As we observed in our sample, patients take anxiolytic-hypnotics, which belong to the class of central nervous system (CNS) depressants, with greater frequency during a depressive episode. They take CNS stimulants (cocaine-amphetamines) at a greater frequency during a hypomanic episode, whereas they tend to take both CNS stimulants and cannabinoids with a greater frequency during a manic episode; lastly, during a mixed episode they take CNS depressants (alcohol), stimulants, and hallucinogens together.
In the case of depressed patients, the use of CNS depressants is consistent with their toxicological status. It should be noted that benzodiazepine use in heroin addicts could be correlated with a condition of opiate dependence improperly compensated by street heroin . From a psychopathological standpoint, depressants may aggravate the slowing of cognitive and physical functions caused by depression, but it remains true that these medications are effective in treating insomnia and anxiety, which are often symptoms of depression. Also, patients may not be seeking an actual ‘lift’ of their depression but be searching for a state of ‘oblivion’ in which the pain of depression is cancelled. In depression, what is seen is not a higher use of stimulant substances, but the use of CNS Depressants that may sometimes relieve some aspects of depression (or may not do so) – a situation that fails to provide support to Khantzian’s hypothesis.
More clearly, Khantzian’s hypothesis does not seem to be supported by the other three kinds of clinical presentations. Patients during a hypomanic, manic or mixed episode, despite experiencing a state of excitement, tend to continue their abuse of psychostimulants, further reinforcing and elevating their mental state. This is consistent with a proposed bipolar-stimulant spectrum where subthreshold bipolar traits are aggravated by stimulant abuse .
If we focus on heroin-dependent subjects, the concomitant use of cocaine is reported to be a relevant phenomenon that will determine negative consequences on social adjustment and outcome. When heroin and heroin-cocaine abusers have been compared, a direct relationship has been found between cocaine abuse and the rate of psychiatric disorders, together with correlation with the severity of self-rated psychopathology . We do not know if this lack of awareness of psychopathological symptoms is due to the use of cocaine or to the underlying excitement that sustains cocaine use. Moreover, if cocaine use represents self-enhancement of one's level of hypomania, cyclothymia or hyperthymia, the craving for hypomania is likely to be particularly strong in heroin addicts, whose level of excitement is lowered by heroin use . In addition, cocaine has been reported to induce a higher frequency of mixed states when abused by bipolar patients . This evidence suggests that some bipolar patients, after deciding to use cocaine instead of being excited, may have shifted from a manic or hypomanic to a mixed episode.
During a manic episode, patients show a high level of consumption of stimulants and cannabinoids. In the literature, the abuse of cannabinoids in bipolar patients has been found to induce manic symptoms , so it is possible that, in our manic patients, as with cocaine use, they may use cannabinoids to optimize their level of excitement. To date, cannabis use is also considered to be one of the most important risk factors for schizophrenia, thanks to its ability to precipitate or exacerbate psychotic symptoms [41–46]. In line with this assumption, in our sample, cannabis is mainly abused in manic and mixed states that, unlike depressive and hypomanic episodes, are often characterized by the presence of psychotic symptoms. Whatever the causes of the use of cannabis, Khantzian’s hypothesis is not supported in its application to cannabis use. For many subjects, ending cannabis use is difficult to achieve, not only because of prior habits of use, but also because of the attendant psychotic symptoms, including poor insight and judgment, lack of impulse control and cognitive impairment. Most of these subjects are unable to understand that cannabis use is connected with the onset of symptoms .
One widely debated issue is whether Khantzian’s hypothesis is a suitable instrument for interpreting alcohol dependence. In examining patients with a mixed episode, we found that, besides their abuse of cocaine-amphetamines and cannabinoids, and in contrast with the other three clinical presentation groups, they often resort to alcohol use. Patients experience their mixed mood as something undesirable and unpleasant, but they still continue to consume substances that tend to preserve their mixed, dysphoric state. Craving for substances and dependence create a loop, a senseless vicious circle in which patients obtain neither satisfaction (mood elation) nor physical benefit (relief). In line with this observation, a past or current alcohol use disorder has proved to raise the likelihood of a switch from depressive to manic, mixed or hypomanic states in patients with bipolar disorder . Nervousness in alcoholic patients has been hypothesized to be the only negative mood state to predict increases in alcohol consumption later in the course of the day. Further examination of this within-person relationship has demonstrated that men were more likely to consume alcohol when nervous than were women, but this association is unrelated to family history of alcoholism, problem drinking patterns, or traits of anxiety and depression. Consistently with the self-medication hypothesis, alcohol consumption has been associated with lower levels of nervousness, but this effect varies in a way dependent on several demographic and clinical variables . Almost one quarter of individuals with mood disorders use alcohol or drugs to relieve symptoms, with the highest prevalence of self-medication in bipolar I disorder. After checking the effects of substance use disorders, self-medication has been associated with higher rates of comorbid anxiety and personality disorders than those found in individuals who do not self-medicate . On the basis of these data we believe that, in the case of alcohol, Khantzian’s hypothesis accounts for anxiety disorders more satisfactorily than mood disorders, although it must be added that it actually explains controlled rather than addictive use. In fact, enduring use, despite the worsening, or the inadequate balance, of symptoms, is inconsistent with a current self-medicating explanation, although that explanation may have been appropriate in a previous stage of controlled use. It should also be remembered that patients were assessed for mood during current drug use, thus ruling out the ambiguity between spontaneous mood swings and substance-induced intoxication. In fact, our patients had been displaying affective dysregulation for some time before being diagnosed, and had been engaged in substance use, which was bound to worsen the affective core of their clinical pictures (i.e. stimulants during excitement and depressants during depression). The hypothesis of symptomatological overlap between temporary substance-related intoxication and mood states is superseded in this way.
After reviewing our data we speculate that, setting aside depressive and mixed episodes, the abuse of substances in hypomanic and manic episodes of bipolar disorder is more probably due to patients’ desire to maintain their current affective state rather than to resolve depressed mood.
These considerations also provide a possible explanation for the fact that bipolar patients tend not to comply with therapy during hypomanic, manic and mixed states [51–53]. In the literature, patients’ lack of compliance with prescribed therapy has been principally associated with their lack of insight into their mental illness [54–56]. We go beyond that in suggesting that, during hypomanic and manic phases, bipolar patients do not comply with prescribed therapies and tend to exacerbate their mental status by means of substances, not only because they are unaware of their mental illness, but also because they somehow live the current episode as a pleasurable and rewarding experience (in cases of addiction to mania).
The obvious limitations of this study are due to the fact that this is a retrospective analysis carried out on a small cohort of patients, rather than a study specifically designed to elucidate this issue. Assessments of the same subject in various different clinical presentations of the natural history of this illness would have provided a better level of information. In addition we must consider the difficulty in determining whether the substance use modifies the mood or the mood state determines the substance used. It is possible that stimulants are seen in those with mania or hypomania because the stimulant produced the mood state. They could have been depressed without it. Lastly, we have no information about the temperament of our subjects. So we cannot exclude the presence of a depressive temperament that is able to moderate the nature of patients’ substance abuse; that would set up the need to modify our hypothesis. One of our earlier findings, however, was that heroin users mainly have a cyclothymic temperament .