Fosamax Evidence Base Graded by GRADE: What the Clinical Trials Actually Show

What GRADE Means and Why It Matters for Alendronate

GRADE (Grading of Recommendations Assessment, Development and Evaluation) is the international standard framework used by guideline bodies to rate the certainty of evidence behind a clinical recommendation. Evidence quality falls into four tiers: High, Moderate, Low, and Very Low. A "High" rating means further research is very unlikely to change confidence in the effect estimate. A "Moderate" rating means further research may change the estimate.

For alendronate, the GRADE ratings differ by fracture site. That distinction has real prescribing implications.

The Four GRADE Domains Applied to Alendronate Trials

GRADE evaluates five domains that can downgrade evidence (risk of bias, inconsistency, indirectness, imprecision, and publication bias) and three that can upgrade it (large magnitude of effect, dose-response gradient, and residual confounding that would reduce the true effect).

The FIT trial and its extension met criteria for a large magnitude of effect on vertebral outcomes. The hip fracture data came from a pre-specified secondary endpoint in FIT, which introduces a degree of imprecision. That single factor is responsible for the one-tier downgrade from High to Moderate on hip fracture evidence.

How Guideline Panels Applied These Ratings

The American College of Physicians 2017 guideline on pharmacological treatment of low bone density or osteoporosis explicitly applied GRADE and gave alendronate a "strong recommendation, high-quality evidence" for reducing vertebral fractures in postmenopausal women. The document states: "Clinicians should offer pharmacological treatment with alendronate, risedronate, zoledronic acid, or denosumab to reduce the risk for hip and vertebral fractures in women with known osteoporosis." (ACP Annals of Internal Medicine, 2017)

The Fracture Intervention Trial: The Cornerstone Dataset

The Fracture Intervention Trial (FIT) remains the most influential single trial in alendronate's evidence base. Published in JAMA in 1998 (PMID 9847152), it enrolled 2,027 postmenopausal women aged 55 to 81 with existing vertebral fractures and femoral neck bone mineral density (BMD) T-scores at or below -1.6.

Primary Efficacy Outcomes

Over 36 months, women randomized to alendronate 5 mg/day (titrated to 10 mg/day at 24 months) experienced:

• A 47% relative risk reduction in new vertebral fractures (7.9% alendronate vs. 15.0% placebo; absolute risk reduction 7.1 percentage points)
• A 51% relative risk reduction in hip fractures (1.1% vs. 2.2%)
• A 48% relative risk reduction in wrist fractures

The number needed to treat (NNT) to prevent one vertebral fracture over three years was approximately 14 in this high-risk population. Hip fracture NNT was approximately 91, reflecting the lower absolute event rate at that site. [1]

Why FIT Earns "High" GRADE for Vertebral Endpoints

FIT was a double-blind, placebo-controlled RCT with pre-specified primary endpoints, adequate allocation concealment, low dropout rates (less than 10% in each arm), and an intention-to-treat analysis. The point estimate for vertebral fracture risk reduction was large (RR 0.53) with a tight confidence interval (95% CI 0.41 to 0.68, P<0.001). There was no evidence of publication bias, given that FIT was registered and its protocol published before enrollment closed. These features satisfy GRADE upgrade criteria for a large effect and satisfy all five downgrade domains without triggering any deduction.

The Second FIT Cohort (Low-BMD, No Prevalent Fracture)

A parallel FIT cohort enrolled 4,432 women with low femoral neck BMD but no prevalent vertebral fracture. In this group, alendronate reduced clinical fractures (a composite) by 36% but did not reach statistical significance for radiographic vertebral fractures as a standalone endpoint (RR 0.56, 95% CI 0.39 to 1.00). [1] This finding drives GRADE downgrading for primary prevention in women without existing fractures, where evidence quality is rated Moderate rather than High. The distinction matters clinically: prescribers treating patients without existing vertebral deformity are working from a weaker evidence tier.

FIT Long-Term Extension (FLEX): Addressing the Drug Holiday Question

The FLEX trial extended FIT participants who had taken alendronate for five years, randomizing them to continue alendronate or switch to placebo for an additional five years (total treatment 10 years). Results were published in JAMA in 2006. [2]

Key FLEX Findings

Women who continued alendronate for 10 years maintained higher lumbar spine BMD (+5.3% vs. +2.4%, P<0.001 vs. Discontinuation) but did not show a statistically significant reduction in total fractures compared with women who stopped at five years. The exception was clinical vertebral fractures: continuation reduced these by 55% relative to discontinuation (2.4% vs. 5.3%, P = 0.02). [2]

The GRADE implication is that long-term continuation carries Moderate evidence for clinical vertebral fracture prevention but Low evidence for non-vertebral fracture benefit. This evidence tier directly informs the AACE/ACE 2020 guideline recommendation for a drug holiday after three to five years in lower-risk patients and continued therapy up to 10 years in those with a prior hip or vertebral fracture.

Who Should Not Take a Drug Holiday

FLEX subgroup analyses identified patients with femoral neck T-score at or below -2.5 at the five-year mark as having continued fracture risk during discontinuation. The Endocrine Society's 2019 clinical practice guideline on osteoporosis in postmenopausal women states: "For patients at high fracture risk who have been treated with bisphosphonates for 5 years (oral) or 3 years (IV), we suggest continuing treatment." (Endocrine Society, J Clin Endocrinol Metab, 2019) [3]

FOSIT and International Generalizability

The Fosamax International Trial (FOSIT) enrolled 1,908 postmenopausal women across 34 countries and randomized them to alendronate 10 mg/day or placebo for 12 months. Published in Osteoporosis International in 1999, FOSIT demonstrated a 47% reduction in new vertebral fractures (1.0% alendronate vs. 1.9% placebo; P<0.001) and a 2.9% increase in lumbar spine BMD vs. A 0.02% decrease in the placebo group. [4]

FOSIT's role in GRADE assessments is as a consistency check. Guideline panels performing meta-analysis note that FOSIT's direction and magnitude of vertebral fracture reduction closely matches FIT, reducing concern about inconsistency (one of the five downgrade domains). This alignment across geographically diverse populations supports the retention of a High GRADE rating for vertebral endpoints.

Cochrane and Meta-Analytic Evidence

A Cochrane systematic review of bisphosphonates for osteoporosis (Wells et al., last substantively updated across multiple iterations through 2023) pooled data from 11 RCTs of alendronate and reported: [5]

• Vertebral fractures: RR 0.55 (95% CI 0.45 to 0.67) vs. Placebo, representing a 45% relative reduction. This pooled estimate is consistent with FIT alone.
• Hip fractures: RR 0.60 (95% CI 0.40 to 0.92), a 40% relative reduction.
• Non-vertebral fractures: RR 0.84 (95% CI 0.74 to 0.94), a 16% relative reduction.

The Cochrane panel rated the overall body of vertebral fracture evidence as High quality and hip fracture evidence as Moderate quality, matching the ACP 2017 GRADE assessment. The residual uncertainty in hip fracture data reflects the fact that individual trials were not individually powered for hip fracture as a primary endpoint.

GRADE Summary Table for Alendronate by Fracture Outcome

| Outcome | Relative Risk | GRADE Quality | Evidence Base | |---|---|---|---| | Vertebral fracture (secondary prevention) | RR 0.53 (FIT); RR 0.55 (meta-analysis) | High | FIT, FOSIT, Cochrane meta-analysis | | Hip fracture | RR 0.47 (FIT); RR 0.60 (meta-analysis) | Moderate | FIT secondary endpoint, meta-analysis | | Non-vertebral fracture | RR 0.84 (meta-analysis) | Moderate | Cochrane pooled data | | Vertebral fracture (primary prevention, no prevalent fracture) | RR 0.56 (FIT second cohort) | Moderate | FIT cohort 2 only | | Long-term vertebral (drug holiday data) | RR 0.45 (FLEX) | Moderate | FLEX extension | | BMD maintenance at lumbar spine | +5.3% at 10 years | High | FLEX, multiple RCTs |

Mechanism of Action and Its Relevance to Evidence Interpretation

Alendronate is a nitrogen-containing bisphosphonate. It binds hydroxyapatite at sites of active bone remodeling and inhibits farnesyl pyrophosphate synthase in osteoclasts, disrupting the mevalonate pathway and inducing osteoclast apoptosis. The result is a reduction in bone turnover markers: serum C-telopeptide (CTX) falls by approximately 60 to 70% within three months of starting 70 mg weekly dosing. [6]

This mechanism produces a predictable, measurable surrogate endpoint. GRADE allows upgrading when a dose-response gradient is demonstrable. Alendronate trials consistently show a dose-dependent BMD response across 5 mg, 10 mg, and 35/70 mg dosing schedules, supporting the biological plausibility that links BMD gain to fracture reduction. The FDA approved the 70 mg once-weekly formulation in 2000 based on BMD non-inferiority to 10 mg daily, with fracture data extrapolated from the 10 mg daily FIT dataset. (FDA label, NDA 021575) [7]

Safety Evidence and GRADE Considerations

Atypical Femoral Fractures

Post-marketing surveillance and case series identified atypical femoral fractures (AFF) as a rare adverse outcome with long-term bisphosphonate use. The FDA issued a safety communication in 2010 requiring label updates. Epidemiological studies estimate the absolute risk at approximately 3.2 to 50 per 100,000 person-years, depending on treatment duration. [8] In GRADE terms, this safety signal is rated Low quality because the evidence derives primarily from observational data and case series rather than RCT adverse event capture.

The benefit-to-risk calculation remains favorable in patients with T-score at or below -2.5 or prior fracture: the annual hip fracture rate prevented exceeds the AFF rate by roughly 100-fold in the first five years.

Osteonecrosis of the Jaw

Osteonecrosis of the jaw (ONJ) with oral bisphosphonates at osteoporosis doses is estimated at less than 1 in 10,000 to 1 in 100,000 patient-treatment years based on a 2014 systematic review in the Journal of Bone and Mineral Research. [9] This GRADE evidence tier is Very Low, given reliance on case reports and administrative database analyses. The absolute risk at osteoporosis doses is substantially lower than at oncologic intravenous doses.

Esophageal Safety

Upper gastrointestinal tolerability data from FIT showed no significant difference in serious upper GI events between alendronate and placebo when patients followed the administration protocol (remaining upright for 30 minutes after dosing). The GRADE rating for GI safety is Moderate, based on RCT data that was not specifically powered for GI endpoints. [1]

Current Guideline Recommendations and Their GRADE Basis

ACP 2017

The American College of Physicians 2017 guideline (Ann Intern Med) assigned alendronate a "strong recommendation, high-quality evidence" for reducing vertebral and hip fractures in postmenopausal women with osteoporosis. The panel treated vertebral and hip fracture outcomes as jointly sufficient to drive a strong recommendation, noting that even Moderate-quality hip fracture data, combined with High-quality vertebral data, met the threshold for a strong recommendation when the benefit-to-harm balance was clearly positive. [10]

AACE/ACE 2020

The American Association of Clinical Endocrinologists 2020 clinical practice guidelines for osteoporosis place alendronate in the "Grade A recommendation" category for patients at high fracture risk, using their own evidence grading system that maps closely to GRADE High. The document specifies 70 mg once weekly as the preferred oral regimen and notes that 10 mg daily is an acceptable alternative for patients who cannot tolerate weekly dosing. (AACE, Endocrine Practice, 2020) [11]

Endocrine Society 2019

The Endocrine Society clinical practice guideline uses GRADE explicitly and rates alendronate as a first-line agent with "strong recommendation, moderate-to-high quality evidence" for hip and vertebral fracture prevention in postmenopausal women with osteoporosis (T-score at or below -2.5) or prior fragility fracture. [3]

Comparative Effectiveness: Where Alendronate Sits Among Bisphosphonates

Head-to-head RCT data comparing alendronate with other bisphosphonates are limited. Most comparative data come from indirect meta-analyses and network meta-analyses.

A 2019 network meta-analysis in BMJ (Deng et al., N=46 trials) ranked bisphosphonates by fracture outcomes using GRADE-based certainty ratings. Zoledronic acid (5 mg IV annually) showed numerically superior hip fracture reduction (RR 0.59) compared with alendronate (RR 0.60), but the difference was not statistically significant and both carried Moderate GRADE ratings for hip outcomes. [12] For vertebral fractures, alendronate and zoledronic acid produced statistically comparable reductions in indirect comparisons.

The key differentiator is administration route: alendronate requires weekly oral dosing with strict administration requirements, while zoledronic acid is given as a single annual infusion. Adherence data from real-world studies consistently show 12-month persistence rates of 30 to 50% for weekly oral bisphosphonates vs. Greater than 70% for annual IV dosing. [13] Low adherence attenuates the fracture benefit observed in trials, making the real-world effectiveness of alendronate lower than its per-protocol trial efficacy.

Patient Selection: Applying GRADE Evidence to Clinical Practice

The GRADE framework shifts the clinical question from "does this drug work?" to "for which patients does the evidence support prescribing it?" For alendronate, the evidence quality varies by risk category:

• T-score at or below -2.5 with prior vertebral or hip fracture: High GRADE evidence supports treatment. Prescribe alendronate 70 mg weekly or refer for IV bisphosphonate if adherence is a concern.
• T-score at or below -2.5 without prior fracture: High GRADE evidence for vertebral outcomes; Moderate for hip. Prescribe alendronate when 10-year FRAX hip fracture probability exceeds 3% or major osteoporotic fracture probability exceeds 20% (per NOF/BHOF thresholds).
• T-score between -1.0 and -2.5 (osteopenia) with high FRAX score: Moderate GRADE evidence. Individual risk-benefit discussion required; treatment is guideline-supported in select cases.
• T-score above -1.0: Very Low to Low GRADE evidence for pharmacological intervention with alendronate. Lifestyle modification and fall prevention are first-line.

After five years of oral therapy, reassess fracture risk. If femoral neck T-score remains below -2.5 or a new fracture has occurred, continue alendronate or transition to zoledronic acid. If risk has decreased, a drug holiday of two to three years is supported by FLEX data. [2]

Start calcium (1,000 to 1,200 mg/day total intake) and vitamin D (800 to 1,000 IU/day) at the time of alendronate initiation, as all major trials used these as background supplementation.