A pilot study of thalidomide in patients with progressive metastatic renal cell carcinoma




The highly vascular nature of renal cell carcinoma (RCC) suggests that angiogenesis inhibition may be therapeutic for patients with this disease. Thalidomide inhibits basic fibroblast growth factor and vascular endothelial growth factor (VEGF)-induced angiogenesis.


In a pilot study, we evaluated the safety and efficacy of escalating doses of thalidomide in patients with progressive metastatic RCC (mRCC), measurable disease, and good organ function. Patients received oral thalidomide starting at 200 mg per day and increasing by 100–200 mg per day weekly until a target dose of 1200 mg per day was reached. Study endpoints were objective tumor response and toxicity.


Of the 20 patients enrolled, 19 were evaluable for response. Eighteen achieved the target dose. The most common, but reversible, toxicities were constipation, somnolence, and fatigue. Peripheral neuropathy was seen after prolonged therapy, necessitating dose reduction. Two patients achieved a partial response and nine had stable disease for a median of 14 months (range, 3–17 months). Median time to progression was 4.7 months (range, 0.7–31.3 months). Fifteen patients died (median survival, 18.1 months; 95% lower confidence bound 10.7). Survival was significantly longer in patients with higher hemoglobin level and longer time from first metastasis to start of thalidomide, but significantly shorter in patients with multiple organ involvement and previous treatments.


Thalidomide at this dose is associated with manageable acute toxicities but long-term dose-limiting neuropathy. Objective responses are rare in patients with mRCC and are characterized by delay in achieving maximum tumor reduction. Prolonged stable disease is seen in some patients, but the benefit of thalidomide, as well as other angiogenesis inhibitors, in that setting needs to be studied in controlled, randomized trials. Cancer 2002;95:758–65. © 2002 American Cancer Society.

DOI 10.1002/cncr.10740

Metastatic renal cell carcinoma (mRCC) is an incurable disease that responds poorly to systemic therapy. It is resistant to most chemotherapeutic agents and only 15–20% of patients experience an objective response to immunotherapy with either interleukin-2 (IL-2) or interferon-alpha (IFN-α).1, 2 A combination of high-dose chemotherapy and allogeneic “mini” (nonmyeloablative) transplantation elicits objective responses, but not without high morbidity and mortality.3 Novel approaches to the treatment of mRCC patients are needed.

One possible approach involves treatment with antiangiogenic agents. Angiogenesis is important in embryogenesis, wound healing, diabetic retinopathy, and cancer progression.4, 5 Overexpression of basic FGF (bFGF), a proangiogenic molecule, in patients with RCC is associated with poor survival rates.6 In seeking a candidate angiogenesis inhibitor, we studied thalidomide because it inhibits angiogenesis induced by bFGF and VEGF,7, 8 induces apoptosis in established neovasculature in experimental models,7, 8 and exerts immunomodulatory effects.9–11 Thalidomide has been approved by the Food and Drug Administration to treat patients with erythema nodosum leprosum.12 It also has antitumor activity against multiple myeloma,13, 14 Kaposi sarcoma,15 and recurrent gliomas.16 In this pilot study, we evaluated its safety and efficacy in patients with progressive mRCC.


Patient Selection

Twenty consecutive eligible patients with progressive mRCC entered this pilot clinical trial. All patients gave their written informed consent on a form approved by the institutional review board at The University of Texas, M. D. Anderson Cancer Center. Patients were required to have histologically confirmed RCC with evidence of progressive, bidimensionally measurable metastatic disease, Zubrod performance status (ZPS) of 0–3, and adequate organ function (serum creatinine level ≤ 2.0 mg/dL, total bilirubin level ≤ 1.5 mg/dL, alanine aminotransferase ≤ four times the upper limit of normal, absolute neutrophil count ≥ 1500/mm2, and a platelet count ≥ 100,000/mm2). There were no exclusions for previous therapy, but at least 4 weeks had to have elapsed since the last dose of treatment, surgery, or radiotherapy. Previously irradiated or angioinfarcted lesions were not used to evaluate response. Patients with brain metastases were eligible provided that they were neurologically stable and did not require intravenous steroids. Patients of child-bearing age were eligible provided that they were practicing adequate contraception. All female patients of child-bearing age had to have been on birth control for 1 month before receiving thalidomide. Two methods of birth control were required, with emphasis on combination of one hormonal and one barrier method. If a female patient had a contraindication to hormone methods, then two barrier methods were recommended. For male patients, condom use was required. Contraception was required to continue for at least 2 months after discontinuation of thalidomide therapy.

Treatment Plan

Thalidomide was supplied initially in 100-mg capsules by Andrulis Pharmaceuticals (Bethesda, MD). After January 2000, the drug was supplied in 50-mg capsules by Celgene Corporation (Warren, NJ). Thalidomide was administered at night, starting at a dose of 200 mg per day. The total daily dose was increased 100–200 mg per day every week until a target maximum dose of 1200 mg per day was achieved. During the dose-escalation phase of the trial, patients were monitored by interim history and physical examination including a neurologic examination, with special attention paid to the development of neuropathy. If no significant toxicities were encountered, the dose of thalidomide was increased by 200 mg per day. If Grade 1 toxicity was observed, the dose was increased by 100 mg per day every week. If Grade 2 peripheral neuropathy was observed, the dose was not changed. At higher doses, patients were permitted to take thalidomide in three divided doses (usually 20% of the drug in the morning, 20% at noon, and 60% at bedtime). Because thalidomide causes constipation, patients were started on a bowel regimen that included sodium docusate daily from Day 1 of the trial to be titrated freely by the patient. Patients were also encouraged to take adequate amounts of dietary fiber and were treated with lactulose as needed. The thalidomide dose was reduced by 100–200 mg per day in patients who developed significant toxicity (mainly neuropathy). Therapy was discontinued for any of the following reasons: disease progression, intolerable or unresolved adverse events, patient's decision, or noncompliance by patient with the protocol's requirements.

Criteria for Response and Schedule for Tumor Evaluation

The primary endpoints for this pilot trial were objective tumor response and toxicity. Toxicity evaluations were performed weekly during the dose-escalation phase and every 4 weeks thereafter. Toxicity was graded according to the National Cancer Institute Common Toxicity Criteria (version 2.0). Following the establishment of a stable daily dose of thalidomide, patients were followed monthly by history and physical evaluation, complete blood count with differential and platelet count, and chemical survey. In addition, serum β-human chorionic gonadotropin level was monitored monthly for female patients of child-bearing age.

Responses were evaluated after the initial 12 weeks of therapy and every 8 weeks thereafter. Complete response was defined as the disappearance of all clinical and radiographic evidence of tumor persisting for at least 4 weeks. Partial response (PR) was defined as a 50% or greater decrease in the sum of the products of diameters of all measurable lesions persisting for at least 4 weeks without the development of any new lesions. Progressive disease (PD) was defined as a 25% or greater increase in the sum of the products of diameters of all measurable lesions or the appearance of any new lesions. All other responses were defined as stable disease (SD). Progression-free survival (PFS) was defined as the time from the start of thalidomide therapy to disease progression, whether on or off thalidomide. Overall survival (OS) was defined as the time from the start of therapy to death. All PFS and OS times were administratively censored as of June 1, 2001.

Statistical Design and Analysis

This pilot trial was designed to accrue 18 patients evaluable for response. Patients who were unable to complete the first 12-week cycle of treatment were inevaluable for response. Due to simultaneous commitment to different patients, 20 patients were accrued in the trial. The trial was designed so that if 2 or more of 18 evaluable patients achieved a PR, thalidomide would be considered worthy of further study. Univariate data were summarized using standard descriptive statistics. Unadjusted PFS and survival probabilities were estimated using the Kaplan–Meier method.17 The Cox proportional hazards regression model18, 19 was used to assess the ability of patient characteristics to predict survival, with goodness-of-fit assessed by the Grambsch–Therneau test, martingale residual plots, and likelihood ratio statistics.19 Predictive variables in the Cox model were selected by performing a backward elimination with P < 0.05 as the cutoff and then allowing any variable previously deleted to reenter the final model once P < 0.05. All computations were carried out on a Gateway PC using the Windows NT operating system and a DEC Alpha 2100 5/250 in Splus20 using both standard Splus functions and the Splus survival analysis package of Therneau.21


Patient Characteristics

Of 20 patients with progressive mRCC (Table 1), 19 were evaluable for response. Response could not be assessed in one patient who stopped taking thalidomide after the first 2 weeks because of rapid deterioration of tumor-related symptoms. The patient did not return for reevaluation. Because there were no exclusion criteria for previous therapy, the study group was pretreated heavily. In fact, 19 patients (95%) had received at least one previous therapy (range, 0–5; median, 2). Previous therapy included IL-2 in 17 patients (for mRCC in 16 and as adjuvant therapy in 1), IFN-α in 19 patients, chemotherapy (usually fluoropyrimidine based in the context of a clinical trial) in 19 patients, radiation therapy to metastatic sites in 2 patients, and resection of brain metastases in 2 patients. Eleven patients (55%) had a ZPS of 1 or higher.

Table 1. Baseline Characteristics of All Enrolled Patients
CharacteristicNo. of PatientsMedian (range)
Total enrolled/evaluable for response20/19 
Gender (male/female)16/4 
Age (yrs)  
 Median (range) 53 (31–71)
Tumor grade (at diagnosis)  
Performance status (Zubrod)  
No. of organs involved  
 ≥ 39 
Sites of metastatic disease  
 Lung/lung only18/3 
 Lymph node (with/without other sites)11 
 Liver (with/without other sites)7 
 Bone (with/without other sites)4 
 Brain (with/without other sites)2 
 Renal bed (with/without other sites)5 
 Adrenal (with/without other sites)3 
 Duodenal mass (with/without other sites)1 
Prior nephrectomy/embolization19/1 
Interval, first diagnosis to therapy (mos)  
 Median (range) 28 (11–92)
Interval, metastatic disease diagnosis to therapy (mos)  
 Median (range) 25 (8–72)
Prior therapy for metastatic disease  
 Any prior systemic treatment19 
 Radiation therapy2 
 Surgery (brain metastases resection)2 
No. of prior regimens  
 Median (range) 2 (0–5)
Adrenal insufficiency2 
Hemoglobin (g/dL)  
 Median (range) 11.9 (7.8–18.4)
Albumin (g/dL)  
 Median (range) 3.85 (3.2–4.5)
Alkaline phosphatase (IU/L)  
 Median (range) 104 (64–155)
Lactate dehydrogenase (IU/L)  
 Median (range) 505 (337–1666)

Most patients had relatively indolent disease. The median interval between diagnosis of metastatic disease and study entry was 24.7 months (range, 7.6–71.5 months). The median interval between initial diagnosis of RCC and start of thalidomide therapy was 27.8 months (range, 10.9–92.1 months). Nevertheless, all patients had a significant disease burden; nine patients (45%) had three or more organs involved with metastatic disease.

Retrospective assignment to Memorial Sloan Kettering Cancer Center (MSKCC) risk group on the basis of pretreatment characteristics22 showed that all but two patients had favorable (n = 3) or intermediate-risk (n = 15) clinical features.

Tumor Response

The primary study efficacy endpoint was objective response. Of the 19 evaluable patients, 2 had a PR, for an overall response rate of 10.5% (95% confidence interval [CI], 1.3–33 %). These two responses were documented at 7 and 11 months of therapy and lasted for 16 and 31 months, respectively (Table 2). The first partial responder, a male, was diagnosed initially with pathologic stage T3 sarcomatoid RCC in January 1997 when he underwent a radical nephrectomy. In June 1997, renal bed recurrence and lymph node metastases were diagnosed. The patient was treated with subcutaneous IL-2/IFN-α and intravenous continuous infusion of 5-fluorouracil (5-FU). SD was achieved, but treatment was discontinued in June 1998 because of drug-related toxicity. When enrolled in the present trial in September 1998, the patient showed evidence of PD in the lung, bone, and adrenal gland and renal bed recurrence. He achieved a PR 7 months after initiation of thalidomide and continued on thalidomide for 9 more months, until his disease progressed and he was taken off the study. He died 2 months later of progressive mRCC.

Table 2. Patient Responses
OutcomeNo. of PatientsTime on thalidomide
  • a


  • b


Complete response0
Partial response216 and 31 mos
Stable disease93–17 mosa
Progression of disease81–6 mosb
Inevaluable13 weeks

The second partial responder, who was also male, had a history of organ-confined, Grade III clear cell carcinoma status postnephrectomy in February 1994. In September 1996, he showed evidence of liver and lung metastases. Between January 1997 and December 1997, he was treated subcutaneously. with IL-2/IFN-α and with intravenous continuous infusion 5-FU and achieved SD. In February 1998, when progression was seen in the lung and liver, he started treatment with intravenous continuous infusion floxuridine, subcutaneous IFN-α, and oral 13-cis-retinoic acid. He continued on this regimen until July 1998, when his metastatic disease progressed again. When the patient was enrolled in the current trial in September 1998, he had new paraneoplastic polycythemia (hemoglobin [Hb] level 18.4 g/dL) and PD in the lung and liver. By August 1999, the patient had a greater than 50% reduction in both liver and lung metastases and his polycythemia had resolved. In June 2000, while still on thalidomide and receiving warfarin anticoagulation for an intercurrent deep venous thrombosis, the patient presented with new-onset seizures. He had a solitary 1-cm hemorrhagic lesion in the left frontal lobe. Because the systemic liver and lung disease continued to regress, he was taken to surgery where the metastatic lesion was resected. Except for a 10-day interruption of treatment perioperatively, the patient continued on thalidomide. Thereafter, his lung and liver metastases remained stable until May 2001 (31 months after initiation of thalidomide), when he showed evidence of PD in the lung.

All 20 patients ultimately discontinued treatment. Reasons for discontinuation included PD in 18 patients (including two early deaths, described below), drug toxicity in 1 patient (Grade 3 peripheral sensory neuropathy after 18 months of therapy, although SD was maintained for 13 months after discontinuation), and mutual decision by the patient and treating physician in 1 patient after 17 months of therapy and before resection of residual stable lung metastases (this patient resumed thalidomide therapy after a 2-month interruption but developed new brain metastases 3 months later).

Time to PD and OS

Disease progressed in all 20 patients. The median PFS was 4.7 months (range, 0.7-31.3 months; Fig. 1). Fifteen patients died. Their median OS was 18.3 months (95% lower confidence bound 10.7 months; Fig. 2).

Figure 1.

Progression-free survival estimated by the Kaplan–Meier method.

Figure 2.

Overall survival estimated by the Kaplan–Meier method.

The variables initially considered for inclusion in the Cox regression model for survival were ZPS (< 2 vs. ≥ 2), number of organs involved (less than vs. three or more), Hb level, time from first diagnosis to initiation of thalidomide, time from first metastasis to initiation of thalidomide (Tmet), and number of treatments between first metastasis and initiation of thalidomide. In the final multivariate Cox model (Table 3), OS was significantly longer with higher Hb and with longer Tmet, but significantly shorter with higher number of previous treatments and multiple organ involvement. On average, patients with three or more organs involved had an approximately 14-fold higher risk of death than patients with two or fewer organs involved. Mortality increased by a factor of 2.5 for each additional previous therapy, decreased by a factor of 0.40 for each unit increase in Hb, and decreased by a factor of 0.92 for each month increase in Tmet. Due to the small sample size, however, each of these values was highly variable and any inferences based on such a small sample are suggestive, not confirmatory. For example, the relative risk of death associated with involvement of three or more organs had a 95% CI of 2.6–78.6.

Table 3. Cox Regression Model for Survival Time
VariableEstimate coefficient (SE)Relative riskP value
  • a

    Time from first metastasis to initiation of thalidomide.

No. of organs (≥ 3)2.65 (.87)14.20.002
No. of previous therapies.91 (.35)2.480.009
Hemoglobin−.92 (.27).40< 0.001
Tmeta−.082 (.028).920.004

Because all patients had received their previous treatment at M. D. Anderson, we were able to analyze in retrospect the time to progression (TTP) with IL-2 versus thalidomide therapy for 16 patients (Fig. 3). Of the 17 patients with previous IL-2 treatment, 16 had received IL-2 for treatment of metastatic disease (1 received it as adjuvant therapy). The median TTP with IL-2 treatment (6.6 months; 95% CI, 1.9–18 months) was not statistically different from the median TTP with thalidomide treatment (4.7 months; 95% CI, 0.7–31.3 months; P = 0.27). However, individual patients seem to have derived “clinical benefit” from thalidomide (i.e., a longer TTP on thalidomide than on IL-2, as seen for patients 2, 4, 9, 16, 17).

Figure 3.

Time to progression with IL-2 versus thalidomide treatment.

Toxicity and Dose Modifications

Table 4 summarizes the toxicities that were associated with thalidomide. Somnolence and constipation were seen in almost all patients. The use of “prophylactic” stool softeners helped to minimize constipation.

Table 4. Adverse Events Related to the Study Drug
ToxicityGrade 1Grade 2Grade 3Grade 4
  1. DVT/PE: deep venous thrombosis/pulmonary embolism.

Sensory neuropathy1041
Skin rash61
Dry skin42
Clinical hypothyroidism1

Thirteen patients remained on study beyond the first 3 months. At 3 months, 12 of these tolerated the maximum daily dose of 1200 mg; the last patient could only tolerate a daily dose of 800 mg because of sedation. Nine patients (45%) remained on study without toxicity or PD for 9 months or more. Six of these nine patients had PD after previous IL-2 therapy, two had PD after previous IFN-α/5-FU combination therapy, and one patient had no previous therapy. All nine patients tolerated the maximum daily dose of 1200 mg for at least 9 months.

Peripheral sensory neuropathy (mostly Grade 1) was also common. Delayed dose reductions were required for three patients because of peripheral sensory neuropathy. One patient's dose was reduced to 600 mg per day 15 months into the study and the doses for two other patients were reduced to 800 and 1000 mg per day, respectively, at 18 months.

The only patient who developed Grade 3 peripheral neuropathy had a history of diabetes mellitus and preexisting Grade 1 neuropathy at study entry. His symptoms improved significantly after discontinuation of thalidomide.

Three patients developed thromboembolic events (deep venous thrombosis in one patient and pulmonary embolism in two patients) believed to be related to thalidomide. Five patients developed new or worsening hypomagnesemia. One patient developed clinical hypothyroidism. Leukopenia, neutropenia, and lymphopenia were also observed but were limited in degree.

Two patients died on study. The first was a 65-year-old female with a 2-year history of mRCC who did not respond to IL-2 and IFN-α–based therapy or to two previous chemotherapy regimens. She was enrolled in this study with a ZPS of 3 and extensive PD in four organs (bone, adrenal gland, renal bed, and lung). She required oxygen supplementation at rest. Because of rapidly progressing disease and deterioration in her performance status, she was unable to return for her scheduled weekly clinic visits after the first 2 weeks of thalidomide therapy. She died of PD 3 weeks after initiation of thalidomide therapy.

The second patient who died on study was a 43-year-old male with a history of Stage III-A Hodgkin disease after total lymph node irradiation in 1974 and MOPP chemotherapy in 1979. He was first diagnosed with Grade 4 mRCC in October 1997 and was enrolled in the current trial in September 1998 having failed IL-2/IFN-α and fluoropyrimidine-based therapy. At study entry, he had multiple organ involvement (lung, liver, lymph node, renal bed, and malignant pleural effusion), adrenal insufficiency, and a ZPS of 3. After 3 months on thalidomide therapy (1200 mg per day), his disease was radiographically stable but his performance status had deteriorated. He was admitted to the hospital with sepsis, adrenal insufficiency, worsening pleural effusion, and malnutrition. Treatment with broad-spectrum parenteral antibiotics, stress-dose intravenous steroids, palliative thoracentesis, parenteral hyperalimentation, and reduction of his thalidomide dose to 600 mg per day resulted in brief symptomatic improvement, but the patient became comatose on the third hospital day and died a few hours later. A limited autopsy showed no evidence of intracranial metastasis, bleeding, or ischemic event in the brain. There was evidence of left lower lobe pneumonia positive for Klebsiella pneumoniae and Klebsiella bacteremia. K. pneumoniae was isolated subsequently from a blood culture that had been obtained on hospital admission. It was sensitive to the intravenous antibiotics used during hospitalization. This patient's death was attributed to sepsis, although the possible contribution of thalidomide could not be excluded, especially because his severe hypoalbuminemia could have resulted in higher free thalidomide bioavailability.23


In the current study, we demonstrated that thalidomide was relatively well tolerated by patients with advanced mRCC. Short-term toxicity (e.g., constipation and somnolence) was observed in almost all patients. Constipation was manageable from the outset with an aggressive bowel regimen. Somnolence, however, was a concern at the doses we used and raised two questions: 1) Is the optimal dose of thalidomide lower than its maximum tolerated dose? 2) Are less sedative thalidomide derivatives equally effective but better tolerated? Long-term treatment with thalidomide at the dose used in this study increases the risk of debilitating peripheral neuropathy that requires dose reduction.

Although objective responses in our trial were infrequent (2 PR; 10.5%) and occurred late, they lasted for long periods (16 and 31 months, respectively). This is in agreement with a study by Eisen et al.24 in which 3 of 18 patients responded late to low-dose thalidomide (100 mg per day) and with recent preliminary reports of thalidomide as a second-line therapy.25, 26 Conversely, others have reported no objective responses.27, 28 The interpretation of our results, however, is limited by the very small sample size, patient heterogeneity, the high rate of pretreatment, and variations in the thalidomide dose among patients.

Motzer et al.28 summarized the results of six studies of single-agent thalidomide in mRCC patients. Although no objective responses were seen in the MSKCC trial, the overall response rate was 6% (range, 0–17%), with 8 of 144 assessable patients achieving a PR. The results of the current study fall within this range. A retrospective assessment of risk factors does not support a difference between our and the MSKCC patient populations,28 at least as per the previously reported criteria.22 The factors that may account for the different response rates between the two trials include a higher rate of control of the primary organ in our trial (100% vs. 58%) and the presence of more indolent disease in our patient population. The latter is difficult to assess because information regarding the time from diagnosis or first metastasis to study entry is not provided in that trial.28 Although differences in the thalidomide dose could also play a role in the observed differences in response rates, this cannot be supported with the currently available information because objective responses have also been reported with very low thalidomide doses.24

A common finding in all clinical trials of thalidomide in RCC patients is the high rate of SD. However, the stability of mRCC may be part of the natural history of the disease.28 Therefore, the endpoint of SD should be viewed with skepticism and subjected to rigorous testing in randomized trials. This skepticism should also extend to putative antiangiogenic, antiinvasive, or antimetastatic agents, for which the response rate may not be the optimal primary endpoint. In the past, positive results in Phase II trials of agents such as IFN-gamma29, 30 or IFN-α31, 32 did not always translate into clinical benefit in Phase III trials.33 Together, our data show some antitumor activity of thalidomide in heavily pretreated patients with mRCC and warrant further evaluation of the role of thalidomide in RCC patients in controlled, randomized trials. We are currently studying the role of low-dose thalidomide in preventing/delaying metastases in a randomized adjuvant trial of patients at high risk of recurrence after nephrectomy. The potential benefit of thalidomide in extending TTP is also being studied in a Phase III trial by the Eastern Cooperative Oncology Group in combination with low-dose IFN-α.


The authors thank Linda Hicks for secretarial assistance.