Addition of pamidronate to chemotherapy for the treatment of osteosarcoma




This study evaluated the safety and feasibility of the addition of pamidronate to chemotherapy for treatment of osteosarcoma.


The authors treated 40 patients with osteosarcoma with cisplatin, doxorubicin, and methotrexate with the addition of pamidronate 2 mg/kg/dose (max dose 90 mg) monthly for 12 doses. Survival, event-free survival (EFS), and durability of orthopedic reconstruction were evaluated.


For patients with localized disease, event-free survival (EFS) at 5 years was 72% and overall survival 93%. For patients with metastatic disease, EFS at 5 years was 45% and overall survival 64%. Toxicity was similar to patients treated with chemotherapy alone. Thirteen of 14 uncemented implants demonstrated successful osteointegration. Among allograft reconstructions, there were 2 graft failures, 4 delayed unions, and 6 successful grafts. Overall, 5 of 33 reconstructions failed. There were no stress fractures or growth disturbances.


Pamidronate can be safely incorporated with chemotherapy for the treatment of osteosarcoma. It does not impair the efficacy of chemotherapy. Pamidronate may improve the durability of limb reconstruction. Cancer 2011. © 2010 American Cancer Society.

The identification of effective chemotherapy for the treatment of osteosarcoma (OS) has led to significant improvement in patient outcome.1–2 There are only 4 chemotherapy agents widely accepted to have efficacy against osteosarcoma: doxorubicin, cisplatin, high-dose methotrexate (HDMTX), and ifosfamide.3-7 The Children's Oncology Group (COG) performed a randomized, multi-institution, cooperative trial to test the benefit of the addition of ifosfamide to a regimen of cisplatin, doxorubicin, and HDMTX.8 The addition of ifosfamide did not result in improved event-free survival (EFS) or overall survival. There is a continued need for new therapeutic approaches.

Bisphosphonates are analogues of endogenous pyrophosphates. In vivo, bisphosphonates bind strongly to hydroxyapatite on bone surfaces and are delivered to sites of increased bone formation or resorption. They are widely used to treat hypercalcemia of malignancy. Initially, it was felt that the mechanism of action of bisphosphonates was exclusively to stabilize bone. More recently it has become apparent that bisphosphonates have direct effects on tumor cells.9 In vitro, bisphosphonate treatment of myeloma cells leads to growth inhibition and induction of apoptosis.10 Bisphosphonates induce apoptosis in human breast cancer cell lines.11 Bisphosphonates appear to inhibit adhesion of tumor cells to bone matrix.12 Bisphosphonates may also inhibit matrix metalloproteinases, which are used by tumor cells to invade tissue and establish metastases.13

Numerous in vitro and xenograft studies support the concept that bisphosphonates have activity against osteosarcoma, alone or in combination with chemotherapy. Investigators have studied human and animal cell lines and spontaneous osteosarcoma arising in murine models. They have studied alendronate, clodronate, minodronate, pamidronate, and zoledronate14-37 (Table 1). Oligonucleotide microarray assays of human osteosarcoma offer additional support for the concept of using bisphosphonates in osteosarcoma. Expression profiling of 30 osteosarcoma tumors from patients identified 104 genes differentially expressed between favorable and unfavorable responses to chemotherapy.36 A striking finding was the significant decrease in osteoprotegerin, an osteoclastogenesis inhibitory factor. Additional genes involved in osteoclastogenesis and bone resorption, which were statistically different, included annexin 2, SMAD, PLA2G2A, and TGFbeta1. ECM remodeling genes included desmoplakin, SPARCL1, biglycan, and PECAM. Overexpression of desmoplakin (P = .008), PECAM (P = .028), and SPARCL1 (P = .00,098) were associated with a favorable chemotherapy response, whereas overexpression of annexin2 (P = .05), biglycan (P = .025), and PLA2G2A (P = .025) were associated with an unfavorable chemotherapy response. This suggests the interaction of the tumor with the microenvironment is a determinant of response to chemotherapy. Bisphosphonates may have activity through disruption of these interactions.

Table 1. Preclinical Investigation of Bisphosphonates in Osteosarcoma
AuthorReferenceYearBisphosphonateChemotherapyModel System
  1. Doxo indicates doxorubicin; Carbo, carboplatin; vcr, vincristine; etop, etoposide; gem, gemcitabine; os, osteosarcoma.

Cheng162004Alendronate Human lines in vitro
Farese192004Alendronate Dog lines in vitro
Molinuevo292007Alendronate Rat cell line in vitro
Heikkila202003Clodronate Human lines in vitro
Kubo272006Minodronate Human in vitro/xenograft
Kubo262008MinodronateDoxoHuman in xenograft
Mackie282001Pamidronate Rat cell line in vitro
Sonnemann352001Pamidronate Human lines in vitro
Ashton142005Pamidronate Dog lines in vitro
Murayama312008RisedronateCarbo doxo vcr etopHuman lines in vitro
Evdokiou182003Zoledronate Human lines in vitro
Heymann212005ZoledronateIfosfamideRat cell line in vivo
Ory332005Zoledronate Mouse cell line in vivo
Horie232006Zoledronate Mouse cell line in vitro
Kubista252006Zoledronate Human lines in vitro
Benassi152007ZoledronateCisplatinHuman lines in vitro
Dass172007Zoledronate Human lines in xenograft
Horie222007ZoledronateDoxo paclitaxel gemMouse cell line in vitro
Iguchi242007Zoledronate Human lines in vitro
Muraro302007Zoledronate Human lines in vitro
Ory322007Zoledronate Human/rat lines in vitro
Ory342008Zoledronate Human/rat lines in vitro
Koto372009Zoledronate Mouse spontaneous os

There are a large number of bisphosphonate compounds in clinical use. Osteosarcoma is most common in the second decade of life. We initiated our clinical trial in 2003. We needed to identify a bisphosphonate which had been used in this younger population and demonstrated a good safety profile. We also needed to identify a dose and schedule of administration which could be applied in children and adolescents as well as in adults. Pamidronate had been the most widely used bisphosphonate in children and young adults (Table 2). Pamidronate has been used to treat cancer-related hypercalcemia in children38-40 Pamidronate has been used to treat fibrous dysplasia in children with the McCune-Albright syndrome.41 Pamidronate has been used to treat a teenager with multifocal Langerhans cell histiocytosis.27 Pamidronate has been used to treat osteogenesis imperfecta, including some very young children.43, 45 Pamidronate has been given to children with osteoporosis resulting from a variety of etiologies.46 Pamidronate has been safely given simultaneously with chemotherapy to children with acute lymphocytic leukemia (ALL).44 Although other bisphosphonates were in clinical use, this drove our selection of this agent for this clinical trial.

Table 2. Pamidronate Experience in Pediatrics
YearIndicationReferenceAge RangeDoseSchedule
  1. ALL indicates acute lymphoblastic leukemia.

1998Hypercalcemia392-151 mg/kg
2001Hypercalcemia4051 mg/kg
2000Fibrous dysplasia410.5-1 mg/kg q.d. x3 dq 6mo
2001Langerhans cell histiocytosis42142 mg/kg q.d. x3 dq 1mo
2002Osteogenesis imperfecta431-160.25-1 mg/kg q.d. x3 dq 2-4mo
2001ALL443-161 mg/kg q.d. x3 dq 3mo

The purpose of this trial was to determine whether pamidronate can safely be given in conjunction with chemotherapy to young patients with osteosarcoma. We wished to determine the feasibility of this approach and to determine whether the combination of chemotherapy with pamidronate results in increased toxicity or in decreased efficacy against tumor. We wished to assess the ability of pamidronate to improve the survival of endoprosthetic reconstruction.



We offered participation in the clinical trial to all patients with newly diagnosed previously untreated osteosarcoma presenting to the Memorial Sloan-Kettering Cancer Center (MSKCC) Department of Pediatrics. This clinical trial was approved by the Memorial Hospital Institutional Review Board. Patients were eligible when they had adequate renal, hepatic, hematopoietic, and cardiac function. Exclusion criteria included prior treatment for any cancer, prior history of Paget disease, prior history of pericarditis, myocarditis or cardiac conduction abnormalities, and pregnancy or lactation. All patients or their guardians were required to provide written informed consent to participate in the study. Because the primary aim of the study was to evaluate the safety and feasibility of contemporaneous administration of pamidronate and chemotherapy, patients with both localized and metastatic osteosarcoma were eligible for participation.

We enrolled 40 patients on study. Twenty-nine patients presented without clinically detectable metastatic disease, and 11 patients had clinically detectable metastatic disease at initial presentation. Patients ranged in age from 7 to 36 years with a median age of 15 years. There were 20 males and 20 females. Two patients had a pathological fracture, each successfully treated with limb preservation.


The treatment plan called for a period of induction, followed by definitive surgical resection of the primary tumor, followed by resection of any pulmonary metastases, followed by a period of maintenance chemotherapy. Chemotherapy consisted of cisplatin 120 mg/m2 administered as a 4 hour infusion 4 times with doxorubicin, twice during induction at Weeks 0 and 5, and twice during maintenance at Weeks 0 and 5. Doxorubicin was administered as a 15-30 minute infusion at a dose of 37.5 mg/m2/day for 2 consecutive days 6 times, twice during induction at Weeks 0 and 5, and 4 times during maintenance at Weeks 0, 5, 10, and 15. We administered dexrazoxane 375 mg/m2 as a 15-30 minute infusion 15 minutes before each dose of doxorubicin. The first 4 courses were administered with cisplatin, twice during induction and twice during maintenance. High dose methotrexate (HDMTX) 12 g/m2 with a maximum dose of 20 g was administered as a 4 hour infusion followed by leucovorin administration at a dose of 10 mg (not adjusted to body surface area) beginning 24 hours from the initiation of the methotrexate infusion and continuing until the serum methotrexate level was less than 1 × 10−7M (100 nanomolars). Serum methotrexate levels and renal function were monitored daily, and hydration, alkalinization, and leucovorin doses were specified in the event of delayed methotrexate excretion.45 HDMTX was administered 12 times, 4 times during induction at Weeks 3, 4, 8, and 9; and 8 times during maintenance at Weeks 3, 4, 8, 9, 13, 14, 18, and 19. In an effort to maintain the dose intensity of doxorubicin, the protocol specified that if there were a delay greater than 1 week between the first and the second of each pair of HDMTX administrations, the second of the pair was to be omitted.

Experience with pamidronate in children has used doses from 1 to 6 mg/kg given over the course of 1 to 3 days. If we use bisphosphonates to treat osteosarcoma, we must administer the bisphosphonate with chemotherapy agents with potential nephrotoxicity. We chose a dose of 2 mg/kg (maximum dose 90 mg) given in 1 day as a safe dose. We specified that administration of pamidronate would be separated from administration of cisplatin or high-dose methotrexate by at least 72 hours. Pamidronate was administered once each month for a total of 12 doses. The first dose of pamidronate was given during the first cycle of chemotherapy. Pamidronate was always separated from cisplatin and HDMTX by a minimum of 72 hours. Pamidronate was administered as a 2-hour infusion.

Orthopedic Reconstruction

Surgery was of curative intent and achieved a wide margin in all but 1 sacral case where the margin was positive on final pathological review. Reconstructions varied based on the location and circumstances, including patient growth potential and the presence of metastatic disease. There were 3 amputations, and 5 patients who did not have any reconstruction. There were 23 prosthetic reconstructions, including 19 pure implants (1 press fit, 14 Compress, 4 cemented stems) and 4 allograft prosthetic composites (APC) where the stem was cemented into the allograft, and the remaining stem was press fit into the host bone (3 patients) or cemented into the host bone (1 patient). There were 11 allografts, including the 4 APCs. Two of the intercalary grafts were coupled with vascularized fibular transplants. We fixed the allografts with plates in all cases, and the APCs also had intramedullary stem transfixation. A vascularized fibula alone was used for 1 intercalary reconstruction.

Mobilization was based on a standard protocol. Cemented implants were allowed immediate weight bearing. Uncemented implants, including the Compress stems, were protected by toe-touch weight bearing for 6 weeks then half weight bearing for 6 weeks. APCs were protected by half weight bearing for 12 weeks. Allografts and vascularized grafts were protected with half-weight bearing until there was painless radiographic union. The vascularized fibula was then braced for an additional 2 years until graft hypertrophy. Five of the 23 prosthetic reconstructions also included extensible shaft segments. These were lengthened as needed to keep the limb length inequality <1 cm. Limb lengths were monitored by physical examination by using blocks under foot to level the pelvis and tape measurements from the anterior iliac spine to the medial malleolus. We rarely used scanograms.

Graft failure was defined by removal for any reason, persistent nonunion of 18 months after surgery, or 12 months after the conclusion of chemotherapy. Prosthetic failure was defined as the removal of an implant, for any reason, or pain with progressive radiolucency. Successful osteointegration was defined by retention of the prosthesis, no pain, no radiolucency, and progressive bone hypertrophy around the implant.

Endpoints and Statistical Methods

Endpoints were EFS, overall survival, toxicity, and success rates for endoprosthetic reconstruction after definitive resection of primary tumor. Actuarial curves were estimated by using the Kaplan-Meier method for EFS and overall survival. Toxicity was monitored with National Cancer Institute (NCI) common toxicity criteria, with special attention to hepatotoxicity, ototoxicity, nephrotoxicity, and incidence of osteonecrosis of the jaw. We compared the incidence of grade III and IV adverse events for these selected toxicities in this regimen to the incidence in 390 patients who received the identical chemotherapy regimen without pamidronate in the prospective randomized trial performed by the pediatric cooperative groups.8


Data were analyzed as of April, 2010. Median follow-up for the entire cohort of 40 patients, including patients who experienced an event, was 53 months.

EFS and Survival

EFS for 29 patients who presented with localized disease was 72% at 5 years from study enrollment (Fig. 1). EFS for the 11 patients who presented with clinically detectable metastatic disease was 45% at the same time point. Overall survival for the localized patients at 5 years was 93%; for patients who presented with metastatic disease, overall survival at 5 years was 64% (Fig. 2). We did not observe local recurrence. Two patients developed myelodysplastic syndrome as the first event. All other first events were metastatic recurrence.

Figure 1.

Event-free survival is shown for patients who presented with localized or metastatic osteosarcoma at initial presentation.

Figure 2.

Overall survival is shown for patients who presented with localized or metastatic osteosarcoma at initial presentation.


The toxicities observed in patients in this study were similar to the toxicities observed in patients treated with the same chemotherapy regimen who did not receive pamidronate. Hypocalcemia was common after administration of pamidronate, but only 2 patients experienced symptoms, including perioral numbness and paresthesias of the hands and feet (Table 3). Symptomatic hypocalcemia responded promptly to oral calcium supplementation, and subsequent administration of pamidronate preceded by oral calcium supplementation was not associated with symptoms. With subsequent administration of pamidronate, the incidence of hypocalcemia decreased. We observed ototoxicity ≥grade 3 in 6 of 40 patients (15%; 95% confidence interval [CI], 6%-30%). Ototoxicity ≥grade 3 was observed in 39 of 390 patients treated with the same chemotherapy regimen without bisphosphonate in the pediatric cooperative group trial (P = .29; Fisher exact test; Table 4). We observed nephrotoxicity ≥grade 3 in 1 of 40 patients (2.5%; 95% CI, 0.1%-13%). Nephrotoxicity ≥grade 3 was observed in 8 of 390 patients treated with the same chemotherapy regimen without bisphosphonate in the pediatric cooperative group trial (Fisher exact test P = .58; Table 4).There were no cases of osteonecrosis of the jaw, either during study therapy or during follow-up. No patients sustained atypical subtrochanteric or other long bone fractures, which have been reported with bone metastasis and osteoporosis patients receiving bisphosphonate therapy for >5 years.48 We observed no difference in toxicities or delays in chemotherapy between patients older or younger than 18 years of age.

Table 3. Hypocalcemia Following Administration of Pamidronate
Pamidronate Cycle123456789101112
  1. SD indicates standard deviation.

Calcium nadir, mg/dL Mean ± SD7.0 ± 1.07.9 ± .78.1 ± 0.78.2 ± 0.68.0 ± 0.68.4 ± 0.38.4 ± 0.48.9 ± 0.58.9 ± 0.59.2 ± 0.49.2 ± 0.49.2 ± 0.5
Hypocalcemia, grade 3/41952010000000
Table 4. Selected Toxicity Comparison
ToxicityCurrent TrialIntergroup Trial
Ototoxicity≥Grade 36/40 (15%)39/390 (10%)
Nephrotoxicity≥Grade 31/40 (2.5%)8/390 (2%)

Orthopedic Reconstruction

The 11 allograft reconstructions included 4 osteoarticular tibial replacements (all plated, 3 with intramedullary cement), 1 intercalary femur replacement (plated without cement), 2 intercalary femoral replacements with intramedullary vascularized fibulas (plated without cement), and 4 alloprosthetic composites (1 proximal humerus,1 proximal femur, and 2 proximal tibias, all of which had intramedullary prosthetic stems; 2 had supplemental plates, and 1of 4 was cemented into the remaining host bone.) The number of variables among the patients is too great to allow meaningful comparison. Due to 3-6–month variation in the intervals between extremity films, the time-to-union results may not be precise. Nevertheless, radiographic union was achieved in 11 of 13 osteosynthesis sites at a mean of 19.4 months (standard deviation [SD], 7.2 months). Our impression was that the healing was at least as fast as what has historically been seen for similar reconstructions during chemotherapy. Ultimate union and graft retention are more reproducible and clinically meaningful outcomes.

Five reconstructions failed. One (of 1) uncemented press-fit stems had aseptic loosening and was converted to a cemented stemmed implant that had a stable 2 mm radiolucent line over one-third of the stem length, 5 years and 5 months after implantation. Four allografts failed, 2 from infection (1 exchanged for a cement intercalary spacer and 1 amputation) and 2 from nonunion, successfully treated by autogenous bone grafting and exchange from plate to rod fixation. A fifth allograft, part of an APC, had a persistent asymptomatic nonunion that had not failed nor required surgery by the time of the patient's death of disease, 18 months after surgery. Of a total of 14 osteosynthesis sites, 9 united.


Successful treatment for osteosarcoma requires the combination of effective systemic therapy and surgery to remove all sites of clinically detectable disease. One treatment strategy for osteosarcoma is to administer chemotherapy, followed by definitive surgical resection of primary tumor and metastatic disease if necessary, followed by additional adjuvant chemotherapy. A second strategy is primary definitive surgery followed by adjuvant chemotherapy. Different combinations of the 4 active chemotherapy drugs have been used in both strategies. Large trials from single institutions and cooperative groups have achieved similar outcomes.1, 8, 49-51 We need to identify new agents and treatment strategies.

The bisphosphonates are good candidates to employ in the treatment of osteosarcoma. Osteosarcoma cells demonstrate upregulation of many genes whose normal functions are to participate in osteogenesis and whose functions can be inhibited by bisphosphonates. In addition, evidence is accruing that bisphosphonates may have the ability to interfere with the processes used by tumor cells to establish metastases. The treatment of osteosarcoma requires surgical resection of tumor-bearing bone and reconstruction of the resected area with metal or bone graft. Osteosarcoma occurs predominantly in young patients and long-term survival of the reconstruction is essential. Bisphosphonates may improve the outcome for osteosarcoma by direct antitumor effects, decreasing the risk of metastasis, and improving the durability of reconstruction following tumor resection.

Our experience represents a pilot study with a single bisphosphonate, pamidronate. We chose pamidronate because there was prior experience with pamidronate in children with cancer. Newer bisphosphonates, such as zoledronate, are significantly more potent than pamidronate. There are more data from preclinical studies for zoledronate in osteosarcoma than for any other bisphosphonate. Future clinical trials of bisphosphonates in osteosarcoma will almost certainly employ zoledronate. It is important to recognize that although zoledronate is more potent than pamidronate, it is also associated with a greater risk of osteonecrosis of the jaw, a significant risk associated with bisphosphonate therapy.52 Its risk in young children with deciduous teeth is unknown.

We treated 40 patients with conventional chemotherapy for osteosarcoma and pamidronate. Given the limitations of small patient sample and limited duration of follow-up, we observed no statistically significant increase in toxicity. We saw no osteonecrosis of the jaw. We believe that pamidronate can safely be incorporated into a multiagent chemotherapy regimen for the treatment of osteosarcoma. If zoledronate replaces pamidronate in the treatment strategy, then we will need to acquire similar safety data. We observed EFS and overall survival for our patients very similar to our prior experience and to the published experience in the literature for similar chemotherapy treatment regimens. This was a small single-arm study that included both localized and metastatic patients. We cannot draw any firm conclusions about the impact of pamidronate on the efficacy of treatment for osteosarcoma, but pamidronate does not appear to have impaired the outcome.

In a series of 108 patients with primary bone sarcoma treated at MSKCC, 15 (14%) patients suffered fractures during treatment.53 These children remain at risk for osteoporosis and insufficiency fractures throughout their lifetimes. Treatment-related osteopenia is an under-recognized problem in young patients receiving chemotherapy.54 In adults with osteoporosis, treatments with bisphosphonates resulted in a 50% decrease in fracture rates after 1 year of treatment. These adults achieved bone mass gains of 2-4% per year during the first 4 years of treatment. Even more devastating for these patients, pathologic fractures through a tumor can affect survival and local recurrence rates. In a multicenter, retrospective, matched-control review, patients with osteosarcoma and a pathologic fracture had a 5-year survival rate of 55% compared with 77% for patients without fracture. Local recurrence at 5 years was also increased in the group with pathologic fractures (25% compared with 4%).55

Most patients with osteosarcoma undergo limb-preserving surgical resection with insertion of an endoprosthesis. During chemotherapy, ingrowth into the surface of these prostheses is delayed. Aseptic loosening, with poor bone/implant interface contact remains 1 of the major factors leading to the necessity for implant revision surgery. At MSKCC, 15.8% of prosthetic knee reconstructions require revision for aseptic loosening. Overall prosthesis failure-free rates were 82%, 71%, and 50% at 3, 5, and 10 years, respectively, after initial implantation.56 At the University of California at Los Angeles (UCLA), 11.6% of hip prostheses required revision for aseptic loosening or fatigue failure, and failure-free survival at 7 years was 69%.57 Even recent uncemented prostheses and implants with novel Compression fixation have 12% early failure by 5 years.58 Currently, young patients who survive osteosarcoma can anticipate multiple revisions of their prostheses during their lifetime. Bisphosphonate therapy may contribute to improved prosthetic longevity by several mechanisms including 1) improved bone density and strength, 2) promoting more robust bone ingrowth into porous surfaces of uncemented prostheses, and 3) stabilization of the bone-prosthesis or bone-cement interface retarding osteoclastic bone resorption stimulated by particulate-wear debris. Bisphosphonates improve the fixation interface and can be predicted to improve failure-free implant survival.

We observed a high rate of successful osteointegration of endoprosthetic reconstructions and union following allografts. Overall durability of reconstructions has been good. We did not formally evaluate bone density in patients receiving pamidronate; such prospective evaluation in future trials of more potent bisphosphonates such as zoledronate may provide additional information on the usefullness of bisphosphonates in patients with sarcomas receiving therapy associated with an increased risk of osteopenia. Again, the limitations of a small single-arm study preclude firm conclusions, but these outcomes compare favorably with our prior institutional experience.

Our experience with pamidronate and chemotherapy for the treatment of osteosarcoma suggests that we can safely incorporate pamidronate with chemotherapy. It suggests the efficacy of therapy is comparable to our prior experience. It suggests that bisphosphonates may improve the durability of reconstruction. These results provide feasibility data, support, and justification for a prospective randomized trial of the addition of bisphosphonates to chemotherapy for the treatment of osteosarcoma. It seems appropriate that such a trial should use a newer, more potent bisphosphonate than pamidronate, such as zoledronate. The use of zoledronate will require vigilant monitoring for toxicity.


This study was supported by NCI grant CA106450.