Preventing atypical femur fractures in patients treated with bisphosphonates

A 62-year-old woman made an appointment to see me because she had read about unusual fractures caused by taking too much alendronate (Fosamax, Merck). She did not discuss this with her treating physician because she did not want to offend him by questioning his treatment (not a great philosophy, in my opinion, because most clinicians would not be offended, but would welcome the opportunity to discuss items such as this in great detail or would recommend a consultation).

Her personal and family history related to osteoporosis was unremarkable and I wondered why she had been on bisphosphonates for 8 years. Two years after her last menstrual period her gynecologist performed a bone density study and recorded a spine T-score of –2.1. It would not be unreasonable to recommend therapy with a T-score that low so early in the menopause because bone loss is most rapid in the first 5 to 6 years postmenopause.

However, the gynecologist neglected one critical factor in interpreting the bone density study. DXA is a two-dimensional tool and cannot take the patient’s “thickness” into account. Slender women, particularly someone as short as this woman (62”, 118 lb.), will often have a low T-score because the third dimension of the vertebral body cannot be determined by a 2-D DXA study.

She tolerated the therapy without concern and was very compliant with her weekly dose. A follow-up bone density had been performed every 2 years and the T-score improved to –1.9 by year 4 and had remained there since.

There was no history of fracture, no height loss and no clinical reason to suspect a secondary cause of bone loss. However, she looked “funny” to me, particularly her fingers which were short and her very marked lumbar lordosis. I decided to send her directly to the radiologist in our area who I believe has more understanding of what the skeleton should look like than any other radiologist with whom I have worked.

He called me to report that the fingers were fine aside from minor degenerative joint disease changes and there were no spine abnormalities he could detect. He was, however, very concerned about what he saw in the femurs — there were possible lesions in the outer margins in the middle of both. He recommended a bone scan that confirmed non-displaced fractures of both femurs as well as similar findings in one toe on each foot.

The patient was fortunate to have these lesions identified before they became clinically apparent. She is now on teriparatide (Forteo, Eli Lilly), calcium, vitamin D (1,000 IU) and “vibration” therapy (to be discussed at a later time when I have more fully understood the literature on this).

I got lucky with this patient by ordering the right imaging study for the wrong reason, and I cannot recall any literature documenting pre-fracture femoral lesions complicating long-term bisphosphonate therapy. I am sure it must happen, but we cannot be ordering radionuclide studies on all patients on bisphosphonates. Given the very low incidence of clinically apparent atypical femoral fractures on long-term bisphosphonate therapy, that would not be practical or cost effective.

What is a reasonable and practical alternative?

To my knowledge there isn’t anything established, but my practice is to monitor bone density every 2 years (yearly if the patient has already had an osteoporosis-related minimal trauma fracture and the insurance will cover the cost). I also monitor a marker of bone resoprtion — usually serum CTX — and bone formation — serum P1NP — at baseline, 3, 6 and 12 months to document that they have reached a nadir and as a fairly reliable indication that the patient is compliant with the therapy.

When, as in this case, two consecutive bone density studies done properly on the same DXA instrument report no change in BMD and there is stable suppression of the markers, I recommend that my patients have a drug-free holiday. There is little, if any, literature in support of this. The incidence of atypical femur fractures is so low that very few would develop such fractures even if not given a drug-free period.

But why take that chance?

There is literature reporting a link between adherence to therapy and future fragility fracture risk, but this literature is mainly restricted to 1 or 2 years of observation. There is a major difference between a physician-recommended drug-free holiday in compliant patients and fracture risk in non-compliant patients.