William K. Oh, Kevin Proctor, Mari Nakabayashi, Carolyn Evan, Lauren K. Tormey, and Timothy Daskivich are employees of Dana-Farber Cancer Institute, which has received research funds from Analysis Group, Inc.
The risk of renal impairment in hormone-refractory prostate cancer patients with bone metastases treated with zoledronic acid†
Article first published online: 20 FEB 2007
Copyright © 2007 American Cancer Society
Volume 109, Issue 6, pages 1090–1096, 15 March 2007
How to Cite
Oh, W. K., Proctor, K., Nakabayashi, M., Evan, C., Tormey, L. K., Daskivich, T., Antràs, L., Smith, M., Neary, M. P. and Duh, M. S. (2007), The risk of renal impairment in hormone-refractory prostate cancer patients with bone metastases treated with zoledronic acid. Cancer, 109: 1090–1096. doi: 10.1002/cncr.22504
Sponsored by F. Hoffmann-La Roche, Ltd., Basel, Switzerland.
- Issue published online: 7 MAR 2007
- Article first published online: 20 FEB 2007
- Manuscript Accepted: 6 DEC 2006
- Manuscript Revised: 23 NOV 2006
- Manuscript Received: 19 SEP 2006
- renal impairment;
- hormone-refractory prostate cancer;
- zoledronic acid;
- risk factors
Bisphosphonates have been used to treat bone metastases in hormone-refractory prostate cancer (HRPC), but certain agents have been associated with renal toxicity. For this observational study, the authors assessed the risk of renal impairment in patients with HRPC who received zoledronic acid from December 1999 to April 2005.
A comprehensive medical records review was performed in a major tertiary oncology center (n = 122 patients). The primary outcome of renal impairment was defined as an increase ≥0.5 mg/dL or ≥1.0 mg/dL over baseline creatinine value if the baseline value was <1.4 mg/dL or ≥1.4 mg/dL, respectively. A risk factor analysis was conducted using the Andersen-Gill extension to the Cox proportional hazards model.
Renal impairment was observed in 23.8% of patients. The risk of renal impairment increased with an extended duration of zoledronic acid therapy (<6 months, 11.1%; ≥24 months, 26.3%) and previous pamidronate treatment (45.5% vs 19.0% for patients with no prior pamidronate). A significantly greater risk of renal impairment was associated with increasing age at zoledronic acid initiation, prior pamidronate use, and a history of renal disease, hypertension, or smoking (P ≤ 0.05).
In an outpatient clinic setting, the risk of renal impairment among patients with HRPC who received zoledronic acid was greater than the risk reported previously in clinical trials. Cancer 2007 © 2007 American Cancer Society.
More than 80% of patients with hormone-refractory prostate cancer (HRPC) eventually develop metastatic bone disease.1 Serious skeletal-related complications from bone metastases, such as bone pain and pathologic fractures,2 are common and cause significant morbidity and mortality.3
Intravenous (IV) zoledronic acid is a bisphosphonate that has demonstrated efficacy in patients with HRPC who have evidence of bone metastases.4, 5 However, treatment with zoledronic acid is associated with a risk of nephrotoxicity, which has resulted in recent changes to its prescribing information. In the United States and the European Union, the product label for zoledronic acid was updated to include additional warnings of nephrotoxicity and restrictions for patients with varying degrees of renal impairment.6, 7 The recommended dosage of zoledronic acid for bone metastases is 4 mg IV over at least 15 minutes repeated every 3 or 4 weeks.6 Nevertheless, because of renal toxicity issues, the dose of zoledronic acid should be adjusted in patients with mild or moderate renal impairment (patients with a creatinine clearance ≥30 mL per minute). The standard 4-mg dose should be given only to patients with a creatinine clearance >60 mL per minute. The patients who are at greatest risk from renal deterioration, as may be expected, include those with advancing or progressive cancer, those who are receiving other potentially nephrotoxic drugs (eg, chemotherapy), and those with renal-related comorbidities (ie, chronic renal failure, hypertension, or diabetes).8 Several studies have suggested that the effect of zoledronic acid on renal deterioration may be greater among elderly patients,9 patients who received pamidronate prior to zoledronic acid,10, 11 and patients with prior renal impairment,12 but none of those studies focused on the HRPC population.
In the pivotal Phase III trial of IV zoledronic acid in 643 patients with HRPC and bone metastases, renal function deterioration occurred in 15.2% of patients who received 4 mg zoledronic acid and in 11.5% of patients who received placebo every 3 weeks for 15 months.4 Kaplan-Meier estimates of the risk of renal function deterioration revealed a relative risk ratio of 1.07 (zoledronic acid 4 mg vs placebo; P = .882). In a 2-year update of that trial, renal function deterioration occurred in 16 of 92 patients (17.4%) who received zoledronic acid and in 10 of 78 patients (12.8%) who received placebo.6 In contrast, a small observational study by Johnson and colleagues suggested that there may be a higher incidence of renal deterioration than reported in clinical trials with zoledronic acid; in that study, the incidence of renal toxicity associated with zoledronic acid was 27.0% in 15 patients with prostate cancer.13
The current analysis is the first observational study to our knowledge that quantifies the rate of renal impairment and identifies the associated risk factors in patients with HRPC who received zoledronic acid for bone metastases in an active, tertiary care clinical practice.
MATERIALS AND METHODS
A comprehensive medical records review was performed using a prospective electronic medical database implemented at the Dana-Farber Cancer Institute (DFCI) and the Brigham and Women's Hospital (BWH) (both located in Boston). The Prostate Cancer Clinical Research Information System was used to collect, store, and access comprehensive clinical, treatment, and outcomes data on patients with prostate cancer who were treated at the DFCI and BWH. Legacy information systems also were queried to determine dispensed drugs in the DFCI outpatient pharmacy, laboratory data across 2 hospitals, and registration data. Information on comorbid diagnoses and oral or outpatient medications was not stored in the system and required a medical records review. Results of creatinine tests that were conducted outside of either center were obtained through patients' community physicians and laboratories. All patients provided written informed consent to participate in this clinical research study. This retrospective study was approved by the DFCI Institutional Review Board.
This was a single-arm, observational follow-up study of patients who had HRPC with bone metastases and were treated with zoledronic acid at the DFCI. Patients were included in the study if they were aged ≥18 years, if they were treated actively at the DFCI, if they had HRPC with bone metastases, if they received ≥1 zoledronic acid infusion in the period from December 1999 to April 2005, and if they had ≥1 creatinine reading before and after the first zoledronic acid infusion. The observation period began on the date of the first zoledronic acid infusion and ended on the last center visit date, the last creatinine test date, or the defined study end date (April 30, 2005), whichever occurred later. The treatment period spanned from the first infusion to the last infusion of zoledronic acid.
The primary endpoint of renal impairment was defined as the proportion of patients with any of the following abnormal serum creatinine elevations: an increase ≥0.5 mg/dL over the baseline serum creatinine value (defined as the last creatinine serum test prior to beginning zoledronic acid treatment) if the baseline value was <1.4 mg/dL or an increase ≥1.0 mg/dL over the baseline serum creatinine value if the baseline value was ≥1.4 mg/dL. Two secondary endpoints of renal impairment were used. The first was defined based on creatinine clearance and was calculated by using the Cockcroft-Gault formula,14 which converts the serum creatinine level to creatinine clearance by factoring in patient sex, age, and weight. Creatinine clearance may be a more sensitive measure than serum creatinine level, because it considers patient-specific factors. Renal impairment was defined as a decrease ≥25% in creatinine clearance from the baseline value. For the secondary endpoint of renal impairment, patients were graded according to serum creatinine levels: grade 1, from the upper limit of normal to 1.5 times the baseline serum creatinine value; grade 2, from 1.5 to 3 times the baseline serum creatinine value; grade 3, from 3.1 to 6 times the baseline serum creatinine value; and grade 4, >6 times the baseline serum creatinine value. If the patient experienced different grades over the course of observation, then the most severe grade was designated for the patient.
The following covariates were considered in the risk-factor analysis: demographics (age and race), comorbidities (eg, history of hypertension, congestive heart failure, diabetes, prior renal disease, hyperlipidemia, and recent surgery), medications associated with acute renal failure15 (eg, the use of nonsteroidal anti-inflammatory agents, antihypertensives, and anticoagulants), treatment with chemotherapy and second-line hormone therapy, and prior use of other bisphosphonates.
The risk of renal toxicity was calculated as the proportion of patients who experienced renal impairment. The risk-factor analysis was conducted using the Cox proportional hazards model for time-related binary outcome data and the Andersen-Gill multiple-event analysis,16 which is an extension to the Cox model. The Andersen-Gill regression models the recurrence of renal impairment outcomes as well as the time to first renal impairment, taking into consideration how persistent the problem has been. In this model, each patient is treated as a multievent counting process. In both models, patients who did not experience any renal impairment during their observation periods were included in the analysis as censored observations. Collinearity among covariates was examined and significant risk factors associated with the primary endpoint of renal impairment were identified through the stepwise selection procedure at P < .05.
Medical records data from a total of 135 patients were collected, and 13 of those patients were excluded from the analysis (10 patients had no baseline creatinine test, and 3 patients had no postzoledronic acid creatinine test). Baseline demographics and characteristics for the 122 evaluable patients are shown in Table 1. The median observation period was 11.7 months. The median patient age at the time of first zoledronic acid infusion was 70.7 years (range, 45.6–88.4 years). Of 122 patients, 94.3% were white. There were 111 patients (91.0%) who had a normal baseline serum creatinine levels <1.4 mg/dL, and the median baseline serum creatinine level was 1.0 mg/dL (range, 0.5–2.0 mg/dL).
|Characteristic, n = 122||Median||Range|
|Duration of observation, mo*||11.7||0.5–35.5|
|Duration of zoledronic acid treatment, mo†||9.8||0.03–33.6|
|Age at time of first zoledronic acid infusion, y||70.7||45.6–88.4|
|No. of zoledronic acid infusions per patient||8||1–34|
|Interval between zoledronic acid infusions, d||34.1||21.8–302.5|
|Prior pamidronate use, no (%)||22 (18)|
|Duration of prior pamidronate treatment, mo‡||2.5||0.03–34.8|
|Comorbidities, no (%)|
|Coronary artery disease||34 (27.9)|
|Renal disease||16 (13.1)|
|Peripheral vascular disease||5 (4.1)|
|Benign prostatic hyperplasia||19 (15.6)|
|Diabetes mellitus||22 (18)|
|Recent surgery||46 (37.7)|
|Baseline serum creatinine, mg/dL||1||0.5–2|
Treatment and Follow-up
The median treatment period for zoledronic acid was 9.8 months. Patients received a median of 8 infusions each (range, 1–34 infusions), and there was a median interval of 34.1 days between infusions (range, 21.8–302.5 days). There were 72 patients (59.0%) who discontinued zoledronic acid treatment. Of these, 24 patients (33%) died during treatment of progressive prostate cancer. Three patients (4.2%) stopped secondary to bone pain attributed to zoledronic acid, and 30 patients (41.7%) stopped for various other reasons, including the end of tumor-related pain, transfer to an assisted-living facility or hospice care, and switch to treatment with pamidronate. Fifteen patients (20.8%) discontinued zoledronic acid because of renal complications that were noted by their treating physicians.
Renal Impairment Rates
Twenty-nine of 122 patients (23.8%) experienced ≥1 renal impairment event during treatment based on serum creatinine elevations (average, 3 events per patient; range, 1–12 events per patient) (Table 2). Of those 29 patients who experienced a worsening of renal function, 27 patients (93.1%) had normal renal function at baseline (baseline serum creatinine <1.4 mg/dL). Renal impairment based on declines in creatinine clearance was observed in 51 patients (41.8%). There was an increasing trend in renal impairment rates with increasing duration of zoledronic acid treatment that was demonstrated whether we used serum creatinine levels or calculated creatinine clearance measures. Renal impairment rates based on serum creatinine increased from 11.1% among patients who received zoledronic acid for <6 months to 26.3% among patients who received zoledronic acid for ≥2 years. Using the calculated creatinine clearance, the renal impairment rate increased from 23.9% among patients who received zoledronic acid for <6 months to 36.8% among patients who received zoledronic acid for ≥2 years (Table 2). However, this increasing trend was not statistically significant when it was evaluated by using either measure of renal impairment.
|Renal impairment||No. of patients (%)|
|Based on serum creatinine*|
|All patients||29/122 (23.8)|
|Among patients treated with zoledronic acid for†|
|<6 mo||13/117 (11.1)|
|≥12 mo||12/55 (21.8)|
|≥24 mo||5/19 (26.3)|
|Based on creatinine clearance‡|
|All patients||51/122 (41.8)|
|Among patients treated with zoledronic acid for†|
|<6 mo||28/117 (23.9)|
|≥12 mo||25/55 (45.5)|
|≥24 mo||7/19 (36.8)|
Figure 1 shows the effects of prior pamidronate treatment on renal impairment rates. Using serum creatinine levels, among 22 patients who previously received treatment with pamidronate, 10 patients (45.5%) experienced renal impairment while they were receiving zoledronic acid compared with 19 of 100 patients (19.0%) who had not previously received pamidronate. Using creatinine clearance measurements, 16 patients (72.7%) who previously received treatment with pamidronate experienced renal impairment after zoledronic acid compared with 35 patients (35.0%) who had not previously received pamidronate. Renal impairment rates also increased with increasing combined duration of pamidronate and zoledronic acid treatment, from 10.6% among patients who received total treatment for <6 months to 16.7% among patients who received total treatment for ≥36 months based on serum creatinine elevations, and from 22.1% among patients who received total treatment for <6 months to 33.3% among patients who received total treatment for ≥36 months based on declines in creatinine clearance (Fig. 2).
Table 3 shows the renal toxicity grade distribution using National Cancer Institute Common Toxicity Criteria. The majority of patients with any renal insufficiency had mild (grade 1) renal toxicity; however, 27 of 122 total patients (22.1%) and 26 of 29 patients (89.7%) who met the definition of renal impairment as defined in the current study had a renal toxicity at grade ≥2.
|Grade level*||No. of patients (%)|
|All patients||Patients with≥1 renal impairmentevent||Patients with no renal impairment events|
|Grade 1||34/122 (27.9)||3/29 (10.3)||31/93 (33.3)|
|Grade 2||23/122 (18.9)||22/29 (75.9)||1/93 (1.1)|
|Grade 3||3/122 (2.5)||3/29 (10.3)||0 (0)|
|Grade 4||1/122 (0.8)||1/29 (3.5)||0 (0)|
Risk Factors for Renal Impairment
Table 4 shows that increasing age and prior pamidronate use were identified consistently as significant risk factors associated with renal impairment in both the Cox model and the Andersen-Gill model. The Andersen-Gill model had greater power to detect additional risk factors, including history of renal disease, history of hypertension, and history of smoking, whereas the Cox model detected history of hypercalcemia as a risk factor.
|Risk factor||Cox proportional hazards model||Andersen-Gill model|
|Increasing age at zoledronic acid initiation per additional y||1.1||.01||1.1||<.01|
|Prior pamidronate use||2.2||.05||2.2||<.01|
|History of hypercalcemia||4.6||.01||NS|
|History of renal disease||NS||2.5||<.01|
|History of hypertension||NS||4.3||<.01|
|History of smoking||NS||2.5||<.01|
To our knowledge, this is the first published study to investigate renal toxicity in patients with HRPC who were receiving zoledronic acid in an observational setting. The results of this study demonstrated that 23.8% of patients had evidence of renal impairment during treatment with zoledronic acid. Furthermore, 20.8% of the patients who discontinued zoledronic acid did so as a result of renal complications. Significant risk factors for renal impairment in this patient population were age at zoledronic acid initiation, prior pamidronate use, and history of renal disease, hypertension, and smoking (P ≤ .05). Patients who received second-line bisphosphonates may have had more advanced disease than patients who received first-line bisphosphonates, which may explain why there was an increased risk of renal toxicity associated with prior pamidronate use. Furthermore, because the studies of pamidronate did not indicate a benefit in patients with HRPC,17 the number of patients who received prior pamidronate in clinical practice may be limited, as demonstrated in this study (≈18% of patients with HRPC).
Our results are consistent with previous studies that reported renal toxicity with approved doses of IV zoledronic acid leading to renal dysfunction and resulting in drug discontinuation, the need for renal dialysis, and, in rare cases, death.8 For this study, we did not investigate the impact of reducing the dose of zoledronic acid in patients with mild-to-moderate renal impairment; however, it is likely that this would reduce the incidence of renal impairment in patients receiving this bisphosphonate. Our findings also are consistent with a preliminary observational study in patients with prostate cancer that showed a greater incidence of renal toxicity (27.0%) than was noted in other clinical trials.13 Approximately 50% of all renal toxicities in the current study were considered grade 1. These less severe renal events may have limited clinical impact in this population with HRPC because of advanced malignancy and reduced life expectancy. However, of the 29 events that met the definition of renal impairment analyzed in the current study, 26 events (89.7%) were grade ≥2. No patients required any short- or long-term renal dialysis.
The risk of renal impairment increased with an extended duration of zoledronic acid therapy and previous pamidronate treatment. The etiologic association between zoledronic acid and renal toxicity is strengthened by the observation of this dose-response relation. Previous studies have shown a close temporal link between zoledronic acid treatment, prior pamidronate exposure, and renal deterioration.8–11 For example, Chang and colleagues suggested that the onset of renal failure and recovery of serum creatinine levels after drug discontinuation pointed to a temporal relation to the use of zoledronic acid. Renal failure occurred as early as 11 days after the first administration of zoledronic acid in 18 patients (25%), although the average time to renal failure was 56 days of zoledronic acid use. Furthermore, 45 of 72 patients in that study previously had received pamidronate on a long-term basis (mean duration of treatment, 600 days).8 In a study of 26 patients with breast cancer who received zoledronic acid, 16 patients were switched from pamidronate to zoledronic acid. This switch was associated with a statistically significant mean increase in serum creatinine (P = .02) compared with a lack of change in serum creatinine while they were receiving pamidronate.11 The mean decrease in renal function was worse in patients who received zoledronic acid after initially receiving treatment with pamidronate compared with the decrease in renal function in patients who received zoledronic acid de novo. Similar results were reported by Mazj and Lichtman: Of 35 patients who experienced bisphosphonate-related renal dysfunction, only 5 reports were attributed to pamidronate, whereas 30 patients developed renal dysfunction with zoledronic acid de novo (10 patients) or with prior pamidronate use (20 patients).9 The increased risk of renal impairment from previous bisphosphonate therapy is reflected in the prescribing information for zoledronic acid.6, 7
The risk of renal deterioration also was greater in elderly patients and in patients with prior renal impairment, renal disease, or hypertension. These results also are consistent with other studies. In a retrospective review of 293 patients who received bisphosphonates from January 2002 to June 2003, 35 patients (12%) developed renal dysfunction.9 The incidence of renal toxicity was more prominent in the elderly group (aged >80 years); and, with the exception of 1 patient, all patients with renal dysfunction in that group were treated with zoledronic acid. An increase in serum creatinine was observed in 7 of 35 patients (20%) in the elderly group compared with other age groups (28 of 258 patients; 11%). Another retrospective study of 234 patients who received zoledronic acid concluded that patients with preexisting renal impairment required close observation; however, in patients with normal baseline renal function, serum creatinine measurement prior to each dose of zoledronic acid probably was unnecessary.12 Nevertheless the prescribing information for zoledronic acid was updated to recommend that physicians routinely monitor renal function in all patients. It is believed that impairment of renal function associated with zoledronic acid is related to chelation of calcium, resulting in the formation of insoluble calcium bisphosphonate precipitates in the renal tubule.18
Potential limitations of this study were that there was no comparative control arm of patients with HRPC who received chemotherapy and supportive measures who were not receiving bisphosphonates. This is especially relevant because patients with prostate cancer may have renal impairment caused by their underlying disease; therefore, a control group would help determine more clearly the extent of renal impairment caused by zoledronic acid treatment. The retrospective nature of this study is another limitation; different incidences of renal impairment have been observed in prospective, randomized studies with more stringent inclusion criteria, longer-follow-up, and greater patient numbers.4, 6 However, it should be noted that there are limited data on bisphosphonates in this patient population. Prospective, randomized studies are needed to confirm the findings of the current retrospective study. An additional limitation is that the current results were based on the data that were available only from the DFCI, except for serum creatinine measurements, which also were obtained through primary care providers. Therefore, there is the potential for missing information on noncancer care. Another limitation is that the generalizability of the current study is limited to patients who are seen in a tertiary cancer center and may not apply to the overall population of patients with HRPC.
Given the evidence about the nephrotoxicity of zoledronic acid, the outcomes of ongoing clinical trials will be important. For example, a randomized trial conducted by the Cancer and Leukemia Group B is comparing the initiation of zoledronic acid in hormone-sensitive, metastatic prostate cancer with standard therapy at the time of HRPC. Because newer anticancer therapies may improve patient survival, it will be important to evaluate whether increased exposure to zoledronic acid with other potentially nephrotoxic medications increases the likelihood of renal toxicity.
Additional studies are warranted. It would be worthwhile to increase the sample size of future trials, which would increase their statistical power to allow for a more sensitive method of examining risk factors for renal impairment.
In summary, the results of the current study demonstrate that, in an outpatient oncology clinical setting, nearly 25% of patients who received zoledronic acid experienced a renal impairment event. This renal risk is significantly higher than was reported previously in clinical trials. Age at zoledronic acid initiation, prior pamidronate use, and a history of renal disease, hypertension, or smoking each significantly increased the risk of renal impairment with zoledronic acid. This suggests that these risk factors likely are important in predicting the renal impairment associated with zoledronic acid use in patients with HRPC who have bone metastases, and they should be incorporated into clinical strategies to reduce renal risk in this population. Our results reinforce existing recommendations to monitor renal function actively in these patients.
We thank Gardiner-Caldwell US for editorial assistance
- 6Novartis Pharmaceuticals Corporation. Zometa (zoledronic acid) [package insert]. East Hanover, NJ: Novartis Pharmaceuticals Corporation; 2004.
- 7Novartis International AG. Zometa (zoledronic acid). EU summary of product characteristics. Basel, Switzerland: Novartis International AG; 2005.
- 11Renal dysfunction with IV bisphosphonates in patients with metastatic breast cancer. J Clin Oncol. 2003; 21( suppl): 745. Abstract 2997., , , et al.
- 12Impact of zoledronic acid (Zol) on renal function in patients (pts) with cancer: is constant monitoring necessary? J Clin Oncol. 2003; 21( suppl): 755. Abstract 3036., , , et al.
- 13Significant deterioration in renal function with the new bisphosphonate, zoledronic acid. J Clin Oncol. 2003; 22: 738. Abstract 2968., , , et al.