Fosamax Real-World Evidence: What Registries and Observational Data Actually Show

How Alendronate Works: The Mechanism Behind Fracture Prevention

Alendronate is a nitrogen-containing bisphosphonate that binds to hydroxyapatite on bone surfaces undergoing active resorption. It inhibits farnesyl pyrophosphate synthase within the mevalonate pathway of osteoclasts, blocking prenylation of small GTPase signaling proteins that osteoclasts need to maintain their ruffled border and resorptive activity [1]. The result: osteoclast apoptosis and a measurable decline in bone turnover markers within weeks of starting therapy.

This mechanism translates to measurable clinical endpoints. In the Fracture Intervention Trial (FIT), alendronate 10 mg daily reduced radiographic vertebral fractures by 47% and hip fractures by 51% over three years in women with existing vertebral fractures (N=2,027) [2]. The trial established alendronate as a first-line treatment for postmenopausal osteoporosis.

But RCTs enroll selected populations. FIT participants had T-scores at or below -1.6 at the femoral neck, attended regular study visits, and achieved adherence rates exceeding 90% [2]. The question that real-world evidence attempts to answer is direct: does alendronate protect bones when prescribed to the broader, messier population that actually fills prescriptions in routine clinical care?

Dr. Dennis Black, the principal investigator of FIT and the FLEX extension, stated: "The efficacy we see in trials sets an upper bound. The effectiveness question, what happens when patients take the drug on their own, requires entirely different study designs" [3]. That is exactly where registry data becomes indispensable.

Danish National Registry: The Largest Single-Country Alendronate Cohort

Denmark's national prescription registry offers population-level data with minimal selection bias because it captures every filled prescription in the country. A 2011 study using this registry followed 64,182 alendronate users and matched them against non-users, finding a 27% reduction in hip fracture risk (HR 0.73, 95% CI 0.67 to 0.79) among those with high adherence [4].

The critical finding was dose-response. Patients filling enough prescriptions to cover at least 80% of days (medication possession ratio, or MPR, of 0.80 or above) had substantially better outcomes than those with lower adherence. Below an MPR of 0.50, fracture protection was statistically indistinguishable from no treatment at all [4].

A separate Danish registry analysis published in Osteoporosis International examined 38,088 patients initiating alendronate and found that only 44.7% remained on therapy at one year [5]. This persistence gap matters enormously. The registry showed that each 10% increase in MPR above the 50% threshold corresponded to roughly a 5% incremental reduction in major osteoporotic fracture risk [5]. The relationship was not linear at the extremes but followed a clear gradient in the clinically relevant range.

These numbers from a universal healthcare system with minimal cost barriers to generic alendronate illustrate a problem that no RCT can fully capture: the drug works, but patients stop taking it.

UK Clinical Practice Research Datalink: Primary Care Observational Evidence

The UK Clinical Practice Research Datalink (CPRD) links general practitioner records to hospital admissions and national mortality data, creating longitudinal patient histories that span decades. Several CPRD-based studies have examined alendronate effectiveness in routine British primary care.

A 2017 CPRD analysis of over 45,000 alendronate-treated patients reported a 39% reduction in clinical vertebral fracture risk (adjusted HR 0.61, 95% CI 0.54 to 0.69) and a 28% reduction in hip fracture risk (adjusted HR 0.72, 95% CI 0.65 to 0.80) compared with untreated matched controls [6]. These effect sizes fall below the FIT trial results but remain clinically significant.

One CPRD study specifically examined the "healthy user" bias that can inflate observational treatment effects. After adjusting for lifestyle factors, socioeconomic status, and healthcare utilization patterns, the fracture reduction estimates remained stable, suggesting that the observed benefit reflects genuine pharmacologic effect rather than confounding by health-seeking behavior [6].

The UK National Institute for Health and Care Excellence (NICE) technology appraisal TA464 cited these CPRD findings alongside trial data when reaffirming generic alendronate as the most cost-effective first-line pharmacologic intervention for osteoporosis in postmenopausal women at elevated fracture risk [7]. NICE's guideline committee noted: "Real-world evidence from UK primary care databases supports the fracture reduction efficacy observed in randomised trials, albeit at modestly lower magnitudes consistent with imperfect adherence in clinical practice" [7].

U.S. Medicare Claims: Fracture Protection in Older Americans

Medicare claims databases provide a window into alendronate use among the population most vulnerable to osteoporotic fractures: Americans aged 65 and older. Several large analyses have used these data to examine real-world effectiveness.

A retrospective cohort study using 2007 to 2012 Medicare Part D claims data identified over 126,000 new alendronate users and found a 34% reduction in hip fracture (adjusted HR 0.66, 95% CI 0.60 to 0.73) among those with MPR above 0.80, compared with non-adherent patients [8]. The analysis adjusted for age, sex, comorbidity burden via the Charlson index, concurrent glucocorticoid use, and prior fracture history.

A separate Medicare analysis focused on the first year of alendronate therapy. Hip fracture reduction was negligible in the first six months (HR 0.94, not statistically significant) but became apparent between months 6 and 12 (HR 0.78, 95% CI 0.68 to 0.89) [8]. This delayed onset of real-world fracture protection mirrors the bone mineral density accrual kinetics seen in trials, where statistically significant BMD gains at the hip require 12 to 24 months of continuous therapy [2].

The Medicare data also revealed stark racial disparities. Black women filled alendronate prescriptions at roughly half the rate of white women despite having lower BMD at diagnosis, and Hispanic patients had the lowest 12-month persistence at 28% [9]. These disparities mean that population-level fracture reduction from alendronate falls well short of its pharmacologic potential, not because the drug fails, but because healthcare delivery fails certain populations.

The Adherence Problem: Why RWE Effect Sizes Shrink

Every major registry analysis arrives at the same conclusion: alendronate works, but only when patients take it. Real-world adherence is poor.

A systematic review and meta-analysis published in Osteoporosis International pooled 40 observational studies from 15 countries and found that 12-month persistence with oral bisphosphonates (predominantly alendronate) ranged from 17.9% to 78.0%, with a pooled estimate of approximately 45% [10]. The variability depended on healthcare system, copayment structure, and how persistence was defined (permissible gap length). The median MPR across studies was 0.61, well below the 0.80 threshold associated with meaningful fracture reduction [10].

Reasons for discontinuation are well documented in patient surveys linked to registry data. Gastrointestinal side effects (esophageal irritation, dyspepsia, nausea) account for 25 to 30% of early discontinuations within the first 90 days [10]. Dosing inconvenience (fasting requirement, staying upright for 30 minutes) contributes another 15 to 20%. Fear of rare adverse effects, particularly osteonecrosis of the jaw and atypical femoral fractures, drives roughly 10 to 15% of discontinuations despite these events occurring at rates of approximately 1 per 10,000 to 1 per 100,000 patient-years for jaw osteonecrosis and 3.2 to 50 per 100,000 patient-years for atypical fractures during standard oral bisphosphonate use [11].

The gap between trial efficacy and real-world effectiveness is not unique to alendronate. But because osteoporosis is asymptomatic until a fracture occurs, patients lack the symptom-driven motivation that improves adherence for conditions like pain or hypertension. A patient who stops alendronate feels no different the next day. The fracture that might have been prevented is invisible.

Long-Term Registry Data: Benefit Duration and the Drug Holiday Question

The FLEX trial extension followed FIT participants for 10 years total and found that women who continued alendronate had fewer clinical vertebral fractures than those switched to placebo, though hip fracture rates did not differ significantly between groups [3]. FLEX informed the concept of a bisphosphonate drug holiday, a period of planned discontinuation after 3 to 5 years of therapy.

Real-world registry data adds nuance. A Danish registry study tracked patients who discontinued alendronate after 5 or more years of use and found that fracture risk began rising within 1 to 2 years of stopping, with hip fracture hazard ratios returning to pretreatment levels by year 3 post-discontinuation [12]. The rate of fracture risk return varied by site: vertebral fracture protection eroded faster than hip fracture protection, likely because alendronate's long skeletal half-life (estimated at over 10 years in cortical bone) provides more durable protection at cortical-dominant sites like the femoral neck [12].

The American Association of Clinical Endocrinology (AACE) 2020 osteoporosis guideline incorporated these registry findings and recommended reassessment after 5 years of oral bisphosphonate therapy, with continuation for patients who remain at high fracture risk (T-score at or below -2.5 at the hip, prior vertebral fracture, or elevated FRAX score) [13]. The guideline explicitly cited Danish and UK registry data as supporting evidence for the duration-of-therapy recommendation, marking one of the clearer examples of RWE directly influencing clinical practice guidelines.

Head-to-Head Registry Comparisons: Alendronate vs. Other Agents

Registry data allows comparisons that would require impractically large and expensive RCTs. Several observational studies have compared alendronate directly against other osteoporosis medications using propensity-score matching or instrumental variable designs.

A South Korean national health insurance database study (N=178,624) compared alendronate with risedronate and found no statistically significant difference in hip fracture risk (adjusted HR 1.03, 95% CI 0.92 to 1.15), consistent with indirect RCT comparisons [14]. However, the same database showed that patients switching from alendronate to denosumab had an additional 16% reduction in nonvertebral fractures (adjusted HR 0.84, 95% CI 0.74 to 0.95), a finding that aligns with the DAPS and DECIDE trial BMD data but had not been confirmed for fracture endpoints until registry analyses made it possible [14].

U.S. commercial insurance claims data comparing alendronate to zoledronic acid (intravenous bisphosphonate given once yearly) found that zoledronic acid users had 22% lower hip fracture rates (HR 0.78, 95% CI 0.67 to 0.91), largely attributable to the elimination of daily or weekly adherence requirements [15]. The once-yearly infusion achieved near-perfect adherence by design, demonstrating that alendronate's pharmacologic ceiling may be higher than registry data suggests if the adherence barrier could be removed.

Limitations of Alendronate RWE: What Registries Cannot Tell Us

Registry studies carry inherent limitations. Confounding by indication is the most persistent: patients prescribed alendronate already have diagnosed osteoporosis or elevated fracture risk, creating baseline differences from untreated comparators that statistical adjustment can reduce but never fully eliminate.

Immortal time bias affects studies that define exposure based on a minimum number of filled prescriptions, artificially inflating the treatment period during which fractures "cannot" occur in the exposed group [16]. Well-designed analyses account for this through time-dependent exposure classification, but not all published registry studies do.

Prescription fill data does not confirm ingestion. A patient who picks up alendronate from the pharmacy but takes it incorrectly (with food, lying down, or sporadically) will appear adherent in claims data while receiving suboptimal pharmacologic effect [10]. Some studies attempt to address this with serum C-telopeptide or urine N-telopeptide measurements, but biomarker data is sparse in administrative claims.

Despite these limitations, the convergence of evidence across multiple countries, healthcare systems, and analytic methods strengthens the conclusion that alendronate provides meaningful fracture protection in routine clinical care. The effect sizes are consistently smaller than RCT estimates, but the direction and clinical significance remain clear.

Synthesizing the Evidence: Where RWE and RCTs Converge

The FIT trial showed a 47% vertebral and 51% hip fracture reduction under near-ideal conditions [2]. The pooled registry evidence suggests real-world hip fracture reductions of 25 to 35% with adequate adherence and vertebral fracture reductions of 30 to 40% [4][6][8]. The gap is almost entirely explained by adherence.

This convergence matters for clinical decision-making. Alendronate remains the most widely prescribed osteoporosis drug globally, with over 20 million annual prescriptions in the United States alone [9]. Its cost (approximately $4 to $10 per month for generic alendronate 70 mg weekly) makes it the default first-line therapy in virtually every cost-effectiveness analysis [7]. Real-world evidence does not challenge that position. What it does is quantify the adherence-dependent nature of the benefit and identify specific populations (racial minorities, patients with GI comorbidities, those on complex polypharmacy regimens) who may need alternative delivery strategies or closer monitoring.

For clinicians prescribing alendronate in 2026, the registry evidence supports three actionable conclusions: confirm that patients can tolerate the fasting and upright dosing requirements before prescribing, reassess adherence at 3 and 6 months using prescription fill data, and consider switching to intravenous zoledronic acid or denosumab if MPR falls below 0.80, because the real-world data consistently shows that a bisphosphonate not taken is a bisphosphonate that does not prevent fractures [13][15].