Cardiac and cognitive effects of androgen deprivation therapy : are they real ?

With androgen deprivation therapy being used ever earlier and longer in the course of prostate cancer, concerns have emerged about a variety of adverse effects, including cardiovascular disease and cognitive dysfunction. Conflicting data in both areas have led to controversy and confusion. Here, we review published data in an attempt to clarify those issues.


INTRODUCTION
Androgen deprivation therapy (adt) has been used to treat prostate cancer since 1941 1 .With expanding use of the therapy, data suggest that 40%-50% of men diagnosed with prostate cancer in the current era will receive adt at some point 2,3 .With increasing experience and use of adt earlier (and for longer) in the disease course, concerns have emerged about a variety of adverse effects, including cardiovascular disease 4 and cognitive dysfunction 5 .However, conflicting data in both areas have led to controversy and confusion.Here, we review published data in an attempt to clarify those issues.

Effects of ADT on Cardiovascular Morbidity and Mortality
In theory, adt increases the risk of cardiovascular events by inducing or aggravating some of the established risk factors for cardiovascular diseases, such as increased low-density lipoprotein (ldl), increased body fat, and reduced insulin sensitivity.A number of studies have demonstrated increased percentage body fat mass in patients initiating adt [6][7][8][9] , often within 3-6 months of use 10 .Insulin sensitivity, hyperglycemia, and a greater incidence of the development of diabetes mellitus have also been demonstrated 2,[11][12][13][14][15] , with some suggestion of greater risk with prolonged adt use 16 .In addition, numerous studies have shown that adt increases levels of serum triglycerides and ldl 8,17,18 .Collectively, these metabolic changes would lead to an increased risk of cardiovascular events in adt users.
Several retrospective studies have shown a positive correlation between adt and both morbidity and mortality from cardiovascular events.The strongest evidence comes from two population-based trials using the Surveillance, Epidemiology, and End Results (seer) database 2,13 .The first paper, an observational study of a population-based cohort of 73,196 patients with localized prostate cancer, found that adt was associated with a greater risk of coronary artery disease, myocardial infarction (mi), and sudden cardiac death.For the adt group, as compared with a matched non-adt control group, adjusted hazard ratios (ahrs) for mi, coronary artery disease, and sudden cardiac death were 1.11 [95% confidence interval (ci): 1.01 to 1.21; p = 0.03], 1.16 (95% ci: 1.10 to 1.21; p < 0.001), and 1.16 (95% ci: 1.05 to 1.27; p = 0.004) respectively 13 .A second, more recent study by the same authors, using current seer data, found similar results 2 : among 37,443 men with nonmetastatic prostate cancer, 14,597 of whom used adt (median of 2.6 years of follow-up), the risk of mi (ahr: 1.28; 95% ci: 1.08 to 1.52) and sudden cardiac death (ahr: 1.19; 95% ci: 1.10 to 1.28) were increased.A third study, by Saigal et al., also using the seer database, reported that compared with 18% of adt nonusers, 24% of men undergoing adt experience a cardiovascular event (not well-defined in the manuscript) within 1-4 years of starting therapy (ahr: 1.2; p < 0.001) 19 .Those investigators also found a positive correlation between duration of adt use and risk of cardiovascular events.
In addition to evidence from the above three studies that used administrative data, one study used a prospectively maintained clinical database, capsure (Cancer of the Prostate Strategic Urologic Research Endeavor).Among 4892 patients undergoing either radical prostatectomy or 1 of 3 types of radiation, adt Cardiac and cognitive effects of androgen deprivation therapy: are they real?use was associated with an increased risk of death from cardiovascular disease.After a median follow up of 3.8 years, the ahr for adt users compared with nonusers was 2.6 (p = 0.002) among radical prostatectomy patients, but only 1.2 among radiation therapy patients (p = 0.40).The total number of events was relatively small at 131 20 .
Supportive evidence also comes from several randomized controlled trials (rcts).In a secondary analysis of pooled data from three rcts of short-term adjuvant adt in men undergoing radiation, D'Amico et al. found that 6 months of adt was associated with a shorter time to fatal mi (p = 0.017), but that the risk seemed to be confined to men over the age of 65 21 .Finally, one prospective rct of radiation therapy with or without 6 months of adt in 206 patients showed that adt increases the risk of cardiac death, but only in patients with moderate-to-severe comorbidities 22 .Notably, few cardiovascular events were observed in these studies, which are summarized in Table i.
In contrast, several studies have demonstrated no relationship between adt and cardiovascular events, also summarized in Table i.Three large phase iii rcts [Radiation Therapy Oncology Group (rtog) 85-31, rtog 86-10, and rtog 92-02)] with sample sizes ranging from 456 to 1554 men and with follow-up of 8-10 years (Table ii) all found no statistically significantly increased risk of cardiovascular mortality with adt use 24,25,27 .Although each of the studies used adt in a different fashion in conjunction with externalbeam radiotherapy (adjuvant vs. salvage adt in rtog 85-31, adt vs. no adt in rtog 86-10, and 4 months vs. 24 months of adjuvant adt in rtog 92-02), the results were remarkably similar across studies.However, nonfatal mis were not recorded, and relatively few fatal mis were observed in each trial.
Two additional rcts led by European groups [European Organisation for Research and Treatment of Cancer (eortc) 22961 and 30891] not only failed to show any increased risk of cardiovascular toxicity from adt, but both studies suggested a trend toward fewer fatal cardiovascular events in adt users 23,26 .In eortc 22961, in which 1113 men with locally advanced prostate cancer received either 6 months or 36 months of adt along with external-beam radiotherapy, 4% of men in the 6-month adt arm experienced cardiovascular mortality as compared with 3% in the 36-month adt arm.In eortc 30891, 985 men who were not suitable to undergo radical therapy were randomized to immediate or deferred adt.Cardiovascular mortality was 17.9% in the immediate adt arm compared with 19.7% in the deferred adt arm (Table i).
Lastly, a large retrospective cohort study recently conducted by our group used administrative data from Ontario, Canada.Among 19,079 matched pairs of adt users and nonusers 66 years of age or older, after extensive adjustment for comorbidities, medication use, socioeconomic factors, and prior treatment, the ahrs for mi and sudden cardiac death were 0.91 (95% ci: 0.84 to 1.00) and 0.96 (95% ci: 0.83 to 1.10) respectively 12 .

Reconciling Differences in Study Findings
How does one reconcile these disparate findings?
Although the highest-quality evidence should be obtained from rcts, the nine rcts to date provide conflicting results.Although the significant differences in study populations and the varying followups across trials might account for the different results, another key factor is the relatively small number of events observed in many studies.Additionally, most studies recorded only cardiovascular mortality, which is obviously clinically important but subject to attribution bias and considerably less common than nonfatal mi.Most of the rcts did not report results stratified by age or comorbidity, but at least one study has suggested that adt-related cardiovascular toxicity is observed predominantly in older adults 40 .Even in the large administrative databases that have sample sizes sufficient and follow-up adequate enough to capture nonfatal mis (as can be seen in Table i), the inconsistent results can only partly be explained by differences in study methodology or definitions of endpoints.Clearly, further studies-including additional analyses of nonfatal cardiovascular endpoints from rcts and prospective cohort studies-are needed.Meanwhile, what are clinicians to do?
Clearly, adt increases the risk of diabetes and leads to unfavourable alterations in lipid profile.Thus, it seems reasonable to obtain a fasting blood glucose and lipid profile before initiating adt and to ensure that patients with metabolic abnormalities are treated based on established guidelines.Screening for diabetes is particularly important, because a significant proportion of diabetics are asymptomatic and unaware of their diagnoses 41 .It is also reasonable to repeat blood glucose and lipid testing every 1-2 years while a man remain on adt, or sooner if he experiences a cardiovascular event.
It is equally important to consider whether the cumulative risks of adt (on outcomes ranging from quality of life to fractures) are outweighed by benefits with respect to cancer control and cancerspecific survival.For men with non-high-risk localized prostate cancer or biochemical relapse after radical therapy in particular, the benefits of adt are far from established 42,43 , but the risks are independent of adt indication.Based on a recent scientific advisory from the American Heart Association, the American Cancer Society, and the American Urological Association, there is no indication for additional cardiovascular evaluation (for example, stress-testing or echocardiography) in men initiating adt beyond the screening already mentioned, a careful clinical assessment, and a consideration of both the risks and the benefits of adt 4 .

Effects of ADT on Cognitive Function
A significant body of literature involving both animal and human studies of hypogonadism treated with testosterone replacement therapy points to cognitive effects of testosterone [44][45][46][47][48] .Neuroanatomic studies have demonstrated both testosterone and estrogen receptors in varying densities in various parts of the brain 49 .Thus, there is good reason to observe cognitive effects with adt, which leads to profound reductions in testosterone and a shift in the testosterone:estrogen ratio.In particular, based on studies in both animals and older men, alterations in testosterone may preferentially affect working memory and visuospatial function more than they affect other cognitive domains 5,49 .
A dozen studies have been published examining the effects of adt on a variety of cognitive domains in men with prostate cancer.Most of these studies were summarized in a recent review, in which Nelson et al. concluded that adt leads to subtle but significant cognitive declines in men with prostate cancer, particularly in the areas of visuospatial ability and executive functioning 5 .However, those conclusions deserve further scrutiny, given the significant methodologic limitations of published studies and the contrasting findings across studies.
Table ii summarizes the published studies, including one that was published after the review by Nelson et al. 39 .In general, studies have included men with nonmetastatic prostate cancer receiving either continuous or intermittent adt.Most studies have included a control group of healthy age-and education-matched men.Studies have generally been small, with sample sizes ranging from 18 to 62 participants.They also have generally been longitudinal and observational in nature, with 2-3 assessments over a 6-to 12-month period.Cognitive assessments have ranged from a traditional battery of pen-andpaper tasks to customized cognitive testing software.Few studies have adjusted for multiple statistical comparisons, a potentially significant issue given the large number of cognitive tests (typically 6-15 different tests); and even fewer studies have adjusted for practice effects (subjects become more familiar with tests over time, leading to better performance in subsequent test sessions) 50,51 .
Considering the findings as summarized in Table ii, most studies have reported few statistically significant differences between adt users and controls, which is a bit surprising given the number of cognitive and statistical tests involved.Moreover, although several studies have reported worse performance in verbal memory and visuospatial abilities 28,29,34,36,37,39 , others have reported no such differences 29,30,33,38 .Several studies reported worse processing speed 28,36 , but those results were not confirmed in other studies 30,33 .At least one study reported worse verbal fluency 36 , but at least one study reported improved verbal fluency with adt use 30 .
Thus, based on our review of published studies, we do not believe that valid conclusions can yet be drawn about cognitive effects of adt.We recently presented results on cognitive outcomes from a matched prospective cohort study of 84 men with nonmetastatic prostate cancer on adt and 2 control groups (adt nonusers with prostate cancer and healthy controls), all matched on age and education and assessed at 3 time points over 12 months.Using a variety of analytic methods, we found no consistent effect of adt on 14 different cognitive tests across 6 cognitive domains 52 .
No studies have examined whether adt leads to impaired function in daily tasks such as financial management, medication management, or shopping.Similarly, no reported study has examined whether adt affects self-reported cognitive function or leads to dementia.Arguably, these endpoints are more clinically relevant than are subtle cognitive impairments that might be detectable only by detailed neuropsychological assessment.Additionally, cognitive effects of adt, if present, may require a follow-up longer than 12 months.Our study is continuing to follow our 3 cohorts of men for 2 more years; it may be able to shed further light in this regard.

SUMMARY
Although adt has a variety of unfavourable metabolic side effects that increase the risk of cardiovascular toxicity, published studies using a variety of designs have demonstrated inconsistent effects of adt on cardiovascular endpoints ranging from mi to cardiovascular mortality.Far fewer, and methodologically weaker, studies have examined effects of adt on cognitive function, and the results are even less conclusive.Further studies are clearly needed.
In the interim, clinicians should screen for, and manage, cardiovascular risk factors such as diabetes and hyperlipidemia according to the usual guidelines for men without prostate cancer, and should carefully consider the risks and benefits of adt before initiating it.With respect to cognition, the evidence is insufficient to determine whether adt has any effects.Patients with complaints of memory loss or cognitive dysfunction should therefore be referred back to their primary care physician or to a relevant specialist (neurologist, geriatrician) for further evaluation-a recommendation that is no different than it would be for a general population of older adults with such complaints.

ACKNOWLEDGMENTS
Dr. Alibhai is a research scientist with the Canadian Cancer Society.

table i
Studies of the cardiovascular (cv) effects of androgen deprivation therapy (adt) in men with prostate cancer (pca)

table ii
Studies of the cognitive effects of androgen ablation therapy in men withprostate cancer (pca) wms-r = Wechsler Memory Scale-Revised; avlt = Auditory Verbal Learning Test, tmt = Trail Making Test; cowat = Controlled Word Association Test; wais iii = Wechsler Adult Intelligence Scale iii; camcog = Cambridge Examination for Mental Disorders; wms iii = Wechsler Memory Scale iii; nart = National Adult Reading Scale; kaca = Kendrick Assessment of Cognitive Ageing; hscs = High Sensitivity Cognitive Screen.