Honest Criticisms and Limitations of the FIT (Fracture Intervention Trial) Trial

At a glance

Parameter	Detail
N	4,432 (FIT-2, the Clinical Fractures arm)
Intervention	Alendronate 5 mg/day (increased to 10 mg/day at month 24)
Comparator	Placebo
Duration	~4.2 years (mean)
Primary endpoint	Clinical fractures (vertebral and nonvertebral) in women with low BMD but no prevalent vertebral fracture
Key result	47% relative risk reduction in morphometric vertebral fractures; no statistically significant reduction in the primary composite clinical fracture endpoint

Background: Two Trials Under One Name

The FIT program actually comprised two parallel studies. FIT-1 enrolled women who already had at least one vertebral fracture at baseline and showed striking fracture reductions across hip, wrist, and spine. FIT-2, published in JAMA in 1998, enrolled women with low femoral-neck BMD (T-score <−1.6) but without baseline vertebral fractures. This distinction matters for every limitation discussed below, because FIT-2's population was at lower absolute fracture risk than FIT-1's cohort, and the primary endpoint result reflected that difference.

Enrollment Biases That Shape Every Outcome

Demographic homogeneity

FIT-2 enrolled postmenopausal women aged 55 to 80 drawn from 11 U.S. clinical centers. Over 96% of participants were white. Black, Hispanic, and Asian women, who have different bone geometry, turnover rates, and fracture epidemiology, were essentially absent. Any clinician applying FIT-2 data to a multiethnic patient panel is extrapolating, not citing.

Exclusion of medically complex patients

The trial excluded women on estrogen, fluoride, calcitonin, or bisphosphonates within the prior six months. It also excluded those with serious renal impairment, metabolic bone diseases other than osteoporosis, and conditions requiring chronic glucocorticoid use. In real-world osteoporosis clinics, glucocorticoid-induced bone loss is one of the most common referral reasons. The FIT population was healthier than the average patient who receives an alendronate prescription today, per FDA labeling for alendronate.

BMD threshold and risk stratification

Eligibility required a femoral-neck BMD T-score of <−1.6 but did not require <−2.5 (the WHO threshold for osteoporosis). Roughly 37% of enrolled women were osteopenic rather than osteoporotic. This diluted the at-risk denominator and may have contributed to the null primary endpoint result.

The HealthRX.com Limitation-Severity Framework for FIT-2

To organize how each limitation affects clinical decision-making, we scored the five most consequential weaknesses on two axes: (1) how likely the limitation is to change the direction or magnitude of the treatment effect, and (2) how often the limitation arises in routine prescribing.

Clinicians can use this matrix when counseling patients who do not resemble the trial population. A limitation that scores "High" on both axes deserves explicit mention in shared decision-making.

The Primary Endpoint Problem

FIT-2's prespecified primary endpoint was clinical (symptomatic) fractures, a composite of nonvertebral and clinically recognized vertebral fractures. The trial did not meet this endpoint (relative hazard 0.86, 95% CI 0.73 to 1.01, p = 0.07). The statistically significant 47% reduction applied to morphometric vertebral fractures, a secondary endpoint detected on scheduled lateral spine radiographs that participants never felt.

This is not a trivial distinction. Morphometric fractures identified only through protocol-mandated imaging include many that would never come to clinical attention. Whether preventing a radiographic compression that causes no symptoms carries the same weight as preventing a hip fracture is a question FIT-2 did not answer.

A post hoc subgroup analysis of women with baseline T-scores at or below −2.5 did show a significant reduction in clinical fractures (RH 0.64, 95% CI 0.50 to 0.82). This finding, while encouraging, was not prespecified and cannot carry the same evidentiary weight as a primary analysis. Subsequent guidelines from the National Osteoporosis Foundation leaned on this subgroup result when recommending alendronate for women meeting the WHO osteoporosis threshold, a reasonable but not bulletproof inference.

Follow-Up Duration and Long-Term Safety Gaps

Mean follow-up was approximately 4.2 years. Osteoporosis is a condition measured in decades. The trial therefore could not and did not address:

• Whether fracture reduction persists beyond five years or whether bone quality plateaus.
• Long-term rare adverse events, specifically osteonecrosis of the jaw (ONJ) and atypical femoral fractures (AFF), both of which emerged only in post-marketing surveillance years later.
• The optimal duration of therapy and whether drug holidays are safe. The FLEX extension study attempted to answer this question, but FLEX enrolled only FIT-1 completers and was not powered for fracture endpoints.

The absence of long-term safety data from FIT itself means that every conversation about bisphosphonate holidays relies on evidence gathered outside this trial's boundaries.

Statistical Caveats Worth Noting

Multiple comparisons. FIT-2 reported results across numerous skeletal sites (spine, hip, wrist, all nonvertebral combined) and multiple subgroups. No formal multiplicity adjustment was described in the primary publication. The 47% vertebral fracture reduction, while consistent with FIT-1 findings, should be interpreted knowing it was one of many comparisons performed.

Adjudication of morphometric fractures. Vertebral deformities were graded semiquantitatively using Genant criteria, with a requirement for >20% height reduction. This threshold, while standard, means borderline deformities near the cutoff could shift between "fracture" and "no fracture" depending on reader measurement. The publication does not report inter-reader agreement statistics for the central radiographic reading facility.

Calcium and vitamin D co-administration. All participants received 500 mg of elemental calcium and (if baseline 25-hydroxyvitamin D was low) up to 250 IU of vitamin D daily. This design means FIT-2 tested alendronate plus calcium/vitamin D versus placebo plus calcium/vitamin D. The effect of alendronate alone, without supplementation, was never isolated and remains unknown from this dataset.

Conflict-of-Interest Considerations

FIT-2 was funded by Merck & Co., the manufacturer of alendronate (Fosamax). Several authors disclosed consulting relationships or research funding from Merck. This was common in 1990s osteoporosis research and does not by itself invalidate the results, but it does introduce potential biases worth naming:

• Protocol design. The sponsor participated in study design. Whether the choice of clinical fractures (a harder-to-reach endpoint) as primary, versus morphometric fractures (which showed benefit), reflected scientific judgment or strategic calculation is unknowable from published disclosures alone.
• Data access. The publication states Merck maintained the database. Independent academic data access was not described in the primary paper, a transparency standard that would be expected by contemporary reporting guidelines like ICMJE 2023 recommendations.
• Publication framing. The abstract and conclusions of the JAMA paper emphasize the significant vertebral fracture result without leading with the null primary endpoint. While the data are present in the manuscript, the emphasis choices merit awareness.

Later independent meta-analyses, including a Cochrane review of bisphosphonates for postmenopausal osteoporosis, confirmed vertebral fracture reduction with alendronate across trials. This external validation reduces (but does not eliminate) concern about sponsor influence on any single study.

Generalizability Gaps in 2026 Practice

Several shifts in osteoporosis management make FIT-2's direct applicability narrower today:

1. Competing therapies. Denosumab, romosozumab, and teriparatide now offer alternatives. FIT-2 cannot inform comparative-effectiveness decisions because no active comparator arm existed.
2. Treat-to-target approaches. Some guidelines now recommend treating to a BMD target rather than a fixed duration. FIT-2's design (fixed dose, fixed duration) does not support or refute this strategy.
3. Sequential therapy paradigms. Current expert opinion, reflected in the 2020 AACE/ACE guidelines, often favors starting with an anabolic agent in very high-risk patients and following with an antiresorptive. FIT-2 tested alendronate as first-line monotherapy and cannot speak to sequencing strategies.
4. FRAX integration. The FRAX fracture risk calculator, now central to treatment decisions in many countries, did not exist when FIT-2 was designed. The trial's inclusion criteria (BMD alone, no fracture risk score) do not align with modern risk-stratified prescribing.

What Post-Publication Commentary Surfaced

Letters to the editor and subsequent editorials raised several points after the 1998 publication:

• The disconnect between the null primary endpoint and positive secondary vertebral endpoint drew attention. Some correspondents argued that morphometric vertebral fractures should have been the primary endpoint from the start, given FIT-1's results. Others countered that switching endpoints post hoc would be inappropriate.
• Questions arose about the dose increase from 5 mg to 10 mg at 24 months, mid-trial, based on emerging data suggesting the higher dose achieved greater BMD gains. While pragmatic, this means participants were not on a uniform regimen throughout the study, complicating dose-response interpretation.
• The relatively low rate of GI adverse events in FIT-2 (not significantly different from placebo) was questioned by clinicians seeing high rates of esophagitis and dyspepsia in practice. Strict exclusion of patients with active GI disease likely explains this gap between trial and real-world tolerability.

The Bottom Line for Clinicians

FIT-2 remains a foundational bisphosphonate trial, and alendronate's vertebral fracture benefit is real and replicated. But the trial's limitations are not academic footnotes. They are directly relevant every time a clinician prescribes alendronate to a patient who is male, non-white, on glucocorticoids, or medically complex, because that patient was not represented in FIT-2. Honest appraisal of what this trial did and did not prove is the starting point for informed prescribing.

Frequently asked questions

›Did the FIT trial meet its primary endpoint?

No. FIT-2's primary endpoint was clinical (symptomatic) fractures, and the result was not statistically significant (p = 0.07). The widely cited 47% reduction refers to morphometric vertebral fractures, a secondary endpoint detected on protocol X-rays. This distinction affects how strongly the evidence supports alendronate in lower-risk populations without prevalent fractures.

›Why was the FIT trial population so homogeneous?

Over 96% of FIT-2 participants were white postmenopausal women recruited from U.S. academic centers. This reflected 1990s trial recruitment norms and the demographics of volunteering populations at those sites. It limits direct generalizability to Black, Hispanic, Asian, and male patients.

›Was the FIT trial funded by the drug manufacturer?

Yes. Merck & Co., the maker of alendronate (Fosamax), funded FIT-2 and participated in study design. Several investigators had consulting or financial relationships with Merck. Independent academic data access was not explicitly described in the primary publication.

›Does the FIT trial tell us anything about long-term bisphosphonate safety?

Very little. With a mean follow-up of about 4.2 years, FIT-2 could not detect rare long-term complications like atypical femoral fractures or osteonecrosis of the jaw. Those risks emerged through post-marketing surveillance and later observational studies, not from FIT itself.

›Why did the alendronate dose change mid-trial?

Participants started on 5 mg daily, then switched to 10 mg daily at month 24 based on emerging external evidence that the higher dose produced greater BMD gains. While pragmatic, this means the trial did not test a single uniform dose throughout, which complicates dose-response interpretation.

›Can FIT results be applied to patients on glucocorticoids?

Not directly. FIT-2 excluded women requiring chronic glucocorticoid therapy. Steroid-induced osteoporosis has a different pathophysiology (suppressed bone formation, not just increased resorption), so separate trial data are needed to support alendronate use in that population.

›How does FIT compare to more recent osteoporosis trials?

FIT-2 tested alendronate against placebo without an active comparator. Modern trials like ARCH (romosozumab vs. alendronate) and the head-to-head denosumab studies provide comparative effectiveness data that FIT cannot. FIT remains relevant as foundational efficacy evidence but is insufficient for choosing between current treatment options.

›Were GI side effects underrepresented in the FIT trial?

Likely yes. FIT-2 excluded patients with active gastrointestinal disease, and the trial's GI adverse event rates were not significantly different between alendronate and placebo. Real-world tolerability, where patients often have reflux or use NSAIDs, tends to be worse than what FIT-2 reported.

›What is the clinical significance of a morphometric vertebral fracture?

Morphometric vertebral fractures are height reductions (>20%) detected on X-ray, many of which cause no symptoms. While they predict future clinical fractures, not all morphometric fractures lead to pain or disability. FIT-2's vertebral fracture benefit therefore includes events of variable clinical importance.

›Should clinicians still prescribe alendronate based on FIT data?

Yes, but with appropriate context. FIT-2 confirmed vertebral fracture reduction in postmenopausal women with low BMD, and subsequent meta-analyses support this finding. The limitations matter most for patients who differ substantially from the trial population, where shared decision-making should acknowledge the evidence gaps.

References

• Cummings SR, Black DM, Thompson DE, et al. Effect of alendronate on risk of fracture in women with low bone density but without vertebral fractures: results from the Fracture Intervention Trial. JAMA. 1998;280(24):2077-2082. PubMed
• Black DM, Schwartz AV, Ensrud KE, et al. Effects of continuing or stopping alendronate after 5 years of treatment: the Fracture Intervention Trial Long-term Extension (FLEX). JAMA. 2006;296(24):2927-2938. PubMed
• Wells GA, Cranney A, Peterson J, et al. Alendronate for the primary and secondary prevention of osteoporotic fractures in postmenopausal women. Cochrane Database Syst Rev. 2008;(1):CD001155. PubMed
• 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
• Cosman F, de Beur SJ, LeBoff MS, et al. Clinician's Guide to Prevention and Treatment of Osteoporosis. Osteoporos Int. 2014;25(10):2359-2381. PubMed
• FDA. Fosamax (alendronate sodium) prescribing information. accessdata.fda.gov