Why did ARCH use alendronate instead of placebo?

By 2017, placebo-controlled fracture trials in high-risk osteoporosis were considered ethically difficult to justify. Alendronate was chosen because it is the most commonly prescribed bisphosphonate globally and has the largest body of fracture-reduction evidence.

What does double-dummy blinding mean in the ARCH trial?

Every participant received both a monthly injection and a weekly pill. One was active drug, the other was matched placebo. This prevented unblinding from the route of administration (injection vs oral tablet).

How were vertebral fractures measured in ARCH?

Lateral spine radiographs at baseline, 12 months, and 24 months were graded by two blinded central readers using the Genant semiquantitative scale. A new fracture required a worsening of at least one grade. Disagreements went to a third reader.

What was the cardiovascular safety signal in ARCH?

Serious cardiovascular events occurred in 2.5% of romosozumab patients vs 1.9% of alendronate patients during the 12-month blinded phase. This led to a boxed warning on the FDA label, though the FRAME trial (vs placebo) did not show the same imbalance.

Does the 48% fracture reduction apply to all types of fractures?

The 48% reduction is specific to new morphometric vertebral fractures. Hip fractures were reduced by 38%, clinical fractures by 27%, and nonvertebral fractures by 19%. Each was statistically significant within a pre-specified hierarchical testing framework.

Were men included in the ARCH trial?

No. ARCH enrolled only postmenopausal women aged 55 and older. Romosozumab's efficacy in men with osteoporosis was studied separately in the BRIDGE trial.

How long did patients receive romosozumab in ARCH?

Romosozumab was given for 12 months only. After that, all patients transitioned to open-label alendronate. The primary endpoint was assessed at 24 months, meaning half the observation period was post-romosozumab.

What happened to bone density after romosozumab was stopped?

Patients who received romosozumab for 12 months and then switched to alendronate maintained and continued to gain bone density. This supports the concept that an anabolic agent followed by an antiresorptive preserves the bone formed during the anabolic window.

Is ARCH applicable to patients already on denosumab?

ARCH excluded patients who had received denosumab within 12 months. The trial does not directly inform the efficacy of switching from denosumab to romosozumab, a sequence that carries theoretical rebound risk.

How does ARCH compare to the FRAME trial?

FRAME tested romosozumab vs placebo and found a 73% vertebral fracture reduction at 12 months. ARCH tested against alendronate and found a 48% reduction at 24 months. The two trials complement each other: FRAME demonstrates absolute efficacy, ARCH demonstrates superiority over standard care.

Inside the ARCH Methodology: What Most Summaries Skip

By HealthRX Medical Team

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

At a glance

| Parameter | Detail | |---|---| | N | 4,093 postmenopausal women | | Intervention | Romosozumab 210 mg SC monthly x 12 months, then alendronate 70 mg weekly | | Comparator | Alendronate 70 mg weekly x 12 months, then continued alendronate | | Duration | Median 33 months (primary endpoint at 24 months) | | Primary endpoint | New vertebral fracture through 24 months | | Key result | 48% relative risk reduction in new vertebral fractures vs alendronate (p <0.001) |

Why an Active Comparator Instead of Placebo?

Most bone-drug trials before ARCH used placebo arms. The FRAME trial had already shown romosozumab's superiority over placebo. ARCH asked a harder question: does romosozumab beat a drug clinicians already prescribe?

The choice of alendronate as the comparator was deliberate. Alendronate is the most widely prescribed bisphosphonate worldwide. It has decades of fracture-reduction data. By selecting an active comparator with a proven track record, the investigators set a high bar. A positive result here would carry more weight in clinical practice than another placebo-controlled win.

This choice also carried risk. Any adverse-event signal would be measured against a drug with a well-characterized safety profile, meaning differences in cardiovascular events (which did emerge) would be harder to dismiss as background noise.

Randomization and Blinding Architecture

ARCH enrolled 4,093 postmenopausal women across 386 sites in 42 countries. Randomization was 1:1, stratified by age (<75 vs ≥75 years) and by the presence of a prevalent vertebral fracture at one vs two or more vertebral levels. This stratification mattered because baseline fracture burden is the single strongest predictor of future fracture risk, and age modifies both drug metabolism and fall probability.

The trial employed a double-blind, double-dummy technique. Every participant received both a monthly subcutaneous injection and a weekly oral tablet. During the first 12 months, the romosozumab group got active injections plus placebo tablets, while the alendronate group got placebo injections plus active tablets. After month 12, both groups received open-label alendronate. This design eliminated the possibility that injection-site reactions or pill-taking behavior could unblind investigators.

The HealthRX Estimand Dissection

Most trial summaries report the hazard ratio and move on. The ARCH methodology deserves a closer read through the ICH E9(R1) estimand framework, which asks five questions of any primary endpoint. Here is how ARCH answers each one:

| Estimand Element | How ARCH Defined It | |---|---| | Population | Postmenopausal women ≥55 years with T-score ≤ −2.5 at total hip or femoral neck AND either ≥1 moderate/severe vertebral fracture or ≥2 mild vertebral fractures | | Treatment | Romosozumab 210 mg monthly x 12 months → alendronate vs alendronate x 12 months → alendronate | | Endpoint | New morphometric vertebral fracture (≥1 grade worsening by Genant semiquantitative method) | | Intercurrent events | Treatment discontinuation handled by treatment-policy strategy (all randomized patients included regardless of adherence) | | Population-level summary | Relative risk at 24 months, analyzed in the ITT population |

This framework reveals something important. The "treatment-policy" strategy for intercurrent events means the trial measured the effect of assigning romosozumab, not the effect of completing romosozumab. Roughly 19% of patients in the romosozumab arm discontinued before 24 months. The ITT analysis dilutes the observed effect. The true on-treatment efficacy is likely larger than the reported 48% reduction.

Inclusion and Exclusion Criteria: Who Got In (and Who Didn't)

The enrollment criteria selected for high-risk patients. Women needed a femoral neck or total hip T-score at or below −2.5, combined with a history of fracture. This is not a mild-osteoporosis population. The ARCH investigators specifically targeted patients where fracture prevention carries the most clinical urgency.

Key exclusions included:

Metabolic bone disease other than osteoporosis (e.g., Paget's disease, osteomalacia)
Current use of bone-active drugs (bisphosphonates within 3 years, denosumab within 12 months, any prior parathyroid hormone analog)
Vitamin D levels below 20 ng/mL (patients could be repleted and re-screened)
History of hip fracture on both sides
Conditions affecting bone metabolism: uncontrolled thyroid disease, malabsorption syndromes, chronic kidney disease stage 4+

The exclusion of prior anabolic therapy matters clinically. Patients who had already received teriparatide were not represented. This limits the trial's applicability to a common real-world scenario: sequencing a sclerostin inhibitor after a PTH analog.

Primary Endpoint Definition

Vertebral fractures in ARCH were assessed by lateral spine radiographs at baseline, 12 months, and 24 months. Two central readers, blinded to treatment assignment and time point, graded each vertebra from T4 to L4 using the Genant semiquantitative scale:

| Grade | Height Reduction | |---|---| | 0 (Normal) | <20% | | 1 (Mild) | 20-25% | | 2 (Moderate) | 26-40% | | 3 (Severe) | >40% |

A "new" vertebral fracture required a vertebra that was grade 0 at baseline to reach grade 1 or higher at follow-up, or any increase in grade for vertebrae already fractured. Discrepancies between readers went to a third adjudicator. This morphometric definition catches fractures that are clinically silent. Roughly two-thirds of vertebral fractures never come to clinical attention. Without radiographic screening, ARCH would have missed most of its primary events.

Statistical Approach and Power

The trial was powered for superiority, not noninferiority. The primary analysis used a logistic regression model adjusting for age stratum and prevalent vertebral fracture stratum. The relative risk and 95% confidence interval were derived from this model.

The sample size of 4,093 assumed a 24-month vertebral fracture rate of approximately 6% in the alendronate arm and was powered to detect a 40% relative reduction with 90% power at a two-sided alpha of 0.05. The actual result (48% reduction) exceeded the design assumption.

Secondary endpoints, including clinical fracture, nonvertebral fracture, and hip fracture, were tested in a pre-specified hierarchical sequence to control the family-wise error rate. This is a conservative approach. Each subsequent endpoint could only be declared statistically significant if the preceding endpoint crossed its threshold.

The hierarchical results through 24 months:

| Endpoint | Romosozumab → Alen | Alendronate | RR (95% CI) | p-value | |---|---|---|---|---| | New vertebral fracture | 4.1% | 8.0% | 0.52 (0.40-0.68) | <0.001 | | Clinical fracture | 9.7% | 13.0% | 0.73 (0.61-0.86) | <0.001 | | Nonvertebral fracture | 8.7% | 10.6% | 0.81 (0.66-0.99) | 0.04 | | Hip fracture | 2.0% | 3.2% | 0.62 (0.42-0.92) | 0.02 |

All four endpoints achieved significance within the hierarchy.

The Cardiovascular Signal

The result that shaped romosozumab's FDA label as much as the fracture data was the cardiovascular imbalance. Through 12 months (the double-blind period), adjudicated serious cardiovascular events occurred in 2.5% of the romosozumab group vs 1.9% of the alendronate group.

This finding requires careful interpretation. Alendronate may have a mild cardioprotective effect, which would make any neutral drug look comparatively worse. The FRAME trial (romosozumab vs placebo) did not show the same imbalance. Whether the signal reflects true romosozumab risk, alendronate benefit, or statistical noise remains debated.

The FDA responded with a boxed warning and a contraindication in patients who have had a myocardial infarction or stroke within the preceding year. The Endocrine Society's 2020 guidelines recommend cardiovascular risk assessment before prescribing romosozumab, citing ARCH directly.

Limitations the Authors Acknowledged

The original publication noted several limitations worth reading past the abstract:

Population specificity. Only postmenopausal women with severe osteoporosis were enrolled. Extrapolation to men, premenopausal women, or milder disease requires caution.
Sequence constraint. ARCH tested romosozumab followed by alendronate vs alendronate alone. It did not test romosozumab monotherapy beyond 12 months, nor did it test romosozumab followed by denosumab (a sequence increasingly used in practice).
Open-label transition phase. After month 12, both groups received open-label alendronate. The loss of blinding could have influenced adverse-event reporting in the extension period.
Cardiovascular adjudication timing. The cardiovascular endpoint was not pre-specified as a primary or key secondary outcome. The adjudication committee was convened after the signal emerged, raising questions about whether the process was optimally powered or designed for this analysis.
Ethnic and geographic diversity. While the trial spanned 42 countries, the majority of sites were in Europe, North America, and Latin America. Representation from East Asia and Sub-Saharan Africa was limited.

What This Means for Prescribing Decisions

ARCH's methodology makes a strong case that romosozumab-to-bisphosphonate is superior to bisphosphonate alone for preventing fractures in high-risk postmenopausal women. The active comparator, double-dummy blinding, central radiographic adjudication, and hierarchical testing all reflect a rigorous design.

The cardiovascular signal, however, means that the decision to prescribe romosozumab cannot be based on skeletal efficacy alone. Clinicians must weigh fracture risk against cardiovascular risk, a trade-off that the trial's design made visible precisely because the comparator was an active drug with a known safety profile rather than a placebo.

For patients with imminent fracture risk and low cardiovascular burden, the ARCH data supports romosozumab as a first-line anabolic agent. For patients with recent cardiovascular events, the trial's own data (and the resulting FDA label) argues against it.

Frequently asked questions

›

References

Saag KG, Petersen J, Brandi ML, et al. Romosozumab or alendronate for fracture prevention in women with osteoporosis. N Engl J Med. 2017;377(15):1417-1427. PubMed
Cosman F, Crittenden DB, Adachi JD, et al. Romosozumab treatment in postmenopausal women with osteoporosis. N Engl J Med. 2016;375(16):1532-1543. PubMed
Evenity (romosozumab-aqqg) prescribing information. Amgen Inc. 2019. FDA Label
Shoback D, Rosen CJ, Black DM, et al. Pharmacological management of osteoporosis in postmenopausal women: an Endocrine Society guideline update. J Clin Endocrinol Metab. 2020;105(3):587-594. PubMed
Genant HK, Wu CY, van Kuijk C, Nevitt MC. Vertebral fracture assessment using a semiquantitative technique. J Bone Miner Res. 1993;8(9):1137-1148. PubMed