Absolute risk reduction to compare efficacy of antiosteoporotic treatments in the absence of head-to-head trials




Bernard CORTET, MD, PhD
Département Universitaire
de Rhumatologie
Université de Lille 2
Lille, FRANCE

Absolute risk reduction
to compare efficacy of
antiosteoporotic treatments
in the absence
of head-to-head trials

by B. Cortet ,France

The last 15 years have seen considerable development in the therapeutic arsenal available for the treatment of osteoporosis. While this is good news, it also implies that therapeutic choices prescribers are required to make are not always easy. Analyzing the major pivotal studies reported in the literature is of great help when faced with these choices. However, the level of proof is not the same for all available molecules, especially as regards prevention of nonvertebral fractures as a whole and hip fractures in particular. Nevertheless, there are several efficient treatments for the prevention of vertebral fractures, nonvertebral fractures as a whole, and hip fractures in particular. Where several possible treatments seem to be efficient, a choice needs to be made. Besides tolerance to antiosteoporotic treatments, which is generally satisfactory, and the practical aspects of administration, other tools need to be employed. Molecules cannot be compared directly since there are no randomized studies in which their efficacies are directly compared. Nor is it possible to consider the reduction in relative risk, since doing so would lead to an error in interpretation as the inclusion criteria for the various studies are not strictly identical. On the other hand, a review of the data from the major pivotal studies shows that the reduction in relative risk depends on the severity of osteoporosis. Thus, when that type of analysis was possible, it was shown that the reduction in relative risk was greater when osteoporosis was less severe. It would be appropriate, therefore, in these conditions, to take into account the improvement in absolute fracture risk (defined as the difference between fracture risk in placebo and treatment groups). This parameter varies widely from one molecule to the next, even though all of the drugs at our disposal have proven antifracture efficacy (for vertebral fractures, in any case), which justified their being granted full market approval. This parameter can also be used to calculate the number needed to treat to prevent a fracture event, defined as the inverse of the reduction in absolute fracture risk, an asset to practitioners for translating clinical trial results into benefits for patients.

Medicographia. 2010;32:41-47 (see French abstract on page 47)

Within the last 15 years, substantial progress has been made in the understanding and treatment of osteoporosis. With the considerable development of the therapeutic arsenal at their disposal, practitioners now find themselves having to choose from among the many different molecules available. The choices they are faced with are not easy, since all of the available molecules have been granted full market approval and have, by definition, demonstrat- ed their usefulness. One of the problems stems from the fact that, as far as osteoporosis is concerned, no head-to-head trials have ever been carried out to evaluate antifracture efficacy. In practice, the practitioner must make a choice, and this will depend on the antifracture efficacy of the molecule for all types of fractures, since this may differ from one molecule to the next. In intertrial comparisons of antifracture efficacy with respect to a placebo group, relative risk is generally taken into account. This approach is open to criticism in that the reduction in relative riskmay depend on the severity of the disease in the selected population. In other words, quite frequently, the reduction in relative risk is found to be greater in less severe cases of osteoporosis. More generally, the lack of strict similarity across trial populations is a serious criticism of this approach. Within the last few years, as observed in other domains and particularly cardiovascular pathology, absolute risk reduction has emerged as a factor to be taken into account. In some situations, for instance, it can serve as the basis for deciding which antiosteoporotic treatment to administer.

Given the broad scope of this topic, we will restrict our discussion to postmenopausal osteoporosis. In doing so, we can suppose that the diagnosis of osteoporosis is established. In other words, initial testing would have already been carried out to eliminate malignant and benign bone fragility diseases other than osteoporosis (osteomalacia, primary hyperparathyroidism). Similarly, in this paper, we will only address the drug treatment of osteoporosis, even though it is clear that nondrug aspects of therapy must be taken into account, regardless of the drug used. Likewise, we will not be considering vitamin D, which is widely used in the treatment of osteoporosis, but in more of a supplementary role (in association with another antiosteoporotic treatment) than as a treatment in its own right.

In the first part of this paper, we will explore the antifracture effects of various treatments, expressed in terms of reduction in relative fracture risk. In the second part, we will examine the concept of absolute risk, and the reduction of the latter during antiosteoporotic treatment. Lastly, we will consider the corollary of absolute fracture risk, ie, the number needed to treat (NNT) to prevent a fracture event. The antifracture efficacy of various antiosteoporotic treatments was the subject of a recent update,1 the main results of which are summarized in Table I.1

Antiosteoporotic drugs are generally classified into three groups, according to their mechanism of action: bone resorption inhibitors, bone anabolic agents, and uncoupling agents. Bone resorption inhibiting drugs include drugs used in menopausal hormone treatment, selective estrogen receptor modulators (SERMs), and bisphosphonates. We will not consider menopausal hormone replacement therapy (HRT) in this paper for two reasons: firstly, the WHI (Women’s Health Initiative) study has clearly demonstrated that its benefit/tolerance ratio is poor; and, secondly, given the wide range of therapeutic solutions available today, it is now very rare to prescribe HRT as part of the treatment of osteoporosis. Where bone-forming drugs are concerned, we will focus particular attention on teriparatide. Lastly, strontiumranelate is the only uncoupling agent at our disposal.

Bone resorption inhibitors

_ Selective estrogen receptor modulators (SERMs)
The only SERMcurrently available is raloxifene. Although basedoxifene and lazofoxifene have recently been granted market approval in Europe, they are not yet available to prescribers in the major European countries.

Raloxifene was examined in MORE (Multiple Outcomes of Raloxifene Evaluation).2 At the end of 3 years of treatment, a reduction in vertebral fracture risk was found in a population of osteoporotic women defined as such according to densitometric criteria and/or the presence of at least one vertebral fracture compared with a placebo group. However, fracture risk reduction varied according to the initial data. Thus, risk reduction was 55% in women without vertebral fractures at inclusion (relative risk [RR], 0.45; 95% confidence interval [CI], 0.29-0.71). On the other hand, in women who had at least one vertebral fracture at inclusion, risk reduction was lower (30%; RR, 0.70; 95% CI, 0.56-0.86). The study was extended for a further year, with the double-blind procedure being maintained. During the fourth year, a 50%reduction in fracture risk was observed in women who had no initial vertebral fractures, as opposed to a 38%reduction in women with a preva- lent vertebral fracture.3 However, efficacy was not demonstrated either in preventing nonvertebral fractures as a whole or hip fractures in particular. Through the MORE study, raloxifene was shown to be effective in preventing breast cancer, but its efficacy varied according to the type of cancer and duration of follow-up (–70%).4 It was found to be effective only in cancers in which estrogen receptors were present. The initial findings on the prevention of cardiovascular morbidity were not confirmed by the RUTH (Raloxifene Use for The Heart) study.5 On the other hand, unlike estrogens, raloxifene has not been demonstrated to have harmful cardiovascular effects.

Table I
Table I. European guidelines for the diagnosis and management of osteoporosis in postmenopausal women.

Abbreviations: HRT, hormone replacement therapy; PTH, parathyroid hormone. Adapted from reference 1: Kanis JA, Burlet N, Cooper C, et al. Osteoporos Int. 2008;19:399-428. Copyright ©
2008, International Osteoporosis Foundation and National Osteoporosis Foundation.

_ Bisphosphonates
Bisphosphonates—structural analogues of pyrophosphates— are powerful inhibitors of bone resorption. Etidronate will not be considered in this review, on account of its modest efficacy.

_ Alendronate
Alendronate was the first available bisphosphonate to demonstrate antifracture efficacy. Thus, in FIT-1 (Fracture Intervention Trial 1), which evaluated patients with at least one vertebral fracture at inclusion, a significant reduction in the risk of vertebral fractures (–47%), wrist fractures (–50%), and hip fractures (–51%) was observed.6 However, the efficacy of the treatment was not demonstrated when nonvertebral fractures were considered globally. In the FIT-2 study,7 alendronate was only found to be effective in preventing morphometric vertebral fractures. The study population, however, was not osteoporotic. On average, at inclusion, the women had osteopenia and, in most cases, without a prevalent fracture. In a subanalysis of the trial, it was demonstrated that when women with densitometric osteoporosis (T-score <–2.5) were considered, alendronate was also effective in preventing wrist and hip fractures. However, as was the case in the FIT-1 trial (and in the subpopulation of women with densitometric osteoporosis), alendronate was not found to be effective in preventing nonvertebral fractures as a whole. _ Risedronate
Risedronate has also proven its efficacy in preventing vertebral fractures. After being administered for 3 years, it was found to reduce vertebral fracture risk by between 41% and 49%, depending on the authors.8,9 It has also been shown to be effective in preventing nonvertebral fractures. However, findings concerning its efficacy vary according to different studies. Thus, the reduction in nonvertebral fracture risk was significant in one of the two pivotal studies (–36%).8 In the other pivotal study conducted in Europe, the reduction was not significant (–33%).9 In a specific study on the efficacy of risedronate in preventing hip fractures,10 the authors reported a global 30% reduction in fracture risk. However, the efficacy of the drug was most apparent in women aged 70 to 79 years old, in whom femoral neck bone density had substantially declined (T-score <–3, and presence of at least one hip-fracture risk factor). Under these conditions, a 40% reduction in hip fracture risk was observed after 3 years of treatment. A comparative meta-analysis of alendronate and risedronate was performed to evaluate their efficacy in preventing nonvertebral fractures, given the results observed in the pivotal studies.11 The meta-analysis showed that both molecules were effective: RR values for alendronate and risedronate were 0.86 (0.76-0.97) and 0.81 (0.71-0.92), respectively.

_ Ibandronate
Ibandronate has also proved to be effective in preventing vertebral fractures, with a 62% risk reduction after 3 years of treatment.12 Its efficacy in preventing nonvertebral fractures was not demonstrated in BONE (oral iBandronate Osteoporosis vertebral fracture trial in North America and Europe), which involved a population of women at risk of vertebral fracture. On the other hand, in a post hoc analysis of women with severe osteoporosis in the population (T-score <–3 or T-score <–2.5 plus a history of bone fragility fractures in the 5 years preceding their inclusion in the study), a 69% reduction in nonvertebral fracture risk was observed in the former case and a 60% reduction in the latter. _ Zoledronate
It was recently demonstrated that zoledronic acid could reduce vertebral fracture risk by 70% (95% CI, 62% to 76%) in patients with osteoporosis, defined either in densitometric terms or by the presence of at least one vertebral fracture. In the same study, the drug was found to be effective in preventing both nonvertebral fractures (–25%) (95% CI, 13% to 36%) and hip fractures (–41%) (95% CI, 17% to 58%).13 In a study involving a population of men and women with recent hip fractures, zoledronate was found to reduce the risk of clinical fractures by 35%(RR, 0.65; 95%CI, 0.50-0.84).14 In the same study, the drug was also shown to be effective in preventing clinical vertebral fractures (–46%) (95% CI, 8% to 28%). Lastly, in a secondary analysis of their data, the authors also found the treatment to be effective in reducing the death rate (–28%; RR, 0.72 [0.56-0.93]).

_ Optimal bisphosphonate treatment duration
This is a difficult question to answer insofar as the pivotal studies were conducted over a period of 3 years. A 5-year study was conducted on the efficacy of risedronate (an extension of the VERT MN [Vertebral Efficacy with Risedronate Therapy, Multinational] study for a further 2 years), but the study is methodologically subject to criticism. In that study,15 the authors showed that the efficacy of treatment: (i) remained unchanged during the 5th year when compared to the previous years, but (ii) declined significantly during the 5th year when compared to the results observed in the placebo group. Lastly, a follow-up study of the effect of alendronate on bone mineral density (BMD) was published in 2004,16 the maximum treatment duration of which was 10 years. In that study, group sizes were small, and, over such a long period of time, the double-blind procedure was not maintained. Nonetheless, the authors16 were able to demonstrate that the evolution in nonvertebral fracture risk between the 1st and the 3rd year and the 6th and the 10th years was identical, suggesting that efficacy did not decline over time. However, the methodology can be criticized, and it is difficult to draw definitive conclusions.

Bone formation stimulants

_ Parathyroid hormone
The leading drug in this category is teriparatide, which is the 1-34 fragment of parathyroid hormone. At a dose of 20 ìg/ day, teriparatide has been shown to be capable of reducing vertebral fracture risk by 65% (after 18 months of treatment, on average) when compared with the results observed in a placebo group. At the end of that period, an equally significant reduction in nonvertebral fracture risk (–53%) was also demonstrated. Moreover, its vertebral antifracture efficacy seemed to be more pronounced in women who had at least two vertebral fractures at inclusion, and it is for this reason that teriparatide is generally indicated for the treatment of the most severe osteoporosis. Parathyroid hormone (1-84) has also been evaluated.17 After 18 months of treatment, a 60% reduction in vertebral fracture risk was observed, but without a significant effect on nonvertebral fracture risk.

_ Strontium ranelate
Strontium ranelate has an original mode of action in that it stimulates bone formation while at the same time inhibiting bone resorption. This has been demonstrated in vitro as well as in vivo by measuring changes in bone-remodeling markers and analyzing bone biopsies. Strontium ranelate has been the subject of a vast development program comprising two pivotal studies: namely, the SOTI (Spinal Osteoporosis Therapeutic Intervention) study,18 which sought to evaluate the efficacy of strontium ranelate in preventing vertebral fractures; and the TROPOS (TReatment Of Peripheral OSteoporosis) study, whose main purpose was to evaluate the efficacy of the molecule on nonvertebral fractures.19 In the SOTI study, after four years of strontium ranelate treatment, a 33% reduction in vertebral fracture risk was observed when compared with the results observed in the placebo group.20 The 5-year data from the TROPOS study have recently been published.21 In that study, the authors observed a significant reduction in nonvertebral fracture risk (–15%). The reduction was somewhat higher (–18%) when major nonvertebral fractures alone were considered. A significant reduction (43%) in hip fracture risk was also observed in patients with low bone densities (T-score <–2.4 at the spine and femoral neck). Strontium ranelate was also found to be effective in patients over 80 years old, in whom a global reduction of 30% in both vertebral and nonvertebral fracture risk was observed.22 Lastly, in a very recent study,23 strontium ranelate was shown to be effective in osteopenic women, regardless of whether they had vertebral fractures at inclusion or not.

Toward a more judicious approach: absolute fracture risk reduction

Globally, these various data show that there are currently several antiosteoporotic molecules with proven therapeutic value. In practice, the question is how does one know which of these molecules is the most relevant for a given patient. The previously mentioned therapeutic trials, in focusing on relative risk only, cannot always provide an answer. Indeed, as mentioned earlier, not all of the molecules were found to be effective in preventing nonvertebral and hip fractures. However, some of the molecules mentioned earlier have clearly been shown to be capable of reducing both vertebral and nonvertebral fracture risks. For practical purposes and to assist prescribers with their choices, appropriate tools are needed. Unfortunately, in the field of osteoporosis, there are no comparative studies on the antifracture efficacy of the various treatments. And given the sizes of the populations required for such studies and the cost of the latter, there is little likelihood that such studies will be undertaken.

Furthermore, it is not possible to compare the reduction in relative risk across different studies. Indeed, the criteria for inclusion in the studies seeking to evaluate the efficacy of a given molecule were quite variable. In some of them, patients with at least one vertebral fracture were included. In others, only patients with densitometric osteoporosis were included. And in other studies, both of these criteria were taken into account. In practice, therefore, other tools are needed to help determine the benefits patients can expect in a given situation. As is the practice in other medical disciplines, the evaluation of absolute risk—and consequently the reduction thereof with treatment—is an important factor to take into consideration.

The evaluation of absolute fracture risk is now possible on an individual basis, thanks to the FRAX® tool.24 More specifically in therapeutic terms, besides what has been previously mentioned, the use of relative risk presents at least two drawbacks. In clinical practice, by definition, there are no placebo groups. The reduction in relative risk observed in certain conditions is therefore of little relevance. Moreover, by definition, the concept of relative risk does not take account of the frequency of the event that one wishes to predict (in our case, the fracture). Taking into consideration relative risk alone to guide therapeutic choices could lead to the treatment of populations in which fracture risk is very low, as illustrated in the article by P. Alonso-Coello.25 Another approach consists of considering the reduction in absolute fracture risk, defined as fracture incidence in the placebo group minus fracture incidence in the treatment group.26 The reduction in absolute risk can also be used to calculate the number needed to treat to prevent the occurrence of the event being considered (in this case, the fracture). This is defined as the inverse of the reduction in absolute risk (expressed as a raw value and not as a percentage). Figure 1 shows the reductions in absolute vertebral fracture risk expressed as percentages for the major clinical trials mentioned in the first part of this paper.

Obviously, this is not a comparative study of the various antiosteoporotic treatments, but this approach underscores the fact that the improvement in absolute risk is very variable from one molecule to the next, for both vertebral and nonvertebral fractures. For vertebral fractures, the reduction in absolute risk varies from 5% for ibandronate to 12% for strontium ranelate.

By definition, the reduction in absolute fracture risk takes account of the fracture risk in the placebo group, which varies widely across different studies. A high incidence of fractures in the placebo group reflects the relevance of the population targeted for assessing the efficacy of the treatment in a clinical trial. Once again, this risk is highest with strontium ranelate and risedronate in the VERT MN study. These results should also be compared with the reduction in absolute risk, which is highest with strontium ranelate as well as with risedronate in the VERT MN study (Figure 1).

Figure 1
Figure 1. Effects of antiosteoporotic treatments on the risk of vertebral fracture (absolute risk reduction).

Evaluation of the vertebral fracture absolute risk reductions of 5 antiosteoporotic
treatments from different controlled studies.
Abbreviations: Al, alendronate; ARR, absolute risk reduction; Iban, ibandronate; Ris, risedronate; SR, strontium ranelate; VERT MN, Vertebral Efficacy with Risedronate Therapy, Multinational; VERT NA, Vertebral Efficacy with Risedronate Therapy, North America; Zol, zoledronate.

Figure 2
Figure 2. Effects of antiosteoporotic treatments on the risk of hip fracture (absolute risk reduction)

Evaluation of the hip fracture absolute risk reductions of 5 antiosteoporotic treatments from different controlled studies.
Abbreviations: Al, alendronate; ARR, absolute risk reduction; Iban, ibandronate; Ris, risedronate; SR, strontium ranelate; Zol, zoledronate.

As regards the reduction in absolute hip fracture risk, less data is available in the literature. A summary of this data is shown in Figure 2. While, by definition, the reduction in absolute hip fracture risk with ibandronate is nil, it is identical and low (–1%) with zoledronate, risedronate, and alendronate. It is most pronounced with strontium ranelate (a 2% reduction in absolute hip fracture risk).

From reduction in absolute fracture risk to number needed to treat to prevent a fracture event As mentioned earlier, the NNT to prevent a fracture event is defined as the inverse of the reduction in absolute fracture risk. This approach has been adopted by several authors27 within the framework of the major therapeutic trials to eval- uate the efficacy of the various treatments. A recent report28 reviews the therapeutic modalities for the prevention and treatment of osteoporosis. One of the interesting things about this report is that it provides NNT values for the studies referred to in the first part of this paper. As an example, it is classically reported that HRT—which was not mentioned previously given the fact that the reasons for not prescribing HRT are numerous—is the only treatment whose antifracture efficacy has been established in the general population. This is true, but aside from the poor benefit/risk ratio of HRT, it is certainly not useful to treat all women with HRT to prevent hip and vertebral fractures. Indeed, for the latter two conditions, the NNTs are 216 and 225, respectively.

As regards the other drugs and following the order in the first part of this paper, the results are as follows: for raloxifene, through the MORE study and taking into consideration the population as a whole, the NNT is 29. Where the bisphosphonates are concerned, the results depend on the molecule. Thus, for alendronate within the framework of the FIT-1 study (patients with at least one vertebral fracture at inclusion), the NNT is 37 for vertebral fractures, 91 for hip fractures, and 53 for wrist fractures. In the FIT-2 study, a post hoc analysis was immediately performed on patients with a T-score <–2.5.When all clinical fractures were taken into consideration, the NNT was 15. For clinical vertebral fractures, the NNT was 34. For risedronate in the VERT NA (Vertebral Efficacy with Risedronate Therapy, North America) study, the NNT to prevent vertebral fracture was 20. In the VERT MN study, it was 10. One explanation for this is that vertebral fracture risk in the placebo group was much higher in the VERT MN study than in the VERT NA study (29%and 16%, respectively). In the previously mentioned HIP (Hip Intervention Program), the NNT was 91. In the BONE study evaluating the efficacy of ibandronate, the NNT for the prevention of vertebral fracture was 20. The most recent bisphosphonate to receive market approval, ie, zoledronate (HORIZON PFT [Health Outcomes and Reduced Incidence with Zoledronic acid ONce yearly Pivotal Fracture Trial]) has an NNT of 13 for vertebral fractures, 91 for hip fractures, and 37 for nonvertebral fractures. These marked differences are not always comparable with the reduction in relative risk. For example, the reduction in hip fracture risk in the HORIZON PFT study was 41%, while the reduction in nonvertebral fracture risk was markedly lower (–25%). In the HORIZON RCT (Health Outcomes and Reduced Incidence with Zoledronic acid ONce yearly Randomized Controlled Trial), the NNT to prevent a clinical fracture was 19. This increased to 48 when clinical vertebral fractures alone were taken into consideration. When all nonvertebral fractures were taken into account, the NNT was 32. As mentioned earlier, in the HORIZON RCT study, a 28% reduction in mortality was observed, corresponding to an NNT of 27. For strontium ranelate, the NNT (SOTI study) to prevent the occurrence of a vertebral fracture was 9,18 and 59 (TROPOS study) to prevent a peripheral fracture.19 In the latter study, the NNT to prevent a hip fracture in women at high risk of fracture at the upper end of the femur was 48.

Conclusion

Studies on osteoporosis are always conducted versus a placebo group. Currently, several antiosteoporotic drugs exist with different mechanisms of action and modalities of administration. These molecules have all demonstrated their usefulness in preventing the occurrence of vertebral fractures, and to a lesser degree—but this is highly dependent on the molecule in question—nonvertebral fractures in general and hip fractures in particular. Taking into consideration the reduction in relative fracture risk alone to guide prescribers’ choices is not sufficient. However, no comparative studies have been conducted with the main goal of evaluating antifracture efficacy. It is known that the populations selected for inclusion in the pivotal studies were quite different. Given these conditions, taking the reduction in absolute fracture risk into consideration seems appropriate.

As mentioned earlier, the reduction in absolute fracture risk varies from one molecule to the next. It is also possible, using this parameter, to calculate the NNT to prevent a fracture event, thus translating the results of clinical trials into current medical practice. These parameters deserve to be taken into consideration as tools to allow a fair and complete comparison of the efficacy of the antiosteoporotic treatment. _

References

1. Kanis JA, Burlet N, Cooper C, et al. European guidance for the diagnosis and management of osteoporosis in postmenopausal women. Osteoporos Int. 2008;19:399-428.
2. Ettinger B, Black DM, Mitlak BH, et al. Reduction of vertebral fracture risk in postmenopausal women with osteoporosis treated with raloxifene: results from a 3-year randomized clinical trial multiple outcomes of raloxifene evaluation (MORE) investigators. JAMA. 1999;382:637-645.
3. Delmas PD, Ensrud KE, Adachi JD, et al. Efficacy of raloxifene on vertebral fracture risk reduction in postmenopausal women with osteoporosis: four-year results froma randomized clinical trial. J Clin EndocrinolMet. 2002;87:3609-3617.
4. Cummings SR, Eckert S, Krueger KA, et al. The effect of raloxifene on risk of breast cancer in postmenopausal women: results from the MORE randomized trial. Multiple Outcomes of Raloxifene Evaluation. JAMA. 1999;281:2189-2197.
5. Barret-Connor E, Mosca L, Collins P, et al. Effects of raloxifene on cardiovascular events and breast cancer in postmenopausal women. N Engl J Med. 2006; 355:125-137.
6. Black DM, Cummings SR, Karpf DB, et al. Randomised trial of effect of alendronate on risk of fracture in women with existing vertebral fractures. Fracture Intervention Trial research group. Lancet. 1996;348:1535-1541.
7. Cummings SR, Black DM, Thompson DE, et al. Effect of alendronate on risk of fracture in women with low bone density but without vertebral fractures: results from the Fracture Intervention Trial. JAMA. 1998;280:2077-2082.
8. Harris ST, Watts NB, Genant HK, et al. Effects of risedronate treatment on vertebral and nonvertebral fractures in women with postmenopausal osteoporosis: a randomized controlled trial. Vertebral Efficacy with Risedronate Therapy (VERT) study group. JAMA. 1999;282:1344-1352.
9. Reginster J, Minne HW, Sorensen OH, et al. Randomized trial of the effects of risedronate on vertebral fractures in women with established postmenopausal osteoporosis. Vertebral Efficacy with Risedronate Therapy (VERT) study group. Osteoporos Int. 2000;11:83-91.
10. McClung MR, Geusens P, Miller PD, et al. Effect of risedronate on the risk of hip fracture in elderly women. Hip Intervention Program study group. N Engl J Med. 2001;344:333-340.
11. Boonen S, Laan RF, Barton IP, Watts NB. Effect of osteoporosis treatments on risk of non-vertebral fractures: review and meta-analysis of intention-to-treat studies. Osteoporos Int. 2005;16:1291-1298.
12. Chesnut III CH, Skag A, Christiansen C, et al. Effects of oral ibandronate administered daily or intermittently on fracture risk in postmenopausal osteoporosis. J Bone Miner Res. 2004;19:1241-1249.
13. Black DM, Delmas PD, Eastell R, et al. Once-yearly zoledronic acid for treatment of postmenopausal osteoporosis. N Engl J Med. 2007;356:1809-1822.
14. Lyles KW, Colon-Emeric CS, Magaziner JS, et al. The HORIZON recurrent fracture trial zoledronic acid and clinical fractures and mortality after hip fracture. N Engl J Med. 2007;357:1799-1809.
15. Sorensen OH, Crawford GM, Mulder H, et al. Long-term efficacy of risedronate: a 5-year placebo-controlled clinical experience. Bone. 2003;32:120-126.
16. Bone HG, Hosking D, Devogelaer JP, et al. Ten years’ experience with alendronate for osteoporosis in postmenopausal women. N Engl J Med. 2004;350: 1189-1199.
17. Greenspan SL, Bone HG, Ettinger MP, et al. Effect of recombinant human parathyroid hormone (1-84) on vertebral fracture and bone mineral density in postmenopausal women with osteoporosis. A randomized trial. Ann Intern Med. 2007;146:326-339.
18. Meunier PJ, Roux C, Seeman E, et al. The effects of strontium ranelate on the risk of vertebral fracture in women with postmenopausal osteoporosis. N Engl J Med. 2004;350:459-468.
19. Reginster JY, Seeman E, De Vernejoul MC, et al. Strontium ranelate reduces the risk of nonvertebral fractures in postmenopausal women with osteoporosis: treatment of peripheral osteoporosis (TROPOS) study. J Clin Endocrinol Metab. 2005;90:2816-2822.
20. Meunier PJ, Roux C, Ortolani S, et al. Effects of long-term strontium ranelate treatment on vertebral fracture risk in postmenopausal women with osteoporosis. Osteoporos Int. 2009;20:1663-1673.
21. Reginster JY, Felsenberg D, Boonen S, et al. Effects of long-term strontium ranelate treatment on the risk of nonvertebral and vertebral fractures in postmenopausal osteoporosis. Results of a five-year, randomized, placebo-controlled trial. Arthritis Rheum. 2008;58:1687-1695.
22. Seeman E, Vellas B, Benhamou C, et al. Strontium ranelate reduces the risk of vertebral and onvertebral fractures in women eighty years of age and older. J Bone Miner Res. 2006;21:1113-1120.
23. Seeman E, Devogelaer JP, Lorenc R, et al. Strontium ranelate reduces the risk of vertebral fractures in patients with osteopenia. J Bone Miner Res. 2008;23: 433-438.
24. Kanis JA, Oden A, Johansson H, Borgström F, Ström O, McCloskey E. FRAX® and its applications to clinical practice. Bone. 2009;44:734-743.
25. Alonso-Coello P, Lopez Garcia-Franco A, Guyatt G, Moynihan R. Drugs for preosteoporosis: prevention or disease mongering? BMJ. 2008;336:126-129.
26. Replogle WH, Johnson WD. Interpretation of absolute measures of disease risk in comparative research. Fam Med. 2007;39:432-435.
27. Nuovo J, Melnikow J, Chang D. Reporting number needed to treat and absolute risk reduction in randomized controlled trials. JAMA. 2002;287:2813-2814.
28. MacLaughlin EJ, Raehl C. ASHP therapeutic position statement on the prevention and treatment of osteoporosis in adults. Am J Health Syst Pharm. 2008; 65:343-357.