Predictive Value of Bone Resorption and Formation Markers in Cancer Patients With Bone Metastases Receiving the Bisphosphonate Zoledronic Ac
http://www.100md.com
《临床肿瘤学》
the Weston Park Hospital, Cancer Research Centre, Yorkshire Cancer Research Academic Unit of Clinical Oncology, Sheffield, United Kingdom
McMaster University, Juravinski Cancer Centre, Hamilton, Ontario
University of Waterloo, Department of Statistics and Actuarial Science, Waterloo, Ontario
Centre Hospitalier de l'Université de Montréal, H?pital Notre-Dame, Montréal, Quebec, Canada
Milton S. Hershey Medical Center, Division of Oncology, Hershey, PA
Massachusetts General Hospital, Department of Hematology/Oncology, Boston, MA
Novartis Pharmaceuticals Corp, East Hanover, NJ
ABSTRACT
PURPOSE: Three large, randomized trials of patients with bone metastases recently demonstrated that zoledronic acid reduces the risk of skeletal-related events. These trials provide an opportunity for investigating the correlation between bone metabolism and clinical outcome during bisphosphonate therapy.
PATIENTS AND METHODS: Urinary measurements of N-telopeptide (Ntx) and serum bone alkaline phosphatase (BAP) were obtained in 1,824 bisphosphonate-treated patients—1,462 with zoledronic acid (breast, 490; prostate, 411; myeloma, 210; non–small-cell lung, 183; other, 168) and 362 with pamidronate (breast, 254; myeloma, 108). This exploratory cohort analysis grouped patients by baseline and most recent levels of Ntx as low (< 50 nmol/mmol creatinine), moderate (50 to 99 nmol/mmol creatinine), or high ( 100 nmol/mmol creatinine), and BAP as low (< 146 U/L) or high ( 146 U/L). The relative risks for negative clinical outcomes were estimated for each group using multiple-event and Cox regression models with time-varying covariates.
RESULTS: Patients with high and moderate Ntx levels had 2-fold increases in their risk of skeletal complications and disease progression compared with patients with low Ntx levels (P < .001 for all). High Ntx levels in each solid tumor category were associated with a 4- to 6-fold increased risk of death on study, and moderate Ntx levels a 2- to 4-fold increased risk compared with low Ntx levels (P < .001 for all). Bone alkaline phosphatase also showed some correlation with risk of negative clinical outcomes.
CONCLUSION: The bone resorption marker Ntx provides valuable prognostic information in patients with bone metastases receiving bisphosphonates.
INTRODUCTION
Bone is the most frequent, typically the first, and often the only site of metastasis in patients with advanced solid tumors.1 The underlying pathophysiology of bone metastases involves the interaction of tumor cells with osteoclasts and osteoblasts, whose bone resorption and formation activities are normally tightly regulated. Bone metastases typically increase both bone resorption and formation rates, and these increases can be evaluated through measurements of biochemical markers of bone metabolism in the serum and urine of patients.2 Highly specific markers of osteolysis include N-telopeptide of type I collagen (Ntx), pyridinoline-crosslinked peptides, and deoxypyridinoline-crosslinked peptides. Markers of bone resorption include bone-specific alkaline phosphatase (BAP).3,4
Levels of biochemical markers of bone metabolism in serum or urine seem to correlate with the severity of bone pain and the extent of bone metastases.5,6 Small preliminary studies also suggest that a return of Ntx levels to normal during bisphosphonate treatment for bone metastases is associated with symptomatic response,7 a significant decrease in the rate of bony disease progression,6 and a trend for a decreased incidence of fractures.6 A recent study in 121 patients with metastatic bone disease (mostly from breast cancer) suggested that Ntx levels are a powerful predictor of skeletal complications.8 However, although other investigations have found correlations between bone markers and clinical outcomes,9-12 this has not been confirmed by a large study involving patients with different tumor types who received treatment with bisphosphonates, and the value of ongoing versus baseline assessments has not been formally assessed.
Recently, the results of three randomized, double-blind trials of the bisphosphonate zoledronic acid (ZOMETA; Novartis Pharma AG, Basel, Switzerland; Novartis Pharmaceuticals Corp, East Hanover, NJ) were published.13-15 For the approved 4-mg dose compared with placebo, reductions in risk of experiencing a skeletal-related event (SRE; based on Andersen Gill multiple event analysis) were 36% for prostate cancer,13 27% for non–small-cell lung cancer (NSCLC),14 and 28% for other solid tumors.14 In breast cancer and multiple myeloma, zoledronic acid reduced the risk of developing SREs by an additional 20% and 7%, respectively, compared with pamidronate.15 These findings set a new standard of care, whereas the placebo arm data served to underscore the incidence and severity of skeletal morbidity in patients with bone metastases. These trials also increased interest in identifying surrogate biologic markers to monitor the pathophysiology and progression of cancer in bone. These trials were the largest bisphosphonate trials ever conducted in metastatic bone disease (N = 3,064) and generated a rich database of bone marker and clinical outcomes statistics. This database provides a unique opportunity to investigate correlations between markers of bone metabolism and the occurrence of disease progression, survival, and clinically important SREs.
Herein, we present evidence of the association between high levels of biochemical markers of bone metabolism and negative clinical outcomes using data from the active treatment arms of the trials reported by Rosen et al14,15 and Saad et al.16 The analyses are based on time-to-event and multiple event regression models in which time-varying covariates were used to assess the risk of clinical events as a function of baseline and most recent (on-study) assessments of markers of bone remodeling.17-21
PATIENTS AND METHODS
Study Population
Recently, three randomized, multicenter, double-blind, phase III clinical trials evaluated the safety and efficacy of zoledronic acid in more than 3,000 patients with malignant bone disease from a broad range of primary cancers.13-15
For the breast cancer and multiple myeloma study, patients had stage IV breast cancer or Durie-Salmon stage III multiple myeloma. For the prostate cancer study, men had experienced disease progression while on hormone therapy and had serum testosterone less than 50 ng/dL. Patients with bone pain requiring strong narcotic therapy were excluded. For the study in other solid tumors, patients had stage IV disease. Patients in all three studies had one or more radiographically confirmed bone metastasis, Eastern Cooperative Oncology Group performance status of 0 to 2, serum creatinine less than 3 mg/dL (265 μmol/L), and all patients provided written informed consent. In all cases, anticancer treatment could be changed as clinically indicated throughout the course of the study.
Patients were randomly assigned to treatment with 4 mg zoledronic acid, 8 mg zoledronic acid, or control (90 mg pamidronate in the breast cancer and multiple myeloma study or placebo in the prostate and other solid tumors studies). Study drug was administered via intravenous infusion every 3 to 4 weeks for up to 24 months. During the course of the trial, the 8-mg zoledronic acid dose was reduced to 4 mg to ensure renal safety in all patients, and this group was then referred to as the 8/4-mg group. Pamidronate was administered at the standard dose (90 mg via 2-hour infusion) for the control group of patients with breast cancer or multiple myeloma. Only patients in the zoledronic acid and pamidronate treatment groups are included in these analyses. Analyses of the placebo-treated patients are reported elsewhere.22,23
Patient Evaluation
The main efficacy end points of these trials compared the number and timing of SREs (defined as pathologic fracture, surgery to bone to treat or prevent an impending fracture, palliative radiotherapy to bone, spinal cord compression, hypercalcemia of malignancy [included in secondary analyses], and changes in antineoplastic therapy for bone pain [only for patients with prostate cancer]) in the zoledronic acid treatment and control arms.
A radionuclide bone scan and a bone radiographic survey were performed at baseline and repeated quarterly or semiannually throughout the study. All imaging studies were reviewed by a central radiology facility blinded to study treatment. Disease progression was defined as the appearance of any new bone lesions or the progression of existing bone metastases. In the present analyses, only deaths occurring on study were considered.
Markers of bone metabolism, including serum BAP and urinary Ntx, were assessed at baseline; 1 month; and at 3, 6, 9, and 12 months during the period on study. Assessment of Ntx and BAP was performed in central laboratories (Mayo Medical Laboratories, Rochester, Minn; Bio Analytical Research Corp, Ghent, Belgium; Laboratorio Bioquimica Medica, Buenos Aires, Argentina; Fleury Laboratories, Sao Paulo, Brazil; Bio-Imaging Technologies Inc, Newtown, PA). Urinary Ntx was normalized to the level of urinary creatinine. Urine was obtained as a morning second-void sample. Serum BAP was measured in International Units per liter (U/L).
An exploratory cohort analysis based on bone marker levels was performed. Bone marker data were available for the majority of patients (Ntx, n = 2,183; BAP, n = 2,452), 1,824 of whom received treatment with bisphosphonates. Of these, 1,462 patients were treated with zoledronic acid (490 breast cancer, 411 prostate cancer, 210 multiple myeloma, 183 NSCLC, and 168 other tumors), and 362 patients were treated with pamidronate (254 breast cancer and 108 multiple myeloma).
Statistical Methods
Urinary Ntx was characterized as low (Ntx < 50 nmol/mmol creatinine), moderate (Ntx = 50 to 99 nmol/mmol creatinine), or high (Ntx 100 nmol/mmol creatinine), and BAP levels were characterized as low (BAP < 146 U/L) or high (BAP 146 U/L). The cutoff values for Ntx were chosen to reflect the approximate upper limit of normal (ULN). However, the normal range for urinary Ntx varies according to age, sex, and endocrine function, with approximately 50 nmol/mmol creatinine being the approximate ULN in young healthy adults and approximately 100 nmol/mmol creatinine being the approximate ULN after menopause in women or during androgen deprivation therapy in men. Therefore, analyses based on both 50 and 100 nmol/mmol cutoff values were performed. The BAP cutoff value was selected based on available data for the normal range of BAP and from our collective clinical experience in the oncology setting.23
Clinical outcomes were similar in both zoledronic acid treatment groups in each trial.13-15 Additionally, the biochemical response to treatment in terms of bone markers was very similar for the two zoledronic acid treatment arms. Therefore, for the purpose of these analyses, the 4- and 8/4-mg zoledronic acid treatment arm data have been pooled. This has two advantages: first, all of the available treatment arm data are considered; second, the increased sample size improves the statistical power of the estimated relative risk ratios.
Intensity-based multiple event regression models were applied to assess the predictive significance of marker levels for all SREs because SREs may occur repeatedly over time.17 Cox regression models were used to assess the correlation between bone marker levels and first SRE, progression of bone disease, and death.17,19,20 For each outcome (all SREs, first SRE, disease progression, and death), regression models were stratified according to the prospective strata of the trials. The predictive significance of on-study marker assessments was assessed by the inclusion of time-varying covariates indicating the most recent marker state (moderate v low and high v low for Ntx; high v low for BAP). Marker assessments were scheduled at baseline, 1 month, and on a quarterly basis thereafter. However, to minimize the impact of missed assessments, values were considered recent for up to 6 months. Following a 6-month period with no marker assessment, patients were no longer included in the at-risk set until they received a subsequent marker level assessment, at which time they were re-entered into the at-risk set.
Robust estimates of the proportion of patients in the low, moderate, and high marker states were obtained by estimating time-varying transition probabilities in a first-order Markov model.21 This characterized each patient into one of four states for Ntx or three states for BAP: alive with a low marker level, alive with a moderate marker level (for Ntx only), alive with a high marker level, or deceased. These methods have recently been shown to provide robust estimates of state occupancy probabilities that reflect the percent of patients in each of the multiple states over time.18,24,25 Log-linear models were also fitted to estimate the effect of zoledronic acid on changes in marker states over time based on the Markov model.26
RESULTS
Patient Demographics and Skeletal Disease
Patient demographics and disease characteristics for the 1,824 patients included in this analysis are presented in Table 1. 14-16 Overall, 2,688 SREs occurred in bisphosphonate-treated patients whose bone marker levels were evaluated.
Bone Markers at Baseline and in Response to Treatment
The median time under observation was approximately 17 months for the patients included in this analysis. The proportions of patients with low, moderate, and high Ntx levels at baseline for each cancer type are shown in Figure 1. The proportions of patients with high and moderate Ntx levels during the study decreased during treatment but showed different patterns of response to bisphosphonate therapy according to the underlying tumor type (Fig 2A for high Ntx and Fig 2B for moderate Ntx). Patients with prostate cancer, on average, had the highest Ntx values at baseline, and approximately 25% of patients with prostate cancer had high or moderate Ntx at each on-study assessment. At the other extreme, patients with multiple myeloma, on average, had the lowest baseline Ntx levels, and the proportion with high or moderate Ntx levels during the studies ranged from 1% to 18%. This finding is at first sight counterintuitive but probably reflects the efficacy of background systemic chemotherapy for multiple myeloma and the exclusion of patients with hypercalcemia from these zoledronic acid trials. Patients with breast cancer, NSCLC, and other solid tumors had mean baseline and on-study Ntx levels that were intermediate compared with the prostate cancer and multiple myeloma groups. As expected, there was a correlation between baseline values of Ntx and BAP levels and surrogates for tumor bulk such as baseline prostate-specific antigen (r = 0.155, P = .002 for Ntx, and r = 0.175, p = < 0.001 for BAP). Within 1 month, the specific effects of zoledronic acid on bone metabolism (as opposed to tumor bulk) led to a weakening of this correlation (r = –0.111, P = .037 for Ntx, and r = 0.040, P = .435 for BAP).
Proportion of Patients With an SRE
In the overall population of patients, moderate or high Ntx or high BAP levels while on study correlated with a higher risk for SREs irrespective of primary cancer type (Fig 3 14-16 and Table 2). 17,20 In each disease group, patients with the highest levels of Ntx had a two- to three-fold higher risk of experiencing SREs compared with patients with low Ntx levels. The relative risk of an SRE for high Ntx versus low Ntx was highest for patients with breast cancer (relative risk ratio [RR] = 2.96; P < .001; Fig 4) and lowest for NSCLC (RR = 1.89; P = .111). The increased risk of SREs associated with moderate Ntx levels compared with low Ntx levels was also significant in each disease category (Fig 5). Although there was some variability in the significance of this correlation between primary cancer types, there was insufficient evidence to suggest heterogeneity of effects by disease type (P = .813).
In the overall population, high BAP was associated with a significant two-fold increase in the risk of SREs compared with low BAP. The correlation was weakest for solid tumors other than breast or prostate cancer (RR = 1.29; P = .180) and strongest for prostate cancer (RR = 3.29; P < .001). There was strong evidence of heterogeneity in the correlation between high BAP and risk of all SREs (P = .001), although the trend was similar in all patient groups.
First SRE
High Ntx levels showed a significant correlation with the risk of experiencing a first SRE on study for the overall population of patients (P < .001) and for each disease type except NSCLC and other solid tumors. Relative risk ratios for patients with high Ntx levels compared with low Ntx levels ranged from 1.56 for NSCLC (P = .306) to 3.01 for multiple myeloma (P = .008; Table 2). Moderate Ntx levels were also associated with a significantly increased risk of experiencing a first SRE on study for the overall patient population (P < .001) and for breast cancer, NSCLC, and other solid tumors. There was no suggestion of heterogeneity in the relative risk between groups by primary disease (P = .929).
High BAP was also associated with an increased risk of experiencing a first on-study SRE in the overall population (RR = 1.95, P < .001) and in all disease types except NSCLC and other solid tumors (Table 3). 17,19 There was significant heterogeneity in the correlation between high BAP and time to first SRE across disease types (P = .0007).
Progression of Bone Lesions
Overall, there was a 2.21-fold increased risk of bone lesion progression for patients with high Ntx versus low Ntx levels (P < .001) and a 1.57-fold increase for patients with moderate Ntx versus low Ntx levels (P < .001 for both comparisons; Table 2). There was consistent and strong evidence (P .001) that high Ntx was associated with an increased risk of disease progression for patients with solid tumors (breast cancer, prostate cancer, NSCLC, and other solid tumors), and there was a trend toward increased risk associated with high Ntx in patients with multiple myeloma (P = .080). However, there was considerable variation within the multiple myeloma group. There was also a trend toward increased risk of progression among patients with moderate Ntx levels in each of the disease types, although this reached statistical significance only for patients with breast cancer (P < .001; Fig 4) or prostate cancer (P = .015). There was no evidence of heterogeneity of effects across cancer types (P = .988).
In the overall patient population, high BAP was associated with a significantly increased risk of progression of bone lesions (P < .001). However, in contrast with the Ntx assessment, there was significant heterogeneity in this correlation (P = .043). Among patients with solid tumors, there was a significant 1.77- to 2.42-fold increased relative risk of bone lesion progression for patients with high BAP, whereas there was no significant relationship between BAP and bone lesion progression in patients with multiple myeloma (P = .881).
Death
In the overall population of patients and in each of the solid tumor disease types, high or moderate Ntx levels correlated with a significantly greater relative risk of dying on study (P < .001), whereas the correlation between Ntx level and death was not significant in patients with multiple myeloma (Table 2). This probably reflected the low frequency of elevated Ntx on treatment in the myeloma cohort (see Differential Response to Zoledronic Acid and Pamidronate section). Overall, high Ntx and moderate Ntx were associated with a 4.80-fold increased risk of death (95% CI, 3.90 to 5.91; P < .001) and a 3.11-fold increased risk of death (95% CI, 2.54 to 3.79; P < .001), respectively, compared with low Ntx levels.
The correlations between high BAP and relative risk of death during the assessment period are summarized in Table 3. There was a consistently significant association between high BAP and mortality, though the RRs showed considerable variation between different cancer types.
Differential Responses to Zoledronic Acid and Pamidronate
Figure 6 shows the proportion of patients with breast cancer or multiple myeloma treated with either zoledronic acid or pamidronate who had elevated Ntx levels during the course of the clinical trials. Figure 6A uses a cutoff of 100 nmol/mmol creatinine (high) for Ntx, and Figure 6B uses a cutoff of 50 to 99 nmol/mmol creatinine (moderate). Zoledronic acid was more effective than pamidronate at suppressing high rates of bone resorption, reducing Ntx levels from a high to a low state (RR = 2.73; 95% CI, 1.60 to 4.69), and was similarly effective at reducing Ntx levels from a moderate to a low level (Table 4). For multiple myeloma, the proportion of patients with high Ntx on study was very low with both treatments, possibly indicating maximal suppression of osteoclast activity by either bisphosphonate. In contrast, in the breast cancer subset, in whom zoledronic acid has demonstrated significant benefits beyond those of pamidronate,15 more patients had high Ntx levels at baseline compared with the multiple myeloma cohort, and a smaller proportion of patients treated with zoledronic acid had high Ntx levels on study compared with pamidronate.
DISCUSSION
The measurement of biochemical markers of bone metabolism has become more common in recent trials in patients with bone metastases. Assessment of these markers provides specific insight into rates of bone formation and resorption,2-4 and levels seem to correlate with bone pain levels and the burden of disease in the skeleton, while being unaffected by the burden of soft tissue or visceral disease.4-7,27 Decreases in marker levels have also been found to correlate with response to systemic anticancer treatments8,28,29 and symptomatic response to skeletal radiotherapy.30 However, all of these reports have been based on small cohorts of patients or are based on short-term (a few weeks) outcomes. As a result, the use of bone markers as prognostic tools during therapy has remained academic,31 pending robust information from large clinical trials.
Recently we reported the correlation between bone marker levels and clinical outcomes in patients with prostate cancer and a variety of other solid tumors (excluding breast cancer) who were receiving concomitant standard antineoplastic therapy.22 These patients were not treated with bisphosphonates, whereas in the exploratory cohort analysis reported herein, all patients received ongoing bisphosphonate treatment (zoledronic acid; some comparisons are also made between zoledronic acid and pamidronate in patients with multiple myeloma or breast cancer), consistent with current standard practice. The correlation between bone marker levels and clinical outcomes, including SREs, disease progression, and death, were analyzed and found to be statistically significant. These analyses used data collected during three large, multicenter, randomized, controlled, phase III trials of zoledronic acid.13-15 Characterizing patients by their most recent Ntx or BAP levels revealed clear relationships between bone marker levels and imminent risk of skeletal morbidity, bone lesion progression, and death. The risk of any SRE, time to first on-study SRE, disease progression, and death were all greater during periods of accelerated bone resorption (as indicated by high Ntx level) compared with periods of lower bone resorption activity (as indicated by low Ntx level).
The relationships between Ntx levels and outcome were broadly similar, irrespective of the underlying tumor type. Additionally, for the breast cancer and multiple myeloma patients, the prognostic significance of high Ntx levels was similar for patients treated with either zoledronic acid or pamidronate. However, because of the increased potency of zoledronic acid over pamidronate, the proportion of patients with high Ntx levels was always lower in the zoledronic acid group compared with pamidronate.32 Perhaps because relatively few of the patients with multiple myeloma had high Ntx at baseline and during the trial, both drugs were able to effectively reduce levels of Ntx in patients with multiple myeloma at more than 95% of the time points. The similarity in biochemical response reflects the similar risk of SREs among patients with multiple myeloma treated with zoledronic acid and pamidronate.15 In contrast, high Ntx was more common in patients with breast cancer, and zoledronic acid was more effective than pamidronate at reducing the proportion of patients with breast cancer in the high Ntx category. Perhaps as a result of this improved biochemical response, breast cancer patients treated with zoledronic acid had an additional 20% reduction in the risk of developing SREs compared with the pamidronate group (RR = 0.799; P = .025).15 Therefore, we believe that these results support the conclusion that the greater potency of zoledronic acid compared with pamidronate is responsible for its increased clinical activity in patients with breast cancer.
Based on these analyses, all patients with malignant bone disease are at risk for SREs, but if patients have high Ntx levels at baseline or on treatment, they are at a higher risk for negative clinical outcomes. Because bone marker levels change during the course of the disease, recent ( 6 months old) marker assessments have greater predictive value than the baseline assessment at later time points during the trial.22 For example, changes in the underlying systemic therapies were allowed during the course of these trials, and these modifications would, in themselves, be expected to influence bone marker levels. Patients who developed or maintained high Ntx or BAP levels during the course of zoledronic acid therapy were most at risk for deterioration in skeletal integrity, SREs, and a poor clinical outcome. Furthermore, because relative risks were calculated based on bone marker values taken up to 6 months before the events, increases in Ntx might precede clinical deterioration by several months. Thus, high Ntx levels could signal the need for more frequent follow-up. However, further studies may be necessary to determine whether the relationships can be generalized to patients treated with other bisphosphonates.
The results of this analysis and those previously published for the placebo arms of the clinical trials in patients with bone metastases from prostate cancer and patients with NSCLC or other solid tumors support the conclusion that recent Ntx levels significantly correlate with risk of negative clinical outcomes.22 Although the RRs for each negative clinical outcome varied depending on primary malignancy, the patterns of risk were consistent throughout each patient subset. Therefore, regardless of the primary malignancy, the level of bone resorption is closely correlated with malignant bone disease pathology, and the prognostic implications of Ntx assessments could be broadly applicable in the oncology setting.
Although these analyses cannot directly address a causal link between decreases in bone metabolism and clinical outcome, they do suggest that reducing bone turnover should have a positive effect by delaying progression of bone lesions and possibly improving survival. These observations will need to be confirmed in prospective randomized studies. We believe that these data support the hypothesis that bone marker measurements can provide meaningful additional data for assessing the risk of negative clinical outcomes in patients with malignant bone disease.
Authors' Disclosures of Potential Conflicts of Interest
Although all authors have completed the disclosure declaration, the following authors or their immediate family members have indicated a financial interest. No conflict exists for drugs or devices used in a study if they are not being evaluated as part of the investigation. For a detailed description of the disclosure categories, or for more information about ASCO’s conflict of interest policy, please refer to the Author Disclosure Declaration and the Disclosures of Potential Conflicts of Interest section in Information for Contributors.
NOTES
Presented in part at the 39th Annual Meeting of the American Society of Clinical Oncology, Chicago, IL, May 31-June 3, 2003.
Authors' disclosures of potential conflicts of interest are found at the end of this article.
REFERENCES
Mundy GR: Metastasis to bone: Causes, consequences and therapeutic opportunities. Nat Rev Cancer 2: 584-593, 2002
Coleman RE: The clinical use of bone resorption markers in patients with malignant bone disease. Cancer 94: 2521-2533, 2002
Coleman RE, Whitaker KB, Moss DW, et al: Biochemical prediction of response of bone metastases to treatment. Br J Cancer 58: 205-210, 1988
Berruti A, Dogliotti L, Gorzegno G, et al: Differential patterns of bone turnover in relation to bone pain and disease extent in bone in cancer patients with skeletal metastases. Clin Chem 45: 1240-1247, 1999
Vinholes J, Guo CY, Purohit OP, et al: Metabolic effects of pamidronate in patients with metastatic bone disease. Br J Cancer 73: 1089-1095, 1996
Lipton A, Demers L, Curley E, et al: Markers of bone resorption in patients treated with pamidronate. Eur J Cancer 34: 2021-2026, 1998
Vinholes JJ, Purohit OP, Abbey ME, et al: Relationships between biochemical and symptomatic response in a double-blind randomised trial of pamidronate for metastatic bone disease. Ann Oncol 8: 1243-1250, 1997
Brown JE, Thomson CS, Ellis SP, et al: Bone resorption predicts for skeletal complications in metastatic bone disease. Br J Cancer 89: 2031-2037, 2003
Glover D, Lipton A, Keller A, et al: Intravenous pamidronate disodium treatment of bone metastases in patients with breast cancer: A dose-seeking study. Cancer 74: 2949-2955, 1994
Berruti A, Dogliotti L, Bitossi R, et al: Incidence of skeletal complications in patients with bone metastatic prostate cancer and hormone refractory disease: Predictive role of bone resorption and formation markers evaluated at baseline. J Urol 164: 1248-1253, 2000
Samma S, Kagebayashi Y, Yasukawa M, et al: Sequential changes of urinary pyridinoline and deoxypyridinoline as markers of metastatic bone tumor in patients with prostate cancer: A preliminary study. Jpn J Clin Oncol 27: 26-30, 1997
Luftner D, Richter A, Geppert R, et al: Normalisation of biochemical markers of bone formation correlates with clinical benefit from therapy in metastatic breast cancer. Anticancer Res 23: 1017-1026, 2003
Saad F, Gleason DM, Murray R, et al: Long-term efficacy of zoledronic acid for the prevention of skeletal complications in patients with metastatic hormone-refractory prostate cancer. J Natl Cancer Inst 96: 879-882, 2004
Rosen LS, Gordon D, Tchekmedyian NS, et al: Long-term efficacy and safety of zoledronic acid in the treatment of skeletal metastases in patients with nonsmall cell lung carcinoma and other solid tumors: A randomized, phase III, double-blind, placebo-controlled trial. Cancer 100: 2613-2621, 2004
Rosen LS, Gordon D, Kaminski M, et al: Long-term efficacy and safety of zoledronic acid compared with pamidronate disodium in the treatment of skeletal complications in patients with advanced multiple myeloma or breast carcinoma: A randomized, double-blind, multicenter, comparative trial. Cancer 98: 1735-1744, 2003
Saad F, Gleason DM, Murray R, et al: A randomized, placebo-controlled trial of zoledronic acid in patients with hormone-refractory metastatic prostate carcinoma. For the Zoledronic Acid Prostate Cancer Study Group. J Natl Cancer Inst 94: 1458-1468, 2002
Andersen PK, Borgan O, Gill RD, et al: Statistical Models Based on Counting Processes. New York, NY, Springer-Verlag, 1993
Datta S, Satten GA: Validity of the Aalen-Johansen estimators of stage occupation probabilities and Nelson-Aalen estimators of integrated transition hazards for non-Markov models. Statistics & Probability Letters 55: 403-411, 2001
Kalbfleish JD, Prentice RL: The Statistical Analysis of Failure Time Data. Hoboken, NJ, John Wiley & Sons Inc, 2002
Lawless JF: Statistical Models and Methods for Lifetime Data. New York, NY, John Wiley & Sons Inc, 2002
Ross SM: Editor: Introduction to Probability Models. San Diego, CA, Academic Press Inc, 1989
Brown JE, Cook RJ, Major P, et al: Bone markers as predictors of skeletal complications in prostate cancer, lung cancer, and other solid tumors. J Natl Cancer Inst 97: 59-69, 2005
Demers LM, Costa L, Lipton A: Biochemical markers and skeletal metastases. Cancer 88: 2919-2926, 2000 (suppl)
Datta S, Satten GA: Estimation of integrated transition hazards and stage occupation probabilities for non-Markov systems under dependent censoring. Biometrics 58: 792-802, 2002
Aalen OO, Borgan O, Fekjaer H: Covariate adjustment of event histories estimated from Markov chains: The additive approach. Biometrics 57: 993-1001, 2001
Agresti A: Categorical data analysis. Honecker, NJ, John Wiley & Sons Inc, 2002
Demers LM, Costa L, Lipton A: Biochemical markers and skeletal metastases. Clin Orthop 415: S138-S147, 2003 (special issue)
Vinholes J, Coleman R, Lacombe D, et al: Assessment of bone response to systemic therapy in an EORTC trial: Preliminary experience with the use of collagen cross-link excretion. European Organization for Research and Treatment of Cancer. Br J Cancer 80: 221-228, 1999
Costa L, Demers LM, Gouveia-Oliveira A, et al: Prospective evaluation of the peptide-bound collagen type I cross-links N-telopeptideand C-telopeptide in predicting bone metastases status. J Clin Oncol 20: 850-856, 2002
Hoskin PJ, Stratford MR, Folkes LK, et al: Effect of local radiotherapy for bone pain on urinary markers of osteoclast activity. Lancet 355: 1428-1429, 2000
Hillner BE, Ingle JN, Chelbowski RT, et al: American Society of Clinical Oncology 2003 update on the role of bisphosphonates and bone health issues in women with breast cancer. Published erratum in: J Clin Oncol 2004;22:1351. J Clin Oncol 21: 4042-4057, 2003
Coleman RE, Cook R, Major P, et al: Why is zoledronic acid (Z) superior to pamidronate (P) for bone metastases from breast cancer (BC) but equivalent in multiple myeloma (MM)? Proc Am Soc Clin Oncol 22: 47, 2003 (abstr 187)(Robert E. Coleman, Pierre)
McMaster University, Juravinski Cancer Centre, Hamilton, Ontario
University of Waterloo, Department of Statistics and Actuarial Science, Waterloo, Ontario
Centre Hospitalier de l'Université de Montréal, H?pital Notre-Dame, Montréal, Quebec, Canada
Milton S. Hershey Medical Center, Division of Oncology, Hershey, PA
Massachusetts General Hospital, Department of Hematology/Oncology, Boston, MA
Novartis Pharmaceuticals Corp, East Hanover, NJ
ABSTRACT
PURPOSE: Three large, randomized trials of patients with bone metastases recently demonstrated that zoledronic acid reduces the risk of skeletal-related events. These trials provide an opportunity for investigating the correlation between bone metabolism and clinical outcome during bisphosphonate therapy.
PATIENTS AND METHODS: Urinary measurements of N-telopeptide (Ntx) and serum bone alkaline phosphatase (BAP) were obtained in 1,824 bisphosphonate-treated patients—1,462 with zoledronic acid (breast, 490; prostate, 411; myeloma, 210; non–small-cell lung, 183; other, 168) and 362 with pamidronate (breast, 254; myeloma, 108). This exploratory cohort analysis grouped patients by baseline and most recent levels of Ntx as low (< 50 nmol/mmol creatinine), moderate (50 to 99 nmol/mmol creatinine), or high ( 100 nmol/mmol creatinine), and BAP as low (< 146 U/L) or high ( 146 U/L). The relative risks for negative clinical outcomes were estimated for each group using multiple-event and Cox regression models with time-varying covariates.
RESULTS: Patients with high and moderate Ntx levels had 2-fold increases in their risk of skeletal complications and disease progression compared with patients with low Ntx levels (P < .001 for all). High Ntx levels in each solid tumor category were associated with a 4- to 6-fold increased risk of death on study, and moderate Ntx levels a 2- to 4-fold increased risk compared with low Ntx levels (P < .001 for all). Bone alkaline phosphatase also showed some correlation with risk of negative clinical outcomes.
CONCLUSION: The bone resorption marker Ntx provides valuable prognostic information in patients with bone metastases receiving bisphosphonates.
INTRODUCTION
Bone is the most frequent, typically the first, and often the only site of metastasis in patients with advanced solid tumors.1 The underlying pathophysiology of bone metastases involves the interaction of tumor cells with osteoclasts and osteoblasts, whose bone resorption and formation activities are normally tightly regulated. Bone metastases typically increase both bone resorption and formation rates, and these increases can be evaluated through measurements of biochemical markers of bone metabolism in the serum and urine of patients.2 Highly specific markers of osteolysis include N-telopeptide of type I collagen (Ntx), pyridinoline-crosslinked peptides, and deoxypyridinoline-crosslinked peptides. Markers of bone resorption include bone-specific alkaline phosphatase (BAP).3,4
Levels of biochemical markers of bone metabolism in serum or urine seem to correlate with the severity of bone pain and the extent of bone metastases.5,6 Small preliminary studies also suggest that a return of Ntx levels to normal during bisphosphonate treatment for bone metastases is associated with symptomatic response,7 a significant decrease in the rate of bony disease progression,6 and a trend for a decreased incidence of fractures.6 A recent study in 121 patients with metastatic bone disease (mostly from breast cancer) suggested that Ntx levels are a powerful predictor of skeletal complications.8 However, although other investigations have found correlations between bone markers and clinical outcomes,9-12 this has not been confirmed by a large study involving patients with different tumor types who received treatment with bisphosphonates, and the value of ongoing versus baseline assessments has not been formally assessed.
Recently, the results of three randomized, double-blind trials of the bisphosphonate zoledronic acid (ZOMETA; Novartis Pharma AG, Basel, Switzerland; Novartis Pharmaceuticals Corp, East Hanover, NJ) were published.13-15 For the approved 4-mg dose compared with placebo, reductions in risk of experiencing a skeletal-related event (SRE; based on Andersen Gill multiple event analysis) were 36% for prostate cancer,13 27% for non–small-cell lung cancer (NSCLC),14 and 28% for other solid tumors.14 In breast cancer and multiple myeloma, zoledronic acid reduced the risk of developing SREs by an additional 20% and 7%, respectively, compared with pamidronate.15 These findings set a new standard of care, whereas the placebo arm data served to underscore the incidence and severity of skeletal morbidity in patients with bone metastases. These trials also increased interest in identifying surrogate biologic markers to monitor the pathophysiology and progression of cancer in bone. These trials were the largest bisphosphonate trials ever conducted in metastatic bone disease (N = 3,064) and generated a rich database of bone marker and clinical outcomes statistics. This database provides a unique opportunity to investigate correlations between markers of bone metabolism and the occurrence of disease progression, survival, and clinically important SREs.
Herein, we present evidence of the association between high levels of biochemical markers of bone metabolism and negative clinical outcomes using data from the active treatment arms of the trials reported by Rosen et al14,15 and Saad et al.16 The analyses are based on time-to-event and multiple event regression models in which time-varying covariates were used to assess the risk of clinical events as a function of baseline and most recent (on-study) assessments of markers of bone remodeling.17-21
PATIENTS AND METHODS
Study Population
Recently, three randomized, multicenter, double-blind, phase III clinical trials evaluated the safety and efficacy of zoledronic acid in more than 3,000 patients with malignant bone disease from a broad range of primary cancers.13-15
For the breast cancer and multiple myeloma study, patients had stage IV breast cancer or Durie-Salmon stage III multiple myeloma. For the prostate cancer study, men had experienced disease progression while on hormone therapy and had serum testosterone less than 50 ng/dL. Patients with bone pain requiring strong narcotic therapy were excluded. For the study in other solid tumors, patients had stage IV disease. Patients in all three studies had one or more radiographically confirmed bone metastasis, Eastern Cooperative Oncology Group performance status of 0 to 2, serum creatinine less than 3 mg/dL (265 μmol/L), and all patients provided written informed consent. In all cases, anticancer treatment could be changed as clinically indicated throughout the course of the study.
Patients were randomly assigned to treatment with 4 mg zoledronic acid, 8 mg zoledronic acid, or control (90 mg pamidronate in the breast cancer and multiple myeloma study or placebo in the prostate and other solid tumors studies). Study drug was administered via intravenous infusion every 3 to 4 weeks for up to 24 months. During the course of the trial, the 8-mg zoledronic acid dose was reduced to 4 mg to ensure renal safety in all patients, and this group was then referred to as the 8/4-mg group. Pamidronate was administered at the standard dose (90 mg via 2-hour infusion) for the control group of patients with breast cancer or multiple myeloma. Only patients in the zoledronic acid and pamidronate treatment groups are included in these analyses. Analyses of the placebo-treated patients are reported elsewhere.22,23
Patient Evaluation
The main efficacy end points of these trials compared the number and timing of SREs (defined as pathologic fracture, surgery to bone to treat or prevent an impending fracture, palliative radiotherapy to bone, spinal cord compression, hypercalcemia of malignancy [included in secondary analyses], and changes in antineoplastic therapy for bone pain [only for patients with prostate cancer]) in the zoledronic acid treatment and control arms.
A radionuclide bone scan and a bone radiographic survey were performed at baseline and repeated quarterly or semiannually throughout the study. All imaging studies were reviewed by a central radiology facility blinded to study treatment. Disease progression was defined as the appearance of any new bone lesions or the progression of existing bone metastases. In the present analyses, only deaths occurring on study were considered.
Markers of bone metabolism, including serum BAP and urinary Ntx, were assessed at baseline; 1 month; and at 3, 6, 9, and 12 months during the period on study. Assessment of Ntx and BAP was performed in central laboratories (Mayo Medical Laboratories, Rochester, Minn; Bio Analytical Research Corp, Ghent, Belgium; Laboratorio Bioquimica Medica, Buenos Aires, Argentina; Fleury Laboratories, Sao Paulo, Brazil; Bio-Imaging Technologies Inc, Newtown, PA). Urinary Ntx was normalized to the level of urinary creatinine. Urine was obtained as a morning second-void sample. Serum BAP was measured in International Units per liter (U/L).
An exploratory cohort analysis based on bone marker levels was performed. Bone marker data were available for the majority of patients (Ntx, n = 2,183; BAP, n = 2,452), 1,824 of whom received treatment with bisphosphonates. Of these, 1,462 patients were treated with zoledronic acid (490 breast cancer, 411 prostate cancer, 210 multiple myeloma, 183 NSCLC, and 168 other tumors), and 362 patients were treated with pamidronate (254 breast cancer and 108 multiple myeloma).
Statistical Methods
Urinary Ntx was characterized as low (Ntx < 50 nmol/mmol creatinine), moderate (Ntx = 50 to 99 nmol/mmol creatinine), or high (Ntx 100 nmol/mmol creatinine), and BAP levels were characterized as low (BAP < 146 U/L) or high (BAP 146 U/L). The cutoff values for Ntx were chosen to reflect the approximate upper limit of normal (ULN). However, the normal range for urinary Ntx varies according to age, sex, and endocrine function, with approximately 50 nmol/mmol creatinine being the approximate ULN in young healthy adults and approximately 100 nmol/mmol creatinine being the approximate ULN after menopause in women or during androgen deprivation therapy in men. Therefore, analyses based on both 50 and 100 nmol/mmol cutoff values were performed. The BAP cutoff value was selected based on available data for the normal range of BAP and from our collective clinical experience in the oncology setting.23
Clinical outcomes were similar in both zoledronic acid treatment groups in each trial.13-15 Additionally, the biochemical response to treatment in terms of bone markers was very similar for the two zoledronic acid treatment arms. Therefore, for the purpose of these analyses, the 4- and 8/4-mg zoledronic acid treatment arm data have been pooled. This has two advantages: first, all of the available treatment arm data are considered; second, the increased sample size improves the statistical power of the estimated relative risk ratios.
Intensity-based multiple event regression models were applied to assess the predictive significance of marker levels for all SREs because SREs may occur repeatedly over time.17 Cox regression models were used to assess the correlation between bone marker levels and first SRE, progression of bone disease, and death.17,19,20 For each outcome (all SREs, first SRE, disease progression, and death), regression models were stratified according to the prospective strata of the trials. The predictive significance of on-study marker assessments was assessed by the inclusion of time-varying covariates indicating the most recent marker state (moderate v low and high v low for Ntx; high v low for BAP). Marker assessments were scheduled at baseline, 1 month, and on a quarterly basis thereafter. However, to minimize the impact of missed assessments, values were considered recent for up to 6 months. Following a 6-month period with no marker assessment, patients were no longer included in the at-risk set until they received a subsequent marker level assessment, at which time they were re-entered into the at-risk set.
Robust estimates of the proportion of patients in the low, moderate, and high marker states were obtained by estimating time-varying transition probabilities in a first-order Markov model.21 This characterized each patient into one of four states for Ntx or three states for BAP: alive with a low marker level, alive with a moderate marker level (for Ntx only), alive with a high marker level, or deceased. These methods have recently been shown to provide robust estimates of state occupancy probabilities that reflect the percent of patients in each of the multiple states over time.18,24,25 Log-linear models were also fitted to estimate the effect of zoledronic acid on changes in marker states over time based on the Markov model.26
RESULTS
Patient Demographics and Skeletal Disease
Patient demographics and disease characteristics for the 1,824 patients included in this analysis are presented in Table 1. 14-16 Overall, 2,688 SREs occurred in bisphosphonate-treated patients whose bone marker levels were evaluated.
Bone Markers at Baseline and in Response to Treatment
The median time under observation was approximately 17 months for the patients included in this analysis. The proportions of patients with low, moderate, and high Ntx levels at baseline for each cancer type are shown in Figure 1. The proportions of patients with high and moderate Ntx levels during the study decreased during treatment but showed different patterns of response to bisphosphonate therapy according to the underlying tumor type (Fig 2A for high Ntx and Fig 2B for moderate Ntx). Patients with prostate cancer, on average, had the highest Ntx values at baseline, and approximately 25% of patients with prostate cancer had high or moderate Ntx at each on-study assessment. At the other extreme, patients with multiple myeloma, on average, had the lowest baseline Ntx levels, and the proportion with high or moderate Ntx levels during the studies ranged from 1% to 18%. This finding is at first sight counterintuitive but probably reflects the efficacy of background systemic chemotherapy for multiple myeloma and the exclusion of patients with hypercalcemia from these zoledronic acid trials. Patients with breast cancer, NSCLC, and other solid tumors had mean baseline and on-study Ntx levels that were intermediate compared with the prostate cancer and multiple myeloma groups. As expected, there was a correlation between baseline values of Ntx and BAP levels and surrogates for tumor bulk such as baseline prostate-specific antigen (r = 0.155, P = .002 for Ntx, and r = 0.175, p = < 0.001 for BAP). Within 1 month, the specific effects of zoledronic acid on bone metabolism (as opposed to tumor bulk) led to a weakening of this correlation (r = –0.111, P = .037 for Ntx, and r = 0.040, P = .435 for BAP).
Proportion of Patients With an SRE
In the overall population of patients, moderate or high Ntx or high BAP levels while on study correlated with a higher risk for SREs irrespective of primary cancer type (Fig 3 14-16 and Table 2). 17,20 In each disease group, patients with the highest levels of Ntx had a two- to three-fold higher risk of experiencing SREs compared with patients with low Ntx levels. The relative risk of an SRE for high Ntx versus low Ntx was highest for patients with breast cancer (relative risk ratio [RR] = 2.96; P < .001; Fig 4) and lowest for NSCLC (RR = 1.89; P = .111). The increased risk of SREs associated with moderate Ntx levels compared with low Ntx levels was also significant in each disease category (Fig 5). Although there was some variability in the significance of this correlation between primary cancer types, there was insufficient evidence to suggest heterogeneity of effects by disease type (P = .813).
In the overall population, high BAP was associated with a significant two-fold increase in the risk of SREs compared with low BAP. The correlation was weakest for solid tumors other than breast or prostate cancer (RR = 1.29; P = .180) and strongest for prostate cancer (RR = 3.29; P < .001). There was strong evidence of heterogeneity in the correlation between high BAP and risk of all SREs (P = .001), although the trend was similar in all patient groups.
First SRE
High Ntx levels showed a significant correlation with the risk of experiencing a first SRE on study for the overall population of patients (P < .001) and for each disease type except NSCLC and other solid tumors. Relative risk ratios for patients with high Ntx levels compared with low Ntx levels ranged from 1.56 for NSCLC (P = .306) to 3.01 for multiple myeloma (P = .008; Table 2). Moderate Ntx levels were also associated with a significantly increased risk of experiencing a first SRE on study for the overall patient population (P < .001) and for breast cancer, NSCLC, and other solid tumors. There was no suggestion of heterogeneity in the relative risk between groups by primary disease (P = .929).
High BAP was also associated with an increased risk of experiencing a first on-study SRE in the overall population (RR = 1.95, P < .001) and in all disease types except NSCLC and other solid tumors (Table 3). 17,19 There was significant heterogeneity in the correlation between high BAP and time to first SRE across disease types (P = .0007).
Progression of Bone Lesions
Overall, there was a 2.21-fold increased risk of bone lesion progression for patients with high Ntx versus low Ntx levels (P < .001) and a 1.57-fold increase for patients with moderate Ntx versus low Ntx levels (P < .001 for both comparisons; Table 2). There was consistent and strong evidence (P .001) that high Ntx was associated with an increased risk of disease progression for patients with solid tumors (breast cancer, prostate cancer, NSCLC, and other solid tumors), and there was a trend toward increased risk associated with high Ntx in patients with multiple myeloma (P = .080). However, there was considerable variation within the multiple myeloma group. There was also a trend toward increased risk of progression among patients with moderate Ntx levels in each of the disease types, although this reached statistical significance only for patients with breast cancer (P < .001; Fig 4) or prostate cancer (P = .015). There was no evidence of heterogeneity of effects across cancer types (P = .988).
In the overall patient population, high BAP was associated with a significantly increased risk of progression of bone lesions (P < .001). However, in contrast with the Ntx assessment, there was significant heterogeneity in this correlation (P = .043). Among patients with solid tumors, there was a significant 1.77- to 2.42-fold increased relative risk of bone lesion progression for patients with high BAP, whereas there was no significant relationship between BAP and bone lesion progression in patients with multiple myeloma (P = .881).
Death
In the overall population of patients and in each of the solid tumor disease types, high or moderate Ntx levels correlated with a significantly greater relative risk of dying on study (P < .001), whereas the correlation between Ntx level and death was not significant in patients with multiple myeloma (Table 2). This probably reflected the low frequency of elevated Ntx on treatment in the myeloma cohort (see Differential Response to Zoledronic Acid and Pamidronate section). Overall, high Ntx and moderate Ntx were associated with a 4.80-fold increased risk of death (95% CI, 3.90 to 5.91; P < .001) and a 3.11-fold increased risk of death (95% CI, 2.54 to 3.79; P < .001), respectively, compared with low Ntx levels.
The correlations between high BAP and relative risk of death during the assessment period are summarized in Table 3. There was a consistently significant association between high BAP and mortality, though the RRs showed considerable variation between different cancer types.
Differential Responses to Zoledronic Acid and Pamidronate
Figure 6 shows the proportion of patients with breast cancer or multiple myeloma treated with either zoledronic acid or pamidronate who had elevated Ntx levels during the course of the clinical trials. Figure 6A uses a cutoff of 100 nmol/mmol creatinine (high) for Ntx, and Figure 6B uses a cutoff of 50 to 99 nmol/mmol creatinine (moderate). Zoledronic acid was more effective than pamidronate at suppressing high rates of bone resorption, reducing Ntx levels from a high to a low state (RR = 2.73; 95% CI, 1.60 to 4.69), and was similarly effective at reducing Ntx levels from a moderate to a low level (Table 4). For multiple myeloma, the proportion of patients with high Ntx on study was very low with both treatments, possibly indicating maximal suppression of osteoclast activity by either bisphosphonate. In contrast, in the breast cancer subset, in whom zoledronic acid has demonstrated significant benefits beyond those of pamidronate,15 more patients had high Ntx levels at baseline compared with the multiple myeloma cohort, and a smaller proportion of patients treated with zoledronic acid had high Ntx levels on study compared with pamidronate.
DISCUSSION
The measurement of biochemical markers of bone metabolism has become more common in recent trials in patients with bone metastases. Assessment of these markers provides specific insight into rates of bone formation and resorption,2-4 and levels seem to correlate with bone pain levels and the burden of disease in the skeleton, while being unaffected by the burden of soft tissue or visceral disease.4-7,27 Decreases in marker levels have also been found to correlate with response to systemic anticancer treatments8,28,29 and symptomatic response to skeletal radiotherapy.30 However, all of these reports have been based on small cohorts of patients or are based on short-term (a few weeks) outcomes. As a result, the use of bone markers as prognostic tools during therapy has remained academic,31 pending robust information from large clinical trials.
Recently we reported the correlation between bone marker levels and clinical outcomes in patients with prostate cancer and a variety of other solid tumors (excluding breast cancer) who were receiving concomitant standard antineoplastic therapy.22 These patients were not treated with bisphosphonates, whereas in the exploratory cohort analysis reported herein, all patients received ongoing bisphosphonate treatment (zoledronic acid; some comparisons are also made between zoledronic acid and pamidronate in patients with multiple myeloma or breast cancer), consistent with current standard practice. The correlation between bone marker levels and clinical outcomes, including SREs, disease progression, and death, were analyzed and found to be statistically significant. These analyses used data collected during three large, multicenter, randomized, controlled, phase III trials of zoledronic acid.13-15 Characterizing patients by their most recent Ntx or BAP levels revealed clear relationships between bone marker levels and imminent risk of skeletal morbidity, bone lesion progression, and death. The risk of any SRE, time to first on-study SRE, disease progression, and death were all greater during periods of accelerated bone resorption (as indicated by high Ntx level) compared with periods of lower bone resorption activity (as indicated by low Ntx level).
The relationships between Ntx levels and outcome were broadly similar, irrespective of the underlying tumor type. Additionally, for the breast cancer and multiple myeloma patients, the prognostic significance of high Ntx levels was similar for patients treated with either zoledronic acid or pamidronate. However, because of the increased potency of zoledronic acid over pamidronate, the proportion of patients with high Ntx levels was always lower in the zoledronic acid group compared with pamidronate.32 Perhaps because relatively few of the patients with multiple myeloma had high Ntx at baseline and during the trial, both drugs were able to effectively reduce levels of Ntx in patients with multiple myeloma at more than 95% of the time points. The similarity in biochemical response reflects the similar risk of SREs among patients with multiple myeloma treated with zoledronic acid and pamidronate.15 In contrast, high Ntx was more common in patients with breast cancer, and zoledronic acid was more effective than pamidronate at reducing the proportion of patients with breast cancer in the high Ntx category. Perhaps as a result of this improved biochemical response, breast cancer patients treated with zoledronic acid had an additional 20% reduction in the risk of developing SREs compared with the pamidronate group (RR = 0.799; P = .025).15 Therefore, we believe that these results support the conclusion that the greater potency of zoledronic acid compared with pamidronate is responsible for its increased clinical activity in patients with breast cancer.
Based on these analyses, all patients with malignant bone disease are at risk for SREs, but if patients have high Ntx levels at baseline or on treatment, they are at a higher risk for negative clinical outcomes. Because bone marker levels change during the course of the disease, recent ( 6 months old) marker assessments have greater predictive value than the baseline assessment at later time points during the trial.22 For example, changes in the underlying systemic therapies were allowed during the course of these trials, and these modifications would, in themselves, be expected to influence bone marker levels. Patients who developed or maintained high Ntx or BAP levels during the course of zoledronic acid therapy were most at risk for deterioration in skeletal integrity, SREs, and a poor clinical outcome. Furthermore, because relative risks were calculated based on bone marker values taken up to 6 months before the events, increases in Ntx might precede clinical deterioration by several months. Thus, high Ntx levels could signal the need for more frequent follow-up. However, further studies may be necessary to determine whether the relationships can be generalized to patients treated with other bisphosphonates.
The results of this analysis and those previously published for the placebo arms of the clinical trials in patients with bone metastases from prostate cancer and patients with NSCLC or other solid tumors support the conclusion that recent Ntx levels significantly correlate with risk of negative clinical outcomes.22 Although the RRs for each negative clinical outcome varied depending on primary malignancy, the patterns of risk were consistent throughout each patient subset. Therefore, regardless of the primary malignancy, the level of bone resorption is closely correlated with malignant bone disease pathology, and the prognostic implications of Ntx assessments could be broadly applicable in the oncology setting.
Although these analyses cannot directly address a causal link between decreases in bone metabolism and clinical outcome, they do suggest that reducing bone turnover should have a positive effect by delaying progression of bone lesions and possibly improving survival. These observations will need to be confirmed in prospective randomized studies. We believe that these data support the hypothesis that bone marker measurements can provide meaningful additional data for assessing the risk of negative clinical outcomes in patients with malignant bone disease.
Authors' Disclosures of Potential Conflicts of Interest
Although all authors have completed the disclosure declaration, the following authors or their immediate family members have indicated a financial interest. No conflict exists for drugs or devices used in a study if they are not being evaluated as part of the investigation. For a detailed description of the disclosure categories, or for more information about ASCO’s conflict of interest policy, please refer to the Author Disclosure Declaration and the Disclosures of Potential Conflicts of Interest section in Information for Contributors.
NOTES
Presented in part at the 39th Annual Meeting of the American Society of Clinical Oncology, Chicago, IL, May 31-June 3, 2003.
Authors' disclosures of potential conflicts of interest are found at the end of this article.
REFERENCES
Mundy GR: Metastasis to bone: Causes, consequences and therapeutic opportunities. Nat Rev Cancer 2: 584-593, 2002
Coleman RE: The clinical use of bone resorption markers in patients with malignant bone disease. Cancer 94: 2521-2533, 2002
Coleman RE, Whitaker KB, Moss DW, et al: Biochemical prediction of response of bone metastases to treatment. Br J Cancer 58: 205-210, 1988
Berruti A, Dogliotti L, Gorzegno G, et al: Differential patterns of bone turnover in relation to bone pain and disease extent in bone in cancer patients with skeletal metastases. Clin Chem 45: 1240-1247, 1999
Vinholes J, Guo CY, Purohit OP, et al: Metabolic effects of pamidronate in patients with metastatic bone disease. Br J Cancer 73: 1089-1095, 1996
Lipton A, Demers L, Curley E, et al: Markers of bone resorption in patients treated with pamidronate. Eur J Cancer 34: 2021-2026, 1998
Vinholes JJ, Purohit OP, Abbey ME, et al: Relationships between biochemical and symptomatic response in a double-blind randomised trial of pamidronate for metastatic bone disease. Ann Oncol 8: 1243-1250, 1997
Brown JE, Thomson CS, Ellis SP, et al: Bone resorption predicts for skeletal complications in metastatic bone disease. Br J Cancer 89: 2031-2037, 2003
Glover D, Lipton A, Keller A, et al: Intravenous pamidronate disodium treatment of bone metastases in patients with breast cancer: A dose-seeking study. Cancer 74: 2949-2955, 1994
Berruti A, Dogliotti L, Bitossi R, et al: Incidence of skeletal complications in patients with bone metastatic prostate cancer and hormone refractory disease: Predictive role of bone resorption and formation markers evaluated at baseline. J Urol 164: 1248-1253, 2000
Samma S, Kagebayashi Y, Yasukawa M, et al: Sequential changes of urinary pyridinoline and deoxypyridinoline as markers of metastatic bone tumor in patients with prostate cancer: A preliminary study. Jpn J Clin Oncol 27: 26-30, 1997
Luftner D, Richter A, Geppert R, et al: Normalisation of biochemical markers of bone formation correlates with clinical benefit from therapy in metastatic breast cancer. Anticancer Res 23: 1017-1026, 2003
Saad F, Gleason DM, Murray R, et al: Long-term efficacy of zoledronic acid for the prevention of skeletal complications in patients with metastatic hormone-refractory prostate cancer. J Natl Cancer Inst 96: 879-882, 2004
Rosen LS, Gordon D, Tchekmedyian NS, et al: Long-term efficacy and safety of zoledronic acid in the treatment of skeletal metastases in patients with nonsmall cell lung carcinoma and other solid tumors: A randomized, phase III, double-blind, placebo-controlled trial. Cancer 100: 2613-2621, 2004
Rosen LS, Gordon D, Kaminski M, et al: Long-term efficacy and safety of zoledronic acid compared with pamidronate disodium in the treatment of skeletal complications in patients with advanced multiple myeloma or breast carcinoma: A randomized, double-blind, multicenter, comparative trial. Cancer 98: 1735-1744, 2003
Saad F, Gleason DM, Murray R, et al: A randomized, placebo-controlled trial of zoledronic acid in patients with hormone-refractory metastatic prostate carcinoma. For the Zoledronic Acid Prostate Cancer Study Group. J Natl Cancer Inst 94: 1458-1468, 2002
Andersen PK, Borgan O, Gill RD, et al: Statistical Models Based on Counting Processes. New York, NY, Springer-Verlag, 1993
Datta S, Satten GA: Validity of the Aalen-Johansen estimators of stage occupation probabilities and Nelson-Aalen estimators of integrated transition hazards for non-Markov models. Statistics & Probability Letters 55: 403-411, 2001
Kalbfleish JD, Prentice RL: The Statistical Analysis of Failure Time Data. Hoboken, NJ, John Wiley & Sons Inc, 2002
Lawless JF: Statistical Models and Methods for Lifetime Data. New York, NY, John Wiley & Sons Inc, 2002
Ross SM: Editor: Introduction to Probability Models. San Diego, CA, Academic Press Inc, 1989
Brown JE, Cook RJ, Major P, et al: Bone markers as predictors of skeletal complications in prostate cancer, lung cancer, and other solid tumors. J Natl Cancer Inst 97: 59-69, 2005
Demers LM, Costa L, Lipton A: Biochemical markers and skeletal metastases. Cancer 88: 2919-2926, 2000 (suppl)
Datta S, Satten GA: Estimation of integrated transition hazards and stage occupation probabilities for non-Markov systems under dependent censoring. Biometrics 58: 792-802, 2002
Aalen OO, Borgan O, Fekjaer H: Covariate adjustment of event histories estimated from Markov chains: The additive approach. Biometrics 57: 993-1001, 2001
Agresti A: Categorical data analysis. Honecker, NJ, John Wiley & Sons Inc, 2002
Demers LM, Costa L, Lipton A: Biochemical markers and skeletal metastases. Clin Orthop 415: S138-S147, 2003 (special issue)
Vinholes J, Coleman R, Lacombe D, et al: Assessment of bone response to systemic therapy in an EORTC trial: Preliminary experience with the use of collagen cross-link excretion. European Organization for Research and Treatment of Cancer. Br J Cancer 80: 221-228, 1999
Costa L, Demers LM, Gouveia-Oliveira A, et al: Prospective evaluation of the peptide-bound collagen type I cross-links N-telopeptideand C-telopeptide in predicting bone metastases status. J Clin Oncol 20: 850-856, 2002
Hoskin PJ, Stratford MR, Folkes LK, et al: Effect of local radiotherapy for bone pain on urinary markers of osteoclast activity. Lancet 355: 1428-1429, 2000
Hillner BE, Ingle JN, Chelbowski RT, et al: American Society of Clinical Oncology 2003 update on the role of bisphosphonates and bone health issues in women with breast cancer. Published erratum in: J Clin Oncol 2004;22:1351. J Clin Oncol 21: 4042-4057, 2003
Coleman RE, Cook R, Major P, et al: Why is zoledronic acid (Z) superior to pamidronate (P) for bone metastases from breast cancer (BC) but equivalent in multiple myeloma (MM)? Proc Am Soc Clin Oncol 22: 47, 2003 (abstr 187)(Robert E. Coleman, Pierre)