Neutral Associations of Testosterone, Dihydrotestosterone and Estradiol With Fatal and Non-fatal Cardiovascular Events, and Mortality in Men Aged 17–97 Years

Yi X. Chan; Matthew W. Knuiman; Joseph Hung; Mark L. Divitini; John P. Beilby; David J. Handelsman; Jonathan Beilin; Brendan McQuillan; Bu B. Yeap


Clin Endocrinol. 2016;85(4):575-582. 

In This Article

Abstract and Introduction


Context Lower testosterone (T) is associated with poorer health outcomes in older men, however, the relationship between T, dihydrotestosterone (DHT) and estradiol (E2) with cardiovascular disease (CVD) in younger to middle-aged men remains unclear.

Objectives We assessed associations between endogenous sex hormones with mortality (all-cause and CVD) and CVD events, in a cohort of men aged 17–97 years.

Participants and methods Sex hormones were assayed using mass spectrometry in 2143 men from the 1994/5 Busselton Health Survey. Outcomes to December 2010 were analysed.

Results Of the 1804 men included in the analysis, mean age was 50·3 ± 16·8 years and 68·9% of men were aged <60. Mean follow-up period was 14·9 years. There were 319 deaths, 141 CVD deaths and 399 CVD events. Compared to the full cohort, men who died had lower baseline T (12·0 ± 4·4 vs 13·6 ± 4·9 nmol/l), free T (181·9 ± 52·9 vs 218·3 ± 63·8 pmol/l) and DHT (1·65 ± 0·64 vs 1·70 ± 0·72 nmol/l), but higher E2 (64·0 ± 32 vs 60·1 ± 30·2 pmol/l). After adjustment for risk factors, T was not associated with mortality (adjusted HR = 0·90, 95% CI 0·79–1·04; P = 0·164 for every increase in 1 SD of T), CVD deaths (adjusted HR = 1·04, 95% CI 0·84–1·29; P = 0·708) or CVD events (adjusted HR = 1·03, 95% CI 0·92–1·15, P = 0·661). No associations were found for free T, DHT or E2. Results were similar for men older and younger than 60 years.

Conclusions In predominantly middle-aged men, T, DHT and E2 do not influence mortality or CVD outcomes. This neutral association of hormones with CVD contrasts with prior studies of older men.


Cardiovascular disease (CVD) is a major contributor to morbidity and mortality, particularly with increasing age.[1] Testosterone (T) levels decrease with age, comorbidities and obesity,[2] and epidemiological studies have reported an inverse relationship between T levels and mortality risk.[3] However, existing randomized controlled trials (RCTs) of testosterone replacement therapy (TRT) have not been powered for cardiovascular end-points. One RCT was terminated early due to adverse cardiovascular-related effects seen with T supplementation in older men with limited mobility.[4] However, higher initiation doses of testosterone gel was administered to men in that study compared to conventional practice.[4] Furthermore, no adverse signals were seen in other large RCTs of TRT.[5,6] Meta-analyses generally do not show increased risk of CVD events.[7] Furthermore, observational studies suggest that TRT with subsequent normalization of T levels is associated with favourable outcomes.[8,9] The use of TRT has increased in the past decade.[10] Recently, the Food and Drug Administration (FDA) has recommended that men be warned of potential adverse cardiovascular effects associated with TRT.[11] Given the prevalence of CVD, and the increase in use of TRT, the relationship between T and CVD risk requires further clarification.

Epidemiological studies assessing the relationship between endogenous T with mortality and CVD outcomes vary significantly according to population size and age range.[3,12] Acknowledging heterogeneity in previous studies, overall results suggest an inverse relationship between T with CVD events and mortality.[3,13] Liquid chromatography-tandem mass spectrometry (LC-MS) provides a more accurate measure of T compared to immunoassay,[14] therefore some earlier studies may be limited by the use of immunoassay for measurement of sex hormones. Studies using LC-MS measured T have generally reported inverse associations between T and CVD events, however these studies have consisted of older populations.[3,13,15–17] Results from middle-aged populations using LC-MS measured T are conflicting.[18,19] Earlier studies of middle-aged men using T measured by immunoassay have reported no association between T and acute myocardial infarction[20] and an inverse association of free and bioavailable T with CVD deaths at 9 but not 18 years,[21] however these studies consisted of a smaller number of outcome events. Smith et al.[22] reported no association of cortisol:T ratio with CVD events, but did not assess associations of T independent of cortisol. Furthermore, T is converted by 5α-reductase into dihydrotestosterone (DHT), which is a more potent ligand for the androgen receptor, and by aromatase into estradiol (E2).[23] Studies assessing the relationship between DHT and E2 with CVD are relatively limited. Our aim was to assess the relationship between T, DHT and E2 assayed using LC-MS with all-cause and CVD related mortality and with CVD events in community-dwelling men across a wide age span.