Method. Patients were randomized to receive either intravenous zoledronic acid 5 mg or placebo within 90 days of surgical repair of a hip fracture and yearly thereafter for the duration of the study. Patients with serum 25-hydroxyvitamin D levels <15 ng/mL received a loading dose of vitamin D (50,000–125,000 IU) 14 days before the first dose of the study drug. All patients were given daily calcium (1000–1500 mg) and vitamin D (800–1200 IU) supplements. Simultaneous treatment with nasal calcitonin, selective estrogen-receptor modulators, hormone replacement, tobolone, and external hip protectors was permitted “at the discretion of the investigators,” and 10.5% of the study patients did receive one of these other osteoporosis therapies.
Patients were followed for up to 5 years. Bone mineral density was tested by DXA at the hip and femoral neck at baseline and then annually. The median follow-up time was 1.9 years; 71.3% of patients completing the trial, and 3% of patients were lost to follow-up.
Results. The patients assigned to zoledronic acid had a 5.3% absolute risk reduction for new clinical fractures, yielding an NNT of 19 over 2 years to prevent one new clinical fracture ( TABLE 1 ). Bone mineral density of the contralateral hip increased in the zoledronic acid group by 2.6% after 1 year, 4.7% after 2 years, and 5.5% after 3 years and declined in the placebo group by 1.0%, 0.7%, and 0.9% respectively.
There was a 3.7% absolute risk reduction of death, with an NNT of 27 for 2 years to prevent one death.
Adverse events that were more common in the zoledronic acid group included fever (8.7% vs 3.1%) and musculoskeletal pain (3.1% vs. 1.2%). There were no reported cases of jaw osteonecrosis in either group and no statistically significant delay in the healing of fractures with zoledronic acid. Both groups had similar rates of renal and cardiovascular events, including atrial fibrillation and stroke.
Novartis provided funding for the study. An independent data and safety monitoring board oversaw the conduct and safety of the study and recommended that it be stopped early after having surpassed the pre-specified efficacy boundaries. Independent statisticians confirmed the data analysis that was performed by the sponsor.1
TABLE 1
The HORIZON study: Rates of fracture and death were lower in the zoledronic group compared to placebo1
| OUTCOME | PLACEBO (N=1062) | ZOLEDRONIC ACID (N=1065) | HAZARD RATIO (95% CI) | P VALUE |
|---|---|---|---|---|
| Any fracture | 139 (13.9%) | 92 (8.6%) | 0.65 (0.50-0.84) | .001 |
| Hip fracture | 33 (3.5%) | 23 (2.0%) | 0.70 (0.41-1.19) | .18 |
| Vertebral fracture | 39 (3.8%) | 21 (1.7%) | 0.54 (0.32–0.92) | .02 |
| Death | 141 (13.3%) | 101 (9.6%) | 0.72 (0.56-0.93) | .01 |
| Rates of fracture were calculated by Kaplan-Meier methods at 24 months and are not simple percentages | ||||
What’s new: The adherence advantage
The obvious advantage of zoledronic acid over other bisphosphonates is the high level of adherence that is possible under the controlled environment of a once yearly infusion administered under medical supervision. Considering the low rates of adherence to oral bisphosphonates, this is a significant medical advance.
This study shows that a yearly infusion of zoledronic acid is highly effective in preventing subsequent clinical fractures in patients who have recently suffered a hip fracture. It is the first randomized-controlled trial of a bisphosphonate for secondary prevention in patients with recent hip fracture, regardless of their bone mineral density status.
In a previous 3-year randomized controlled trial of 5mg zoledronic acid once yearly vs placebo for postmenopausal women with osteoporosis, the risk of vertebral fractures was reduced by 70% (3.3% vs 10.9% placebo) and the risk of hip fracture was reduced by 41% (1.4% vs 2.5% placebo).9
The NNT of 19 for 2 years to prevent one clinical fracture and NNT of 27 for 2 years to prevent one death strikes us as a good bargain compared to many modern medical interventions. Are these results too good to be true? We don’t think so.
No serious design flaws
This was a well done trial with no serious flaws in design. The number of deaths, however, was relatively small, so the NNT may be as high as 50 by our rough calculations. As a “worst case” for the benefit, however, this still seems worthwhile.
Caveats
Patients with uncorrected hypocalcemia or creatinine clearance <35 mL/min should not take bisphosphonates, including zoledronic acid. Although this study did not show any evidence of an increased risk of atrial fibrillation or report any cases of osteonecrosis of the jaw, providers should monitor patients for these potential adverse events.
