Illicit anabolic-androgenic steroid (AAS) use represents a growing worldwide public health problem (1, 2). Some AAS users consume only a few courses of these drugs in a lifetime, but others progress from discrete courses of use to a maladaptive pattern of almost continuous use, despite adverse medical, psychological, and social effects (3, 4). In the last 20 years, accumulating animal and human studies have documented and characterized this syndrome of AAS dependence. For example, rats and mice will select AAS in conditioned place preference models (5), and hamsters will self-administer testosterone even to the point of death (6). Unlike rodents, humans may initially develop a pattern of AAS dependence as a result of “muscle dysmorphia” – a form of body dysmorphic disorder where they become preoccupied that they do not look adequately muscular (7). In later stages, however, AAS dependence comes to resemble “classical” drug dependence, with a well-defined withdrawal syndrome mediated both by neuroendocrine factors and by a variety of cortical neurotransmitter systems, especially the opioidergic system (5, 8). AAS dependence may be associated with substantial medical and psychiatric morbidity, including hypertension, dyslipidemia, cardiomyopathy, persistent hypogonadism, major mood disorders, and progression to other forms of substance abuse and dependence, especially opioid dependence (2). The full magnitude of these risks is still unknown, because widespread AAS abuse did not spread from the athletic world to the general population until the 1980s (2), and only now are many AAS users becoming old enough to have established a dependence pattern and to have entered the age of risk for some of these adverse outcomes. Although AAS users historically have been reluctant to seek treatment (1, 9), these adverse outcomes may bring increasing numbers to clinical attention. Importantly, unlike classical drugs of abuse, AAS are not ingested to achieve an immediate “high” of acute intoxication, but instead are consumed over a preplanned course of many weeks to achieve a delayed reward of increased muscularity. Therefore, the existing DSM-IV criteria for substance dependence, which were designed primarily for acutely intoxicating drugs, do not apply precisely to AAS. For example, criteria such as “using the substance in larger amounts than was intended, ” or “giving up or reducing important activities because of substance use, ” apply more easily to alcohol or cocaine than to AAS. But these considerations should not obscure the fact that AAS have definite psychoactive effects, including a potential for addiction, that is likely underestimated because attention has focused on the drugs’ muscle-building properties (1). On the basis of the available literature (2–4, 10) and clinical experience with AAS-dependent individuals, we would suggest that the existing DSM criteria could be adapted for diagnosing AAS dependence with only small interpretive changes (Table 1). AAS are presently the only major class of drugs scheduled by the Drug Enforcement Administration for which DSM-IV does not explicitly recognize a dependence syndrome (11); this omission could be rectified in DSM-V by offering these proposed interpretations for AAS dependence in a small table or in the accompanying text of the substance dependence section. Alternatively, DSM-V could initially propose these criteria only for research purposes, pending further evidence of their reliability and validity. In either case, clarified criteria for AAS dependence will likely improve recognition of this diagnosis among clinicians and researchers encountering the syndrome, and stimulate increased attention to this emerging public health problem. TABLE 1 DSM Substance Dependence Criteria (Shown in Bold), Interpreted for Diagnosing AAS Dependence (Shown in Plain Text)