How long can you safely take denosumab based on extension data?

The FREEDOM extension provides controlled data for up to 10 years of continuous use. Fracture rates remained low and BMD continued increasing throughout. ONJ and AFF events were rare. No pre-specified stopping rule exists, and many guidelines support indefinite treatment in patients at persistently high fracture risk.

What happens to bone density when you stop denosumab?

BMD at the lumbar spine and hip returns to near or below pre-treatment levels within 12 to 24 months of stopping. Bone turnover markers rebound above baseline, creating a temporary window of accelerated bone loss. This effect is more pronounced after longer treatment durations and higher cumulative BMD gains.

Are rebound fractures after denosumab discontinuation common?

In the FREEDOM post hoc analysis, approximately 6.2% of women who discontinued denosumab sustained vertebral fractures within 12 to 24 months. Multiple simultaneous vertebral fractures occurred at a rate higher than expected from the natural disease course. This risk is specific to denosumab and not observed with bisphosphonate discontinuation.

Can you switch from denosumab to a bisphosphonate safely?

Yes, and guidelines recommend it. A single dose of IV zoledronic acid given approximately 6 months after the last denosumab injection prevents most of the rebound bone loss. Some patients need a second dose. Oral alendronate is an alternative, though data are less strong. The transition should be planned before stopping denosumab, not improvised afterward.

Did the FREEDOM extension find increased infection risk with denosumab?

The parent trial showed a small numerical increase in serious infections (cellulitis, erysipelas) with denosumab. The extension did not show a progressive increase in infection rates over 10 years. Current guidelines do not contraindicate denosumab based on infection risk, though RANKL's role in immune function remains a theoretical concern.

How does the 10-year ONJ rate with denosumab compare to bisphosphonates?

The cumulative ONJ rate in the FREEDOM extension was about 5.2 per 10,000 patient-years, which is comparable to rates reported with long-term oral bisphosphonate use. All cases in the extension were associated with dental procedures and resolved with conservative management.

Is denosumab still considered first-line after the extension data?

Yes, for patients at high fracture risk who are prepared for long-term treatment. The 2020 AACE/ACE and Endocrine Society guidelines list denosumab as a first-line option for postmenopausal osteoporosis. The extension data strengthened its efficacy profile but added important caveats about treatment commitment and discontinuation planning.

Why can't you take a 'drug holiday' from denosumab like you can with bisphosphonates?

Bisphosphonates bind to bone mineral and continue suppressing resorption for months to years after the last dose. Denosumab is a circulating antibody that clears within weeks. Once it is gone, osteoclast activity rebounds rapidly because there is no skeletal reservoir of drug to buffer the transition. This pharmacological difference makes unsupervised drug holidays unsafe with denosumab.

Were atypical femoral fractures common in the 10-year extension?

Only 2 confirmed atypical femoral fractures occurred across the entire 10-year extension. This is a very low incidence given the sample size and duration, though the risk may continue to increase beyond 10 years. Current data do not support a mandatory stopping point based on AFF risk alone.

Did the crossover group from placebo catch up in bone density?

Women who switched from placebo to denosumab at year 4 gained BMD at a rate comparable to what the long-term group achieved during their first years of treatment. By 7 years of active treatment, the crossover group had lumbar spine gains of approximately 9.4% from their crossover baseline, suggesting prior untreated bone loss did not permanently impair the treatment response.

FREEDOM Extension Data and What Happened After the Trial Ended

By HealthRX.com Medical Team

Published May 25, 2026Updated May 25, 2026Last reviewed May 25, 2026

Clinical image for FREEDOM Extension Data and What Happened After the Trial Ended Image: HealthRX.com clinical image

FREEDOM Extension Data and What Happened After the Trial Ended

At a glance

| Detail | Value | |---|---| | Original trial N | 7,868 postmenopausal women with osteoporosis | | Extension N | 4,550 (enrolled); 2,626 completed 10 years | | Intervention | Denosumab 60 mg SC every 6 months | | Comparator | Original placebo group crossed over to denosumab at year 4 | | Duration | Up to 10 years total (3-year parent + 7-year extension) | | Primary endpoint (parent) | New vertebral fractures at 36 months | | Key result (parent) | 68% relative risk reduction in vertebral fractures (Cummings et al., NEJM 2009) | | Key extension result | Lumbar spine BMD increased 21.7% over 10 years with continuous treatment |

Why the Extension Mattered More Than the Original Trial

The parent FREEDOM trial answered the efficacy question: denosumab cut vertebral fractures by 68%, hip fractures by 40%, and nonvertebral fractures by 20% over three years compared with placebo. That was enough for FDA approval in 2010. What the parent trial could not answer was whether those benefits would hold beyond three years, whether rare adverse events would accumulate, and what would happen when patients stopped the drug.

The FREEDOM Extension (published across several reports from 2012 through 2017) enrolled 4,550 women from the parent trial. Women originally randomized to denosumab continued treatment (the "long-term" group, up to 10 years of exposure). Women originally on placebo crossed over to active denosumab (the "crossover" group, up to 7 years of exposure). There was no longer a placebo arm, which means efficacy comparisons relied on within-group trends and historical fracture rates rather than concurrent controls.

Bone et al. reported the full 10-year dataset in the Lancet Diabetes & Endocrinology (2017), providing the longest continuous denosumab exposure data available for postmenopausal osteoporosis.

BMD Trajectory: Gains That Kept Accumulating

Unlike bisphosphonates, which typically plateau in BMD effect after 3 to 5 years, denosumab showed continued BMD accrual throughout the extension period.

| Timepoint | Lumbar Spine BMD Change (from parent baseline) | Total Hip BMD Change | |---|---|---| | 3 years (end of parent trial) | +9.2% | +6.0% | | 5 years | +13.7% | +7.0% | | 8 years | +18.4% | +8.3% | | 10 years | +21.7% | +9.2% |

The crossover group (originally placebo) showed a similar trajectory once started on denosumab, gaining about 9.4% at the lumbar spine by year 7 of their active treatment. These gains tracked closely with what the long-term group had achieved at the same duration of exposure, suggesting that prior placebo assignment did not permanently impair the bone's response to denosumab.

Bone turnover markers (serum CTX and P1NP) remained suppressed for the full 10 years. CTX stayed approximately 80% below baseline, and P1NP remained about 50% to 60% below baseline, with no evidence of "escape" or tachyphylaxis.

Fracture Rates During Long-Term Treatment

Without a concurrent placebo control in the extension, fracture incidence could not be expressed as a relative risk reduction. Instead, the investigators reported annualized fracture rates and compared them against the parent trial rates and against expected rates from epidemiological cohorts.

| Fracture type | Annual rate, extension years 4-10 | Annual rate, parent trial (denosumab arm, years 1-3) | |---|---|---| | New vertebral | 1.0-1.5% | 2.3% | | Nonvertebral | 1.4-2.0% | 6.5% (cumulative 3-yr) | | Hip | 0.1-0.2% | 0.7% (cumulative 3-yr) |

Vertebral fracture incidence in the extension years remained below what would be expected based on the cohort's age and T-scores. The authors noted that fracture rates did not increase over time despite the cohort aging by a decade, which they interpreted as evidence of sustained anti-fracture protection. Critics pointed out that this cohort was self-selected (women who tolerated the drug and stayed in the trial), making direct comparisons with population-based fracture rates unreliable.

The Discontinuation Problem: Rebound Vertebral Fractures

The most consequential finding from the FREEDOM program did not come from the extension itself but from analyses of women who discontinued denosumab during or after the study.

Cummings et al. (2018) published a post hoc analysis in JBMR showing that BMD gains reversed rapidly after the last denosumab dose. Within 12 months of discontinuation, BMD at the lumbar spine dropped to near or below the original baseline, and bone turnover markers rebounded to levels above their pre-treatment values, a phenomenon termed "rebound resorption."

More alarming: women who stopped denosumab had a significantly higher rate of multiple vertebral fractures than women who had never received the drug. Among discontinuers, 6.2% sustained vertebral fractures in the 12 to 24 months after their last dose, compared with approximately 2% per year in the original placebo group. The clustering of multiple vertebral fractures (two or more at once) was particularly striking and not seen at this frequency with bisphosphonate discontinuation.

Why Rebound Happens

Denosumab is a monoclonal antibody against RANKL with a circulating half-life of approximately 25 to 28 days. Once the drug clears, osteoclast precursors that were suppressed during treatment differentiate rapidly and begin resorbing bone. The bone that was deposited during treatment may be inadequately mineralized or structurally vulnerable to this sudden increase in resorptive activity. The net effect is an accelerated bone loss that overshoots the pre-treatment trajectory.

This mechanism differs fundamentally from bisphosphonates, which bind to the hydroxyapatite mineral surface and continue to suppress resorption for months to years after the last dose. That persistent skeletal reservoir is why bisphosphonate "drug holidays" are feasible. Denosumab has no such reservoir.

The Sequencing Problem: What to Do After Stopping

The rebound data created an immediate clinical problem. Denosumab was approved as a long-term therapy, but indefinite treatment raises concerns about rare cumulative risks (discussed below). Clinicians needed a safe off-ramp.

Multiple small studies and expert consensus from ASBMR, the Endocrine Society, and the European Calcified Tissue Society converged on the same recommendation: transition patients to a bisphosphonate (typically zoledronic acid 5 mg IV or oral alendronate) within 6 months of the last denosumab injection to consolidate BMD gains and prevent rebound.

Optimal timing remains debated. Some data suggest administering zoledronic acid 6 months after the last denosumab dose (when bone turnover markers start to rise), while other protocols give it at month 7 or 8. A 2020 randomized study by Everts-Graber et al. found that a single infusion of zoledronic acid at 6 months prevented most but not all of the BMD loss seen with unprotected discontinuation. Some patients required a second infusion.

The clinical takeaway: denosumab is effectively a commitment. Starting it without a plan for transition or indefinite continuation creates avoidable risk.

Long-Term Safety: ONJ, AFF, and Infection Signals

Osteonecrosis of the Jaw (ONJ)

No cases of ONJ were reported during the 3-year parent FREEDOM trial. During the extension, 13 cases of confirmed ONJ occurred among women treated with denosumab. Most cases were associated with invasive dental procedures (tooth extraction, implant placement). The cumulative incidence was approximately 5.2 per 10,000 patient-years of exposure, consistent with the rate seen with long-term bisphosphonate use. All 13 cases resolved with conservative management (antibiotics, debridement), and none required major surgical intervention.

Atypical Femoral Fractures (AFF)

Two confirmed AFFs occurred during the 10-year extension period. This very low incidence is notable given the sample size and duration. Whether the risk increases beyond 10 years of exposure is unknown, as no controlled data exist past this horizon. The FDA label for denosumab (Prolia) includes AFF as a warning based on postmarketing reports rather than trial data.

Infections and Malignancy

The parent trial reported a small numerical excess of serious infections (including skin infections like cellulitis and erysipelas) in the denosumab arm. RANKL is expressed on immune cells, and its inhibition could theoretically impair innate immunity. In the extension, the overall incidence of serious infections did not increase over time, and no clear pattern emerged. The Endocrine Society's 2019 guidelines concluded that the infection signal was not strong enough to change prescribing behavior, though it remains a monitored concern.

Malignancy rates in the extension were comparable to age-matched population norms. No signal for increased cancer risk appeared across any organ system over 10 years.

What the Extension Could Not Answer

Several questions remain unresolved after the FREEDOM program.

First, the extension was open-label with no concurrent placebo arm. All efficacy estimates rely on within-group comparisons and historical controls, which are inherently weaker evidence than the randomized design of the parent trial.

Second, the extension suffered substantial attrition: of 4,550 women enrolled, only 2,626 (58%) completed the full 10 years. Dropouts may have differed systematically from completers (sicker, less tolerant, or less adherent), introducing survivorship bias that could inflate the apparent safety and efficacy of long-term treatment.

Third, the extension did not include men, premenopausal women, or patients with secondary osteoporosis (glucocorticoid-induced, transplant-related). Long-term denosumab data in those populations remain limited.

Fourth, the rebound fracture analyses were post hoc and observational. A prospective, randomized discontinuation study with pre-specified transition protocols has not been conducted, though it would be considered unethical to randomize patients to unprotected discontinuation given what is now known.

How Extension Data Changed Clinical Practice

The FREEDOM extension reshaped denosumab prescribing in three specific ways.

Treatment duration conversations changed. Clinicians now discuss denosumab as a long-term or indefinite commitment at the point of initiation, not after several years of treatment. The 2020 AACE/ACE guidelines specifically recommend against starting denosumab without a plan for either continuation or bisphosphonate transition.

Transition protocols became standard of care. Before the rebound data emerged, many clinicians simply stopped denosumab after 5 years (similar to bisphosphonate holiday protocols). That practice is now recognized as potentially dangerous. Bisphosphonate bridging after denosumab discontinuation is endorsed by ASBMR, the Endocrine Society, and most national guidelines.

Risk-benefit stratification shifted. For patients at very high fracture risk who need rapid, reliable BMD improvement, denosumab remains a first-line option. For patients with moderate risk who may not want a long-term injectable commitment, bisphosphonates are often preferred as initial therapy precisely because they allow drug holidays without rebound.

Frequently asked questions

›

References

Cummings SR, San Martin J, McClung MR, et al. Denosumab for prevention of fractures in postmenopausal women with osteoporosis. N Engl J Med. 2009;361(8):756-765. PubMed
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
Cummings SR, Ferrari S, Eastell R, et al. Vertebral fractures after discontinuation of denosumab: a post hoc analysis of the randomized placebo-controlled FREEDOM trial and its extension. J Bone Miner Res. 2018;33(2):190-198. PubMed
Prolia (denosumab) prescribing information. Amgen Inc. Revised 2020. FDA Label
Everts-Graber J, Reichenbach S, Ziswiler HR, Stuber E, Lehmann T. A single infusion of zoledronate in postmenopausal women following denosumab discontinuation results in partial conservation of bone mass gains. J Bone Miner Res. 2020;35(7):1207-1215. PubMed
Camacho PM, Petak SM, Binkley N, et al. American Association of Clinical Endocrinologists/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