Alendronate Prevents Postmenopausal Bone Loss in Women without Osteoporosis: A Double-Blind, Randomized, Controlled Trial

Methods

Study Participants

Healthy women aged 40 to 59 years who had experienced menopause 6 to 36 months before enrollment were eligible to participate in this study. Women were excluded if their spine bone mineral density was more than 2 SDs above or below normal peak bone mineral density or if they had a history of nontraumatic spine or hip fracture. Women with disorders of bone and mineral metabolism were also excluded, as were those with recent (within 1 year of study entry) major upper gastrointestinal diseases (such as peptic ulcer, esophageal disease, and malabsorption). Other exclusion criteria were previous treatment with bisphosphonates or fluoride (>1 mg/d) or treatment within the 12 months before enrollment with estrogen, progestin, calcitonin, glucocorticoids, anticonvulsant agents, phosphate-binding antacids, or excessive vitamin A or vitamin D. Women who regularly used (>four times per week) any medication that had the potential to cause gastrointestinal irritation (such as aspirin), who smoked more than 20 cigarettes per day, or who drank three or more alcoholic beverages per day were also excluded.

Four hundred forty-seven women met the inclusion criteria and were enrolled at 15 centers throughout the world. The target sample size of 250 women completing the study was selected to detect a difference in bone density of 2.4% between an individual alendronate dosage and placebo at a P value of 0.05 or less, with 95% power using an SD of 3.3% that was obtained from 1-year bone density data in placebo recipients [17]. All centers conducted the study with appropriate approval from the institutional review boards, and all participants gave informed consent.

Treatment

Participants were randomly assigned (in allocation blocks of five) to one of five regimens: placebo for 3 years; alendronate at 1, 5, or 10 mg/d for 3 years; or alendronate at 20 mg/d for 2 years followed by placebo for 1 year (20/0 mg/d). In all groups, double-blinding was maintained for all 3 years. Alendronate and placebo tablets were identical in size, shape, and color. The women were instructed to take the study drug daily at least 1 hour before breakfast or, as a less desirable alternative, at least 2 hours after a meal and 1 hour before the next meal. All participants also received a daily supplement of calcium carbonate (Os-Cal 500, Smith-Kline Beecham Consumer Brands, LP, Pittsburgh, Pennsylvania, or the equivalent) unless dietary calcium intake exceeded 1000 mg/d. This supplement was usually taken with the evening meal. After beginning therapy, each participant was seen at months 1 and 3 and every 3 months thereafter. Eight participants (four in the placebo group and four in the 1-mg/d group) lost more than 6% of their spine bone density at 24 months relative to their baseline measurements and were therefore designated “fast bone losers.” Seven of these participants completed the study; from month 24 to the end of the study, they received 5 mg of open-label alendronate per day. The other patient discontinued therapy at month 24.

Bone Mineral Density

Bone mineral density of the spine, proximal femur, total body, and forearm was measured every 6 months by dual-energy x-ray absorptiometry with a Hologic QDR-1000, 1000/W, or 2000 densitometer (Waltham, Massachusetts) or a Lunar DPX-L densitometer (Madison, Wisconsin). One bone density quality assurance center (Oregon Osteoporosis Center) that remained blinded to treatment allocation was responsible for the quality control of all participant and phantom calibration scans [18]. Factors to correct for machine calibration drift were applied as necessary. The primary end point was bone mineral density of the lumbar spine; the other most important end points were bone mineral density of the femoral neck, trochanter, and total body. Bone mineral density of the total body, total hip, and forearm was measured at the centers that had densitometers capable of performing these measurements (11 centers measured bone mineral density of the total body; 10 centers measured bone mineral density of the total hip and forearm).

Biochemical Markers and Indices of Mineral Metabolism

Fasting serum and urine samples (second morning void) were obtained at all clinic visits except those in months 27 and 33. The resorption markers included urine deoxypyridinoline, measured by high-pressure liquid chromatography, and (in a subgroup of 268 participants) urine N-telopeptide cross-links of type I collagen (Ostex, Seattle, Washington). Each of these markers was expressed as a ratio to urine creatinine. Serum osteocalcin levels, measured by radioimmunoassay (INCSTAR, Stillwater, Nebraska), and serum bone-specific alkaline phosphatase levels were used to assess the rate of bone formation. Levels of serum calcium, phosphorus, intact parathyroid hormone (measured by immunoradiometric assay), and 1,25 dihydroxyvitamin D (measured by competitive binding assay) were also determined to assess the effects of treatment on mineral metabolism. N-telopeptide levels were measured by Medical Research Laboratories (Highland Heights, Kentucky). All other assays were done at Corning-Nichols Institute (San Juan Capistrano, California).

Safety Evaluations

At each visit, vital signs were measured and any new or worsening symptoms were recorded. Physical examinations were performed at the baseline, month 3, and yearly visits. Standard laboratory safety evaluations (including evaluations of hematologic, renal, and liver function) were performed at every visit. Investigators reported any unfavorable or unintended clinical or laboratory events as adverse experiences. After 3 years of therapy, biopsy specimens of transiliac bone were obtained for histologic and histomorphometric assessment from 55 women who provided specific informed consent for this procedure. These analyses were performed, as previously described, by a single investigator who was blinded to treatment groups [19].

Statistical Analysis

The Tukey trend test [20] was used to assess the trend in response with increasing alendronate dosages. The 20/0-mg/d group was excluded from this test because of the change in dosage. This test uses the minimum P value, adjusted for multiplicity [21], that was obtained from tests of the regression sloped on three dosage scalings. The test is done in a stepwise manner, with the highest dosage eliminated, until the minimum P value exceeds 0.05. Dose-response relations were examined for the responses of the percentage change in bone density variables, natural logarithm (fraction of the baseline value) for biochemical variables, and change in clinical safety variables.

Role of Funding Source

Employees of Merck & Co., Inc., participated in the study as co-investigators. After designing the study with the input of other investigators, these employees implemented the protocol and coordinated data collection and statistical analysis. They also contributed to the writing of this paper, but data interpretation and decisions about the content of the paper and submission for publication resided with the entire group of investigators.

Results

Baseline Clinical and Laboratory Characteristics

At baseline, all treatment groups had similar demographic characteristics and bone density values (Table 1) and biochemical markers (data not shown). Of the 447 participants, 89.3%, 71.8%, and 69.6% completed at least 1, 2, and 3 years of treatment, respectively. The rate of withdrawal was similar in all treatment groups.

Bone Mineral Density

By 36 months, placebo recipients had lost approximately 3% to 4% of bone density at the spine, femoral neck, and trochanter (Figure 1 and Table 2). In contrast, participants in the 5-, 10-, and 20/0-mg/d groups had increased their bone density by approximately 1% to 4% at these sites relative to baseline. All alendronate dosages differed significantly from placebo, including the 1-mg/d dosage, which attenuated bone loss. Total body bone mineral density (measured at 11 of the 15 centers) increased significantly only with 10 mg of alendronate per day (increase of 1%) and was maintained with 5 mg/d; in contrast, progressive and significant losses of bone mineral density occurred in the placebo and 1-mg/d groups. Treatment with 10 mg of alendronate per day produced slightly greater mean increases in bone mineral density than did 5 mg/d, but the differences were significant only at the trochanter site. Most of the increase in bone mineral density of the spine and proximal femur occurred in the first 12 months of the study and was maintained thereafter with continuous alendronate treatment (Figure 1). The bone mineral density changes in the 20/0-mg/d group at 24 months were similar to those in the 5- and 10-mg/d groups. After 24 months, when women in the 20-mg/d group were switched to placebo, loss of bone density resumed in this group at a rate similar to that seen throughout the entire study in the placebo group.

Bone density increments were also achieved in the other regions of the proximal femur (Table 2). Forearm measurements were done in slightly more than half of the participants. Placebo led to loss of bone mass in the forearm sites, whereas alendronate significantly attenuated loss of bone mineral density at both subregions of the forearm (Table 2).

Biochemical Markers and Indices of Mineral Metabolism

Both urinary markers of bone resorption decrease (deoxypyridinoline levels decreased by 35% to 45%; N-telopeptide levels decreased by 65% to 70%) within 1 month of the initiation of alendronate at 5, 10, and 20 mg/d and reached a new steady state by 3 months (Figure 2). This decrease was sustained for the duration of treatment with 5 or 10 mg of alendronate per day. Bone-specific alkaline phosphatase and osteocalcin levels began to decrease at 1 month, reaching a steady state during treatment between 6 and 12 months at values of 40% to 60% below baseline (Figure 2). Upward trends in both deoxypyridinoline (between months 24 and 30) and bone-specific alkaline phosphatase (after 6 months) were seen in all groups, but the relation between active treatment groups and the placebo group was maintained. These changes are most consistent with a drift in the results of the assays during the study. After patients receiving 20 mg of alendronate per day were switched to placebo during the third year, the suppression of bone turnover was attenuated.

Women treated with at least 5 mg of alendronate per day had modest decreases in serum calcium levels (1% to 3% from baseline) and serum phosphate levels (5% to 9% from baseline) within the first few months of therapy, a finding consistent with the antiresorptive effect of the drug. These changes did not result in clinical signs or symptoms, and the values gradually returned toward baseline by 12 to 24 months. In response to these changes, serum parathyroid hormone levels increased by approximately 50% to 60% and 1,25-dihydroxyvitamin D levels increased by 20% in the 5-, 10-, and 20/0-mg/d alendronate groups by 1 month; by 12 months, these levels returned toward baseline values (data not shown).

Histomorphometry

Alendronate treatment did not impair skeletal mineralization; no increase in the mean thickness of osteoid seams (osteoid thickness) (Figure 3) or the proportion of unmineralized cancellous bone (osteoid volume/total volume) was documented. No differences were seen in the rate at which bone mineral was deposited (mineral apposition rate) for any treatment group. The local bone turnover rate was assessed by measuring the proportion of bone surface that was undergoing mineralization (mineralizing surface). This variable decreased, a finding consistent with the ability of alendronate to decrease the overall rate of bone turnover; however, mineralizing surface remained measurable in all participants (Figure 3). In addition, mineralizing surface was higher in the 20/0-mg/d group than in any other alendronate group, reflecting the resolution of the effect of alendronate to decrease bone turnover 1 year after discontinuation of therapy. No qualitative abnormalities of bone or bone marrow were seen on histologic examination, and the normal lamellar structure was preserved in all cases.

Adverse Experiences

Alendronate was well tolerated by participants. Clinical adverse events (any unexpected symptom ranging from mild, common symptoms, such as headache and upper respiratory infection, to infrequent but more significant problems requiring hospitalization) occurred in more than 90% of each treatment group. However, the incidence of all adverse experiences, including events defined by the protocol as serious (cancer, life-threatening events, or events resulting in hospitalization or prolongation of hospitalization) and adverse experiences leading to withdrawal from the study, was similar in the placebo and alendronate groups (Table 3).

Of the clinical adverse experiences seen in our sample, only flatulence and odynophagia showed dose-related increases (Table 3). Six cases of odynophagia occurred, three in the 10-mg/d group and three in the 20/0-mg/d group. None was serious, and only one (in a 20-mg/d recipient with esophagitis) resulted in discontinuation of study therapy. The remaining five participants had transient (lasting 2 to 11 days) episodes of odynophagia that resolved despite continued therapy. One of these participants (in the 20/0-mg/d group) was receiving placebo at the time of the adverse experience. No consistent differences in the frequency of other upper gastrointestinal symptoms (such as nausea, vomiting, and heartburn) or any other type of clinical or laboratory adverse experience were seen among the treatment groups.

Discussion

At baseline, bone mineral density in the participants was approximately 10% below the mean values for young adult women. This finding suggests that participants may have already experienced substantial bone loss during the perimenopausal and early postmenopausal period and is consistent with results of other recent studies that showed early postmenopausal bone loss [22, 23]. The average bone mineral density of the spine at baseline in our sample was 10% to 15% greater than levels considered to indicate osteoporosis [24]. If rates of bone loss were to continue at the rate of approximately 1% per year (the rate seen in the placebo group), a substantial number of these women would develop osteoporosis in the next 10 to 15 years.

In contrast to the therapeutic objective of increasing bone mass and strength in women with established osteoporosis, the aim of therapy in these women is simply the prevention of bone loss. Calcium administration did not achieve this goal; decreases of 3% to 4% were seen at several skeletal sites over the 3 years of observation, a finding consistent with those of earlier studies [25-27]. In our study, alendronate at dosages of 5 mg/d or greater given for 3 years met the objective of preventing bone loss at the spine, hip, and total body. The 10- and 20-mg/d dosages seemed to have slightly greater effects than the 5-mg/d dosage, but no clinically significant difference was apparent. Alendronate at 1 mg/d significantly attenuated but did not fully prevent bone loss, suggesting that 5 mg/d is the lowest of the tested dosages that can achieve the objective of preserving bone mass in this group of early menopausal women. We did not evaluate the effect of dosages between 1 and 5 mg/d. Compared with placebo, alendronate attenuated but did not prevent bone loss in the forearm. The clinical significance of this finding is unclear because alendronate therapy reduces the rate of wrist fractures in women with osteoporosis [10].

The skeletal response to alendronate at different sites was heterogeneous; greater increases in bone mass were seen at more trabecular sites, such as the spine. The effect of alendronate on bone mineral density of the spine was similar to or greater than the effects of estrogen or other bisphosphonates in early postmenopausal women [28-31]. The effect at the hip was of the same magnitude as that seen with hormone replacement therapy [28].

The decreases in biochemical markers of bone turnover in response to alendronate were similar to those seen with estrogen therapy [32, 33]. As was expected, bone resorption decreased more quickly than bone formation as a result of alendronate's direct ability to suppress osteoclastic activity. During this early transient period, formation exceeds resorption; this “remodeling space filling” probably explains the fact that the greatest increase in bone mass occurs during the first 6 to 12 months of therapy.

Biochemical markers achieved a new steady state after approximately 3 to 6 months of therapy and did not continue to decrease with continued therapy. The progressive return of the markers of bone resorption toward baseline values after the 20-mg/d group was switched to placebo in the third year suggests that it is the current (or at least recent) alendronate dosage, rather than the cumulative dose, that determines the degree of suppression of bone turnover. These findings are consistent with observations in animals that osteoclasts are exposed only to the drug present on resorption surfaces and that alendronate incorporated into bone is not pharmacologically active [34]. The resumption of bone loss after the discontinuation of alendronate therapy suggests that long-term therapy may be necessary for persistent inhibition of bone loss and reduction of the incidence of fractures in older women.

In addition to documenting increases in bone mass, our study shows the continued formation of bone of normal quality without histomorphometric evidence of osteomalacia. This is consistent with findings seen in more than 500 bone biopsies done in patients with osteoporosis and Paget disease of bone who were receiving daily alendronate dosages as high as 80 mg/d [35, 36]. The normal quality of alendronate-treated bone is further supported by preclinical studies that showed normal bone histologic findings [16] and by the decreased incidence of fractures associated with increased bone density in postmenopausal women with osteoporosis [9, 10].

Alendronate was well tolerated by participants in our study. The proportion of women who discontinued therapy because of an adverse experience was similar among the placebo and alendronate groups. Although the total incidence of upper gastrointestinal symptoms did not differ between the alendronate and placebo groups, the incidence of odynophagia was significantly increased in the 10- and 20-mg/d groups; one woman in the 20-mg/d group had esophagitis. No cases of esophagitis developed in the 5- and 10-mg/d groups, although this potential clearly exists with the 10-mg/d dosage, especially if the patient does not adhere to dosing instructions [37]. The safety profile of all doses was favorable; among the effective dosages, however, 5 mg/d clearly has the least potential for any adverse effects (including gastrointestinal). Therefore, on the basis of both efficacy and safety considerations, 5 mg/d seems to be the optimal dosage for the prevention of postmenopausal bone loss.

The limitations of our study include the relatively short duration of follow up. Understanding the long-term effect of strategies to prevent a process that occurs over many decades is important, and we plan to continue our study for at least 7 total years. Studies such as this one cannot assess the effect of therapy on risk for fracture: A much larger study conducted over many years would be required. Our data, however, combined with the documented effectiveness of alendronate therapy in reducing the incidence of fracture in older women with osteoporosis, suggest that preventing bone loss could ultimately protect against fractures. The participants in our study were healthy women who had recently undergone menopause. Thus, our results are not applicable to premenopausal women, men, or women with secondary causes of bone loss.

In summary, our study shows that alendronate treatment given for 3 years at dosages of at least 5 mg/d is effective in preventing bone loss at the spine, hip, and total body in early postmenopausal women. Bone quality is preserved, and treatment is associated with a favorable safety and tolerability profile. Alendronate shows promise as a nonhormonal, bone-specific therapy for the prevention of postmenopausal bone loss.

Appendix

The Alendronate Osteoporosis Prevention Study Group comprised the following:

Investigative Centers and Clinical Investigators: Richard Bauer, MD (University of Texas Health Science Center, San Antonio, Texas); Claus Christiansen, MD, Bjorg Clemmesen, MD, and Pernille Ravn, MD (Center for Clinical and Basic Research, Ballerup, Denmark); Jan Dequeker, MD, PhD (Universitair Ziekenhuis, Pellenberg, Belgium); John A. Eisman, MD, and Philip Sambrook, MD (Garvan Institute of Medical Research, St. Vincent's Hospital, Sydney, Australia); Ignac Fogelman, MD, and Paul Ryan, MD (Guy's Hospital, London, United Kingdom); Ghada El Hajj Fuleihan, MD (Brigham and Women's Hospital, Boston, Massachusetts); Piet Geusens, MD (Limburgs Universitair Centrum, Diepenbeek, Belgium); Nigel L. Gilchrist, FRACP (Princess Margaret Hospital, Christchurch, New Zealand); Marjorie Luckey, MD (Mount Sinai Medical Center, New York, New York); Michael R. McClung, MD (Oregon Osteoporosis Center, Providence Health System, Portland, Oregon); Charles J. Menkes, MD (Service de Rhumatologie A, Hopital Cochin, Paris, France); Jose A. Rodriguez-Portales, MD (Catholic University of Chile, Santiago, Chile); Stephan Goemaere, MD, and Jean Marc Kaufman, MD (UZ Gent Department of Endocrinology, Gent, Belgium); Richard Wasnich, MD (Hawaii Osteoporosis Center, Honolulu, Hawaii); Stuart R. Weiss, MD (San Diego Endocrine and Medical Clinic, San Diego, California). Bone Density Quality Assurance Center: Michael McClung, MD, and Leslic Roberts, BS (Oregon Osteoporosis Center, Providence Health System, Portland, Oregon). Histomorphometry Center: Robert S. Weinstein (University of Arkansas Medical School, Little Rock, Arkansas). Merck Research Laboratories, Rahway, New Jersey: Clinical Research: Anastasia Daifotis, MD, Celia Reda, BS, Enza Ricerca, BS, and A. John Yates, MD; Clinical Biostatistics: Erika Hale, MS, and Tracy Survill, MS.

From Providence Health System, Portland, Oregon; Center for Clinical and Basic Research, Ballerup, Denmark; Merck & Co., Rahway, New Jersey; Princess Margaret Hospital, Christchurch, New Zealand; St. Vincent's Hospital, Sydney, Australia; University of Arkansas for Medical Sciences, Little Rock, Arkansas; and Brigham and Women's Hospital, Boston, Massachusetts.

Drs. Clemmesen and Ravn: Center for Clinical and Basic Research, Ballerup Byvej 222, Ballerup, Denmark 2750.

Drs. Daifotis and Yates and Ms. Reda: Merck & Co., Inc., 126 East Lincoln Avenue (RY32-537), Rahway, NJ 07065.

Dr. Gilchrist: Princess Margaret Hospital, Cashmere Road, 8002 Christchurch, New Zealand.

Dr. Eisman: Garvan Institute of Medical Research, St. Vincent's Hospital, 384 Victoria Street, Sydney, NSW 2010, Australia.

Dr. Weinstein: University of Arkansas for Medical Sciences, 4301 West Markham-Slot 587, Little Rock, AK 72205.

Dr. El Hajj Fuleihan: Brigham and Women's Hospital, 221 Longwood Avenue, Boston, MA 02215.