Changes in prostate-specific antigen and hormone levels following withdrawal of prolonged androgen ablation for prostate cancer

Changes in prostate-specific antigen and hormone levels following withdrawal of prolonged androgen ablation for prostate cancer S Egawa 1 *, H Okusa 1, K Matsumoto 1, K Suyama 1 & S Baba 1 1 Department of Urology, Kitasato University School of Medicine, Sagamihara, Kanagawa, Japan (2003) 6, 245 249 & 2003 Nature Publishing Group All rights reserved 1365-7852/03 $25.00 www.nature.com/pcan We conducted a study in order to characterize changes after withdrawal of androgen ablation (AA) for prostate cancer. AA was withdrawn in 38 Japanese patients with prostate cancer who had undergone this therapy for various periods. Patients were stratified into those who had undergone AA for less than 24 months (Group 1, n ¼ 12) and those with longer periods of AA (Group 2, n ¼ 26). Serial changes in hormones and prostate-specific antigen (PSA) were prospectively monitored following cessation of AA. The median durations of AA in the two groups were 8.5 and 54.5 months, respectively. Levels of total testosterone (T), luteinizing hormone and PSA increased significantly with time. At the end of 2 y, 30/38 patients (78.9%) had T levels above 50 ng/dl and 19/38 (50%) had levels above 320 ng/dl. Patients in Group 2 required significantly longer duration for T recovery. Complete T recovery is not always accompanied by rising PSA. Recovery of T levels is often slow following cessation of prolonged AA. Expression of PSA after AA is often variable and unpredictable. Thus, interpretation of outcomes in clinical trials incorporating AA needs caution and careful consideration. (2003) 6, 245 249. doi:10.1038/sj.pcan.4500675 Keywords: androgen ablation; testosterone; prostate-specific antigen; bone mineral density Introduction Since the introduction of endocrine therapy by Huggins and Hodges 1 in 1941, androgen ablation (AA) has been the mainstay for treatment of advanced prostate cancer treatment. Today, AA is used in multiple clinical settings that include intermittent hormonal therapy 2 4 and neoadjuvant or adjuvant therapy combined with definitive local treatment. 5 7 Use of luteinizing hormone releasing hormone agonists (LHRHa) has become the preferred method of AA. Knowledge of long-term outcome is important to assess critically the clinical impact of prolonged adjuvant therapy, and several important *Correspondence: S Egawa, Department of Urology, Kitasato University School of Medicine, 1-15-1 Kitasato, Sagamihara, Kanagawa 228-8555, Japan. E-mail: s-egpro@jcom.home.ne.jp Received 10 January 2003; revised 10 May 2003; accepted 17 June 2003 observations have recently been made regarding the reversibility of LHRHa effects after cessation of AA. 6,8 11 The objectives of the present study are to characterize changes in levels of hormones and prostate-specific antigen (PSA) after withdrawal of AA in 38 Japanese patients with prostate cancer. Patients and methods Patients Between March 1999 and January 2002, 38 Japanese patients who had undergone AA for histologically confirmed prostate cancer were invited to join this study at Kitasato University Hospital. AA had been maintained with a combination of 1-month depot injection of LHRHa (leuprolide acetate 3.75 mg or goserelin acetate 3.6 mg) and antiandrogen (flutamide 375 mg/day) in 16 patients

246 and with LHRHa alone in 21 patients. One patient had been maintained on a daily dose of diethylstilbestrol diphosphate (DESD), 200 mg. The patient group included 29 men with advanced prostate cancer who were recruited for a prospective trial of intermittent endocrine therapy; 17 of these patients had multiple bone metastases. 2 These intermittent therapy patients had received at least 15 months of AA. All drugs were discontinued at the time of study entry and serial data were collected during the off-therapy period. Androgen ablation was resumed 2 months after PSA reached levels greater than 10 ng/ml, when indicated clinically, or on patient request. The study also included nine patients who underwent 8 or more months of neoadjuvant endocrine therapy prior to radical prostatectomy for resectable prostate cancer. No patient showed any evidence of clinical or biochemical progression during endocrine therapy. Before entering the study, all 38 patients gave signed informed consent after discussion. Study design and assessment In order to investigate the impact of length of AA, patients were further stratified into those whose AA duration was less than 24 months (Group 1, n ¼ 12) and those with longer duration AA (Group 2, n ¼ 26). Patients were seen in monthly follow-up visits to determine treatment-related side effects and to measure serum PSA, total testosterone (T) and luteinizing hormone (LH) levels. Blood samples were collected between 0900 and 1200. Serum PSA after withdrawal of AA was quantitated by an Immulite third-generation hypersensitive assay (Iatron Laboratories, Inc., Tokyo, Japan). Values prior to AA were measured by AxSYM assay (Dinabot, Tokyo, Japan). Total T and LH were measured by immunoradiometric assay. The normal ranges for T and LH in this study are 320 1030 ng/dl and 1.1 8.8 miu/ml, respectively. Castrate serum T was defined as less than 50 ng/dl. Statistical analysis Differences between values at each time point were assessed by the Wilcoxon signed rank test, while other values were assessed using the Mann Whitney U or Kruskal Wallis test. Standard Kaplan Meier actuarial statistics were used to generate curves of T-level recovery, with the log-rank test being used to compare recovery of hormone values. The impact of covariates on recovery of T levels was assessed using multivariate logistic regression analysis. Covariates included age, clinical stage, pretreatment PSA, biopsy tumor grade, gland volume, total T levels prior to AA, and type and duration of endocrine therapy. P-values o0.05 were considered significant. Results Baseline clinical findings Table 1 shows patient demographics at entry and duration of follow-up. Table 2 shows baseline PSA, T and LH levels as well as serial changes between baseline and end of follow-up. Serum T levels prior to AA were available for 22 patients, with a median value of 398.5 ng/ dl (range 84.0 710.0). Five patients had subnormal T levels prior to treatment; values in three of these reached the normal range at some point during follow-up. Changes of hormone levels after withdrawal of AA Levels of T and LH increased significantly with time after cessation of AA (Table 2). Figure 1 illustrates the time course of T recovery in Groups 1 and 2, as well as in a subgroup of patients who had AA for 36 months or longer. Patients in both Group 2 subgroups required significantly longer duration for T recovery than those in Group 1. Within 2 years, T levels recovered above 50 ng/dl in 30 patients (78.9%), including 11 (92%) in Group 1 and 19 (73%) in Group 2; the respective median times to recovery were 2.0 (1 5) and 4.0 (1 36) months (P40.05). Normal T levels (4320 ng/dl) were achieved by 19 patients (50%), including 10 (83%) in Group 1 and nine (35%) in Group 2; respective median times to recovery were 3.0 (2 9) and 11.0 (2 19) months (Po0.01). For Group 2, the median T value at 6 months (33.0 ng/dl) was in the castrate range, but by 9 months had recovered slightly to 143.5 ng/dl. Eight patients remained at castrate T levels after a median follow-up time of 27.0 (2 36) months. Recovery of T generally paralleled LH recovery. Six of seven patients who had castrate levels of T after 17 36 months of follow-up had above-normal LH. In the remaining patient, who had undergone prolonged DES treatment, LH levels remained subnormal 19 months after AA discontinuation. Changes of PSA after withdrawal of AA Longer AA duration tended to delay the increase in PSA. While half (2/4) of the nonsurgical patients in Group 1 reached PSA levels of 10 ng/dl following discontinuation Table 1 Patient demographics Total (n=38) Group 1 (n=12) Group 2 (n=26) P-value Age at entry (y) 68.071.5 (48 84) 66.572.4 (52 81) 69.071.9 (48 84) NS AA duration (months) 37.075.0 (8 126) 8.571.9 (8 23) 54.575.2 (26 126) o0.0001 Follow-up duration (months) 16.572.0 (2 36) 6.572.7 (2 35) 22.072.2 (2 36) 0.002 PSA prior to AA (ng/ml) 3 1.0754.6 (0.6 1690.0) 33.8755.2 (8.0 591.0) 28.2776.5 (0.6 1690.0) NS Biopsy Gleason score 7.070.2 (3 9) 5.570.4 (3 8) 7.070.2 (4 9) 0.04 AA, androgen ablation; PSA, prostate-specific antigen; s.e., standard error, NS, not significant.

Table 2 Changes in PSA and hormone levels after withdrawal of androgen ablation 247 Months after off Tx PSA (ng/ml) Total testosterone (ng/dl) LH (miu/ml) 0 0.0270.1 (0.003 2.2, 36) 6.573. 1(5 89, 28) 0.570.1 (0.1 1.9, 18) 1 0.0270.1* (0.003 3.4, 30) 14.0715.4* (5 276, 22) 1.370.3* (0.5 3.5, 14) 3 0.370.4** (0.004 7.4, 28) 64.5743.4** (5 665, 26) 5.972.0** (0.8 29.5, 14) 6 0.670.7** (0.006 1 6.2, 29) 123744.6** (5 697, 23) 7.272.2** (1 33.5, 15) 9 0.871.3** (0.003 17.8, 22) 236755.2** (5 770, 16) 10.372.1* (0.8 24.9, 12) 12 0.770.6** (0.003 10.3, 18) 56746.1** (5 453, 15) 1172.5* (0.6 26.8, 9) 15 3.471.5** (0.006 1 5.5, 12) 975 1.8* (5 481, 9) 11.771.9 (3.9 18.4, 7) 18 1.271.5** (0.006 19, 14) 48770.1** (8 624, 9) 12.172.2* (1 18.6, 8) 24 0.472.2* (0.005 16.2, 8) 2757103.5* (9 708, 6) 12.6719.1 (7.9 87.1, 4) 30 0.0473.0 (0.007 17.9, 6) 837123.2 (5 660, 5) 14.470.8 (13 16.6, 4) 36 0.2710.3 (0.049 0.2, 3) 58713.4 (20 62, 3) 13.971.8 (10.4 16.6, 3) PSA, prostate-specific antigen; LH, luteinizing hormone; Hb, hemoglobin; BMD, bone mineral density; Tx, endocrine therapy; s.e., standard error. *Po0.05, **Po0.01, vs baseline levels. The Values are expressed as median7s.e. (range, no. of patients). Figure 1 Recovery of total testosterone levels of 50 ng/ml or greater after withdrawal of AA for prostate cancer according to group stratification: T, total testosterone levels; o24m, patients who had undergone AA less than 24 months; X24M, patients who had undergone AA 24 months or longer; and X36M, patients who had undergone AA 36 months or longer. of therapy, only 36.0% (9/25) of those in Group 2 did so. All but one patient responded to resumption of AA. As shown in Figure 2a, b and Table 3, levels of T did not necessarily parallel PSA values. Of the 18 patients with subnormal T recovery, six (33.3%) reached PSA levels above 10.0 ng/ml, while in nine (50.0%) the PSA level remained below 1.0 ng/ml. Those who reached values of 10.0 ng/ml did so at a median time of 19.0 (2 30) months, while median follow-up in those with very low values was 33.0 (2 36) months. Of the 11 patients whose T recovered to normal levels, five (45.5%) reached PSA levels above 10.0 ng/ml (median time 11.0 (10 16) months), while in three (27.3%) the PSA remained below 1.0 ng/ml at a median follow-up time of 31.0 (30 34) months. Rising PSA was noted despite castrate levels of T in two patients. One of these appeared to be hormone resistant. Results of logistic regression multivariate analysis All parameters tested failed to predict the delayed recovery of T at any time point after withdrawal of AA (P40.05). Figure 2 Changes of PSA after withdrawal of AA: M, months. (a) Patients with subnormal recovery of total testosterone (o320 ng/dl). Figures in the graph indicate testosterone levels at the last observation (ng/dl). (b) Patients with normal recovery of total testosterone (X320 ng/dl). Discussion AA with LHRHa, in various clinical settings, has increasingly been used to treat prostate cancer. 2 7 Knowledge of hormone and PSA changes after cessation of AA is important for accurate interpretation of the impact of newly developed strategies for using this modality.

248 Table 3 PSA values after withdrawal of androgen ablation stratified by total testosterone recovery Total testosterone recovery (ng/dl) No. pts. (%) PSA range (ng/ml) Total o1.0 1.1 10.0 >10.1 o50.0 7 (24.1) 5 (71.4) 1 (14.3) 1 (14.3) X50.0 and o320.0 11 (37.9) 4 (36.4) 2 (18.2) 5 (45.5) X320.0 11 (37.9) 3 (27.3) 3 (27.3) 5 (45.5) No. pts., number of patients; PSA, prostate-specific antigen. Studies have suggested that recovery of T levels after short-term AA is relatively rapid. For example, Oefelein 11 reported a 6-month median duration of castrate level T (r20 ng/dl) after a single 3-month LHRHa injection. Similarly, Nejat et al 10 reported a median time from withdrawal to T normalization (4270 ng/dl) of 7 months following a median of 9 months of AA produced by 1- or 3-month depot LHRHa and antiandrogen. Recovery was often delayed, however, in the relatively small number of patients treated for X24 months. This was seen in the study of Nejat et al, although only seven of 68 patients underwent the longer-term treatment. Likewise, Hall et al 9 reported that 9 months after withdrawal, the median T values remained in the castrate range (o50 ng/dl) in 14 patients who had undergone an average of 38.6 (25 82) months of 3-month LHRHa depot therapy with or without antiandrogen. There was no correlation between duration of therapy and time course of T recovery. Following adjuvant AA of between 3 months and 3 years duration in patients undergoing definitive radiotherapy, Pickles et al 6 reported that 79% recovered normal T levels (10 nmol/l) and 93% recovered levels of at least 5 nmol/l. In multivariate analysis, advanced age, low pretherapy T and use of 3-month LHRHa were all significantly associated with delayed T recovery. Duration of AA was not significantly associated with time to T recovery, but only nine men had undergone AA for more than 2 y. We conducted a similar investigation using predominantly monthly LHRHa injections with or without antiandrogen. Testosterone levels in 10 patients from Group 1 (83%) and nine from Group 2 (35%) recovered to the normal range at a median time of 3.0 (2 9) months and 11.0 (2 19) months, respectively (Po0.01). In Group 2, the median T value was in the castrate range at 6 months, but recovered above this level at 9 months. This suggests that the time course of T recovery may differ between monthly and 3-month depot injections, since Hall et al 9 reported a more durable response with a 3-month depot preparation. Pickles et al 6 have suggested a switch from 3-month to monthly depot injections for the last year of adjuvant therapy to encourage prompt T recovery, and our findings support their suggestion. Although significant in univariate analysis, the duration of AA was not a significant independent predictor of delayed T recovery. This lack of significance may be due to the small number of patients in this study. Group 2 patients with more than 36 months of AA were not notably different from the rest of the group. This is in accord with the experience of Hall et al. 9 It seems that once AA duration reaches 24 months, further increases may have little impact on T recovery. Recovery of T paralleled LH levels in the majority of patients. Six of seven patients who had castrate T levels after at least 17 months off therapy had elevated levels of LH. All these had received LHRHa alone or with antiandrogen. Testosterone levels prior to AA had been normal in the two patients for whom they were available. Although it has been suggested that the effect of LHRHa on spermatogenesis and Leydig cell function may be reversible, 12,13 prolonged administration may lead to permanent impairment in at least some patients. However, as described by Pickles et al, 6 T recovery may occur in some patients after periods as long as 5 y. The level of LH was suppressed in our remaining patient with castrate T levels, who had been under DESD for 126 months. Long-term DESD may have affected the recovery of biological function in the hypothalamic pituitary testicular axis. The mechanism is not certain, but a similar observation with diethylstilbestrol has been reported. 14 Oefelein 8 suggested higher-than-castrate T levels as a threshold for resuming long-acting LHRHa therapy. Although this suggestion seems reasonable, the most appropriate threshold value of T remains to be determined. Although secretion and production of PSA are known to be under androgenic control, levels of PSA in our study did not always parallel T recovery (Figure 2, Table 3). Nejat et al 10 suggested caution in the interpretation of PSA levels in men with subnormal T. In this study, approximately one-third of patients with normal T had suppressed PSA levels (less than 1 ng/ml) after prolonged AA. This effect has been explained by altered expression of PSA by tumor cells following AA. 2 4 Host factors, including deferred recovery of free testosterone and sex hormone binding globulin, may be involved in those with the lowest total testosterone levels. 13,15 Clinically, this implies that interpretation of neoadjuvant and adjuvant endocrine therapy needs careful consideration and long-term follow-up. Our study is limited by its nonrandomized nature and the small number of patients. The number of patients with subnormal T prior to AA, which is known to occur in some elderly men, 16 is unknown in this study but is also lacking in all previous studies. 6 10 It is also unknown whether our current findings may be applicable to other races. The interpretation of our current findings thus needs caution and careful consideration. Further work is warranted. Conclusions Recovery of T levels is delayed in a substantial number of patients when the duration of AA exceeds 24 months. More information, including certain host factors, is needed to interpret correctly PSA changes in various clinical settings.

Acknowledgements We thank WA Thomasson, PhD, for expert editorial assistance. This work was supported in part by a Grant from the Ministry of Health and Welfare of Japan and the Foundation for Promotion of Cancer Research in Japan. References 1 Huggins C, Hodges CV. Studies on prostatic cancer. I. The effect of castration, of estrogen, and of androgen injection on serum phosphatases in metastatic carcinoma of the prostate. Cancer Res 1941; 1: 293 297. 2 Egawa S et al. A pilot study of intermittent androgen ablation in advanced prostate cancer in Japanese men. Jpn J Clin Oncol 2000; 30: 21 26. 3 Akakura K et al. Effects of intermittent androgen suppression on androgen-dependent tumours: apoptosis and serum prostate specific antigen. Cancer 1993; 71: 2782 2790. 4 Gleave M, Bruchovsky N, Goldenberg SL, Rennie P. Intermittent androgen suppression: rationale and clinical experience. In: Schroder T-H. (ed). Recent Advances in Prostate Cancer and BPH. Parthenon: New York, 1997, pp 109 120. 5 Moul JW. Contemporary hormonal management of advanced prostate cancer. Oncology 1998; 12: 499 505. 6 Pickles T et al. Testosterone recovery following prolonged adjuvant androgen ablation for prostate carcinoma. Cancer 2002; 94: 362 367. 7 Bolla M et al. Improved survival in patients with locally advanced prostate cancer treated with radiotherapy and goserelin. N Engl J Med 1997; 337: 295 300. 8 Oefelein MG. Serum testosterone-based luteinizing hormonereleasing hormone agonist redosing schedule for chronic androgen ablation: a phase I assessment. Urology 1999; 54: 694 699. 9 Hall MC et al. Prospective determination of the hormonal response after cessation of luteinizing hormone-releasing hormone agonist treatment in patients with prostate cancer. Urology 1999; 53: 898 903. 10 Nejat RJ et al. A prospective analysis of time to normalization of serum testosterone after withdrawal of androgen deprivation therapy. JUrol2000; 164: 1891 1894. 11 Oefelein MG. Time to normalization of serum testosterone after 3-month luteinizing hormone-releasing hormone agonist administered in the neoadjuvant setting: implications for dosing schedule and neoadjuvant study consideration. JUrol1998; 160: 1685 1688. 12 Kuber W, Viehberger G, Zeillinger R, Spona J. Effects of the duration of therapy with the LHRH agonist D-ser (BUT)6 Azgly10-LHRH (ICI 118-630) on the steroid hormone content and the morphology of human testicular tissue in the treatment of patients with advanced prostate cancer. Urol Res 1991; 19: 19 24. 13 Lunglmayr G et al. Effects of long term GnRH analogue treatment on hormone levels and spermatogenesis in patients with carcinoma of the prostate. Urol Res 1988; 16: 315 319. 14 Tomic B, Bergman B. Hormonal effects of cessation of estrogen treatment for prostatic carcinoma. J Urol 1987; 138: 801 803. 15 Hoffman MA, DeWolf WC, Morgentaler A. Is low serum free testosterone a marker for high grade prostate cancer? JUrol2000; 163: 824 827. 16 Vermeulen A, Kaufman JM. Ageing of the hypothalamopituitary-testicular axis in men. Horm Res 1995; 43: 25 28. 249