Tolerability of Nonsteroidal Antiandrogens in the Treatment of Advanced Prostate Cancer

COL. DAVID G. MCLEOD

Walter Reed Army Medical Center, Washington, DC
Uniformed Services University of the Health Sciences, Bethesda, Maryland

Correspondence:
David G. McLeod, M.D.,
Walter Reed Army Medical Center, Urology Service
6825 Georgia Avenue NW
Washington, DC 20307-5001, USA.

ABSTRACT

INTRODUCTION

The efficacy of CAB is still under investigation. Although some studies ^{4, 5} have shown significant survival benefit of flutamide over placebo when given as part of CAB, other studies could not confirm this ^{6,7,8,9,10,11}. However, there is heterogeneity in trial methodology and, in most of the studies that failed to show significant benefit, the statistical power was not sufficient to show a difference ¹². One of two pivotal trials with nilutamide showed benefit in terms of time to progression ¹³ and survival after long-term follow-up ¹⁴, while the other did not ¹⁵. Furthermore, a meta-analysis of seven nilutamide trials, including the two pivotal trials, also concluded that CAB with nilutamide has advantages over castration alone with respect to progression-free survival ¹⁶ and overall survival after long-term follow-up ¹⁷. A meta-analysis of 22 randomized trials comparing CAB (castration plus flutamide, nilutamide or cyproterone acetate) with castration failed to confirm a survival benefit for CAB ¹⁸. The methodology of this analysis has been challenged, and a subsequent sensitivity analysis of randomized trials has highlighted a two-year survival benefit of CAB for published studies using nonsteroidal antiandrogens but no advantage for CAB in several smaller studies ¹⁹. As part of CAB, bicalutamide has been demonstrated to be at least equivalent to flutamide with regard to time-to-treatment failure, time to progression, and survival, after a median of 95 weeks follow-up ^{20, 21}. This is the only trial reported to date that compares two antiandrogens in a double-blind manner. A large intergroup study of orchiectomy plus placebo versus orchiectomy plus flutamide is expected to provide the answer on the benefits of CAB utilizing orchiectomy and flutamide. Initiated by the National Cancer Institute in 1989 and closed in 1994, this study (NCI INT-0105) has the statistical power to detect a 25% improvement in the median survival time achieved with orchiectomy.

Despite the ongoing discussions, CAB is being used by many clinicians in the management of advanced prostate cancer. With several antiandrogens now available, there is the option of choosing the drug to derive the maximum benefit from CAB with minimum side effects. This review seeks to aid that choice by addressing the common side effects of the nonsteroidal antiandrogens and highlighting differences in their tolerability profiles.

REPORTING OF TOLERABILITY

More importantly, the reporting of adverse event data can be biased by the method of assessment. The most objective assessment, which is required by most regulatory authorities, is one in which there is no attempt to assign causality; hence all adverse events are reported, irrespective of their relationship to the study medication. This is the method used in the recent trial of bicalutamide versus flutamide ²². The practice of assigning causality may lead to biased reporting of adverse events, particularly when done by a third party such as the trial sponsor. In most trials, this bias does not happen and causality is assessed only by the investigator. Even so, different investigators may make different decisions when faced with the same data, and in doing so may introduce a measure of subjective bias. The most widely quoted antiandrogen trial, that of a luteinizing hormone-releasing hormone-A (LHRH-A) with or without flutamide ⁴, reported only those adverse events that were considered treatment-related.

Tolerability reporting in publications may comprise all adverse events, only treatment-related adverse events, or only events requiring withdrawal of study medication. Sometimes it may not even be possible to determine which method of assessment was used. Literature comparisons of tolerability of different drugs are therefore fraught with difficulty. For this reason, no attempt has been made in this review to pool any of the data from different studies in order to arrive at a mean incidence for adverse events. Instead, ranges of incidences that reflect the differing methodologies have been quoted.

TOLERABILITY OF ANTIANDROGENS

Pharmacologic Events

Gynecomastia and breast pain are seen more frequently during antiandrogen monotherapy than during CAB or castration alone. Hot flashes are a side effect of castration, but are seen most often with medical castration using LHRH-A. For example, pooled data from bicalutamide monotherapy studies show that breast pain was reported in 36% of patients taking 50 mg and 39% of patients taking 150 mg. This compares with only 4% of patients on bicalutamide plus LHRH-A and 2% of patients with castration alone ⁴³. The figures for hot flashes were 9% for 50 mg monotherapy, 12% for 150 mg monotherapy, 49% for bicalutamide plus LHRH-A, and 43% for castration alone ^{22, 43}. The comparability of the findings for bicalutamide plus LHRH-A and LHRH-A alone indicates that in combination therapy the pharmacologic adverse event profile is dominated by that of the LHRH-A. This finding holds true for all the nonsteroidal antiandrogens. The data also show that for pharmacologic events, bicalutamide is not associated with a dose-related increase over the range of dosages from 50-150 mg ⁴³.

In the only double-blind, comparative study of two antiandrogens reported to date, that of flutamide plus LHRH-A versus bicalutamide plus LHRH-A, treatment-related pharmacologic adverse events were similar in the two groups (gynecomastia 6% versus 5%; hot flashes 49% versus 50%, respectively) ²².

Pharmacologic adverse events seldom require withdrawal of therapy, although severe gynecomastia has been reported in 2% of 599 UK patients in a post-marketing surveillance study of flutamide, the first antiandrogen to be introduced ⁴⁷.

Non-Pharmacologic Events

Gastrointestinal Events

Adverse events affecting the gastrointestinal (GI) system are reported with all three non-steroidal antiandrogens, but there are differences between the drugs with respect to the nature, frequency, and severity of such events.

Nausea, Vomiting and Abdominal Discomfort

In combination studies of nilutamide plus surgical castration ^{13, 45, 48} the 3%-13% incidence of these events is comparable with the 4%-20% incidence reported with combinations of flutamide plus medical castration ^{4, 6, 7, 10}. The incidences of nausea and general GI problems were also similar in both groups in a direct comparison of nilutamide plus surgical castration versus bicalutamide 150 mg monotherapy (5% versus 3% and 8% versus 7%, respectively) ³¹.

In comparative studies, all the nonsteroidal antiandrogens, whether used as monotherapy or part of CAB, generally give rise to slightly more nausea, vomiting, and GI distress than castration alone ^{6, 7, 10, 13, 29, 30, 45, 48}.

In the double-blind comparison of flutamide versus bicalutamide, each in combination with an LHRH-A ²², nausea was reported, regardless of causality, in over 10% of patients in both groups and considered related to therapy in 7% and 8% of patients, respectively ⁴³.

Diarrhea

In large studies comparing flutamide plus LHRH-A with LHRH-A alone, diarrhea was more common in the flutamide groups ^{4, 7, 8, 10}. The difference achieved statistical significance in one such double-blind, placebo-controlled study (14% versus 5%; p < 0.001) ⁴. Although there is no access to the underlying data, the transcript of the proceedings of the 1988 FDA Advisory Committee Meeting ⁵⁰ included discussion of data from this trial; the incidence of diarrhea, irrespective of relationship to therapy, was presented as 24% in the flutamide group compared with 11% in the placebo group. The difference between the two reports of the same study illustrates the importance of understanding the reporting methodology. The FDA figures confirm the incidence of diarrhea reported in other flutamide studies ^{22, 35, 42, 49}.

In contrast, diarrhea has been reported in only 2%-5% of patients in bicalutamide monotherapy studies ^{26,27,28,29,30,31}, which include one high-dose (150 mg) study ³¹, and in only 2%-4% of patients in nilutamide studies, which include monotherapy and combinations with surgical castration ^{31, 32, 49}. A higher incidence (10%) was seen when bicalutamide was combined with medical castration, but this was significantly lower (p < 0.001) than the 24% incidence that occurred with flutamide plus medical castration in this double-blind, comparative study ²². The figures from this study are for adverse events irrespective of relationship to therapy. These data ^{20, 43} illustrate the nature of flutamide-induced diarrhea which led to withdrawal of therapy in 6% of flutamide patients compared with 0.5% of bicalutamide patients (Fig. 1). However, in this study, dose modification was not allowed; patients were withdrawn if they could not tolerate their randomized therapy.

Figure 1. Occurrence of diarrhea in a double-blind comparison of flutamide plus LHRH-A versus bicalutamide plus LHRH-A ²².

It has been suggested that the lactose present in the flutamide formulation may contribute to flutamide-induced diarrhea ⁵¹. Reducing the intake of dairy products or supplementation with oral lactase have therefore been suggested as possible strategies for the management of this side effect. However, lactose intolerance is unlikely to account for more than a small proportion of the reported cases, since its prevalence is very low in North American whites and Northern Europeans ⁵², who constitute the majority of patients in the flutamide studies. It has been reported that 98% of 54 patients who could not tolerate flutamide due to diarrhea did not withdraw because of diarrhea from subsequent therapy with bicalutamide ⁵³.

Hepatic Toxicity

Laboratory Test Abnormalities

Incidences of abnormal liver function test results have been variously reported from 2%-33% in nilutamide groups ^{13, 32, 33, 46} and from 4%-62% in flutamide groups ^{5, 7, 9, 11, 34, 38,39,40,49} in trials of monotherapy and CAB. In a study of LHRH-A plus flutamide or placebo, elevated amino-transferases were significantly more common in the flutamide group than in the placebo group (12% versus 3%; p = 0.01) ⁹. Furthermore, the transcript of the proceedings of the 1988 FDA Advisory Committee Meeting ⁵⁰ included discussion of data from the large study comparing flutamide plus LHRH-A with LHRH-A alone ⁴. When adverse events irrespective of causality were considered, a marginal treatment effect was detected with regard to hepatic adverse events in the flutamide group compared with the placebo group (16% versus 10%; p = 0.05).

In the double-blind, comparative study of flutamide plus LHRH-A versus bicalutamide plus LHRH-A, elevated transaminases occurred in slightly more patients in the flutamide group than in the bicalutamide group (10% versus 6%; p = 0.07). This increased incidence was also seen for patients with greatly elevated (>5normal) transaminase values (2% versus 0.5%) ^{22, 43}.

Clinical Hepatotoxicity

One case of acute, reversible hepatitis and one case of fatal, fulminant hepatitis have been described in the literature for nilutamide ^{54, 55}. In the latter, the rapidly fatal outcome may have been promoted by phenobarbital, which was administered after the appearance of symptoms and before discontinuation of nilutamide. In addition to these case reports, hepatitis (not defined) has occurred in 2 of 125 (3%) and 1 of 112 (1%) patients in the nilutamide plus castration groups in double-blind comparisons with castration alone and bicalutamide monotherapy, respectively ^{31, 48}.

Flutamide has been described as a potent hepatotoxin in certain patients ⁵⁶ and clinical hepatotoxicity is well documented in clinical trial reports ^{11, 23, 24, 42, 9}, individual case reports, and series of case reports ^{56,57,58,59,60,61,62,63,64,65}. Flutamide has a short half-life (six to eight hours for the active metabolite 2-hydroxyflutamide). However, if patients fail to have hepatic enzymes monitored frequently and use of flutamide continues if the enzymes become elevated, (more than twice the upper limits of normal), clinical hepa-totoxicity and death can result. Wysowski and Fourcroy reported on liver failure and death in 17 patients ⁵⁶. In a large series of 1,091 consecutive patients treated with flutamide in combined therapy, two developed clinical manifestations of liver disease and both had a return to normal liver enzyme levels with discontinuation of flutamide ⁵⁷. It is now recommended that baseline liver tests should be drawn and levels determined frequently thereafter for patients taking flutamide. Presenting symptoms in both non-fatal and fatal cases have generally been nausea, vomiting, lethargy, and jaundice. Laboratory abnormalities have included elevated amino-transferases and bilirubin, and, in some cases, prolonged prothrombin times. The predominant features of biopsies in five patients have been cholestasis and hepatocellular necrosis. The dose of flutamide has generally been 750 mg daily, and duration of treatment has ranged from 5 to 300 days. Workups to exclude other possible causes of hepatotoxicity were negative, and, in the non-fatal cases, prompt discontinuation of flutamide resulted in normalization of laboratory tests and resolution of clinical symptoms even when therapy with LHRH-A was continued.

Wysowski and Fourcroy ⁵⁶ have compared the observed reporting rates for flutamide-induced hepatotoxicity (deaths and hospitalizations) with the expected rates of hospitalization due to liver toxicity in an elderly male population. Between February 1989 and December 1994, the FDA received reports of 20 patients who died and 26 who were hospitalized for hepatotoxicity due to flutamide. This observed rate of approximately 3 per 10,000 treated patients exceeds the expected rate of 2.5 per 100,000 hospitalizations for non-infectious liver injury by 10-fold or more.

Liver function should be monitored periodically in patients on antiandrogens, and therapy should be discontinued if it is persistently abnormal or accompanied by symptoms such as anorexia, nausea, vomiting, fatigue, discolored urine, pruritus, or jaundice.

Cardiovascular Events

The nonsteroidal antiandrogens, in contrast to steroidal antiandrogens such as cyproterone acetate, do not appear to be associated with significant clinical cardiovascular toxicity ^{27, 29, 48}. Although the prescribing information for nilutamide mentions tachycardia and hypertension, while that for bicalutamide mentions angina pectoris and congestive heart failure, the incidences of these events in published studies have not differed significantly from the incidences in the castration-only groups ^{29, 48}.

In an open monotherapy study using 50 mg daily bicalutamide in 267 patients with advanced prostate cancer, the drug did not appear to affect cardiac parameters as assessed by either 12-lead (all patients) or 24-hour (20 patients) ECG recording. Many of the elderly patients in the study had underlying cardiac abnormalities that remained unchanged ²⁷. Indeed, detailed cardiac assessments were made in all the bicalutamide dose-ranging and phase II studies. In contrast to the shortening of the P-R interval and increased heart rate found in preclinical studies with this drug in dogs, these studies in humans revealed no evidence of any such adverse effects ⁴³. In an analysis of the most frequently reported cardiac adverse events for patients who received bicalutamide in controlled clinical studies (hypertension, heart failure, angina pectoris, myocardial infarction), the incidence was comparable with the prevalence of cardiovascular diseases in the elderly male population ⁶⁶. In a comparison of high-dose (150 mg) bicalutamide against nilutamide plus castration, cardiovascular adverse events occurred with a similar frequency in both groups (7% versus 8%) ³¹.

The prescribing information for flutamide mentions hypertension in 1% of patients and also refers to preclinical studies in beagle dogs in which serious cardiac lesions indicative of chronic injury and repair processes occurred. These included chronic myxomatous degeneration, intra-arterial fibrosis, myocardial acidophilic degeneration, vasculitis and perivasculitis. The doses at which these lesions occurred were associated with levels of the active metabolite 2-hydroxyflutamide that were 1- to 12-fold greater than those observed in humans at therapeutic levels.

While serious cardiac lesions have not been reported in humans, one study comparing flutamide with the estrogen estramustine phosphate found no difference between the two groups with respect to adverse cardiovascular events ⁶⁷. In a double-blind comparison of flutamide monotherapy versus diethylstilbestrol, diethylstilbestrol produced a higher incidence and severity of cardiovascular and thromboembolic complications. However, presumed treatment-related cardiovascular or thromboembolic toxicity was reported in the flutamide group, and was severe, life-threatening, or fatal in 18% ⁶⁸. Specifically, hypertension occurred in 13 (38%) patients and was severe or life-threatening in 4 (12%), fatal stroke occurred in 1 (3%) patient, peripheral edema in 9 (26%) and angina pectoris in 2 (6%). The investigators may have been more inclined to assess causality as related to therapy in these double-blind studies because the estrogen comparators are known to be associated with significant cardiovascular toxicity. This might explain the contrast between these findings and those of the majority of published studies of flutamide in which cardiovascular toxicity is seldom among the most commonly reported adverse events. The transcript of the proceedings of the 1988 FDA Advisory Committee Meeting ⁵⁰ included discussion of data from a large study comparing flutamide plus LHRH-A with LHRH-A alone ⁴. When adverse events irrespective of causality were considered, a marginal treatment effect was detected with respect to cardiovascular experiences in the flutamide group compared with the placebo group (12% versus 7%; p = 0.05). In the double-blind comparison of flutamide versus bicalutamide, each in combination with an LHRH-A ²², cardiovascular adverse events occurred in a similar proportion of patients in both groups when all adverse events were considered, irrespective of causality (Table 2).

Hematological Toxicity

Similarly, anemia was slightly less common in the nilutamide-plus-orchiectomy group than in the placebo-plus-orchiectomy group (7% versus 4%) in one comparative study ¹³. In bicalutamide monotherapy and combination therapy studies, anemia has occurred in 7%-8% of patients, compared with a 7% incidence seen with castration alone ⁴³. In the double-blind comparison of flutamide or bicalutamide, each in combination with an LHRH-A, the incidence of anemia in the flutamide group was slightly higher than that in the bicalutamide group (10% versus 7%) ⁴³. This low incidence contrasts with a higher incidence reported in two uncontrolled studies of flutamide plus LHRH-A when used in addition to radiotherapy ^{49, 70}.

The prescribing information for flutamide lists hemolytic anemia, macrocytic anemia, methemoglobinemia, leukopenia, neutropenia, and thrombocytopenia, but only methemoglobinemia ^{71, 72}, sulphemoglobinemia ⁷³ and neutropenia ⁷⁴ have been described in the literature. A recent update to the prescribing information for flutamide requires methemoglobin levels to be monitored in patients who are susceptible to aniline toxicity and in patients who smoke. Although all three of the currently available nonsteroidal antiandrogens are structurally related to aniline, amide hydrolysis of bicalutamide does not occur in humans who are not exposed to an aniline derivative of bicalutamide ⁷⁵.

Ocular Events