Normalization of bone markers is associated with improved survival in patients with bone metastases from solid tumors and elevated bone resorption receiving zoledronic acid

Authors

  • Allan Lipton MD,

    Corresponding author
    1. Department of Hematology-Oncology, Milton S. Hershey Medical Center, Pennsylvania State University, Hershey, Pennsylvania
    • Milton S. Hershey Medical Center, Pennsylvania State University, 500 University Dr., Hershey, PA 17033
    Search for more papers by this author
    • Allan Lipton is on the speaker's bureau for Norvartis. He also has served as a consultant for and given testimony on behalf of Novartis. Richard Cook receives honoraria from and has a consultant's agreement with Novartis. Fred Saad has served on advisory boards of and receives honoraria and research funds from Novartis. Pierre Major has no financial interests in Novartis. Patrick Garnero has received consultant fees from Novartis. Evangelos Terpos has received an honorarium for participation in an advisory board from Novartis. Janet Brown has participated in an advisory board for Novartis. Robert Coleman receives honoraria, consultancy payments, and research funds from Novartis. He has also given expert testimony on behalf of Novartis.

    • Fax: (717) 531-5076

  • Richard Cook PhD,

    1. Department of Statistics and Actuarial Science, University of Waterloo, Waterloo, Ontario, Canada
    Search for more papers by this author
  • Fred Saad MD,

    1. Department of Surgery/Urology, Central Hospital of the University of Montreal, Hospital Notre-Dame, Montreal, Quebec, Canada
    Search for more papers by this author
  • Pierre Major MD,

    1. Division of Medical Oncology, Juravinski Cancer Center, Hamilton, Ontario, Canada
    Search for more papers by this author
  • Patrick Garnero MD, PhD,

    1. INSERM Research Unit 664 and SYNARC, Lyon, France
    Search for more papers by this author
  • Evangelos Terpos MD, PhD,

    1. Department of Hematology and Medical Research, 251 General Air Force Hospital, Athens, Greece
    Search for more papers by this author
  • Janet E. Brown MD,

    1. Cancer Research UK Clinical Center in Leeds, University of Leeds, Leeds, UK
    Search for more papers by this author
  • Robert E. Coleman MD

    1. Academic Unit of Clinical Oncology, Weston Park Hospital, Cancer Research Center, University of Sheffield, Sheffield, UK
    Search for more papers by this author

Abstract

BACKGROUND.

For patients with bone metastases, high N-telopeptide of type I collagen (NTX) levels correlate with increased risks of skeletal-related events and death. However, the relation between NTX decreases and clinical benefits is unclear.

METHODS.

Correlations between NTX normalization during treatment and clinical outcome were retrospectively analyzed in 3 large, phase 3 trials. Urinary NTX levels were measured at baseline and at Month 3 in patients with bone metastases from breast cancer (BC; n = 578), hormone-refractory prostate cancer (HRPC; n = 472), or nonsmall-cell lung cancer and other solid tumors (NSCLC/OST; n = 291) who received zoledronic acid or control (pamidronate for BC; placebo for HRPC and NSCLC/OST) for up to 24 months. NTX levels were characterized as normal (N; <64 nmol/mmol creatinine) or elevated (E; ≥64 nmol/mmol creatinine).

RESULTS.

After 3 months of zoledronic acid, most N-group patients maintained normal levels; however, most E-group patients normalized their NTX levels (BC, 81%; HRPC, 70%; NSCLC/OST, 81%). In contrast, NTX levels normalized with pamidronate in 65% of BC, with placebo in 8% of HRPC, and in 17% of NSCLC/OST E-group patients. Normalized NTX correlated with improved overall survival versus persistently elevated NTX (significant for zoledronic acid-treated patients; trend for placebo-treated patients). Moreover, percentage reductions from baseline NTX levels correlated with benefits regardless of whether patients transitioned from E to N.

CONCLUSIONS.

Zoledronic acid normalizes or maintains normal NTX levels in most patients with bone metastases. Normalized NTX within 3 months of treatment, versus persistently elevated NTX, was associated with reduced risks of skeletal complications and death. Cancer 2008. © 2008 American Cancer Society.

Skeletal metastases are common in patients with solid tumors, affecting approximately 75% of patients with advanced breast cancer (BC) or hormone-refractory prostate cancer (HRPC) and approximately 40% of patients with advanced nonsmall-cell lung cancer (NSCLC).1 Median survival after the diagnosis of bone metastases from BC or HRPC is approximately 2 to 3 years.2 In contrast, median survival in advanced NSCLC is less than 1 year.3, 4 However, novel anticancer therapies are increasing survival in these settings. Supportive therapies to prevent pain and skeletal-related events (SREs) from bone metastases may help to preserve quality of life (QOL) throughout each patient's lifetime.

For many patients, bone metastases are diagnosed only after they become symptomatic, typically with bone pain. Malignant bone disease is also associated with potentially debilitating SREs including pathologic fracture, spinal cord compression, the need for palliative radiotherapy or surgery to bone, and hypercalcemia of malignancy (HCM).5 Without bone-directed therapy, patients with bone lesions experience, on average, 2 to 4 SREs annually.5 After an SRE, QOL typically decreases and subsequent SRE risk increases.6–8 Moreover, pathologic fractures are associated with reduced survival.9, 10

Malignant bone disease is associated with increased levels of bone resorption markers, especially N-telopeptide of type I collagen (NTX), which correlates with both the presence and extent of bone metastases.11, 12 Patients with bone metastases and elevated NTX levels have significantly increased risks of SREs, bone lesion progression, and death versus patients with low NTX levels (P ≤ .011 for each).13 Bisphosphonates inhibit osteoclast-mediated osteolysis and reduce bone marker levels.14, 15 Urinary NTX normalization with pamidronate has been associated with significant palliative responses (P = .03),16 significant delays in bone disease progression (P = .03),17 and a trend toward fewer fractures (P = .07),17 and NTX decreases with zoledronic acid have been significantly associated with longer time to bone lesion progression (P = .001).18 Therefore, NTX may have utility for monitoring therapeutic response in bone.

Among the bisphosphonates approved for malignant bone disease, only zoledronic acid has received international approval for all solid tumors.19 Although its phase 3 trials were not designed to assess survival, trends toward improved survival with 4 mg zoledronic acid versus placebo were detected. In HRPC patients, zoledronic acid increased survival by approximately 2.5 months versus placebo (P = .103).20 A similar trend was observed in the subset of patients with renal cell carcinoma in the NSCLC/other solid tumors (OST) trial (P = .179).21 Zoledronic acid decreased NTX levels in all these subsets. Therefore, the objective of this study was to assess whether normalization of elevated NTX levels is associated with improved event-free and overall survival in patients with bone metastases receiving standard systemic therapy with or without zoledronic acid.

MATERIALS AND METHODS

Patients and Study Design

This was an exploratory analysis correlating NTX patterns and outcomes in 3 separate randomized, double-blind, international, phase 3 trials of zoledronic acidfor the treatment of bone metastases.22–27 The trials had a parallel design: patients were randomized to 4-mg or 8-mg zoledronic acid or control (placebo or 90-mg pamidronate) for up to 24 months. The 8-mg zoledronic acid dose was reduced to 4 mg during the trials to ensure renal safety and was referred to as the 8/4-mg group in the primary publications.22, 25, 26 All patients provided written informed consent.

Only patients with bone metastases from BC (n = 1130) were included from the trial that compared zoledronic acid with pamidronate in BC and multiple myeloma,22, 23 whereas all patients in the HRPC (N = 643) and NSCLC/OST (N = 773) trials were evaluated.24–27 The current analyses included NTX data for baseline and 3 months only and excluded patients missing either assessment.

Study Assessments

The primary endpoint of each trial was the proportion of patients who developed an SRE (not including HCM). Secondary endpoints included HCM as an SRE and evaluated the time to first SRE, bone-disease progression, and death. Radionuclide bone scans and surveys were performed quarterly. Bone-disease progression was defined as the appearance of a new bone metastasis or progression of existing bone metastases. Urinary NTX levels were assessed and corrected for creatinine levels at international central laboratories.13, 28

Statistical Methods

Urinary NTX levels were stratified as normal (<64 nmol/mmol creatinine) or elevated (≥64 nmol/mmol creatinine) based on the approximate upper limit of normal for disease-free postmenopausal women; most enrolled patients were elderly.29 Clinical outcomes and NTX response to treatment were similar in the 4- and 8/4-mg zoledronic acid arms; thus, they were pooled to increase the sample size and improve statistical power.22–27 Patients were grouped by their baseline and 3-month NTX levels: normal baseline NTX (N) group; elevated baseline NTX (E) group; elevated baseline and normalized 3-month NTX (E-N) group; elevated baseline and 3-month NTX (E-E) group.

Cox regression models were used to calculate relative risks of clinical events between NTX groups and risk of persistently elevated NTX during bisphosphonate therapy. Cumulative incidence functions were used to estimate proportions of patients with ≥1 SRE on study. Covariates associated with normalization of urinary NTX were assessed via logistic regression models with backward elimination at the .05 significance level to generate reduced models. The survival-time distribution was estimated by Kaplan-Meier methodology. For each population studied the effect of 3-month NTX changes from baseline was assessed as a continuous variable (percentage reduction = 100 × [baseline NTX – 3-month NTX] ÷ baseline NTX) and related to survival time for the 25th percentile, median, and 75th percentile baseline NTX level. All statistical tests were 2-sided (alpha = .05 significance level).

RESULTS

Demographics and Baseline Characteristics

Baseline and 3-month NTX data were available for 379 patients with BC, 314 with HRPC, and 204 with NSCLC/OST who were treated with zoledronic acid. Demographics and baseline disease characteristics are shown in Table 1. The E group tended to have poorer performance status and higher mean Brief Pain Inventory (BPI) scores than the N group. Baseline and 3-month NTX data were also available for 199 patients with BC treated with pamidronate and 158 patients with HRPC and 87 with NSCLC/OST treated with placebo. Demographics and baseline disease characteristics were similar to those of the respective zoledronic acid groups (data not shown). Among patients with bone metastases from BC treated with zoledronic acid, 1 or more metastases in the lung, liver, and other sites (excluding brain) were present in approximately 20% of patients each (Table 2). Less than 5% of patients with bone metastases from HRPC had distant visceral metastases; however, approximately 20% had lymph node metastases. Regardless of primary tumor, less than 1% of patients had brain metastases. In patients with bone metastases from NSCLC/OST, sites of metastases were not reported.

Table 1. Patient Demographics and Baseline Disease Characteristics for the Zoledronic Acid Treatment Arms by Tumor Type
 BC n = 379HRPC n = 314NSCLC/OST n = 204
 Normal NTX* n = 159Elevated NTX* n = 220Normal NTX* n = 121Elevated NTX* n = 193Normal NTX* n = 117Elevated NTX* n = 87
  • BC indicates breast cancer; HRPC, hormone-refractory prostate cancer; NSCLC/OST, nonsmall-cell lung cancer and other solid tumors; NTX, N-telopeptide of type I collagen; SE, standard error; SRE, skeletal-related event; FACT-G, Functional Assessment of Cancer Therapy–General; BPI, Brief Pain Inventory.

  • *

    Normal NTX <64 nmol/mmol creatinine; elevated NTX ≥64 nmol/mmol creatinine.

Mean age, y ± SE58 ± 11.159 ± 12.771 ± 6.971 ± 8.563 ± 9.861 ± 11.2
Sex, no. (%)      
 Female157 (98.7)220 (100.0)0037 (31.6)35 (40.2)
 Male2 (1.3)0121 (100. 0)193 (100.0)80 (68.4)52 (59.8)
Performance score, no. (%)      
 Fully active78 (49.1)61 (28.0)69 (57.5)85 (44.0)36 (31.3)19 (21.8)
 Some restriction81 (50.9)157 (72.0)51 (42.5)108 (56.0)79 (68.7)68 (78.2)
Predominantly blastic, no. (%)      
 No89 (56.0)135 (61.3)25 (20.7)20 (10.4)81 (69.2)55 (63.2)
 Yes70 (44.0)85 (38.6)96 (79.3)173 (89.6)36 (30.8)32 (36.8)
Prior SRE, no. (%)      
 No82 (51.6)70 (31.8)90 (74.4)129 (66.8)35 (29.9)33 (37.9)
 Yes77 (48.4)150 (68.2)31 (25.6)64 (33.2)82 (70.1)54 (62.1)
FACT-G, score ± SE80 ± 14.874 ± 17.283 ± 13.683 ± 13.677 ± 14.568 ± 16.0
Baseline mean NTX, nmol/mmol creatinine ± SE41.9 ± 14.2203.6 ± 691.443.6 ± 13.8207.7 ± 227.041.2 ± 14.4151.3 ± 235.6
Baseline mean BPI, score ± SE2.7 ± 2.23.3 ± 2.21.9 ± 2.02.4 ± 2.03.1 ± 2.03.4 ± 1.9
Table 2. Stratification of Patients With Breast Cancer by Metastasis Site (All Patients Had Bone Metastases)
Site of metastasisTreatmentBaseline NTXNTX transitions at 3 months on study
No. of patients (% of total)E patients (% of total)E-N patients (% of total)E-E patients (% of total)Normalized, %
  1. NTX indicates N-telopeptide of type I collagen; N, Normal NTX levels (<64 nmol/mmol creatinine); E, Elevated NTX levels (≥64 nmol/mmol creatinine).

BonePamidronate108 (100.0)145 (100.0)73 (100.0)40 (100.0)64.6
 Zoledronic acid195 (100.0)292 (100.0)180 (100.0)40 (100.0)81.8
LungPamidronate31 (28.7)28 (19.3)11 (15.1)6 (15.0)64.7
 Zoledronic acid42 (21.5)66 (22.6)48 (26.7)5 (12.5)90.6
LiverPamidronate27 (25.0)37 (25.5)15 (20.5)11 (27.5)57.7
 Zoledronic acid39 (20.0)63 (21.6)38 (21.1)5 (12.5)88.4
BrainPamidronate5 (4.6)4 (2.8)3 (4.1)0100
 Zoledronic acid5 (2.6)4 (1.4)1 (0.6)1 (2.5)50.0
OtherPamidronate23 (21.3)29 (20.0)14 (19.2)8 (20.0)63.6
 Zoledronic acid38 (19.5)61 (20.9)38 (21.1)8 (20.0)82.6

Baseline NTX Change in Response to Zoledronic Acid

At baseline, 220 (58%) BC, 193 (61%) HRPC, and 87 (43%) NSCLC/OST patients randomized to zoledronic acid had elevated NTX levels. Among the E groups, zoledronic acid normalized NTX levels (E-N group) within 3 months in 179 (81%) patients with BC, 135 (70%) with HRPC, and 70 (81%) with NSCLC/OST (Table 3). Thirty-seven (17%) patients with BC, 56 (29%) with HRPC, and 11 (13%) with NSCLC/OST in the E groups had persistently elevated NTX after 3 months of zoledronic acid (E-E group). Normalization was observed in greater than 80% of patients with BC regardless of metastasis site except for brain, but there were too few patients with brain metastases for statistical comparisons (Table 2). Among the E groups in the control arms, pamidronate normalized NTX levels (E-N group) within 3 months in 74 (65%) patients with BC, and NTX levels normalized in only 8 (8%) of HRPC and 7 (17%) of NSCLC/OST patients in the E groups who received placebo. (Note that patient percentages do not add up to 100% because of deaths during the assessment period.)

Table 3. NTX Transitions at 3 Months on Study
 Baseline NTX*No. of patientsAliveDied within 111 days
Normal NTX,* no. (%)Elevated NTX,* no. (%)All patients no. (%)
  • NTX indicates N-telopeptide of type I collagen; HRPC, hormone-refractory prostate cancer; NSCLC/OST, nonsmall-cell lung cancer and other solid tumors.

  • *

    Normal NTX <64 nmol/mmol creatinine; Elevated NTX ≥64 nmol/mmol creatinine.

  • Only trial with pamidronate as comparator.

Zoledronic acid    
Breast cancerNormal159155 (97.5)3 (1.9)1 (0.6)
Elevated220179 (81.4)37 (16.8)4 (1.8)
HRPCNormal121117 (96.7)3 (2.5)1 (0.8)
Elevated193135 (69.9)56 (29.0)2 (1.0)
NSCLC/OSTNormal117110 (94.0)3 (2.6)4 (3.4)
Elevated8770 (80.5)11 (12.6)6 (6.9)
Pamidronate or placebo     
Breast cancerNormal8583 (97.6)2 (2.4)0
Elevated11474 (64.9)38 (33.3)2 (1.8)
HRPCNormal5939 (66.1)20 (33.9)0
Elevated998 (8.1)88 (88.9)3 (3.0)
NSCLC/OSTNormal4635 (76.1)9 (19.6)2 (4.3)
Elevated417 (17.1)32 (78.0)2 (4.9)

In the N groups, approximately 96% of zoledronic acid-treated patients maintained normal NTX at 3 months. In contrast, NTX elevations were approximately 10-fold more common in N-group patients treated with placebo versus zoledronic acid.

Baseline Characteristics Associated With Persistently Elevated NTX Levels

For the E groups, in both univariate and multivariate analyses, baseline NTX >150 nmol/mmol creatinine was associated with increased odds of persistently elevated NTX levels after 3 months of bisphosphonate therapy for all tumors analyzed (P < .01; Table 4). Among BC patients, lower baseline BPI scores (P= .0213) in the zoledronic acid arm, higher Functional Assessment of Cancer Therapy-General (FACT-G) total score (P = .0280), and osteoblastic versus osteolytic metastases (P = .0344) in the pamidronate arm were associated with reduced odds of persistently elevated NTX. No other baseline variables showed significant correlation with odds of NTX normalization with 3 months of bisphosphonate therapy.

Table 4. Baseline Disease Characteristics Associated With Persistently Elevated 3-Month NTX Levels in Patients Treated With Zoledronic Acid
 Odds ratio (P)
 BC N = 578HRPC n = 314NSCLC/OST n = 204
 PamidronateZoledronic acidZoledronic acidZoledronic acid
 Univariate analysisReduced multivariate modelUnivariate analysisReduced multivariate modelUnivariate analysisReduced multivariate modelUnivariate analysisReduced multivariate model
  • NTX indicates N-telopeptide of type I collagen; BC, breast cancer; HRPC, hormone-refractory prostate cancer; NSCLC/OST, nonsmall-cell lung cancer and other solid tumors; BPI, Brief Pain Inventory; FACT-G, Functional Assessment of Cancer Therapy–General; SREs, skeletal-related events.

  • *

    Continuous baseline variables.

  • Some restrictions versus fully active in physical activity.

  • Odds ratio is undefined because of small patient numbers.

  • §

    Yes versus no.

  • ||

    Mild/strong analgesics versus none/tranquilizers.

Baseline NTX >150 nmol/mmol creatinine6.58 (<.0001)8.52 (<.0001)4.80 (<.0001)4.73 (<.0001)14.13 (<.0001)14.13 (<.0001)10.97 (.001)10.97 (.001)
BPI score*1.16 (.137)1.22 (.013)1.22 (.021)1.29 (.005)1.02 (.931)
FACT-G total score*0.97 (.018)0.96 (.028)0.97 (.004)0.99 (.405)0.99 (.526)
Age*0.98 (.196)0.99 (.534)1.02 (.415)1.01 (.741)
Baseline performance status1.46 (.397)1.75 (.195)1.04 (.913)
Weight*0.98 (.062)0.99 (.520)0.99 (.275)1.01 (.800)
Baseline serum creatinine*0.32 (.307)0.59 (.572)0.83 (.695)1.17 (.891)
Cancer duration*0.97 (.347)0.95 (.172)0.94 (.264)0.62 (.210)
Time from initial cancer to bone metastases*0.98 (.608)0.97 (.352)0.93 (.169)0.38 (.162)
Antineoplastic therapy§0.57 (.171)1.78 (.108)
History of SREs§2.18 (.088)1.02 (.968)2.16 (.028)0.45 (.226)
Versus lytic metastases        
 Blastic0.41 (.057)0.31 (.034)0.72 (.426)1.09 (.918)2.00 (.364)
 Other0.93 (.895)1.41 (.619)0.96 (.924)0.56 (.608)0.70 (.705)
Versus white race        
 black/Asian/other3.31 (.051)1.67 (.289)1.83 (.166)
Analgesic score||2.91 (.091)2.01 (.076)

NTX Normalization and Clinical Outcome

For the study populations analyzed, there were no statistically significant differences in survival between patients with BC or NSCLC/OST who had baseline bone marker data and the overall patient population in the 2 trials. However, survival between patients with HRPC who had baseline bone marker data and the overall patient population was significantly better (P = .0241) for reasons unknown. In the E groups the survival curves for the patients whose NTX levels normalized (E-N groups) appeared closer to those of the N groups, indicating a possible transition in risk status for death (Fig. 1). Patients in the E-N groups had significantly reduced risks of death of 48% for BC (relative risk [RR] = 0.52; 95% confidence interval [CI] = 0.34, 0.78; P = .0017), 59% for HRPC (RR = 0.41; 95% CI = 0.29, 0.59; P < .0001), and 57% for NSCLC/OST (RR = 0.43; 95% CI = 0.22, 0.83; P = .0116) compared with respective E-E groups. Although far fewer of the placebo-treated E-group patients had normalized NTX at 3 months, the correlations between NTX normalization and improved survival were statistically significant in placebo-treated HRPC patients (n = 99; P = .0286) and reflected a strong trend in pamidronate-treated BC patients (n = 114; P = .0554). There was no association between NTX normalization and improved survival in placebo-treated NSCLC/OST patients (n = 41; P = .5643); however, patient numbers were small.

Figure 1.

Kaplan-Meier survival estimates stratified by baseline and 3-month N-telopeptide of type I collagen (NTX) levels. (a) Patients with bone metastases from breast cancer treated with zoledronic acid. (b) Patients with bone metastases from breast cancer treated with pamidronate. (c) Patients with bone metastases from hormone-refractory prostate cancer treated with zoledronic acid. (d) Patients with bone metastases from hormone-refractory prostate cancer treated with placebo. (e) Patients with bone metastases from nonsmall-cell lung cancer and other solid tumors treated with zoledronic acid. (f) Patients with bone metastases from nonsmall-cell lung cancer and other solid tumors treated with placebo. E-E indicates patients with elevated baseline and 3-month NTX; E-N, patients with elevated baseline and normalized 3-month NTX; N, patients with normal baseline NTX.

Similar associations were observed for SRE-free and bone lesion progression-free survival. Patients whose NTX normalized during zoledronic acid treatment had an approximately 50% improved SRE-free survival compared with the respective E-E groups. Relative increases in SRE-free survival were 51% for BC (P = .0004), 49% for HRPC (P = .0009), and 45% for NSCLC/OST (P = .1309) for E-N versus E-E groups. Among patients with NSCLC/OST, NTX normalization was also associated with a significant 61% relative increase in bone lesion progression-free survival (P = .0227). Relative increases in bone lesion progression-free survival in BC or HRPC did not achieve statistical significance (26%; P = .1824; and 15%; P = .4102, respectively).

Relative Risks of Clinical Events After Zoledronic Acid-Mediated Normalization of NTX

Normalization of NTX was associated with a ≥40% reduction in risk of first SRE compared with the E-E group. The risk of a first SRE for the E-N group was reduced by 49% for BC (RR = 0.51; 95% CI = 0.33, 0.78; P = .0020), 38% for HRPC (RR = 0.62; 95% CI = 0.39, 0.98; P = .0411), and 39% for NSCLC/OST (RR = 0.61; 95% CI = 0.25, 1.49; P = .2803) versus the E-E group. In BC the risk of first fracture or surgery to bone was also significantly reduced by 53% (RR = 0.473; 95% CI = 0.30, 0.76; P = .0018), and the rate of first fractures had nonsignificant 29% to 49% decreases for HRPC and NSCLC/OST patients in the E-N versus E-E groups (P = .3653 and P = .2986, respectively).

Correlations Between Clinical Outcome and Percentage Reduction in NTX

Both patient groups (baseline N and E) were pooled for each cancer type and percentage reduction from baseline NTX at 3 months was compared with the RR of death (Fig. 2). To analyze correlations in this model, data were graphed for 3 hypothetical patients with the following baseline NTX values (in nmol/mmol creatinine): the lowest quartile value (BC, 47; HRPC, 52; NSCLC/OST, 38), the median value (BC, 76; HRPC, 80; NSCLC/OST, 59), and the highest quartile value (BC, 125; HRPC, 153; NSCLC/OST, 87). Among all patients with BC a 40% decrease in NTX at 3 months resulted in significant reductions in risk of death, ranging from 11% to 17% for these 3 scenarios (homogeneity test: P = .1835). Among HRPC patients, a 40% decrease in NTX at 3 months resulted in significant approximate 17% reductions in risk of death regardless of baseline NTX levels (homogeneity test: P = .8415). Among all NSCLC/OST patients a 40% decrease in NTX at 3 months significantly reduced risk of death by approximately 22% regardless of baseline NTX values (homogeneity test: P = .8758).

Figure 2.

Relative risk of death stratified by N-telopeptide of type I collagen (NTX) decrease. (a) Breast cancer. (b) Hormone-refractory prostate cancer. (c) Nonsmall-cell lung cancer and other solid tumors. *([Baseline NTX – 3-month NTX] ÷ baseline NTX) × 100. Top x-axis represents patients with NTX change ≤ the corresponding value on bottom x-axis.

Validity tests of the proportional hazards assumptions found no model violations.

DISCUSSION

Metastatic bone disease disturbs bone metabolism, which can be monitored through biochemical markers. For example, NTX is released during osteolysis, can be measured in serum or urine, and may provide insight into ongoing interactions between tumor and bone. Recently, elevated NTX levels have been associated with increased risks of SREs, disease progression, and death compared with normal NTX levels in patients with malignant bone disease.13

In this analysis, approximately 50% of patients had elevated NTX at baseline and had significantly reduced survival compared with patients with normal baseline NTX. Among patients with elevated baseline NTX, 75% normalized NTX levels within 3 months of zoledronic acid treatment. In contrast, only 8% (HRPC) to 17% (NSCLC/OST) of placebo-treated patients achieved NTX normalization. Survival curves in the E-N groups approached those for the N groups. The E-E groups had significantly shorter survival compared with the E-N groups, consistent with earlier observations.13, 15 In the zoledronic acid treatment arms, compared with the E-E groups, median survival in the E-N groups was longer by >9 months for BC, approximately 12 months for HRPC, and approximately 4 months for NSCLC/OST. Median survival was also longer in the E-N groups of patients treated with pamidronate or placebo compared with the E-E groups for BC and HRPC. A survival benefit was not observed in the E-N group of placebo-treated NSCLC/OST patients compared with the E-E group, although at the median survival timepoint of 6 months there was a separation of the survival curves. Moreover, the majority of analyzed patients were receiving concomitant cytotoxic or hormonal chemotherapy; therefore, the survival benefits associated with NTX normalization were in addition to anticancer treatment benefits.

Reductions in NTX levels provided a continuum of SRE risk reduction and survival benefits regardless of baseline NTX levels. Consistent with earlier findings in BC patients,30 a 40% NTX reduction within 3 months of zoledronic acid treatment in these analyses generally achieved the same magnitude of clinical benefits regardless of baseline NTX level in patients with any solid tumor. Monitoring NTX changes may provide insight into therapy response regardless of baseline NTX levels.

The survival benefit in the NTX transition group could stem from multiple factors. Zoledronic acid significantly delays the time to onset of potentially life-limiting SREs. Normalization of NTX corresponded with a 50% reduction in overall risk of SREs, although no significant decrease in SRE risk for E-N versus E-E was detected for the NSCLC subset. Radiotherapy to bone and pathologic fractures were the most common SREs.23, 24, 26 Pathologic fractures have been associated with a higher risk of death in BC (32% increase; P = .0031) and HRPC (23% increase; P > .05).9

Another explanation for the observed survival advantage is zoledronic acid-mediated antitumor effects, which have been detected in preclinical models.31 In patients with bone metastases (N = 26), Santini et al.,32 reported that 1-mg zoledronic acid weekly for 4 weeks then every 4 weeks thereafter lowered serum levels of vascular endothelial growth factor, consistent with antiangiogenic effects. In a pilot study (N = 40), results suggested that zoledronic acid delayed the development of bone metastases in patients with high-risk tumors compared with no treatment (P = .0002 at 18 months).33 This could be mediated by direct effects or by blocking tumor-induced osteolysis, thereby preventing the release of growth factors from bone.34 Ongoing trials are assessing whether zoledronic acid can delay metastasis to bone in patients with high-risk cancers (eg, does Adjuvant Zoledronic acid redUce REcurrence [AZURE] in BC, the Zometa EUropean Study [ZEUS] in PC, and G2914 in NSCLC).

A small percentage of patients treated with zoledronic acid had elevated NTX levels that remained persistently elevated (E-E group). These patients had a poorer prognosis than patients in the N or E-N groups and may have had more aggressive disease or osteoclast-independent osteolysis. Indeed, poor prognostic factors such as severely elevated NTX (>150 nmol/mmol creatinine), poor FACT-G scores, and higher BPI scores each correlated with significantly increased odds of persistent NTX elevation despite bisphosphonate therapy in some or all groups. Alternative zoledronic acid schedules or combination therapies may be necessary for these patients.32 Early treatment before severely increased bone metabolism and symptomatology develop could optimize therapy.35, 36

Together, these analyses suggest that breaking the cycle of bone destruction and tumor growth in bone with bisphosphonates can produce profound benefits. Indeed, in this exploratory analysis normalization of elevated baseline NTX within 3 months was associated with significant improvements in survival compared with persistent NTX elevation, and zoledronic acid normalized NTX in the majority of treated patients.

Acknowledgements

We thank Tamalette Loh, PhD, ProEd Communications, for medical editorial assistance with the article.

Ancillary