Phase II study of transdermal estradiol in androgen-independent prostate carcinoma


  • Lisa B. Bland M.D.,

    1. Division of Urology, Oregon Health and Science University and Portland Veterans Affairs Medical Center, Portland, Oregon
    Search for more papers by this author
  • Mark Garzotto M.D.,

    1. Division of Urology, Oregon Health and Science University and Portland Veterans Affairs Medical Center, Portland, Oregon
    Search for more papers by this author
  • Thomas G. DeLoughery M.D.,

    1. Division of Hematology and Medical Oncology, Department of Medicine, Oregon Health and Science University, Portland, Oregon
    Search for more papers by this author
  • Christopher W. Ryan M.D.,

    1. Division of Hematology and Medical Oncology, Department of Medicine, Oregon Health and Science University, Portland, Oregon
    Search for more papers by this author
  • Kathryn G. Schuff M.D.,

    1. Division of Endocrinology and Metabolism, Department of Medicine, Oregon Health and Science University, Portland, Oregon
    Search for more papers by this author
  • Emily M. Wersinger M.P.H.,

    1. Division of Hematology and Medical Oncology, Department of Medicine, Oregon Health and Science University, Portland, Oregon
    Search for more papers by this author
  • Dianne Lemmon R.N.,

    1. Division of Hematology and Medical Oncology, Department of Medicine, Oregon Health and Science University, Portland, Oregon
    Search for more papers by this author
  • Tomasz M. Beer M.D.

    Corresponding author
    1. Division of Hematology and Medical Oncology, Department of Medicine, Oregon Health and Science University, Portland, Oregon
    • Division of Hematology and Medical Oncology, Department of Medicine, Oregon Health and Science University, Mail Code CR-145, 3181 SW Sam Jackson Park Road, Portland, OR 97239
    Search for more papers by this author
    • Fax: (503) 494-6197

  • The current study was performed at the Oregon Health Science University General Clinical Research Center.



Oral estrogen therapy has activity in patients with hormone-naive and androgen-independent prostate carcinoma (AIPC), but its utility is limited by the associated risk of thromboembolic toxicity. Parenteral administration may be safer as it avoids “first pass” liver exposure to estrogen. The authors tested the safety and efficacy of transdermal estradiol (TDE), as well as the effect of therapy on hot flashes, sex hormones, the procoagulant cascade, and bone turnover in patients with AIPC.


Patients with prostate carcinoma progressing after primary hormonal therapy received TDE 0.6 mg per 24 hours (administered as six 0.1 mg per 24-hour patches replaced every 7 days). Serum prostate-specific antigen (PSA) and hormone levels, coagulation factors, markers of bone turnover, bone density measurements, and a hot flash diary were collected at regular intervals.


Three of 24 patients (12.5%; 95% confidence interval [CI], 0–26%) had a confirmed PSA reduction > 50%. The Kaplan–Meier estimate of median time to disease progression was 12 weeks (95% CI, 4.6–19.4 weeks). Toxicity was modest and no thromboembolic complications occurred. The mean (±95% CI) serum estradiol level increased from 17.2 pg.mL (range, 14.8–19.6 pg/mL) to 460.7 pg/mL (range, 334.6–586.7 pg/mL). The total testosterone level remained stable in the anorchid range during treatment, but the free testosterone level decreased as a result of increased sex hormone binding globulin. No change in factor VIII activity, F 1.2, or resistance to activated protein C was observed, whereas a modest decrease in the protein S level was observed.


In patients with APIC, TDE was well tolerated and produced a modest response rate, but was not associated with thromboembolic complications or clinically important changes in several coagulation factors. Cancer 2005. © 2005 American Cancer Society.

Androgen deprivation therapy (ADT) has been shown to be effective in improving survival when administered relatively early in the course of prostate carcinoma.1, 2 However, this survival advantage may be offset by a reduction in quality of life due to the adverse effects of ADT that include loss of libido and erectile function, fatigue, depression, anemia, osteoporosis, obesity, hot flashes. and others. Thus, there is a pressing need to develop less toxic forms of ADT.

Diethylstilbesterol (DES), an oral estrogen, has long been known to be active in hormone-naive prostate carcinoma and recently has been reported to have activity in androgen-independent prostate carcinoma (AIPC).3 The mechanism of DES activity in this setting is uncertain and could be a result of direct inhibitory effects on tumor cells or DES-induced changes in circulating hormone concentrations that include reduction of levels of leutenizing hormone, follicle-stimulating hormone, free and total testosterone, estradiol, and dehydroepiandrosterone sulfate and increases in the levels of sex hormone binding globulin (SHBG) and cortisol.4, 5

The utility of oral estrogen therapy is limited by the associated risk of thromboembolism.6–9 Mechanisms of increased thrombotic risk with oral estrogens are believed to include induction of resistance to the natural anticoagulant protein C, reduced protein S, a cofactor for protein C, and increased concentrations of the procoagulant factor VIII.10–13

Unlike oral estrogens, parenterally administered estrogens do not increase liver synthesis of most proteins and, therefore, may be less prothrombotic.14, 15 In women, transdermal estrogens do not lead to acquired protein C resistance, decreases in protein S, nor increases in factor VIII and F1.2, which have been observed with oral estrogen therapy.16, 17 As a result, in women receiving estrogen replacement therapy, clinical thrombotic risk is increased with oral but not with transdermal therapy.18

Parenteral estrogens have been studied in the treatment of hormone-naive prostate carcinoma but not in AIPC. Intramuscular polyoestradiol phosphate performed similarly to conventional ADT with regard to both efficacy and cardiovascular mortality rates.19–21 In a pilot study of 20 men, transdermal estradiol (TDE) therapy produced anorchid testosterone concentrations and PSA responses in all 20 patients without thromboembolic complications.22

In the current study, we evaluated TDE safety, efficacy, and impact on hot flashes, coagulation factors, and bone turnover in patients with AIPC.



Eligibility criteria included histologically or cytologically confirmed prostate carcinoma progressing despite standard hormonal therapy including antiandrogen withdrawal (6 weeks for bicalutamide or nilutamide, 4 weeks for flutamide) in patients who received an antiandrogen in addition to a gonadotropin-releasing hormone (GnRH) agonist or orchiectomy. Disease progression was defined as new or increased measurable or evaluable disease and/or disease progression by serum prostate-specific antigen (PSA) level as described in the consensus criteria.23 Additional entry criteria were Eastern Cooperative Oncology Group performance status score ≤ 2, serum testosterone level ≤ 50 ng/dL, serum PSA level ≥ 5 ng/mL, and age ≥ 18 years. Patients were excluded if they had received previous chemotherapy, DES or another estrogen, or PC-SPES for the treatment of prostate carcinoma. In addition, patients were excluded if they had major surgery within 4 weeks, serious medical illnesses, New York Heart Association class III or IV congestive heart failure (CHF), unstable angina, myocardial infarction within 6 months, acute deep venous thrombosis, acute pulmonary embolism, or active second malignancy other than nonmelanoma skin carcinoma.

The protocol was approved by the institutional review boards of the Oregon Health and Science University and the Portland VA Medical Center (Portland, OR). Written informed consent was obtained from all patients.


Standard prophylactic breast irradiation was offered to all patients before therapy. Six 7.6-mg transdermal estradiol patches (Climara, 0.1 mg per 24 hours; Berlex, Montville, NJ) were applied weekly for ≤ 12 months or until disease progression occurred, unacceptable toxicity, or a patient request to withdraw from the study. PSA progression was defined by consensus criteria.23 GnRH agonist therapy was not continued during the study period. Initially, generic (Mylan Laboratories, Inc., Canonsburg, PA) patches were used briefly and discontinued in favor of Climara due to poor skin adherence. This dose of TDE was modeled after the dose reported by Ocrim et al.22


Pretreatment evaluation included a history and physical examination, radionuclide bone scan, computed tomography scan of the abdomen and pelvis, bone mineral density dual energy x-ray absorptiometry (DEXA) scan of the hip and lumbar spine, complete blood count with automated differential (CBC), and monitoring of serum PSA levels, serum chemistries, and urine n-telopeptide levels. Hormones tested included total testosterone, estradiol, and SHBG. Free testosterone was calculated from the method of Vermeulen et al.24 using testosterone, albumin, and SHBG concentrations. Coagulation tests included factor VIII activity, protein S activity, resistance to activated protein C (APC resistance), and prothrombin activation measured by F1.2. As previously described, a 7-day hot flash diary25 was collected.

Adverse events were recorded monthly. Serum PSA and chemistries and a CBC were measured monthly. Measurable disease was reexamined every 8 weeks. Radionuclide bone scans were repeated as clinically indicated. Bone density scans were planned after 52 weeks of study treatment. Hormone and coagulation factor measurements were repeated after 4 and 8 weeks of therapy and for a subset of patients, at the time of discontinuation of therapy. Urine n-telopeptide levels were measured after 8 weeks of therapy. Hot flash diaries were collected after 4 and 8 weeks.

Statistical Considerations

The primary end point of the current study was PSA response rate as defined by consensus criteria.23 A sample size of 25 was chosen based on a 2-stage study design with the goal of rejecting the regimen if the response rate was < 10% and recommending the regimen for further study if the response rate was ≥ 30%. The minmax design with α = 0.05 and β = 0.20 (power of 80%) would reject the regimen if < 2 of 15 patients responded in the first stage. It would recommend the regimen for further study if ≥ 5 of 25 patients responded in the second stage.26 Enrollment continued during the evaluation of the first 15 patients.

Hot flash scores (HFS) were calculated as described by Sloan et al.25 For assays that yielded a result below the lower limit of detection (i.e., testosterone and estradiol), the lower limit of detection was used for analyses of means. Means were compared using t tests, whereas the Pearson chi-square test was used for categorical measures. Repeated-measures analysis of variance was used to analyze serially measured variables. The Kaplan–Meier method was used to determine the time to disease progression (TTP) with a log-rank statistical test of significance where comparisons were made.



Twenty-five patients were recruited. One patient was deemed ineligible after an abnormal stress electrocardiogram before beginning treatment. The remaining 24 patients were enrolled and provide the basis for this analysis. Baseline patient characteristics are summarized in Table 1. The median age of the patients was 75 years. Seventy-five percent had radiographically identified metastases and 25% had PSA-only disease. The group was largely composed of patients with good performance status scores. The median serum PSA level was 22.3 ng/mL (range, 6.6–560.3 ng/mL). Ten patients reported some degree of difficulty with patch adherence. Two patients completed the planned 12 months of treatment, 21 withdrew due to disease progression, and 1 withdrew due to toxicity.

Table 1. Patient Characteristics
  • ECOG: Eastern Cooperative Oncology Group; PSA: prostate-specific antigen; GnRH: gonadotropin-releasing hormone.

  • a

    One patient had an orchiectomy after 5 years of receiving a gonadotropin-releasing hormone agonist.

No. of patients24
Median age (range)75 (49–91 yrs)
ECOG performance status 
Median PSA level (range)22.3 (6.6–560.3 ng/mL)
Median alkaline phosphatase level (range)74 (37–488 U/L)
Median hemoglobin level (range)13.5 (11–16.3 g/dL)
Site of metastases 
 Bone only50%
 Lymph only13%
 Bone and lymph nodes13%
 None (PSA only)25%
Previous therapy 
 GnRH agonista75%
 Previous antiandrogen use79%
 ≥ 2 previous hormonal therapies79%


Treatment was generally well tolerated. Observed toxicities are shown in Table 2. There were no deaths or thromboembolic events. One patient with a previous history of coronary artery disease and intermittent angina suffered an episode of angina. Coronary angiography revealed no new or acute lesions and the patient was stabilized medically and remained in the study. Another patient with a previous history of intermittent CHF suffered an exacerbation of CHF. This patient was also treated medically and myocardial infarction was ruled out, but study treatment was discontinued. In both cases, independent consulting cardiologists attributed these events to fluctuations in existing cardiac disease and did not diagnose new thromboembolic events. One patient developed acute cholecystitis and had a cholecystectomy. Grade 1 anemia, hyponatremia, and hypocalcemia were the most common laboratory abnormalities. There were no clinical cases of deep venous thromboses or pulmonary embolism.

Table 2. Toxicities that were at Least Possibly Related to Drug Treatmenta
ToxicityGrade 1 (%)Grade 2 (%)Grade 3 (%)Grade 4 (%)Any (%)
  • a

    All toxicities ≥ Grade 2 are shown. Grade 1 toxicities are shown if they occurred in >1 patient (n = 24).

 Dermatitis (patch site)50.000050.0
 Dyspnea on exertion8.34.20012.5
 Mood swings8.30008.3
 Pleural effusion004.204.2
 Congestive heart failure004.204.2
 Thromboembolic events00000

Thirteen patients chose to receive prophylactic breast irradiation. Overall, mild to moderate gynecomastia and gynecodynia were common. The incidence rates of any breast complaints were 69.2% and 81.8% for men who received and did not receive prophylactic radiotherapy (P = 0.48). The incidence of Grade 2 breast toxicity in men who received prophylactic radiation was 15.4% and it was 27.3% in men who did not (P = 0.48). Grade 1 dermatitis at the patch site was reported by 50% of patients.

Response and Time to Disease Progression

Three of 24 patients (12.5%; 95% confidence interval [CI], 0–26%) had a confirmed PSA reduction > 50%. The Kaplan–Meier estimate of median TTP was 12 weeks (95% CI, 4.6–19.4 weeks). Disease progression was documented solely by an increasing PSA level in 18 patients, by imaging (new bone lesions) in 1 patient, and by clinical evaluation (clinical deterioration) in 2 patients. Five patients had measurable disease (lymph node metastases) at baseline. Two of these five patients progressed by measurable or evaluable disease as determined by imaging studies, two remained stable, and one had insufficient follow-up.

Although confirmed PSA responses were infrequent and the median TTP was brief in the entire study group, potentially clinically useful disease stabilization was observed in a subset of patients. Thirty-eight percent of patients (n = 9) were disease progression free at 20 weeks, 25% (n = 6) were disease progression free at 28 weeks, and 8% (n = 2) remained disease progression free at the end of 52 weeks. Confirmed PSA reduction > 25% was observed in 7 patients (29%; 95% CI, 11–48%).

PSA responses (> 50%) were less frequent in men with distant metastases (5.6% [1 of 18]), compared with men without distant metastases (33% [2 of 6]; P = 0.08). A confirmed 25% PSA reduction occurred in 3 of 18 (16.7%) patients with metastases and in 4 of 6 (67%) patients without metastases (P = 0.02). Considering the small patient sample, these hypothesis-generating exploratory analyses should be viewed with caution. Previous therapy, estradiol levels, and free testosterone levels were not predictive of response or TTP.


Results of hormone analyses are shown in Table 3. Estradiol concentrations at baseline and after 4 and 8 weeks were available for 20 patients. Elevated estradiol levels were reached within 4 weeks and maintained at 8 weeks in all but 2 patients. However, concentrations varied considerably within and between subjects and ranged from 70 to 1045 pg/mL in patients with elevated levels. Estradiol ranged from 193 to 646 pg/mL at 4 weeks and from 153 to 527 pg/mL at 8 weeks in the 3 responders. We were unable to detect a relation between reported difficulty with patch adherence and estradiol concentrations.

Table 3. Levels of Hormonesa
Hormone levelsBaseline4 Weeks8 WeeksOff studybP valuec
  • SHBG: sex hormone binding globulin.

  • a

    Levels shown are mean values ± the 95% confidence interval values. The lower limit of detection was assigned to values reported to be below the lower limit of detection.

  • b

    N = 8 for estradiol and total testosterone; N = 7 for SHBG and free testosterone.

  • c

    Repeated-measures analysis of variance comparing baseline, Week 4, and Week 8.

Estradiol (pg/mL)17.2 ± 2.4460.7 ± 126.1370.8 ± 114.6498.4 ± 243.1< 0.0001
SHBG (μg/dL)1.0 ± 0.22.7 ± 0.62.2 ± 0.42.0 ± 0.7< 0.0001
Total testosterone (ng/dL)11.0 ± 2.510.1 ± 1.611.3 ± 2.611.8 ± 4.00.41
Free testosterone (pg/mL)200.2 ± 52.3101.6 ± 25.1117.7 ± 30.0139.6 ± 47.40.0005

A significant increase in SHBG concentrations was observed within 4 weeks of TDE therapy. Total serum testosterone level did not change and remained in the anorchid range in all patients during 8 weeks of treatment. As a result of the increase in SHBG levels, free testosterone levels significantly decreased after 4 and 8 weeks (P = 0.0005).

Longer-term hormone data are available for a subset of patients who had the hormones assayed again at the time therapy was discontinued (12–52 weeks). As is illustrated in Table 3, testosterone levels remained in the anorchid range in all patients, seven of whom had not received an orchiectomy. Estradiol, SHBG, and free testosterone concentrations also differed little from the value obtained at week 8.

Coagulation Factors

Results of coagulation factor analysis are shown in Table 4. Treatment did not affect factor VIII activity, F1.2, or resistance to APC and was associated with a modest reduction in the level ofprotein S. Although a statistically significant reduction in protein S levels was observed, it is not clear whether this reduction was sufficient to be clinically significant considering that protein S levels remained in the normal range for all patients.

Table 4. Coagulation Measures
CharacteristicsBaseline4 Weeks8 WeeksOff studybP valuec
  • APC: activated protein C.

  • a Values shown are mean values ± 95% confidence interval values.

  • b

    N = 7 for F1.2, N = 6 for all others.

  • c

    Repeated-measures analysis of variance comparing baseline, Week 4, and Week 8.

Factor VIII (U/mL)234 ± 30237 ± 25222 ± 48187 ± 820.74
Protein S (U/mL)110 ± 1493 ± 992 ± 1284 ± 13< 0.001
F1.2 (ng/mL)1.31 ± 0.581.19 ± 0.391.11 ± 0.460.92 ± 0.210.82
APC resistance ratio2.10 ± 0.232.12 ± 0.232.27 ± 0.072.13 ± 0.250.35

Hot Flashes

Eleven of 24 (46%) patients reported hot flashes at study entry and 10 of these patients completed the pretreatment hot flash diary. Eight of these 10 patients provided hot flash diaries at Week 4 and 5 at both Weeks 4 and 8. We observed a trend towards reduction in hot flashes at Week 4 (HFS reduced 57%; P = 0.08) and a significant reduction (HFS reduced 77%; P = 0.02) at Week 8. One patient who reported no hot flashes before entry complained of mild hot flashes during treatment.

Bone Turnover Markers

Urine n-telopeptide levels were examined at baseline and after 8 weeks of therapy in 16 patients. When all the patients were included in the analysis, there was no significant difference in urine n-telopeptide levels (P = 0.78). One patient was an outlier. His baseline n-telopeptide levels were 32 nM bone collagen equivalents (BCE)/mM creatinine and increased to 518 nM BCE/mM creatinine at Week 8. When this outlier was excluded from the analysis, the mean value decreased from 116.8 (range, 74.2–159.4) nM BCE/mM creatinine at baseline to 94.5 (range, 58.1–130.9) nM BCE/mM creatinine after 8 weeks (P = 0.04). Bone density data are available for both patients who continued to receive treatment for the full 52 weeks. The mean bone density increased by 5.8% and 13.2% in the lumbar spine and 1.8% and 2.1% at the hip, respectively, in these 2 patients.


TDE therapy generally was well tolerated in patients with AIPC. The overall activity of this regimen was low, as often occurs with second-line hormonal maneuvers in patients with AIPC. Lower PSA reductions (> 25%) occurred more commonly but such changes are of uncertain significance and may be of interest for hypothesis generation. A small subset of patients had modest disease control with TDE therapy.

As suggested by Hedlund et al.21 and Ockrim et al.,22 the initial hormonal management of metastatic prostate carcinoma may be a more useful setting for parenteral estrogen therapy. In this setting, parenteral estradiol may offer important advantages over conventional ADT, including fewer hot flashes (and possibly improved sleep quality) and less osteoporosis. Furthermore, ADT is being used earlier in the course of disease including during the adjuvant setting, during early disease progression after a curative attempt, and as primary therapy for patients with localized disease.27 The availability of low-toxicity ADT likely would translate into meaningful quality of life improvements in these asymptomatic patient populations. Both the practical observations and the correlative results of the current study will aid in planning future studies of TDE as initial therapy.

Although none of the patients discontinued therapy due to breast toxicity, this adverse effect was common. It is noteworthy that there was relatively little difference in the breast toxicity reported by patients who received prophylactic breast irradiation when compared with those who did not. The absolute benefit of prophylactic X-ray therapy in this setting cannot be determined outside the context of a randomized controlled trial.

Our results also suggest that elements of the coagulation cascade that are impacted adversely by oral estrogens were not altered substantially with transdermal therapy. Consistent with these and previous findings in patients treated with parenteral estrogen, we saw no thromboembolic complications. Thus, our results add to the growing body of evidence that parenteral administration of estrogens may circumvent the thromboembolic toxicity long associated with oral therapy.

Our data regarding hot flashes and bone density are limited by small patient samples, but are consistent with the hypothesis that TDE reduces hot flashes and preserves bone density in this patient population.

With our transdermal delivery system, unlike Ockrim et al.,22 we reported considerable variation in serum estradiol concentrations. We detected no clear relation between estradiol concentrations and cancer response. Only a few patients received testosterone monitoring after 8 weeks. None of these patients had elevations in total testosterone levels > 30 ng/dL, suggesting that this regimen is sufficient to maintain androgen deprivation. As previously observed with DES, we observed a significant increase in SHBG concentrations that, in turn, produced a reduction in free testosterone levels, but we could not detect a relation between free testosterone levels and response status. Recent data from Chen et al.28 using matched androgen-sensitive and androgen-insensitive prostate carcinoma xenografts show that low levels of circulating testosterone may be a significant factor in hormone-refractory prostate carcinoma progression due to amplification of androgen receptor levels in resistant clones. Thus, any reduction in free testosterone levels may offer therapeutic benefits in subsets of patients who have resistant prostate carcinoma clones with androgen receptor amplifications. Further study of this concept is needed.

Finally, our results suggest that even in studies of second-line hormonal therapy, patients with PSA-only AIPC probably should be studied separately from those with established metastases. Both with regard to response rate and TTP, patients with PSA-only AIPC tended to have better outcomes than those with metastases, although the number of patients studied was small. The outcomes of studies of second-line hormonal therapy are likely to be impacted by patient selection.

In conclusion, this regimen of transdermal estrogen was well tolerated in AIPC and was associated with a modest PSA response rate. Treatment was not associated with thromboembolic complications, did not affect factor VIII activity, resistance to APC or F1.2, and was associated with a modest reduction in protein S that is unlikely to be clinically significant. Hot flashes were reduced and in 2 patients who continued to receive therapy for 52 weeks, bone density improved. Based on these results further study in patients with androgen-sensitive prostate carcinoma and PSA-only AIPC is warranted.