Reviews/Osteoporosis Fracture Prevention/Bisphosphonate benefit versus atypical femoral fracture risk
Bisphosphonate Safety · ostfrac_04 · v1
Borderline: within 4 points of the Established/Likely boundary, so a re-grade could land either side.
Several large, independent studies looked at how often bisphosphonates cause an unusual thigh-bone (atypical femoral) fracture compared with how many ordinary hip and bone fractures they prevent. They agree that the unusual fracture is rare while the fractures prevented are many: one study [2] found that treating would cause one atypical fracture for every 1424 people over five years, while only 56 people need treating to stop one hip fracture, and another [1] modelled that prevented fractures far outweighed the added risk. A randomised trial [7] separately confirmed that the drug meaningfully reduces fractures. The panel judged the overall balance favorable, but noted it is weaker in Asian patients and after very long treatment, and that the direct trade-off comes from observational data rather than a single randomised comparison. That is why it stopped short of the strongest tier.
Multiple independent datasets converge that atypical femoral fracture (AFF) absolute risk is low relative to fractures prevented [1][2][3][4][5]. [2] quantifies the trade-off directly, reporting a 5-year number needed to harm of 1424 for one AFF against a number needed to treat of 56 to prevent one hip fracture, and concludes benefits 'dramatically exceed' the risk; [1] models that reductions in hip and osteoporotic fractures 'far outweighed' the increased AFF risk. [3] gives an increased absolute risk of only 5 cases per 10,000 patient-years, [4] finds AFF rates 'much lower than the expected rate of devastating hip fractures', and [5] pairs a subtrochanteric/shaft OR of 2.74 at ≥5 years with a reduced typical-fracture OR of 0.76 while noting absolute risk is low. The benefit arm is anchored by RCT [7], where zoledronate gave a fracture hazard ratio of 0.63 and a number needed to treat of 15. The contested point: The evidence divides on uniformity: [1] shows 149 hip fractures prevented versus 2 AFF in Whites but only 91 versus 8 in Asians, and AFF risk climbs steeply with duration (HR 43.51 at ≥8 years in [1]; OR 2.74 at >5 years in [5]). The panel also noted the direct trade-off is built from observational cohorts and case-control data ([1],[4],[5]), while the one RCT [7] measures benefit alone and reports no AFF data.
All 9 papers behind this claim with their DOIs and what each does to it, the panel’s arbitration, the evidence gap map and the open research questions. Free, and sent to your inbox.
We’ll also send you other claims from Osteoporosis Fracture Prevention, one at a time, and tell you when a release or a pivotal trial changes a grade. Unsubscribe by replying.
One reviewer's only job is to attack the finding. The challenges below were weighed before the grade was set.
The favorable balance is population-dependent: [1] reports 149 hip fractures prevented vs 2 AFF in Whites but only 91 vs 8 in Asians, so an unqualified 'benefits outweigh' claim overreaches for Asian populations.
AFF risk rises steeply with duration (HR 43.51 at >=8 years in [1]; OR 2.74 at >5 years in [5]); a net statement conceals that the benefit margin compresses with prolonged use.
Where it remains uncertain. The margin narrows in Asians (149 vs 2 prevented/AFF in Whites but 91 vs 8 in Asians [1]) and compresses at long durations (HR 43.51 at ≥8 years [1]), and the direct trade-off is observational, not a single randomised comparison.
Aleatory
AFF is intrinsically rare (189 AFF among 4973 shaft fractures in [2]; 59 atypical in [3]; 142 in [4]), so counts and resulting ratios carry irreducible random variation.
Epistemic
The balance differs by race; [1] reports a 4.84 hazard ratio for AFF in Asians versus Whites, leaving the Asian trade-off less favorable and less certain.
The load-bearing papers from a corpus of 9, each classified by what it does to the claim. Every DOI is taken from the corpus the claim was graded on.
Direct modeled risk-benefit comparison favoring prevention, with explicit population caveat.
“Decreases in the risk of osteoporotic and hip fractures during 1 to 10 years of bisphosphonate use far outweighed the increased risk of atypical fracture among Whites but less so among Asians.”
Panel auditInternal validity: medium · scope of inference: Women aged 50 or older receiving bisphosphonates in a US integrated health system, with a race-stratified balance that differs for Asians.Key limitations: observational, residual confounding · race subgroup imbalance · net balance modeled, not a measured endpoint
Quantitative NNH-vs-NNT showing benefit vastly exceeds harm over 5 years.
“The 5-year number needed to harm for one AFF was 1424, while the 5-year number needed to treat to prevent one hip fracture was 56.”
Panel auditInternal validity: medium · scope of inference: Danish adults aged 50 or older, relating patterns of bisphosphonate use to AFF and hip fracture.Key limitations: observational design · net balance from NNT/NNH, not a trial · nearly a third of AFF had no bisphosphonate exposure
Independent nationwide confirmation that absolute AFF risk is small despite strong relative association.
“Although there was a high prevalence of current bisphosphonate use among patients with atypical fractures, the absolute risk was small.”
Panel auditInternal validity: medium · scope of inference: Swedish women 55 or older sustaining femoral fractures, speaking to the magnitude of atypical-fracture risk only.Key limitations: no benefit side measured · small atypical-fracture sample · registry-based exposure data
The ecological trends are correlational at population level and do not measure the benefit-versus-atypical-fracture tradeoff the claim requires.
Panel auditInternal validity: low · scope of inference: US population-level trends in prescriptions, media interest and fracture hospitalisations.Key limitations: ecological, no causal inference · aggregate data only · confounded by secular trends
4 of 9 papers shown. The full review lists all of them, with the panel’s arbitration →
Claims are graded at each quarterly release, and re-graded between releases only when a pivotal trial lands. Every change is a new version of the claim.
Consistent, quantified, multi-cohort support for a net benefit, tempered by observational trade-off evidence and population- and duration-dependent caveats.
Release 2026-q4 re-graded every claim from its corpus under protocol v0.19. An earlier release's grade is shown for the record; it came from an earlier protocol and is not a movement.
Assumptions that carry weight but are not themselves established.
The panel records the gap that would most move the claim, and the study that would close it.
Race-stratified net-benefit estimate. The balance is markedly less favorable in Asians (91 prevented vs 8 AFF versus 149 vs 2 in Whites in [1]), yet rests on small counts in one dataset. The panel rates this the highest-importance open gap for the claim, and names the study design that would close it: Large multi-ethnic cohort powered to estimate prevented-fracture-to-AFF ratios separately by race over defined durations.
Open question: Does the net fracture benefit remain favorable for Asian patients, and over what treatment duration? Why it is still open: [1] shows only 91 hip fractures prevented versus 8 AFF in Asians (against 149 vs 2 in Whites) with a 4.84 AFF hazard ratio versus Whites, from a single dataset.
Priors is a sceptical first reader, not peer review, and not medical advice.
Release 2026-q4 · graded 21 Oct 2026 · protocol v0.19