Safety and Efficacy of Risedronate in Patients With Age-Related Reduced Renal Function as Estimated by the Cockcroft and Gault Method: A Pooled Analysis of Nine Clinical Trials

Authors


  • Dr Miller receives funding and serves as a consultant for Procter and Gamble. Dr Roux serves as a consultant for Procter and Gamble. Drs Barton, Dunlap, and Burgio are employees of Procter and Gamble. All other authors have no conflict of interest.

Abstract

The incidences of osteoporosis and renal insufficiency increase with age. We studied the influence of renal function on the safety and efficacy of risedronate 5 mg daily in osteoporotic women. Risedronate was safe and effective in osteoporotic women with mild, moderate, or severe age-related renal impairment.

Introduction: The incidences of both osteoporosis and renal insufficiency increase with age; thus, the effect of renal impairment on the safety and efficacy of osteoporosis treatments is a clinical concern. Risedronate is a pyridinyl bisphosphonate well established as safe and effective in the treatment and prevention of osteoporosis. Currently, there is little available information about the effect of bisphosphonate treatment in patients with renal insufficiency. This retrospective analysis was conducted to study the influence of renal function on the safety and efficacy of risedronate in a population of osteoporotic women.

Materials and Methods: Combined data from nine randomized, double-blind, placebo-controlled phase III risedronate trials were analyzed. The patients in these studies had no markedly abnormal laboratory parameters that were considered clinically significant and no evidence of significant disease. This analysis included patients who received placebo (n = 4500) or risedronate 5 mg (n = 4496) for up to 3 years (average duration of exposure, 2 years) and who had renal impairment (creatinine clearance [CrCl] < 80 ml/min). CrCl was estimated by the Cockcroft and Gault method, based on age, weight, and serum creatinine. Patients were categorized as having mild (CrCl ≥50 to <80 ml/min), moderate (CrCl ≥30 to <50 ml/min), or severe (CrCl < 30 ml/min) renal impairment.

Results: Of the patients studied, renal impairment at baseline was mild in 48% (mean [range] serum creatinine, 0.9 [0.4–1.6] mg/dl), moderate in 45% (1.1 [0.6–1.9] mg/dl), and severe in 7% (1.3 [0.7–2.7] mg/dl). In both the placebo and risedronate treatment groups, the patients with the most severe renal impairment were older and had more severe osteoporosis. The incidences of overall adverse events and of renal function-related adverse events were similar in the placebo and risedronate 5 mg groups regardless of renal function. Furthermore, evaluation of changes from baseline in serum creatinine revealed no difference in renal function between the placebo and risedronate 5 mg groups in any of the renal impairment subgroups at any time-point. In all three subgroups, risedronate effectively preserved BMD and reduced the incidence of vertebral fractures.

Conclusions: These findings show that risedronate is safe and effective in osteoporotic women with age-related mild, moderate, or severe renal impairment.

INTRODUCTION

BISPHOSPHONATES ARE WELL established as safe and effective for the treatment and prevention of postmenopausal and glucocorticoid-induced osteoporosis.(1–3) Bisphosphonates bind strongly to hydroxyapatite and are therefore taken up and sequestered by bone.(4) The fraction of systemically available bisphosphonate not taken up by the skeleton is excreted unaltered through glomerular filtration(5) and an active secretion process in the kidney.(6) Because bisphosphonates are eliminated almost exclusively by the kidney, they have the potential to accumulate in patients with compromised renal function. Therefore, the effect of renal insufficiency on the safety and efficacy of bisphosphonate treatment is an important consideration in clinical practice.

Because both renal function and bone mass tend to decline with age,(7–11) it is likely that many older individuals will have comorbid renal insufficiency and osteoporosis. In fact, in a study of the relationship between age, renal function, and BMD, Klawansky et al.(12) found that 85% of women and 58% of men with osteoporosis have some degree of renal compromise. Because many bisphosphonate clinical trials excluded patients with abnormal renal function based on serum creatinine,(13–15) there is little published information on the effect of renal compromise on the safety and efficacy of bisphosphonate treatment.

Risedronate is a third-generation pyridinyl bisphosphonate effective in the treatment and prevention of postmenopausal osteoporosis and glucocorticoid-induced osteoporosis at doses of 5 mg daily or 35 mg weekly.(15–22) Like other bisphosphonates, risedronate is eliminated primarily by the kidney. The risedronate phase III clinical trial database contains data from nearly 10,000 women up to 100 years of age treated with placebo or risedronate 5 mg daily for up to 3 years. We performed a retrospective analysis of these data to study the influence of baseline renal function on the safety and efficacy of risedronate in a population of osteoporotic women. For this analysis, we estimated renal function using the Cockcroft and Gault formula,(23) which uses serum creatinine, age, and weight to determine creatinine clearance. Although measurement of serum creatinine is most commonly used in the clinical setting, this method may lead to overestimation of renal function, especially in elderly patients with low muscle mass.(8,23) In contrast, creatinine clearance values determined by the Cockcroft and Gault formula have been shown to be highly correlated with glomerular filtration rate (GFR) as measured by creatinine clearance.(24)

MATERIALS AND METHODS

Data included in the analyses

The analysis was based on combined data from 9 of the 11 randomized, double-blind, placebo-controlled, parallel-group, phase III trials of risedronate administered daily. One study of risedronate in the treatment of patients with Paget's disease was not included because its focus was not osteoporosis. A second study of risedronate plus estrogen in the treatment of patients with osteoporosis was not included because patients received estrogen in addition to risedronate. The nine studies included in the analysis were conducted at multiple centers in Europe, North America, Australia, and Asia from November 1993 to April 1998. Risedronate was evaluated in the treatment of postmenopausal osteoporosis in six studies,(15–17,21,22) in the prevention of postmenopausal osteoporosis in one study,(20) and in the prevention and treatment of glucocorticoid-induced osteoporosis in one study each.(18) In all nine studies, patients received risedronate 5 mg or placebo daily for up to 3 years. The average duration of exposure to treatment across the nine studies was 2 years. In all nine studies, all patients received up to 1 g of elemental calcium daily, and in seven of the nine studies, patients with low serum 25-hydroxyvitamin D levels at baseline (<40 nM; 16 ng/ml) also received up to 500 IU of vitamin D. Patients were excluded from these studies if they had evidence of clinically significant systemic disease, such as history of hyperparathyroidism, hyperthyroidism, or osteomalacia, within 1 year before enrollment or markedly abnormal laboratory values, including serum creatinine levels >1.1 times the upper limit of normal or alkaline phosphatase levels >1.5 times the upper limit of normal. All studies were conducted according to the Declaration of Helsinki and approved by the appropriate ethics committee. All patients gave written informed consent. The sponsor designed this study and analyzed the data. The study design, patients studied, and methods for each study have been described in detail elsewhere.(15–18,20–22)

Measurements

Adverse event information was collected at study visits made every 3 months during the trials. At each visit, patients were asked if they had experienced any medically related changes in their well being, and a physical examination was performed as required. The investigator recorded adverse events reported by the patients, as well as adverse events he or she observed. “Specific renal function-related adverse events” included hematuria, hydronephrosis, kidney failure, acute kidney failure, abnormal kidney function, uremia, oliguria, polyuria, glomerulitis, and nephritis. “Urinary- and renal function-related adverse events” included these specific renal function-related adverse events plus other events from the COSTART urogenital body system that were related to urinary function or kidney disease.(25) Events were included whether or not they were considered related to study treatment.

Clinical laboratory evaluations, including measurement of serum creatinine, calcium, and phosphorus, were performed at baseline, at months 6, 12, and 24, and at the last postbaseline visit (endpoint) as part of the safety observations. PTH levels were measured at baseline in a subset of patients in two studies.(16,18) Creatinine clearance (CrCl) at baseline was estimated for each patient on the basis of serum creatinine (SCr), age (years), and body weight (kg) using the Cockcroft and Gault method(23) as follows:

equation image

BMD was measured by DXA at baseline, at months 6, 12, and 24, and at the endpoint. Lateral thoracolumbar (T4-L4) radiographs were obtained at baseline and annually in four studies and at baseline and at study completion in the other five. Incident vertebral fractures were assessed quantitatively and semiquantitatively with adjudication (89% of patients), or quantitatively, with or without visual verification (11% of patients). An incident new vertebral fracture was defined quantitatively as a decrease of 15% or more in vertebral height relative to baseline in a vertebra that was normal at baseline(26) and semiquantitatively as a change from grade 0 (normal) at baseline to a grade of 1 or greater.(27) Visual verification consisted of visual confirmation of fractures diagnosed by quantitative morphometry. The radiologists remained blinded to treatment while performing all vertebral fracture assessments.

Transiliac bone biopsies and histomorphometric evaluations were performed as part of three of the nine trials in this analysis.(16,18) Bone biopsy samples were taken at baseline and after 3 years of treatment at a subset of study centers from women who consented to the procedure. Details of the biopsy procedures, histology, and quantitative assessments are described elsewhere.(28)

Biochemical markers of bone turnover were assessed at baseline, at 3, 6, and 12 months, and at endpoint in patients enrolled at a subset of study centers. Urine samples were collected for 2 h after the overnight void in the morning, and serum samples were collected in the morning after an overnight fast. Stored samples were analyzed in batches at central laboratories over the course of the studies. Bone-specific alkaline phosphatase was determined with the Tandem R-Ostase immunoradiometric assay (Hybritech, San Diego, CA, USA), and urinary deoxypyridinoline/creatine ratios were measured by high-pressure liquid chromatography.

Statistical analysis of safety and efficacy outcomes

Assessment of safety was the primary objective of these analyses; assessment of efficacy was a secondary objective. Data from the nine phase III osteoporosis studies were pooled and evaluated. The protocols for these nine studies were similar with respect to the dosages administered, the inclusion and exclusion criteria, and the assessments and their timing. Homogeneity was assumed for the safety responses, as the treatment-by-trial interactions were not statistically significant (p ≥ 0.328).

The primary analysis population was drawn from the 9883 osteoporotic women who were enrolled and randomly assigned to receive placebo or risedronate in one of the nine phase III clinical trials. Of these, 9662 (98%) had postmenopausal osteoporosis, and 221 (2%) had glucocorticoid-induced osteoporosis. To be included in this analysis, patients were required to have received at least one dose of study treatment and to have some degree of renal impairment (i.e., estimated CrCl < 80 ml/min). “Sensitivity” analyses were also performed based on the population of 5023 patients who completed the trials (per protocol analysis). Both analyses were based on data from patients who received the 5 mg dose (the U.S. marketed daily dose).

Patients were categorized as having mild, moderate, or severe renal impairment on the basis of CrCl at baseline, according to Food and Drug Administration (FDA) criteria (severe, CrCl < 30 ml/min; moderate, CrCl >30 to <50 ml/min; mild, CrCl >50 to <80 ml/min).(29) Descriptive statistics were generated for baseline demographic and disease characteristics.

Safety was evaluated on the basis of the incidence of overall, urinary-, and renal function-related, specific renal function-related adverse events, and change from baseline in serum creatinine, calcium, and phosphorus levels. The percentages of patients reporting at least one adverse event were summarized across treatment groups and renal impairment subgroups for each category of adverse events. These percentages were compared between treatment groups by calculating the relative risk and 95% CI using the Cochran-Mantel-Haenszel methodology, stratifying for trial. The mean percent changes from baseline in serum creatinine, calcium, phosphorus, bone alkaline phosphatase, and deoxypyridinoline were compared between treatment groups using a two-way ANOVA model, with trial and treatment group as factors. The mean difference (risedronate versus placebo) and 95% CI were summarized for each renal impairment subgroup. Because of the limited amount of histomorphometric data from paired biopsy samples, descriptive statistics were used to summarize these data across treatment groups.

Efficacy was evaluated on the basis of change from baseline in BMD and the incidence of new vertebral fractures. The percent change from baseline in BMD at the lumbar spine, femoral neck, and femoral trochanter was compared between treatment groups using the same methodology as that used to compare the serum variables. Homogeneity was assumed for the efficacy responses, because the treatment-by-trial interactions were not statistically significant for vertebral fractures (p = 0.928) or for BMD at the lumbar spine (p = 0.255), femoral neck (p = 0.361), or femoral trochanter (p = 0.334). Time-to-first fracture statistical methodology was used to estimate the cumulative vertebral fracture incidence for each treatment group within each renal impairment subgroup using Kaplan-Meier estimates. Cox regression was used to estimate the treatment effect (risedronate versus placebo) and corresponding 95% CI within each renal impairment subgroup. The analysis model was stratified for trial to account for separate underlying hazards functions. For each treatment group, a Cox regression model that included a factor for renal severity subgroup (i.e., two degrees of freedom) was used to determine whether fracture risk differed significantly across the three renal severity subgroups.

RESULTS

Patient characteristics

Of the 9883 patients evaluated in the primary intent-to-treat analysis, 8996 (91%) had some degree of renal impairment (Table 1). Renal impairment was classified as severe in 572 (7%), moderate in 4071 (45%), and mild in 4353 (48%). Within each renal impairment subgroup, the two treatment groups (placebo and risedronate 5 mg) were very similar with respect to baseline demographic and disease characteristics (Table 1). Similar percentages of patients in the two treatment groups were categorized as having severe, moderate, or mild renal impairment. In both treatment groups, the patients with the most severe renal impairment were older, more frail, and had more severe osteoporosis. At baseline, the overall median (interquartile range) creatinine clearance was 49.5 (40.0–59.6) ml/min (range, 14.1–79.9 ml/min) for the placebo group and 49.2 (39.7–59.5) ml/min (range, 13.2–79.9 ml/min) for the risedronate 5 mg group. Mean PTH levels at baseline were similar in the placebo (n = 89) and risedronate 5 mg (n = 105) treatment groups and were within the normal range (data not shown). For all patients with severe renal impairment for whom baseline PTH data were available (n = 4), PTH levels were within the normal range.

Table Table 1.. Baseline Characteristics for Patients With Renal Impairment Treated With Placebo or Risedronate 5 mg Daily
original image

The percentages of patients who discontinued increased with the severity of renal impairment; however, within each renal impairment subgroup, the percentages of patients who withdrew were similar in the risedronate and placebo groups. The most commonly cited reason for discontinuation was voluntary withdrawal (19% in the placebo group and 18% in the risedronate group), followed by adverse events (17% in each treatment group).

A total of 5023 patients completed the trial and were included in the per protocol “sensitivity” analysis. The overall mean (SD) duration of treatment in these patients was 34 (6.1) months. The results of the per protocol analyses were consistent with those of primary intent-to-treat analysis (data not shown).

Adverse events

The incidence of overall, urinary- and renal function-related, and specific renal function-related adverse events was similar within and between treatment groups in the subgroups of patients with severe, moderate, or mild renal impairment (Table 2). Statistically and clinically, there were no significant differences. The most common urinary- and renal function-related adverse event was urinary tract infection, which was noted in 10–13% in each treatment group in each renal impairment subgroup.

Table Table 2.. Incidence of Adverse Events in Patients With Renal Impairment Who Were Treated With Placebo or Risedronate 5 mg Daily
original image

Serum creatinine, calcium, and phosphorus

The differences between the placebo and risedronate 5 mg groups in the changes from baseline in serum creatinine at 6, 12, or 24 months or at the endpoint (Fig. 1A) were not significant in any of the renal impairment subgroups.

Figure FIG. 1..

(A) Mean (95% CI) difference between the placebo and risedronate 5 mg treatment groups in the percent change from baseline in serum creatinine at 6, 12, and 24 months and at the endpoint (last postbaseline visit). Mean differences are shown for the severe, moderate, and mild renal impairment subgroups. Negative values imply that decreases from baseline in serum creatinine were greater in the risedronate 5 mg group than in the placebo group. (B) Mean (95% CI) difference between the placebo and risedronate 5 mg treatment groups in the percent change from baseline in serum calcium at 6, 12, and 24 months and at the endpoint (last postbaseline visit). Mean differences are shown for the severe, moderate, and mild renal impairment subgroups. Negative values imply that decreases from baseline in serum calcium were greater in the risedronate 5 mg group than in the placebo group. (C) Mean (95% CI) difference between the placebo and risedronate 5 mg treatment groups in the percent change from baseline in serum phosphorus at 6, 12, and 24 months and at the endpoint (last postbaseline visit). Mean differences are shown for the severe, moderate, and mild renal impairment subgroups. Negative values imply that decreases from baseline in serum phosphorus were greater in the risedronate 5 mg group than in the placebo group.

Although there were sporadic statistically significant differences between the treatment groups with respect to the mean percent changes from baseline in serum calcium and serum phosphorus, these differences were small and an expected consequence of changes in calcium homeostasis during therapy and were not considered to be clinically meaningful (Figs. 1B and 1C). These differences were small, early, transient, and asymptomatic and required no intervention (Figs. 1B and 1C).

Efficacy: BMD and new vertebral fractures

The mean percent increase from baseline to endpoint in BMD at the lumbar spine and femoral neck and trochanter was significantly greater in the risedronate 5 mg group than in the placebo group in all three renal impairment subgroups, except at the femoral neck in the severe renal impairment subgroup. Mean percent changes (SE) from baseline to endpoint in lumbar spine BMD were −1.37% (1.72%) in the placebo group versus +4.23% (1.82%) in the risedronate group in patients with severe renal impairment; −0.47% (0.50%) in the placebo group versus +4.33% (0.51%) in the risedronate group in patients with moderate renal impairment; and −0.14% (0.19%) in the placebo group versus +3.96% (0.18%) in the risedronate group in patients with mild renal impairment (p < 0.001 for all comparisons). The increases from baseline in BMD were greater in the risedronate 5 mg group than in the placebo group at all three skeletal sites at 6, 12, and 24 months in each renal impairment subgroup. In addition, the incidence of new vertebral fractures in the risedronate 5 mg group was significantly lower than that in the placebo group within each renal impairment subgroup (Fig. 2). The incidence of new vertebral fractures in risedronate-treated patients was similar across renal impairment subgroups (p = 0.124), whereas the incidence in placebo-treated patients increased significantly with the severity of renal impairment (p < 0.001).

Figure FIG. 2..

Percentages of patients in the placebo (cross-hatched bars) and risedronate 5 mg (white bars) treatment groups with new vertebral fractures within each renal impairment subgroup. (N is the number of patients with paired radiographs that could be evaluated).

Histomorphometry

Histomorphometric data from paired biopsy samples (baseline and end of treatment) were available for 57 subjects with mild (placebo, n = 22; risedronate, n = 21) or moderate (placebo, n = 6; risedronate, n = 8) renal impairment. The mean (SD) ages of these 57 subjects were 66 (9.1) years in the placebo group and 65 (10.8) years in the risedronate group. The mean (SD) serum creatinine levels were 1.03 (0.142) mg/dl in the placebo group and 1.02 (0.127) mg/dl in the risedronate group, and the median (range) calculated creatinine clearances were 60.2 (35.1–78.7) ml/min in the placebo group and 54.1 (41.5–78.9) ml/min in the risedronate group. In these renally impaired patients, risedronate treatment led to decreases from baseline of 68% in mineralizing surface and 54% in activation frequency (Table 3).

Table Table 3.. Histomorphometry Values at Baseline and at Endpoint for Patients With Mild or Moderate Renal Impairment (Paired Data)
original image

Bone turnover markers

Bone turnover marker data were available for a subgroup of the patients in the analysis (20–30% of the intent-to-treat population, depending on the trial). As expected, the decreases from baseline in serum bone alkaline phosphatase levels and in urinary deoxypyridinoline/creatinine ratios were greater than those in the placebo group in all three renal impairment subgroups (data not shown). Importantly, the difference between the placebo and risedronate treatment groups in the severe renal impairment group was no greater than those in the mild and moderate renal impairment subgroups for either bone turnover marker.

DISCUSSION

This analysis showed that treatment with risedronate 5 mg daily for up to 3 years was safe and effective in this population of renally compromised patients, even in those with an estimated baseline CrCl of 15–30 ml/min. Evaluation of changes from baseline in serum creatinine over 24 months of treatment revealed no difference in renal function between the placebo and risedronate groups in any of the renal impairment subgroups. Small differences in the changes from baseline in serum calcium and phosphorus in the placebo and risedronate groups occurred as a result of changes in calcium homeostasis caused by inhibition of bone resorption, the desired therapeutic effect of risedronate. These changes were small, transient, and asymptomatic and required no intervention, and thus were considered clinically insignificant. The incidences of overall adverse events, urinary- and renal function-related adverse events, and specific renal function-related events were similar in patients in the placebo and risedronate groups regardless of renal function. The findings from this pooled analysis are consistent with those from the individual studies included in the analysis (Table 4), as reflected in the absence of any significant treatment-by-trial interaction (p ≥ 0.328). Generally, in the individual studies, the risk of adverse events in the risedronate group relative to that in the placebo group was close to 1. In a few studies, the numbers of patients were small, and between-group differences of just one or two patients led to disproportionately large differences in the risk of adverse events. Analysis of bone turnover markers, performed in a subgroup of the patients studied, showed no evidence of oversuppression of bone turnover in any renal function subgroup. Also, suppression was no greater in the severe renal impairment subgroup than in the mild or moderate impairment subgroups.

Table Table 4.. Comparison of Safety Endpoints Between Groups of Renally Impaired Patients Treated With Risedronate (RIS) or Placebo (PLAC) Across the 9 Individual Phase III Clinical Trials Used in the Analysis
original image

Our safety findings are consistent with those of a small retrospective study by Lewiecki and Rudolph.(30) In that study, the authors found that of 124 women 65 years of age or older who had been treated with a bisphosphonate for osteoporosis or osteopenia, 16 (13%) had previously unrecognized severely impaired renal function (CrCl < 35 ml/min). Despite this, no serious adverse events were attributed to bisphosphonate therapy in this group of patients. Our findings are also in line with long-term safety data from women treated with risedronate 5 mg for up to 7 years.(31,32) These data revealed no evidence of an effect of long-term risedronate treatment on renal function.

In our analysis, risedronate effectively increased BMD and reduced the incidence of vertebral fractures in all three subgroups. Interestingly, although the incidence of vertebral fractures in risedronate-treated patients was similar across renal impairment subgroups, the incidence in placebo-treated patients increased significantly with the degree of renal impairment. This finding is consistent with the fact that the more severely renal impaired patients were older and had more severe osteoporosis.

An important finding in this analysis was that 91% of the osteoporotic women from the risedronate phase III clinical trial database had some degree of renal impairment, as determined by creatinine clearance estimated with the Cockcroft and Gault method. These women had been presumed to have normal renal function on entry to the primary studies on the basis of serum creatinine levels. Our findings are consistent with those of Klawansky et al.,(12) who showed that 85% of women with osteoporosis have some degree of renal impairment and suggested that unrecognized renal insufficiency may be common in clinical practice.

In our study and in others,(33) the prevalence of renal impairment was shown to increase with age. As patients age, they tend to lose muscle mass. Because creatinine generation is proportional to muscle mass, muscle wasting is associated with a decreased creatinine pool, leading to decreased creatinine generation and excretion. As a result, patients with low muscle mass could have “normal” serum creatinine levels and substantially reduced GFR. Serum creatinine determinations, therefore, may not always reveal renal dysfunction in elderly patients, leaving this condition to go unrecognized.

In our analysis, renal function was assessed on the basis of the Cockcroft and Gault formula,(23) which uses serum creatinine, age, and weight to predict creatinine clearance and is highly correlated with GFR as measured by creatinine clearance.(24) Nevertheless, the Cockcroft and Gault method has been shown to both overestimate(34) and underestimate(35) GFR. In a recent study, Bostom et al.(36) compared the predictive performance of renal function equations in patients with chronic kidney disease (CKD) and normal serum creatinine. They found that the most accurate results were obtained with the Cockcroft and Gault equation. Thus, although not perfect, the Cockcroft and Gault equation is one of the best predictors of GFR among the currently available prediction equations.

Because renal clearance values determined from the Cockcroft and Gault equation are affected by changes in serum creatinine and body weight, changes in these variables over the course of the study are relevant. As shown in Fig. 1A, serum creatinine remained relatively constant from baseline through 24 months and the endpoint. Body weight data were collected at baseline and postbaseline in only four of the trials included in the analysis.(16–18) Analysis of these data (not shown) revealed no significant differences between the placebo and risedronate groups with respect to the changes from baseline in mean body weights; this was true in all of the renal impairment subgroups.

An important aspect of the use of bisphosphonates in renally compromised patients is the effect of renal insufficiency on the skeletal retention of drug and on bone histomorphometry.(37) Studies specifically measuring the skeletal retention of risedronate have not been conducted. However, in a pharmacokinetics study in patients with varying degrees of renal impairment, a decrease in creatinine clearance from 120 to 20 ml/min was associated with a reduction in clearance of risedronate of 64% after single dose oral administration.(38) Despite this reduction in clearance, there was no clinically or statistically significant increase in risedronate exposure. The authors concluded, therefore, that no adjustment in risedronate dose was necessary for patients with renal impairment with creatinine clearance >20 ml/min (excluding patients with end stage renal disease [ESRD]).

There is concern that reduced renal function would result in risedronate producing greater suppression of bone turnover and possibly impairing mineralization. Available histomorphometry data suggest that this is not the case. The decreases from baseline in mineralizing surface (68%) and activation frequency (54%) observed in the renally impaired subjects treated with risedronate in this study (Table 3) are similar to those observed among all patients from the vertebral (VERT)-North America study who had paired histomorphometric data. In those patients, risedronate treatment led to decreases of 58% in mineralizing surface and 47% in activation frequency after 3 years.(28) These findings are consistent with the antifracture efficacy of risedronate and suggest that risedronate treatment is not associated with a deleterious effect on bone strength. In our evaluation of renally impaired patients, double tetracycline-labeled surface was found in all of the post-treatment biopsies. Evaluation of post-treatment mineral apposition rates, osteoid surface, and osteoid thickness revealed no effect of risedronate treatment on mineralization (Table 3).

The safety of bisphosphonate treatment in patients with renal osteodystrophy has not been established. It is important to note that the results of our analyses apply only to patients with stage 2, 3, or 4 age-related renal disease. As outlined in the recently published guidelines from the National Kidney Foundation,(39) there is a spectrum of bone disease across the stages of progressive CKD (stage 1: normal GFR to stage 5: GFR <15 ml/min to dialysis). There is concern that bisphosphonates could impair bone strength in patients with a specific form of renal osteodystrophy—adynamic renal bone disease—characterized by very low bone turnover and increased bone fragility.(40) Because bisphosphonates reduce bone turnover, administering bisphosphonates to patients with preexisting very low bone turnover is a concern, with an outcome that is as yet unknown. Whereas most adynamic renal bone disease occurs in patients with stage 5 CKD (with ESRD or on dialysis), adynamic renal bone disease may also be seen in patients with stage 4 CKD (GFR 15–30 ml/min).(41) In these latter cases, adynamic renal bone disease is more likely to occur because of oversuppression of PTH caused by excessive treatment of secondary hyperparathyroidism with vitamin D analogs that have indirect as well as direct effects on PTH production.(42–44) On the basis of their medical histories and laboratory findings, we do not believe that any of the patients in our studies had adynamic bone disease. Nevertheless, we acknowledge that even age-related reductions in renal function are a form of chronic renal failure that may lead to increased bisphosphonate retention. If a clinician has any concern that a patient with osteoporosis who also has stage 4 CKD may have adynamic renal bone disease, a double tetracycline-labeled bone biopsy should be done to definitively exclude this possibility.(45–47)

The patients included in this analysis received risedronate 5 mg daily. Risedronate may also be taken once weekly at a dose of 35 mg. The 5 mg daily and 35 mg weekly dosages have been shown to provide similar risedronate exposure over a 1-week dosing interval,(48) with linear risedronate pharmacokinetics up to 50 mg/week in postmenopausal women. In addition, most recently, the renal and hepatic safety of risedronate 30 mg/day was shown over a wide range of steady-state concentrations.(49) Additional study is needed to further characterize the safety profile of the 35 mg weekly dosage in patients with renal impairment.

The strength of these findings may be limited by the fact that the study was a posthoc analysis of pooled data from nine trials. However, the prevalence of renal compromise in the population of patients studied was consistent with that in the larger population of osteoporotic patients studied by Klawansky et al(12); therefore, the population of patients in our analysis is most likely representative of the patients who receive treatment for osteoporosis in clinical practice. The findings from this analysis would be expected to be applicable to the general population of patients who would be treated with risedronate. Although age-related declines in GFR are a form of chronic renal failure, our results may not apply to patients who have renal failure known to be related to a systemic or renal-specific intrinsic cause, patients with ESRD, patients on dialysis, or patients pre- and postrenal transplantation. Such patients require a different approach to differentiating among the various forms of renal osteodystrophy, including double-tetracycline labeled quantitative bone histomorphometry.(33,46) An additional limitation of our pooled analysis of efficacy data are the differences between studies in the methods used for fracture assessment. In four of the nine studies (89% of patients), fractures were assessed by both quantitative morphometry and by a semiquantitative assessment, and discrepancies between the two methods were adjudicated by an expert radiologist. In the remaining five studies (11% of patients), fractures were assessed on the basis of quantitative morphology with or without visual verification. Despite these methodological differences, all nine studies showed similar reductions in the incidence of vertebral fractures in risedronate-treated patients. Also, there were no significant treatment-by-trial interactions. It should be noted that the patients included in this analysis were treated for up to 3 years, with an average duration of treatment of 2 years. Because the various antiresorptive agents used to treat osteoporosis have different characteristics, our findings cannot be assumed to apply to other orally or intravenously administered bisphosphonates.

Patients with mild, moderate, or severe renal impairment who were treated with risedronate 5 mg daily experienced no significant increase in the incidence of overall, urinary- and renal function-related, or specific renal function-related adverse events compared with placebo-treated controls. Risedronate significantly reduced the incidence of new vertebral fractures within each renal impairment subgroup, and the incidence of new vertebral fractures in risedronate-treated patients was similar across renal impairment subgroups. Of particular significance are our findings that treatment with risedronate 5 mg daily was safe and effective even in women with severe renal impairment (CrCl < 30 ml/min). These data are important because a large percentage of patients with osteoporosis have been shown to have some degree of renal impairment, and there is currently little published information about the effect of risedronate treatment in patients with renal insufficiency. The results of our analyses apply only to patients whose renal function is compromised as a result of aging, although there may not be differences in the skeletal retention of bisphosphonates associated with different etiologies of decreased GFR. Additional studies are needed to establish the safety and efficacy of risedronate treatment in patients ESRD or stage 5 CKD (GFR < 15 ml/min).

Acknowledgements

SB is a Senior Clinical Investigator of the Fund for Scientific Research Flanders, Belgium (F.W.O. Vlaanderen). The authors are grateful to Roger Phipps for interpretation of the histomorphometry data. The authors acknowledge Mary G Royer for assistance in preparing the manuscript. This work was supported by Procter & Gamble Pharmaceuticals (Mason, OH, USA) and sanofi-aventis (Bridgewater, NJ, USA).

Ancillary