Effect of oral alendronate on bone mineral density and the incidence of fractures in postmenopausal osteoporosis.

Authors: Liberman U, Weiss S, Bröll J, Minne H, Quan H, Bell N et al.
Source: New England Journal of Medicine. 333:1437-43. November 30, 1995.
Institutions: Multi-institutional in the US and other countries.
Financial support: Merck Research Laboratories.



The bisphosphonates etidronate and the newer compound alendronate are inhibitors of both bone mineralization and resorption. Because of its negative effect on mineralization, etidronate is administered cyclically in the treatment/prevention of osteoporosis. Alendronate, although similarly inhibiting mineralization at an equivalent dose, is a much more potent inhibitor of bone resorption, and can thus be administered continuously, at lower doses that do not significantly inhibit bone mineralization. This paper reports on the pooled results of two identical, randomized, double-blind, placebo-controlled, multi-institutional trials, initiated in 1990. The purpose of these trials was to assess the efficacy of several dosages of alendronate in the treatment of postmenopausal osteoporosis. One trial was conducted in the US, one in a number of other countries.


Subjects: Women 45-80 years old, at least 5 years post-menopause, with documented osteoporosis (lumbar bone mineral density < 2.5 SD below normal). Exclusion criteria included other causes of osteoporosis (steroids), prior use of biphosphonates, treatment within the previous 12 months with estrogen/progesterone, calcitonin, fluoride.

Treatment: Women were randomized to treatment for three years with placebo (40%), or alendronate at one of three different doses (20% each): 5 mg, 10 mg and a third group at 20 mg for 2 years followed by 5 mg for one year. All women were given 500 mg of calcium daily.

Measurements: Bone mineral density of the lumbar spine, femoral neck, trochanter, forearm and total body were measured with dual-energy x-ray absorptiometry (intervals and frequency of measurements not specified in the article). Lateral spine films were obtained at baseline and at 1, 2 and 3 years and were digitized and analyzed for: fractures at baseline, new fractures and a calculated Spine Deformity Index. Height was measured at baseline and eight times during the study. Symptomatic non-vertebral fractures were recorded.


Data from the two studies (US and non-US) were pooled, as had been originally planned. 994 women were randomized (87.4% white, 0.4% black, 12.2% other races); 909 patients completed at least one year of the study.

At 36 months, there were significant increases in bone mineral density compared with baseline, at all alendronate doses and at all sites (except the mid-forearm, where there was a significant increase only at the 10 mg dose). There were significant losses at all sites in the placebo group. The 10 mg dose was significantly more effective than the 5 mg dose and was as effective as the 20 mg / 5 mg dose. At 10 mg daily, there was an increase in bone mineral density of 8.8% in the spine, and 2.5% in the total body,

During the study, 6.2% of the women in the placebo group had at least one new vertebral fracture, compared with 3.2% in the alendronate groups (P=0.03). The total number of vertebral fractures per 100 women was 4.2 vs. 11.3, treatment vs. placebo.

The Spine Deformity Index increased in 33% of women on alendronate, vs. 41% of women on placebo (P=0.028). The mean loss of height after three years was 3.0 mm in the treatment group vs. 4.6 mm in the placebo group; this difference was mainly due to vertebral fractures.

There was a non-significant trend towards fewer non-vertebral fractures in the alendronate group (3.0 women with fractures per 100 patient-years at risk, vs. 3.7).

About 16% of patients discontinued the assigned treatment, about 6% for clinical adverse effects in both the placebo and the treatment groups (mainly GI side-effects).


This study clearly demonstrates the efficacy of oral alendronate, particularly at the 10 mg dose, in preventing / reversing the bone loss related to post-menopausal osteoporosis. The effect was demonstrated on bone mineral density, but also on the clinical event of vertebral fractures and there was a suggestion of an effect on non-vertebral fractures. From a methodologic standpoint, the study seems well-designed, although there is no explanation for the subdivision, from the outset, into two, separate studies which were then pooled. Also, although the frequency of spine films is described and the numbers of patients with x-ray studies is given, no such detail is given for the bone mineral density measurements.

The actual role of alendronate in the treatment of postmenopausal osteoporosis remains unclear, however, since other agents (calcium, fluoride, vitamin D analogues, estrogen replacement therapy and calcitonin) are available, and the relative effectiveness of alendronate compared with them, singly and in combination, is not yet known. As Sambrook points out in an accompanying editorial, randomized clinical trials will be necessary to clarify this point. Also, the trend towards fewer non-vertebral fractures will need to be confirmed, since these comprise a larger group than the vertebral fractures. Finally, although alendronate appears to be more effective than etidronate, it is a newer drug and long-term safety remains an issue.


Reader comments

Date: Sun, 10 Dec 1995
From: 102032.3014@compuserve.com

 I agree with your comments about the article except the comments about the "trend" toward a decrease in the incidence of non-vertebral fractures. Every study will have a trend in one direction or the other from the baseline or compared against placebo. The reason we use a statistic (usually the P value) to determine is keep from jumping to conclusions about "trends". So why comment on trends? They are by definition, not statistically significant. Paying attention to trends in data have gotten us into a lot of trouble over the years. Let's determine what statistically significant is going to be before the study is done and then stick with it when we draw conclusions about the study. Ignore the trends. Those can be dealt with by the folks who do meta-analyses later on.

Winston Warr, M.D., M.P.H.

March 14, 1996
Letters to the Editor about this article from the NEJM

Journal Club on the Web home page

Submit a comment about this article

Site Meter