Side effects and complications — All androgens have some side effects when taken in high doses; other side effects depend upon the structure of the androgen or the steroids to which it is converted (table 1). Some side effects occur only in women.

Cardiovascular — The effect of high doses of androgens on cardiac function is uncertain.

●Cardiac hypertrophy – Several case reports describe sudden death in young athletes who had no previously known heart disease but who were taking androgens; cardiac hypertrophy or myocarditis were found at autopsy [42,43]. It is not possible to establish causality in these sporadic cases (see "Athletes: Risk of sudden cardiac death"). There are also reports of left ventricular hypertrophy in body builders and power lifters, but most of these studies have not been randomized or controlled for degree of exercise, which itself can affect the degree of cardiac hypertrophy [44]. In one randomized, placebo-controlled trial, eight body builders treated with nandrolone decanoate showed no difference in several echocardiographic parameters at the end of eight weeks from those treated with placebo, but this study was limited by the small numbers of subjects and short duration [45].

●Lipids – Although physiologic doses of testosterone have no consistent effects on serum lipid concentrations, pharmacologic doses of androgens, especially 17-alpha-alkylated androgens administered orally, decrease serum high-density lipoprotein (HDL) cholesterol and increase low-density lipoprotein (LDL) cholesterol concentrations (table 1) [46].

In a study of normal men aged 30 to 56 years given androstenedione (300 mg/day for 28 days), serum HDL cholesterol concentrations decreased by 15 percent, a change that, in the general population, would predict an increase in risk of coronary heart disease [25].

●Hemostatic system – Androgen administration is also associated with activation of the hemostatic system, as illustrated in a study of 49 weight lifters in whom androgen use was ascertained by history and urine testing. The confirmed steroid users had a higher percentage of abnormally high thrombin-antithrombin complexes in plasma than nonusers (16 versus 6 percent, p = 0.01); higher plasma concentrations of prothrombin fragment 1 (44 versus 24 percent, p<0.001), antithrombin III (22 versus 6 percent, p = 0.005), and protein S (19 versus 0 percent); and lower plasma concentrations of tissue plasminogen activator and its inhibitor [47]. The importance of hemostatic system activation with regard to risk of thrombosis is unclear.

●Erythropoiesis – Testosterone stimulates erythropoiesis, and in men made hypogonadal by administration of a GnRH agonist, it increases hemoglobin and hematocrit in a dose-dependent manner [48]. Erythrocytosis is a well-recognized side effect of treatment of hypogonadism with physiologic doses of testosterone. Erythrocytosis, sometimes to a severe degree, has also been reported in association with administration of pharmacologic doses of androgens but only in case reports [49]. (See "Testosterone treatment of male hypogonadism", section on 'Erythrocytosis'.)

Neuropsychiatric — Many psychological abnormalities have been described, both in the medical literature and anecdotally, in men taking high doses of androgens (table 1). Most descriptions are uncontrolled, although in one study an attempt was made to compare men taking and not taking androgens [50]. One hundred sixty men recruited from gyms responded to a questionnaire about androgen use and psychiatric symptoms. Psychiatric symptoms, including major mood disorders and aggressive behavior, were more common in the men who had taken androgens than in those who had never taken androgens, and among the former, the symptoms were more common when they were taking androgens. Controlled studies using supraphysiologic doses of testosterone enanthate, although lower than those athletes often use, for up to six months, demonstrate no [39,51,52] or minimal [53] psychological abnormalities.

Several studies describe an association between nonmedical use of androgens and risky or even criminal behavior. In mail surveys of approximately 10,000 to 15,000 college students from 1993 to 2001, nonmedical use of androgens was associated with cigarette smoking, other illicit drug use, drinking and driving, and Diagnostic and Statistical Manual of Mental Disorders, 4th edition (DSM-IV) alcohol use disorder [8,54].

Examination of data from the Longitudinal Study of Adolescent Health, in which 6823 adolescents were surveyed on three occasions between 7th and 12th grade, demonstrated that nonmedical use of androgens was associated with violent behavior such as physical fighting [55].

Women who use anabolic steroids have described both hypomanic and depressive symptoms [56,57]. In addition, some women report rigid dietary practices (referred to as "eating disorder, bodybuilder type"), nontraditional gender roles, and dissatisfaction/preoccupation with their bodies (referred to as "muscle dysphoria").

Reproductive (women) — In women, side effects of androgens include acne, hirsutism, temporal hair recession in a male pattern, clitoromegaly, and deepening of the voice (table 1). Voice changes are irreversible [7]. Although not well studied, many women also develop oligomenorrhea or amenorrhea. Breast atrophy may also be seen.

Reproductive (men)

Hypogonadism following discontinuation of exogenous androgens — In men, all androgens suppress gonadotropin secretion and endogenous testicular function, both testosterone and sperm production (table 1). Testicular volume eventually decreases with chronic use of androgens, and spermatogenesis and fertility become greatly diminished. Chronic exogenous androgen use can cause prolonged suppression of gonadotropins and, therefore, result in hypogonadism after the exogenous androgens have been discontinued [17].

The prevalence of hypogonadism among exogenous androgen users is not known. However, one retrospective series found that 21 percent of 382 men with hypogonadism presenting for testosterone therapy had previously taken these drugs [58]. This observation highlights the importance of obtaining a careful drug history before prescribing testosterone to men with apparent hypogonadism. (See "Testosterone treatment of male hypogonadism", section on 'Appropriate candidates'.)

Among men who stop using androgens, the sperm count usually returns to normal within four months after discontinuation [59] but may take more than a year [60]. Gonadotropin and testosterone secretion remain suppressed for a few months after androgens are discontinued. Younger men may recover more quickly than older men [17].

Gynecomastia — Gynecomastia occurs because testosterone is converted to estradiol via the action of the aromatase enzyme complex, so that high doses of testosterone result in high serum estradiol concentrations. Androgens that have been 5-alpha reduced, such as dihydrotestosterone, and synthetic androgens in which the A ring has been modified cannot be aromatized and therefore cannot be converted to estrogens and do not cause gynecomastia.

As noted above, many male AS users also take antiestrogens or aromatase inhibitors to prevent gynecomastia. (See 'Estrogen blockade' above.)

Other — AS have been associated with a number of other adverse effects as well, including (table 1):

●Infection – Sporadic case reports describe infections due to injection of androgens, including local abscess at the site of injection, septic arthritis, hepatitis B and C, and human immunodeficiency virus (HIV) infection from sharing of needles [61].

●Prostate – Because the prostate is a testosterone-dependent gland, there is concern that the high doses of androgens that athletes take might increase the risk of benign prostatic hypertrophy and prostate cancer. A meta-analysis of replacement doses of testosterone in hypogonadal men did not show such increased risks [62], but the risk in athletes who take large doses of androgens has not been reported.

●Tendon rupture – The risk of tendon rupture (eg, triceps or biceps tendon rupture) appears to be increased in weight lifters who use androgens [63-65].

●Epiphyseal closure – Pharmacologic doses of testosterone or other androgens that can be aromatized to estrogens hasten epiphyseal closure in adolescents whose epiphyses have not yet closed. The frequency of androgen use in adolescents is discussed above. (See 'Epidemiology' above.)

●Hepatic side effects – Hepatic side effects occur only with oral 17-alpha-alkylated androgens and include high serum concentrations of liver enzymes, cholestatic jaundice, and peliosis hepatis, characterized by blood-filled hepatic cysts. Hepatomas have also been reported, but the number of cases is few and causality is uncertain. (See "Testosterone treatment of male hypogonadism", section on 'Oral preparations'.)