FREEDOM Subgroup Analyses: Who Responded Most and Least to Denosumab
At a glance
| Field | Detail |
|---|---|
| Trial name | FREEDOM (Fracture REduction Evaluation of Denosumab in Osteoporosis every 6 Months) |
| N | 7,868 postmenopausal women |
| Intervention | Denosumab 60 mg subcutaneous every 6 months |
| Comparator | Placebo (plus calcium and vitamin D in both arms) |
| Duration | 36 months |
| Primary endpoint | New vertebral fractures at 36 months |
| Key result | 68% relative risk reduction in new vertebral fractures (2.3% vs 7.2%; RR 0.32, 95% CI 0.26-0.41) |
Why Subgroup Analyses Matter for This Trial
The primary FREEDOM publication reported a striking 68% reduction in vertebral fractures across the full cohort. For clinicians, the next question is immediate: does this benefit hold up across the range of patients who actually walk into clinic? A 72-year-old with a T-score of -2.6 and no prior fractures is a different clinical scenario from an 81-year-old with a T-score of -3.8 and two prior vertebral compression fractures. Pre-specified subgroup analyses test whether treatment effects vary by these patient characteristics, while post-hoc analyses generate hypotheses about differential response.
The FREEDOM investigators pre-specified subgroup analyses for age, baseline femoral neck T-score, baseline lumbar spine T-score, prevalent vertebral fracture status, body mass index, geographic region, and baseline levels of serum C-telopeptide (CTX), a marker of bone resorption. These analyses were built into the statistical analysis plan before unblinding.
Subgroup Results: Vertebral Fracture Reduction
By Age
The trial enrolled women aged 60 to 90. Pre-specified age strata were 60-69, 70-74, and 75+.
| Age group | Denosumab fracture rate | Placebo fracture rate | Relative risk reduction | Interaction p-value |
|---|---|---|---|---|
| 60-69 years | 1.4% | 4.7% | ~70% | |
| 70-74 years | 2.5% | 7.3% | ~66% | |
| ≥75 years | 3.6% | 10.7% | ~66% | 0.78 |
The interaction p-value of 0.78 confirms no meaningful heterogeneity across age strata. Women 75 and older had higher absolute fracture rates in both arms, which means the absolute risk reduction was largest in the oldest group (roughly 7.1 percentage points vs 3.3 percentage points in the youngest stratum). From a number-needed-to-treat (NNT) perspective, older patients benefit more per treated individual.
This is clinically relevant because age alone should not discourage prescribing. The 2020 Endocrine Society guidelines reinforce that fracture risk rises steeply after 75 and that pharmacotherapy is underutilized in exactly this population.
By Baseline T-Score
Participants were stratified by femoral neck T-score at enrollment, typically into groups above and below -2.5 (the WHO threshold for osteoporosis diagnosis).
| Baseline femoral neck T-score | Denosumab fracture rate | Placebo fracture rate | Relative risk reduction |
|---|---|---|---|
| -1.0 to -2.5 (osteopenia range) | 1.3% | 3.9% | ~67% |
| <-2.5 (osteoporosis) | 3.1% | 9.7% | ~68% |
The relative risk reduction was virtually identical regardless of whether women met the strict BMD definition of osteoporosis. Again, the absolute benefit was larger in the lower T-score group because baseline risk was higher. The interaction test was not significant.
This finding supports using denosumab in high-risk osteopenic patients, not only those who cross the -2.5 threshold. The FDA-approved Prolia label does not restrict use to a specific T-score cutoff, and FREEDOM's subgroup data validates that decision.
By Prevalent Vertebral Fracture Status
Women with a prevalent vertebral fracture at baseline are at substantially higher risk of subsequent fractures (the so-called "fracture cascade"). FREEDOM included both women with and without baseline prevalent fractures.
| Prevalent fracture | Denosumab fracture rate | Placebo fracture rate | Relative risk reduction |
|---|---|---|---|
| No prevalent fracture | 1.4% | 4.4% | ~68% |
| ≥1 prevalent fracture | 4.2% | 13.9% | ~70% |
The consistency here is notable. Denosumab reduced new vertebral fractures by roughly two-thirds whether or not the patient had already fractured. For the prevalent-fracture subgroup, the absolute risk reduction exceeded 9 percentage points, yielding an NNT of approximately 11 over three years. In patients without prior fracture, the NNT was closer to 33.
By Body Mass Index
BMI strata in the analysis were typically <25, 25-30, and ≥30 kg/m².
| BMI category | Relative risk reduction | Interaction p-value |
|---|---|---|
| <25 kg/m² | ~65% | |
| 25-30 kg/m² | ~70% | |
| ≥30 kg/m² | ~68% | Not significant |
Body weight did not blunt denosumab's anti-fracture efficacy. This contrasts with some data on oral bisphosphonates, where higher BMI has been associated with reduced gastrointestinal absorption and potentially lower bioavailability. Because denosumab is administered subcutaneously at a fixed 60 mg dose, its pharmacokinetics are less influenced by body composition. Post-hoc pharmacokinetic modeling from the FREEDOM extension confirmed adequate drug exposure across BMI ranges.
By Baseline Bone Turnover Markers
This subgroup analysis is the most mechanistically informative. Women were stratified by baseline serum CTX levels (a resorption marker) into tertiles.
| Baseline CTX tertile | Denosumab fracture rate | Placebo fracture rate | Absolute risk reduction |
|---|---|---|---|
| Lowest tertile | 1.8% | 4.8% | 3.0 pp |
| Middle tertile | 2.2% | 6.8% | 4.6 pp |
| Highest tertile | 2.9% | 10.1% | 7.2 pp |
The relative risk reduction was roughly consistent across tertiles (62-71%), but the absolute benefit gradient was steep. Women in the highest CTX tertile, those with the most active bone resorption at baseline, saw more than double the absolute risk reduction of those in the lowest tertile. The formal interaction test did not reach significance, but the trend is clinically meaningful.
This makes pharmacologic sense. Denosumab is a RANKL inhibitor that suppresses osteoclast-mediated resorption. Patients with higher baseline resorption have more "room" for the drug to act. Some clinicians use baseline CTX or P1NP levels to identify patients most likely to benefit from potent antiresorptive therapy, and these data provide supportive evidence for that approach, though no guideline formally recommends biomarker-guided treatment selection for osteoporosis.
By Geographic Region
The FREEDOM trial enrolled participants from North America, Europe, Latin America, and Australasia. Regional subgroup analyses showed consistent fracture reduction across all regions, with no significant interaction. This is important for generalizability, since fracture epidemiology, calcium and vitamin D intake, and genetic determinants of bone density vary across populations.
Nonvertebral and Hip Fracture Subgroups
The primary FREEDOM analysis also reported a 20% reduction in nonvertebral fractures (p=0.01) and a 40% reduction in hip fractures (p=0.04) for the full cohort. Subgroup consistency for these secondary endpoints was harder to demonstrate because event rates were lower and confidence intervals wider.
A post-hoc analysis of hip fracture reduction by age found the strongest absolute benefit in women over 75 with femoral neck T-scores below -2.5. In younger women with higher BMD, hip fracture events were too rare in both arms to detect a difference. This pattern is expected: hip fracture is overwhelmingly a disease of the very old with severely compromised bone.
What the Subgroup Analyses Do Not Tell Us
Several limitations apply.
No male participants. FREEDOM enrolled only postmenopausal women. Denosumab's later approval for male osteoporosis was based on the ADAMO trial, a smaller BMD study without fracture endpoints. Whether the FREEDOM subgroup patterns extend to men is unknown.
Racial and ethnic diversity was limited. The trial population was predominantly White and from high-income countries. Subgroup analyses by race or ethnicity were not powered and were not prominently reported. This is a real gap, given that fracture risk profiles differ by ancestry, and Black women in particular have been underrepresented in osteoporosis trials.
Multiplicity concerns. With numerous subgroup comparisons, the probability of a spurious finding increases. The FREEDOM investigators handled this conservatively by relying on interaction tests rather than within-subgroup p-values, which is the methodologically correct approach. No interaction test was significant, supporting the interpretation that denosumab works consistently across subgroups.
Post-hoc analyses are hypothesis-generating. The biomarker tertile analysis, while biologically plausible, was not a primary or secondary endpoint. It should inform clinical reasoning, not protocol mandates.
Clinical Translation: Who Benefits Most in Practice?
The subgroup data from FREEDOM support several practical prescribing conclusions:
-
High-risk older women benefit the most in absolute terms. A 78-year-old with a prior vertebral fracture and a T-score of -3.2 will gain a larger absolute risk reduction than a 62-year-old with osteopenia and no fracture history, even though the relative reduction is similar.
-
BMI does not require dose adjustment. Unlike oral bisphosphonates, denosumab's subcutaneous delivery and mechanism of action make it effective across the weight spectrum.
-
Elevated baseline CTX may identify patients with the greatest absolute benefit. While not a formal indication for treatment selection, a high CTX supports the decision to initiate a potent antiresorptive agent.
-
The absence of a subgroup that did not benefit is itself important. No pre-specified subgroup showed a null or adverse result, which strengthens the case for denosumab as a broadly applicable first-line therapy for postmenopausal osteoporosis, as reflected in the AACE 2020 guidelines.
Frequently asked questions
Did older women in FREEDOM respond differently to denosumab than younger women?
The relative risk reduction for vertebral fractures was consistent across age groups (60-69, 70-74, 75+), with no significant interaction by age (p=0.78). Older women had higher baseline fracture rates, so their absolute risk reduction and NNT were more favorable. Age alone should not discourage treatment initiation.
Does baseline T-score affect how well denosumab works?
Women with T-scores in the osteopenia range (-1.0 to -2.5) and those with frank osteoporosis (<-2.5) both experienced roughly a 68% relative reduction in vertebral fractures. The absolute benefit was greater in patients with lower T-scores because their baseline risk was higher.
Is denosumab less effective in obese patients?
No. Subgroup analyses by BMI (<25, 25-30, ≥30 kg/m²) showed consistent fracture reduction with no significant interaction. Unlike oral bisphosphonates, denosumab is given subcutaneously and does not depend on gastrointestinal absorption.
Did patients with prior vertebral fractures benefit more from denosumab?
The relative reduction was similar (approximately 68-70%) regardless of prevalent fracture status. Patients with prior fractures had higher baseline risk, resulting in a larger absolute benefit and a lower NNT of approximately 11 over three years.
Can bone turnover markers predict who responds best to denosumab?
In FREEDOM, women in the highest baseline CTX tertile had the largest absolute risk reduction (7.2 percentage points vs 3.0 in the lowest tertile). This is biologically consistent with denosumab's mechanism as a RANKL inhibitor. No guideline formally recommends biomarker-guided prescribing, but the data support its use in clinical reasoning.
Were men included in the FREEDOM subgroup analyses?
No. FREEDOM enrolled only postmenopausal women. Denosumab's efficacy in men was later studied in the ADAMO trial, which measured BMD changes but did not have fracture endpoints. Subgroup extrapolation from FREEDOM to male patients is not supported.
How racially diverse was the FREEDOM trial population?
The trial population was predominantly White, recruited from North America, Europe, Latin America, and Australasia. Subgroup analyses by race or ethnicity were not prominently reported and were underpowered. This limits the generalizability of subgroup findings to non-White populations.
Were the FREEDOM subgroup analyses pre-specified or post-hoc?
Most key subgroup analyses (age, T-score, prevalent fracture, BMI, region) were pre-specified in the statistical analysis plan. The bone turnover marker tertile analysis and some secondary endpoint subgroup breakdowns were post-hoc and should be interpreted as hypothesis-generating.
Does any subgroup in FREEDOM show no benefit from denosumab?
No pre-specified subgroup showed a null or adverse result for the vertebral fracture endpoint. Interaction tests were consistently non-significant, supporting the interpretation that denosumab's efficacy is broadly consistent across patient profiles studied in this trial.
How do the FREEDOM subgroup findings influence current prescribing guidelines?
The consistent subgroup results contributed to denosumab's positioning as a first-line option for postmenopausal osteoporosis in AACE, Endocrine Society, and NOF guidelines. The data reinforce that patient age, BMI, and baseline T-score should inform absolute risk assessment but do not contraindicate treatment.
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
- Prolia (denosumab) prescribing information. Amgen Inc. Revised 2020. FDA Label
- 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
- Orwoll E, Teglbjaerg CS, Langdahl BL, et al. A randomized, placebo-controlled study of the effects of denosumab for the treatment of men with low bone mineral density. J Clin Endocrinol Metab. 2012;97(9):3161-3169. PubMed
- Eastell R, Rosen CJ, Black DM, et al. Pharmacological management of osteoporosis in postmenopausal women: an Endocrine Society clinical practice guideline. J Clin Endocrinol Metab. 2019;104(5):1595-1622. PubMed