INTRODUCTION

Osteoporotic fractures affect millions of postmen-opausal women, and can have a substantial negative effect on daily functioning and quality of life. These fractures are associated with an increased risk of future fracture, clinical complications, and a higher risk of death. However, osteoporosis remains woefully under-diagnosed and undertreated, even among patients who already have fractures. When treatment is provided, adherence and persistence are often suboptimal, reducing the likelihood of a positive outcome.

A variety of treatments for osteoporosis are available, many of which seek to improve compliance by reducing the frequency of dosing and the resulting side effects. All currently approved oral bisphosphonates have an indication for reducing the risk of vertebral fracture, but they have shown mixed results for efficacy against nonvertebral fractures, including hip fractures. Because oral bisphosphonates may be associated with upper gastrointestinal side effects, adherence can be reduced. Other concerns with bisphosphonates include osteo-necrosis of the jaw and reduced renal function.

The latest advance in the field of osteoporosis is the FDA approval of zoledronic acid, administered once a year via intravenous infusion for treatment of post-menopausal osteoporosis. In the Health Outcomes and Reduced Incidence With Zoledronic Acid Once Yearly (HORIZON)-Pivotal Fracture Trial, treatment with zoledronic acid over a 3-year period reduced the risk of vertebral fractures and nonvertebral fractures, including hip fractures, when compared with placebo. Zoledronic acid was also evaluated in a previously understudied group of hip fracture patients. The HORIZON-Recurrent Fracture Trial examined the risk of new fractures in men and women who had hip fractures but could not tolerate oral bisphosphonates. The results of this trial were reported at the 2007 American Society for Bone and Mineral Research (ASBMR) meeting.

The diagnosis and treatment of osteoporosis face continued challenges. This monograph discusses some of the obstacles that lead to underdiagnosis and under-treatment; provides an overview of current bisphos-phonate treatments; identifies opportunities to improve treatment adherence and persistence; and presents recent data from clinical studies in osteoporosis patients.

OBSTACLES TO OPTIMAL OSTEOPOROSIS CARE

Nelson B. Watts, MD

Osteoporotic fractures are a major healthcare problem in the United States and around the world. A study in the United States estimated that there were more than 2 million osteoporotic fractures in 2005, nearly 73% of them at nonvertebral sites.¹ Vertebral fractures were probably underestimated because they are often not recognized clinically. More than two‑thirds of the fractures—about 1.4 million—occurred in women, with 89% of those among white women.¹ In fact, osteoporotic fractures are more common among older women of all ages than breast cancer, strokes, and heart attacks combined (Figure 1).^1‑3

The direct cost of caring for patients with osteoporotic fractures in 2005 was estimated to be almost $17 billion.¹ This does not include the costs of lost pro-ductivity, loss of independence, unpaid caregiver time, transportation, and social services. Hip fractures account for nearly 75% of the total cost, although they represent only 14% of all fractures. Based on the aging of the US population, the annual incidence of fractures is predicted to increase to more than 3 million by 2025, with annual costs of $25.3 billion.¹

Obstacles to Optimal Osteoporosis Care: Underdiagnosis and Undertreatment

Osteoporosis is often underdiagnosed and undertreated, even though it is common and associated with high mor-bidity and cost. There is general agreement among the various clinical guidelines that all women ≥65 years of age should be screened for osteoporosis. There is less agreement on when younger postmenopausal women should be screened. The US Preventive Services Task Force recommends screening women 60 to 64 years of age who weigh less than 70 kg (154 lbs), and makes no recommendation for or against testing in postmeno-pausal women <60 years of age.⁴ The National Osteoporosis Foundation (NOF) and the American Association of Clinical Endocrinologists (AACE) recom-mend screening postmenopausal women <65 years of age if clinical risk factors, such as prior fracture, low body weight (<127 lbs), current smoking, or family history of fracture, are present.^5,6 The AACE puts part-icular emphasis on bone mineral density (BMD) testing in women who have had a prior non traumatic fracture.

Despite these recommendations, few women are screened or treated, even among those who have experienced a fragility fracture. The National Committee for Quality Assurance (NCQA), which monitors how well healthcare providers meet recommended standards, reported that in 2003, only 18% of women ≥67 years of age who had suffered a fracture received either a BMD test or a prescription medication to treat osteoporosis within 6 months of the fracture.⁷ The rate of screening or treatment improved incrementally to 19.0% in 2004 and to 20.1% in 2005.⁸ The rate of osteoporosis man-agement compares poorly with rates for other serious and chronic diseases (Table 1).⁸

King et al conducted a 3-year study to examine the hypothetical effect of increased osteoporosis screening and treatment among Medicare beneficiaries.⁹ This study estimated that in 2001, 71.6% of women ≥65 years of age with osteoporosis (N=5.1 million) received neither treatment nor a Medicare-reimbursed BMD test. Among US women aged ≥65 years, only 8.8% had a BMD test in 2001. The study projected that testing an additional 1 million women would lead to treatment of 440,000 new patients and prevent more than 35,000 fractures over 3 years, producing a net Medicare savings of $77.86 million.⁹ Despite these potential savings, proposed cuts to Medicare reimbursement for BMD screening may reduce the availability of this key test in the near future. (For more information, go to the International Society for Clinical Densitometry Web site: www.iscd.org.)

In addition to the low rates of screening, osteoporosis is also undertreated, even when diagnosed, as illustrated by the NCQA data presented previously. The various guidelines (AACE, ACOG, NAMS, NOF) agree that pharmacologic therapy is rarely needed for women with T‑scores of ‑1.5 or higher, and that treatment is indicated for women with T‑scores ≤-2.5, regardless of whether other risk factors are present.^5,6,10,11 There is less agreement regarding which patients with T‑scores between -1.5 and -2.5 should be treated
(Figure 2).^5,6,10,11 How-ever, there are data to show that at least half of women who experience low-trauma fractures have a BMD in this range. Using data from the National Osteoporosis Risk Assessment (NORA) study (149,524 postmenopausal women), Siris et al showed that while the fracture rate was higher at lower T‑scores, as expected, the overall number of fractures roughly followed the bell curve of the BMD population distribution but was shifted to the left (Figure 3).¹² Most women had T‑scores above ‑2.5, and approximately half of all women with fractures had BMD in that range. Based on numbers of fractures alone, it behooves healthcare providers to identify women with T‑scores >‑2.5 who have other risk factors for fractures.

Age is an important independent risk factor for fracture. Kanis et al determined the 10‑year absolute probability of fracture in the Swedish population. Within a given age group of women, lower T‑scores correlated with an increased risk of fracture (Figure 4).¹³ For a given T‑score, older age was linked with substantially increased risk. At a T‑score of ‑2.5, for example, the 10‑year probability of osteoporotic fracture approx-imately doubles, from 11.3% at 50 years of age to 22.8% at 70 years of age.¹³ The World Health Organization (WHO) is developing an absolute risk prediction model that will consider not only BMD and age, but also risk factors such as prevalent fracture, corticosteroid use, cigarette smoking, alcohol use, and secondary causes of osteoporosis such as rheumatoid arthritis. This model should soon be available and will provide 10‑year fracture probabilities that can be combined with national guidelines to determine threshold interventions.¹⁴

Obstacles to Optimal Osteoporosis Care: Poor Patient Adherence

Unfortunately, even when patients are diagnosed with osteoporosis and receive appropriate therapy, many do not take their medication as they should. Compliance was the old term for following the doctor’s instructions; the preferred term today is adherence, which indicates an active collaboration between patient and provider in a mutually acceptable course of behavior.¹⁵ Persist-
ence refers to how long a patient maintains adherence or compliance.

Poor adherence and poor persistence are common with chronic, “silent” diseases such as hypertension, hyper-cholesterolemia, and osteoporosis. Among 15,175 patients with hypertension, 1‑year persistence with antihypertensive medications ranged from a low of 21% to 67%; 4‑year persistence ranged from 16% to 51%.¹⁶ In a retrospective cohort study of 34,501 patients starting statin therapy, the proportion who were adherent over time was 60% at 3 months, 43% at 6 months, and 26% at 60 months.¹⁷

Adherence and persistence rates for osteoporosis are no better. In various studies, adherence rates to osteo-porosis medication varied from <25% to 90%, depending on the therapy in question and the definition of adherence used.^18‑22

Better adherence to osteoporosis treatment is associ-ated with better clinical outcomes.^19,23 Siris et al conducted a retrospective cohort study among 35,537 women taking bisphosphonates.²⁴ Over the 2‑year follow‑up period, only 43% of subjects were considered compliant, defined as a medical possession ratio (MPR; days of medication available/study period) of 0.80 or higher. Women who achieved compliance had a 21% reduction in the overall fracture rate compared with those who were not compliant (P<.001). For patients who were compliant, the relative risk of fracture was also significantly lower for vertebral (37%; P<.001), nonvertebral (20%; P<.001), and hip fractures (37%; P<.001), but not for wrist fractures. Persistence—defined as no more than a 30‑day gap in refills—was also key to reducing fracture risk. Over the 24‑month study period, 80% of subjects were nonpersistent. Persistence was associated with a 29% reduction in overall fracture risk and a 45% reduction in the risk of hip fracture (P<.001 for both comparisons). Figure 5 shows the relationship between compliance and fracture risk in this study.²⁴ The probability of fracture is high for MPR values below 0.50. A substantial reduction in fracture risk is first seen at an MPR of approximately 0.60, and then fracture risk drops more sharply at MPR values ≥0.75.

It seems intuitive that more consistent adherence to a medication regimen would improve the outcomes that medication is intended to treat, but adherence by itself may be a marker for better outcomes. In 1980, the Cor-onary Drug Project reported findings of an investigation on the effects of several lipid‑modifying drugs on mortality and secondary coronary events in men who had had a heart attack.²⁵ Mortality among patients treated with clofibrate (n=1103) was not significantly different than for patients who received placebo (n=2789). However, patients who had good adherence to clofibrate (took ≥80% of medication over 5 years) had a lower risk of death compared with patients who had poor adherence (15.0% vs 24.6%; P=.00011); patients who adhered to the placebo regimen had a similar reduction in mortality compared with those who had poor adherence (15.1% vs 28.3%; P<.0001). These results suggest that patients who have good adherence to a medication regimen likely also exhibit other charact-eristics, such as better overall health or better adherence to dietary changes that can influence clinical outcomes.

Setting the stage for good adherence begins when treatment is initiated. Patients must be ready to accept treatment. Stages of readiness to consider osteoporosis treatment were explored in a study of 21 postmeno-pausal women (mean age 85) hospitalized with a low‑ impact hip fracture (Figure 6).²⁶ Three‑ fourths of the patients were in the first 3 stages: unaware of treatment, not seriously considering treatment, or having decided against it. Only 15% of patients had actually been taking medication for osteoporosis for any length of time. These results demonstrate that there is room for improvement in patient education and in understanding the nature and consequences of osteoporosis.

After the initiation of treatment, the patient must continue with therapy for it to be effective. In a telephone survey of 956 women, Tosteson et al investigated early dis-continuation of osteoporosis treatment.²⁷ An average of 7 months after treatment initiation, 22% of the women had discontinued treatment (including hormone therapy, raloxifene, and alendronate). The most common reason for discontinuation was the occurrence of side effects; two‑thirds of women cited this concern as their reason for stopping. Among women who reported extreme or very bothersome side effects, 71% discontinued therapy. Safety concerns (breast cancer for hormone therapy or blood clots for raloxifene) were the second most common reason for discontinuation. Women who thought their bone density tests didn’t show osteoporosis were 60% more likely to stop treatment,²⁷ which highlights the need for women to be educated about their BMD tests, and about what side effects to expect and how to minimize them.

Ongoing feedback about the effects of therapy can have an influence on treatment adherence. In the IMPACT (Improving Measurements of Persistence on Actonel Treatment) study, bone turnover markers were assessed in 2382 women after starting osteoporosis therapy (risedronate, vitamin D, and calcium).²⁸ Centers were randomized to receive information about changes in bone turnover markers and provide this feedback to patients (RE+; n=1189), or to not receive this information (RE–; n=1113). Overall, persistence was high in this trial (80% in the RE+ group and 77% in RE–), perhaps due to patient awareness of electronic monitoring. Patients in the RE+ group who had improvement in bone turnover markers were significantly more likely to stay on therapy compared with patients in the RE– group (HR=0.71; 95% CI, 0.53‑0.95; P=.02). However, RE+ patients who had a poor biomarker response were significantly more likely to discontinue treatment compared with patients in the RE– group (HR=2.22; 95% CI, 1.27‑3.89; P=.005).

Poor adherence may be due, at least in part, to difficult‑ to‑follow treatment regimens. In the past, bisphos-phonates required daily oral dosing first thing in the morning, with no food for 30 minutes afterwards. Cramer et al compared adherence with once‑weekly and once‑daily bisphosphonates in 2741 postmenopausal women.²¹ Weekly users had significantly higher compliance than daily users (69.2% vs 57.6% MPR; P≤.0001). Weekly users also had significantly longer persistence with therapy than daily users (P<.0001). Figure 7 shows persistence over a year of follow‑up.²¹ There was a 20% to 30% drop‑off for both regimens within about a month of initiating treatment. At the end of the first year, only 44% of patients given a weekly agent were still on therapy, whereas just slightly over 30% of those given daily therapy were still persistent.

Summary

There are multiple obstacles to optimal care of osteo-porosis. The guidelines for diagnosis and treatment could be improved. It is expected that the WHO guidelines regarding absolute fracture risk will help in making treatment decisions for those patients most at risk of fracture. Unfortunately, gains in this area may be offset by reduced reimbursement for BMD testing. Healthcare providers should be educated about the need to treat osteoporosis once it is diagnosed. Patients also should be educated about the need for treatment and the risks of avoiding it. Any side effects or safety concerns must be addressed to encourage adherence and persistence. Feedback about the effectiveness of treatment offers an incentive to continue therapy for patients with a positive response, and an opportunity to adjust therapy for patients with a poor response.

Table 1. Medicare Disease Management Rates in 2005⁸

^*Women aged 50 to 69 who had ≥1 mammogram in past 2 years.
^†Adults aged 46 to 85 with hypertension who had a blood pressure reading ≤140/90 mm Hg during past year.
^‡Adults aged ≥65 who had ≥1 eye exam for glaucoma in past 2 years.

Figure 1. Osteoporotic fractures are a more common health problem among women of all ages than heart attack, stroke, and breast cancer.^1‑3

^*2005 annual incidence estimate, women ≥50 years of age.
†2006 new cases estimate, women of all ages.
‡2004 estimate, new and recurrent myocardial infarction among women ≥35 years of age.
§2004 estimate, women of all ages.

Figure 2. When to treat postmenopausal osteoporosis.^5,6,10,11

ACOG = American College of Obstetrics and Gynecology; AACE = American Association of Clinical Endocrinologists; BMD = bone mineral density; NAMS = North American Menopause Society; NOF = National Osteoporosis Foundation.

Figure 3. Population distribution of BMD and fractures among postmenopausal US women.¹²

BMD = bone mineral density.
Adapted with permission from Siris ES et al. Arch Intern Med. 2004;164:1108‑1112.

Figure 4. The relationship between BMD and 10‑year fracture probability in women according to age.¹³

BMD = bone mineral density.

Figure 5. Probability of fracture over 24 months in women treated with bisphosphonates.²⁴

Adapted with permission from Siris ES et al. Mayo Clin Proc. 2006;81:1013‑1022.

Figure 6. Stages of readiness to accept treatment for osteoporosis in the modified precaution adoption process model.²⁶

Figure 7. Persistence (percentage of patients remaining on therapy) with daily versus weekly bisphosphonate therapy.²¹

Adapted with permission from Cramer JA et al. Curr Med Res Opin. 2005;21:1453‑146

OVERVIEW OF CURRENT BISPHOSPHONATE TREATMENT ARMAMENTARIUM

Michael R. McClung, MD, FACP

The first drugs documented to reduce fractures in patients with osteoporosis were introduced in the mid 1990s. Today, we have a long and expanding list of available treatments for osteoporosis, ranging from the bisphosphonates to hormonal therapies. Among these various agents, bisphosphonates have become the mainstay of treatment for osteoporosis and will be the focus of this discussion.

Bisphosphonates are stable synthetic analogs of pyro-phosphate. They are strong chelators of divalent cations including calcium and magnesium, and bind tightly to bone mineral (hydroxyapatite). Simple bisphosphon-ates, such as etidronate, inhibit bone mineral growth and dissolution.²⁹ Upon release from bone mineral, bisphos-phonates are taken up by endocytosis into osteoclasts.³⁰ Bisphosphonate accumulation reduces osteoclast act-ivity and causes apoptosis of osteoclasts, resulting in reduced bone resorption.

All bisphosphonates have in common a P‑C‑P back-bone, which is responsible for bone mineral binding.²⁹ The side chain substituents affect the antiresorptive potency of the particular bisphosphonate. The addition of nitrogen groups has increased the potency up to 1000‑fold more than the early-generation bisphos-phonates clodronate and etidronate. Pamidronate, alendronate, and ibandronate have a basic amino group on the alkyl side chain, while the most potent anti-resorptive bisphosphonates, zoledronic acid and ris-edronate, contain nitrogen within a heterocyclic ring.^29,31 Nitrogen‑containing bisphos-phonates exert their antiresorptive effect through the inhibition of isoprenoid (farnesyl and geranylgeranyl) biosynthesis, and the resulting inhibition of prenylation of small guanosine triphosphate (GTP)‑binding mole-cules such as cdc42, rac, and rho.^30,31 These molecules play key roles in signal transduction pathways that affect a variety of important cellular processes. Inhibition of protein pre-nylation disrupts these processes and leads to loss of osteoclast function, as well as apoptosis.³¹ It has been shown that bisphosphonates can inhibit farnesyl diphos-phate synthase (FPPS) in vitro.³⁰

These 2 attributes—binding to bone mineral and inhibiting FPPS—account for the clinical efficacy of bisphosphonates. Of the 4 bisphosphonates most com-monly used, differences exist in the tightness with which they bind to bone mineral (zoledronic acid the most and risedronate the least) and in the potency with which they inhibit enzymatic activity (zoledronic acid the greatest and alendronate the lowest) (Figure 1).^29,30 Each drug has its own unique combination of mineral‑binding affinity and osteoclast‑inhibiting activity. These differ-ences among the bisphosphonates could lead to clinical differences in the speed of onset and reversibility of effect, the degree of bone turnover suppression, differ-ences in bisphosphonate uptake in trabecular and cortical bone, antifracture efficacy (vertebral versus nonvertebral), and safety and tolerability. However, it is not known whether there are true differences in the clinical effectiveness of the drugs.

In the clinical setting, bisphosphonates substantially re-duce biochemical indices of bone remodeling, as shown in Figure 2.³² They predominantly inhibit osteoclast‑ mediated bone resorption, but due to osteoclast‑ osteoblast crosstalk, there is also delayed inhibition of bone formation. These effects on bone remodeling are beneficial in patients with high bone turnover. The im-balance that occurs when resorption exceeds formation is corrected, bone remodeling activity is normalized, and the loss of bone mass and deterioration of bone archit-ecture is inhibited. As a result of those effects, bone mineral density (BMD) increases modestly at important skeletal sites.

The net result of these beneficial effects on bone is that bisphosphonates provide increased protection from fractures to patients with osteoporosis. The approved doses of the 3 oral bisphosphonates that are commonly used to treat postmenopausal osteoporosis in the United States are different (Table 1),^33‑35 based on doses used in pivotal clinical trials rather than on proven differences in potency. The optimal doses of any of these agents for fracture reduction have not been carefully evaluated. All 3 drugs are approved for the reduction of vertebral fractures, which was the primary end point for the pivotal studies leading to approval of these agents. In addition, risedronate is approved for the reduction of nonvertebral fractures, and alendronate for preventing hip fractures. All of these drugs are approved for daily dosing. Alendronate and risedronate are approved for weekly dosing, and ibandronate can be dosed monthly. Risedronate may also be dosed monthly at 75 mg on each of 2 consecutive days. An intravenous (IV) form of ibandronate is administered once every 3 months.³⁶

Data from the registration trials evaluating the effects of these agents on risk of vertebral fracture after at least 3 years of treatment, compared with placebo, are shown in Table 2.^37‑41 As is clear from the incidence of vertebral fractures in the placebo group of each trial, there were substantial differences in the populations of patients studied, which preclude making any comparisons among the studies. For example, in the Clinical Fracture Arm of the Fracture Intervention Trial (FIT), which included women without preexisting vertebral fractures but with low BMD, the incidence of vertebral fractures in the placebo group was about 4% over 4 years.³⁸ The Vertebral Fracture Arm of the same trial included women with prevalent vertebral fractures; in this arm the placebo group had a 15% incidence of new vertebral fractures over 3 years.³⁷ Despite the differences, the relative reduction in fracture risk was similar among the trials. The bisphosphonates offer clear protection against vertebral fracture for women with osteoporosis or prior vertebral fractures.

The most impressive evidence of bisphosphonate efficacy is the effect on the risk of experiencing multiple vertebral fractures. Before these drugs were available, it was not uncommon to see an otherwise healthy postmenopausal woman recover completely from a first vertebral fracture, but begin a downward spiral of lost function and impaired quality of life after experiencing a second and third fracture. In clinical trials, the probability of experiencing multiple vertebral fractures after being randomized to oral bisphosphonate treatment was reduced by between 77% and 96% compared with placebo treatment (Figure 3).^37,42,43

Assessing the effect on nonvertebral fractures is more difficult than assessing vertebral fracture risk for several reasons. Nonvertebral fracture risk was not a primary end point in any of the pivotal trials. Therefore, the studies were not powered for this outcome. Also, nonvertebral fractures are substantially affected by nonskeletal risk factors, such as dementia and falling, which would not be expected to respond to bone‑strengthening drugs, making it more difficult to demonstrate a reduction in risk for these fractures. Finally, the trials used different definitions of nonvertebral fracture, further confounding comparisons between the studies. Among the pivotal trials for alendronate, ibandronate, and risedronate, only the risedronate North American Vertebral Fracture Trial (VERT‑NA) showed a significant effect on the risk of nonvertebral fracture.⁴⁰ In that study, patients ran-domized to risedronate 5 mg daily (n=812) had a lower incidence of nonvertebral fractures over 3 years than patients randomized to placebo (n=815) [5.2% vs 8.4%, respectively; RR=0.6; 95% CI, 0.39‑0.94; P=.02].

Significant reductions in the incidence of nonvertebral fractures have been shown in nonregistration trials with bisphosphonates. In the Fosamax International Trial (FOSIT), 2.0% of patients randomized to alendronate 10 mg daily (n=950) experienced at least 1 nonvertebral fracture compared with 3.9% of patients in the placebo group (n=958) at 12 months (RR=0.53; 95% CI, 0.3‑0.9; P=.021).⁴⁴ In the risedronate Hip Intervention Program (HIP) study, the risk of nonvertebral fracture was reduced from 11.2% in the placebo group (n=3134) to 9.4% among women taking risedronate (n=6197; 2.5 or 5.0 mg daily) [RR=0.8; 95% CI, 0.7‑1.0; P=.03].⁴⁵

Two of the bisphosphonate agents discussed here—alendronate and risedronate—have been shown to reduce the incidence of hip fractures in postmenopausal women with osteoporosis. In the Vertebral Fracture Arm of FIT (patients with prevalent vertebral fractures), hip fracture risk was a secondary end point.³⁷ There was a 51% reduction in the risk of hip fractures, from 2.2% with placebo (n=1005) to 1.1% with alendronate (n=1022) [RR=0.49; 95% CI, 0.23‑0.99; P=.047]. The HIP trial with risedronate is the only trial discussed here in which hip fracture was the primary end point.⁴⁵ Among women 70 to 79 years of age with osteoporosis treated with risedronate (n=3624), there was a 40% reduction in the risk of hip fracture over 3 years compared with those treated with placebo (n=1821; 1.9% vs 3.2%, respectively; RR=0.6; 95% CI, 0.4‑0.9; P=.009).

The results presented above are all based on daily oral dosing of bisphosphonates. Weekly and monthly dosing regimens are also approved and are the most commonly used due to their increased convenience for patients. These doses were approved based on noninferiority studies comparing BMD responses to daily and nondaily regimens. None of the nondaily doses have been demonstrated to reduce fracture risk.

As discussed in the previous section by Dr. Watts, concerns about the potential or real side effects of these agents can be a major barrier to their acceptance and continued use by patients. Many of the concerns with bisphosphonates have not been documented in large, well‑designed, randomized clinical trials. In general, these drugs were very well tolerated.

Upper gastrointestinal (GI) side effects are a major concern for many patients, to the point that they are hesitant to start bisphosphonate therapy. Upper GI intolerance has not been documented in any clinical trial to occur more frequently in patients randomized to oral bisphosphonates compared with patients randomized to placebo (Table 3).^{37‑41,44,45} The background incidence of these symptoms is very high, but there is no evidence from large clinical trials that it is higher when taking bisphosphonates. However, upper GI intolerance has been observed in clinical practice. The perception that these drugs are associated with GI side effects is
very strong and continues to be a barrier to therapy
for osteoporosis.

Bone and muscle pain have been reported in clinical trials in association with these agents. These side effects tend to occur relatively soon after therapy is initiated and tend to be relatively mild. Flu‑like symptoms due to an acute phase reaction have also been observed in a small proportion of patients taking monthly oral doses (150 mg) and every–3‑month IV doses (3 mg) of ibandronate.

There is also concern with bisphosphonate therapy regarding osteonecrosis of the jaw (ONJ), perhaps more in the United States than elsewhere. In the clinical trials of the drugs described here, there were no identified cases of ONJ, although those trials comprised about 25,000 patients who had received bisphosphonate treatment. A task force of the American Society for Bone and Mineral Research, appointed to evaluate the risk of ONJ with bisphosphonate therapy, concluded that the risk of ONJ with oral bisphosphonates is between 1/10,000 and 1/100,000.⁴⁶ Although the risk is very low, ONJ remains a concern among patients and healthcare providers alike.

Summary

Bisphosphonates are potent inhibitors of osteoclast activity and bone resorption. These agents reduce bone remodeling to the lower half of the normal range, restore the balance between resorption and formation, and, as a consequence, have been proven to be effective in reducing fracture risk in patients with osteoporosis. The consistency of results across multiple studies with multiple agents increases the confidence that we have in the effectiveness of these drugs. As a group, the bisphosphonates are well tolerated. These agents are the primary therapies for the management of osteo-porosis at this time. 

Table 1. Overview of Bisphosphonate Options in Early 2007 for Postmenopausal Osteoporosis

*On 2 consecutive days per month.

Table 2. Reduction in Risk of Vertebral Fracture After at Least 3 Years of Treatment in Postmenopausal Women With Osteoporosis and/or Prior Vertebral Fractures*

BONE = oral iBandronate Osteoporosis Vertical Fracture in North America and Europe; CFA = Clinical Fracture Arm; FIT = Fracture Intervention Trial; MN = multinational; NA = North America; VERT = Vertebral Efficacy with Risedronate Therapy; VFA = Vertebral Fracture Arm.
^*These studies were conducted in separate populations and cannot be compared directly.
†Patients were followed for an average of 4.2 years.

Table 3. Patients Reporting Upper Gastrointestinal (GI) Adverse Events (AEs) in Trials of Daily Oral Bisphosphonates*

No significant differences between rates for bisphosphonates and placebo.
BONE = oral IBandronate Osteoporosis Vertical Fracture in North America and Europe; CFA = Clinical Fracture Arm;
FIT = Fracture Intervention Trial; FOSIT = Fosamax International Trial; MN = Multinational; HIP = Hip Intervention Program; NA = North America; NR = not reported; VERT = Vertebral Efficacy with Risedronate Therapy; VFA = Vertebral Fracture Arm.
^* Definitions of moderate to severe upper GI AE varied between trials.
^† Overall GI AEs not reported. No significant differences between treatment arms for any individual GI AE.

Figure 1. The panel on the left shows the in vitro binding affinity of bisphosphonates for bone, which determines attachment to bone and duration of effect.²⁹ The panel on the right shows the in vitro inhibition of FPP synthase by bisphosphonates, which determines antiresorptive potency. The measure shown is the dose of drug that is required to inhibit half the activity, with lower concentrations indicating more potent agents.³⁰

ALN = alendronate; FPP = farnesyl diphosphate; IBA = ibandronate; rhFPP = recombinant human FPP; RIS = risedronate; ZA = zoledronic acid.

Figure 2. Change from baseline over time in biochemical markers of bone resorption (urinary deoxypyridinoline/ creatinine) and formation (serum osteocalcin) among postmenopausal women treated with daily alendronate.³²

Adapted with permission from Chesnut CH III et al. Am J Med. 1995;99:144‑152.

Figure 3. Incidence of multiple vertebral fractures among postmenopausal women with previous vertebral fractures who were treated for at least 1 year with placebo, alendronate, and risedronate in the FIT³⁷ and VERT⁴³ studies.

FIT = Fracture Intervention Trial; MN = Multinational; NA = North America; VERT = Vertebral Efficacy with Risedronate Therapy.

OSTEOPOROSIS: OPPORTUNITIES FOR BETTER OUTCOMES

Felicia Cosman, MD

Poor adherence to therapy is common among patients with osteoporosis, as is typical for a chronic disease with no symptoms. Adherence is further reduced by the inconvenience of daily dosing regimens. Recent years have seen a trend towards less frequent administration to help improve adherence, with dosing regimens moving from daily to weekly and then monthly. Recently, extended‑dosing osteoporosis agents have been approved for quarterly and once‑yearly intravenous (IV) treatment, and another agent is under investigation for biannual subcutaneous (SC) administration.

Tolerability is also a major underlying reason for poor adherence. Oral bisphosphonates are associated with the perception of increased upper gastrointestinal (GI) irritation, with side effects of heartburn, dysphagia, and odynophagia. These side effects occur no more commonly in the bisphosphonate‑treatment arms of clinical trials than in the placebo arms. However, there is some suggestion of an increase in postmarketing studies, including rare cases of bleeding and esophageal perforation.^47,48 With regard to patient adherence, the question of whether upper GI symptoms are actually caused by these agents is not very relevant; it is clear that this is one of the reasons patients stop taking their medications. Intravenously administered bisphos-phonates avoid GI irritation, although they have been associated with an acute phase reaction following infusion.⁴⁸ Both oral and IV bisphosphonates are associated with osteonecrosis of the jaw (ONJ). Although this side effect occurs with both forms of bisphosphonate, it rarely occurs at the doses used to treat osteoporosis and is more common in patients receiving high‑dose bisphosphonates to treat cancer—especially multiple myeloma.^46,48 The same is true for renal insufficiency, which occurs only rarely at the highest doses.⁴⁸

This section will focus on some of the newest treatment options for osteoporosis, which include extended‑ dosing regimens and parenteral administration routes: quarterly IV ibandronate, biannual SC denosumab, and annual IV zoledronic acid. Given the barriers to adherence dis-cussed previously, it is hoped that these treatment options will lead to improved adherence and persistence and, ultimately, improved outcomes against fracture.

Ibandronate

Intravenous ibandronate was investigated in the Dosing IntraVenous Administration (DIVA) study.⁴⁹ DIVA was a noninferiority study with daily oral ibandronate as the comparator and bone mineral density (BMD) as a surrogate end point for fracture risk. The objective was to identify the optimal IV dosing regimen for ibandronate in the treatment of osteoporosis in postmenopausal women. Women aged 55 to 80 years with lumbar spine T‑score <‑2.5 (N=1395) were randomized to 2 mg IV ibandronate every 2 months, 3 mg IV ibandronate every 3 months, or 2.5 mg oral daily ibandronate.⁴⁹

At 1 and 2 years’ follow‑up, both IV dosing groups showed significantly greater improvement from baseline in lumbar spine BMD compared with the daily oral group (P<.001; Figure 1).^49,50 For the currently approved dose of 3 mg every 3 months, the mean treatment differences versus oral daily therapy were 1.0% at 1 year and 1.5% at 2 years. Hip BMD was also improved in the IV groups compared with oral treatment: total hip and hip trochanter BMD increased significantly more with IV versus oral ibandronate (P<.001 for all comparisons). Femoral neck BMD showed a numerically greater increment, but was not statistically higher in the IV groups versus oral ibandronate.^49,50

The effect of extended‑dose ibandronate on the risk of nonvertebral or clinical fractures was investigated in a meta‑analysis of pooled data from the DIVA, MOBILE (Monthly Oral iBandronate In LadiEs), BONE (oral iBandronate Osteoporosis vertebral fracture in North America and Europe), and IV fracture prevention trials.⁵¹ All of these trials included postmenopausal women aged 55 to 80 years with osteoporosis. The BONE and IV fracture prevention studies were placebo controlled and investigated vertebral fracture as a primary end point. These studies included only patients with prevalent vertebral fractures. The MOBILE and DIVA studies compared extended‑dose ibandronate regimens (oral versus IV administration) to once‑daily ibandronate 2.5 mg, and did not require prevalent vertebral fractures as entry criteria. The primary end point in both of these studies was lumbar spine BMD.

The meta‑analysis included 8710 patients from the 4 trials. The primary end point of the meta‑analysis was key nonvertebral fractures (NVFs), which included clavicle, humerus, wrist, hip, pelvis, and leg. Patients were grouped according to the annual cumulative exposure (ACE) of ibandronate. The once‑monthly 150‑mg oral dose and the quarterly 3‑mg IV dose provide ACE within the high‑dose range; the once‑daily 2.5‑mg oral dose is in the low‑dose range.⁵¹ The high‑dose group (ACE ≥10.8 mg) showed significantly reduced risk relative to placebo for key NVFs (HR=0.656; 95% CI, 0.45‑0.96; P=.032), all NVFs (HR=0.701; 95% CI, 0.50‑0.99; P=.041), and clinical fractures (HR=0.730; 95% CI, 0.56‑0.95; P=.019). The high‑dose group also had a significantly longer time to fracture versus placebo for key NVFs (P=.031), all NVFs (P=.025), and clinical fractures (P=.002). The mid‑ and low‑dose groups did not show a significant reduction in fracture risk.⁵¹

Denosumab

Denosumab is a human monoclonal antibody that binds to the receptor activator of nuclear factor‑κB ligand (RANKL).⁵² RANK plays a key role in the signal transduction pathways that mediate osteoclast differentiation, activation, and survival. Denosumab binding blocks the interaction of RANKL with RANK, inhibiting this signaling pathway. Denosumab is administered subcutaneously twice a year. The efficacy and safety of denosumab was evaluated over 2 years in 412 postmenopausal women with low BMD.^52,53 Subjects were randomized to receive SC denosumab 6 mg, 14 mg, or 30 mg every 3 months; SC denosumab 14 mg, 60 mg, 100 mg, or 210 mg every 6 months; oral alendronate 70 mg once weekly; or placebo. The primary end point was change from baseline in lumbar spine BMD.

After 24 months of treatment, all doses of denosumab produced significant increases in lumbar spine BMD compared with placebo (P<.001; Figure 2).⁵³ The 60‑mg twice‑yearly dose was selected for phase III trials, and is therefore shown in the graph. The change in BMD with this dose of denosumab was comparable to that with alendronate at 2 years, with some greater benefit for denosumab after 1 year of treatment. Denosumab treatment was also associated with significantly greater improvements in hip and radius BMD compared with alendronate at 2 years (P=.001). Denosumab treatment showed significant decreases in the bone resorption marker serum C‑telopeptide (CTX) compared with pla-cebo over the 2‑year study period (P<.001; Figure 3).⁵³ There was partial reversal of this effect on CTX prior to each dose. It is hoped that these findings on bone turnover and bone density will translate to reduced fracture risk; results from the phase III trials are expected in late 2008.

Zoledronic Acid

Zoledronic acid is administered intravenously once annually. Such a long dosing interval is possible due to the high potency of this bisphosphonate, high binding affinity for bone, and probably local recirculation of the agent. In preclinical studies, zoledronic acid has shown a higher binding affinity for hydroxyapatite, more potent inhibition of farnesyl diphosphate synthase, and more potent inhibition of osteoclast‑mediated bone resorption than any other bisphosphonate.^29,30,54 In a dose‑finding trial of zoledronic acid, 351 post-menopausal women with low BMD were randomized to receive either placebo or IV zoledronic acid at doses of 0.25 mg, 0.5 mg, or
1 mg every 3 months; 2 mg every 6 months; or 4 mg once a year. As shown in Figure 4, the once‑ annual regimen effectively suppressed markers of bone turnover to within a normal premenopausal range over the entire 12‑month period.⁵⁵

HORIZON‑PFT

As a result of these findings, a phase III trial called Health Outcomes and Reduced Incidence with Zoledronic Acid Once Yearly—Pivotal Fracture Trial (HORIZON‑PFT) was designed to evaluate once‑yearly zoledronic acid for prevention of fractures in women with osteoporosis.⁵⁶ HORIZON‑PFT was a 3‑year, random-ized, double‑blind, placebo‑controlled study. Postmeno-pausal women (65 to 89 years of age) with osteoporosis (N=7736) were randomized to receive either an annual infusion of 5 mg zoledronic acid over 15 minutes or placebo at baseline, 12 months, and 24 months. All patients also received daily oral calcium and vitamin D. Patients were followed at 6, 12, 24, and 36 months. Patients were stratified on the basis of whether they were taking any other osteoporosis medications at baseline: stratum 1 included patients not taking any osteoporosis medications; stratum 2 was comprised of patients taking an allowed medication, which included hormone therapy or estrogen therapy, raloxifene, tibolone, or calcitonin (not documented to reduce hip fracture risk at the time of study design). Previous use of parathyroid hormone or strontium ranelate resulted in exclusion. Past bisphosphonate use was allowed with a washout period. The primary end points in HORIZON‑PFT were new morphometric vertebral fractures in stratum 1, and hip fracture in both strata. Secondary end points included nonvertebral and clinical vertebral fractures, and changes in BMD and markers of bone turnover.⁵⁶

Results

Efficacy

Results for the primary end points are presented in Table 1.⁵⁶ Within stratum 1, patients who received zoledronic acid had a 70% reduction in new vertebral fractures after 3 years compared with patients who received placebo (P<.001). Significant risk reduction was seen as early as 1 year and maintained throughout the 3‑year trial. Furthermore, there was a 41% reduction in the occurrence of hip fractures (P=.002). There were significant risk reductions for secondary end points as well, including clinical vertebral (77%) and nonvertebral fractures (25%) in both strata (P<.001).⁵⁶

Between stratum 1 and stratum 2, there are very consistent risk reductions both for hip and for nonvertebral fractures (Table 2): about 41% and 42% reduction for hip fracture, and 22% and 26% reduction for nonvertebral fracture. There was a suggestion of difference in the magnitude of risk reduction for clinical vertebral fractures—83% in stratum 1 compared with 66% in stratum 2—although the reduction within each stratum was significant (P<.001 and P=.0035, respect-ively). A slightly lower risk reduction in stratum 2 was expected, as these patients were receiving ongoing osteoporosis therapy known to reduce the risk of vertebral fractures. This was the first demonstration that another antiresorptive drug could reduce the incidence of clinical vertebral fracture on top of a reduction already apparent due to another drug.

Subgroup Analysis

A separate prospectively planned analysis was con-ducted of fracture risk reduction by subgroups in HORIZON‑PFT.⁵⁷ Subgroups included age distribution, disease severity at baseline as defined by prevalent fractures and BMD, and kidney function as defined by creatinine clearance. The original study was not powered to look at these subgroups.

Table 3 shows the reduced risk of vertebral fracture by subgroup. Compared with placebo, zoledronic acid significantly reduced risk by 64% to 72% across all subgroups of patients with different baseline prevalence of vertebral fractures (P<.0001). This finding was also true when patients were grouped according to baseline BMD. Patients with a baseline T‑score ≤‑2.5 and patients with T‑scores >‑2.5 had reduced risks of vertebral fractures of 72% and 65%, respectively (P<.0001 vs placebo). The effect of zoledronic acid against vertebral fracture was also consistent across age groups, and for patients with creatinine clearance
<60 mL/min and ≥60 mL/min (all P<.0001).

The risk of hip fracture in the same subgroups is shown in Table 4. There was some variation from group to group, but, overall, the consistency of effect was shown by relative risk reductions with zoledronic acid among all subgroups examined. Similarly, the incidence of any clinical fracture was lower with zoledronic acid across age groups (Table 5).

Safety

Bone safety and remodeling was monitored in the HORIZON‑PFT trial in a subgroup of patients. By the end of the study, iliac bone biopsies were obtained from 152 patients (82 zoledronic acid and 70 placebo), of which 111 were readable (59 zoledronic acid and
52 placebo).⁵⁸ Overall, the biopsy data showed ongoing bone formation in 81 of 82 biopsies that were evaluable for this end point (tetracycline label visible).

Figure 5 shows microcomputed tomography biopsies from 2 patients, 1 on placebo and the other treated with zoledronic acid.⁵⁸ The sample from the zoledronic acid patient shows a greater amount of bone and better preserved structure in both the cancellous region and the cortical compartment compared with the sample from the placebo patient. These differences were quantified in the biopsy population and many were found to be significantly different on average between the zoledronic acid and placebo groups (P<.05). There was a slight increase in mineral apposition rate among patients treated with zoledronic acid compared with those who received placebo (0.60 µm/day vs 0.53 µm/day; P<.001). This finding is somewhat surprising, but suggests at least that zoledronic acid did not suppress osteoblast activity. There was greater bone volume among subjects treated with zoledronic acid compared with the placebo group (median 16.6% vs 12.8%, respectively; P=.020), and there were no abnormalities on the bone biopsy specimens.⁵⁸

In HORIZON‑PFT, serious adverse events occurred in 29.2% of patients in the zoledronic acid group compared with 30.1% in the placebo group.⁵⁶ Among patients in the zoledronic acid group, 2.1% discontinued due to adverse events (AEs) compared with 1.8% in the placebo group. The total number of deaths was also similar between groups: 3.4% for zoledronic acid and 2.9% for placebo. None of these rates were significantly different between groups. The overall rate of any AE was significantly higher among zoledronic acid patients than among placebo patients (95.5% vs 93.9%; P=.002), mainly due to postdose symptoms following zoledronic acid infusion. These symptoms, which included fever, myalgia, flu‑like symptoms, headache, and arthralgia,⁵⁶ generally occur-red within 3 days of infusion, resolved within a few days, and were much more common after the first infusion than after subsequent administrations. Anti-inflammatory agents, such as ibuprofen and NSAIDs, are highly effective for reducing the incidence of these symptoms when given within a few hours of the infusion.

Renal safety was also monitored in the HORIZON‑PFT study. Within 9 to 11 days postinfusion, 1.3% of patients in the zoledronic acid group had a rise in serum creatinine levels of >0.5 mg/dL, compared with 0.4% of patients in the placebo group (P=.001).⁵⁶ This increase, however, was transient and within 30 days creatinine levels had returned to within 0.5 mg/dL of preinfusion levels in more than 85% of patients. The rest had returned to normal by the next annual follow‑up visit.
At 3 years, there was no significant difference in creatinine clearance.

Hypocalcemia (serum calcium <2.075 mmol/L) occurred in 49 patients in the zoledronic acid group within 9 to 11 days after the first infusion, compared with 1 patient in the placebo group. The cases were all transient, asymptomatic, and self‑resolving.

With regard to cardiac safety, atrial fibrillation was not significantly more common in the zoledronic acid group compared with placebo, but cases of serious atrial fibrillation requiring hospitalization occurred more frequently with zoledronic acid than with placebo
(1.3% vs 0.5%, respectively; P<.001).⁵⁶ Risk factors for serious atrial fibrillation in HORIZON‑PFT included congestive heart failure, tachyarrhythmia, age, and past bisphosphonate use (Figure 6).⁵⁹ The clinical sign-ificance of these findings is not yet clear for zoledronic acid or for other bisphosphonates that may be assoc-iated with atrial fibrillation.⁶⁰ Other studies of zoledronic acid have shown similar incidence rates of atrial fibril-ation between the placebo and zoledronic acid arms.⁶¹

There were no spontaneous reports of ONJ, but a subsequent search of the trial database found 2 potential cases: 1 in the zoledronic acid group and 1 in the placebo group. Both cases resolved after antibiotic therapy and debridement.⁵⁶

Summary

In women with postmenopausal osteoporosis, once‑ yearly infusion of zoledronic acid over 3 years was shown to reduce vertebral fractures (morphometric and clinical), hip fractures, and nonvertebral fractures. There is evidence of a consistent treatment effect across subgroups. Treatment with zoledronic acid was assoc-iated with reduced bone turnover and increased bone density. In conclusion, novel extended‑dosing regimens with excellent efficacy have the potential to make a major public health impact on osteoporosis due to substantially improved adherence and persistence.

Table 1. Relative Risk of Fracture in HORIZON‑PFT*⁵⁶

HORIZON‑PFT = Health Outcomes and Reduced Incidence with Zoledronic Acid Once‑Yearly—Pivotal Fracture Trial.
^*The percentage of morphometric fractures is based on the proportion of patients with a baseline radiograph, at least 1 follow‑up radiograph, and a fracture (2853 patients in the placebo group and 2822 patients in the zoledronic acid group). The percentage of clinical fractures is based on Kaplan‑Meier estimates of the 3‑year cumulative incidence (3875 patients with clinical fractures in the placebo group and 3861 in the zoledronic acid group).
^†For morphometric vertebral fractures, the relative risk is presented; for all other end points, the adjusted hazard ratio is presented.
Adapted with permission from Black DM et al. N Engl J Med. 2007;356:1809‑1822.
Copyright © 2007 Massachusetts Medical Society. All rights reserved.

Table 2. Fracture Risk Reduction by Stratum in HORIZON‑PFT

HORIZON‑PFT = Health Outcomes and Reduced Incidence with Zoledronic Acid Once‑Yearly—Pivotal Fracture Trial.
^*Zoledronic acid versus placebo.

Table 3. Risk of New Vertebral Fractures by Subgroup in HORIZON‑PFT

BMD = bone mineral density; HORIZON‑PFT = Health Outcomes and Reduced Incidence with Zoledronic Acid Once‑Yearly—Pivotal Fracture Trial.
^*Femoral neck BMD.

Table 4. Risk of Hip Fractures by Subgroup in HORIZON‑PFT

BMD = bone mineral density; HORIZON‑PFT = Health Outcomes and Reduced Incidence with Zoledronic Acid Once‑Yearly—Pivotal Fracture Trial.
^*Femoral neck BMD.

Table 5. Risk of Clinical Fractures by Age in HORIZON‑PFT

BMD = bone mineral density; HORIZON‑PFT = Health Outcomes and Reduced Incidence with Zoledronic Acid Once‑Yearly—Pivotal Fracture Trial.

Figure 1. Mean change from baseline in lumbar spine BMD after 1 and 2 years of treatment with ibandronate in DIVA study.^49,50

Per‑protocol analysis.
BMD = bone mineral density; DIVA = Dosing IntraVenous Administration.
^*P<.05 vs 2.5 mg daily oral ibandronate.
^†P<.001 vs 2.5 mg daily oral ibandronate.

Figure 2. Mean change from baseline in lumbar spine BMD after 2 years of treatment with denosumab 60 mg SC every 6 months, alendronate 70 mg once weekly, or placebo.⁵³

BMD = bone mineral density; SC = subcutaneous.
^*P<.001 vs placebo.
Adapted with permission from Lewiecki EM et al. J Bone Miner Res. 2007;22:1832-1841.

Figure 3. Median change in serum CTX over 24 months of treatment with denosumab 60 mg SC every 6 months, alendronate 70 mg once weekly, or placebo.⁵³

CTX = C‑telopeptide; SC = subcutaneous.
^*P<.001 vs placebo.
Adapted with permission from Lewiecki EM et al. J Bone Miner Res. 2007;22:1832-1841.

Figure 4. Effect of once‑yearly zoledronic acid on markers of bone resorption and bone formation.⁵⁵

BSAP = bone‑specific alkaline phosphatase; NTx = N‑telopeptide; ZA = zoledronic acid.
Adapted with permission from Reid IR et al. New Engl J Med. 2002;346:653‑661.
Copyright © 2002 Massachusetts Medical Society. All rights reserved.

Figure 5. Bone biopsies from a patient treated with zoledronic acid 5 mg and a patient treated with placebo in the HORIZON‑PFT study. The figure shows microcomputed tomography renditions of the whole biopsy core (left panel), thick sections (middle panel), and thin sections (right panel).58

HORIZON‑PFT = Health Outcomes and Reduced Incidence with Zoledronic Acid Once‑Yearly—Pivotal Fracture Trial. PBO = placebo; ZA = zoledronic acid.
Reprinted with permission from Recker RR et al. J Bone Miner Res. 2008;23:1‑16.

Figure 6.Risk factors for serious atrial fibrillation in the HORIZON‑PFT study.⁵⁹

HORIZON‑PFT = Health Outcomes and Reduced Incidence with Zoledronic Acid Once‑Yearly—Pivotal Fracture Trial.
*Assignment to zoledronic acid vs placebo.

THE POST–HIP FRACTURE PATIENT: NOVEL INTERVENTIONS

Kenneth W. Lyles, MD

Although hip fractures are less common than other types of osteoporotic fractures, they are associated with considerable morbidity and mortality and overall use of healthcare resources compared with other types of fracture.^1,62 Mortality may be increased by as much as 20% to 25% over the 5 years following a hip fracture, with the greatest increase in risk during the first 6 to
12 months.^62‑64 There is also substantial functional dis-ability following a hip fracture,⁶⁵ and up to half of all subjects never regain their former level of function.^66,67 The incidence of nursing home admissions may be up to 50% higher in the year following a hip fracture as well. Approximately 15% of hip‑fracture patients will remain in a nursing home for the rest of their lives.⁶⁸ The risk of a new clinical fracture is 2.5 times greater after a hip fracture, independent of risk factors present before the fracture.⁶⁹ Bone mineral density (BMD) shows a significant decline following hip fracture compared with matched controls, likely due to reduced activity and hyperparathyroidism, among other reasons.⁷⁰ The many serious consequences of hip fracture emphasize the importance of optimizing therapies to treat this important clinical concern.

The Health Outcomes and Reduced Incidence With Zoledronic Acid Once Yearly—Recurrent Fracture Trial (HORIZON‑RFT): Study Design

HORIZON‑RFT was a double‑blind, placebo‑controlled, randomized clinical trial, conducted in 148 clinical cen-ters in 23 countries throughout the world. This trial was designed to investigate the efficacy of zoledronic acid for the prevention of osteoporotic fractures of the hip.⁶¹ Subjects were 2127 men and women, aged 50 years or older, with a recent (within 90 days) surgical procedure for a low‑trauma hip fracture. Subjects had to have been ambulatory prior to the hip fracture, and had to be unwill-ing or unable to take oral bisphosphonates. Exclusion criteria included current use of oral bisphos-phonates or use of intravenous (IV) bisphosphonates within the previous 2 years, and any use of teriparatide or its analogs for more than 1 week. Subjects also could not have a calculated creatinine clearance of <30mL/min, and could not have hypercalcemia, hypocalcemia, or metabolic bone disease.⁶¹

Study treatments were an annual infusion of zoledronic acid, 5 mg over 15 minutes, or a placebo solution admin-istered in the same way. Almost all subjects received a single loading dose of 50,000 to 125,000 international units (IU) of vitamin D 2 weeks before they received their study drug. During the study period, subjects received calcium supplements, 1000 to 1500 mg a day, and vitamin D supplementation, 800 to 1200 IU daily. Patients could receive concomitant osteoporosis therapy with the exception of teriparatide or bisphosphonates. Permissable medications included selective estrogen receptor modulators, estrogen therapy, and calcitonin therapy.⁶¹

The primary end point of HORIZON‑RFT was the rate of new clinical fractures. Secondary objectives included the rates of clinical vertebral, hip, and nonvertebral fractures, and the levels of BMD at the hip and femoral neck in the nonfractured hip. HORIZON‑RFT was an event‑driven trial, with 90% power to detect a 20% reduction in fracture rates with vitamin D supplement, and a 35% reduction in fracture rates after the hip fracture with the use of zoledronic acid. Follow‑up was done at 6, 12, 24, and 36 months.⁶¹

HORIZON‑RFT: Results

Patients

The 2 treatment groups were well balanced in terms of baseline demographics (Table 1).⁶¹ The study population was 91% Caucasian and 24% male. The mean age in the study was 74 years. At baseline, femoral neck BMD and total hip BMD were comparable in both groups of patients. Among all patients, 41.8% had a T‑score <‑2.5 at the femoral neck.

Outcomes

Figure 1 shows the Kaplan‑Meier curve for new clinical fractures.⁶¹ Treatment with zoledronic acid was assoc-iated with a 35% reduction in the relative risk of new fractures (HR=0.65; 95% CI, 0.50‑0.84; P=.001). Absolute risk reduction was 5.3% among patients who had a fracture, with a mean time to clinical fracture of 39.8 months in the zoledronic acid group and 36.4 months in the placebo group. Secondary variables are shown in Table 2.⁶¹ Compared with placebo, nonvert-ebral fractures were reduced by 27% (P=.03) with zoledronic acid, and clinical vertebral fractures by 46% (P=.02). Hip fractures were reduced by 30%. However, there was a low occurrence of hip fractures overall in this trial, which contributed to the lack of statistical signif-icance for this decrease.

There was a significant 28% reduction in the risk of death among patients who received zoledronic acid compared with those who received placebo (P=.01) [Figure 2].⁶¹ There were 101 deaths in the zoledronic acid group compared with 141 in the placebo group. The reasons for the reduced risk of death are not fully understood, but the decrease was due at least in part to the reduction in fractures. In the zoledronic acid group, 11 deaths (1.0%) were from cardiovascular disease and 7 deaths (0.7%) were from cerebrovascular disease, compared with 18 deaths (1.7%) and 7 deaths (0.7%), respectively, in the placebo g