How to Safely Stop Fosamax (Alendronate): A Physician-Level Discontinuation Protocol

What the FDA Label Covers, and What It Doesn't

The FDA-approved label for alendronate covers treatment and prevention of osteoporosis and does not itself specify a discontinuation protocol or a "drug holiday" [3]. The holiday concept comes from professional society guidance, chiefly the American Society for Bone and Mineral Research (ASBMR) Task Force report [4], the American College of Physicians living guideline [5], and the Endocrine Society's clinical practice guideline [7], all of which interpret long-term trial data rather than restate label instructions. That distinction matters: what follows is a synthesis of guideline recommendations and trial evidence, applied through clinical judgment. It is not an FDA-mandated sequence, and an individual prescriber may reasonably deviate from it based on a patient's full history.

How Alendronate Works at the Bone Level

Alendronate is a nitrogen-containing bisphosphonate. It binds hydroxyapatite in bone mineral with high affinity, and osteoclasts take up the drug during normal bone resorption [1][2]. Once internalized, alendronate inhibits farnesyl pyrophosphate synthase in the mevalonate pathway, which disrupts osteoclast function and promotes osteoclast apoptosis [2]. The practical effect is that bone resorption slows and bone density is preserved or modestly increased over time.

This binding mechanism is why stopping the drug does not cause an immediate rebound in bone loss. Alendronate remains embedded in bone matrix for a long time; the drug's plasma half-life is only a few hours, but its skeletal retention is measured in years [2][3]. As old bone is slowly remodeled, stored drug re-enters the local environment and continues to act on newly recruited osteoclasts. This pharmacokinetic behavior is the basis for the bisphosphonate holiday concept: without it, stopping any anti-resorptive drug would carry an immediate fracture-risk penalty.

The original Fracture Intervention Trial (FIT) established alendronate's clinical benefit in women with existing vertebral fractures, showing a roughly 50% reduction in new vertebral fractures and a benefit against hip fracture in this higher-risk group over about 3 years [1]. Any discontinuation strategy is judged against how much of that benefit it preserves.

Who Is Generally a Candidate for a Bisphosphonate Holiday

Not every patient should stop. The 2016 ASBMR Task Force on bisphosphonate-associated atypical femoral fractures recommended reassessing fracture risk after about 5 years of oral bisphosphonate therapy and considering a holiday for patients at comparatively lower risk [4]. The 2023 American College of Physicians guideline adds more specific criteria: patients treated for 3 to 5 years with an oral bisphosphonate, with a femoral neck T-score above -2.5 and no history of vertebral or hip fracture, are reasonable candidates to pause [5]. FRAX scoring can add nuance: a 10-year probability of major osteoporotic fracture below roughly 20% and hip fracture probability below roughly 3% supports holiday eligibility, though exact thresholds vary by guideline and by country-specific FRAX calibration [8].

Patients who are generally poor candidates for a holiday include those with:

• A femoral neck T-score at or below -2.5
• A prior vertebral compression fracture, including one found incidentally on imaging
• A prior hip fracture
• Current glucocorticoid therapy above roughly 5 mg prednisone-equivalent daily
• A 10-year FRAX hip fracture probability at or above 3%

The femoral neck T-score threshold comes largely from the FLEX extension study, where women who stopped alendronate and already had a femoral neck T-score below -2.5 carried the highest subsequent vertebral fracture risk [6].

The FLEX Trial: What Actually Happens When You Stop

FLEX (the FIT Long-term Extension) is the most informative dataset on alendronate discontinuation. It enrolled women who had already taken alendronate for about 5 years in FIT, then randomized them to continue alendronate or switch to placebo for another 5 years [6].

At year 10 (5 years after stopping, in the placebo arm):

• Total hip bone density declined modestly in the placebo group, while the continuation group saw a small additional gain [6]
• Lumbar spine bone density declined further in the placebo group, while the continuation group continued to gain [6]
• Bone turnover markers rose gradually after stopping but generally stayed below pre-treatment levels through year 10 [6][9]
• Clinical vertebral fracture risk was higher in the placebo group than the continuation group, a statistically significant difference [6]
• Nonvertebral fracture rates did not differ significantly between the two groups [6]

A key detail is that the placebo group did not experience a sharp bone-loss rebound. The decline was gradual, which is consistent with a slowly waning skeletal drug reservoir rather than an abrupt loss of protection. The Endocrine Society guideline describes this as a residual effect that distinguishes bisphosphonates from denosumab, where stopping produces a much faster and sometimes more pronounced increase in bone resorption [7].

Step-by-Step: How This Decision Is Typically Structured

This sequence reflects how ASBMR, ACP, and Endocrine Society guidance is generally applied in practice [4][5][7]. It is a framework for the discussion with your prescriber, not a self-directed protocol, and an individual clinician may reasonably adjust the order or add steps based on your history, labs, and imaging.

1. Confirm duration of therapy. The standard reassessment point is around 5 years of continuous oral bisphosphonate use, though some guidance accepts 3 years as a minimum when fracture risk is clearly low [5].

2. Obtain a current DXA scan. A scan close to the planned stop date, ideally within the prior 6 months, provides the reference point for future comparisons, with T-scores recorded at the spine, femoral neck, and total hip.

3. Calculate FRAX. Using femoral neck bone density in a country-specific FRAX calculation helps quantify 10-year hip and major osteoporotic fracture probability [8].

4. Review exclusion criteria. A single prior vertebral fracture, even an asymptomatic one found on imaging, generally argues against a standard holiday.

5. Stop the drug without a taper. Because alendronate's effect depends on skeletal binding rather than ongoing circulating drug, there is no pharmacologic withdrawal to manage; the last scheduled dose is simply the last dose.

6. Maintain calcium and vitamin D intake. Guidance generally supports calcium intake around 1,000 to 1,200 mg daily from diet plus supplements, and vitamin D sufficiency, throughout a holiday period [7]. This does not replace anti-resorptive therapy on its own; inadequate intake can accelerate bone loss regardless of prior bisphosphonate use.

7. Schedule monitoring in advance, rather than leaving the next check-in open-ended (see the framework below).

Monitoring During the Holiday

Bone turnover markers respond faster than DXA and can serve as an early signal. Serum C-terminal telopeptide (CTX), a resorption marker, typically stays suppressed for roughly 6 to 12 months after the last dose and then rises gradually as the skeletal drug reservoir depletes [9]. Procollagen type I N-terminal propeptide (P1NP) tends to follow a similar pattern. A marker that returns toward pre-treatment levels suggests the drug's effect is fading, though single time-point marker values are noisy and are generally used alongside DXA rather than in place of it.

DXA remains the more definitive check. Densitometry consensus guidance describes a significant change as one that exceeds the scanner-specific least significant change, often cited in the range of roughly 3 to 5% at the total hip, though the exact figure depends on the individual scanner and technologist precision and should be confirmed with the imaging center [11].

Clinician Discussion and Monitoring Framework

This is not a fixed treatment schedule. It is a structure for the conversation with your prescriber at each checkpoint, distinguishing what comes from guideline consensus versus what depends on your individual case.

The right-hand column is the important one. Guideline thresholds (T-score cutoffs, FRAX percentages, 5-year ceilings) are population-level defaults built from trial data; they are a reasonable starting point but not a diagnosis of your individual risk. Your prescriber may weigh factors the guidelines do not fully capture, such as fall risk, family history, or how your bone density has trended over years rather than a single scan.

When to Discuss Restarting Therapy

Restarting alendronate, or switching to another agent, is generally worth discussing if any of the following occur during a holiday:

1. A new fragility fracture at the hip, spine, proximal humerus, or wrist [5].
2. A meaningful BMD decline, confirmed against the scanner's precision error rather than a single raw percentage [11].
3. Bone turnover markers rising well above the expected range within the first year or so off therapy, particularly if both resorption and formation markers move together [9].
4. A new risk factor, such as starting systemic glucocorticoids, a new diagnosis affecting bone metabolism, or substantial unintentional weight loss [7].
5. Reaching the guideline-cited 5-year ceiling without another trigger, at which point most guidance calls for a fresh evaluation rather than an open-ended continuation [4][6].

The re-treatment choice does not have to be alendronate again. Patients with GI side effects on oral bisphosphonates sometimes move to annual zoledronic acid infusion. Patients whose reassessment shows very high fracture risk may be candidates for an anabolic agent (teriparatide or romosozumab) before returning to anti-resorptive maintenance [7]. This decision depends on individualized risk and tolerability and is not something this article can determine for a specific patient.

Why the Holiday Exists: Duration-Linked Risks

Bisphosphonate holidays exist largely because of two rare but serious risks that appear to track with cumulative duration of use: atypical femoral fracture (AFF) and osteonecrosis of the jaw (ONJ).

AFFs are stress fractures of the femoral shaft or subtrochanteric region that occur with little or no trauma. A widely cited epidemiologic review describes AFF incidence rising with longer bisphosphonate exposure and declining after the drug is stopped [12]. Different cohort studies report somewhat different absolute incidence figures and risk ratios, including variation by ethnicity, so a specific number (for example, an exact incidence-per-100,000 figure or a specific hazard ratio) should be checked against the primary study behind it before being presented to a patient as a fixed statistic; this article intentionally does not present one as settled. What is consistent across the literature is the direction of the effect: risk is low overall, rises with years of continuous use, and drops meaningfully within roughly a year of stopping.

ONJ is the other duration-linked concern, though its incidence in patients taking oral bisphosphonates at osteoporosis doses is very low, generally estimated in the range of 1 per 10,000 to 1 per 100,000 patient-years [13]. Risk is substantially higher with intravenous bisphosphonates used at oncologic doses, which is a different clinical context. For oral alendronate at standard osteoporosis dosing, ONJ risk alone rarely drives the decision to pause therapy, but it reinforces the value of reassessing duration around the 5-year mark.

For patients who meet holiday criteria, a pause preserves most of the anti-fracture benefit through the residual skeletal drug reservoir while reducing cumulative exposure tied to these rare risks. For patients who do not meet the criteria (low T-scores, prior fractures), the fracture risk from stopping generally outweighs the small absolute risk of AFF or ONJ.

Special Populations

Premenopausal women on alendronate for glucocorticoid-induced osteoporosis: this population was not represented in FLEX. If the glucocorticoid is discontinued and bone density has stabilized outside the osteoporotic range, a holiday may be reasonable; if glucocorticoids continue, stopping alendronate is generally not advised without an individualized discussion [7].

Men with osteoporosis: the evidence base for bisphosphonate holidays in men is thin. The Endocrine Society guideline suggests applying similar T-score and FRAX thresholds used for postmenopausal women, while noting that direct trial data in men are limited [7].

Patients switching from denosumab to alendronate: this is a different clinical scenario, not a holiday. Denosumab discontinuation causes rapid resorption rebound, and a course of a bisphosphonate is often used specifically to blunt that rebound rather than as standalone long-term therapy [14].

Patients with reduced kidney function (eGFR below roughly 35 mL/min): alendronate is contraindicated at this level of renal impairment. If kidney function declines during therapy, the drug should be stopped for safety reasons, not as a planned holiday, and this requires coordination with the prescribing clinician and often nephrology [3].

Alendronate Discontinuation vs. Denosumab Discontinuation

These two commonly prescribed anti-resorptive drugs behave very differently when stopped, which matters because patients sometimes assume all osteoporosis medications can be paused the same way.

Alendronate's long skeletal retention means its effect persists for a meaningful period after the last dose [2][3]. Denosumab (Prolia), a RANKL inhibitor, has no comparable skeletal reservoir; its effect wanes within about 6 months of the last injection, and bone turnover can rebound above pre-treatment levels, a phenomenon associated with vertebral fractures clustering in the months after a missed or stopped dose [14]. Stopping alendronate after an appropriate holiday evaluation is a comparatively low-risk maneuver for eligible patients. Stopping denosumab without a bridging bisphosphonate is a different and potentially higher-risk situation. These two protocols should not be treated as interchangeable.