Over 80% of patients with prostate cancer respond to androgen deprivation using drugs or surgery that marked reduce male sex hormone (testosterone) and related androgen levels from the body. Anti-androgens are a class of drugs that specifically block the entry of testosterone into cells of the body, thus preventing its biological effects. Examples of such drugs that are available on the U.S. market include flutamide (Eulexin®), bicalutamide (Casodex®) and nilutamide (Nilandron™). Cyproterone (Androcur) is another anti-androgen that is not available in the U.S.

In approximately 50% of patients whose cancer has started to grow again despite treatment that includes an antiandrogen, the cancer has been noted to regress by simply stopping the anti-androgen. This unusual response is referred to as the Anti-Androgen Withdrawal Response (AAWR). Scientists theorize that prostate cancer cells exposed to antiandrogens for a prolonged time may mutate (undergo genetic change) that paradoxically causes the anti-androgen to stimulate cancer growth.

Although more research is needed, we believe a trial of stopping the anti-androgen for a period of time is warranted in order to determine whether or not an AAWR may occur before considering other types of treatment. A review of the scientific studies that support this treatment approach follows below:

The anti-androgen withdrawal response (AAWR) may be seen, in up to 50% of patients who have PSA relapse while receiving androgen blockade. An AAWR response usually lasts 6 to 8 months, but for some patients, may last as long as 2 years.

Reports of AAWR in the medical literature

Prostate specific antigen (PSA) decline after anti-androgen withdrawal was first reported with flutamide (Eulexin®) by Dupont et.al. (Dupont A, Gomez J. et al. Response to flutamide withdrawal in advanced prostate cancer in progression under combination therapy. J Urol 150:908-913, 1993). These authors observed an AAWR in 30 of 40 of patients (75%) (1 complete, 3 partial, 26 stable disease). The average AAWR duration was 14.5 months (range 3.6-29.9 months). Patients in this study had reportedly developed a rising PSA after initially responding combined hormone blockade (CHB) for a relatively long time (average duration 46.8 +/- 4.5 months).

Kelly and Scher reported an AAWR in 10/35 of patients (29%) after 3 months of flutamide withdrawal (Kelly WK and Scher H: Prostate specific antigen decline after antiandrogen withdrawal: the flutamide withdrawal syndrome. J Urol 149:607-609, 1993). They defined response by a PSA decline of 50% or more (range 37% to 89% decline. In contrast to Dupont, their median response was only 5+ months compared to 14.5 months reported by DuPont. Twenty-five of these patients received CHB as initial treatment, of whom 10 (40%) had an AAWR. Ten patients who received flutamide after PSA relapse on “monotherapy” (a LHRH agonist or orchiectomy alone) did not have an AAWR, nor did any of the other patients who did not receive CHB as their initial hormone treatment.

A similar withdrawal response was reported with another non-steroidal anti-androgen, bicalutamide (Casodex®) by Small and Carroll at UCSF (Small EJ and Carroll PR: Prostate-specific antigen decline after Casodex withdrawal: evidence for an antiandrogen withdrawal syndrome. Urology, 43:408, 1994) and also by Nieh at the Lahey Clinic (Nieh PT: Withdrawal phenomenon with the antiandrogen Casodex, J Urol, 153:1070-1073, 1994). In these studies, the authors referred to this PSA response as the “antiandrogen withdrawal syndrome.” 

An AAWR using chlormadinone acetate (a steroidal anti-androgen available only in Japan) was reported in 2 patients who had greater than 50% PSA declines and significant symptom improvement when this antiandrogen was withdrawn (Akakura K, et.al.: Anti-androgen withdrawal syndrome in prostate cancer after treatment with the steroidal antiandrogen chlormadinone acetate. Urol, 45:700-5, 1995). The above findings provide evidence that a PSA decline is a general response to withdrawal of several classifications of anti-androgens,

Further generalization of withdrawal syndromes has been demonstrated with other hormonal agents active in prostate cancer. Dawson and McLeod from Walter Reed Army Medical Center in Washington DC reported a similar withdrawal response to megestrol acetate (Dawson NA and McLeod DG: Dramatic PSA decline in response to discontinuation of megestrol acetate in advanced prostate cancer; expansion of the antiandrogen withdrawal syndrome. J Urol, 153:1946-47, 1995)

The molecular basis for an AAWR

Moul, et.al. presented a summary and a discussion of various mechanisms that had been found to explain this phenomenon in May 1995 (Moul JW, Srivastava S and McLeod DG: Molecular implications of the antiandrogen withdrawal syndrome. Semin Urol, 13:157-163, 1995). Three series of patients in which various anti-androgen medications were withdrawn were summarized, including data from Kelly and Scher, Nieh and Dawson and McLeod from NCI. In the 105 patients reported in this series, 50-75% exhibited a withdrawal response. Several examples of good responses were cited, including patients having > 80% declines in PSA and remissions of many objective symptoms of the disease that lasted for 5-14 months.

Moul stated that the exact molecular mechanism behind a withdrawal response was unknown, although he listed several mechanisms that have been postulated. The most popular mechanism has centered on mutations occurring in the androgen receptor. While clonal differences among prostate cancer cells may be partially responsible for the time-limited response to anti-androgens observed in some prostate cancers, it has been shown that during or as a consequence of progression to an androgen independent state upward of 20% to 30% of tumors develop mutations in the androgen receptor.

Clinically, the frequency and degree of AAWR was somewhat correlated to the duration of anti-androgen exposure in a low androgen environment. In this condition, an increased rate of androgen receptor mutations are evident and appear to lead to a partial agonist (stimulator) activity of the anti-androgen used, thereby resulting in a decline in its clinical activity.

These assumptions have been supported by in-vitro (laboratory) studies of the LNCaP prostate cancer cell line in which codon 877 reveals a mutation also seen in several other codon point mutations in exon H that have been similarly observed in clinical prostate cancer tissue. Other proposed mechanisms for an AAWR include anti-androgen-induced tumor enzyme changes that may affect the local hormonal milieu and/or possible interactions of anti-androgen metabolites with other developing genetic changes which are known to occur naturally as prostate cancer progresses.

Another reasonable mechanism for serum PSA levels to increase in patients treated with CHB was reported in a 1994 abstract by Herrada, et.al. (Herrada J. Hossan B. Amato R. et.al.: Adrenal androgens predict for early progression to flutamide withdrawal in patients with androgen-independent prostate carcinoma. Proc Am Soc Clin Oncol 13:237, 1994). In this study, an AAWR was not seen in patients with high serum levels of the adrenal androgen, DHEA, at the time of PSA relapse. A summary of the data in the 10 patients studied are summarized below:

Data Form Herrada

As the above data show, none of the patients with DHEA levels > 75 ng/ml had an AAWR, whereas 3 of 5 patients (60%) who had an AAWR had DHEA levels < 75 ng/ml. These observations suggests that if the antiandrogen is in fact acting as an agonist at the androgen receptor, negative feedback (a reflex stimulation) takes place that inhibits pituitary gland release of ACTH (adreno-corticotropic hormone) thereby decreasing DHEA production by the adrenal glands. Conversely, if the anti-androgen is not acting as an agonist, there would be no suppression of adrenal DHEA production. This would predict that an AAWR would be unlikely to occur and that PSA progression is due to hormone-refractory prostate cancer.


Anti-androgen withdrawal should be the first step in patients progressing under CHB. Whether this should be attempted as a solitary maneuver or in conjunction with a second-line hormonal manipulation such as high-dose ketoconazole (Nizoral®) plus hydrocortisone or aminoglutethimide (Cytadren®) plus hydrocortisone is unknown. Many recent trials in prostate cancer patients with PSA relapse on CHB have “exploited” a possible onset of an AAWR by combining simultaneous anti-androgen discontinuation with Nizoral plus hydrocortisone. The results of such trials should be interpreted carefully given our appreciation for the relative frequency and occasional long duration of AAWR in prostate cancer patients.

Regardless of the possible molecular mechanisms for an AAWR, the response is reproducible and clinically significant in a large percentage of cancer patients. With flutamide withdrawal, a significant decrease in PSA levels usually occurs within a few weeks and can potentially last up to 1-2 years. Due to the slower elimination half-life of Casodex from the body, the onset of an AAWR may take up to 4 to 8 weeks with this anti-androgen.

Please refer to our separate booklet entitled “High-dose ketoconazole and hydrocortisone, an effective treatment for prostate cancer” for more information about studies that combine AAWR with this second-line hormone treatment.