Comment on ‘Anabolic–androgenic steroids and cardiac function’

We appreciate Dr Victor Čulić for his “Letter to Editor” and his interest in our research paper ‘Severe biventricular cardiomyopathy in both current and former long-term users of anabolic–androgenic steroids’.^1,2 In our cross-sectional study, we compared 101 weightlifting AAS users with at least one year of cumulative AAS use (mean 11 ± 7 accumulated years of AAS use) to 71 non-using weightlifting controls (WLC). We showed that long-term AAS users compared to WLC had signs of biventricular cardiomyopathy with worse LV ejection fraction (LVEF) (49 ± 7% vs. 59 ± 5%, P < 0.001) and compromised right ventricular function by global longitudinal strain (RVGLS) (−17.3 ± 3.5 vs. −22.8 ± 2.0%, P < 0.001). The reduced systolic function was also found in former users. Regression analyses of the total population showed that the strongest determinant of reduced LVEF was neither coexisting strength training nor hypertension, but history of AAS use (β −0.53, P < 0.001). Importantly, these cardiac impairments persisted even among former AAS users, emphasizing the enduring impact of AAS use on cardiovascular health evident in former users that had quit 6 ± 6 years earlier [median 3.75 (1–25)].

Čulić highlights the modest impairment in left ventricle diastolic function observed among AAS users in our study. While the differences in the E/A and E/e′ ratios were significant, they were within the range as reported in earlier studies.³ Moreover, according to the ASE/EACVI guidelines, only 3 out of 101 AAS users (3%) had Grade II left ventricle diastolic dysfunction (LVDD) in our study, and no users had Grade I or III.⁴ We did not elaborate further on LVDD in the manuscript, but we welcome Čulić’s input.

Myocardial hypertrophy, a known factor that may induce LVDD,⁴ was evident in our study and comparable with patients with hypertensive heart disease, as reported in a study of Lamb et al.⁵ Although the AAS users in our study had increased blood pressure, the hypertensive heart disease patients in Lamb’s study had more pronounced LVDD. Direct comparison between a weightlifting population and a patient population poses challenges, and the disparity could suggest that AAS users may have lower risk of developing LVDD despite myocardial hypertrophy and hypertension.

Čulić discusses factors that may have favourable effects on the diastolic function. Athletes have normal or supranormal diastolic filling parameters, but studies have shown that the supranormal effect is more pronounced in endurance athletes than in power athletes.^6,7

While medical use of testosterone may have some favourable effects on diastolic dysfunction in some subpopulations, it is unclear whether this is applicable to otherwise healthy power athletes, especially considering the supraphysiological doses of AAS that are used. Long-term use of supraphysiological AAS doses impacts the hypothalamic–pituitary–gonadal axis, compromising testosterone biosynthesis, are associated with negative correlations between AAS use and brain volume and cortical thickness, and with elevated levels of inflammatory markers relevant to cardiovascular pathology.^8,9 Hence, while medical use of testosterone may offer benefits in select subpopulations, it is speculative to suggest that a mean weekly AAS dose of around 1000 mg/week would confer such protective effects.

In our study, all the participants were younger than 50 years of age, and AAS users were at the same age as the WLC`s, indicating that the age-effect on diastolic function was negligible.

Among AAS users, history of smoking, amphetamine, and cocaine use was more common. We did not perform additional regression analyses for determinants of slightly reduced E/A and increased E/e′ ratios, but in the multivariate linear regression analyses for determinants of reduced systolic function, neither cocaine nor amphetamine was significant determinants, only AAS use itself and increased left ventricular mass index. Studies have shown that cocaine and amphetamine impact LVDD negatively,¹⁰ and as Čulić stresses, these factors may have contributed to the reduced E/A and increased E/e′ ratios, in addition to hypertension and myocardial hypertrophy.

Table 1 shows the results of the diastolic parameters when excluding three AAS users (two current and one former) with LVDD Grade II. The new analyses show only minor differences.

Table 1

When excluding three AAS users with left ventricle diastolic dysfunction Grade II, of which two current and one former AAS users, the total AAS-population consists of 98 participants

	AAS users (n = 98)	Current users (n = 67)	Former users (n = 29)	WLC (n = 71)
E/A ratio	1.24 ± 0.34a	1.21 ± 0.34c	1.29 ± 0.35	1.49 ± 0.42
E/e′ ratio	6.5 ± 1.8a	6.6 ± 1.8*	6.3 ± 1.8	6.0 ± 1.2

The table shows comparisons of left ventricle diastolic dysfunction parameters in the population of 98 AAS users, and in current or former AAS users, with 71 non-using WLC. Ninety-six AAS users were available for classification into current (n = 67) and former (n = 29) AAS users. Data are presented as mean ± SD and n (%) when appropriate. Superscripts a, b, c, and d mean significant difference (P < 0.05) when comparing the total population of 101 AAS users vs. 71 WLC; current users vs. former users; current users vs. WLC; and former users vs. WLC, respectively. If no superscript, there was no significant difference.

AAS, anabolic–androgenic steroids; WLC, weightlifting controls; E, maximum E wave velocity at transmitral blood flow Doppler; A, maximum A wave velocity at transmitral blood flow Doppler; e′, maximum E wave velocity at tissue Doppler imaging.

^*When comparing E/e′ ratio in current users vs. WLC, P = 0.07.

Reflecting the multifactorial and complexity of the aetiology of LVDD, our study shows that weightlifting long-term AAS users with significant heart muscle hypertrophy and increased blood pressure, all 97% had no definite LVDD, only modest reduced E/A and increased E/e′ ratios compared to the non-using WLC. Smoking and cocaine and amphetamine use may have contributed to the observed changes. In concordance with Čulić, we stress the need for further studies on this subject.

Acknowledgements

Institution where the work was performed: Department of Cardiology, Oslo University Hospital, Rikshospitalet, Norway.

Author notes

Conflict of interest: Vibeke M. Almas has received financial support from Novartis. Kristina H. Haugaa has received financial support from Bristol Meyer Squibb, Boehringer Ingelheim Norway, Agiana Pharmaceuticals, Novartis and Solid Bioscience.