In this cross-sectional study we found two distinct patterns of substance use, the first distinguished by the use of both stimulants and alcohol, and the second by the use of THC among A-ADHD patients. Consistently with the literature on ADHD, the substances mostly used by our patients were THC, CNS stimulants and alcohol [20], with less frequent use of BDZs and opioids. To the best of our knowledge this is the first study on A-ADHD patients which has analyzed, by applying factorial analysis, substance use typologies including both past and current use. The concomitant use of stimulants and alcohol is popular and could be due to the specific properties of the substances involved, in so far as psychostimulants are known to produce a rapid addictive feeling of a “high” which can be dampened by the sedative properties of alcohol. Patients belonging to this pattern were, in fact, significantly younger when they attracted our attention by comparison with THC users—a difference that could reasonably be attributed to a higher severity of substance addiction, together with its poly-use consequences. Similarly, legal issues were significantly better represented in the S-CNS/ALC group. Multiple studies [21,22,23,24] showed a high prevalence of ADHD diagnosis in prisoners distinguished by being particularly young when they committed their first crime [25] and showing high rates of recidivism [26]. Previous reports showed that SUDs in ADHD patients, with or without antisocial disorders, seem to increase the risk of violent crimes, especially drug-related ones [27]. The latter could be seen as consequences of the marked emotional dysregulation and impulsivity; affecting subjects with ADHD [28]. Furthermore, substance intoxication and, above all, abstinence could increase the deficits in emotional regulation and therefore increase the risk of aggressive behaviors [29]. Even if substance use in ADHD patients might be seen as a way to reduce manifestations of neurocognitive distress, the combined use and misuse of non-medical prescribed stimulants and alcohol could have the opposite effect, leading to behavioral disturbances, aggressiveness and legal consequences.

In our sample we found that while all patients showed an intermediate sleep chronotype, patients who used mainly THC go to bed earlier. This was expected, as the use of cocaine and/or stimulants, often with subsequent withdrawal, are usually associated with more sleep problems and increased sleep latency [30]. More importantly, the ones who mainly used THC displayed significantly higher total scores, which suggests a probable major role of THC in earlier sleep induction. Current research on cannabinoids use and sleep has given mixed results, with several works supporting the role of the endocannabinoid system in the regulation of circadian rhythm sleep. ADHD patients with poor sleep quality may use cannabinoids for their short-term benefits on sleep, so activating a therapeutic function. However, long-term cannabinoid use may cause addiction to the sleep effects with increased risk of dependence and more sleep disturbances if the drug is stopped. Overall, the global effect seems to be related to the type of cannabinoids, their dosage and route of administration [31].

From a clinical standpoint the most interesting finding was the similar presentation of ADHD symptoms and their severity, independently of substance typologies, at treatment entry. The two patterns were also similar in terms of psychiatric comorbidities, impulsivity emotional dysregulation, general psychopathology, hypomanic symptoms and global functioning. Overall, in the medical literature many authors have studied the influence of ADHD symptomatology on the course of drug addiction in substance abuser samples by mainly focusing on a single substance of abuse, rather than comparisons between different patterns of use. SUDs are often comorbid with a variety of psychiatric disorders, in particular the bipolar ones, in which addiction is associated with more manic switches, a worse psychopathology, less treatment adherence and greater functional impairment [32,33]. A worse course of SUD in ADHD patients compared with other substance users is well established, with an earlier onset of SUD, a more severe symptomatology, higher resistance to therapies and worse drop-out rates [34,35].

Looking at our results we might hypothesize that different SUD typologies do not influence ADHD symptomatology because of the specific effect of substances on ADHD subjects. Even if the results reported in the literature are still controversial as regards the efficacy of stimulants on people with both SUD and ADHD, a recent study from our research group showed a marked reduction of cocaine use concomitant with ADHD improvement in a small sample of A-ADHD patients treated with either atomoxetine or methylphenidate [36]—a finding that suggests the possible benefit of these medications not only in A-ADHD but in SUD components, too. When we talk about stimulants in this specific population, we should take care to correctly differentiate between the two constructs of “relief” and “reward” craving. According to the self-medication theory, a certain number of ADHD patients may choose stimulants driven by a “relief” craving as an attempt to reduce symptoms of inattentiveness and restlessness. On the other hand, other patients who use psychostimulants mostly report a “reward” craving associated with a recreational, pleasurable use. Thus, in ADHD patients, we believe that different substances, in particular stimulants, could be used according to the substance-specific dopaminergic action with autotherapeutic purposes (bringing “self-medication”, to cite Khantzian). In clinical practice we therefore suggest evaluating the type of craving that underlies the use of substances to allow an early diagnosis and a targeted treatment for dual disorder (A-ADHD/SUD) patients. The main result of this study is the evident absence of any symptomatology beyond the two substance use typologies found at treatment entry. One possible explanation for this lack of differences is the following. When entering treatment, the balance of the pre-treatment phase is likely to be lost both for those who “self-medicate” with stimulants, and those who use cannabis for its rewarding effects. The psychopathological decompensation experienced by the patient becomes generalized, and the potential differences can no longer be highlighted. So, it cannot be ruled out that in the pre-treatment phase, these differences do exist. This opens up the possibility that at treatment entry only the differences in MEQ will be noticed, as a result of the chronic use of cocaine and cannabinoids, which give rise to different effects.

We should, in any case, bear in mind that the prolonged use of fast-acting stimulants and polysubstance use in association with the psychopathology specific to ADHD will only lead to more intense dysphoria, more severe mood instability and more deeply negative outcomes.

Polysubstance use, reward and relief craving often coexist in A-ADHD patients but a good knowledge of the dominant typologies of substance use can help clinicians, independently of the quality and severity of symptoms, to better elaborate a treatment strategy and expect different treatment outcomes. If type 1 is dominant, we can anticipate a response to the A-ADHD medications (stimulant drugs), and we should use anti-craving medications for cocaine only if the cessation of cocaine use is still incomplete. Conversely, if type 2 is dominant, we should use anti-craving medications for THC, and, unfortunately, given the scarcity of scientific evidence in this kind of treatment, we shouldn’t expect any important positive pharmacological outcome, especially in dual disorder patients. Our clinical experience usually limits us to combating the psychiatric complications of the THC use. Regarding dominant type 2 A-ADHD patients, there is also the risk of a frequency increase as a result of the new leniency and legal innovations associated with cannabis procurement and use. Nevertheless, cannabis had not been legalized, in Italy, for the time being.

Limitations: Several limitations should be noted in the present study. First, our sample size was small, a datum that limits the generalizability of our results and the power of the study. Indeed, the patients recruited came to our attention as ADHD treatment naïve patients, neither diagnosed nor treated before adulthood. Many patients with ADHD are often misdiagnosed and reached our attention only when the clinical picture was severe and had been complicated by other comorbidities. Consequently, to confirm our hypothesis, more adult ADHD patients with or without a dual disorder need to be enrolled.

Secondly, some of the features assessed, such as impulsiveness, emotional dysregulation and overall functioning, were determined by the answers given to self-report questionnaires, so raising the danger of reporting biases. However, all participants completed the reports under the supervision of a psychiatrist who evaluated them clinically and conducted other valid interviews.

Thirdly, our findings showed a preliminary baseline evaluation of A-ADHD patients at treatment entry, without including a control group. Indeed, this cross-sectional picture is a part of a longitudinal study with multiple follow-up steps that facilitated the evaluation of the trajectories of Substance Use Disorders in A-ADHD patients and helped to make clear how the two disorders are influenced by different psychopharmacological treatments. Moreover, besides SUD typology comparisons, we are also analyzing both similarities and differences in A-ADHD patients with and without a dual disorder at treatment entry.

In conclusion, another significant limitation is the use of a non-random sample. Our hospital is located in Central Italy, which exerts a specific nationwide attraction on patients who have a dual disorder with predominantly psychopathological aspects. Besides, Italian legislation lists centers of excellence for the treatment of adult ADHD in various areas of Italy, and patients typically apply for treatment to the nearest one. So, a bias in favor of the closest officially recommended area cannot be excluded. The risk of a self-selection bias is related to the type of treatment we offer (stimulant medications). So, only patients interested in this kind of medicine may have requested treatment. Finally, our sample did not include patients who choose an “addiction unit” for their treatment; this applied, specially, to those currently using opioids, who, in fact, were only minimally represented in our sample. In Italy, psychiatric units and addiction units work independently, the result being that the level of communication between them is poor. These sample selection biases limit the generalization of our data.