FIT Extension Data and What Happened After the Trial Ended

At a glance

| Detail | Value | |---|---| | Original Trial | FIT (Fracture Intervention Trial), two arms: Vertebral Fracture Arm (FIT-1, N = 2,027) and Clinical Fracture Arm (FIT-2, N = 4,432) | | Extension Study | FLEX (Fracture Intervention Trial Long-term Extension), N = 1,099 | | Intervention | Alendronate 5 mg/day (later 10 mg/day) vs. placebo | | Duration | FIT: 3-4 years; FLEX: additional 5 years (total ~10 years exposure) | | Primary Endpoint (FIT) | New morphometric vertebral fractures | | Key Result (FIT) | 47% relative risk reduction in vertebral fractures (RR 0.53 to 95% CI 0.41-0.68) | | Key Result (FLEX) | Continuing alendronate for 10 years vs. stopping at 5 reduced clinical vertebral fractures (2.4% vs. 5.3%, RR 0.45) but not total nonvertebral fractures |

Why Extension Data Matters More Than the Original Abstract

The FIT trial answered whether alendronate works. It did not answer how long to keep prescribing it, what happens when patients stop, or whether rare harms accumulate with prolonged use. These are the questions clinicians actually face in practice, and they required a decade of post-trial observation to address.

Bisphosphonates bind to hydroxyapatite in bone and persist for years after the last dose. That pharmacokinetic reality means the offset curve after discontinuation looks nothing like stopping a statin or an antihypertensive. The drug is still present, still suppressing osteoclasts, long after the prescription ends. Understanding durability required the FLEX trial and subsequent registry analyses.

The FLEX Extension: Study Design and Who Was Enrolled

FLEX (published 2006, JAMA) re-enrolled 1,099 women from the original FIT cohorts who had taken alendronate for a mean of 5 years. Participants were randomized to continue alendronate (5 or 10 mg daily) or switch to placebo for an additional 5 years.

Key design details that matter clinically:

• Only women who tolerated alendronate for 5 years and were willing to continue were eligible. This creates a healthy-adherer bias that inflates the safety profile compared to real-world use.
• Women with femoral neck BMD T-scores below -3.5 at FLEX enrollment were excluded, removing the highest-risk patients from the continuation vs. discontinuation comparison.
• The study was not powered for hip fracture as an endpoint. Total nonvertebral fracture was a secondary endpoint, and the trial had limited statistical power for rare events.

What Happened to Bone Density After Stopping

BMD trajectories after alendronate discontinuation followed a predictable pattern based on the drug's skeletal half-life.

| Site | Alendronate Group (10 yr) | Placebo Group (stopped at 5 yr) | Difference | |---|---|---|---| | Total hip BMD change (years 5-10) | +0.46% | -2.36% | ~2.8 percentage points | | Lumbar spine BMD change (years 5-10) | +3.74% | -1.48% | ~5.2 percentage points | | Femoral neck BMD change (years 5-10) | +0.46% | -1.71% | ~2.2 percentage points |

The placebo group (those who stopped) lost BMD gradually, not precipitously. After 5 years off drug, their hip BMD returned roughly to levels seen at FIT randomization a decade earlier. Spine BMD declined but remained above original baseline. This pattern confirmed that bisphosphonate offset is slow and partial, unlike the rapid rebound seen with denosumab discontinuation.

Fracture Outcomes: Who Needed to Continue?

The headline fracture results from FLEX require careful reading.

| Outcome | Continue Alendronate (N=662) | Switch to Placebo (N=437) | Relative Risk (95% CI) | |---|---|---|---| | Clinical vertebral fractures | 2.4% | 5.3% | 0.45 (0.24-0.85) | | Morphometric vertebral fractures | 9.8% | 11.3% | 0.86 (0.60-1.22) | | All nonvertebral fractures | 19.0% | 18.9% | 1.00 (0.76-1.32) |

The statistically significant benefit of continuation was limited to clinical vertebral fractures. For nonvertebral fractures, including hip fractures, there was no measurable difference between continuing and stopping. A pre-specified subgroup analysis found that women who entered FLEX with a femoral neck T-score at or below -2.5 had higher fracture rates when switched to placebo, suggesting this subgroup benefits most from extended treatment.

This subgroup finding became the basis for current American Association of Clinical Endocrinology guidelines recommending reassessment after 5 years of oral bisphosphonate therapy, with continuation reserved for patients who remain at high fracture risk.

The Bisphosphonate Holiday Concept

FLEX did not use the phrase "drug holiday." That language emerged from clinical interpretation of the data over the following years. The reasoning: if stopping alendronate after 5 years does not increase nonvertebral fracture risk for most women, and if rare cumulative toxicities (atypical femoral fractures, osteonecrosis of the jaw) correlate with duration of use, a planned pause could optimize the risk-benefit ratio.

The FDA's 2012 review of bisphosphonate safety cited the FLEX data explicitly when discussing duration of therapy. The agency concluded there was no clear benefit to continuing beyond 3 to 5 years for patients not at high risk, though it stopped short of mandating discontinuation.

Current practice patterns, shaped directly by FLEX, typically follow this approach:

• Lower-risk patients (T-score above -2.5 at reassessment, no prior fractures): consider stopping after 5 years of oral therapy. Reassess with DXA in 2 to 3 years.
• Higher-risk patients (T-score at or below -2.5, prior vertebral fracture, ongoing glucocorticoid use): continue for up to 10 years, then reassess.

Late Safety Signals the Original FIT Could Not Detect

The original FIT publication reported a side-effect profile dominated by upper GI complaints. Two serious adverse events emerged only with prolonged post-marketing surveillance.

Atypical Femoral Fractures (AFFs)

AFFs are stress fractures of the femoral shaft that occur with minimal trauma, often preceded by prodromal thigh pain. They were not identified in FIT or FLEX because they are rare (estimated 3.2 to 50 cases per 100,000 person-years of bisphosphonate use) and were not a pre-specified endpoint in either trial.

A Swedish registry study of over 12,000 women with subtrochanteric fractures found that AFF risk increased with duration of bisphosphonate use, peaking after 4 to 7 years. Risk declined after discontinuation. This duration-dependent pattern is the strongest argument for time-limited therapy.

Osteonecrosis of the Jaw (ONJ)

ONJ was first reported in 2003, years after FIT. Incidence in oral bisphosphonate users for osteoporosis is very low (estimated 1 in 10,000 to 1 in 100,000 patient-years), but it is substantially higher in cancer patients receiving high-dose intravenous bisphosphonates. Neither FIT nor FLEX captured ONJ events. The condition's association with dental procedures means it functions as a practical consideration for treatment duration rather than a common clinical complication.

Bone Turnover Markers: The Suppression Plateau

FLEX provided important mechanistic data through bone turnover markers (BTMs). After 5 years of alendronate, markers of bone resorption (urinary NTX, serum CTX) were suppressed 50% to 70% below untreated baseline. When alendronate was stopped, these markers rose gradually over 1 to 2 years but did not fully return to pre-treatment levels even after 5 years off drug.

This persistent suppression distinguishes bisphosphonates from denosumab, where BTMs rebound above baseline within 3 to 6 months of the last injection. The slow return of bone turnover after bisphosphonate discontinuation explains why fracture protection partially persists during a drug holiday and why the concept of a holiday is pharmacologically plausible only for this drug class.

Limitations the Authors Acknowledged

The FLEX investigators were transparent about several constraints:

1. Selection bias. Only adherent, tolerant patients from FIT entered FLEX. Real-world populations include patients who discontinued for GI intolerance, which limits generalizability.
2. Inadequate power for hip fracture. The sample size could not detect clinically meaningful differences in hip fracture rates between continuation and discontinuation.
3. No true treatment-naive placebo arm. Every FLEX participant had already received 5 years of alendronate, so the "placebo" group was not untreated. The study cannot address whether residual drug effect after 5 years differs from never having been treated.
4. Exclusion of highest-risk women. Those with T-scores below -3.5 were excluded, precisely the group most likely to fracture and most likely to benefit from continuation.

What This Means for Current Practice

The combined FIT-FLEX evidence base established three principles that still guide osteoporosis management.

First, alendronate works. The 47% vertebral fracture reduction from FIT remains one of the strongest effect sizes for any osteoporosis intervention and has been replicated across populations.

Second, indefinite treatment is not necessary for most patients. The residual benefit after 5 years of therapy provides a pharmacologic buffer that allows planned interruptions without immediate fracture consequences for moderate-risk patients.

Third, treatment duration decisions should be individualized based on femoral neck T-score at reassessment, fracture history, and ongoing risk factors. The FDA's alendronate label reflects this by recommending periodic reassessment of the need for continued therapy.

Frequently asked questions

›

›

›

›

›

›

›

›

›

›

References

1. Black DM, Cummings SR, Karpf DB, et al. Randomised trial of effect of alendronate on risk of fracture in women with existing vertebral fractures. Lancet. 1996;348(9041):1535-1541. PubMed
2. 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
3. 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
4. Schilcher J, Michaelsson K, Aspenberg P. Bisphosphonate use and atypical fractures of the femoral shaft. N Engl J Med. 2011;364(18):1728-1737. PubMed
5. Whitaker M, Guo J, Engel T, Brown A. Bisphosphonates for osteoporosis: where do we go from here? N Engl J Med. 2012;366(22):2048-2051. PubMed
6. Fosamax (alendronate sodium) prescribing information. FDA. Label