this study, the bisphosphonate-naive group achieved increases in BMD at the spine and hip that are comparable to those seen in the phase 3 SOTI study.1 However, the prior bisphosphonate group had significant blunting of the BMD response to strontium ranelate at all three sites studied. More than 50% of the BMD response to strontium ranelate is thought to be due to the attenuation artefact caused by strontium's high atomic mass.2, 15 Therefore, a large proportion of the blunting of the BMD response to strontium is likely to reflect poor strontium uptake into the bone. This is likely to occur because the bisphosphonate-induced suppression of bone turnover reduces the formation of new bone, which is the site at which strontium is predominantly deposited.9, 10 Also, if strontium ranelate does have anabolic properties, then part of the blunting of the BMD response may be due to inhibition of the bone-forming effects of strontium ranelate, leading to smaller gains in bone mass similar to the observed effects of prior bisphosphonate therapy on teriparatide.11, 12 Reassuringly, the bone turnover markers all increased significantly after bisphosphonate discontinuation, reflecting increased bone turnover, and by 6 months, all bone markers were similar in the two groups. Since it is the suppressed bone turnover that is likely to cause the blunting of the BMD response, then the normalization of the bone markers suggests that the blunting of the response to strontium is likely to be temporary. This is supported by the observation that the increase in BMD achieved at the spine during the second 6 months of the study was the same in both groups.

In contrast to the spine, women with prior bisphosphonate exposure had no increase in total hip or heel BMD over the 12 months of the study. Therefore, the blunting of the BMD response appears to be more persistent at the heel and hip than at the spine, and this study has been extended into a second year to investigate this further. The hip is predominantly cortical bone, which is less metabolically active than trabecular bone.16 As such, bone turnover may take longer to recover after bisphosphonate therapy, resulting in a more prolonged blunting of the BMD response to strontium ranelate. Furthermore, the preferential incorporation of strontium into new bone, compared with old bone, is greater in cortical than in trabecular bone. This has been demonstrated in monkeys, with the strontium content in new cortical bone being three to four times greater than that in old cortical bone, whereas new trabecular bone contained only 2.5 times more strontium.17 Therefore, reduced new bone formation may cause a greater impedance of strontium uptake at cortical sites than trabecular sites.

The heel, like the spine, is predominantly comprised of trabecular bone. In contrast to the spine, in women with prior bisphosphonate exposure, there was no increase in heel BMD throughout the whole year. The more persistent bisphosphonate-induced blunting of the BMD response at the heel may reflect the fact that the heel is a site of yellow (fatty) bone marrow, whereas the spine has greater red bone marrow content.18 Red bone marrow is the source of both osteoclasts and osteoblasts, as well as a variety of cytokines, including macrophage colony-stimulating factor and receptor activator of nuclear factor κB ligand, which influence osteoclast differentiation.19 Therefore, the relative lack of these cells and cytokines in the yellow marrow of the heel may lead to more prolonged suppression of bone turnover after bisphosphonate discontinuation, which may account for the longer-lasting blunting of the BMD response to strontium ranelate.

The clinical implications of this blunting of the BMD response by prior bisphosphonate therapy are uncertain because it is difficult to assess whether this has any effect on the ability of strontium ranelate to reduce fracture incidence. Theoretically, if strontium uptake into the skeleton is reduced, then it is plausible that this will reduce the effect of strontium on bone strength. However, it is also likely that prior bisphosphonate therapy will have a period of residual effect on fracture risk7 that may provide protection from fractures while the strontium has time to overcome the blunting effect. While there were significantly more fractures in the women with prior bisphosphonate exposure, it is important to remember that this was not a randomized trial. The prior bisphosphonate women were older and had a lower spine BMD at baseline, suggesting that they had a greater risk of fracture, which may account for the observed differences in fracture incidence.

The blunting of the BMD response to strontium ranelate may have other implications for clinical practice. First, in treatment-naive women, strontium ranelate causes a large increase in BMD, making it possible to detect a treatment response as early as 1 year after the initiation of therapy. However, the 2.1% increase in BMD at 1 year observed in the prior bisphosphonate group is less than the most optimistic estimate of least significant change for spine BMD.20 Therefore, in clinical practice, it may be necessary to allow a greater time period before performing a follow-up DXA scan to assess a woman's treatment response to strontium ranelate if there is a history of recent bisphosphonate use. Second, if bisphosphonates inhibit the response to strontium ranelate, then it may be appropriate to consider strontium ranelate as a first-line treatment, especially in patients likely to tolerate strontium ranelate better than bisphosphonates, such as women with coexisting gastrointestinal disease or other medication associated with dyspepsia and women over age 80 in whom strontium has very good evidence for fracture reduction.21 Conversely, if a woman is intolerant of generic alendronate, the recommended first-line bisphosphonate in the UK,6 after more than 12 months on treatment, then it may be prudent to try a better-tolerated bisphosphonate or an intravenous bisphosphonate before switching to strontium ranelate. Whether bisphosphonate exposure for less than 12 months results in blunting of the BMD response to strontium ranelate is uncertain because such women were excluded from our study.

This study also uses bone markers to provide insight into the changes in bone turnover in each group of patients. CTX and BSAP were measured because these markers allow comparison of our results with the SOTI study.1 P1NP was measured because it is potentially a better marker of bone formation owing to the fact that it is derived directly from type I collagen synthesis and has the lowest degree of analytic and biologic variability.22 Unsurprisingly, in the prior bisphosphonate group, all three bone turnover markers were significantly suppressed at baseline and increased progressively throughout the study. This is consistent with the increase in bone turnover that has been observed after the withdrawal of bisphosphonate therapy.7, 8, 23 In the bisphosphonate-naive group, the change in CTX and BSAP were comparable to the findings of the SOTI study, which reported an increase in BSAP and a reduction in CTX.1 In our study, similar divergent changes in these markers were observed. The 5.6% increase in BSAP was not significant in our study but is of a similar magnitude to the 8.1% increase reported in the SOTI study. The reduction observed in CTX also was somewhat smaller in our study (8.1% versus 12.2%) but was still significant.1 Using these markers, our study supports the theory that strontium ranelate is a “dual-action bone agonist.”

The change in P1NP in the bisphosphonate-naive group is unexpected and interesting. Like BSAP, it is a marker of bone formation, but in contrast to BSAP, P1NP reduced progressively throughout the study, with a significant 12.2% reduction after 12 months. Previous studies with antiresorptives and anabolic agents demonstrate that changes in BSAP and P1NP usually mirror each other.24, 25 BSAP is produced by osteoblasts and is a measure of osteoblast activity and number.22 The increase in serum BSAP is consistent with reports that strontium ranelate increases osteoblast proliferation and differentiation and increases osteoblast expression of alkaline phosphatase.26 P1NP is a direct quantitative measure of type I collagen synthesis.22 The reduction in P1NP in the face of increased BSAP suggests that collagen synthesis is not actually increased despite the increased osteoblast activity. In fact, the P1NP response is similar to the CTX response. This would be more consistent with an overall mild antiresorptive effect leading to a reduction in bone turnover with reduced collagen breakdown and synthesis. This is somewhat at odds with the histomorphometric data, which suggests that there is an increase in bone-formation parameters as well as evidence of increased osteoblast proliferation and differentiation.3, 4, 27 Furthermore, in vitro studies suggest that strontium increases collagen synthesis,28 which would be expected to increase P1NP. Further studies of the effect of strontium ranelate on P1NP are required to confirm our findings, and ultimately, paired human bone biopsies are required to accurately determine the effect of strontium ranelate therapy on bone microarchitecture.

There are limitations to this study. This was not a randomized study because women were allocated into one of the two groups according to their prior bisphosphonate use. The lack of randomization did result in baseline differences in age and BMD. However, adjustment for differences at baseline had no effect on the results, and otherwise the two groups were well matched. Another limitation was that different bisphosphonates, predominantly alendronate and risedronate, were used immediately prior to enrollment in the study. Alendronate suppresses bone turnover for several years after discontinuation,7, 8 whereas with risedronate bone turnover normalizes within 1 year.23 This difference in offset time may mean that women with prior alendronate use may experience greater blunting than those with risedronate. Since this study assesses the first year after discontinuation, the effects of this should be minimized because even bisphosphonates with a rapid offset of action are still likely to cause blunting for a large proportion of the first year. Furthermore, data with teriparatide suggest that the BMD response to teriparatide is the same after risedronate and alendronate.29 Finally, for ethical reasons, most of the women in the prior bisphosphonate group were switched to strontium ranelate on the basis of a poor clinical response to therapy. As such, there may have been selection bias leading to the recruitment of women who, for some reason, are more resistant to treatment. While this cannot be ruled out, the baseline bone markers do demonstrate that the bisphosphonates were successfully suppressing bone turnover prior to the study, suggesting a therapeutic effect and compliance with treatment. Also against this is the observed increase in spine BMD during the second 6 months of the study. Furthermore, causes for a poor treatment response, such as malabsorption and conditions affecting bone metabolism, were excluded at baseline, and the compliance with strontium ranelate was the same in each group.

In conclusion, this study demonstrates that after discontinuation of bisphosphonates and switching to strontium ranelate, there remains a significant suppression of bone turnover for 3 to 6 months. This is associated with a blunting of the BMD response to strontium ranelate for 6 months at the spine and for longer at the hip and heel. The clinical implications of this in terms of fracture are uncertain, but it does imply that after switching from bisphosphonates to strontium ranelate, a greater time period should be allowed before performing a follow-up DXA to assess the treatment response.

Disclosures

Our center has received an educational grant from Servier Laboratories that funded the costs of this study. However, the study design, the data collection, analysis and interpretation, and the writing of the paper all were carried out by the authors independent of Servier. All authors had full access to the data and were involved in manuscript preparation. The strontium ranelate was provided by Servier Laboratories, and ProStrakan provided the calcium tablets (Adcal D3) for the study. Servier has paid speaker fees to ETM (<£1000) and MA (£<2000). All the authors state that they have no conflicts of interest.

Author contributions: Conception and design: ETM, SAS, and SMD; analysis and interpretation of data: ETM, MA, SAS, and SMD; drafting of the paper or critically revising it: ETM, MA, SAS, and SMD; approval of the final version submitted: ETM, MA, SAS, and SMD.