Too Much of a Good Thing: The Evidence Behind the Need for a Bisphosphonate Holiday

May 9, 2013

By Jenna Piccininni

Faculty Peer Reviewed

Bisphosphonates are a relatively new medication having only been approved to treat osteoporosis in the US since 1995 [1]. In addition, large placebo controlled trials have, at most, 10 years of follow-up data. Thus, there are still questions regarding the long-term use of these agents. There are a few well-established side effects of bisphosphonates including rare osteonecrosis of the jaw and more common esophageal irritation. However, several more recent case reports suggest a correlation between prolonged bisphosphonate use and atypical femoral fracture [2]. This brings into question whether too much antiresorptive activity can lead to increased fracture risk and if a medication holiday is therefore appropriate.

The pharmacology and mechanism of bisphosphonates is integral in understanding the biological underpinnings of this potential association and the implications of discontinuing the medication. Once absorbed into the blood stream there is no systemic metabolism and the molecules directly bind to active bone [1]. These binding sites on bone are so numerous that saturation is physiologically unrealistic. In the setting of osteoclast activity an acidic microenvironment is created which causes the bisphosphonates to be released and taken up by these cells, leading to loss of osteoclast function and ultimately apoptosis. This decrease in bone resorption leads to increased bone density but also suppress the repair of microdamage that occurs with normal daily activities. One study in Beagle dogs examined how the use of bisphosphonates, at six times the clinical dose, affected bone structure [3]. In control dogs resorption spaces were more numerous but more likely to be associated with areas of microdamage cracks whereas treated dogs had a higher proportion of cracks that were not associated with resorption spaces. These findings suggest that bisphosphonates impair targeted repair of microdamage, which is the proposed mechanism by which long-term use is thought to lead to increased fracture risk.

Based on these case reports, controlled studies have looked for a significant association between prolonged bisphosphonate use and risk of atypical fracture. In 2010 a NEJM study used data from the FIT, FLEX, and HORIZON PFT large randomized controlled trials to assess this association and found no increased risk [4]. However in 51,287 patient years only 12 atypical fractures were identified resulting in wide confidence intervals and low power. Despite the lack of evidence for definitive causation, in 2010 the FDA issued a label change for all bisphosphonates requiring the risk of atypical subtrochanteric femur fractures to be included [5]. Then in 2011 a paper in JAMA reported the results of an epidemiological population based case-control study showing that women using bisphosphonates for more than five years have 2.74 times the odds of being hospitalized for a subtrochanteric or femoral neck fracture than women with only sporadic use (<100 days) [6]. Following 5 years of use the risk of fracture was .13% in the first year and .22% in two years. 64% of the fractures in long-term users were attributable to bisphosphonate use, but because the total incidence of these fractures is so low eliminating exposure would lead to only an 11% decrease in the total rate of fracture. While this study shows a significant association between prolonged bisphosphonate use and atypical fracture, the evidence should be interpreted keeping in mind that confounders may be present in this observational study. Larger controlled studies must be done confirm these findings. If this association is present, the effects of limiting bisphosphonate exposure though medication holidays could be beneficial.

The idea of a bisphosphonate holiday is particularly appealing due to the unsaturable accumulation of the molecules in bone allowing for continued effectiveness even after discontinuation of the medication. The Fracture Intervention Trial Long-term Extension (FLEX) study randomized women who previously had three to five years of alendronate in the FIT trial to receive either 5 more years of 10 mg alendronate/day, 5mg alendronate/day, or placebo [7]. Their results showed that while bone mineral density (BMD) at the hip decreased 2-3% in the placebo group this was less BMD loss than expected in women never treated with bisphosphonates and that their BMD remained above baseline levels at the start of the FIT trial. In addition, there was no difference in the rate of non-vertebral fractures between any of the groups, showing that loss of hip BMD in the placebo group was not clinically relevant. However, the women in the alendronate arms had a 2.9% absolute risk reduction and 55% relative risk reduction for clinical vertebral fractures compared to placebo, but the rate of morphometric fracture was equal. This benefit was most appreciable in women with past vertebral fractures and very low BMD. Another study in NEJM similarly followed women for 10 years assigned to either 10mg/day alendronate, 5mg/day alendronate, or 5mg/day alendronate for 5 years followed by 5 years of placebo [8]. They similarly found that changes in BMD in the placebo group varied depending on the site, but that the risk of morphometric vertebral fractures did not differ between the three groups. This data shows that after treatment for five years with a bisphosphonates a five-year holiday is only detrimental to women at high risk for fracture with very low BMD and previous fractures.

Although no official recommendations have been made regarding bisphosphonate holidays, some authors have proposed personal guidelines based upon the patient’s risk of fracture. Watts and Diab suggest an indefinite holiday for patients at mild risk after 5 years of treatment, a two to three year holiday for moderate risk patients after 5-10 years of treatment, and a 1-2 year “holiday” for high-risk patients after 10 years of treatment during which they receive an alternative medication such as a SERM [1]. During these holidays bisphosphonates should be restarted if BMD decreases significantly or if fracture occurs. I believe that the aforementioned studies support this type of algorithm. Discontinuing bisphosphonates after several years of treatment does not appear to have clinical implications in most low risk patients most likely due to their storage in bone and residual effects. Data regarding risk of atypical fracture with long-term treatment is not as clear but in the setting of this potential debilitating side effect physicians that are risk-averse are justified in giving their patients a medication holiday based on their clinical scenario. More controlled studies must be done examining the clinical outcomes of specific holiday durations, the optimal time to begin a holiday and indications for ending one. With that information more solid evidence based guidelines can be constructed, patient safety can be improved, and the use of these relatively new medications can be optimized.

Jenna Piccininni is a 3rd year medical student at NYU School of Medicine

Peer reviewed by Michael Tanner, Associate Editor, Clinical Correlations

Image courtesy of Wikimedia Commons


1. Watts N, Diab D. Long-term use of bisphosphonates in osteoporosis. Journal of Clinical Endocrinology and Metabolism. 2010;95(4):1555-1565.

2. Lenart B, Lorich D, and Lane J. Atypical fractures of the femoral diaphysis in postmenopausal women taking alendronate. New England Journal of Medicine. 2008;358:1304-1306.

3. Li J, Mashiba T, Burr D. Bisphosphonate treatment suppresses not only stochastic remodeling but also the targeted repair of microdamage. Calcification Tissue International. 2001;69:281-286.

4. Black D, Kelly M, Genant H, et al. Bisphosphonates and fractures of the subtrochanteric or diaphyseal femur. The New England Journal of Medicine. 2010;362(19):1761-1771.

5. Bisphosphonates 9osteoporosis drugs): Label change-Atypical fractures update. Food and Drug Administration Website.  Accessed July 25, 2012.

6. Park-Wyllie L, Mamdani M, Juurlink D, et al. Bisphosphonate use and the risk of subtrochanteric for femoral shaft fractures in older women. Journal of the American Medical Association. 2011;305(8): 783-789.

7. Black D, Schwartz A, Ensrud K, et al. Effects of continuing or stopping alendronate after 5 years of treatment The fracture intervention trial long-term extension (FLEX): A randomized trial. Journal of the American Medical Association. 2006;296(24):2927-2938.

8. Bone H, Hosking D, Devogelaer J, et al. Ten years’ experience with alendronate for osteoporosis in postmenopausal women. The New England Journal of Medicine. 2004;350:1189-1199.

4 comments on “Too Much of a Good Thing: The Evidence Behind the Need for a Bisphosphonate Holiday

  • Avatar of sorin herscovici
    sorin herscovici on

    In Watts and Diab holiday duration suggestion, what are the definitions for mild, moderate and high risk patients ?

  • Avatar of Sabahat H Zaidi
    Sabahat H Zaidi on

    But I see a huge sample size with little reports of subtrochantaric fractures. In this senario, the confounders are definitely strong enough to mislead the authors. Well, not going to change my practice because of this but will keep my eyes open.

  • Avatar of Jenna Piccininni
    Jenna Piccininni on

    There were no specific definitions of each risk stratified group within the paper. However, examples of each type of patient were included in the paper and are provided below. I have also written the author for more clarification and will get back to you when I receive a response.

    Low risk “Example: 53-yr-old
    woman, menopause at age 50 yr, lowest T-score
    1.6, no risk factors, bisphosphonate therapy for 2
    yr. Treatment was not indicated in the first place and
    can be discontinued.”

    Mild Risk “Example: 65-yr old woman, menopause at age 52
    yr, initial lowest T-score2.6,norisk factors, bisphosphonate
    treatment for 5 yr,BMDstable over that time.
    Treatment was indicated, but after 5 yr of treatment, a
    drug holiday might be considered.”

    Moderate Risk “Example: 70-yr-old
    woman, menopause at age 49 yr, lowest initial Tscore
    2.7, no risk factors, bisphosphonate therapy
    for 8 yr, BMD increased over that time so lowest
    T-score now is 2.3. Treatment was indicated, but
    after 8 yr of treatment, a drug holiday might be

    High Risk “Example: 72-yr-old woman, menopause at age 43 yr, lowest
    initial T-score3.8, rheumatoid arthritis requiring
    ongoing corticosteroid therapy for 12 yr, 3-in.
    height loss and two vertebral fractures byVFA,treatment
    with bisphosphonate therapy for 10 yr Treatment
    was indicated. After 10 yr, she remains at high
    risk of fracture. If a holiday from the bisphosphonate
    is considered, interval treatment with teriparatide or
    raloxifene would be prudent.”

  • Avatar of Jenna Piccininni

Comments are closed.