SEARCH

SEARCH BY CITATION

Keywords:

  • Prostate;
  • cancer;
  • stilboestrol;
  • salvage therapy;
  • outcome

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. References
  8. Authors

Objective To investigate the efficacy of low-dose stilboestrol (SB) with hydrocortisone in patients with advanced prostate cancer refractory to androgen suppression.

Patients and methods Thirty-four consecutive patients (median age 70 years, range 51–83) with metastatic disease who progressed on hormone therapy, as shown by recurrent/worsening symptoms and an increase in prostate-specific antigen (PSA) level, were recruited and discontinued hormonal treatment before starting SB. Patients received SB (1 mg/day) combined with hydrocortisone (40 mg/day). In an attempt to reduce the incidence of thrombo-embolic events, aspirin (75 mg/day) was also added.

Results Stilboestrol was the second-line treatment in 19 patients and the third or fourth in 15. The median (range) duration of treatment with SB was 5 (0.5–21) months and the median follow-up 7.5 months, with 18 patients still alive and 14 still on treatment. Of 29 symptomatic patients, 24 had symptomatic improvement and five had no clear benefit; the median duration of benefit was 6 (2–21) months. The PSA level decreased by 0–50% in six patients, by 50–90% in 13 and by > 90% in eight, while there was symptomatic improvement in these three categories in five, 11 and seven patients, respectively. The median times to PSA nadir and progression were 4 and 6 months, respectively. Some thrombo-embolic events and fluid retention occurred but overall the treatment was well tolerated.

Conclusion Low-dose SB with hydrocortisone is effective in refractory prostate cancer, although there is some toxicity. Randomized studies against hydrocortisone or SB alone are needed to establish the cost/benefit ratio of this combination.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. References
  8. Authors

Patients with metastatic prostate cancer who progress during androgen suppression provide a therapeutic challenge. Current therapy in this setting is purely palliative and includes corticosteroids, analgesics, EBRT and intravenous radioactive isotopes such as strontium. Chemotherapy has little benefit and corticosteroids are the most commonly used systemic treatment.

Oestrogens are a cheap and effective treatment for metastatic prostate cancer. Higher doses of oestrogens, e.g. 3–5 mg of stilboestrol (SB) per day, were as effective as androgen ablation [1,2] and superior to flutamide [3] as first-line therapy for metastatic disease. SB at 3 mg/day is also active as second-line therapy after hormonal escape of disease [4], but at the doses used in these studies, there was considerable morbidity and mortality from thrombotic and cardiovascular events. Indeed, in the earlier Veteran Administration Cooperative Urologic Research Group studies, gains in survival from prostate cancer were offset by more cardiovascular deaths [1]. The incidence of thrombo-embolic episodes increased with increasing doses of oestrogens which, in some of the earlier studies, were as high as 75 mg of SB a day [5]. Low-dose SB (leqslant R: less-than-or-eq, slant 1 mg/day) was associated with fewer thrombo-embolic and ischaemic events, and retained activity as primary therapy [1,6,7].

Additionally, although the vascular complications of oestrogen therapy may not justify their use in the first-line treatment of prostate cancer, given the more favourable toxicity profile of LHRH agonists, low-dose oestrogens may still be beneficial in disease refractory to androgen deprivation [8]. We investigated the efficacy of low-dose SB in hormone-refractory advanced prostate cancer. Previous studies showed that with SB doses of 1 mg/day, only ≈ 25% of patients maintained castrate levels of testosterone [7]. Therefore, as an additional therapeutic manoeuvre, SB was combined with low-dose corticosteroids to suppress adrenal androgen synthesis [9,10]. Low-dose aspirin was added as prophylaxis against cardiovascular events.

Patients and methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. References
  8. Authors

The study included 34 patients (median age 70 years, range 51–83), all with metastatic prostate cancer refractory to androgen suppression, who were treated with SB (1 mg/day) with cortisone (25 mg) or hydrocortisone (20 mg) twice daily and aspirin (75 mg daily). Patients had received at least one line of hormonal therapy including androgen ablation, either chemically or surgically.

Twenty-six patients had documented increasing serum levels of PSA despite hormonal therapy. The lowest median level of PSA during their previous hormonal treatment was 5 ng/mL (range 0.1–390). This had risen by 58-fold to a median (range) of 291 (10–10 500) ng/mL by the time SB was started, with 14 patients having a PSA level before SB of > 500 ng/mL, and eight of these being > 1000 ng/mL. In addition, 29 patients had documented worsening of symptoms or appearance of new symptoms; these included 18 with bone pain, two of whom had severe malaise, and another with transfusion-dependent anaemia secondary to marrow failure. Five patients had haematuria, causing clot retention in one and requiring repeated transfusion in another. Four patients had severe malaise as the main symptom, two of whom also complained of LUTS not secondary to infection, and one had malignant ascites. A fifth patient had severe oedema of the lower extremities, probably secondary to lymphatic/venous obstruction by tumour.

Previous hormonal therapy was discontinued before or at the time of starting SB. Thirteen patients stopped their previous treatment a median (range) of 1 (0.25–4) month before starting SB, while 13 stopped at the time of starting SB. Seven patients had previously undergone orchidectomy, of whom six were also on other agents which were discontinued before SB. Five of the 13 patients who stopped at the time of starting SB had been on flutamide either as a single agent or combined with LHRH agonists. Because of the severity of disease progression it was felt unsafe to delay starting treatment, as in the one patient who was left off treatment to see if he responded to flutamide withdrawal there had been a rapid deterioration in symptoms within days of stopping flutamide, and which improved within days of starting SB.

The response to SB included subjective improvement in symptoms and/or a decrease in PSA level. Progressive disease during SB included subjective deterioration in symptoms, or need for other treatment such as radiotherapy. A sustained rise in PSA (for at least two values), or a rise of > 20 ng/mL in patients with normal levels (< 4.0 ng/mL), also constituted progressive disease. The progression-free interval was calculated from the start of SB to the first documented symptom progression or any intervention such as new hormone treatment or radiotherapy. If these were not present, then the date of the first significant PSA elevation according to the above criteria was taken as the progression date. Survival was also estimated from the start of SB.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. References
  8. Authors

Table 1 shows the patient characteristics on entry to the study. The median (range) duration from diagnosis of prostate cancer to the start of SB was 23.5 (3–110) months and the median follow-up from the start of SB was 7.5 (1–21) months, during which the median duration of SB therapy was 5 (0.5–21) months. At the time of analysis, 18 of 34 patients were alive (53%), and 14 of these were still on treatment; 32 patients were evaluable for disease progression and of these, 12 (38%) were progression-free at analysis.

Table 1.  Patient characteristics
Characteristic No. of patients
  • *

    Some patients had more than one symptom.

No. of previous hormonal treatments
 119
 212
 33
Main symptoms at start of SB*
 Bone pain20
 Haematuria5
 Malaise6
 Anaemia1
 LUTS2
 Leg oedema1
 Symptoms unspecified6
PSA level (ng/mL) at start of SB
 < 1008
 101–50012
 > 50014

Of the 34 patients, 29 were evaluable for symptomatic change with treatment; of these 29, 25 reported an improvement in symptoms, with the remaining four reporting no change or worsening symptoms. Allowing for some patients reporting the relief of more than one symptom, 19 reported relief or amelioration of pain, eight reported improved well-being, three had a reduced transfusion requirement, haematuria resolved in three and one who presented with severe ascites had complete resolution. The median duration of symptom response was 6 (2–21) months.

Serum PSA levels were the main objective criterion for tumour response; the median (range) PSA level before treatment was 292 (10–10 500) ng/mL and the median PSA value at the nadir was 66 (0.1–6110) ng/mL, giving a median decrease in PSA of 84%. The median time from the start of treatment to PSA nadir was 4 (2–8) months and the median duration of PSA response was 6 (0–21) months.

Table 2 shows the absolute reduction in PSA level in three categories of patients according to their PSA level before treatment. There was a similar proportion of patients with a > 90% decrease in PSA in the first group (baseline PSA < 100 ng/mL) as in the other two categories. This group had the largest relative decrease in PSA, with a 14-fold median reduction, compared with 3.6-fold and 3.8-fold for the second and third categories, respectively. Paradoxically, the time to symptom progression in the first category was at least half that in the other two; this was statistically significant when category 1 was compared with category 3 (P = 0.01, Student's t-test).

Table 2.  Absolute decrease in PSA level in relation to baseline level, and the correlation with the proportion and duration of response
  Baseline PSA category (ng/mL) Variable < 100
No. of patients (no. evaluable)8 (7)12 (10)14 (10)
Median (sem) PSA (ng/mL)
 At baseline36 (9)224 (31)1237 (713)
 At nadir2.5 (5)63 (16)327 (567)
No. with > 90% PSA decline3/82/103/10
No. with symptom response6/89/109/10
Median (range) time to symptom progression8 (3-12)5.5 (2–21)8 (3–12)

Interestingly, there was a high proportion of patients reporting symptom relief in all three categories. Indeed, taking all categories together, almost all patients (seven of eight) with a > 90% decrease in PSA level reported symptom improvement, compared with 17 of 19 with a lower magnitude of PSA reduction.

Table 3 shows the relationship between the magnitude of PSA reduction and relief of symptoms. Twenty-four of 29 evaluable patients had significant symptomatic improvement, giving an overall response rate of 83%. Of 29 patients, 22 had a > 50% decrease in PSA level, giving a PSA response rate of 76%. Allowing for the relatively few patients in each group, the frequency of symptomatic response was similar in the three arbitrarily defined PSA response groups. There was a trend for the subgroup with the highest decrease in PSA level (> 90%) to have the longest progression-free interval, but this was not quite significant (P = 0.07).

Table 3.  The correlation between the decrease in PSA level and the symptomatic progression-free interval; seven patients were not evaluable
% decrease in PSA
Variable 0–50 51–90 > 90 Total
No. of patients613827
Proportion with symptom response511723
Symptomatic progression-free interval (months)2,2+,5,8,9,16 0,0,2+,3,3+,5,5, 6,6+,7,7,7,82+,5,6,6+,9+, 9+,10,21+
Median progression-free interval (months)6.557.5

Fluid retention with dependent oedema, often responding poorly to diuretics, was the commonest toxicity encountered (Table 4). This was sufficiently severe to precipitate heart failure in three patients, although in one of these heart failure was manifest after an acute myocardial infarct. Fluid retention often diminished when medication was stopped. There were two episodes of deep vein thrombosis, but the contribution of SB to these events is difficult to establish given the background of active pelvic malignancy in both of these patients.

Table 4.  Toxicity from stilboestrol and hydrocortisone
Description No. of events*
  • *

    Some patients reported more than one toxicity;

  • †one of these is the same patient who had a myocardial infarct.

Fluid retention11
Breast symptoms5
Cardiac failure3
Deep vein thrombosis2
Myocardial infarction1
Cerebrovascular event1
Upper GI bleed1
Hot flashes1
Malaise1
No toxicity7
Not evaluable3

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. References
  8. Authors

Earlier studies showed that oestrogens were equally as effective as androgen suppression and superior to antiandrogens such as flutamide in the first-line systemic management of metastatic prostate cancer. However, they became unpopular because of the cardiovascular and thrombo-embolic side-effects associated with the doses used [1–3]. When lower doses were used, this appeared to ameliorate the toxicity profile while preserving the antitumour activity [1]. It was this activity which we investigated in the context of disease resistant to androgen ablation.

Several mechanisms of action have been proposed for the effect of oestrogens on prostate cancer cells. In the pioneering days of hormone therapy for prostate cancer, oestrogens were thought to act through suppressing the pituitary secretion of LH [11]. Oestrogen doses as low as 1 mg/day of SB reduced circulating testosterone to castrate levels in ≈ 25% of treated patients with advanced prostate cancer [7]. Work reported by Plowman et al.[9] and Dowsett et al.[10] showed that as little as 30–40 mg/day of hydrocortisone significantly suppressed adrenal androgen synthesis. We therefore added 40 mg/day of oral hydrocortisone to SB in an attempt to produce more universal adrenal androgen blockade than would be produced with low-dose SB alone. More recently, direct inhibitory effects of oestrogens on the ability of prostatic cells to metabolize testosterone have been reported [12]. Oestrogens were also shown to directly influence cellular metabolic pathways, as shown by changes in lactate dehydrogenase isoenzymes in prostate cancer cells in vitro, an effect which was more evident in malignant than in benign prostatic cells [13]. Additionally, oestrogenic antiproliferative effects were seen in prostate cancer cells, which may be mediated through mechanisms such as increased levels of TGF-β1 [14] and up-regulation of cell-to-cell adhesion molecules [15]. In the setting of androgen-resistant disease, these mechanisms may partly explain how oestrogens could circumvent androgen insensitivity in vivo. Given these androgen-independent effects, the contribution of hydrocortisone to the antitumour effect of this combination, through its suppression of adrenal androgens, remains unclear and requires evaluation in a randomized study of this combination vs SB alone.

There was a worthwhile symptomatic improvement in most of the present patients and a large proportion of them had significant decreases in plasma PSA levels. However, a substantial decrease in PSA level was not necessary for a symptomatic response. There was a tendency for patients with the largest reductions in PSA level to have more durable symptom control (Table 3), although this was not statistically significant because there were too few patients. Interestingly, patients with the highest baseline PSA level (> 500 ng/mL) had a significantly longer time to progression (Table 2). This could be a chance finding, given the few patients, but could also suggest that patients with highly secretory tumours, and possibly more differentiated ones, may benefit most from this form of treatment and hence progress more slowly.

The protective effect of low-dose aspirin against thrombotic events is difficult to evaluate. Previous series quoted cardiovascular and thrombo-embolic incidence rates of 35% when 3 mg/day SB was used [3], and 8% after 1 mg/day [7], whereas in the present series there were six events in 34 treated patients (18%). This may suggest that the toxicity from low-dose SB was not ameliorated by low-dose aspirin, but again there were few patients. Fluid retention may have been aggravated by the mineralocorticoid properties of hydrocortisone, and it may be better to use an equivalent dose of prednisolone or dexamethasone, which have less intrinsic mineralocorticoid activity, or to reduce the dose of hydrocortisone further. Gynaecomastia and breast tenderness are common with oestrogen therapy, even at low doses. Breast enlargement may be prevented by a single fraction of radiotherapy before starting SB [16]. The contribution of single-agent low-dose SB to the overall efficacy of this combination also requires study in a randomized prospective trial.

In summary, this phase II study showed that the combination of SB and hydrocortisone was highly effective and had an acceptable side-effect profile as a salvage regimen, although probably not as first-line therapy. This regimen is currently our standard second-line therapy for patients who are no longer sensitive to androgen ablation. However, it may yet be possible to adapt such an inexpensive and effective regimen for use at an earlier stage in advanced prostate cancer. Thus, with increasing evidence in favour of the efficacy of intermittent hormone therapy in the primary treatment of advanced disease [17], the use of low-dose SB for short periods, possibly combined with subcutaneous heparin, separated by intervals off-treatment, could be associated with even less toxicity than has so far been reported.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. References
  8. Authors
  • 1
    Byar DP. The Veteran Administration Cooperative Urological Research Group's Studies of cancer of the prostate. Cancer 1973; 32: 112630
  • 2
    Leuprolide Study Group. Leuprolide versus diethylstilboestrol for metastatic prostate cancer. N Engl J Med 1984; 311: 12816
  • 3
    Chang A, Yeap B, Davis T et al. Double-blind, randomized study of primary hormonal treatment of stage D2 prostate carcinoma: flutamide versus diethylstilboestrol. J Clin Oncol 1996; 14: 22507
  • 4
    Dearnaley D, Capp A, Gadd J, Huddart R. Stilboestrol in the management of refractory prostate cancer. Br J Cancer 1997; 76: 22
  • 5
    Shearer R, Hendry W, Somerville IF, Fergusson J. Plasma testosterone: an accurate monitor of hormone treatment in prostate cancer. Br J Urol 1973; 45: 66877
  • 6
    Robinson MRG. A further analysis of European Organisation for Research and Treatment of Cancer protocol 30805. Cancer 1993; 72: 38557
  • 7
    Bishop MC. Experience with low-dose oestrogen in the treatment of advanced prostate cancer: a personal view. Br J Urol 1996; 78: 9218
  • 8
    Harland S, Mufti G, Gelister J, Anjum I, Nicholson P. Hydrocortisone and stilboestrol in combination for castration relapsed prostate cancer. Proc Am Soc Clin Oncol 1998; 17: 1251
  • 9
    Plowman P, Perry L, Chard T. Androgen suppression by hydrocortisone without aminoglutethimide in orchiectomised men with prostate cancer. Br J Urol 1987; 59: 2557
  • 10
    Dowsett M, Shearer R, Ponder B, Malone P, Jeffcoate S. The effects of aminoglutethimide and hydrocortisone, alone and combined, on androgen levels in post-orchiectomy prostatic cancer patients. Br J Cancer 1988; 57: 1902
  • 11
    Huggins C & Hodges C. Studies on prostate cancer: effect of castration, of estrogen and of androgen injection on serum phosphatases in metastatic carcinoma of the prostate. Cancer Res 1941; 1: 2937
  • 12
    Farnsworth W. A direct effect of estrogens on prostatic metabolism of testosterone. Invest Urol 1969; 6: 4237
  • 13
    Belitsky P, Elhilali M, Oliver J. Effect of stilbestrol on the isoenzymes of lactate dehydrogenase in benign and malignant prostatic tissue. J Urol 1969; 104: 4536
  • 14
    Castagnetta L & Carruba G. Human prostate cancer: a direct role for oestrogens. Ciba Foundation Symposium 1995; 191: 2869
  • 15
    Carruba G, Miceli D, D'Amico D et al. Sex-steroid upregulate e-cadherin expression in hormone responsive LNCaP human prostate cancer cells. Biochem Biophys Res Commun 1995; 212: 62431
  • 16
    Alfthan O & Holsti L. Prevention of gynaecomastia by local roentgen irradiation in estrogen-treated prostatic carcinoma. Scand J Urol Nephrol 1969; 3: 1837
  • 17
    Gleave M, Goldenberg SL, Bruchovsky N, Rennie P. Intermittent androgen suppression for prostate cancer. Rationale and clinical experience. Prostate Cancer Prostatic Dis 1998; 1: 28996

Authors

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. References
  8. Authors

D. Farrugia, MD, MRCP, PhD, Macmillan Specialist Registrar in Urological Oncology.

W. Ansell, BA, RGN, Onc. Nurs. Cert., Macmillan Urology Nurse Specialist.

M. Singh, MBBS, FRCS, Consultant Urologist.

T. Philp, MA, Mchir, FRCS, Consultant Urologist.

F. Chinegwundoh, MS, FRCS(Urol), FRCS(Edin), FEBU, Consultant Urologist.

R.T.D. Oliver, FRCP, MD, Sir Maxwell Joseph Professor in Medical Oncology.