Secondary Logo

Journal Logo

Original Articles

Aggressive Behavior and Substance Use Disorder: The Heroin Use Disorder as a Case Study

Maremmani, Icro MD*,†,‡; Avella, Maria T. MD§; Novi, Martina MD§; Bacciardi, Silvia MD; Maremmani, Angelo G.I. MD

Author Information
Addictive Disorders & Their Treatment: September 2020 - Volume 19 - Issue 3 - p 161-173
doi: 10.1097/ADT.0000000000000199
  • Free


The study of aggression is crucial in psychiatry and psychopathology because hostile behaviors are involved in various physical pathologies and because aggression plays an essential role in the genesis of mental pathology. In addition to the importance of pharmacology, the role of psychosocial factors in favoring the onset and influencing the course of diseases, by expressing or controlling aggressive bursts, has been enhanced.

Aggression refers to a wide range of human behaviors that can perform different functions in enabling an individual to adapt to his or her circumstances. The term aggression can describe both adaptation to the environment in a practical, creative, and available manner, and harmful and destructive, socially deplorable behavior. Apart from the difficulty of defining the term itself, there are various different theoretical approaches to the study of the origins and causes of aggressive behavior. From time to time, starting with Freud conception and arriving at the ethological view, aggression was considered to be an aspect of libido, a desire for control over external reality, a tool for achieving gratification or overcoming frustration. In contrast, aggression was often perceived as an autonomous death instinct, already detectable in the last writings of Freud,1 and taken up in a specific way by Klein.2 However, some psychoanalysts, such as Adler3 and later Fromm colleagues,4,5 highlighted the constructive aspect of aggression that appears to express the thirst for the domination of the individual. Ethologists, such as Lorenz,6 underscored its importance for the survival of the individual and the whole species. In the behaviorist view, the contributions of Dollard et al7 of the Yale School were the most important. According to this perspective, attention is focused on defining the contingent modalities rather than the original causes of aggression, which are understood to be a consequence of a state of frustration. Berkowitz8 introduced the concept of disposition favorable to aggressive acts represented by the emotional reaction produced by frustration and by aggressive clothes (indicative of readiness) acquired previously. Bandura9 argued that aggression is a class of responses that the individual learns by more or less direct imitation of models. Buss10 emphasized, above all, the instrumentality of the aggression oriented to overcoming pathogenic noxae or the acquisition of benefits, and investigated the typology and its acceptability from the social point of view.

In this perspective review, we analyze the aggressive behaviors of our patients suffering from heroin use disorder (HUD) using the concepts of the behavioral school and the diagnostic criteria of the various Diagnostic and Statistical Manual of Mental Disorders (DSM).11 On the basis of the many years of experience of the Vincent P. Dole Research Group at Santa Chiara University Hospital in Pisa, we first analyzed the aggressive behaviors displayed by HUD patients at treatment entry, and then compared them with those of individuals of the same social extraction who, by contrast, were not substance users. We then addressed the problem of aggression in patients with dual disorder (DD) and suicidal risk in HUD patients. Finally, we studied the role of the violence/suicidality (V/S) dimension that we found in HUD patients,12 and the effects of opioid medications (methadone and buprenorphine) on aggression.


As investigating tools, our research group used the Buss-Durkee Questionnaire, in the Italian version curated by us, the Hostility factor of Symptom Checklist-90 (SCL-90), which was obtained in psychiatric patients from Derogatis et al,13 and the V/S factor highlighted by our research group in patients suffering from substance use disorder (SUD).

Buss-Durkee Hostility Inventory

Buss-Durkee Hostility Inventory (BDHI) defines the subclasses of hostility that are typically delineated in everyday clinical situations.10

  • Assault: physical violence against others. This includes getting into fights with others, but not the destruction of objects.
  • Indirect hostility: both roundabout and undirected aggression. Roundabout behavior like malicious gossip or practical jokes is indirect, in the sense that the hated person is not attacked directly, but by devious means. Undirected aggression, such as temper tantrums and slamming doors, consists of a discharge of negative affect against no one in particular; it is a diffuse rage reaction without a target.
  • Irritability: a readiness to explode with negative affect at the slightest provocation. It includes quick temper, grouchiness, exasperation, and rudeness.
  • Negativism: oppositional behavior, usually directed against authority. It involves a refusal to cooperate that may vary from passive noncompliance to open rebellion against rules or conventions.
  • Resentment: jealousy and hatred of others. It refers to a feeling of anger at the world over real or fantasized mistreatment.
  • Suspicion: projection of hostility onto others. It varies from merely being distrustful and wary of people to beliefs that others are being derogatory or are planning harm.
  • Verbal hostility: negative affect expressed in both the style and the content of speech. This style includes arguing, shouting, and screaming; its content includes threats, curses, and being overcritical.

The variable of guilt was added because the relationship of guilt to the total score may be of clinical interest. Accordingly, items were compiled for a Guilt Scale, with guilt being defined as feelings of being inadequate, having done wrong, or suffering pangs of conscience.

The items of BDHI have been translated into Italian (QTA version) and then translated into English by a translator who had no knowledge of the English version. The accuracy of the translation and its conformity with the original version have been checked by a bilingual English expert native speaker, who confirmed the original meaning of the items of the QTA.14 QTA was standardized in the Italian population (N=861)15; standardized T points were corrected by sex and age (≤31 vs. >31 y). Factor analysis revealed 2 dimensions: without (type 1) and with (type 2) physical contact. The first dimension is characterized by verbal hostility, irritability, negativism, and indirect hostility; the second dimension is characterized by suspicion, resentment, assault, and guilt. The ratio Guilt/Total BDHI clustered patients into: (1) ego-dystonic; and (2) ego-syntonic aggression groups. Patients obtaining Total BDHI scores >50 were considered to display highly aggressive behavior; patients who had scores of <50 were considered to have a low level of aggressive behavior compared with the general population (standardization sample).

SCL-90 Anger-Hostility Factor

Developed by Derogatis et al,16 the SCL-90 is made up of 90 symptoms, each of which is divided into 5 levels of severity. Usually, these items can be grouped into 9 factors: Somatic Symptoms, Obsessive-Compulsive Symptoms, Interpersonal Sensitivity, Depression, Anxiety, Anger and Hostility, Phobic Anxiety, Paranoid Ideation, and Psychoticism. The sum of all the symptoms represents the total of the scale; ancillary indices are the number of symptoms present at the time of examination and their severity, calculated by dividing the total score by the number of recorded symptoms. The stereotype of the Hostility factor identifies a patient who feels easily annoyed or irritated, has uncontrolled tantrums, often feels the urge to hit, hurt someone, feels the impulse to break objects, or engage in other frequent arguments, screams, and hurls objects.

SCL-90 V/S Factor

Maremmani and colleagues12,17 have instead identified a 5-factor structure that seems to be specific to the SUD. The 5 factors were renamed on the basis of the symptoms carrying higher weight among the factors. The first factor is characterized by a Worthlessness/Being Trapped (W/BT) depressive dimension, which represents ∼30% of the total variance of the instrument. The second factor, which explains about 4% of the total variance, is represented by a dimension of Somatic Symptoms (SS), in which somatic symptoms of opioid withdrawal appear. The third factor identifies a dimension of Sensitivity/Psychoticism (S/P), which represents about 3% of the total variance. Symptoms of panic and agoraphobia are present in the fourth dimension, named Panic Anxiety (PA), which accounts for about 2% of the total variance. Finally, the fifth dimension, with its 2% of the total variance, can be called V/S. These 5 factors taken together cover ∼40% of the total variance of the instrument. On the basis of the highest Z score obtained by the patients in the dimensions, it is possible to typify the addicted patients into 5 subgroups, mutually exclusive and distinguished by the symptoms mentioned above. About 14% of patients can be classified as W/BT; 24% are distinguished by the presence of SS; 19% show predominantly S/P symptoms; 22% show PA symptoms, and 19% show a preponderance of V/S feelings.

The V/S dimension qualifies patients with a tendency to scream and throw objects, to feel the impulse to break objects, and be dyscontrolled. Expressed more accurately, this dimension is mainly characterized by aspects that, according to the standard factorization, belong to the hostility, depression, and anxiety factors (above all, verbal and indirect aggressiveness). Patients scream and throw objects, often feel the impulse to destroy things, have to deal with uncontrolled anger, engage in frequent arguments, and feel inclined to hit, hurt themselves, and hurt others. Death ideation and suicidal ideation are also included in this dimension. Note that ideas of death, in the standard version, do not fit into any factor.


Aggressive Behavior in Heroin Addicts at Treatment Entry

Our group investigated the psychopathologic dimension in a sample of 1055 heroin addicts entering an opioid-agonist treatment (OAT). In particular, patients were asked to fill in the SCL-90. On the basis of the scores obtained, all patients were assigned to 1 of 5 mutually exclusive groups: W/BT (14.2%), SS (24.4%), S/P (19.4%), PA (22.3%), and V/S (19.7%). In particular, youngsters showed higher scores for the last 3 factors listed.12

Subsequently, we evaluated aggressiveness in a sample of 252 heroin-addicted patients entering OAT treatment from 1994 to 2012 by means of the Drug Addiction History Questionnaire (DAH-Q) and the BDHI. Overall, 81.3% of the participants showed highly aggressive behavior. Of these participants, 23.8% showed a type 1 dominant profile of aggressive behavior, consisting of verbal hostility, irritability, negativism, and indirect hostility (aggressive behavior without physical contact), whereas 76.2% showed a type 2 dominant profile, including suspicion, resentment, assault, and guilt (aggressive behavior with physical contact). More than half of the patients perceived aggressive behavior as being ego-dystonic (55.6%), whereas the remaining (44.4%) perceived it as ego-syntonic. On the whole, heroin-addicted patients showed higher scores in all factors of BDHI, compared with the general population, with type 2 emerging as the predominant profile. As to addiction history, type 2 patients showed altered mental status and unsatisfactory social leisure activity, whereas patients with a low level of aggressiveness showed fewer legal problems and more frequent self-detoxifications.18

During the treatment, Gerra and colleagues investigated aggressiveness in a sample of 18 abstinent HUD patients and 18 controls. SCL-90 evaluated all participants, whereas BDHI and the Child Experience of Care and Abuse-Questionnaire were applied to test the perception of parental neglect retrospectively. Oxytocin serum levels were measured too. Results revealed that SCL-90, BDHI, and Child Experience of Care and Abuse-Questionnaire scores were significantly higher in probands than in controls. Furthermore, oxytocin serum levels were positively correlated with psychiatric symptoms, aggressiveness, and motherly neglect. By contrast, no correlation between oxytocin and extent of exposure to heroin, or heroin dosages, was reported.19

Aggressive Behavior of HUD Participants Compared With Nonsubstance-User Peers

A more recent study evaluated aggressiveness in a sample of 73 HUD patients and 45 substance nonuser peers, matched for sociodemographic features. Compared with the general population, HUD patients were deviant in all the items included in the BDHI: Suspiciousness (T=9.31, P<0.001), Indirect Aggression (T=8.59, P<0.001), Negativism (T=6.80, P<0.001), Verbal Aggression (T=6.73, P<0.001), and Resentment (T=6.38, P<0.001), whereas Nonsubstance-User (NSU) peers proved to be deviant in Indirect Aggression (T=4.46, P<0.001) and Suspiciousness (T=2.65, P<0.011). Furthermore, HUD participants turned out to be significantly more aggressive than their NSU peers in Assault (F=19.89), Negativism (F=14.34), Suspiciousness (F=14.34), Resentment (F=7.82), Guilt (F=6.26), and Verbal Aggression (F=6.10).20

Aggressive Behavior in DD-HUD

In another study on 1090 HUD patients entering OAT, we focused on dual diagnosis. Patients were assessed using the DAH-Q, and divided into 3 groups according to their degree of aggressiveness: no lifetime self-harm or assault episodes (n=808), at least 1 moderate/superficial self-harm episode in the last month before requesting treatment (n=30), and at least 1 assault episode in the previous month (n=162). DAH-Q was use to evaluate the patients. Dual diagnosis, in particular bipolar spectrum, was found to be the most significant risk factor for aggressive behavior. Other factors included unstable modality of heroin use, a diagnosis of chronic psychosis, and the simultaneous use of central nervous system stimulants and depressants. Moderate/superficial self-harm was associated with chronic psychosis, whereas depressive (nonbipolar) or anxiety disorders and older age at first use of heroin were correlated with a lower risk of aggressive behavior.21

Subsequently, we assessed 1195 patients entering Therapeutic Community (TC). Overall, 25.9% of these (n=309) were affected by 1 or more psychiatric disorders (DD-HUD), whereas 74.1% (n=886) did not show any psychiatric comorbidities (NDD/HUD). SCL-90 total and single domain scores turned out to be higher in DD-HUD patients than in NDD-HUD patients (P=0.000, <0.001).22

Gerra and colleagues also investigated the aggressive responses showed in a sample of 20 HUD patients in OAT with methadone, comparing them with 20 healthy controls. Notably, 13 of the heroin addicts showed psychiatric symptoms (anxiety, depression, antisocial personality traits), whereas only 2 of 20 fulfilled the diagnostic criteria for a psychiatric disorder [one with major depression (MD) and the other with obsessive-compulsive disorder]. Patients were evaluated using the Italian version of the BDHI and underwent the Point Subtraction Aggression Paradigm (PSAP), an experimental model to induce aggressiveness. As expected, aggressive responses were significantly higher in patients than in the control group (P<0.001) and they were correlated with BDHI assault and irritability scores. By contrast, no correlation was found with methadone treatment.23 Similarly, the same authors reported no impact of buprenorphine on aggressiveness. In particular, PSAP aggressive responses were evaluated in a sample of 30 HUD patients under treatment with buprenorphine (BUP, n=15) or methadone (METH, n=15), and in 15 healthy controls. Among HUD participants, 15 showed psychiatric symptoms, although only 7 fulfilled the criteria for an established diagnosis. As reported previously, aggressiveness was significantly higher in the HUD group, with no differences between the BUP and METH groups.23 Both studies found that aggressiveness in HUD patients was correlated with personality traits and monoamine plasma levels, whereas no role of OAT emerged.23,24

Suicidality in Heroin Addicts

Our group investigated suicidality in a sample of 616 heroin addicts in OAT. Suicidal thoughts during the previous week were reported in less than one third of the sample (29.1%). Overall, 2.8% of the patients showed the highest severity of suicidal thoughts, according to the SCL-90, whereas they were light or moderate in about one fifth of the sample. Moreover, depression and hostility dimensions were correlated with suicidal thoughts. In terms of sociodemographic variables, a significant association between being unemployed, not receiving welfare benefits, living alone, belonging to blue-collar or unemployed families, and suicidality was found.25

In 2008, Trémeau and colleagues reported significantly higher scores in the assault and irritability subscales of the BDHI in HUD with a personal history of suicidal attempts or family history of suicide. The latter also showed higher BDHI and Barratt-Impulsiveness Scale (BIS) total scores.26

Some years later, Kazour and colleagues investigated suicidality and other psychopathologic dimensions in a sample of 61 HUD inpatients (vs. 61 controls) with the Barratt-Impulsiveness Scale, version 11 (BIS-11), the Hamilton Depression Rating Scale (HAMD), and the Beck Suicidal Ideation Scale (BSI). HUD patients showed significantly higher scores than controls on BIS-11 (82.11 vs. 45.74), HAMD (14.7 vs. 2.43), and BSI (12.55 vs. 1.47). In line with the previously mentioned study, scores were higher in patients with a personal history of attempts at suicide compared with those who had never attempted suicide (HAMD, P=0.004; BSI, P<0.001; BIS-11, P=0.014). Furthermore, a positive correlation between HAMD, BIS-11, BSI scores, and the number of lifetime suicidal attempts emerged (BSI, r=0.51, P<0.001; HAMD, r=0.38, P=0.002; BIS-11, r=0.28, P<0.03).27

In a broad sample of 3949 HUD patients, the standardized mortality rate for suicide was found to be 4-fold higher than that in the general population, with sex differences (3.7 higher for men compared with the general male population and 2.2 higher than women; 7.0 higher in women, compared with the general female population). Older age was correlated with a higher risk of suicide, whereas no correlation was reported between male sex, living alone, and unemployment. By contrast, being at first OAT seemed to play a protective role.28

Zhong and colleagues investigated nonsuicidal self-injury (NSSI) in HUD. In HUD patients (n=603) already in methadone maintenance treatment (MMT), 13.8% of patients reported NNSI during the previous month, whereas the most prevalent types of self-injury were burning (59%), cutting (19.3%), hitting (9.6%), and carving (6.0%). A positive correlation was found between NSSI and unemployment [odds ratio (OR)=2.54, P=0.009], a short duration of MMT (OR=1.04, P=0.034), pain (OR=2.3, P=0.028), depression (OR=4.32, P<0.001), anxiety (OR=3.74, P=0.002), and loneliness (OR=3.04, P=0.012)29. According to a meta-analysis on Chinese HUD patients in compulsory or voluntary detoxification treatment (15 studies included, total sample n=37,243), the pooled prevalence of NSSI was 4.4% (2.8% to 6.2%), including swallowing foreign objects, especially those designed for cutting, hitting, or burning, jumping from a height, cutting off fingers, apastia, attempts to break bones, knocking a nail into the head, and drug overdoses.29

Role of the V/S Dimension in HUD Patients

V/S typology is related to a natural history of HUD. Older patients were more in number in the W/BT group, whereas younger patients were more in number among S/P and V/P patients.30 The sample consisted of 455 heroin-dependent patients (according to the DSM III/IIIR/IV/IVR criteria), of whom 340 (74.9%) were men and 115 (25.3%) were women. The average age of the patients was 28±7 years (range: 16 to 50). Most of the patients were single (N=295; 64.8%), had had <8 years of education (N=346; 76.0%), and were unemployed (N=212; 46.6%).

There is a significant association between V/S severity and the choice of OAT.31 We investigated psychopathologic features in a sample of 1195 patients entering a TC. Participants were clustered into 5 groups: W/BT (16.7%), SS (17.3%), S/P (21.1%), PA (29.3%), and V/S (15.5%). In particular, V/S and W/BT, SS, and PA dimensions were more severe in OAT patients and women, compared with the TC group and the men in the TC group, respectively. V/S was more severe for OAT versus TC male patients, but did not differ in TC versus OAT women.

V/S was more severe in NDTX versus DTX HUD patients, but the typology did not differ,32 in a sample of 1015 HUD patients who underwent detoxification (DTX, n=374) or not (NDTX, n=641). On the whole, SCL-90 total scores, and the V/S dimension, were significantly higher in NDTX patients than in DTX ones.32

The V/S typology was less frequent in alcohol use disorder (AUD) versus HUD and cocaine use disorder (CUD) patients.33 According to the VOECT cohort study (Evaluation of Therapeutic Community Treatments and Outcomes), carried out in 2008-2009 on a sample of 2533 patients diagnosed with SUD (AUD, n=449; CUD, n=670; HUD, n=1195) admitted to a TC, the V/S dimension was higher in HUD and CUD patients than in those diagnosed with AUD. Furthermore, no differences were reported in V/S domain scores between HUD patients with CUD or AUD and patients without any other SUD.

The V/S typology did not differ among HUD patients according to the secondary substance of use (alcohol, cocaine, or none).33 In the second part of the same study, the most frequent psychopathologic dimension was PA for all the 3 groups of patients. The least frequent were V/S and S/P for the group of patients with alcohol as the secondary abused drug, V/S for the group of patients with no alcohol or cocaine as the secondary drug, and W/BT for the group of patients with cocaine as the secondary abused drug. No statistically significant differences between the 3 groups were observed in any of the 5 SCL-based psychopathologic dimensions.

V/S severity and typology did not differ in AUD, HUD, and CUD monousers.34 In one of our studies on 256 patients diagnosed with HUD, CUD, or AUD, with no other substance use comorbidities, no differences in the V/S dimension emerged among the 3 groups.

Traumatic life-events can also influence aggression in HUD patients. V/S severity was higher in posttraumatic stress disorder spectrum (H/PTSD-S) HUD patients; conversely, the V/S typology did not differ.35 In terms of the role of life-events in patients with HUD, we found a higher V/S severity in those with heroin posttraumatic stress disorder (PTSD) spectrum H/PTSD-S symptoms. In particular, we investigated psychopathologic and PTSD symptoms in a sample of 93 HUD patients using the SCL-90 and the Trauma and Loss Spectrum Questionnaire. Patients were divided into 2 groups according to their Trauma and Loss Spectrum Questionnaire scores, and the presence or absence of H/PTSD-S (>32, presence, and <32, absence, respectively). H/PTSD-S HUD patients reported more severe scores in all the domains of SCL-90, including V/S, compared with no H/PTSD-S patients.

Ethnicity does not seem to interfere with the V/S dimension. In fact, in a naturalistic case-control study, V/S severity did not differ in 30 migrant HUD patients heading for Italy, compared with 30 Italian counterparts matched for age and sex.36 Conversely, V/S severity was higher in Slovenian HUD patients compared with Italian ones. Sixty-six Slovenians and 66 Italian HUD patients matched according to age and sex, at univariate analysis, showed a more severe V/S dimension, whereas multivariate discriminant analysis revealed a less significant difference between the 2 groups (P<0.026).37 In the same study, V/S typology did not differ either in migrant versus Italian HUD patients or in Slovenian versus Italian ones.

The 5-factor SCL-90 psychopathologic dimensions are strongly correlated with HUD trait-conditions.38 The presence and severity of addictive behaviors were recorded utilizing CRAV-HERO, an inventory for assessing the behavioral covariates of craving in HUD patients. Thirteen behaviors were selected. We clustered the 13 behaviors into 6 operating models.

  • Exchange-related addictive behaviors (EXC-BEHAV) that aimed to reveal the hierarchical approach applied by a patient to his/her values.
  • Time-related addictive behaviors (TIME-BEHAV) that test the patient’s ability to wait and manage the substance, and how much time is taken up by thinking about the substance.
  • Risk-related addictive behaviors (RISK-BEHAV) that are linked to the theme of risk in which the choice of whether to engage in substance use raises the question of endangering health, even life itself.
  • Cue-induced/Environment-related addictive behaviors (CUE/ENV-BEHAV).
  • Reward-craving-induced behaviors (REW-BEHAV).
  • Relief-obsessive craving-induced behaviors (REL/OBS-BEHAV).

On the basis of the highest Z scores obtained on the CRAV-HERO clusters, patients can be assigned to 1 of these 6 mutually exclusive groups. For details, see the study by Maremmani et al.39

The V/S dimension was strongly correlated with heroin-addictive behavior (exchange and time items). The results of the canonical correlation analysis between 5 psychopathologic dimensions and 5 different types of addictive behaviors showed that only 1 canonical variate was significant. Canonical variate set-1 (SCL-90), which accounts for 20.11% of the total variance, was saturated negatively by V/S, SS, and PA dimensions and positively (but at a superficial level) by the S/P dimension. Set-2 (CRAV-HERO), which accounts for 83.68% of the total variance, was saturated negatively by TIME-BEHAV, EXC-BEHAV, RISK-BEHAV, and CUE/ENV-BEHAV related to REL/OBS-BEHAV. The addictive behavior set was saturated positively (but at a superficial level) by CUE/ENV-BEHAV related to REW-BEHAV. These sets were significantly correlated. V/S severity is, therefore, closely related to the severity of SS (withdrawal syndrome) and linked to the hierarchical approach applied by a patient to his/her values and to that patient’s ability to wait and manage the substance, and to how much time is taken up by thinking about the substance.

The 5-factor SCL-90 psychopathologic dimensions can differentiate HUD patients from other psychopathologic patients, but V/S was more severe and more frequent in MD patients.40 We compared 972 HUD patients with 504 MD patients to estimate the magnitude of the differences in terms of psychopathologic symptoms. Prominent psychopathologic domains were more frequent in HUD patients, in particular, W/BT, SS, and S/P. The V/S dimension was more frequent in MD patients, whereas the PA dimension failed to differentiate between the 2 groups. The prominent psychopathologic groups are the most crucial factor in significantly differentiating between the 2 groups when drawing comparisons on the basis of age, male sex, and the severity of psychopathologic symptoms. The frequency of the association between addiction and mood disorders may be explained on both the neurobiological41 and clinical levels.42,43 This study suggests that, in HUD patients, depressive symptomatology remains the most critical and frequent psychopathologic aspect of HUD. Moreover, this symptomatology is less closely related to suicidal ideation than in depressed patients.

V/S severity and typology were shown to be more marked in nonpsychiatric obese patients by comparing 972 HUD patients with 106 obese participants; the severity of all psychopathologic dimensions was, in fact, significantly higher in obese individuals. Discriminant analysis showed that PA and V/S severity were greater in obese patients, to a degree sufficient to allow differentiation between HUD (lower severity) and obese individuals (greater severity). At the reclassification level, 70.8% of obese individuals in the sample were reclassified as HUD patients. Psychopathologic subtypes characterized by PA and V/S typology were more frequent in obese patients, sufficiently so as to allow differentiation between groups. Psychopathologic subtypes characterized by W/BT, SS, and S/P symptomatology were more frequent among HUD patients, whereas PA and V/S symptomatology were more frequent among obese individuals.44

V/S severity and typology were more marked in HUD than in gambling disorder (GD) patients.45 We compared the severity and frequency of each of the 5 aspects found by us in 972 HUD and 110 GD patients at univariate and multivariate levels. HUD patients showed higher general psychopathology indexes than GD patients. The severity of all 5 psychopathologic dimensions was significantly greater in HUD patients. Discriminant analysis revealed that SS and V/S severity could differentiate between HUD (higher severity) and GD patients (lower severity), whereas PA and S/P could not. V/S correlated positively with SS and negatively with the W/BT dimension. HUD patients were distinguished by high scores for SS, whereas high scores for V/S were associated with low ones for W/BT. Psychopathologic subtypes marked out by V/S and SS symptomatology were better represented in HUD patients, whereas PA symptomatology occurred more frequently in GD individuals. The similarities to be found on neurobiological, genetic, and clinical grounds between SUD and GD seem to be confirmed once again at the psychopathologic level. Interestingly, the highest scores in the W/BT dimension can differentiate GD from HUD patients when they are associated with a low degree of severity of the SS and V/S dimensions. To sum up, HUD patients are distinguished by a more severe withdrawal syndrome and more severe aggressiveness, and GD patients by a higher degree of severity of the W/BT dimension.

By contrast, V/S severity and typology failed to differentiate HUD from chronic psychotic patients, when 40 chronic psychotic patients (CHR-PSY) were matched with 33 HUD patients according to age and sex, and compared, at univariate and multivariate levels, on the severity and typology of 5 dimensions of SCL-90.46


Effect of Opioid Medications on Aggression

The presence of SCL-90 hostility predicts an inadequate clinical response in naltrexone-treated HUD patients. Naltrexone has been shown to have poor results on unselected populations of heroin addicts. Its use is mostly confined to detoxification-related procedures, whereas its long-term effects and properties have been largely neglected. In one of our studies, we investigated the predictors of a successful outcome in a population of 149 patients diagnosed as heroin addicts on the basis of DSM-IV criteria and undergoing long-term naltrexone treatment (naltrexone maintenance). Favorable outcomes were related to ongoing treatment, whereas negative outcomes were due to treatment discontinuation through addictive relapse. Retained individuals are more likely to have no problems at work and to be psychosocially adjusted. Earlier substance users are those most likely to dropout. Global psychopathologic impairments, with particular reference to mood, aggressiveness, and delusions, are negatively related to treatment retention. Naltrexone maintenance appears to be suitable for a subgroup of heroin abusers whose clinical pictures combine a low level of addictive disease with the absence of significant dysphoria, aggressive behavior, and psychosis.47

SCL-90 Hostility severity predicts an inadequate clinical response in general practitioners’ (GPs’) Office-based MMT. In one of our studies, we evaluated the effectiveness of methadone treatment carried out by GPs in Trieste, Italy, and we identified response treatment factors. Thirty-three patients with heroin addiction according to the DSM-IV-R criteria were placed in an observational protocol with an average duration of 429±273 days. The retention rate was used as a measure of outcome. Patients with higher severity of illness, with problematic relationships with their spouse/partner, difficulty with socialization and organization of leisure, with an altered mental state at the beginning of treatment, patients with dual diagnosis (especially bipolar disorder), with greater severity of obsessive-compulsive symptoms, interpersonal sensitivity, depression, violence, with greater severity of psychopathologic symptoms, with the most significant number of problematic areas in terms of quality of life, and patients with a low dose of methadone administered for treatment were those considered to be at the highest risk of abandoning treatment.48

Higher dosages of methadone are needed to stabilize violent patients. Using the SCL-90’s hostility factor and the BDHI, in one of our studies, we verified that methadone dosages depend on the grade of psychopathology and aggressiveness at treatment entry. A sample of 20 patients was divided into 2 clusters according to the baseline SCL-90 score (high psychopathology vs. low psychopathology). All these patients had been abstinent from various substances for a long time and had achieved a satisfactory level of psychosocial adaptation after a treatment period of variable length (1 to 96 mo). Stabilization dosages ranged from 7 to 80, averaging 39±23 mg/d. A higher degree of psychopathology corresponded to higher stabilization dosages (60 vs. 30 mg/d on average, the latter corresponding to a lower degree of psychopathology); similarly, higher aggressiveness accounted for higher stabilization dosages (50 vs. 30 mg/d for mildly aggressive patients). Neither psychopathology nor aggressiveness appeared to vary with treatment duration. Methadone-sensitive psychopathology appeared to comprise depression, phobic anxiety, paranoia, physical features, and psychotic symptoms, with the latter 2 showing the strongest correlations. In terms of BDHI-recorded aggressiveness, methadone dosage seemed to be related to unexpressed aggression, irritability, and violence, the most active associations emerging for the latter 2. In conclusion, the higher the level of psychopathology and aggressiveness at treatment entry, the higher the methadone dosage required for stabilization.49

The question to be asked at this point is whether these psychopathologic typologies, which can typify patients with SUD, are affected by different medications used in the treatment of addiction. In the case of heroin addiction, for example, are methadone and buprenorphine equally effective on these psychopathologic types, and what happens if the patient is not treated with medications, but, for instance, in a therapeutic community? In previous studies, we were able to demonstrate that, when the 9 standard dimensions of the SCL-9016 were used, the effect was not specific, because almost all 9 psychopathologic factors improved.50,51

In a sample of 213 HUD patients treated with opiate agonists (methadone or buprenorphine), at the end of the 12 months of observation, no significant outcome differences could be found in patients who showed prominent symptoms of W/BT, SS, and PA, a result that could depend on the use of 1 of these 2 medications.50

Patients who showed the presence of S/P symptomatology also showed higher retention in treatment if treated with methadone, irrespective of sex, educational level, marital status, the presence of somatic and psychiatric comorbidity, social adjustment, legal problems, and polysubstance use at treatment entry. Achievement of a positive outcome was also independent of patients’ age, age at the onset of heroin use, age at the onset of continuous use, how long dependence lasted, and age at first treatment. During follow-up, no differences in the use of opiates or cocaine were observed. It should be borne in mind that, at the start of treatment, working conditions in methadone-treated patients were worse than in those treated with buprenorphine. By contrast, patients showing V/S symptoms achieved significantly higher retention in treatment if they were treated with buprenorphine. This result was independent of patients’ employment status, their educational level, civil status, social adjustment, the presence of somatic and psychiatric comorbidity, or legal problems, even polysubstance use. For these patients too, no differences were found in drug history or the efficacy of the 2 medications on substance use during treatment. In the sample group treated with buprenorphine, men were preponderant, showing more severe maladjustment in the social/leisure area. In the sample examined by us, therefore, methadone and buprenorphine showed the same impact on addiction pathology, as demonstrated by urinary tests, but different types of impact on the psychopathologic subtypes that we consider specific to SUD.52

Gerra and colleagues investigated aggressive behavior in a sample of 20 heroin-addicted patients in methadone or buprenorphine maintenance treatment, using a laboratory task: the PSAP. The aggressive response was found to be higher in the heroin-dependent patient group than in the control group, independent of agonist treatment.53 Gerra and colleagues found a possible role of olanzapine in improving aggressiveness in HUD patients. A sample of 67 HUD patients in OAT with methadone or buprenorphine was co-administered with olanzapine (OLA group) or selective serotonin reuptake inhibitors (SSRIs) (fluoxetine or paroxetine) and clonazepam (SSRI+benzodiazepine group) for 12 weeks. Among those who completed the observational period (n=33, 49.2%), a significant decrease in BDHI total scores and direct, indirect, verbal aggressiveness, irritability, and resentment subscales scores were found in the OLA group, whereas no differences were reported in the SSRI+benzodiazepine group.54 Similarly, according to a 12-week Iranian trial, augmentation therapy with olanzapine or valproate can reduce aggressive behavior in HUD patients in MMT. Notably, 201 HUD patients in MMT were randomized into 2 groups, the first group receiving olanzapine (2.5 to 15 mg/d) plus placebo valproate (n=101, OLA group) and the second group receiving valproate (600 to 1000 mg/d) plus olanzapine placebo (n=100). Patients were evaluated on the Overt Aggression Scale-Modified at baseline and weekly. Among those who completed the trial (n=53), Overt Aggression Scale-Modified total score and aggression, irritability, and suicidality subscales decreased significantly, with olanzapine being more effective than valproate.55 Evren and colleagues studied aggressiveness and impulsivity in a sample of 52 HUD patients in maintenance treatment (maintenance group, MG) followed up for 12 months. Of these, 44.23% (n=23) relapsed during the follow-up (relapse group, RG). Patients were assessed using the BIS-11, and the Buss-Perry Aggression Questionnaire at baseline (T0) and 12 months (T1). Compared with MG, RG showed higher mean verbal aggression (T0), physical aggression, and impulsivity scores (T1). On the whole, at T1, aggression and impulsivity were lower in MG (hostility, motor, and nonplanning impulsivity), whereas they were higher in the RG patients (motor and verbal aggression, attentional and nonplanning impulsivity). Interestingly, low verbal aggression and high motor impulsiveness at T0 and anger and motor impulsiveness at T1 may predict a relapse.56

Thus, a series of studies indicate that opiate agonists are likely to be effective in controlling aggressive behavior in opiate-addicted patients, as confirmed by the decrease in levels of aggressiveness that followed adequate methadone treatment.57,58 Moreover, aggressive symptoms are among the features that may be found frequently in cases of self-medication.59 In our study, buprenorphine showed better results than methadone in patients with prominently aggressive characteristics (in the V/S dominant group).

The observations reported in the literature and the results of our studies suggest that opioid agonists should be reconsidered, as they not only possess an anticraving activity but are also able to act as psychotropic instruments in treating mental illness, with particular reference to mood, anxiety, and psychotic syndromes. In particular, methadone seems to be more effective in treating S/P aspects, whereas buprenorphine appears to be more effective in acting against aggressive behavior (V/S). As a result, some dual-diagnosis patients may benefit from a treatment (methadone or buprenorphine) that not only targets their addictive problem but is also active in attenuating their mental disorder.

Is the psychopathologic SUD typology proposed by us able to show its involvement in the results of residential treatment? To answer this question, 2016 patients treated in various therapeutic communities were classified psychopathologically and monitored for up to 16 months.60 Abandonment of the treatment was considered the focal event; it determined the level of retention in treating these patients. The sample consisted of 2016 SUD patients diagnosed according to a clinical judgment; there were 1693 men and 323 women. At the end of the study, the cumulative retention rate was 0.39. The W/BT dimension was prominent in 298 patients, SS in 456; S/P in 406; PA in 518; and V/S in 338. Retention rates differed statistically between the 5 subgroups. In particular, V/S patients showed a lower retention rate than those for the W/BT, SS, S/P, and PA dimensions. Not having been detoxified on entering residential treatment and the presence of psychopathologic symptoms were 2 factors that had a negative influence on the outcome. Prominent SS and V/S symptomatology at treatment entry both correlated positively with dropout from the therapeutic community.

In conclusion, TC programs show a considerable variety of results, depending on the psychopathologic typology of patients. Length of retention in the treatment of patients entering TC treatment is significantly lower for those who have a more severe psychopathology. Moreover, V/S and SS patients may leave the treatment earlier than those allocated to the other 3 psychopathologic dimensions resulting from the application of PCA to the SCL-90 responses (ie, W/BT, S/P, and PA). The SCL-90 5-factorial structure of the psychopathology of substance dependence could turn out to be a useful tool when applied as a prognostic factor, together with age, detoxification status, and kind of substance of abuse, all of which have been shown to influence retention in treatment.

In summary, our studies show that 8 out of every 10 patients entering treatment show highly aggressive behavior; in 2.5 out of every 10, this occurs without physical contact; in 7.5 out of every 10, this occurs with physical contact; and 4.5 out of every 10 patients have an ego-syntonic perception of their aggressive behavior. At treatment entry, BDHI Negativism and Assault clearly differentiate HUD patients from their NSU peers; violence and self-harm are found to be correlated in DD/HUD patients; in DD/bipolar-HUD suicidality increases; predominantly V/S patients are more frequent in younger HUD age brackets; the V/S dimension only minimally affects HUD state-conditions (choice of treatment, active substance use, psychiatric comorbidity, primary and secondary used substance, mono-substance use and ethnicity); the V/S dimension is strictly correlated with HUD trait-conditions (addictive behavior and PTSD spectrum); and the V/S dimension can differentiate HUD from other psychiatric patients (depressed and psychotics) and nonpsychiatric ones (obese), but is not able to differentiate patients affected by behavioral addictions (GD). On the therapeutic level, the presence of aggressive behavior negatively influences GPs’ office-based OAT and naltrexone maintenance. A higher methadone dosage is needed to stabilize AO-treated aggressive patients; predominantly V/S patients are well stabilized when in buprenorphine treatment.


We recognize that our conclusions are based on clinical considerations. In contrast, innovative strategies such as those presented here are often necessary in clinical practice.


Our studies suggest an in-depth psychopathologic evaluation of the HUD patients, with particular attention to aggressive symptoms, before their entry into long-term OAT. Moreover, our data are compatible with Khantzian self-medication theory, which views heroin addiction as a way of controlling violent manifestations. For this reason, the use of an adequate amount of opioid agonist seems to be crucial in the management of violent opioid addicts.


1. Freud S. Beyond the Pleasure Principle. London, UK; Vienna, Austria: The International Psycho-Analytical Press; 1922.
2. Klein M. The Psychoanalysis of Children. New York, NY: Grove Press; 1960.
3. Adler A. Understanding Human Nature. New York, NY: Greenberg; 1927.
4. Fromm E. The Anatomy of Human Destructiveness. London, UK: Cape; 1974.
5. Storr H. Human aggression. London, UK: Penguin; 1995.
6. Lorenz K. On aggression. New York, NY: Harcourt, Brace& World; 1966.
7. Dollard J, Doob LW, Miller NE, et al. Frustration and Aggression. New Haven, CT: Yale University; 1939.
8. Berkowitz LBerkowitz L. The frustration-aggression hypothesis revisited. Roots of Aggression. New York, NY: Atherton Press; 1969:1–28.
9. Bandura A. Aggression: A Social Learning Analysis. Englewood Cliffs, NJ: Prentice Hall; 1973.
10. Buss AH, Durkee A. An inventory for assessing different kinds of hostility. J Consult Clin Psychol. 1957;21:343–349.
11. Regier DA, Kuhl EA, Kupfer DJ. The DSM-5: classification and criteria changes. World Psychiatry. 2013;12:92–98.
12. Maremmani I, Pani PP, Pacini M, et al. Subtyping patients with heroin addiction at treatment entry: factor derived from the self-report symptom inventory (SCL-90). Ann Gen Psychiatry. 2010;9:15.
13. Derogatis LR, Lipman RS, Covi L. Scl-90: an outpatient psychiatric rating scale—preliminary report. Psychopharmacol Bull. 1973;9:13–28.
14. Castrogiovanni P, Andreani MF, Maremmani I, et al. Per una valutazione dell’aggressività nell’uomo: Contributo alla validazione di un questionario per la tipizzazione del comportamento aggressivo [Assessing aggression in humans: Validation of a questionnaire for assessing different kinds of hostility]. Riv Psichiatr. 1982;17:276–295.
15. Maremmani I, Massimetti G, Bozzi G, et al. Un questionario per la tipizzazione del comportamento aggressivo dal possibile impiego in psicosomatica. Prima standardizzazione nella popolazione italiana [A questionnaire to typify aggressive behavior usable in psychosomatics. First standardization in the Italian population]. Boll Soc Med Chir Pisa. 1988;1:37–43.
16. Derogatis LR, Lipman RS, Rickels K. The Hopkins Symptom Checklist (HSCL). A self report symptom inventory. Behav Sci. 1974;19:1–16.
17. Maremmani I, Maremmani AGI, Pani PP. Psicopatologia del disturbo da uso di sostanze. Pisa and Pietrasanta, Italy: Pacini Editore Medicina & AU-CNS; 2016.
18. Bacciardi S, Maremmani AGI, Rovai L, et al. Aggressive behavior in heroin dependent subjects at treatment entry. Heroin Addict Relat Clin Probl. 2013;15:5–13.
19. Gerra LM, Gerra G, Mercolini L, et al. Increased oxytocin levels among abstinent heroin addicts: association with aggressiveness, psychiatric symptoms and perceived childhood neglect. Prog Neuropsychopharmacol Biol Psychiatry. 2017;75:70–76.
20. Conversano C, Belcari I, Marchi L, et al. Personality profiles and aggressive behavior of heroin use disorder patients compared with non-substance-use peers. Heroin Addict Relat Clin Probl. 2018;20:45–54.
21. Maremmani AG, Rugani F, Bacciardi S, et al. Does dual diagnosis affect violence and moderate/superficial self-harm in heroin addiction at treatment entry? J Addict Med. 2014;8:116–122.
22. Pani PP, Maremmani AGI, Trogu E, et al. Psychic structure of opioid addiction: impact of lifetime psychiatric problems on SCL-90-based psychopathologic dimensions in heroin-dependent patients. Addict Disord Their Treatment. 2016;15:6–16.
23. Gerra G, Zaimovic A, Raggi MA, et al. Aggressive responding of male heroin addicts under methadone treatment: psychometric and neuroendocrine correlates. Drug Alcohol Depend. 2001;65:85–95.
24. Gerra G, Zaimovic A, Raggi MA, et al. Experimentally induced aggressiveness in heroin-dependent patients treated with buprenorphine: comparison of patients receiving methadone and healthy subjects. Psychiatry Res. 2007;149:201–213.
25. Maremmani I, Pani PP, Canoniero S, et al. Is the bipolar spectrum the psychopathological substrate of suicidality in heroin addicts? Psychopathology. 2007;40:269–277.
26. Tremeau F, Darreye A, Staner L, et al. Suicidality in opioid-dependent subjects. Am J Addict. 2008;17:187–194.
27. Kazour F, Soufia M, Rohayem J, et al. Suicide risk of heroin dependent subjects in Lebanon. Community Ment Health J. 2016;52:589–596.
28. Lovrecic M, Lovrecic B, Maremmani I, et al. Excess suicide mortality in heroin use disorder patients seeking opioid agonist treatment in Slovenia and risk factors for suicide. Heroin Addict Relat Clin Probl. 2018;20:35–40.
29. Zhong BL, Xu YM, Zhu JH, et al. Non-suicidal self-injury in chinese heroin-dependent patients receiving methadone maintenance treatment: prevalence and associated factors. Drug Alcohol Depend. 2018;189:161–165.
30. Maremmani AGI, Rovai L, Maremmani I. Heroin addicts’ psychopathological subtypes. Correlations with the natural history of illness. Heroin Addict Relat Clin Probl. 2012;14:11–22.
31. Pani PP, Trogu E, Vigna-Taglianti F, et al. Psychopathological symptoms of patients with heroin addiction entering opioid agonist or therapeutic community treatment. Ann Gen Psychiatry. 2014;13:35.
32. Pani PP, Maremmani AGI, Trogu E, et al. Psychopathological symptoms in detoxified and non-detoxified heroin-dependent patients entering residential treatment. Heroin Addict Relat Clin Probl. 2015;17:17–24.
33. Pani PP, Maremmani AGI, Trogu E, et al. Psychopathology of addiction: may a SCL-90 based five dimensions structure be applied irrespectively of the involved drug? Ann Gen Psychiatry. 2016;15:13.
34. Carbone MG, Maiello M, Spera V, et al. The SCL90-based psychopathological structure may be applied in substance use disorder patients irrespectively of the involved drug also in heroin, alcohol and cocaine mono-drug users. Heroin Addict Relat Clin Probl. 2018;20:29–34.
35. Maremmani AGI, Maiello M, Carbone MG, et al. Towards a psychopathology specific to substance use disorder: should emotional responses to life events be included? Compr Psychiat. 2018;80:132–139.
36. Carbone MG, Tagliarini C, Ricci M, et al. Ethnicity and specific psychopathology of addiction. Comparison between migrant and Italian heroin use disorder patients. Heroin Addict Relat Clin Probl. 2019;21:61–66.
37. Maremmani AGI, Lovrecic M, Lovrecic B, et al. Ethnicity and specific psychopathology of addiction. Comparison between Slovenian and Italian heroin use disorder patients. Heroin Addict Relat Clin Probl. 2019;21:35–39.
38. Della Rocca F, Maremmani A, Rovai L, et al. Further evidence of a specific psychopathology of heroin use disorder. Relationships between psychopathological dimensions and addictive behaviors. Heroin Addict Relat Clin Probl. 2017;19:13–20.
39. Maremmani AGI, Rovai L, Bacciardi S, et al. An inventory for assessing the behavioral covariates of craving in heroin substance use disorder. Development, theoretical description, reliability, exploratory factor analysis and preliminary construct validity. Heroin Addict Relat Clin Probl. 2015;17:51–60.
40. Maremmani AGI, Cerniglia L, Cimino S, et al. Towards a specific psychopathology of heroin addiction. Comparison between heroin use disorder and major depression patients. Heroin Addict Relat Clin Probl. 2015;17:9–16.
41. Turgeon SM, Pollack AE, Fink JS. Enhanced CREB phosphorylation and changes in c-Fos and Fra expression in striatum accompany amphetamine sensitization. Brain Res. 1997;749:120–126.
42. Maremmani I, Pacini M, Perugi G, et al. Addiction and bipolar spectrum: dual diagnosis with a common substrate? Addict Disord Their Treatment. 2004;3:156–164.
43. Maremmani I, Perugi G, Pacini M, et al. Toward a unitary perspective on the bipolar spectrum and substance abuse: opiate addiction as a paradigm. J Affect Disord. 2006;93:1–12.
44. Maremmani AGI, Cerniglia L, Cimino S, et al. Further evidence of a specific psychopathology of addiction. Differentiation from other psychiatric psychopathological dimensions (such as obesity). Int J Environ Res Public Health. 2017;14:943.
45. Maremmani AGI, Gazzarrini D, Fiorin A, et al. Psychopathology of addiction: is the SCL90-based five dimensional structure applicable to a non-substance-related addictive disorder such as gambling disorder? Ann Gen Psychiatry. 2018;17:1–9.
46. Rugani F, Paganin W, Maremmani AGI, et al. Towards a specific psychopathology of heroin addiction. Comparison between heroin use disorder and chronic psychotic patients. Heroin Addict Relat Clin Probl. 2019;21:53–59.
47. Maremmani I, Pacini M, Giuntoli G, et al. Naltrexone as maintenance therapy for heroin addiction: predictors of response. Heroin Addict Relat Clin Probl. 2004;6:43–52.
48. Michelazzi A, Vecchiet F, Leprini R, et al. GPs’ office based metadone maintenance treatment in Trieste, Italy. Therapeutic efficacy and predictors of clinical response. Heroin Addict Relat Clin Probl. 2008;10:27–38.
49. Maremmani I, Zolesi O, Agueci T, et al. Methadone doses and psychopathological symptoms during methadone maintenance. J Psychoactive Drugs. 1993;25:253–256.
50. Maremmani I, Pani PP, Pacini M, et al. Substance use and quality of life over 12 months among buprenorphine maintenance-treated and methadone maintenance-treated heroin-addicted patients. J Subst Abuse Treat. 2007;33:91–98.
51. Pani PP, Maremmani I, Pirastu R, et al. Buprenorphine: a controlled clinical trial in the treatment of opioid dependence. Drug Alcohol Depend. 2000;60:39–50.
52. Maremmani AGI, Rovai L, Pani PP, et al. Do methadone and buprenorphine have the same impact on psychopathological symptoms of heroin addicts? Ann Gen Psychiatry. 2011;10:17.
53. Gerra G, Zaimovic A, Moi G, et al. Aggressive responding in abstinent heroin addicts: neuroendocrine and personality correlates. Prog Neuropsychopharmacol Biol Psychiatry. 2004;28:129–139.
54. Gerra G, Di Petta G, D’Amore A, et al. Effects of olanzapine on aggressiveness in heroin dependent patients. Prog Neuropsychopharmacol Biol Psychiatry. 2006;30:1291–1298.
55. Zarghami M, Sheikhmoonesi F, Ala S, et al. A comparative study of beneficial effects of olanzapine and sodium valproate on aggressive behavior of patients who are on methadone maintenance therapy: a randomized triple blind clinical trial. Eur Rev Med Pharmacol Sci. 2013;17:1073–1081.
56. Evren C, Yilmaz A, Can Y, et al. Severity of impulsivity and aggression at 12-month follow-up among male heroin dependent patients. Bull Clin Psychopharmacol. 2014;24:1.
57. Haney M, Miczek KA. Morphine effects on maternal aggression, pup care and analgesia in mice. Psychopharmacology. 1989;98:68–74.
58. Shaikh MB, Dalsass M, Siegel A. Opioidergic mechanisms mediating aggressive behavior in the cat. Aggress Behav. 1990;16:191–206.
59. Khantzian EJ. Psychological (structural) vulnerabilities and the specific appeal of narcotics. Ann NY Acad Sci. 1982;398:24–32.
60. Maremmani AGI, Pani PP, Trogu E, et al. The impact of psychopathological subtypes on retention rate of patients with substance use disorders entering residential therapeutic community treatment. Ann Gen Psychiatry. 2016;15:29.

aggressive behavior; heroin use disorder; clinical aspects; therapeutic aspects

Copyright © 2020 Wolters Kluwer Health, Inc. All rights reserved.