An estimated 20 million individuals worldwide used cocaine according to the world drug report 2021. This global level of cocaine use continues to increase every year (1), and its impact in terms of public health is major, particularly with regard to overdoses (2). The treatment of this addiction is complex, especially when it is associated with a psychiatric disorder, with a high relapse rate and a worse follow-up addiction severity (3, 4). Cocaine, as well as other psychostimulants, can also cause stroke and alterations in mood and cognition (5).

Relapsing patients are more likely to declare serious life-time psychiatric symptoms, including depressive symptoms (6). These patients rated their psychiatric problems as more severe and reported a greater need for treatment for these problems (6). Some studies suggested that depressive symptoms are specifically and significantly associated with an increased risk of relapse after treatment in substance users (7). These mood fluctuations are important because they are a pejorative prognostic factor in cocaine-dependent patients (8) and are associated with increased suicidal risk (9). Furthermore, patients who drop out early have more depressive symptoms than the later dropouts (10). It has been shown that worse depressive symptoms represent a significant predictor of worse medical severity at 12-months follow-up (4). Depressive symptoms seem to play a key role in the process of relapse and so have been chosen in our study as a discriminative factor.

A particular form of transient depressive symptoms is observed during cocaine withdrawal. It is often the subsyndromic form and does not correspond to the well-known timeline of the major depressive disorder. There are several other specificities: (1) these depressive symptoms disappear when taking cocaine, (2) treatment of depression in cocaine users with selective serotonin reuptake inhibitor (SSRI) seems to underperform (11, 12). The mechanism therefore seems different and raises the question of the involvement of dopamine in the emergence of depressive symptoms during cocaine withdrawal. This is especially relevant considering that dopamine plays a central role in the mechanism of cocaine, which works by blocking the dopamine transporter (DAT) and increasing the brain's dopamine level (13). This involvement of dopamine had already been suggested by Dackis and Gold (14).

A potential model of dopaminergic depression can be observed in Parkinson's disease, where symptoms are correlated with dopamine levels (15). In this disease, the hypodopaminergic behaviors results in depression, anxiety, apathy, anhedonia, cognitive dysfunctions, and sleep disorders (15, 16). In the particular case of Parkinson's disease, the hyperdopaminergic behaviors, which is induced by dopaminergic treatments in Parkinson's disease, is characterized by hallucinations, delusions, and compulsive behaviors, such as pathological gambling, hypersexuality, shopping, binge eating, and punding (17, 18). These “dopaminergic behaviors” could be used as indirect clinical markers of dopamine activity.

The psychoactive effects of cocaine are fairly well described today (19). Among these symptoms are those that may serve as indirect markers of dopaminergic activity. Cocaine use induces a brief “peak” of pleasure, lasting a few minutes, associated with subjective stimulating effects. Sometimes, there are also psychotic symptoms, especially during high consumption: hallucinations, delusions, or consumption associated behavior (20). This phase then quickly gives way to a withdrawal characterized by contrary symptoms, such as depressive symptoms, anhedonia, and anxiety (21). The intensity of depressive and psychotic symptoms seems to be related to the severity of addiction and the level of use (22).

Based on the Parkinson's disease model and what is clinically observed in cocaine users, we hypothesized that variation in dopamine activity is implicated in the development of depressive symptoms during cocaine withdrawal. Consumers may first experience hyperdopaminergic symptoms upon substance use (psychotic symptoms, or stereotypes), then hypodopaminergic symptoms upon substance withdrawal (depressive symptoms, apathy, and anxiety).

Our objective is to investigate whether patients reporting depressive symptoms during cocaine withdrawal (RDS+) have a different profile of clinical markers of dopaminergic behaviors from those who do not (RDS–).

This is one of the first studies to have evaluated mechanisms underlying the onset of depressive symptoms during cocaine withdrawal. The characteristics of our sample corresponded to those of European cocaine users (22, 27). Patients were predominantly single men, with an average age of onset of use of 22 years. They mainly used cocaine by inhalation, then by smoking and finally by injection. Most use was daily, and levels of polydrug use were high.

Our data showed that the RDS+ (Reported Depressive Symptoms +) and RDS– (Not Reported Depressive Symptoms) groups were homogeneous for age, gender, marital status, education, and history of hospitalization for withdrawal. The age of onset, frequency, and type of product most commonly used were also comparable in both groups. Levels of use of other psychoactive substances were identical in both groups, as was the level of polydrug use. These variables did not seem to have an influence on depressive symptoms, as reported by Uslaner et al. (28).

Concerning the modality of use, snorting was found more frequently in the RDS+ group, with depressive symptoms, (63.90 vs. 48.61%). Injectable or smoked pathways were more frequent in the RDS– group (26.39 vs. 23.65% for smoked; 18.06 vs. 7.86 % for injectable). This could be explained by the different pharmacokinetics of cocaine, depending on the form used. Clinical effects appear 3 mins after a cocaine snort and can last up to an hour. For the smoked route, they last only 10–30 mins, but are perceived in 5–10 s. The intravenous route acts in 16–20 s and has an effect lasting for 10–30 mins. In our depressed patients, snorting may therefore be a way to optimize intake so as to manage depressive symptoms more sustainably. This would be consistent, as the euphoric effects of cocaine are also perceived more intensely in depressed subjects compared to non-depressed subjects (29).

Depressive symptoms are more common among cocaine users than in the general population, with lifetime prevalence ranging from 25 to 61%, depending on the study (30, 31). Our patients have a higher rate of depressive symptoms (77%). Our recruitment sites could explain this phenomenon, as they were centers specialized in the management of opioid users. This corresponds to the data in the literature, which shows that depressive symptoms are found more particularly in patients who have entered the care process, probably because they are more symptomatic, and are associated with a more severe use disorder (31, 32). Our patients were therefore probably at a more severe stage than in other studies with higher recruitment. They had an average of 5.53 DSM IV dependence criteria. Bipolar disorders are also common in this population (33), and some authors confirm a predominance of depression among drug users (34).

The psychotic symptoms associated with cocaine use that we observed were mainly hallucinations, delusions, or stereotypies. They corresponded to those reported in the literature. These psychotic symptoms are found during consumption in 54–80% of patients (35, 36), and during dopaminergic treatments in Parkinson's disease (17, 37–39).

The choice of variables included in our analyses was crucial. We wanted to model the concepts of hyper and hypodopaminergy as well as possible by using indirect clinical markers of dopaminergic behaviors. Severity of dependence (p < 0.001), anhedonia (p < 0.01), and withdrawal symptoms (p < 0.01) were relevant for modeling clinical hypodopaminergic reactions in cocaine users. Psychotic symptoms and particularly hallucinations, delusions, and physical symptoms before use were criteria integrated into our modeling of clinical hyperdopaminergic disease. All of these criteria are consistent with those mentioned in several articles that have tried to clinically characterize these concepts in Parkinson's disease (PD), and define symptoms associated with hyperdopaminergia and hypodopaminergia. During the hyperdopaminergic phases in PD, certain researchers have observed delusions, hallucinations and motor stereotypes (17, 37, 38). More recently, hyperdopaminergia has also been linked to the development of behavioral addictions and still appears to be underestimated. They include compulsive purchases, pathological gambling and sexual behavior disorders (15). Symptoms associated with hypodopaminergia in PD are similar to depressive symptoms (15, 40), anxiety (41) and apathy (42).

Several studies have also investigated the release kinetics of dopamine in the nucleus accumbens during drug intake. In human subjects, a relationship between the subjective effects of cocaine and DA transporter (DAT) occupancy in the striatal areas has been demonstrated (13). Cerebral level curves for cocaine in the striatum are related to the cocaine-induced “high” behavior, with a peak at 10 mins then a progressive decrease of cerebral cocaine level and cocaine effects (13). These observations have also been found in animal models (43, 44). They show that the dopamine peak induced by stimulant intake in mice is almost immediate, followed by a progressive decrease over several tens of minutes or even hours.

Our patients were well discriminated in the analysis of the hyperdopaminergic component, and a large overall effect size of the hyperdopaminergic variables was observed. The most important criteria capable of explaining this difference between the groups were the existence of psychotic symptoms, particularly hallucinations and delirium, which is consistent with the scientific data (35, 36). Associated movements and physical symptoms before use seem to be involved but to a lesser extent. These Cocaine-Related Behavioral (CRB) symptoms are very frequently found in cocaine users, in particular for repetitive/stereotyped behaviors (45). The type of cocaine used does not influence these stereotypes (46), but on the other hand, these symptoms are found more frequently among cocaine injecting users, who were more numerous in the RDS– group. It therefore seems difficult to highlight a difference between our two groups on the basis of the sole criterion of depressive symptom. It would be interesting to study this phenomenon specifically in these injecting vs. non-injecting patients.

Patients were also well separated in the analysis of the hypodopaminergic component and there was a large overall effect size for these variables. Anhedonia, severity of dependence and symptoms appear to be good markers of this dimension. Our polyconsumption criterion, on the other hand, had only a small effect size. This could be explained by the low precision of the data for this variable, which is based on lifetime use of substances and not current or recent use. It might have been interesting to study comorbid opioid use disorder or Opioid Maintenance Treatment (OMT). A previous study using the same sample highlighted the role of OMT on the subjective effects produced by cocaine (23). Patients receiving an OMT at the time of their first cocaine use reported significantly less tachypsychia during this first cocaine intake, suggesting a specific protective effect of OMT on cocaine-induced “high” hyperdomapaminergic effects. Thus, a possible preventive effect on cocaine withdrawal would merit investigation.

These results show the interest of our two dimensions for exploring the phenomenon of depressive symptoms in cocaine users. When we integrated all these variables to perform a global analysis, discriminative capacity was even better, with a large overall effect size. This could indicate that rather than hypo or hyperdopaminergic changes occurring separately, these thymic variations might be explained by the switch from one to the other.

One of the main limitations of this study is that it was based on a cross-sectional study, and therefore causality cannot be established. It would have been interesting to perform a longitudinal evaluation of our two components. Even so, as we are interested in a short-term phenomenon, namely cocaine “comedown,” this cross-sectional analysis appears justified.

We chose to use a subjective assessment of depressive symptoms in cocaine users. This measure is more practical to use in this population, but it would have been appropriate to supplement this self-assessment with a more precise and detailed scale (24). Furthermore, using clinician-rated scales rather than self-reports could be useful if the tool is sufficiently time-sensitive.

Our sample may not perfectly represent the general population of cocaine users, but rather the most severe patients due to selection bias. Indeed, all of our patients were recruited from specialized addiction centers. The small number of subjects in our RDS- group (N = 72) is also a limitation.

In our overall model, we highlighted a homogeneous subgroup among RDS+ patients (see Figure 3A), but with a few scattered subjects. Other variables not available in our protocol (apathy and anhedonia) might play a role in this model to better condense the group, or to reveal subgroups that were not discriminated by our analyses.

In conclusion, our study was able to better characterize patients presenting depressive symptoms during cocaine withdrawal by comparing them with those who do not, according to their dopaminergic activity profile assessed by indirect clinical markers.

Early identification in the management process of these patients at risk of pejorative evolution or relapse could help to adapt the therapeutic strategy envisaged. The evaluation of depressive disorders in substance abuse patients might be helpful in designing and implementing specialized interventions to reduce the likelihood of relapse (7).

Further studies would be needed to study this phenomenon in more detail, with longitudinal evaluation and more accurate monitoring of symptoms. It would also be interesting to specifically study the very homogeneous group of RDS+ patients highlighted (Figure 3A). We could, for example, investigate physiological specificity by analyzing the genetic polymorphisms of the dopaminergic pathway.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

The studies involving human participants were reviewed and approved by Research Ethics Committee of Ile de France (Paris area). The patients/participants provided their written informed consent to participate in this study.

JC: conceptualization, methodology, data curation, writing—original draft preparation, writing—reviewing and editing, and visualization. GB: conceptualization, investigation, supervision, and writing—review and editing. BP: methodology, data curation, and formal analysis. NC and PL: conceptualization and writing—review and editing. EK, E-HZ, RI, and VB: conceptualization. FV: conceptualization, investigation, and writing—review and editing. ID: conceptualization, supervision, methodology, data curation, writing—review and editing, and formal analysis. All authors contributed to the article and approved the submitted version.

The research project PSYCHOCOKE received a grant from the French Ministry of Health (Programme Hospitalier de Recherche Clinique National 2010, AOM10165), and was promoted by the DRCD (Direction de la Recherche Clinique et du Developpement) of the Assistance Publique–Hopitaux de Paris.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.