madman
Super Moderator
Abstract
Introduction
The cardiovascular (CV) safety of testosterone (T) replacement therapy (TRT) is still conflicting. Recent data suggested a TRT-related increased risk of atrial fibrillation (AF). To systematic review and meta-analyze CV risk related to TRT as derived from placebo controlled randomized trials (RCTs).
Areas covered
An extensive Medline, Embase and Cochrane search was performed. All placebo controlled RCTs reporting data on TRT-related CV safety were considered. To better analyze the role of T on AF, population-based studies investigating the relationship between endogenous circulating T levels and AF incidence were also included and analyzed.
Expert opinion
Out of 3.615, 106 studies were considered, including 8.126 subjects treated with TRT and 7.310 patients allocated to placebo. No difference between TRT and placebo were observed when major adverse CV events were considered. Whereas the incidence of non-fatal arrhythmias and AF was increased in the only trial considering CV safety as the primary endpoint,this was not confirmed when all other studies were considered (MH-OR 1.61[0.84;3.08] and 1.44[0.46;4.46]). Similarly no relationship between endogenous T levels and AF incidence was observed after the adjustment for confounders Available data confirm that TRT is safe and it is not related to an increased CV risk.
1.0 Introduction
During the last three decades, there has been an increasing awareness concerning age-associated testosterone (T) deficiency and its potential treatment. Although the phenomenon of an age associated T decline was supported by several epidemiological studies [1-4], its clinical relevance is still unclear and a matter of intense debate. In fact, it is not clear whether the reduced T levels observed in the aging male contribute to the age-related morbidities and symptoms or whether low T and associated morbidities are concomitant conditions, both associated with the aging process [5].Nonetheless, the pharmaceutical sales of T-containing medications dramatically increased worldwide with a 100-fold rise in thirty years [6]. This trend was clearly apparent in the North American market. In fact, the defined monthly doses per 1000 population per year of T transdermal preparation rose over 11 years from 10 and 10.3 to 385 and 98.5 in Canada and the US, respectively[6]. This figure was also apparent in Northern Europe with a raise from 4.5 to 22.1, but less in Southern Europe, where the increasing trend over 11 years was relatively modest (from 5.6 to 7.2). In North America, the major factors responsible for this phenomenon were not an increased prevalence of T deficiency - which indeed is very similar worldwide - but direct-to-consumer advertising (DTCPA), the presence of Internet pharmacies and intense disease mongering [6]. In fact, T treatment was often promoted as a fountain of youth [5,6]. A survey, conducted in several American market areas, has demonstrated that between 2009 and 2013 exposure to televised DTCPA was associated with greater T testing, new initiation of therapy and, especially, initiation of therapy without prior T testing [7].
To partially limit this T overuse, the US Food and Drug Administration (FDA) cautioned that the benefits and safety of treatment with T products have not been clearly established for the treatment of low T levels due to aging [8,9]. In particular, the FDA stated that testosterone replacement therapy (TRT) should be considered only for men with “classical hypogonadism”, i.e. due to primary or secondary T deficiency resulting from known problems within the testis, pituitary, or hypothalamus [8]. Later on, the concept of non-classical hypogonadism (HG) or functional HG was partially incorporated in the clinical practice guidelines of the US Endocrine Society [10] as well as in the Australian one [11], as opposed to organic (classic) HG. In addition, the FDA issued a safety notification regarding the misuse of T-containing products due to a potential risk of cardiovascular(CV) harm [8]. In particular, FDA recommended against the use of T supplementation in men with a hormone deficiency not due to organic causes but age or comorbidities, i.e. functional HG [8]. Infact, Grossmann & Matsumoto, in a perspective article [12], considered the risks of CV diseases(CVD) as “unknown” in functional HG, while they were “low relative to benefits” in the organic one. This statement is not evidence-based because we are not aware of any trials enrolling subjects with organic HG and facing the risk over the benefit of TRT. Conversely, the majority of available trials with TRT were enrolling subjects with functional HG, such as the recently published Testosterone Treatment in Older Men trials (TTrials; [13]), the Testosterone treatment to prevent or revert type 2 diabetes in men enrolled in a lifestyle program trial (T4DM;[14]), and the Testosterone Replacement Therapy on the Incidence of Major Adverse Cardiovascular Events (MACE) and Efficacy Measures in Hypogonadal Men trial (TRAVERSE; [15]). In the real world, the large majority (85%) of subjects visiting an outpatient clinic for sexual dysfunctions, which represent the most specific complex of symptoms associated with adult-onset HG, are classified as functional HG, as demonstrated by an analysis of more than 4000 individuals [16].
The FDA’s position on possible CV risk associated with TRT was essentially based on three studies published almost 10 years ago [17-19]. Two of them were pharmaco-epidemiological studies[18,19] and one was prematurely interrupted randomized placebo-controlled trial not powered to detect differences in CV and not having CV events as a primary end-point [17]. Specific criticisms of these three studies are reported elsewhere [16]. Of note, the European Medicine Agency (EMA), after conducting a similar review of the data as the FDA, did not find sufficient evidence for declaring TRT to be associated with an increased CV risk [20]. The large majority of the meta analyses conducted so far on randomized placebo-controlled trials (RCT) failed to demonstrate an increase in CV risk in subjects with functional HG receiving TRT [21-29], including a recent individual patient and aggregate meta-analysis [30]. However, studies included in the aforementioned meta-analyses do not have CV events as primary end-points, used a broad definition of them and CV events were often not adjudicated.
Until recently, only few RCTs were available with subjects with CV outcomes as the primary endpoint in patients with coronary heart disease (CHD); the meta-analysis of these trials did not provide any signal for an increase in CV risk associated with T administration [31]. In fact, in the six RCTs meta-analyzed, enrolling 258 patients with CHD with a mean follow-up of 23 weeks,TRT was positively associated with a significant increase in treadmill test duration and time to 1 mm ST segment depression[31]. However, the small number of patients and the short follow-up strongly limited the interpretation of the results.
Considering the intense debate around the safety of TRT, in March 2015 the FDA released a requirement that manufacturers of FDA-approved T therapies conduct prospective studies of sufficient size and duration to evaluate whether their products are associated with an elevated risk of CV events. Hence, the TRAVERSE study was there after prompted [15]. The TRAVERSE study recruited 5246 men aged 45-80, predominantly with pre-existing or high-risk CV disease and a total T<10.4 nmol/L with self-reported sexual dysfunction or other symptoms characteristic of functional HG. Men with congenital or acquired severe organic HG were not included in the study. Patients were randomly allocated to receive transdermal T gel (1.6%) or placebo for an average of 21.7± 14.1months and after 33 months of post-treatment follow-up, with a discontinuation rate of 61% in both groups. The primary CV end-point was the first occurrence of any CV events of a composite of cardiovascular death, nonfatal myocardial infarction, and nonfatal stroke. The secondary outcome was the first recurrence of the composite outcome plus coronary revascularization. The first adjudicated major adverse CV event (MACE) occurred in 182 patients (7.0%) in the T group and in 190 patients (7.3%) in the placebo group. No apparent clinically meaningful differences in the incidence of secondary CV end-point events were observed between the trial groups. Hence, no increase in MACE risk was detected in this large study [15]. However, in the T arm of the TRAVERSE trial non fatal arrhythmias occurred in 134 patients (5.2%) and atrial fibrillation (AF) in 91 patients (3.5%), while the same events in the placebo group occurred in 87 (3.3%) and 63 patients (2.4%), respectively, reaching statistical significance [15].
The aim of this study is to compare results obtained in the TRAVERSE trial on CV risk [15] with those available in other trials, with MACE not as a primary end-point, by using a meta-analytic approach. In addition, it will also be investigated whether, as revealed by the TRAVERSE trial[15], there is a potential negative effect of TRT on nonfatal cardiac arrhythmias by using data from other available trials. A possible association between endogenous T and atrial fibrillation will also be explored in available epidemiological studies.
4.0 Conclusions
Present data show that TRT has a neutral effect on MACE risk, either when TRAVERSE or other (non-TRAVERSE) trials are considered. Similar results were observed when individual CV events composing MACE were considered. In addition, among non-TRAVERSE studies, no difference in MACE risk was observed when several aspects, including subject age, duration of follow-up,reported industry support, baseline T levels, and type of T preparation used, were considered. Conversely, a higher MACE risk was detected in those studies reporting T dosages above the suggested recommendations. Although a lower risk of MACE was observed when trials enrolling
obese (BMI> 30 kg/m2) patients were analyzed in non-TRAVERSE trials, the inclusion of the TRAVERSE results into the meta-analysis did not confirm the association.
The present meta-analysis of epidemiological studies investigating relationships between endogenous T levels and non-fatal arrhythmias suggests a possible protective role for T on AF,although the association was not confirmed in a fully adjusted model or when outliers’ studies were excluded. In contrast, the TRAVERSE study (15) showed an increased, and not a decreased, risk of arrhythmias and, in particular, of AF. It is important to note that, in the TRAVERSE trial [15],cardiac arrhythmias (AF and non-fatal arrhythmias) were not trial endpoints but investigator reported adverse events. When a similar analysis of investigator-reported arrhythmias was performed in other non-TRAVERSE trials, we found a non-significant trend towards an increased risk of TRT-related cardiac arrhythmias (including AF), which was further substantially smoothed when trials involving frail subjects were excluded.
5.0 Expert opinion
The TRAVERSE trial, a properly powered placebo-RCT with a primary CV endpoint specifically designed for middle-aged and older symptomatic men with functional hypogonadism [15], substantially confirms the results of the previous [21-29] and the present (updated) meta-analyses.The odds ratio for AMI, acute coronary syndrome, stroke, heart failure, MACE, and overall CV mortality were equally null for either the present meta-analytic results or the TRAVERSE trial. Although meta-analyses are often considered the highest level of evidence for evaluating interventions in healthcare, they suffer from several methodological issues. In fact, they summarize CV events in RCTs designed for other purposes, with possible inconsistencies in the reporting of adverse events often without prior formal adjudication, along with limited reliability of diagnostic criteria for diagnosing and interpreting CV events. Hence, it is important that the TRAVERSE trial, having MACE as a primary endpoint, confirms present and previous [21-29]meta-analytic results .In fact, the present head-to-head comparison between the TRAVERSE and other meta-analyzed trials substantially halts the previous dubs on the relative CV safety of TRT in middle-aged or older men with symptomatic functional hypogonadism. It is important to note that the TRAVERSE trial enrolled only men with acquired functional HG, with those with congenital or severe acquired HG explicitly excluded. The large majority of the other trials included in the present meta-analysis also enrolled men with characteristics compatible with functional hypogonadism. The present meta analysis supports an even more protective role for TRT against MACE in trials enrolling obese men, as previously reported [28] or suggested by other authors in observational studies [142,143]. Considering that TRT can improve body composition in hypogonadal men by reducing fat mass and improving lean mass [14,144-146], this can explain its protective effect on CV risk in obese individuals. However, the introduction of TRAVERSE data into the meta-analysis did not substantiate the negative association between obesity and CV risk, most probably because the enrolled subjects in the TRAVERSE study were already at high CV risk. It is important to note that all analyzed trials, including the TRAVERSE one, on TRT and CV risk are of relatively short duration (i.e., a mean of 34 and 87 weeks, respectively), not allowing the inferences on longer-term effects of TRT. However, the number of observations collected in the present analysis is relevant. In fact, it includes 2,601 and 2,603 patients in the active and comparator arms of the TRAVERSE trial, along with 5,525 and 4,707 subjects treated with T and placebo, respectively, in the other trials here meta-analyzed. Hence, in total, we here report 15,436 observations from the available RCT on TRT-related CV risk, suggesting the overall safety of the treatment. However, this is not the case for non fatal cardiac arrhythmias, in particular atrial fibrillation.
AF is the most common arrhythmia and affects more than 33 million people worldwide [147]. The prevalence of AF is higher in men than in women probably due to different CV risks that are gender-related [148]. In addition, polymorphic ventricular tachycardia torsades de pointe (TdP) is in some way gender-specific, being less frequent in men than in women [149]. Preclinical studies suggested that T ablation promotes arrhythmias in aged mice by increasing late inward sodium current and prolonging repolarization in mouse ventricular myocytes [150,151]. Accordingly,androgen deprivation therapy (ADT) in patients with prostate cancer increases overall CV risk and the risk of arrhythmias, including TdP [149,152,153]. It is suggested that T shortens the action potential interval duration with a shortening of the QTc interval [151], acting through distinct signaling pathways, including inhibition of L-type calcium current, enhancing IKr and the slow component of the delayed rectifier calcium current (IKs) [154]. In an RCT three-way crossover study, transdermal T attenuates drug-induced QT lengthening in older men [155] by affecting both early and late ventricular repolarization [156].
Epidemiological studies reported conflicting results, most probably because of a U-shaped effect of androgens, as recently demonstrated [138]. In fact, analysis of 173,498 men from a UK Biobank showed that both low and high free T were associated with AF, while bradycardia was more frequent in those with low free T and the reverse was observed for ventricular arrhythmia (VA).The present meta-analysis of available epidemiological studies shows a significant association between low T, and a higher risk of AF, when an age-adjusted model is considered. The latter observation was attenuated in a fully adjusted model, where only a trend for an association between low T and AF was observed. As reported in the TRAVERSE trial [15], the present meta-analytic results show a trend towards an increased risk (p=0.07) of overall arrhythmia for TRT in intervention trials. However, the latter association was further blunted in studies including only hypogonadal men or excluding frail men when TRT was prescribed at dosages above the suggested recommendations. Considering that the TRAVERSE trial enrolled only frail men at high CV risk, it is conceivable that the association between TRT and arrhythmia or AF was more evident in the TRAVERSE trial than in other trials. Finally, as stated before, both non-fatal arrhythmia and AF were not trial endpoints of the TRAVERSE study but only investigator-reported adverse events [15], and hence they should be interpreted with caution. It is our expert opinion that a possible association between TRT and cardiac arrhythmias should be more deeply investigated in future dedicated trials, with AF and other arrhythmias as end-points. Overall, the evidence here reported, although caution suggests evaluating the presence of high CV risk when prescribing TRT in hypogonadal men, as those enrolled in the TRAVERSE study, along with a baseline and follow-up ECG and/or cardiological evaluation, to screen for atrial fibrillation.
Several limitations should be recognized. Data derived from studies with non-CV primary endpoints should be interpreted with caution due to the lack of pre defined diagnostic criteria and screening methods for incident CVD and the risk of misdiagnosis and under-diagnosis. However, our analysis shows that the final result is similar when TRAVERSE and non-TRAVERSE studies are considered. The duration of available RCTs, including the TRAVERSE study, is still limited, not allowing inferences on long-term effects. The information related to AF as derived from non TRAVERSE studies is limited and many studies reporting data on the risk of overall arrhythmias do not clarify the specific type of event observed. In addition, the correct assessment of AF can be problematic in many studies if a baseline and post-treatment ECG was not correctly performed. Furthermore, even if an ECG was performed, it should be recognized that AF may be often intermittent and asymptomatic limiting its correct evaluation.
In conclusion, available data show that TRT is not associated with an increased risk of CV events when hypogonadism is properly diagnosed and treated. In fact, the three largest T RCTs reported to date, T Trials [13], T4DM [14], and TRAVERSE [15], showed no signal for adverse CV events, as does the present meta-analysis. The possible advantages of TRT observed in uncomplicated subjects could be blunted when TRT is used in patients at high CV risk. Considering that the majority of subjects involved in the present meta-analyzed trials are classified as having functional HG, it is difficult to extend this CV safety signal to other forms of hypogonadism, including the genetic forms. For instance, in the most common genetic condition, Klinefelter’s Syndrome, TRT was not able to restore the levels of control subjects, altered body composition and glycol-metabolic parameters, as recently reported in a meta-analysis of observational studies [157].
Age-dependent decline of T observed in aging men (the so-called functional hypogonadism) is the result of a combination of genetic background, accumulation of morbidities, and wrong lifestyle behaviors [10,158,159]. Hence, correcting unsafe lifestyle behavior is mandatory in these hypogonadal subjects. Other studies are advisable to better clarify whether or not early identification of hypogonadism and an early TRT approach can reduce long-term CV risk.
Introduction
The cardiovascular (CV) safety of testosterone (T) replacement therapy (TRT) is still conflicting. Recent data suggested a TRT-related increased risk of atrial fibrillation (AF). To systematic review and meta-analyze CV risk related to TRT as derived from placebo controlled randomized trials (RCTs).
Areas covered
An extensive Medline, Embase and Cochrane search was performed. All placebo controlled RCTs reporting data on TRT-related CV safety were considered. To better analyze the role of T on AF, population-based studies investigating the relationship between endogenous circulating T levels and AF incidence were also included and analyzed.
Expert opinion
Out of 3.615, 106 studies were considered, including 8.126 subjects treated with TRT and 7.310 patients allocated to placebo. No difference between TRT and placebo were observed when major adverse CV events were considered. Whereas the incidence of non-fatal arrhythmias and AF was increased in the only trial considering CV safety as the primary endpoint,this was not confirmed when all other studies were considered (MH-OR 1.61[0.84;3.08] and 1.44[0.46;4.46]). Similarly no relationship between endogenous T levels and AF incidence was observed after the adjustment for confounders Available data confirm that TRT is safe and it is not related to an increased CV risk.
1.0 Introduction
During the last three decades, there has been an increasing awareness concerning age-associated testosterone (T) deficiency and its potential treatment. Although the phenomenon of an age associated T decline was supported by several epidemiological studies [1-4], its clinical relevance is still unclear and a matter of intense debate. In fact, it is not clear whether the reduced T levels observed in the aging male contribute to the age-related morbidities and symptoms or whether low T and associated morbidities are concomitant conditions, both associated with the aging process [5].Nonetheless, the pharmaceutical sales of T-containing medications dramatically increased worldwide with a 100-fold rise in thirty years [6]. This trend was clearly apparent in the North American market. In fact, the defined monthly doses per 1000 population per year of T transdermal preparation rose over 11 years from 10 and 10.3 to 385 and 98.5 in Canada and the US, respectively[6]. This figure was also apparent in Northern Europe with a raise from 4.5 to 22.1, but less in Southern Europe, where the increasing trend over 11 years was relatively modest (from 5.6 to 7.2). In North America, the major factors responsible for this phenomenon were not an increased prevalence of T deficiency - which indeed is very similar worldwide - but direct-to-consumer advertising (DTCPA), the presence of Internet pharmacies and intense disease mongering [6]. In fact, T treatment was often promoted as a fountain of youth [5,6]. A survey, conducted in several American market areas, has demonstrated that between 2009 and 2013 exposure to televised DTCPA was associated with greater T testing, new initiation of therapy and, especially, initiation of therapy without prior T testing [7].
To partially limit this T overuse, the US Food and Drug Administration (FDA) cautioned that the benefits and safety of treatment with T products have not been clearly established for the treatment of low T levels due to aging [8,9]. In particular, the FDA stated that testosterone replacement therapy (TRT) should be considered only for men with “classical hypogonadism”, i.e. due to primary or secondary T deficiency resulting from known problems within the testis, pituitary, or hypothalamus [8]. Later on, the concept of non-classical hypogonadism (HG) or functional HG was partially incorporated in the clinical practice guidelines of the US Endocrine Society [10] as well as in the Australian one [11], as opposed to organic (classic) HG. In addition, the FDA issued a safety notification regarding the misuse of T-containing products due to a potential risk of cardiovascular(CV) harm [8]. In particular, FDA recommended against the use of T supplementation in men with a hormone deficiency not due to organic causes but age or comorbidities, i.e. functional HG [8]. Infact, Grossmann & Matsumoto, in a perspective article [12], considered the risks of CV diseases(CVD) as “unknown” in functional HG, while they were “low relative to benefits” in the organic one. This statement is not evidence-based because we are not aware of any trials enrolling subjects with organic HG and facing the risk over the benefit of TRT. Conversely, the majority of available trials with TRT were enrolling subjects with functional HG, such as the recently published Testosterone Treatment in Older Men trials (TTrials; [13]), the Testosterone treatment to prevent or revert type 2 diabetes in men enrolled in a lifestyle program trial (T4DM;[14]), and the Testosterone Replacement Therapy on the Incidence of Major Adverse Cardiovascular Events (MACE) and Efficacy Measures in Hypogonadal Men trial (TRAVERSE; [15]). In the real world, the large majority (85%) of subjects visiting an outpatient clinic for sexual dysfunctions, which represent the most specific complex of symptoms associated with adult-onset HG, are classified as functional HG, as demonstrated by an analysis of more than 4000 individuals [16].
The FDA’s position on possible CV risk associated with TRT was essentially based on three studies published almost 10 years ago [17-19]. Two of them were pharmaco-epidemiological studies[18,19] and one was prematurely interrupted randomized placebo-controlled trial not powered to detect differences in CV and not having CV events as a primary end-point [17]. Specific criticisms of these three studies are reported elsewhere [16]. Of note, the European Medicine Agency (EMA), after conducting a similar review of the data as the FDA, did not find sufficient evidence for declaring TRT to be associated with an increased CV risk [20]. The large majority of the meta analyses conducted so far on randomized placebo-controlled trials (RCT) failed to demonstrate an increase in CV risk in subjects with functional HG receiving TRT [21-29], including a recent individual patient and aggregate meta-analysis [30]. However, studies included in the aforementioned meta-analyses do not have CV events as primary end-points, used a broad definition of them and CV events were often not adjudicated.
Until recently, only few RCTs were available with subjects with CV outcomes as the primary endpoint in patients with coronary heart disease (CHD); the meta-analysis of these trials did not provide any signal for an increase in CV risk associated with T administration [31]. In fact, in the six RCTs meta-analyzed, enrolling 258 patients with CHD with a mean follow-up of 23 weeks,TRT was positively associated with a significant increase in treadmill test duration and time to 1 mm ST segment depression[31]. However, the small number of patients and the short follow-up strongly limited the interpretation of the results.
Considering the intense debate around the safety of TRT, in March 2015 the FDA released a requirement that manufacturers of FDA-approved T therapies conduct prospective studies of sufficient size and duration to evaluate whether their products are associated with an elevated risk of CV events. Hence, the TRAVERSE study was there after prompted [15]. The TRAVERSE study recruited 5246 men aged 45-80, predominantly with pre-existing or high-risk CV disease and a total T<10.4 nmol/L with self-reported sexual dysfunction or other symptoms characteristic of functional HG. Men with congenital or acquired severe organic HG were not included in the study. Patients were randomly allocated to receive transdermal T gel (1.6%) or placebo for an average of 21.7± 14.1months and after 33 months of post-treatment follow-up, with a discontinuation rate of 61% in both groups. The primary CV end-point was the first occurrence of any CV events of a composite of cardiovascular death, nonfatal myocardial infarction, and nonfatal stroke. The secondary outcome was the first recurrence of the composite outcome plus coronary revascularization. The first adjudicated major adverse CV event (MACE) occurred in 182 patients (7.0%) in the T group and in 190 patients (7.3%) in the placebo group. No apparent clinically meaningful differences in the incidence of secondary CV end-point events were observed between the trial groups. Hence, no increase in MACE risk was detected in this large study [15]. However, in the T arm of the TRAVERSE trial non fatal arrhythmias occurred in 134 patients (5.2%) and atrial fibrillation (AF) in 91 patients (3.5%), while the same events in the placebo group occurred in 87 (3.3%) and 63 patients (2.4%), respectively, reaching statistical significance [15].
The aim of this study is to compare results obtained in the TRAVERSE trial on CV risk [15] with those available in other trials, with MACE not as a primary end-point, by using a meta-analytic approach. In addition, it will also be investigated whether, as revealed by the TRAVERSE trial[15], there is a potential negative effect of TRT on nonfatal cardiac arrhythmias by using data from other available trials. A possible association between endogenous T and atrial fibrillation will also be explored in available epidemiological studies.
4.0 Conclusions
Present data show that TRT has a neutral effect on MACE risk, either when TRAVERSE or other (non-TRAVERSE) trials are considered. Similar results were observed when individual CV events composing MACE were considered. In addition, among non-TRAVERSE studies, no difference in MACE risk was observed when several aspects, including subject age, duration of follow-up,reported industry support, baseline T levels, and type of T preparation used, were considered. Conversely, a higher MACE risk was detected in those studies reporting T dosages above the suggested recommendations. Although a lower risk of MACE was observed when trials enrolling
obese (BMI> 30 kg/m2) patients were analyzed in non-TRAVERSE trials, the inclusion of the TRAVERSE results into the meta-analysis did not confirm the association.
The present meta-analysis of epidemiological studies investigating relationships between endogenous T levels and non-fatal arrhythmias suggests a possible protective role for T on AF,although the association was not confirmed in a fully adjusted model or when outliers’ studies were excluded. In contrast, the TRAVERSE study (15) showed an increased, and not a decreased, risk of arrhythmias and, in particular, of AF. It is important to note that, in the TRAVERSE trial [15],cardiac arrhythmias (AF and non-fatal arrhythmias) were not trial endpoints but investigator reported adverse events. When a similar analysis of investigator-reported arrhythmias was performed in other non-TRAVERSE trials, we found a non-significant trend towards an increased risk of TRT-related cardiac arrhythmias (including AF), which was further substantially smoothed when trials involving frail subjects were excluded.
5.0 Expert opinion
The TRAVERSE trial, a properly powered placebo-RCT with a primary CV endpoint specifically designed for middle-aged and older symptomatic men with functional hypogonadism [15], substantially confirms the results of the previous [21-29] and the present (updated) meta-analyses.The odds ratio for AMI, acute coronary syndrome, stroke, heart failure, MACE, and overall CV mortality were equally null for either the present meta-analytic results or the TRAVERSE trial. Although meta-analyses are often considered the highest level of evidence for evaluating interventions in healthcare, they suffer from several methodological issues. In fact, they summarize CV events in RCTs designed for other purposes, with possible inconsistencies in the reporting of adverse events often without prior formal adjudication, along with limited reliability of diagnostic criteria for diagnosing and interpreting CV events. Hence, it is important that the TRAVERSE trial, having MACE as a primary endpoint, confirms present and previous [21-29]meta-analytic results .In fact, the present head-to-head comparison between the TRAVERSE and other meta-analyzed trials substantially halts the previous dubs on the relative CV safety of TRT in middle-aged or older men with symptomatic functional hypogonadism. It is important to note that the TRAVERSE trial enrolled only men with acquired functional HG, with those with congenital or severe acquired HG explicitly excluded. The large majority of the other trials included in the present meta-analysis also enrolled men with characteristics compatible with functional hypogonadism. The present meta analysis supports an even more protective role for TRT against MACE in trials enrolling obese men, as previously reported [28] or suggested by other authors in observational studies [142,143]. Considering that TRT can improve body composition in hypogonadal men by reducing fat mass and improving lean mass [14,144-146], this can explain its protective effect on CV risk in obese individuals. However, the introduction of TRAVERSE data into the meta-analysis did not substantiate the negative association between obesity and CV risk, most probably because the enrolled subjects in the TRAVERSE study were already at high CV risk. It is important to note that all analyzed trials, including the TRAVERSE one, on TRT and CV risk are of relatively short duration (i.e., a mean of 34 and 87 weeks, respectively), not allowing the inferences on longer-term effects of TRT. However, the number of observations collected in the present analysis is relevant. In fact, it includes 2,601 and 2,603 patients in the active and comparator arms of the TRAVERSE trial, along with 5,525 and 4,707 subjects treated with T and placebo, respectively, in the other trials here meta-analyzed. Hence, in total, we here report 15,436 observations from the available RCT on TRT-related CV risk, suggesting the overall safety of the treatment. However, this is not the case for non fatal cardiac arrhythmias, in particular atrial fibrillation.
AF is the most common arrhythmia and affects more than 33 million people worldwide [147]. The prevalence of AF is higher in men than in women probably due to different CV risks that are gender-related [148]. In addition, polymorphic ventricular tachycardia torsades de pointe (TdP) is in some way gender-specific, being less frequent in men than in women [149]. Preclinical studies suggested that T ablation promotes arrhythmias in aged mice by increasing late inward sodium current and prolonging repolarization in mouse ventricular myocytes [150,151]. Accordingly,androgen deprivation therapy (ADT) in patients with prostate cancer increases overall CV risk and the risk of arrhythmias, including TdP [149,152,153]. It is suggested that T shortens the action potential interval duration with a shortening of the QTc interval [151], acting through distinct signaling pathways, including inhibition of L-type calcium current, enhancing IKr and the slow component of the delayed rectifier calcium current (IKs) [154]. In an RCT three-way crossover study, transdermal T attenuates drug-induced QT lengthening in older men [155] by affecting both early and late ventricular repolarization [156].
Epidemiological studies reported conflicting results, most probably because of a U-shaped effect of androgens, as recently demonstrated [138]. In fact, analysis of 173,498 men from a UK Biobank showed that both low and high free T were associated with AF, while bradycardia was more frequent in those with low free T and the reverse was observed for ventricular arrhythmia (VA).The present meta-analysis of available epidemiological studies shows a significant association between low T, and a higher risk of AF, when an age-adjusted model is considered. The latter observation was attenuated in a fully adjusted model, where only a trend for an association between low T and AF was observed. As reported in the TRAVERSE trial [15], the present meta-analytic results show a trend towards an increased risk (p=0.07) of overall arrhythmia for TRT in intervention trials. However, the latter association was further blunted in studies including only hypogonadal men or excluding frail men when TRT was prescribed at dosages above the suggested recommendations. Considering that the TRAVERSE trial enrolled only frail men at high CV risk, it is conceivable that the association between TRT and arrhythmia or AF was more evident in the TRAVERSE trial than in other trials. Finally, as stated before, both non-fatal arrhythmia and AF were not trial endpoints of the TRAVERSE study but only investigator-reported adverse events [15], and hence they should be interpreted with caution. It is our expert opinion that a possible association between TRT and cardiac arrhythmias should be more deeply investigated in future dedicated trials, with AF and other arrhythmias as end-points. Overall, the evidence here reported, although caution suggests evaluating the presence of high CV risk when prescribing TRT in hypogonadal men, as those enrolled in the TRAVERSE study, along with a baseline and follow-up ECG and/or cardiological evaluation, to screen for atrial fibrillation.
Several limitations should be recognized. Data derived from studies with non-CV primary endpoints should be interpreted with caution due to the lack of pre defined diagnostic criteria and screening methods for incident CVD and the risk of misdiagnosis and under-diagnosis. However, our analysis shows that the final result is similar when TRAVERSE and non-TRAVERSE studies are considered. The duration of available RCTs, including the TRAVERSE study, is still limited, not allowing inferences on long-term effects. The information related to AF as derived from non TRAVERSE studies is limited and many studies reporting data on the risk of overall arrhythmias do not clarify the specific type of event observed. In addition, the correct assessment of AF can be problematic in many studies if a baseline and post-treatment ECG was not correctly performed. Furthermore, even if an ECG was performed, it should be recognized that AF may be often intermittent and asymptomatic limiting its correct evaluation.
In conclusion, available data show that TRT is not associated with an increased risk of CV events when hypogonadism is properly diagnosed and treated. In fact, the three largest T RCTs reported to date, T Trials [13], T4DM [14], and TRAVERSE [15], showed no signal for adverse CV events, as does the present meta-analysis. The possible advantages of TRT observed in uncomplicated subjects could be blunted when TRT is used in patients at high CV risk. Considering that the majority of subjects involved in the present meta-analyzed trials are classified as having functional HG, it is difficult to extend this CV safety signal to other forms of hypogonadism, including the genetic forms. For instance, in the most common genetic condition, Klinefelter’s Syndrome, TRT was not able to restore the levels of control subjects, altered body composition and glycol-metabolic parameters, as recently reported in a meta-analysis of observational studies [157].
Age-dependent decline of T observed in aging men (the so-called functional hypogonadism) is the result of a combination of genetic background, accumulation of morbidities, and wrong lifestyle behaviors [10,158,159]. Hence, correcting unsafe lifestyle behavior is mandatory in these hypogonadal subjects. Other studies are advisable to better clarify whether or not early identification of hypogonadism and an early TRT approach can reduce long-term CV risk.
