Daily Oral Pamidronate in Women and Men With Osteoporosis: A 3-Year Randomized Placebo-Controlled Clinical Trial With a 2-Year Open Extension

Abstract

The efficacy and safety of oral pamidronate was examined in a double-blind, placebo-controlled trial in women and men with established osteoporosis. Seventy-eight postmenopausal women and 23 men with at least one prevalent vertebral fracture were randomized separately to 150 mg/day of pamidronate or placebo for 3 years followed by 150 mg/day of pamidronate for an additional 2 years. In addition, all patients received 400 U/day of cholecalciferol and 500 mg/day of elemental calcium. Pamidronate increased significantly bone mineral density of the lumbar spine (LS-BMD) and of the femoral neck (FN-BMD). The total increase in BMD of the spine after 5 years of treatment was 14.3%. Lateral spine radiographs were obtained at baseline and after 3 years of treatment. Fractures of previously normal vertebrae occurred in 15 of 45 patients treated with placebo (33.3%) and in 5 of 46 patients treated with pamidronate (11%). The relative risk was 0.33 (95% CI, 0.14-0.77). Treatment was well tolerated and there was no difference in gastrointestinal toxicity between pamidronate and placebo-treated patients. One hundred fifty milligrams daily of pamidronate is an effective and safe treatment of women and men with established osteoporosis.

INTRODUCTION

BISPHOSPHONATES ARE the most extensively studied therapies of osteoporosis under controlled conditions. In addition, the nitrogen-containing bisphosphonates alendronate and risedronate are the only approved treatments shown in randomized clinical trials to reduce the risk of vertebral and nonvertebral fractures, including those of the hip, in calcium- and vitamin D-replete women with postmenopausal osteoporosis.^1-8

Pamidronate has been used for many years in Leiden as a model compound to study the action and efficacy of nitrogen-containing bisphosphonates^9-12 and was the first bisphosphonate of this class to be given to patients with osteoporosis.^{13, 14} Various modes of administration of oral or intravenous pamidronate were shown to prevent bone loss in patients with postmenopausal, idiopathic and glucocorticoid-induced osteoporosis in open or controlled studies.^13-20 Early open studies suggested, in addition, that daily oral pamidronate induces a significant increase in bone mineral density of the spine (LS-BMD) that is not confined to the first 2 years of therapy, as expected from an antiresorptive agent, but continues for at least 4 years of continuous treatment.^{14, 21} Recently, this was shown in a controlled study with alendronate, which progressively increased LS-BMD in women with postmenopausal osteoporosis for up to 7 years of treatment.²² However, there are no data from controlled studies about the long-term effects of pamidronate and on its efficacy in fracture prevention.

In this study, we examined the efficacy and safety of daily oral pamidronate in women and men with osteoporosis during 5 years of continuous treatment, the first 3 years of which were placebo-controlled.

MATERIALS AND METHODS

This was a 3-year randomized, double-blind, placebo-controlled, clinical trial conducted in The Netherlands. At the end of the 3-year trial, patients were offered the option to participate in a 2-year extension during which both groups received active treatment. Patients were recruited from those referred to the participating centers for investigation and treatment of osteoporosis. All patients provided informed written consent and the study was approved by the Ethical Committees of the participating centers.

Included were women younger than 75 years and at least 5 years postmenopausal and men aged between 40 and 75 years with at least one atraumatic radiologically documented vertebral fracture. Patients had to have a life expectancy of at least 5 years. Excluded were patients with abnormal liver function; serum creatinine >140 μM; history of any malignant disease; disorders of calcium and bone metabolism other than osteoporosis; endocrine disorders; treatment with antiepileptics or glucocorticoids at a dose of >7 mg/day of prednisolone or equivalent for 1 week or longer in the year preceding the trial; or prior treatment with bone-acting drugs such as sodium fluoride during the previous 9 months or any use for >3 months, bisphosphonates in the previous 3 years, anabolic steroids or estrogens in the previous 6 months, or calcitonin in the last 3 months before entry into the study. Peptic ulcer or gastrointestinal diseases, other than malabsorption syndromes, were not considered exclusion criteria.

Treatment

Women and men with osteoporosis were randomized separately per center to receive either 150 mg/day of pamidronate or placebo for 3 years, followed by 150 mg/day of pamidronate for an additional 2 years. Trial medication consisted initially of two enteric-coated pellets of 75 mg of pamidronate or placebo and thereafter of 150 mg of enteric-coated tablets or placebo. Enteric-coated pellets were provided by Ciba-Geigy (currently Novartis), Basel, Switzerland, and enteric-coated tablets were prepared by the hospital pharmacy of the Leiden University Medical Center. The patients were instructed to take the trial medication on an empty stomach at least 30 minutes before breakfast, lunch, or dinner with a full glass of water leaving the choice of time to the patients. Generally, patients adhered to the same intake routine during the study. In addition, all patients received cholecalciferol (Devaron; Solvay Pharma, Weesp, The Netherlands) tablets, 400 IU/day, and calcium supplements (Calcium Sandoz Forte; Novartis), 500 mg/day, starting at least 3 months before the beginning of the trial. Compliance with therapy was tested by tablet counting.

Study outcomes

The primary outcome of the study was the percent change in LS-BMD during the 3-year blinded period. Secondary efficacy measures were changes in BMD of the femoral neck (FN-BMD), incidence of vertebral fractures after 3 years, changes in stature, and changes in biochemical indices of bone turnover. Investigators were not blinded to results of BMD or biochemistry during the trial. However, assessment of BMD changes at the end of the blinded treatment was done without knowledge of treatment assignment. In addition, radiological evaluation was performed in a center that did not participate in the trial by personnel blinded to treatment assignment and other study outcomes. BMD was measured by dual-energy X-ray absorptiometry (DXA) at baseline and at yearly intervals by Hologic, Inc., (Waltham, MA, USA) QDR-1000 or Norland Medical Systems (Fort Atkinson, WI, USA) densitometers that were cross-calibrated during the study. Every patient was measured with the same instrument from the beginning to the end of the study. At each time point, a double measurement was performed after repositioning of the patient and the average value was used in the analysis. Scans were analyzed in one center. Fractured vertebrae within the region of interest were excluded from the analysis. Prevalent and incident fractures were assessed morphometrically by the McCloskey-Kanis method²³ on lateral spine radiographs taken at baseline and after 3 years of therapy. Briefly, prevalent vertebral fractures were characterized by comparison of vertebral height ratios with “gender” and “level”-specific reference data. The method has high specificity and takes account of common minor variations in vertebral shape. Incident vertebral fractures were defined using the same method (point prevalence, i.e., any incident fracture had to be identified as a prevalent fracture on the follow-up radiographs) together with a required minimum change from the baseline films. This minimum change was derived from an analysis of measurement errors undertaken in a random subset of paired radiographs. This analysis suggested that a new or incident vertebral deformity would require a reduction in anterior, central, or posterior vertebral height of at least 15% and of at least 4.6 mm in absolute height. The stipulation of a minimum change combined with the point prevalence method prevents the characterization of minor changes in vertebral height ratios as a new fracture. Only fractures of previously normal vertebrae were counted as incident fractures. The incidence was expressed as the proportion of patients with new vertebral fractures after 3 years. Symptomatic fractures occurring during the trial were confirmed radiologically. Height was measured by a Harpenden stadiometer. The ratio of urinary hydroxyproline to creatinine (OHP/Cr) was used as an index of bone resorption and total serum alkaline phosphatase was used as an index of bone formation.

Statistical analysis

Baseline data are expressed as means ± SEM. All data of outcomes, with the exception of changes in height, are expressed as mean percent change from baseline ± SE. Analyses of efficacy data were by intention-to-treat. Differences within groups in BMD, height, and biochemical parameters were tested by two-sided t-tests. Differences between groups were analyzed using a repeated-measurements analysis of variance (ANOVA); a value of p < 0.05 was considered significant. Difference in incident vertebral fractures between groups was assessed by the Mantel-Haenszel test.

Safety evaluations

Physical examination, standard hematology, and clinical chemistry were performed every 6 months and all adverse events were documented. In patients who reported severe gastrointestinal complaints, endoscopy was performed. Iliac crest bone biopsy specimens were obtained at baseline and either after 1 year or 2 years of treatment. These results were reported previously.²⁴

RESULTS

Patients

One hundred one patients (78 postmenopausal women and 23 men) with osteoporosis were randomized to receive placebo or pamidronate. Baseline characteristics of the two groups were similar and are shown in Table 1. Ten patients, 5 patients in each group, were lost to follow-up before the end of 3 years. Reasons were loss of interest (n = 4), Alzheimer's disease (n = 2), moved out of the area (n = 3), and sudden death while on vacation (n = 1). Ninety-one patients (90%) completed the 3-year study period and were included in the intention-to-treat analysis. At the end of 3 years, 83/91 patients were still on medication, 43 patients (33 women and 10 men) were on pamidronate, and 40 patients (31 women and 9 men) were on placebo. Eight patients stopped taking the trial medication but continued to be followed. Fifty-three patients continued after 3 years, 27 patients from the placebo group and 26 patients from the pamidronate group.

Table Table 1.. Baseline Characteristics of Patients Randomized to Placebo or Pamidronate (Mean ± SE)

Changes in BMD

Changes in BMD are shown in Fig. 1. LS-BMD increased by 9.02 ± 1.34% (p < 0.001) in the pamidronate group and by 3.70 ± 0.93% in the placebo group (p < 0.01) after 3 years of treatment. The difference between the two treatment groups was significant (p < 0.001). The increase on pamidronate was most prominent in the first year of treatment but continued during the second and the third year; the differences between the second and the first year and between the third and the second year were significant (p < 0.001 and p < 0.01, respectively). In the placebo group LS-BMD, after an initial increase the first year, remained unchanged during the following 2 years.

Figure FIG. 1.

Percent changes in (A) LS-BMD and (B) FN-BMD during 5 years of treatment with 150 mg daily of oral pamidronate or placebo for 3 years followed by 2 years of 150 mg/day of pamidronate.

FN-BMD increased significantly after 3 years in pamidronate-treated patients (1.60 ± 0.66%; p < 0.05) and in placebo-treated patients there was a small, nonsignificant decrease (−0.21 ± 0.68%). The difference between the two groups was not significant.

Fifty-three patients agreed to participate in the 2-year open extension of the study (27 patients from the placebo group and 26 patients from the pamidronate group). Baseline characteristics (gender, height, BMD, prevalent fractures, and indices of bone turnover) did not differ between patients who continued the study and patients who did not. Patients included in the extension were slightly younger at baseline (63.5 vs. 65.6 years). In those patients who had received placebo for 3 years, treatment with 150 mg/day of pamidronate for another 2 years induced a significant increase in LS-BMD by 6.21% (p < 0.001). Similarly, FN-BMD increased by 3.67% (p < 0.05). In the patients who continued taking pamidronate, there were additional small but significant increases in LS-BMD during the fourth and fifth year of treatment (2.27% and 2.78%, respectively; p < 0.001, 5 years compared with 3 years). FN-BMD also increased slightly (1.32%) but not significantly. Overall continuous administration of pamidronate for 5 years increased LS-BMD and FN-BMD by 14.34 ± 2.24% and 2.92 ± 1.32%, respectively. The corresponding increases in the patients treated with placebo followed by pamidronate were 9.81 ± 1.51% and 3.46 ± 1.04%, respectively.

Responses were similar in women and men. During the 3-year controlled study the percentage of increase in LS-BMD in the pamidronate-treated group was higher in women (10.13 ± 1.67% vs. 5.98 ± 1.49% in men) but this was largely because of the lower baseline BMD of the women; compared with placebo, absolute increases were 0.047 g/cm² in women and 0.040 g/cm² in men.

Incident fractures

Assessment of the incidence of vertebral fractures was done by vertebral morphometry only for the blinded period of the study on X-rays taken at baseline and after 3 years of treatment. Evaluable X-rays were available from 91 patients: 45 patients from the placebo group and 46 patients from the pamidronate group. New vertebral deformities occurred in 15 patients (33.3%) who received placebo and in 5 patients (11.0%) who received pamidronate. The relative risk for vertebral fractures was 0.33 (95% CI, 0.14-0.77), accounting for a significant reduction (67%) of the relative risk for new vertebral fractures in patients treated with pamidronate. There were 5 patients needed to treat (NNT) with pamidronate for 3 years to prevent one vertebral fracture.

Symptomatic vertebral fractures occurred in 6 patients treated with placebo and in 3 patients treated with pamidronate. Nonvertebral fractures occurred in 4 patients. Two patients had a hip fracture: 1 patient in the placebo group, 1 patient in the pamidronate group, and 2 patients had wrist fractures, one of whom fractured both wrists consecutively.

Both of these patients were treated with placebo. Two patients on placebo developed multiple symptomatic spinal fractures and subsequently were treated with pamidronate. Both were analyzed in their original treatment group.

Height

There was a progressive decrease in height in both groups, being greater in the placebo-treated patients (Fig. 2), although the difference did not reach statistical significance.

Figure FIG. 2.

Changes in height during the 3-year placebo-controlled study with 150 mg/day of oral pamidronate.

Biochemistry

The changes in urinary OHP/Cr are depicted in Fig. 3. In the pamidronate-treated patients, OHP/Cr, after an initial significant decrease, reached a plateau that was sustained during the whole period of observation. In the placebo-treated patients, there was a small, nonsignificant decrease in urinary OHP/Cr during the first 3 years. Subsequent pamidronate therapy for 2 years induced a significant decrease in urinary OHP/Cr of a magnitude similar to that achieved in patients treated continuously for 5 years. The changes in serum alkaline phosphatase followed the same pattern (mean decrease of 16%; p < 0.001).

Figure FIG. 3.

Percent changes in urinary OHP excretion, expressed as OHP/Cr ratio during 5 years of treatment with 150 mg/day of oral pamidronate or placebo for 3 years followed by 2 years of 150 mg/day of pamidronate.

Adverse effects

Six patients, 3 patients in each treatment group (6%), stopped treatment during the 3-year blinded period because of gastrointestinal complaints. Three of these patients underwent endoscopy because of severe symptoms. Esophagitis was diagnosed endoscopically in 2 patients on placebo, one of whom had a known hiatus hernia. One patient on pamidronate with epigastric complaints before the trial was found to have two duodenal ulcers and esophagitis. In 1 patient treated with pamidronate, a colon carcinoma was diagnosed after 3 years of therapy. No patient reported symptoms suggestive of an acute-phase reaction. No significant biochemical abnormalities were recorded during the trial except for a transient increase in serum levels of transaminases and/or γ-glutamyl transferase activity in 3 patients. One of these patients was on active treatment and 2 patients were on placebo, 1 of whom was found to have obstructive cholestatic disease on ultrasound.

DISCUSSION

This study shows that 150 mg/day of oral pamidronate is an effective and safe treatment for women and men with osteoporosis. Pamidronate increased LS-BMD progressively over 5 years and reduced the incidence of new vertebral fractures by 67% after 3 years. Although the number of patients studied was small and antifracture efficacy was not a primary outcome, this is the first study to report a significant effect of pamidronate on the incidence of vertebral fractures under controlled clinical conditions. Apart from sample size, several aspects of our study differentiate it from other clinical trials of nitrogen-containing bisphosphonates in osteoporosis.

First, we included both women and men in the same study. Our early open studies with pamidronate indicated that there was no gender difference in the response to the bisphosphonate.^{14, 21} This has been shown repeatedly in glucocorticoid-induced osteoporosis^{13, 25-28} and recently was demonstrated also in a specific study of male osteoporosis by Orwoll et al.²⁹ These investigators showed that alendronate was as effective in male osteoporosis as shown previously in postmenopausal osteoporosis, which led to the registration of alendronate in the United States for the treatment of male osteoporosis. The approved treatment schedule with alendronate was the same as that for postmenopausal osteoporosis. We used separate stratification for women and men but no separate analysis was planned because of the small numbers. As expected, the overall response was similar in men and women.

Second, we used a flexible dosing schedule giving the patients the choice to take the drug at times other than the morning that may have been inconvenient to some. Generally, it is advocated that for optimal effect bisphosphonates should be given in the morning before breakfast because of their poor intestinal absorption and their interaction with food, particularly that containing calcium. We show here that this flexible administration of pamidronate was associated with increases in spine BMD at least comparable with those obtained with more rigid administration of other nitrogen-containing bisphosphonates. Our findings suggest that for long-term treatment, it may not be necessary to adhere to a strict dosing timing protocol if the bisphosphonate is taken on an empty stomach at least 30 minutes before a meal. In another study with oral pamidronate, in which the timing of drug intake was electronically registered, we found no difference in BMD responses between patients who took pamidronate between 6 and 10 a.m. and those who took it between 3 and 7 p.m. (G.L. Leusink, E. de Klerk, J.A. Knottnerus, D. van der Heyde, H.A.P. Pols, and S.E. Papapoulos, unpublished data, 2000). A preliminary report of a study with oral alendronate showed that administration of this bisphosphonate 1 hour before lunch or dinner induced a similar reduction in bone turnover, biochemically assessed, as when given before breakfast.³⁰ However, caution is needed before extrapolating these results to all bisphosphonates. Currently, used nitrogen-containing bisphosphonates bind calcium with equal affinity that is independent of their secondary structure, which determines their antiresorptive potency.³¹ When a relatively high dose of a bisphosphonate is given, in the case of pamidronate (150 mg), the presence of a small amount of calcium in the gastrointestinal tract may have a trivial effect on the amount of the bisphosphonate that fails to be absorbed. However, the same small amounts of calcium may have more pronounced effects when the dose of the bisphosphonate is low, as is the case with the more potent compounds. For example, the currently used doses of alendronate, risedronate and ibandronate in osteoporosis are 10, 5, and 2.5 mg, respectively. Thus, it may be that in osteoporosis the potential disadvantage of using a bisphosphonate with a lower antiresorptive potency may turn out to represent an advantage in daily practice when oral pharmacokinetics are considered. However, this is speculative because no specific studies have addressed this issue, but it represents an intriguing possibility.

Third, patients with a high risk of fractures were included in this study, namely, at least one prevalent fracture with a mean number of three fractures. The high risk was confirmed by the high incidence of vertebral fractures in the placebo-treated group (33.3%), which is the highest reported so far in placebo groups of all bisphosphonate trials with the exception, perhaps, of the first cyclical etidronate trial.^{1, 2, 6, 7, 32-35} Notwithstanding, this patient selection allowed the detection of a significant decrease in fracture risk even in a study of this size. Therefore, inclusion of a high-risk population in a clinical trial in osteoporosis can result in a substantial reduction in the necessary number of studied patients when an effective intervention is used. However, it is no longer ethical to treat such patients with placebo. The high basal fracture risk of the patients studied combined with the effectiveness of pamidronate was responsible for the documented low NNT to prevent a fracture.

Despite its size, our study met the criteria of a properly performed clinical trial.³⁶ For example, the dropout rate was small with 90% of the patients completing the 3-year blinded study, all analyses were by intention-to-treat and the primary and secondary endpoints were predefined in the protocol. Certain issues need further discussion. The increase in spine BMD, the primary endpoint of the study, followed the predicted and recently established (with alendronate) long-term pattern.^{22, 37} The initial increase in LS-BMD after 1 year was followed by a progressive gradual significant increase for 4 consecutive years resulting in an overall gain of ∼14% after 5 years. The mechanism of this response is not entirely clear and may be caused by improvement of the secondary mineralization of bone as suggested for alendronate in animal and human studies.^{38, 39} A trend for an increase in mean wall thickness observed in bone biopsy specimens of pamidronate-treated patients²⁴ may have contributed to the increase in bone mass. The placebo-treated patients showed a significant increase in spine BMD during the first year but no change thereafter. Such increases have been observed in most clinical trials in patients with osteoporosis, although the magnitude of the increase was greater in our study. Treatment with vitamin D and calcium may have contributed to that. The changes in FN-BMD, although significant, are difficult to interpret. For example, the magnitude of the increase during the first 3 years of treatment was smaller than that observed between years 3 and 5 in patients who originally received placebo than that reported with other nitrogen-containing bisphosphonates. We do not have a good explanation for that but potential confounding factors (e.g., gender, center, and type of densitometer) were excluded. In addition, the increases between years 3 and 5 in placebo-treated patients were very similar to those reported for other potent bisphosphonates excluding also the possibility that the dose used may have not been optimal.

Oral pamidronate had an excellent safety profile. The incidence of dropouts caused by gastrointestinal toxicity was 6% and was distributed equally between the placebo and active treatment groups. It should be stressed that gastrointestinal diseases, except for malabsorption syndromes, were no exclusion criteria. In addition, of those patients who underwent endoscopy, two cases of esophagitis were diagnosed in the placebo group and one in the pamidronate group. This favorable safety profile contrasts with our experience using pamidronate at a dose of 300 mg/day, which when compared with placebo was associated with a higher incidence of dropouts caused by gastrointestinal side effects.⁴⁰ With daily oral pamidronate, gastrointestinal toxicity thus is dose-dependent, as has been suggested already in studies of other bone disorders.^{41, 42} Bone safety examined with paired bone biopsy specimens after 1 and 2 years of therapy has been reported previously.²⁴ In summary, pamidronate did not adversely affect mineralization of newly formed bone and induced changes that were compatible with the primary action of the bisphosphonate on bone turnover.

Despite early encouraging results with the use of pamidronate in osteoporosis, its further development for this indication was pursued only outside Europe and the United States mainly because of few reports of gastrointestinal toxicity.⁴³ Our data suggest that this may not have been a wise decision. In conclusion, daily oral pamidronate increases progressively spine BMD and reduces the incidence of new vertebral fractures in women and men with established osteoporosis.

Acknowledgements

We thank P. Vermeij for the preparation of pamidronate and placebo tablets used subsequently in the trial. Initially, this study was supported by Ciba-Geigy (currently Novartis).