What Happened After HORIZON-PFT Ended? Extension Data, Durability, and Long-Term Safety of Yearly IV Zoledronic Acid

At a glance

| Field | Detail | |---|---| | Trial | HORIZON-PFT (Health Outcomes and Reduced Incidence with Zoledronic Acid Once Yearly, Key Fracture Trial) | | Original N | 7,765 postmenopausal women with osteoporosis | | Extension N | 1,233 women re-randomized after 3 years of zoledronic acid | | Intervention | Zoledronic acid 5 mg IV once yearly | | Comparator (extension) | 3 additional years of zoledronic acid vs. switch to placebo | | Duration | 6 years total (3-year core + 3-year extension) | | Primary endpoint (core) | New morphometric vertebral fracture at 3 years | | Key result (core) | 70% relative risk reduction in vertebral fracture (3.3% vs. 10.9%) | | Key result (extension) | Continuing zoledronic acid to 6 years reduced new vertebral fractures vs. stopping at 3 years (3.0% vs. 6.2%) |

Why an Extension Study Mattered

The original HORIZON-PFT trial established zoledronic acid as the most potent antiresorptive bisphosphonate available at the time. A 70% vertebral fracture reduction and a 41% hip fracture reduction over three years were striking. But clinicians immediately faced a practical question: what happens when you stop?

Oral bisphosphonates like alendronate had some residual data from the FLEX trial suggesting that stopping after five years was reasonable for many patients. Zoledronic acid's longer skeletal half-life (estimated at over 10 years for the bisphosphonate bound in bone) raised the possibility that shorter treatment courses might provide durable benefit. Conversely, there were legitimate concerns about rare but serious adverse events accumulating with prolonged exposure, including osteonecrosis of the jaw and atypical femoral fractures.

The extension study, published by Black et al. in 2012, directly tested whether continuing annual infusions beyond three years offered meaningful additional protection.

Extension Study Design

After the core trial ended, 1,233 women who had received zoledronic acid for three years consented to re-randomization. Half continued with three more annual infusions (Z6 group, total six years of treatment). The other half switched to placebo for years 4 through 6 (Z3P3 group, three years of treatment followed by three years off).

Women originally randomized to placebo in the core trial received three years of open-label zoledronic acid during the extension, forming a delayed-treatment group that provided additional safety data but was not part of the primary efficacy comparison.

Key design features:

• Double-blind, placebo-controlled re-randomization
• Spine radiographs at year 6 compared to year 3
• BMD measured at total hip, femoral neck, and lumbar spine
• Bone turnover markers (serum CTX, P1NP) tracked annually

The HealthRX Durability Framework: Parsing What "Residual Protection" Actually Means

Understanding the extension results requires separating three distinct phenomena that get conflated in clinical summaries:

1. Pharmacologic persistence refers to the drug still sitting in bone and still suppressing resorption. CTX levels in the Z3P3 group remained 15-20% lower than pretreatment values even at year 6, confirming that zoledronic acid's skeletal binding produces measurable pharmacologic activity years after the last infusion.

2. BMD retention refers to whether the bone density gained during treatment erodes after stopping. Total hip BMD declined modestly in the Z3P3 group (about 1.4% loss from year 3 to year 6) while continuing to rise slightly in the Z6 group. Lumbar spine BMD remained essentially stable in both groups.

3. Fracture protection is the outcome that matters. BMD retention does not automatically equal fracture protection, and the extension showed exactly this complexity. Morphometric vertebral fractures were significantly higher in the Z3P3 group (6.2% vs. 3.0%, OR 1.86 to 95% CI 1.12-3.08). But clinical (symptomatic) vertebral fractures, nonvertebral fractures, and hip fractures did not differ significantly between groups.

This three-layer distinction matters for clinical decision-making. A patient whose primary concern is hip fracture has a different risk calculus than one with prior vertebral compression fractures.

Results in Detail

Fracture Outcomes at Year 6

| Outcome | Z6 (continued) | Z3P3 (stopped) | Odds Ratio (95% CI) | P value | |---|---|---|---|---| | New morphometric vertebral fracture | 3.0% | 6.2% | 1.86 (1.12-3.08) | 0.02 | | Clinical vertebral fracture | 0.5% | 1.2% |, | NS | | Any nonvertebral fracture | 9.4% | 10.2% |, | NS | | Hip fracture | 1.3% | 1.4% |, | NS | | Any clinical fracture | 11.9% | 13.1% |, | NS |

The morphometric vertebral fracture difference was the primary finding. These are fractures detected on scheduled radiographs. Many are asymptomatic. The clinical significance of purely radiographic vertebral fractures is debated, though they do predict future symptomatic fractures and progressive height loss.

Bone Mineral Density Trajectories

| Site | Z6 change (year 3 to 6) | Z3P3 change (year 3 to 6) | |---|---|---| | Total hip | +0.24% | -1.36% | | Femoral neck | +0.54% | -0.87% | | Lumbar spine | +1.29% | +0.67% |

The BMD differences were statistically significant at the hip but the absolute magnitudes were small. Lumbar spine BMD was well preserved in both groups. This is consistent with the known pharmacokinetics of zoledronic acid: the drug accumulates preferentially in trabecular bone (spine) and is released slowly enough to maintain density there even after stopping.

Bone Turnover Markers

Serum CTX rose in the Z3P3 group after the last infusion but did not return to pretreatment levels by year 6. P1NP (a formation marker) showed a similar partial rebound. In the Z6 group, both markers remained suppressed. This confirmed ongoing pharmacologic activity in the stopped group, just at a reduced level, supporting the concept of a "partial drug holiday" rather than complete washout.

Safety Signals From the Extension and Beyond

Atypical Femoral Fractures

No confirmed atypical femoral fractures (AFFs) occurred in the HORIZON extension. However, AFFs are extremely rare events (estimated at 3.2-50 per 100,000 person-years of bisphosphonate use) and the extension enrolled only ~1,200 women. The study was not powered to detect this signal. Subsequent observational studies and the 2010 ASBMR task force report confirmed a duration-dependent association between bisphosphonates and AFFs, with risk rising after 3-5 years of use and declining after discontinuation. This became one of the strongest arguments for drug holidays.

Osteonecrosis of the Jaw

One case of ONJ was reported in the extension (in the Z6 group). The original HORIZON-PFT core trial reported one delayed-healing oral event that was later adjudicated as possible ONJ. These numbers are consistent with the very low incidence of ONJ in osteoporosis patients (approximately 1 in 10,000 to 1 in 100,000 patient-years) as opposed to the higher rates seen with oncologic dosing of IV bisphosphonates.

Renal Safety

Zoledronic acid is contraindicated in patients with creatinine clearance <35 mL/min. In the extension, transient serum creatinine elevations occurred more frequently with active treatment, but no cases of renal failure were attributed to the drug. Post-marketing surveillance and the FDA-approved label for Reclast added warnings about acute renal impairment, particularly in patients with pre-existing renal insufficiency or dehydration at the time of infusion.

Acute Phase Reactions

The flu-like reaction (fever, myalgia, arthralgia) that affected roughly 30% of patients after the first infusion in the core trial was markedly reduced with subsequent infusions. By year 4 in the extension, fewer than 5% of patients reported acute phase symptoms. Pre-treatment with acetaminophen further attenuated this effect.

What the Extension Could Not Answer

The HORIZON extension had meaningful limitations that shape how we should interpret the findings.

Sample size attrition. Only 1,233 of the original 3,889 zoledronic acid patients entered the extension. Selection bias is inevitable: healthier, more adherent patients are more likely to continue in a trial. The extension population may not represent the full range of patients who received zoledronic acid in the core study.

Three years may not be the right stopping point for comparison. The FLEX trial of alendronate tested five vs. ten years. Zoledronic acid's greater potency and longer half-life mean the optimal treatment duration could differ substantially from oral agents. Three years was the default stopping point because that was how long the core trial ran, not because pharmacologic reasoning suggested it was optimal.

Morphometric vs. clinical fractures. The statistically significant finding was in morphometric vertebral fractures, not clinical fractures. Whether preventing asymptomatic radiographic fractures justifies three additional years of an IV infusion is a clinical judgment call, not a question the trial can answer definitively.

No head-to-head against oral bisphosphonate holidays. We still lack direct comparison data between a zoledronic acid holiday and an alendronate holiday. Clinical decisions rely on indirect comparisons across trials with different populations and different durations.

How These Data Shaped Clinical Practice

The HORIZON extension data, combined with FLEX (alendronate) and FREEDOM extension (denosumab), formed the evidence base for current American Association of Clinical Endocrinology guidelines and Endocrine Society recommendations on bisphosphonate holidays.

The consensus that emerged:

• Patients at moderate fracture risk after 3 years of zoledronic acid can consider a drug holiday with monitoring (BMD every 2-3 years, clinical fracture surveillance).
• Patients at high fracture risk (T-score still below -2.5 at the hip, prior vertebral fracture, or age over 70 with additional risk factors) should generally continue treatment or switch to an anabolic agent.
• If BTMs (bone turnover markers) rise substantially during a holiday, restarting therapy is recommended even if BMD has not declined significantly.

This approach remains the clinical standard as of 2026, though a growing body of evidence suggests that sequential therapy (anabolic agent first, then bisphosphonate to maintain gains) may be superior to prolonged bisphosphonate monotherapy in high-risk patients.

The Denosumab Contrast

The HORIZON extension is best understood alongside the FREEDOM extension of denosumab. Unlike bisphosphonates, denosumab does not bind to bone and its offset is rapid. Stopping denosumab leads to rebound bone turnover, accelerated BMD loss, and an increased risk of multiple vertebral fractures. This created clinical urgency around denosumab discontinuation that does not exist with zoledronic acid.

The practical result: zoledronic acid's skeletal persistence, once seen as a theoretical concern, became a clinical advantage. Patients who need to stop therapy (for cost, preference, or safety reasons) face a gentler transition with zoledronic acid than with denosumab.

Frequently asked questions

›

›

›

›

›

›

›

›

›

›

References

1. Black DM, Delmas PD, Eastell R, et al. Once-yearly zoledronic acid for treatment of postmenopausal osteoporosis. N Engl J Med. 2007;356(18):1809-1822. PubMed
2. Black DM, Reid IR, Boonen S, et al. The effect of 3 versus 6 years of zoledronic acid treatment of osteoporosis: a randomized extension to the HORIZON-Key Fracture Trial (PFT). J Bone Miner Res. 2012;27(2):243-254. PubMed
3. Shane E, Burr D, Abrahamsen B, et al. Atypical subtrochanteric and diaphyseal femoral fractures: second report of a task force of the American Society for Bone and Mineral Research. J Bone Miner Res. 2014;29(1):1-23. PubMed
4. Reclast (zoledronic acid) prescribing information. Novartis Pharmaceuticals. FDA Label
5. Camacho PM, Petak SM, Binkley N, et al. American Association of Clinical Endocrinology/American College of Endocrinology clinical practice guidelines for the diagnosis and treatment of postmenopausal osteoporosis, 2020 update. Endocr Pract. 2020;26(Suppl 1):1-46. PubMed
6. Bone HG, Wagman RB, Brandi ML, et al. 10 years of denosumab treatment in postmenopausal women with osteoporosis: results from the phase 3 randomised FREEDOM trial and open-label extension. Lancet Diabetes Endocrinol. 2017;5(7):513-523. PubMed