Testosterone recovery and changes in bone mineral density after stopping long-term luteinizing hormone-releasing hormone analogue therapy in osteoporotic patients with prostate cancer

Abstract

OBJECTIVE

To investigate the rate of testosterone recovery and changes in bone mineral density in patients found to be osteoporotic while receiving luteinizing hormone-releasing hormone (LHRH) analogues after changing to antiandrogen monotherapy in an attempt to reduce further demineralization.

PATIENTS AND METHODS

Fifteen patients receiving LHRH analogue therapy for ≥ 1 year were identified as osteoporotic by distal forearm dual X-ray densitometry. They were then converted to antiandrogen monotherapy, and prostate specific-antigen (PSA) and total testosterone monitored at 3-monthly intervals. The forearm densitometry was repeated at 1 year.

RESULTS

All patients had some testosterone recovery; the mean (range) duration to initial detectable testosterone was 12.8 (6–22) months. Six patients had a normal testosterone level after a mean of 17.5 (14–30) months. In the year after stopping LHRH analogue therapy the mean bone mineral density (t-score) decreased by 7.2%.

CONCLUSIONS

Osteoporotic patients, after stopping LHRH analogues, continue to have suppressed levels of testosterone which have a detrimental effect on bone mineral density. We therefore would not advocate conversion to antiandrogen monotherapy to improve bone density, and suggest alternative therapeutic intervention e.g. bisphosphonate therapy, for these patients.

Abbreviations

BMD

bone mineral density

DEXA

dual energy X-ray absorptiometry

INTRODUCTION

Depot LHRH analogues are currently the most popular method of hormone manipulation for patients with carcinoma of the prostate. They are effective in delaying the progression of the disease and hence patients are often on this treatment for many years [1]. However, LHRH analogues, because of their testosterone ablative effects, disturb bone metabolism, resulting in osteoporosis, a widely recognized complication of long-term testosterone ablation [2,3]. Osteoporosis is characterized by low bone mineral density (BMD) and is asymptomatic, only clinically manifesting itself when a low-trauma ‘osteoporotic fracture’ occurs. Several studies show that the incidence of osteoporotic fractures in patients receiving LHRH analogues is far higher than the incidence of pathological fractures [4,5]. An alternative method of hormone manipulation which maintains testosterone levels is antiandrogen therapy. This is increasingly popular as monotherapy in prostate cancer, as it offers comparable survival outcomes and potential quality-of-life benefits [6,7]. In the present study we investigated the time taken for testosterone levels to recover in osteoporotic patients established on long-term LHRH therapy, after changing to antiandrogen monotherapy, and assessed the changes in BMD.

PATIENTS AND METHODS

The study comprised 15 osteoporotic patients enrolled between November 2000 and July 2002. The mean (range) follow-up was 18 (12–30) months. The BMD of the subordinate forearm was measured by dual-energy X-ray absorptiometry (DEXA) using a densitometer; the system used has a unique protocol for evaluating BMD at the ultra-distal radius and ulna. The BMD result is expressed as the sd about the mean for a healthy age-matched population (the t-score). Osteoporosis was defined using the WHO criteria as a BMD of ≤− 2.5 sd (below) the mean for a healthy population group aged 20–40 years (t-score ≤− 2.5), with or without pre-existing fragility fractures. Osteopenia is defined by a t-score of −1.0 to −2.4 and a normal BMD t-score as >−1.0.

Eight patients were found to be osteoporotic while already receiving LHRH analogues, and a further seven were osteopenic at the initial scan before treatment, and subsequently went on to develop osteoporosis while on LHRH analogues. Six patients developed osteoporosis within a year; their mean (sd) t-score on initial DEXA scanning was −2.07 (0.15), decreasing to −2.72 (0.13) after a year, while one further patient progressed to osteoporosis after 2 years. All patients had received 3-monthly depot goserelin 10.8 mg for ≥ 1 year, replaced by bicalutamide 150 mg once daily as an alternative monotherapy after osteoporosis was diagnosed, the first dose being 3 months after the last depot injection. They also received calcium and vitamin D supplement (Ca²⁺ 12.6 mmol and cholecalciferol 400 units once daily). Total testosterone and PSA levels were measured at baseline and at ≈ 3-month intervals thereafter. A further forearm DEXA scan was taken a year after stopping LHRH analogue therapy.

To determine the normal testosterone levels in patients with prostate cancer before hormonal treatment, 30 consecutive patients presenting to our unit had their testosterone levels measured and forearm densitometry assessed (control group). The laboratory reference range for a normal testosterone level is 9–40 nmol/L. The two-tailed Student's t-test was used for statistical comparison, with a statistically significant result assumed at P ≤ 0.05.

RESULTS

The mean (range) age of the 15 osteoporotic patients was 72 (55–86) years, the median Gleason grade at diagnosis 6 (2–10), median tumour stage T3 (T1–T4), mean PSA 57  (1.7–312) ng/mL and the mean duration on goserelin therapy 27 (12–96) months. At the time of converting to bicalutamide monotherapy the mean (sd) t-score was −3.0 (0.73).

The mean age of the control group was 73 (56–85) years, with a mean (sd) testosterone level of 11.8 (3.5) nmol/L. The mean t-score was − 1.6, with only six being osteoporotic. However, there was no statistical difference in the testosterone levels between the osteoporotic control patients and the remainder (P = 0.86).

All study patients had undetectable testosterone levels at the time of monotherapy conversion (<1.5 nmol/L). They all had some degree of testosterone recovery (Fig. 1), the mean (range) time to initial detectable testosterone level (≥1.5 nmol/L) being 12.8 (6–22) months. Six patients had normal testosterone levels after a mean of 17.5 (14–30) months. The remaining patients are yet to reach normal testosterone levels.

Figure 1. Testosterone recovery after stopping LHRH analogue therapy in 15 men with osteoporosis. The black horizontal line is the normal level (>9 nmol/L).

Ten patients received goserelin therapy for 1–2 years and five for ≥ 2 years; the mean duration to initial detectable testosterone was 12.6 and 13.0 months, respectively, with no statistical difference between the groups (P = 0.70).

The mean (range) PSA at baseline (before conversion) was 0.65 (undetectable−3.3)  ng/mL; 11 patients had a baseline PSA of ≤ 0.5 ng/mL and after a mean of 17.5 months only one increased from 0.5, to 3.2 ng/mL (after 15 months). The remaining four patients had an initial mean (range) PSA of 1.9 (0.7–3.3) ng/mL, which increased to 4.8 (0.1–9.8) ng/mL after a mean of 14 months. Two of these patients were converted back to goserelin injections because of a persistently rising PSA level. The PSA level in these two patients has continued to rise, despite achieving castrate levels of testosterone after reinstating LHRH analogue.

In the year after discontinuing LHRH analogue therapy the mean (sd) decrease in t-score was 7.2% (0.11) during this period of hypogonadism, from − 3 (0.73) to − 3.2 (0.70) (P = 0.02).

DISCUSSION

The BMD is testosterone-dependent and both the administration of LHRH analogues [8,9] and orchidectomy [10] have been associated with severe osteoporosis. The exact mechanism by which testosterone maintains BMD is not fully understood, but local aromatization of testosterone to oestradiol is necessary for normal bone homeostasis [11]. There is also evidence that prostate cancer itself is a significant risk factor for osteoporosis, and hence fractures, by causing disturbances in bone turnover and mineralization even before androgen-deprivation therapy [12–14].

Bone densitometry scanning amongst men being treated with LHRH analogues for prostate cancer will inevitably identify patients who are already osteoporotic. As there is an increased rate of fragility fracture resulting in significant morbidity and mortality [15], some form of therapeutic intervention is therefore indicated. Currently no therapy has been convincingly confirmed to be effective in preventing osteoporotic fractures in men. However, as it is known that LHRH therapy is responsible for BMD loss by testosterone suppression, we investigated whether changing from goserelin to bicalutamide would be beneficial. Bicalutamide is a pure nonsteroidal antiandrogen which inhibits the action of dihydrotestosterone and testosterone at target sites, by competitively binding to the cytosolic androgen receptor [16]; however, because of its effect on the hypothalamic-pituitary axis it causes an increase in circulating testosterone and oestrogen levels. We therefore postulated that if testosterone levels could sufficiently recover after the change in hormone manipulation, we might confirm a positive effect on BMD.

There are no published studies examining testosterone recovery in osteoporotic patients after stopping LHRH analogues and subsequently starting antiandrogen therapy. The largest study [17] examining testosterone recovery after androgen ablation examined 267 patients, all of whom received radiotherapy, and most (67%) received stilboestrol and cyproterone, with only nine having >2 years of androgen ablation. The authors showed that over half the men had normal testosterone levels after a year, and 83% after 3 years. Nejat et al.[18] reported on 68 men after withdrawing androgen deprivation therapy, 60% of whom had received external beam radiotherapy. They found that the median time to testosterone recovery was 7 months, after a median duration of androgen deprivation of 9 months. In contrast to the present study, they reported a statistically significant delay in testosterone recovery in patients receiving >2 years of LHRH analogue compared with treatment for <2 years. Radiation scatter affecting the testes may be a confounding factor in both these studies. Hall et al.[19] investigated 14 patients after a mean duration of 38.6 months on LHRH analogues, with castrate levels of testosterone being reported for a median of 6 months; however, the follow-up was limited to a year and therefore normalization of testosterone was not assessed.

The present results show that patients who develop osteoporosis while being treated with LHRH analogues will continue to have hypogonadism for a mean of 17.5 months after conversion to an antiandrogen. The control group showed that some patients may have mild hypogonadism before initial hormone therapy, and therefore we would not expect all the study group to achieve normal testosterone levels. However, Amin et al.[20], using data from the Framingham study, indicated that low levels, as opposed to castrate levels, of testosterone did not correlate with a decrease in BMD, although oestradiol levels have a strong and positive association with BMD in men. This raises the point that normal levels of testosterone may not be necessary for maintaining BMD, as has been shown with physiological functions such as potency, but the availability of some testosterone for aromatization to oestradiol is essential for normal bone homeostasis.

In the year after stopping LHRH analogues the decrease in bone mineral density was consistent with the low or castrate testosterone levels. Therefore, with castrate levels of testosterone remaining for a mean of 12.8 months, the present study showed that discontinuing LHRH analogue therapy is not sufficient to protect osteoporotic patients from further demineralization.

Although becoming more popular, the use of antiandrogen monotherapy in advanced prostate cancer is still controversial. Tachyphylaxis did not appear to be a problem in the present study, with PSA levels remaining low despite testosterone recovery. The two patients who had had a persistent rise in PSA levels to >0.5 ng/mL have failed to show a PSA response despite being rendered pharmacologically castrate by reinstating LHRH analogues. We suggest that this probably reflects the natural history of the prostate cancer in these patients, as opposed to the effect of changing their hormone manipulation.

There is no doubt from increasing reports that LHRH analogues cause a decrease in BMD and therefore increase the risk of osteoporotic fracture. Men are twice as likely as women to die within a year of a hip fracture and are more likely to become dependent on nursing homes after such fractures [15]. Urologists should be more aware of the risk of osteoporosis amongst their patients with prostate cancer receiving hormone manipulation.

The present study highlights the continued testosterone suppression after stopping long-term LHRH analogues and the detrimental effect on BMD. Conversion to antiandrogen therapy seems to offer no significant short-term benefit to these patients, and we therefore suggest alternative therapeutic intervention, e.g. bisphosphonate therapy; further studies are required to investigate this.

CONFLICT OF INTEREST

None declared.