Flare Associated with LHRH-Agonist Therapy

Abstract

The most common form of hormonal treatment for prostate cancer is luteinizing hormone-releasing hormone (LHRH)-agonist therapy. During the first 1 to 3 weeks of LHRH-agonist therapy, an initial increase in testosterone is associated with a condition known as “flare.” Blockade of flare can be accomplished with a number of agents, including flutamide, bicalutamide, nilutamide, diethylstilbestrol, ketoconazole, and cyproterone acetate. Evidence from the early use of LHRHagonists suggested that flare could be serious in nature, with an exacerbation of pain, increase in uremia, development of neurologic sequelae, and possibly death. These events have been uncommonly reported of late, most probably owing to the use of flare blockade in most patients with advanced disease, as well as the fact that many patients are currently being treated with much earlier disease. Evidence is conflicting as to whether flare makes a difference in less advanced disease. A few reports have noted complications during flare in patients in whom a blockade of flare was not required, including two deaths from one institution. Reported series, however, seem to suggest that with flare blockade, acute complications are extremely uncommon. Evidence suggests that 1) advanced disease should be blocked; 2) blockade should probably include an antiandrogen beginning about 1 week prior to administration of the LHRH-agonist; and 3) that patients without advanced disease but with very high PSA levels should be considered for flare blockade.

Despite the development of a variety of novel agents for the treatment of advanced prostate cancer, the foundation of treatment for this stage of disease remains androgen-deprivation hormonal therapy. Although orchiectomy may be desirable for some patients, the emotional and psychologic effects of this treatment render it unacceptable to most patients in the United States. As a result, luteinizing hormone-releasing hormone (LHRH) analogs have been used as primary therapy for most such patients. Indeed, evaluations of these agents in clinical trials of patients with advanced prostate cancer have found them to be equivalent to castration.¹

LHRH analogs cause their effect by interfering with the normal pulsatile release of LHRH from the hypothalamus; eventually, after a period of about 2 weeks, downregulation of LH secretion occurs, eventually resulting in a castrate level of testosterone. However, during the initial 1 to 2 weeks, a biochemical and, on occasion, a clinical “flare” can occur. It is now generally accepted that such flare can be blunted by the administration of various agents, but the question has been raised as to whether the flare is of clinical significance. In this review, we will examine available evidence that addresses this question.

Biochemistry of Flare

Kreis and colleagues have investigated the change in gonadotropins and androgens in response to LHRH-agonist therapy.² These changes can be compared to the change in symptoms seen during the “flare” period, based on observations by Waxman and colleagues.³ These relationships are graphically displayed in Figure 1. As can be seen from this figure, after the administration of an LHRH agonist, LH peaks at about three times the baseline level at 24 hours after administration. The testosterone rise is somewhat delayed, with a peak reached at approximately day 3, falling to pretreatment levels by day 7. Ultimately, castrate levels are achieved by 3 weeks following initiation of therapy.

Figure 1.

A graphical description of the timing of changes in testosterone, luteinizing hormone (LH), and pain after first unblocked injection of LH-releasing hormone (LHRH) agonist. Changes are not to scale.

Symptoms during unblocked flare parallel the change in LH and testosterone. Pain is generally first seen 12 hours after initiation of the LHRH analog, peaks at 36 hours, and disappears by the end of the first treatment week.

Blockade of Flare

Several clinical experiences have demonstrated that blockade of clinical flare can be accomplished with a variety of agents. Treatments using cyproterone acetate (CPA),^4,5 diethylstilbestrol (DES),^6,7 flutamide,⁸ or nilutamide⁹ all appear to be effective in preventing clinical flare. Although not often used, treatment with ketoconazole will also prevent flare.¹⁰

An insightful experience with a small sample of patients compared hormonal levels in men treated with one of four regimens: Group I-Zoladex 3.6 mg given monthly; Group II-Zoladex plus CPA 100 mg twice daily, begun 7 days prior to LHRH agonist; Group III-Zoladex plus flutamide 250 mg t.i.d., begun 1 day before LHRH agonist; and Group IV-Zoladex plus flutamide, begun 7 days prior to LHRH agonist.¹¹ The authors found that testosterone and LH levels increased in all groups but were most blunted with CPA. The least effective method to blunt the flare was noted in patients treated with flutamide for only 1 day prior to LHRH-agonist therapy. In these men, LH levels increased over 800%.

A similar interesting observation was made by Ferrari and colleagues. ¹² In this study, 122 men with advanced (node-positive and metastatic) prostate cancer received either buserelin or buserelin with flutamide. In the buserelin-alone group, CPA (250 mg per day) was added for 3 weeks to prevent disease flare. The authors noted an increase in testosterone levels in both groups from about 12 nmol/L to just over 14 nmol/L. A total of four patients had a transient increase in bone pain (three in the buserelin-CPA arm and one in the buserelin-flutamide arm). Although flare reaction was less in the buserelin-CPA arm, systemic adverse reactions were more common in the arm with flutamide (one patient in the former as opposed to 11 patients in the latter).

Like the experience of Ferrari, Kuhn and colleagues analyzed the results of buserelin with or without nilutamide and found that with nilutamide, instead of the increase in average extent of bone pain seen with buserelin alone at days 2 to 7 of treatment, the average degree of pain began to fall on day 2.⁹ Nevertheless, over the initial 29-day period of treatment, although new or decreased pain was less in the combination group, it was not eliminated. Among 17 men treated with buserelin and nilutamide, new or increased pain was noted in 5. On a visual-analogue scale, the maximal increase was at or less than 10 mm in 2 patients, 11 mm to 20 mm in 1, 21 mm to 30 mm in 1, and over 30 mm in 1.

Although some studies have reported individual patients with some exacerbation of pain despite flare blockade, others have not had similar findings. In a series of 70 patients with previously untreated metastatic prostate cancer, Labrie and colleagues found that coadministration of flutamide with the LHRH agonist did not cause an increase of pain in any patient.¹³

It appears that it is possible to block both endocrinologic and clinical manifestations of flare by some treatment approaches. Bruchovsky and colleagues used an innovative, although somewhat complicated, approach: 1) Initial therapy with CPA and DES for 4 weeks; followed by 2) goserelin acetate (Zoladex) begun after the 4-week initial CPA-DES therapy; with 3) DES stopped after a total of 8 weeks but maintenance of CPA to help manage vasomotor side effects.¹⁴ Unlike other series, testosterone was noted to decrease from an initial value of 13.3 ± 1.0 nmol/L to 1.4 ± 0.2 nmol/L after 1 week and thereafter to 0.8 ± 0.2 nmol/L after 4 weeks of therapy. After the first administration of goserelin acetate at 4 weeks, a slight increase in testosterone to 1.2 ± 0.2 nmol/L was seen but was not a statistically significant change. The authors did not note any type of flare symptoms with this regimen.

Complications of Flare

Although we will review the early experience with LHRH flare, it is important to recognize that the frequency of this occurrence is significantly less than when LHRH agonists were first employed. The reason is probably twofold: the prevalence of the use of flare-blocking agents, and the general use of LHRH agonists earlier in the course of the disease due to the ubiquitous use of prostate-specific antigen (PSA) for prostate cancer screening. With regard to the latter, many urologists will not see a de novo patient with metastatic prostate cancer during an entire year of practice-most patients with metastatic disease will have been treated hormonally long before the development of distant disease.

Early reports of LHRH-agonist flare suggested that adverse events were significant and highly prevalent. Waxman’s experience with 46 patients with symptomatic locally advanced or metastatic prostate cancer is instructive.³ Patients received a variety of dosing schedules of buserelin without flare blockade. A total of 18 patients had a transient increase in pain, with four noting increased lymphedema. A significant increase in creatinine (from 1.92 mg/100 mL to 6.79 mg/100 mL) was seen in one patient, and one suffered a cord compression with complete sphincter dysfunction and weakness. (In this patient, radiotherapy to the spine cleared all symptoms and signs.) It is of interest that of 14 patients with no pain at diagnosis, one developed pain during the “flare period.”

In an early experience at our own institution, beginning in 1986, although flare was blocked (according to recommendations at that time) for patients with impending cord compression or ureteral obstruction, routine blockade in all patients was not practiced. We subsequently reported two patients who experienced sudden death during flare.¹⁵ The first patient with low back pain and a positive bone scan was put onto daily leuprolide injections on December 13, 1986 and developed shortness of breath on January 5, 1987, succumbing 3 days later. The second patient with back pain and a positive bone scan but no neurologic symptoms was treated on December 8, 1987 and experienced increased back pain after the first dose, requiring initiation of narcotics. The patient suffered respiratory arrest on December 13, 1987, and died. Autopsy was not performed on either patient.

A review of flare conducted by us at that time found it to be a relatively frequent phenomenon. Of a total of 765 patients in 9 series, 84 (10.9%) suffered an initial flare in symptoms.^{2,3,8,9,16–21} In a review in 1990, we found a total of 17 patients (of a total number of somewhat more than 300 patients treated) who expired during the flare period.

A wide variety of adverse events has been reported during initial therapy and possibly related to flare. These include ureteral obstruction, urinary retention, spinal cord compression, lymphedema, and death.^2,15,22

Impact of Flare in Randomized Clinical Trials

Several randomized clinical trials have compared the effectiveness of hormonal therapy for patients with prostate cancer. Although in general there have been few differences between various forms of hormonal therapy (if these treatments were begun simultaneously), small differences may be attributable to the effect of testosterone and tumor flare. In the Leuprolide Study Group trial, comparing leuprolide with DES, treatment failure at the 3-month mark was noted in 10 patients who received leuprolide alone, compared to only two patients receiving DES.¹⁸ Similarly, in the Southwest Oncology Group study 8494, comparing leuprolide alone to leuprolide plus flutamide, although combination therapy was associated with a slight improvement in survival, the survival curves “separated” very early (within 3 months of treatment) and appeared to parallel each other thereafter.⁸

The Bottom Line: Is Flare Important?

Although for certain patients-those with markedly advanced prostate cancer who are at risk of catastrophic events from transient disease flare-there may be significant risks associated with testosterone flare, for the majority of patients, the phenomenon is probably only recognizable if hormone levels are measured. Certainly, if a major risk were present, its frequency would have been noted increasingly as the use of LHRH agonists proliferated over the past decade. A recent publication by the Agency for Health Care Policy and Research (renamed the Agency for Healthcare Research and Quality) had two striking conclusions that are germane: 1. there is no statistically significant difference in survival for patients treated with an LHRH agonist compared to patients treated with orchiectomy or DES; and 2. no comment was made in the Summary Statement regarding the impact of disease flare with LHRH-agonist therapy.¹ With regards to the former, 10 trials involving 1098 patients, comparing LHRH-agonist therapy with either orchiectomy or DES, found at the 2-year mark that survival was essentially equivalent (hazard ratio = 1.1262; 95% confidence interval = 0.915−1.386). With regards to the latter, the AHCPR actually concludes: “The evidence comparing adverse effects is limited but favors monotherapy over combined androgen blockade.” In the text of the report, the AHCPR acknowledges the potential adverse sequelae of flare but notes, as above, that it is exceedingly rare that such LHRH-agonist monotherapy is used for advanced disease today.

On the basis of the available literature, the following conclusions seem reasonable:

1. With a patient with advanced prostate cancer with symptoms from his disease, almost regardless of the type of symptoms, some form of therapy devoid of flare disease exacerbation should be used. Thus the patient with significant obstructive voiding symptoms merits a flare-free induction of hormonal therapy, as does the patient with significant back pain and early neurologic sequelae: both urinary retention and paraplegia are side effects to be avoided.

2. Flare-free induction hormonal therapy should also be considered for the patient with extensive, asymptomatic disease. In such a patient, flare could induce undesirable symptoms. A reasonable metric would be a significant elevation in PSA (for example, levels above 50 ng/mL to 100 ng/mL) or any evidence of bulky disease.

3. For the patient to be treated with an LHRH agonist in whom flare blockade is desired, optimal blockade is achieved by pretreatment with an appropriate agent (eg, antiandrogen). For the patient in whom maximum blockade is desirable immediately, consideration could be given to the use of ketoconazole, which causes a dramatic reduction in testosterone in only a matter of hours.

Main Points.

• Androgen-deprivation hormonal therapy remains the foundation of treatment for advanced prostate cancer, and luteinizing hormone-releasing hormone (LHRH) analogs are the primary therapy.

• During the initial 1 to 2 weeks of LHRH-agonist therapy, a biochemical and, on occasion, a clinical “flare” can occur.

• Treatments using cyproterone acetate, diethylstilbestrol, flutamide, or nilutamide all appear to be effective in preventing clinical flare; ketoconazole will also prevent flare.

• Flare is significantly less frequent nowadays than when LHRH agonists were first employed, probably because of the use of flareblocking agents, and the use of LHRH agonists earlier in the course of the disease due to prostate-specific antigen screening.

• Adverse events possibly related to flare include ureteral obstruction, urinary retention, spinal cord compression, lymphedema, and death.

• For patients with markedly advanced prostate cancer who are at risk of catastrophic events from transient disease flare, there may be significant risks associated with testosterone flare, but for the majority of patients the phenomenon is probably only recognizable if hormone levels are measured.