Calcified Tissue International

, Volume 81, Issue 5, pp 341–351

The Use of Bone Markers in a 6-Week Study to Assess the Efficacy of Oral Clodronate in Patients with Metastatic Bone Disease

Authors

    • Academic Unit of Clinical OncologyWeston Park Hospital University of Sheffield
    • Cancer Research UK Clinical Center in LeedsJIF Building, St. James’s University Hospital
  • E. V. McCloskey
    • Metabolic Bone UnitUniversity of Sheffield
  • J. A. Dewar
    • Department of OncologyNinewells Hospital
  • J. J. Body
    • Department of OncologyInstitut Jules Bordet
  • D. A. Cameron
    • Department of OncologyWestern General Hospital
  • A. N. Harnett
    • Norfolk and Norwich University Hospital
  • M. Ruutu
    • Department of OncologyMaria Hospital
  • O. P. Purohit
    • Academic Unit of Clinical OncologyWeston Park Hospital University of Sheffield
  • R. Tähtelä
    • Mehiläinen Oy Laboratoriopalvelut
  • R. E. Coleman
    • Academic Unit of Clinical OncologyWeston Park Hospital University of Sheffield
Article

DOI: 10.1007/s00223-007-9061-x

Cite this article as:
Brown, J.E., McCloskey, E.V., Dewar, J.A. et al. Calcif Tissue Int (2007) 81: 341. doi:10.1007/s00223-007-9061-x

Abstract

Biochemical markers of bone metabolism are strongly associated with skeletal complications in metastatic bone disease. The bisphosphonate clodronate reduces skeletal morbidity by inhibiting bone resorption. This study investigated the use of bone markers to assess the efficacy of oral clodronate across a range of clinically relevant doses. There were 125 patients with metastatic bone disease randomized to daily oral clodronate (800, 1,600, 2,400 and 3,200 mg) or placebo in a double-blind, multicenter study. Urinary N-terminal telopeptide of type I collagen (U-NTX), serum C-terminal telopeptide of type I collagen (S-CTX), urinary calcium (U-Ca), and bone alkaline phosphatase were measured weekly for a 6-week treatment period. Doses of ≥1,600 mg clodronate produced mean reductions of >40% in U-NTX, S-CTX and U-Ca, all significantly different from placebo (P = 0.0015, 0.001, 0.0036, respectively), after 6 weeks. Evaluation of least significant changes in markers suggested that the commonly used 1,600 mg dose was most appropriate for breast cancer patients. However, this dose was suboptimal for other (mainly prostate cancer) patients, who showed better response to 2,400 mg. The number of adverse events in the treatment arms was not significantly different from that in placebo, but a higher number of patients had diarrhea in the 3,200 mg arm and withdrew from the study. This trial is the first to explore the dose-response relationship of clodronate in oncology using specific markers of bone turnover. It has confirmed that the 1,600 mg dose is safe and effective for breast cancer patients but may be suboptimal for the other tumors studied.

Keywords

ClodronateMetastatic bone diseaseUrinary N-terminal cross-linking telopeptide of type I collagenSerum C-terminal cross-linking telopeptide of type I collagenBone alkaline phosphatase

The skeleton is the most frequent site for metastasis from a range of common cancers, and bone metastases continue to present a major challenge in cancer management. Increased bone resorption (osteolysis) is the “hallmark” of bone metastasis, and a consequent increase in bone formation often follows. In recent years, several biochemical markers have been developed which provide information on the current status of bone metabolism. In particular, urinary N-terminal cross-linking telopeptide of type I collagen (U-NTX) produced by breakdown of bone collagen has been shown to be a sensitive and specific marker of the extent of bone resorption in metastatic bone disease, being elevated in most patients with detectable metastases to bone and normal in patients with no metastases or metastatic disease at other sites [13].

Recent studies [4, 5] have shown that there are strong correlations between levels of bone markers and the occurrence of skeletal morbidity, confirming earlier small case series from a variety of centers [3, 6, 7]. This has led to the possibility of using bone marker changes as surrogate end points for skeletal morbidity in clinical trials involving bisphosphonate drugs and other bone-specific therapies under development.

Clodronate is a non-nitrogen-containing bisphosphonate which has been widely studied for use in the management of metastatic bone disease [811], and oral clodronate has been shown to decrease the risk of skeletally related events in breast cancer [9, 12]. Although clodronate is less potent than some of the newer bisphosphonates such as zoledronic acid, the oral formulation is convenient and remains widely used in the treatment of metastatic bone disease. Clodronate has also shown promise in the adjuvant treatment of early breast cancer [13, 14], and definitive trials to investigate this potential use are in progress.

The dosing of oral clodronate in metastatic bone disease was investigated almost 10 years ago in a double-blind, placebo-controlled study of 84 patients, who were randomized to receive placebo or 400, 1,600, or 3,200 mg of daily oral clodronate for 4 weeks [15]. Fasting urinary calcium (U-Ca) excretion was used as the primary end point for bone resorption. A significant difference was seen between placebo and 1,600 mg clodronate and between placebo and 3,200 mg clodronate, but no significant difference was observed between placebo and 400 mg clodronate or between 1,600 mg and 3,200 mg clodronate. On the basis of these results and accumulated clinical experience, the most commonly used oral clodronate dose is 1,600 mg. However, it is now generally believed that U-Ca is an inferior bone resorption marker to U-NTX and serum C-terminal telopeptide of collagen (S-CTX), particularly in patients receiving concomitant bisphosphonates [16, 17]. This is partly due to the fact that calcium excretion reflects the net effect of bone resorption and formation and to the dependence of U-Ca on dietary and other metabolic factors.

The overall purpose of the current study was to use bone-specific markers to reassess the commonly used 1,600 mg dose of oral clodronate in relation to other possible doses for the treatment of metastatic bone disease. The study took into account both efficacy using the more sensitive and specific biochemical bone markers and adverse event profile.

Patients and Methods

Study Design

Patients with tumor-induced osteolysis were randomized, with equal weighting, to one of four doses of clodronate (800, 1,600, 2,400, and 3,200 mg) or placebo in a double-blind, placebo-controlled, multicenter study. Full ethical approval was obtained in each center. The duration of treatment was 6 weeks. The randomization scheme was designed to distribute patients equally across treatment groups and stratified by center and within each center by cancer types (breast cancer vs. other types).

Patients included in the study were >18 years of age and had cytologically or histologically confirmed cancer with bone metastases diagnosed by plain X-ray imaging, computerized tomography (CT) scan, or magnetic resonance imaging (MRI). Patients were also required to have a Zubrod index of 0–2 [18] (equivalent to Eastern Cooperative Oncology Group (ECOG) performance status), to be psychologically and physically suitable for oral medication with clodronate or placebo for 6 weeks, and to provide written informed consent.

Exclusion criteria included known metabolic bone disease other than bone metastases or osteoporosis, treatment with bisphosphonates within 6 months prior to randomization, or other treatments with effects on bone, e.g., calcitonin, fluoride, and hormone replacement therapy, within 3 months prior to randomization. The exclusion criteria also included a change in systemic chemotherapy or hormonal antitumor therapy within 2 months prior to randomization or local radiotherapy within 2 weeks prior to randomization, hypercalcemia, impaired renal function, liver dysfunction, life expectancy of less than 3 months, pregnancy, or lactation. Local field radiotherapy was allowed during the study. Patients withdrawing from the study were not replaced.

Treatment and Compliance

Since the absorption of clodronate from the gastrointestinal tract decreases significantly in the presence of divalent cations such as calcium, it was recommended that oral clodronate not be taken with foods or liquids containing calcium, iron, or other divalent cations, e.g., antacids. Each patient took four tablets orally, once daily at least 1 hour before breakfast. The tablets were either clodronate (Bonefos®, Turku, Finland) 800 mg or placebo and were combined in such a way as to achieve the desired dosing regimen. Two different measures of compliance were used. In one, the patient was considered compliant if he or she had taken at least 75% of the required medication over the 6-week study, based on residual tablet counting, assessed weekly. A further indication of compliance was obtained through a single measurement of urinary clodronate at the final study visit (i.e., after 6 weeks) or at the final study visit for premature termination. This was carried out on a 2-hour fasting urine sample before the daily clodronate dose, i.e., approximately 24 hours after the previous clodronate dose. Urinary clodronate was measured in the Bioanalytical Laboratories of Leiras Oy (Turku, Finland) by a validated gas chromatographic method, with mass selective detection [19, 20]. The interassay precision as coefficient of variation (CV%) of quality-control samples varied between 4.8% and 8.4%, and interassay accuracy varied 94% and 102% from nominal concentration. The lower limit of urinary clodronate concentration to satisfy minimum compliance was taken as 240 μg/mmol creatinine. This was based on previous studies using an oral clodronate dose of at least 800 mg/day for 7 days [20].

Bone Marker Measurements and Assessments at Study Visits

All bone marker measurements were carried out at United Laboratories (Helsinki, Finland). Samples were stored locally at each treatment site for less than 2 months (at −20° to −70°C) before shipment to the central laboratory and storage in a −70°C freezer. All samples of a given patient were analyzed in the same run. The reference ranges used in this laboratory were determined in a local Finnish Caucasian population for males (older than 25 years) and for premenopausal and postmenopausal females. Urinary marker measurements were carried out on second voided urine samples, and serum samples for markers were fasting morning samples before 10 a.m. All samples were taken before administration of the study drug on that day.

U-NTX was measured with an automated chemiluminescent immunoassay on a Vitros ECI analyzer (Ortho Clinical Diagnostics, Raritan, NJ). The measuring range of the assay was 5–3,000 nmol bone collagen equivalents (BCEs)/L. Samples with U-NTX concentrations exceeding 3,000 nmol/L were reanalyzed from a prediluted sample. Intra- and interassay precision as CVs over a concentration range 35–1,250 nmol BCE/L were 5.1–9.8% and 6.9–16.7%, respectively. The results were expressed relative to creatinine concentration, and the reference ranges were for premenopausal females 15–80 nmol BCE/mmol creatinine, for postmenopausal females 20–120 nmol BCE/mmol creatinine, and for males 20–100 nmol BCE/mmol creatinine.

S-CTX was measured with an enzyme-linked immunosorbent assay (ELISA) in microtiter format using Serum CrossLaps reagent (Nordic Bioscience Diagnostics, Herlev, Denmark) with a measuring range of 1.00–19.0 nmol/L. Samples with S-CTX concentrations exceeding 19.0 nmol/L were reanalyzed from a prediluted sample. The intra- and interassay CVs, measured over a concentration range of 2.5–8 nmol/L, were 5.2–5.5% and 7.1–9.9%, respectively. The reference ranges of S-CTX were for premenopausal females 1.00–4.50 nmol/L, for postmenopausal females 1.20–5.50 nmol/L, and for males 1.30–5.00 nmol/L.

Serum bone alkaline phosphatase (ALP) was measured with an immunoradiometric assay using Tandem-R Ostase reagents (Beckman Coulter, Fullerton, CA). The measuring range of the assay was 1–120 μg/L. Samples with bone ALP concentrations exceeding 120 μg/L were reanalyzed from a prediluted sample. The intra- and interassay CVs over a concentration range of 10–90 μg/L were 4.1–8.1% and 9.1–25%, respectively. The reference ranges for bone ALP were for premenopausal females 3–15 μg/L, for postmenopausal females 3–22 μg/L, and for males 4–21 μg/L.

The analytical sensitivity of the bone marker assays was established from replicates (10–16) of the lowest standard/zero standard of the assay and was 5 nmol BCE/L, 1 nmol/L, and 1 μg/L for U-NTX, S-CTX, and bone ALP, respectively.

U-Ca was measured with a colorimetric method (o-cresolphthalein, Roche Diagnostics, Mannheim, Germany) on a Hitachi 912 analyzer with a measuring range of 0.07–25.0 mmol/L. The intra- and interassay CVs were from 2.8% and 3.5%, respectively. The measured values were proportioned to creatinine excretion, and the reference range used was 0.13–0.53 mmol/mmol creatinine.

Pain was assessed at each weekly study visit using a visual analogue scale. Patients were asked to mark on a 10-cm horizontal line the average level of pain during the previous week, without access to their previous scores. This was then scored from 0 to a maximum of 100 based on the level they had marked on the 10-cm line, each centimeter being counted as 10 units.

Blood tests included full blood count, routine electrolytes, liver function tests, bone profile, and random blood glucose and were repeated at each subsequent on-study visit. In addition, physical examination was carried out at baseline and at the 3-week and 6-week visits. Any adverse events (AEs) were recorded at each visit.

Statistical Methods

For estimation of the required patient numbers in each arm of the study, the following parameters were used: α risk (type I error) was set to 0.05; β risk (type II error) was set to 0.20, i.e., the power (1 − β) was set to 0.80; standard deviation of observations in U-NTX was estimated from the literature to be 25% [3, 21]. The clinically meaningful effect size was estimated to be 30%, the dropout rate was anticipated to be 20%, and the subgroup proportion (breast cancer patients) was predicted to be 70%. Thus, a minimum sample size of 20 patients per group was required.

An intent-to-treat analysis was to be performed, but in addition a subgroup analysis of the breast cancer cohort was planned. Since there were no clinically significant major protocol deviations in the study, no per-protocol analysis was performed. All analyses were performed as two-sided tests, with Tukey’s method being employed to adjust confidence intervals. Baseline comparisons between treatment groups and cancer types (breast cancer/other) were done for primary and secondary efficacy variables either by multiway analysis of variance or by the extended Mantel-Haenszel test. The level of significance was taken as 0.05. The primary efficacy variable, U-NTX, was analyzed using analysis of variance for repeated measurements (baseline, 1, 2, 3, 4, 5, and 6 weeks). In addition, 95% confidence intervals for the percentage change between baseline and other time points within each group were calculated. The same approach was used for analysis of the breast cancer data alone.

For analyses of the biochemical markers, the geometric means were used to display the bone marker test statistics, and P values as these are less sensitive to extreme values than the arithmetic mean. If the normal distribution assumption was not fulfilled in the analysis of variance for repeated measurements, log transformations to the base (e) were attempted. If the normal distribution assumptions were not met even after log transformations, the nonparametric Mantel-Haenszel method was used. Primary comparisons were without adjustment of the overall significance level. All other comparisons in parametric methods were adjusted using Tukey’s method. If pairwise comparisons were made nonparametrically, no adjustment was done. The secondary efficacy variables bone ALP, S-CTX, U-Ca, and bone pain measured by VAS, were analyzed in the same way as in the primary efficacy variable.

For each of the markers with the exception of bone ALP, the least significant change (LSC) was determined as 2.8 x the maximum of the intra-assay CV%. The numbers of patients meeting and failing to meet this value were determined for each marker at each dose.

Results

Patient Recruitment and Baseline Characteristics

Of 128 patients initially screened at six European centers, 125 were recruited. Breast cancer patients accounted for most of those recruited (62%), with prostate cancer patients making up most of the remainder (27%). The patient characteristics sex, age, body mass index (BMI), primary cancer at study entry, and performance status (Table 1) were well matched for the five treatment groups. Table 1 also shows that one-third of patients had sites of metastatic disease other than bone at study entry and that 10% of patients had received prior bisphosphonates. The median time since diagnosis of bone metastasis was approximately 11 months. On imaging, the majority of patients had multiple bone metastases, but each treatment group was well matched. There were approximately equal numbers of lytic and sclerotic lesions with some mixed lesions.
Table 1

Baseline demographics

Characteristic

Clodronate dose

Overall (n = 125)

Placebo (n = 24)

800 mg (n = 27)

1,600 mg (n = 25)

2,400 mg (n = 26)

3,200 mg (n = 23)

Male (%)

38 (30)

8 (33)

9 (33)

7 (28)

6 (23)

8 (35)

Female (%)

87 (70)

16 (67)

18 (67)

18 (72)

20 (76)

15 (65)

Median age, years (range)

64 (28–81)

67.5 (44–79)

64 (28–80)

57 (46–79)

65 (37–78)

64 (44–81)

Performance status

 

  Zubrod 0 (%)

33 (26)

2 (8)

10 (37)

5 (20)

9 (35)

7 (30)

  Zubrod 1 (%)

73 (58)

18 (75)

12 (44)

19 (76)

13 (50)

11 (48)

  Zubrod 2 (%)

19 (15)

4 (17)

5 (19)

1 (4.0)

4 (15)

5 (22)

Median BMI, kg/m2 (range)

26.4 (15.8–40.5)

27.1 (21.2–40.5)

26.2 (19.7–32.0)

25.8 (18.3–38.5)

27.4 (18.1–38.9)

27.3 (15.8–35.5)

Primary cancer

 

  Breast (%)

87 (70)

16 (67)

18 (67)

18 (72)

19 (73)

16 (70)

  Prostate (%)

33 (26)

7 (29)

7 (26)

6 (24)

6 (23)

7 (30)

  Other (%)

5 (4)

1 (4)

2 (7)

1 (4)

1 (4)

0 (0)

Median duration of bone metastases (months)

10.9

9.6

11

15.2

12.4

6.5

Nonbone metastases

 

  Liver (%)

11 (9)

3 (13)

3 (11)

1 (4.)

2 (8)

2 (9)

  Lung (%)

12 (10)

3 (13)

2 (7)

2 (8)

3 (12)

2 (9)

  Other (%)

19 (15)

5 (21)

2 (7)

2 (8)

4 (15)

6 (26)

Previous bisphosphonates*

13 (10)

3 (13)

1 (4)

4 (16)

2 (8)

3 (13)

Previous radiation therapy

103 (82)

20 (83)

21 (78)

21 (84)

22 (85)

19 (83)

Multiple bone metastases (%)

107 (86)

21

23

22

22

19

Type

 

  Lytic/mixed (%)

72 (58)

16

15

12

17

12

  Sclerotic (%)

49 (39)

8

12

11

8

10

  Missing (%)

4 (3)

0

0

2

1

1

Bone markers

 

  Elevated U-NTX

83 (68)

15 (65)

16 (59)

16 (67)

20 (77)

16 (73)

  Elevated S-CTX

75 (62)

14 (61)

17 (65)

13 (57)

21 (81)

10 (44)

  Elevated bone ALP

85 (70)

14 (61)

20 (74)

15 (65)

19 (73)

17 (74)

  Elevated U-Ca

22 (18)

4 (17)

4 (15)

7 (29)

2 (8)

5 (23)

Data are presented as number of patients, with percentage in parentheses.

*Patients had received either oral clodronate 1,600 mg or intravenous pamidronate 60–90 mg (as these were the only licensed bisphosphonates prior to this study for metastatic bone disease). All patients had received less than 1 year of treatment, and no patients had received bisphosphonates within 6 months of study entry.

For each marker, the numbers and percentages of patients above the relevant reference ranges at baseline are displayed in Table 1. For females, the upper limit of the premenopausal ranges were used, thereby excluding the effects of postmenopausal bone loss on bone marker levels. For U-NTX, 68% of the overall study population was above these reference ranges at baseline, ranging 59–73% when broken down into the placebo and treatment groups. A similar pattern was seen for S-CTX and bone ALP, but for U-Ca in the overall population, only 18% (7–29%) of patients were above the reference range at baseline. The results for breast cancer patients alone showed a similar distribution (data not shown).

Using log-transformed data, there were no statistically significant differences in mean values of U-NTX, S-CTX, or bone ALP at baseline between the dosing groups. Mean values for all three biochemical markers were significantly higher in the non-breast cancer group than in the breast cancer patients (P < 0.001 for all markers, data not shown). Within the breast cancer group, no significant differences were found between the dosing groups for U-NTX, the primary outcome variable, and bone ALP. There was, however, a significant difference in S-CTX (P = 0.011), with the 3,200 mg group having a lower mean value than the 2,400 mg group (3.5 vs. 7.6 nmol/L, respectively, P = 0.004), but no other observed differences.

Drug Compliance

Table 2 shows the compliance data for the study by treatment group according to tablet counting and by biochemical measurement of urinary clodronate. The two methods report on different aspects of compliance, with the tablet counting relating to the whole 6-week study and the biochemical measurement being a one-off check at the end of the study. Nevertheless, there was good agreement between the two methods, and compliance was seen to be high across each of the treatment arms.
Table 2

Measurement of treatment compliance at each dose

Measure of compliance

Clodronate dose

Placebo

800 mg

1,600 mg

2,400 mg

3,200 mg

Overall

>75% of tablets taken

21/21 (100)

20/22 (91)

22/22 (100)

23/24 (96)

19/20 (95)

105/109 (96)

U-Clodronate >240 μg/mmol creatinine

24/27 (89)

23/23 (100)

23/24 (96)

21/21 (100)

91/95 (96)

Data are presented as number of patients complying/total number of patients for whom compliance data were available, with percentage in parentheses.

Efficacy of Oral Clodronate: U-NTX

Table 3 shows the percentage change in geometric mean U-NTX values compared with baseline for placebo and each of the treatment groups over the 6-week period for both the overall patient population and the breast cancer patients separately. While there is considerable variability, it is clear that, for all the treatment arms, there was a rapid decrease in U-NTX value within 1 week, with further, smaller decreases out to 6 weeks. For the geometric mean U-NTX values, from baseline to 6 weeks, there was a significant change between placebo and the 1,600 mg dose (P = 0.0015) for the whole study population. There was a correspondingly significant change for the breast cancer group alone (P < 0.0001).
Table 3

Urinary N-telopeptide: percentage change in geometric means and 95% confidence intervals for each of the treatment groups during the study

Time point

Clodronate dose

Placebo (n = 23)

800 mg (n = 27)

1,600 mg (n = 24)

2,400 mg (n = 26)

3,200 mg (n = 22)

Baseline

0

0

0

0

0

Week 1

 

  All

−5.3 (−25, 20)

−23 (−38, −4.0)

−26 (−41, −7.2)

−52 (−61, −40)

−53 (−63, −41)

  Breast

−2.9 (27, 28)

−28 (−44, −7.9)

−33 (−48, −14)

−50 (−61, −36)

−47 (−60, −31)

Week 2

 

  All

−5.6 (−27, 23)

−27 (−42, −7.3)

−32 (−47,−13)

−55 (−65,−43)

−53 (−64,−39)

  Breast

−9.6 (−34, 23)

−32 (−49, −11)

−41 (−55, −21)

−56 (−66, −43)

−51 (−64, −33)

Week 3

 

  All

−7.0 (−32, 26.)

−22 (−40, 1.3)

−42 (−57, −21)

−58 (−68,−45)

−51 (−64,−34)

  Breast

−7.8 (−36, 32)

−29 (−48, −3.1)

−45 (−61, −23)

−59 (−70, −44)

−48 (−64, −27)

Week 4

 

  All

−5.9 (−31, 28)

−31 (−47, −9.5)

−35 (−52, −13)

−56 (−67, −41)

−47 (−61, −27)

  Breast

3.4 (−23, 62)

−38 (−55, −15)

−39 (−56, −13)

−56 (−68, −39)

−47 (−64, −24)

Week 5

 

  All

0.1 (−27, 38)

−30 (−47, −7.8)

−42 (−57, −2.7)

−52 (−64, −35)

−53 (−67, −34)

  Breast

12 (−23, 62)

−34 (−53, −9.2)

−48 (−64, −27)

−51 (−65, −32)

−50 (−67, −26)

Week 6

 

  All

11 (−21, 55)

−30 (−47, −5.4)

−32 (−51, −5.1)

−50 (−61, −28)

−59 (−72, 40)

  Breast

17 (−23, 75)

−35 (−55, −6.4)

−44 (−62, −16)

−47 (−63, −23)

−60 (−75, −36)

From Table 3, there is a trend toward a dose-response relationship between dose and inhibition of resorption. However, the relationship is relatively flat above 1,600 mg. Because of the spread of the data and the numbers of patients in each group, it was not possible to show a statistically significant difference between adjacent doses for 1,600 mg and above. In order to investigate possible differences between the doses further, analyses were performed on the numbers and proportions of patients with resorption marker changes failing to exceed the LSC for each marker, with the exception of bone ALP. These results are presented in Table 4, with only data for the final two time points (5 and 6 weeks) shown. For the breast cancer population, there was a clear increase in response above 800 mg but little or no difference between the 1,600 mg dose and higher doses. However, for those patients with other solid tumors (mainly prostate cancer), the proportion responding clearly increased from 800 mg to the higher doses; but there also appears to be an increase in response between 1,600 and 2,400 mg.
Table 4

Number of patients failing to respond in excess of the least significant change for each resorption marker at each dose at 5 and 6 weeks

Dose

U-NTX

S-CTX

U-Ca

5 weeks

6 weeks

5 weeks

6 weeks

5 weeks

6 weeks

Breast cancer patients

  Placebo

13/14

13/15

12/14

11/14

6/14

7/14

  800

8/20

9/21

6/19

5/19

5/19

5/19

  1,600

2/13

4/13

2/14

1/13

4/14

2/13

  2,400

2/14

2/14

1/15

1/14

4/15

3/15

  3,200

2/9

2/8

0/9

0/8

2/8

1/7

Patients with other tumors

  Placebo

3/5

4/5

3/7

4/5

2/5

3/5

  800

1/4

2/4

2/4

4/4

2/5

3/6

  1,600

3/5

5/6

1/6

2/5

2/6

1/6

  2,400

0/6

2/6

0/8

0/7

0/6

0/6

  3,200

0/5

0/5

0/5

0/5

0/5

0/5

Figure 1 displays the numbers of patients for whom U-NTX values were above the reference range for placebo and each of the treatment groups at baseline, 1 week, and the end of the study at 6 weeks. There was little change in the placebo group during the study. However, there was a reduction in the numbers of patients above the reference range in the treatment groups. Moreover, this reduction in numbers was already apparent after 1 week, with relatively little change between 1 week and 6 weeks.
https://static-content.springer.com/image/art%3A10.1007%2Fs00223-007-9061-x/MediaObjects/223_2007_9061_Fig1_HTML.gif
Fig. 1

Percentage of patients with elevated U-NTX before and during treatment

Efficacy of Oral Clodronate: Other Markers

The percentage changes in geometric mean of S-CTX, U-Ca, and bone ALP over the 6 weeks of the study are shown in Figure 2. For S-CTX, as with U-NTX, clodronate resulted in a rapid and substantial reduction in the marker value at 1 week, which then did not change appreciably over the remaining weeks of the study. For the 1,600 mg dose compared with placebo, there was a significant difference between mean S-CTX values for both the overall study population (P = 0.0002) and the breast cancer patients alone (P < 0.0001). For U-Ca, there was a rapid fall in marker level with the higher clodronate doses, but with lower doses the response was slower than for U-NTX and S-CTX. Again, the mean value for the 1,600 mg dose was significantly different from the placebo value for both the overall population (P = 0.0246) and the breast cancer patients (P = 0.05).
https://static-content.springer.com/image/art%3A10.1007%2Fs00223-007-9061-x/MediaObjects/223_2007_9061_Fig2_HTML.gif
Fig. 2

Percentage change in S-CTX (a), U-Ca (b), and bone ALP (c) at different clodronate doses over 6 weeks

The LSC data for markers S-CTX and U-Ca show a similar pattern to that seen with U-NTX, with no apparent increase in response at doses greater than 1,600 mg when only the breast cancer patients are considered but an apparent increasing dose response up to 2,400 mg for the other solid tumor population.

Bone ALP was the only formation marker measured in this study. No significant change was observed in the placebo group or lowest (800 mg) dose of clodronate. In the other clodronate groups, a modest increase (10–30% at 6 weeks) was observed.

Pain Score VAS Data

The great majority of the pain scores lay between 10 and 30 at baseline, signifying low- to medium-grade pain, with relatively few patients with no pain or severe pain. There was no significant difference in pain score at 6 weeks between the placebo and clodronate-treated groups, but within each treatment group there did appear to be a trend to lowering of pain score from baseline to 6 weeks (data not shown).

AEs and Toxicity

One hundred and five patients (84% of the study population) reported a total of 432 AEs at some time during the study. The majority of these (65% of the study population) were associated with gastrointestinal disorders. Table 5 shows a breakdown of the numbers of patients reporting AEs, where these were reported by more than 10% of the study population and the number of premature terminations of study due to AE by treatment group and by AE type. It should be noted that, even in the placebo group, substantial numbers of patients reported AEs. The numbers of terminations due to gastrointestinal problems (probably related to clodronate) were greater in the 2,400 mg and 3,200 mg arms than in the placebo, 800 mg, and 1,600 mg arms.
Table 5

Adverse events reported by more than 10% of the whole study population

Event

Clodronate dose

Placebo (n = 24)

800 mg (n = 27)

1,600 mg (n = 25)

2,400 mg (n = 26)

3,200 mg (n = 23)

Diarrhea

6 (25)

2 (7)

8 (32)

8 (31)

11(48)

Nausea

7 (29)

8 (30)

6 (24)

7 (27)

6 (26)

Vomiting

3 (13)

7 (26)

6 (24)

5 (19)

5 (22)

Pain

8 (33)

8 (30)

9 (36)

8 (31)

7 (30)

Dyspepsia

1 (4)

4 (15)

2 (8)

2 (8)

4 (17)

Withdrawal–gastrointestinal

2

1

2

4

4

Withdrawal–musculoskeletal

1

0

0

1

1

Withdrawal–other

1

1

1

0

2

Data are presented as number of patients, with percentage in parentheses.

Three patients experienced serious AEs that were possibly or probably drug-related: one patient in the 2,400 mg group (constipation, gastrointestinal hemorrhage and anemia) and two patients in the 3,200 mg group (one patient with hypocalcemia and the other with urinary retention).

Discussion

Despite the availability of oral clodronate for 20 years, this placebo-controlled trial is the first randomized and double-blind study to explore the dose-response relationships of this agent in oncology, using specific markers of bone metabolism. Baseline characteristics were well balanced, with the percentage of patients with marker values above the reference range at baseline consistent with previous studies of patients with metastatic bone disease when not receiving bisphosphonate therapy [3, 4].

Efficacy was assessed primarily in terms of the change in marker value with time on treatment compared with the pretreatment value, rather than the absolute value of markers or the numbers of patients for whom treatment reduced the markers into the reference range. In this sense, therefore, each patient was acting as his or her own control. For each of the resorption markers and in all of the treatment groups, there was a significant reduction compared to baseline over the 6-week study period. In the case of U-NTX and S-CTX, clodronate doses of ≥1,600 mg produced decreases of >40% from baseline value. Despite the oral route of administration, the majority of the reduction at doses of 1,600 mg and above occurred after only 1 week of treatment, with only small further changes during the following 5 weeks of study. Indeed, the “flatness” of the further response after the 1-week period suggests that substantial inhibition of bone resorption may occur in even less than 1 week and that measurements of bone markers at earlier time points would be of interest.

Although there appears to be a trend toward greater reduction in marker values as the dose increases from 1,600 mg upward because of the large confidence intervals and the relative flatness of the dose responses, it was not possible to determine statistically significant differences between the 1,600 mg dose and either of the two higher doses. However, analysis of the proportion of patients in each group failing to reach the LSC was informative. For the breast cancer subgroup (which comprises 87 out of the 125 patients in the study), while there is a clear increase in response from 800 to 1,600 mg, there appears to be no further increase in response at higher doses. This was seen for all three resorption markers. It is reassuring that the currently used 1,600-mg dose appears to be appropriate for breast cancer, which is the main licensed indication for clodronate in metastatic bone disease. However, in patients with other solid tumors (the great majority of whom were prostate cancer patients, for whom the drug is not currently licensed), the data suggest that higher doses may be needed for adequate response. This is consistent with observations by Coleman et al. [4] and supported by this study, that prostate cancer patients with metastatic bone disease have the highest level of bone resorption of any tumor type. Currently, only zoledronic acid has been shown to give significant reduction in skeletally related events, particularly fractures, in prostate cancer patients [22].

The “normalization” of bone resorption marker levels is now a clear, practical goal in directing the dosing and timing of bisphosphonate therapy in metastatic bone disease [4, 5]. In this study, approximately 50% of the patients with elevated bone markers achieved a sufficient response to clodronate for the bone resorption marker U-NTX to reach the defined reference range for normal individuals. While this 50% response rate appears somewhat less than that seen with more potent aminobisphosphonates such as pamidronate or zoledronic acid [4], it should be noted that direct comparisons of clodronate to aminobisphosphonates are very limited. In a previous small, open, randomized trial of oral clodronate vs. pamidronate, the biochemical response rate to pamidronate was higher, though the clodronate (1,600 mg daily) was given as two divided doses rather than a single dose and bone turnover was slightly higher in the clodronate group prior to treatment [23]. Indirect comparisons with the large randomized trials assessing zoledronic acid and pamidronate suggest that the response to clodronate is somewhat less. In the current study a maximum reduction in U-NTX with oral clodronate of around 50–60% was observed compared to approximately 70% and 60% with zoledronic acid [22] and pamidronate [24], respectively. However, such comparisons can be misleading because of differences in study populations, underlying cancer treatments, dosing schedules, and the timing of response assessment and should be interpreted with great caution.

The increases in bone ALP seen in this study are more difficult to explain. The changes were modest and took longer to occur than those seen with U-NTX, but there was a significant difference in the change in bone ALP levels at 6 weeks between 1,600 mg and placebo. Presumably, this reflects increased bone formation, and a similar rise was observed over this time scale by O’Rourke et al. [15], who interpreted this in terms of increased bone repair. A rise in bone ALP was also noted during the treatment of hypercalcemia with pamidronate [25]. Although the bone ALP rise observed in our study may represent progression of disease, the apparent dose-response effect argues against a disease effect. There may be transient stimulation of osteoblast activity, which is subsequently reversed or masked by the decrease in bone resorption. With longer follow-up and just as has been seen with other bisphosphonates, the coupling between resorption and formation would be expected to lead to a slowing of bone formation and subsequent reduction in levels of bone ALP. However, there are some in vitro data suggesting that bisphosphonates do have a stimulatory effect on osteoblast proliferation and maturation [26]. Alternatively, bisphosphonates may have effects on the recently described Dickkopf family of proteins and Wnt signaling pathways that regulate osteoblast differentiation and mineralized matrix formation [27].

The data for the effects of clodronate on pain were inconclusive. The VAS scores were highly scattered, and treatment groups did not show any significant effects vs. placebo, though there was some trend toward an overall reduction in pain score in the treatment groups as the study progressed. The lack of sensitivity of the score was probably related to the relatively low starting level of pain. It may also be that the 6-week period of the study was too short a time to expect any significant effect on pain scores. However, other studies have also failed to show convincing effects on pain following oral clodronate [12, 24, 28], unlike intravenous therapy with clodronate [29] or other intravenous bisphosphonates [3032].

Clodronate was generally well tolerated, and at doses of 800–2,400 mg the incidence of gastrointestinal AEs was similar to that of placebo. However, severe diarrhea occurred more frequently with the 3,200-mg dose. These data do not, however, indicate the seriousness of particular AEs, which is probably more appropriately reflected in the proportion of patients who discontinued the study because of an AE. Although total withdrawal numbers were small, there is some evidence that withdrawal may be dose-related, at 7%, 12%, 19%, and 30% in the 800, 1,600, 2,400, and 3,200 mg groups, respectively, compared to 8.3% in the placebo group.

Clodronate is still widely used in Europe for the management of metastatic bone disease to reduce skeletal morbidity. Although the newer agents delivered intravenously, such as zoledronic acid, are now making substantial contributions to bisphosphonate therapy, oral agents are more convenient and more easily used in less specialized environments or when patients have difficulty attending hospital. Until the recent approval of oral ibandronate, clodronate was the only oral bisphosphonate to be approved for metastatic bone disease. With the promising initial results of oral clodronate in the adjuvant setting and the large-scale ongoing studies of the use of adjuvant clodronate at 1,600 mg daily, it is likely that this drug will continue to have a role in oncology.

This study has confirmed that the 3,200-mg dose is associated with high AE and discontinuation rates and is probably not appropriate for general use. For breast cancer patients, the study suggests that there is little or no response benefit in increasing the dose beyond 1,600 mg and, therefore, gives reassurance that the commonly used 1,600-mg dose is appropriate. However, there was evidence that for non-breast cancer patients in the study (mainly prostate cancer patients) the 1,600-mg dose may be suboptimal and that a better response may be achieved with 2,400 mg. For such patients who may have rapid progression of bone disease and high rates of bone resorption, more potent bisphosphonates may be preferable [4], although recent evidence suggests that they may be associated with an increased risk of osteonecrosis of the jaws [33, 34].

Copyright information

© Springer Science+Business Media, LLC 2007