Psychopathology of Addiction: May a SCL-90-Based Five Dimensions

Pani et al. Ann Gen Psychiatry (2016) 15:13 DOI 10.1186/s12991-016-0100-8 Annals of General Psychiatry

PRIMARY RESEARCH Open Access Psychopathology of addiction: May a SCL‑90‑based five dimensions structure be applied irrespectively of the involved drug? Pier Paolo Pani1, Angelo G. I. Maremmani9,10, Emanuela Trogu4, Federica Vigna‑Taglianti2,3, Federica Mathis2, Roberto Diecidue2, Ursula Kirchmayer5, Laura Amato5, Joli Ghibaudi6, Antonella Camposeragna6, Alessio Saponaro7, Marina Davoli5, Fabrizio Faggiano8 and Icro Maremmani9,10,11*

Abstract Background: We previously found a five cluster of psychological symptoms in heroin use disorder (HUD) patients: ‘worthlessness-being trapped’, ‘somatic-symptoms’, ‘sensitivity-psychoticism’, ‘panic-anxiety’, and ‘violence-suicide’. We demonstrated that this aggregation is independent of the chosen treatment, of intoxication status and of the pres‑ ence of psychiatric problems. Methods: 2314 Subjects, with alcohol, heroin or cocaine dependence were assigned to one of the five clusters. Dif‑ ferences between patients dependent on alcohol, heroin and cocaine in the frequency of the five clusters and in their severity were analysed. The association between the secondary abuse of alcohol and cocaine and the five clusters was also considered in the subsample of HUD patients. Results: We confirmed a positive association of the ‘somatic symptoms’ dimension with the condition of heroin ver‑ sus cocaine dependence and of the ‘sensitivity-psychoticism’ dimension with the condition of alcohol versus heroin dependence. ‘Somatic symptoms’ and ‘panic anxiety’ successfully discriminated between patients as being alcohol, heroin or cocaine dependents. Looking at the subsample of heroin dependents, no significant differences were observed. Conclusions: The available evidence coming from our results, taken as a whole, seems to support the extension of the psychopathological structure previously observed in opioid addicts to the population of alcohol and cocaine dependents. Keywords: Psychopathology, Addiction, Heroin, Alcohol, Cocaine, SCL-90

Background craving, dyscontrol, and so on. Other distinctive psy- The relationship between substance use and other psychopathological symptoms of people with addictive chiatric disorders is still an open question. According to behaviours are included in separate disorders pertain- the current nosographic approach, substance use disor- ing to the domain of psychiatric ‘comorbidity’. Several ders are defined by the presence of the core symptoms findings challenge this fragmentary approach to unitary of addiction, such as continuation of use despite con- clinical presentations and call for a revision of the pre- sequences, reduction of other interests and activities, sent nosology. Besides the high degree of association between core symptoms of addiction and symptoms of other psychiatric diseases [1–3], further neurobiological *Correspondence: [email protected] and clinical considerations highlight the strong sharing of 9 Vincent P. Dole Dual Diagnosis Unit, Department of Neurosciences, features between addiction per se and other psychopath- Santa Chiara University Hospital, University of Pisa, Via Roma, 67, 56100 Pisa, Italy ological disorders, especially in the mood, anxiety and Full list of author information is available at the end of the article impulse/control domains [4]. On these bases, a unitary

© 2016 Pani et al. This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/ publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Pani et al. Ann Gen Psychiatry (2016) 15:13 Page 2 of 9

perspective in the approach to patients with addiction, psychopathological dimensions. On the basis of our including psychological and psychiatric vulnerability, the unitary hypothesis, we expect that the use of other sub- effect of substances, the characteristics of addictive pro- stances is likely to have cesses, and their interactions has been proposed [4]. In view of the persistent uncertainty in classifying •• A pathoplastic effect on psychological/psychiat- psychiatric symptomatology as intrinsic to the addic- ric dimensions, related to the specific psychoactive tive disorder or due to psychiatric comorbidity, a low properties of each substance; level of inference on the psychopathology of addicted •• No substantial effect on the five-dimension structure people has been applied recently, so that the symptoms of psychopathology, as we believe this structure is expressed by patients are examined independently of intrinsic to the condition of addiction per se and is a pre-established syndromic level [such as that of the independent of the substance. Diagnostic and Statistical Manual of Mental Disorders (DSM)]. In a previous study that took this approach, Methods we investigated the psychopathological dimensions of Study design and population 1055 heroin addicts entering opioid agonist treatment The VOECT (Evaluation of Therapeutic Community (OAT) [5]. By applying an exploratory principal com- Treatments and Outcomes) cohort study was conducted ponent factor analysis (PCA) to the 90 items on the in 8 Italian regions in 2008–2009, recruiting a total of SCL-90 checklist, a 5-factor solution was identified: 2533 patients admitted to a Therapeutic Community the first factor reflected a depressive ‘worthlessness (TC) treatment for a substance use disorder [9]. To ful- and being trapped’ dimension; the second factor picked fil the aims of the present analysis only baseline data out a “somatic symptoms” dimension; the third identi- were used, implementing a cross-sectional approach. fied a ‘sensitivity-psychoticism’ dimension; the fourth a Moreover, specific inclusion criteria were applied: at ‘panic anxiety’ dimension; and the fifth a ‘violence-sui- least 18 years old, with a diagnosis of heroin, cocaine or cide’ dimension. These same results were recently repli- alcohol dependence based on a clinical judgement, and cated by applying the PCA to another Italian sample of information available from the SCL-90 questionnaire. 1195 heroin addicts entering a Therapeutic Community Reflecting these criteria, the sample consisted of 2314 Treatment [6]. Although this second sample differed subjects. Each participant was included in the sample from the first one in important sociodemographic and once only. Out of 2314 patients, 449 (19.4 %) were diag- clinical factors, treatment setting, and programme char- nosed as dependent on alcohol, 670 (28.9 %) on cocaine, acteristics, the same five psychopathological dimen- and 1195 (51.6 %) on heroin. sions were found, suggesting the existence of specific The study was conducted according to the WMA clusters of psychological/psychiatric features within the Declaration of Helsinki—Ethical Principles for Medical category of opioid addicts. Research Involving Human Subjects. All subjects exam- Further analyses confirmed the clusters of symptoms, ined filled up an informed consensus to participate in this independently of demographic and clinical character- study. Both the consent form and the experimental pro- istics, active heroin use, lifetime psychiatric problems, cedures were approved by the pertinent ethics commit- and kind of treatment received by the patient [6–8]. The tees in accordance with internationally accepted criteria available evidence, taken as a whole, seems to support the for ethical research. trait dependent, rather than the state dependent, nature of the proposed factorial dimensions of the psychopa- Instruments thology of opioid addiction. Self‑report symptom inventory (SCL‑90) Our previous considerations on the nature of the Developed by Derogatis et al. [10], the SCL90 consists of relationship between addiction and other psychiatric 90 items, each rated on a 5-point scale of distress. It is a domains, corroborated by neurobiological, epidemiologi- self-report clinical rating scale oriented to the collection cal, and clinical evidence do not apply to heroin depend- of symptomatic behaviours of psychiatric outpatients. ence only, but also to the other addictive disorders, which Among heroin-dependent patients, the 90 items reflected suggests that the investigation should be extended to the 5 primary symptom dimensions which are believed other addictions. to underlie the large majority of symptom behaviours The primary aim of the present study is to investi- observed in this kind of patient: worthlessness-being gate whether the five-factor psychic structure found in trapped, somatic symptoms, sensitivity-psychoticism, opioid dependents also applies to cocaine- or alcohol- panic-anxiety, and violence-suicide [5]. dependent patients. A secondary aim is to evaluate the The worthlessness-being trapped dimension reflects impact of the use of other substances on the five cited a broad range of the symptoms typical of the clinical Pani et al. Ann Gen Psychiatry (2016) 15:13 Page 3 of 9

depressive syndrome. This dimension mirrors feelings between patients with primary dependence on opioids of worthlessness and of being trapped or caught. The and secondary use of cocaine, alcohol, or neither of these somatic symptoms dimension reflects distress arising substances. from perceptions of body dysfunctions. The sensitivity- Stepwise multinomial logistic regression analysis was psychoticism dimension focuses on feelings of a full then conducted with the primary substance of abuse continuum of psychotic behaviours. The panic-anxiety (heroin, cocaine, or alcohol) as dependent variable and dimension subsumes a set of symptoms and experiences the psychopathological dominant groups as predictors. usually clinically associated with a high level of manifest In our model, we included sociodemographic and clinical anxiety. The violence-suicide dimension is organized variables that may act as confounding factors (age, gen- around three categories of hostile behaviour: thoughts, der, marital status, detoxification status, living condition). feelings, and actions; it also comprises thoughts of death A second logistic regression investigating cocaine ver- and suicidal ideation. sus alcohol as primary substance of abuse and including In a previous study, these five dimensions were empiri- the same sociodemographic and clinical variables as had cally established and validated [7, 8, 11, 12]. In the pre- been used in the previous model as possible confounders sent study, SCL-90 was administered to the patients was also carried out. within 15 days from the entry into TC. SCL-90 5-factor solution scores were compared on the base of the primary or secondary abused substance at Other instruments univariate (t test) and multivariate (discriminant analysis) Information on the sociodemographic and clinical char- level. Discriminant analysis is useful to statistically distin- acteristics of the patients included in the study was col- guish between two or more groups of cases, and is also a lected through a research questionnaire administered powerful classification technique. By classifying the cases at the time of entering TC. The secondary substance of used to derive the functions, and comparing predicted abuse (one or more) was self-reported by the patient. group membership with actual group membership, one can empirically measure the success in discrimination by Data analysis observing the proportion of correct classifications. The Two analysis sets were obtained. The principal analy- purpose is to see how effective the discriminating vari- sis was run on the full sample of subjects entering TC ables are. If a large proportion of misclassification occurs, treatment for opioid, cocaine, or alcohol dependence then the selected variables are poor discriminators. We (n = 2314); this sample was divided into three groups used the stepwise procedure to select the best discrimi- according to the primary substance of abuse. The sec- nating variables. ondary analysis was performed on subjects entering a TC All analyses were carried out using SPSS v. 4.0 (SPSS, treatment for opioid dependence. Only opioid addicts Chicago, IL, USA). Statistical significance was set at the reporting cocaine, alcohol, or neither of these as second- p = 0.05 level. ary substances of abuse were included (n = 947), while those reporting both as secondary substances of abuse Results were excluded. Patients were compared as regards demo- Sociodemographic and addiction characteristics graphic and clinical variables by means of the Chi-square In the overall sample (n = 2314), mean age was test (with Bonferroni’s correction) for categorical vari- 35.2 ± 8.5 years, 84.2 % of the subjects were male, 87.9 % ables, and Student’s t test for continuous variables. were single, separated, divorced, or a widow(er), 20.5 % SCL-90 factors were standardized into z-scores in order had a high educational level (with duration over 8 years), to make scores comparable. Each subject was assigned 79.1 % of subjects were unemployed, and 26.5 % lived to one of the five subtypes on the basis of the highest alone. z-score achieved (named “dominant SCL-90 factor”). When looking at differences between alcohol, heroin This procedure gives the opportunity to classify subjects and cocaine patients, mean age was 41.4 ± 8.8 among on the basis of the highest symptomatological cluster, alcohol-dependent, 33.8 ± 7.7 among heroin-depend- overcoming the problem of identifying a cut-off point for ent, and 33.6 ± 7.7 among cocaine-dependent patients. the inclusion of patients in the clusters. The subtypes are This difference was statistically significant (F = 168.87; clearly distinct, as shown by analysing the mean z-scores p = 0.000). As regards gender, the frequency of males was and 95 % confidence interval (CI 95 %) across the factors 85.2 % among patients with alcohol dependence, 81.4 % for each dominant group [5]. Psychopathological symp- among those with heroin dependence, and 88.7 % among toms were then compared by means of the Chi-square those with cocaine dependence (Chi-square = 17.30; test, with Bonferroni’s correction, between patients with p = 0.000). The frequency of single/separated/divorced/ primary dependence on alcohol, heroin, or cocaine; and widow(er) subjects was 87.5, 90.2 and 84.1 % among Pani et al. Ann Gen Psychiatry (2016) 15:13 Page 4 of 9

alcohol-, heroin-, and cocaine-dependent patients, Table 1 SCL90-based predominant psychopathological respectively (Chi-square 14.88; p 0.001). Educa- symptoms according to the primary substance of abuse = = at treatment entry tional level was high (covering over 8 years) in 20.5 % of the alcohol sample, 21.5 % of the heroin sample, and Dominant Total Alcohol Heroin Cocaine group N 2314 N 449 N 1195 N 670 18.7 % of the cocaine sample, without statistically sig- = = = = nificant differences between groups (Chi-square 2.1; N (%) N (%) N (%) p = 0.366). 77.7 % of patients with alcohol dependence, 1. Worthlessness- 349 (15.1) 62 (13.8)a 180 (15.1)a 107 (16.0)a 79.8 % of those with heroin dependence, and 78.8 % of being trapped those with cocaine dependence were unemployed (Chi- 2. Somatic 509 (22.0) 102 (22.7)a,b 291 (24.4)b 116 (17.3)a symptoms square = 0.95; p = 0.621). 37.5, 24.1, and 23.4 % for the alcohol, heroin, and cocaine group of patients, respec- 3. Sensitivity- 465 (20.1) 110 (24.5)a 218 (18.2)b 137 (20.4)a,b psychoticism tively, were living alone (chi 23.71; p 0.000). = = 4. Panic-anxiety 605 (26.1) 124 (27.6)a 300 (25.1)a 181 (27.0)a Among patients with opioid dependence (n 1195), = 5. Violence-suicide 386 (16.7) 51 (11.4)a 206 (17.2)b 129 (19.3)b mean age was 34.45 7.1 for the group with alcohol as ± Chi square 28.61; df 8; p < 0.001 secondary substance of abuse, 33.64 ± 7.5 for those with cocaine, and 35.61 7.6 for those with neither of these Each letter denotes a subset of categories whose column proportions do not ± differ significantly from each other at the 0.05 level substances as secondary substance of abuse (F = 2.91; p = 0.055). The frequency of males was 78.5 % among patients who used alcohol as secondary substance of abuse, •• ‘Sensitivity-psychoticism’, which was more frequent 82.9 % among those who used cocaine, and 76.5 % among in the group of patients with alcohol than in the those who used neither of the two (Chi-square = 2.83; group with heroin dependence; p = 0.243). The frequency of subjects who were single, sep- •• The ‘violence-suicide’ dimension, which had a higher arated, divorced, or a widow(er), was 88.6, 90.9, and 84.7 %, frequency in the groups of patients with heroin or respectively (Chi-square = 3.53; p = 0.170). Educational cocaine than in the group with alcohol dependence. level was high (covering over 8 years) in 31.6 % of those whose secondary substance of abuse was alcohol, 19.5 % The multinomial logistic regression confirmed a sig- with cocaine, and 28.2 % of the patients using neither as nificant positive association of the ‘somatic-symptoms’ secondary substance (Chi-square = 9.06; p = 0.011). dimension with the heroin versus the cocaine group and 81.0 % of patients whose secondary substance of abuse of the ‘sensitivity-psychoticism’ dimension with the alco- was alcohol, 82.5 % of those using cocaine, and 65.9 % of hol versus the heroin group (Table 2). The further logistic those using neither as secondary substance were unem- regression investigating cocaine versus alcohol as pri- ployed (Chi-square = 13.65; p = 0.001). The percentages mary substance of abuse did not detect any significant for those living alone were 32.9, 24.4, and 19.3 % for those association between the five SCL 90 dominant groups using alcohol, cocaine, or neither of the two as secondary and the primary substance of abuse. substance of abuse, respectively (chi = 4.08; p = 0.130). Table 3 shows differences in psychopathological symptoms between patients with opioid dependence and Psychopathological dimensions secondary abuse of alcohol, or else cocaine or neither Table 1 shows differences in psychopathological symp- of these drugs. The most frequent psychopathological toms between patients with addiction to alcohol, her- dimension was ‘panic-anxiety’ for all the three groups oin, or cocaine as primary substance of abuse. The most of patients The less frequent were ‘violence-suicide’ and frequent psychopathological dimension was ‘panic- ‘sensitivity-psychoticism’ for the group of patients with anxiety’ for all the three groups of patients. The least alcohol as secondary abused drug, ‘violence-suicide’ for frequent psychopathological dimensions were ‘violence- the group of patients with no alcohol or cocaine as sec- suicide’ for the group of patients with alcohol depend- ondary drug, and ‘worthlessness-being trapped’ for the ence (11.4 %) and ‘worthlessness-being trapped’ for the group of patients with cocaine as secondary abused drug. patients with heroin or cocaine dependence (15.1 and No statistically significant differences between the three 16.0 %, respectively). groups were observed in any of the five SCL-based psy- The only statistically significant differences observed chopathological dimensions. between the three groups were in the frequency of With reference to psychopathological severity, Table 4 shows the results of the univariate and multivariate (step- •• The ‘somatic symptoms’ dimension, which was better wise discriminant) analysis comparing patients with represented in the group of patients with heroin than dependence on alcohol, heroin, or cocaine. All SCL-90 in the group with cocaine dependence; dimensions, with the exception of ‘violence-suicide’, were Pani et al. Ann Gen Psychiatry (2016) 15:13 Page 5 of 9

Table 2 Risk factors for alcohol or cocaine versus heroin •• ‘Somatic symptoms’, where patients with alcohol dependence at treatment entry, according to logistic or heroin dependence showed higher scores than regression (only significant results are reported), 2314 patients with cocaine dependence; patients •• ‘Panic-anxiety, where patients with alcohol depend- Primary substance of abusea B Odds 95 % CI p ence showed higher scores than patients with cocaine ratio or heroin dependence. Alcohol Age 0.11 1.11 1.09–1.13 0.000 The stepwise analysis identified the severity of ‘somatic Living alone 0.32 1.38 1.06–1.80 0.017 symptoms’ and ‘panic anxiety’ as the factors best discrim- Detoxified status 0.61 1.84 1.44–2.35 0.000 inating between the three groups of patients. However, SCL90 dominant groups according to this analysis, the percentage of the origi- Prominent sensitivity- 0.43 1.54 1.02–2.32 0.039 nally grouped cases correctly classified is only 35.7 %. psychoticismb The other SCL-90-based psychopathological dimensions Cocaine were unable to improve the value of the discrimination. Male gender 0.53 1.70 1.27–2.27 0.000 Looking at the 5-factor solution SCL-90 scores, Table 5 Single marital status 0.53 0.59 0.43–0.80 0.001 − shows univariate and multivariate (stepwise discrimi- Detoxified status 0.45 1.57 1.29–1.92 0.000 nant) analysis for the sample of patients with opioid SCL90 dominant groups dependence and secondary abuse of alcohol, cocaine, Prominent somatic 0.39 0.68 0.48–0.95 0.023 or neither of these drugs. All the SCL-90 dimensions b − symptoms were higher in patients who were co-abusing alcohol Statistics: Chi-square 375.88; df 20; p < 001 or cocaine compared with those not adding these other This table includes SCL-90 factors as exposure variables, other demographic drugs. In any case, none of the differences observed and clinical variables as possible confounders, and primary substance of abuse (alcohol or cocaine vs. heroin) as a dependent variable reached statistical significance and none of the five psy- a Considering heroin as reference primary substance of abuse chopathological dimensions showed a discriminant b Considering prominent worthlessness-being trapped symptoms as reference value. psychopathology Discussion

Table 3 SCL90-based predominant psychopathologi- Patients with alcohol, heroin, or cocaine dependence cal symptoms in heroin-dependent patients according differ in most of the demographic characteristics inves- to their secondary substance of abuse tigated in our study. Differences were observed in age, frequency of male gender, marital status, and living con- SCL90-based factors Secondary substance of abuse ditions. Alcohol dependents were older and a higher Alcohol Cocaine Neither alcohol percentage of these subjects lived alone than in the case N 79 N 783 nor cocaine = = N 85 of heroin or cocaine dependents; heroin dependents = were less frequently male and more frequently single N (%) N (%) N (%) than cocaine-dependent patients. In patients with pri- 1. Worthlessness-being 15 (19.0) 113 (14.4) 12 (14.1) mary opioid dependence, those with cocaine as second- trapped ary substance of abuse had received less education than 2. Somatic symptoms 17 (21.5) 191 (24.4) 18 (21.2) those with alcohol as secondary substance of abuse and 3. Sensitivity- 14 (17.7) 138 (17.6) 20 (23.5) they were more frequently unemployed than those who psychoticism had no secondary substance of abuse. These differences 4. Panic-anxiety 19 (24.1) 199 (25.4) 25 (29.4) are consistent with the sociodemographic characteristics 5. Violence-suicide 14 (17.7) 142 (18.1) 10 (11.8) of the diverse populations of addicts reported in previous Chi-square 5.38; df 8; p 0.715 studies [13–15]. = = = Post-hoc test (Bonferroni’s procedure): no differences Looking now at the predominant psychopathological symptoms of the full sample of patients according to higher in the group of patients with alcohol dependence the primary substance of abuse at treatment entry, few than in the group with heroin dependence, and in this significant differences were found. Taking into account second group than in the group of patients with cocaine sociodemographic and clinical confounding factors, dependence. Scores for the ‘violence-suicide’ dimension the only significant difference was found in the ‘somatic were higher in patients with cocaine dependence or her- symptoms’ dimension, which had a higher frequency oin dependence than in those with alcohol dependence. in the heroin versus the cocaine group of patients and Statistically significant differences were found for in the ‘sensitivity-psychotic’ dimension, which was Pani et al. Ann Gen Psychiatry (2016) 15:13 Page 6 of 9

Table 4 SCL90-based psychopathological factor severity in 2314 drug addicts according to the primary substance of abuse at treatment entry Variables Alcohol Heroin Cocaine Univariate dfa N 449 N 1195 N 670 F; df; p = = = M SD M SD M SD ± ± ± 1. Worthlessness-being trapped 1.17 0.7 1.14 0.8 1.11 0.7 0.73; 2; 0.481 0.16 ± ± ± 2. Somatic-symptoms 1.02 0.7a 1.00 0.7a 0.88 0.7b 7.21; 2; 0.001 0.52 ± ± ± 3. Sensitivity-psychoticism 0.82 0.6 0.77 0.6 0.76 0.6 1.10; 2; 0.330 0.17 ± ± ± 4. Panic-anxiety 0.47 0.5a 0.43 0.6a,b 0.38 0.5b 3.72; 2; 0.024 0.38 ± ± ± 5. Violence-suicide 0.75 0.6 0.79 0.7 0.80 0.7 0.85; 2; 0.425 0.16 ± ± ± − Centroids 0.191 0.047 0.211 − Wilks Lambda 0.97; Chi-square 55.70; df 10; p < 0.001 Correctly classified: 35.7 % Each letter denotes a subset of categories whose column proportions do not differ significantly from each other at the 0.05 level a Discriminant function (structured matrix)

Table 5 SCL90-based psychopathological factor severity in heroin use disorder according to the secondary substance of abuse SCL90-based factors Heroin alcohol Heroin cocaine Heroin (neither F p N 79 + N 783+ alcohol nor cocaine) = = N 85 = M SD M SD M SD ± ± ± 1. Worthlessness-being trapped 1.13 0.7 1.15 0.8 1.02 0.8 1.01 0.362 ± ± ± 2. Somatic symptoms 0.99 0.7 1.02 0.7 0.91 0.7 0.81 0.444 ± ± ± 3. Sensitivity-psychoticism 0.70 0.5 0.81 0.7 0.66 0.6 2.28 0.102 ± ± ± 4. Panic-anxiety 0.41 0.5 0.45 0.6 0.34 0.5 1.30 0.271 ± ± ± 5. Violence-suicide 0.69 0.5 0.83 0.7 0.64 0.6 3.72 0.024 ± ± ± Post-hoc test (Scheffe’s procedure): no differences between groups in any psychopathological factor Discriminant analysis FD1 Wilks lambda 0.98; Chi-square 13.28; df 10, p 0.208 = = = FD2 Wilks lambda 0.99; Chi-square 1.26; df 4, p 0.867 = = = more frequent in the alcohol versus the heroin group of As regards psychiatric severity, considering the full patients. The higher allocation of heroin-dependent ver- sample, the SCL-90 average scores seem to follow a sus cocaine-dependent patients to the ‘somatic symp- decreasing order in four of the five psychopathologi- toms’ group may be explained by the pre-eminence of cal dimensions, with patients who have primary alcohol the withdrawal symptomatology in the everyday life of dependence showing the highest and those with cocaine heroin-dependent as opposed to cocaine-dependent dependence the lowest severity. However, significant dif- patients [5, 16]. The higher allocation of patients depend- ferences between the three groups were observed for ent on alcohol versus those dependent on heroin to the the ‘somatic symptoms’ and the ‘panic-anxiety’ dimen- ‘sensitivity-psychotic’ dimension may be explained by sions only, with alcohol or heroin dependence associated the pre-eminence of the psychotic component of alcohol with higher scores for ‘somatic symptoms’ than in the withdrawal as well as by the likely association of alco- case of cocaine dependence, and with alcohol depend- hol dependence, when compared with heroin dependence associated with higher scores for the ‘panic-anxiety’ ence, including the presence of overt psychotic disorders dimension than in the case of cocaine dependence. Con- [17–22]. sistently with the foregoing, when discriminant analysis Turning now to the frequency of the five predominant was applied, these two dimensions were the only factors psychopathological dimensions in patients with primary that were able to distinguish between substance-defined opioid dependence, the use of alcohol or cocaine as groups of patients, although the discriminant power of secondary substance of abuse does not lead to any sig- these dimensions was rather weak, with only 35.7 % of nificant difference; thus the use of these additional sub- these cases, as originally grouped, proving to be correctly stances appears to have no impact on these dimensions. classified. Pani et al. Ann Gen Psychiatry (2016) 15:13 Page 7 of 9

The discriminating feature of the ‘somatic symptoms’ by the impact of different substances on specific psy- dimension may be explained by the higher burden of chological/psychiatric dimensions. One of these dimen- physical concerns that accompanies heroin or alcohol sions—somatic symptoms—proved to be the strongest dependence, compared with dependence on cocaine, discriminating factor between the three populations con- especially when considering the importance of with- sidered. Even so, the best discriminating solution, which drawal-related physical symptoms in heroin and alcohol also included the panic-anxiety dimension, was only able dependence. Indeed, the SCL-90 items included in the to correctly classify 35.7 % of cases. This low discrimi- ‘somatic-symptoms’ dimension correspond to a number nating power of the five SCL-90-based dimensions may of complaints (in particular, muscle aches, back pain, hot be attributed to a psychopathological structure of drug flushes and cold shivers, nausea, and disturbed sleep) addiction that is shared by the various forms of depend- which are usually prominent in opioid or alcohol with- ence, so that the structure is largely independent of the drawal [5, 16]. Conversely, in cocaine dependence, with- specific drug used; this result allows to define addiction drawal is frequently associated with mild or no somatic as a unitary condition. symptoms. It appears to be more difficult to explain the higher Limitations severity induced by alcohol dependence than by cocaine In this study, we have found evidence of the similar- dependence with reference to the ‘panic-anxiety’ dimen- ity of the SCL-90-based five-factor psychopathological sion. Both forms of dependence, on alcohol and on structure in the three cases of heroin, cocaine, and alco- cocaine, have been associated with a strikingly high odds hol dependence, but, given the cross-sectional design ratio for the presence of anxiety disorders, and anxi- of the study, we still do not know whether the previous ety has been identified as a determinant of the progres- presence and severity of symptoms included in the five sion from use of substances to dependence [1, 23, 24]. psychopathological dimensions may be a predictor of However, anxiety is a common symptom of the effects or predispose to dependence on specific substances, or of cocaine and of cocaine intoxication, whereas in alco- whether, conversely, the use of, or dependence on these hol, dependence symptoms of the anxiety spectrum substances may condition the appearance of the spe- are expressed instead as a component of withdrawal cific psychopathological picture. Moreover, although where, even in mild forms of alcohol dependence, anxi- we did take into account sociodemographic and clinical ety appears as an essential element. Moreover, we have variables as confounding factors, other primary potential also to take into account that alcohol withdrawal-related interfering factors, such as formal psychiatric diagnoses, anxiety, unlike most of the other physical symptoms were unavailable and could not be taken into account. accompanying withdrawal, may last for months [25, Other limits to the validity of the SCL-90-based psy- 26]. Persistent changes in the GABA and NMDA cir- chopathological 5-dimension solution were commented cuits associated with tolerance and withdrawal may sup- on in our previous studies on the SCL-90 defined struc- port the persistence of anxiety-related symptoms [27, ture of the psychopathology of opioid addiction [5, 8, 11]. 28]. Therefore, the highest scores observed in patients Among these, there is the lack of any observer-related belonging to the alcohol-dependents group, as well as the ‘objective’ evaluation, as SCL-90 is a self-administered contribution of the ‘panic-anxiety’ dimension to the dis- instrument that may be affected by the voluntary or crimination between dependence to different substances, involuntary hiding of some symptoms. may reflect the weight of withdrawal-related anxiety in Lastly, a further limitation is that in this analysis we the everyday life of patients affected by different forms of considered only the SCL-90 questionnaire that was drug addiction. administered at entry into treatment, so our results can Looking at the analyses carried out with heroin- only be considered to be representative of subjects who dependent patients only, no difference was shown on the had an addiction at that particular moment. Some symp- basis of the use of alcohol or cocaine as secondary sub- toms may vary at different stages of the disease, so that stance of abuse, either in the frequency of the five psy- they may prove to be underweighted or overweighted in chopathological dimensions or in the severity of SCL-90 our sample. scores. Considering as a whole the results obtained by our Conclusions study, it seems that when the five-factor SCL-based In our study, the frequencies of the five clusters of cluster of symptoms obtained with a population of her- symptoms resulting from the application of PCA to oin-dependent patients is compared with a different the SCL-90 responses of opioid addicts turn out to dif- population of cocaine or alcohol dependents, only slight fer, in a statistically significant way, only in the case of differences are observed, a result that can be explained the ‘somatic symptoms’ dimension between cocaine Pani et al. Ann Gen Psychiatry (2016) 15:13 Page 8 of 9

and opioid-dependent patients and with the ‘sensitivity- National Epidemiologic Survey on Alcohol and Related Conditions. Arch Gen Psychiatry. 2004;61(8):807–16. psychoticism’ dimension between alcohol- and opioid- 4. Pani PP, Maremmani I, Trogu E, Gessa GL, Ruiz P, Akiskal HS. Delineating dependent subjects. Moreover, the ‘somatic symptoms’ the psychic structure of substance abuse and addictions: should anxiety, and ‘panic-anxiety’ dimensions seem to be the only mood and impulse-control dysregulation be included? J Affect Disord. 2010;122:185–97. psychopathologically discriminant factors between the 5. Maremmani I, Pani PP, Pacini M, Bizzarri JV, Trogu E, Maremmani AGI, three populations. Looking only at the population of opi- Perugi G, Gerra G, Dell’Osso L. Subtyping patients with heroin addiction oid addicts, no differences in the psychic structure were at treatment entry: factors derived from the SCL-90. Ann Gen Psychiatry. 2010;9(1):15. shown on the basis of the co-abuse of alcohol or cocaine. 6. Pani PP, Trogu E, Vigna-Taglianti F, Mathis F, Diecidue R, Kirchmayer U, These findings should be added to those of previous Amato L, Davoli M, Ghibaudi J, Camposeragna A, et al. Psychopatho‑ studies that had shown the considerable stability of the logical symptoms of patients with heroin addiction entering opioid agonist or therapeutic community treatment. Ann Gen Psychiatry. five-factor structure of opioid addiction. At the same 2014;13(1):35. time, our latest results support its extension to the popu- 7. Pani PP, Maremmani AGI, Trogu E, Vigna-Taglianti F, Mathis F, Diecidue R, lation of alcohol and cocaine dependents, where the role Kirchmayer U, Amato L, Davoli M, Ghibaudi J, et al. Psychic structure of opioid addiction: impact of lifetime psychiatric problems on SCL-90- played by the different substances seems to be that of based psychopathologic dimensions in heroin-dependent patients. having a pathoplastic effect on some specific dimensions, Addict Disord Treat. 2015. doi:10.1097/ADT.0000000000000072. without substantially interfering with the proposed five- 8. Pani PP, Maremmani AGI, Trogu E, Vigna-Taglianti F, Mathis F, Diecidue R, Kirchmayer U, Amato L, Davoli M, Ghibaudi J, et al. Psychopathologi‑ dimension structure, since this structure belongs to the cal symptoms in detoxified and non-detoxified heroin-dependent condition of addiction per se and is independent of the patients entering residential treatment. Heroin Addict Relat Clin Probl. various substances considered. 2015;17(2–3):17–24. 9. Mathis F, Vigna-Taglianti F, Decidue R, Kirchmayer U, Piras G, Amato L, Authors’ contributions Ghibaudi J, Camposeragna A, Saponaro A, Faggiano F, et al. Studio “Val‑ PPP coordinated the VOECT study. ET, FVT, FM, RD, JG, AC, and AS coordinated utazione dell’Offerta e dell’Esito dei trattamenti in Comunità Terapeutiche the data collection. FM performed the data management of VOECT dataset. (VOECT), Ministero della Salute, Ricerca finalizzata. Monografia n. 1, PPP, ET, AGIM, and IM drafted the strategy of analysis and the present manu‑ Analisi descrittiva della coorte arruolata. Cagliari: Centro Stampa Regione script. PPP and ET reviewed the literature. IM made statistical analyses. FVT, FM, Sardegna; 2013. RD, AS, UK, LA, JG, AC, MD, and FF critically revised the article. All authors read 10. Derogatis LR, Lipman RS, Rickels K. The Hopkins Symptom Checklist and approved the final manuscript. (HSCL)—a self report symptom inventory. Behav Sci. 1974;19:1–16. 11. Pani PP, Trogu E, Vigna-Taglianti F, Mathis F, Diecidue R, Kirchmayer U, Author details Amato L, Ghibaudi J, Camposeragna A, Saponaro A, et al. Psychopatho‑ 1 Social and Health Services, Cagliari Public Health Trust (ASL Cagliari), Cagliari, logical symptoms of patients with heroin addiction entering opioid Italy. 2 Piedmont Centre for Drug Addiction Epidemiology, ASLTO3, Grugliasco, agonist or therapeutic community treatment. Ann Gen Psychiatry. Province of Turin, Italy. 3 Department of Clinical and Biological Sciences, San 2014;13:35. Luigi Gonzaga University, Turin, Regione Gonzole 10, 10043 Orbassano, Prov‑ 12. Maremmani AGI, Rovai L, Maremmani I. Heroin addicts’ psychopathologi‑ ince of Turin, Italy. 4 Department of Psychiatry, Cagliari Public Health Trust (ASL cal subtypes. Correlations with the natural history of illness. Heroin Addict Cagliari), Cagliari, Italy. 5 Department of Epidemiology, Latium Regional Health Relat Clin Probl. 2012;14(1):11–22. Service, Rome, Italy. 6 National Coordination Hospitality Communities (CNCA), 13. Maremmani I, Pani PP, Mellini A, Pacini M, Marini G, Lovrecic M, Perugi Rome, Italy. 7 Regional Epidemiological Observatory, Emilia Romagna Regional G, Shinderman M. Alcohol and cocaine use and abuse among opioid Health Service, Bologna, Italy. 8 Department of Translational Medicine, Avoga‑ addicts engaged in a methadone maintenance treatment program. J dro University, Novara, Italy. 9 Vincent P. Dole Dual Diagnosis Unit, Department Addict Dis. 2007;26(1):61–70. of Neurosciences, Santa Chiara University Hospital, University of Pisa, Via 14. Lopez-Goni JJ, Fernandez-Montalvo J, Arteaga A. Differences between Roma, 67, 56100 Pisa, Italy. 10 Association for the Application of Neuroscientific alcoholics and cocaine addicts seeking treatment. Span J Psychol. Knowledge to Social Aims (AU-CNS), Pietrasanta, Lucca, Italy. 11 G. De Lisio 2015;18:E2. Institute of Behavioural Sciences, Pisa, Italy. 15. Swendsen J, Conway KP, Degenhardt L, Dierker L, Glantz M, Jin R, Merikangas KR, Sampson N, Kessler RC. Socio-demographic risk factors Competing interests for alcohol and drug dependence: the 10-year follow-up of the national The authors declare that they have no competing interests. IM served as Board comorbidity survey. Addiction. 2009;104(8):1346–55. Member for Indivior, Molteni, Mundipharma, D&A Pharma, and Lundbeck. 16. Knapp CM, Ciraulo AM, Jaffe JH. Opiates: clinical aspects. In: Lowinson JH, Ruiz P, Millman RB, Langrod JG, editors. Substance abuse: a comprehen‑ Received: 30 October 2015 Accepted: 14 April 2016 sive textbook. Phyladelphia, PA: Williams & Wilkins; 2005. 17. Maremmani AGI, Rovai L, Rugani F, Bacciardi S, Dell’Osso L, Maremmani I. Substance abuse and psychosis: the strange case of opioids. Eur Rev Med Pharmacol Sci. 2014;18:287–302. 18. Smith MJ, Thirthalli J, Abdallah AB, Murray RM, Cottler LB. Prevalence of psychotic symptoms in substance users: a comparison across substances. References Compr Psychiatry. 2009;50(3):245–50. 1. Regier DA, Farmer ME, Rae DS, Locke BZ, Keith SJ, Judd LL, Goodwin FK. 19. Jordaan GP, Nel DG, Hewlett RH, Emsley R. Alcohol-induced psychotic Comorbidity of mental disorders with alcohol and other drug abuse. disorder: a comparative study on the clinical characteristics of patients JAMA. 1990;19(264):2511–8. with alcohol dependence and schizophrenia. J Stud Alcohol Drugs. 2. Kessler RC, McGonagle KA, Zhao S, Nelson CB, Hughes M, Eshleman S, 2009;70(6):870–6. Wittchen H-U, Kendler KS. Lifetime and 12-month prevalence of DSM- 20. Greenberg DM, Lee JW. Psychotic manifestations of alcoholism. Curr IIIR psychiatric disorders in the United States: results from the National Psychiatry Rep. 2001;3(4):314–8. Comorbidity Survey. Arch Gen Psychiatry. 1994;51:8–19. 21. Perala J, Kuoppasalmi K, Pirkola S, Harkanen T, Saarni S, Tuulio-Henriksson 3. Grant BF, Stinson FS, Dawson DA, Chou SP, Dufour MC, Compton W, A, Viertio S, Latvala A, Koskinen S, Lonnqvist J, et al. Alcohol-induced Pickering RP, Kaplan K. Prevalence and co-occurrence of substance use psychotic disorder and delirium in the general population. Br J Psychiatry. disorders and independent mood and anxiety disorders: results from the 2010;197(3):200–6. Pani et al. Ann Gen Psychiatry (2016) 15:13 Page 9 of 9

22. Jordaan GP, Emsley R. Alcohol-induced psychotic disorder: a review. 26. Carlson RW, Kumar NN, Wong-Mckinstry E, Ayyagari S, Puri N, Jack‑ Metab Brain Dis. 2014;29(2):231–43. son FK, Shashikumar S. Alcohol withdrawal syndrome. Crit Care Clin. 23. Regier DA, Rae DS, Narrow WE, Kaelber CT, Schatzberg AF. Prevalence 2012;28(4):549–85. of anxiety disorders and their comorbidity with mood and addictive 27. Heilig M, Egli M, Crabbe JC, Becker HC. Acute withdrawal, protracted disorders. Br J Psychiatry Suppl. 1998;34:24–8. abstinence and negative affect in alcoholism: are they linked? Addict Biol. 24. Lopez-Quintero C, Perez de los Cobos J, Hasin DS, Okuda M, Wang S, 2010;15(2):169–84. Grant BF, Blanco C. Probability and predictors of transition from first use 28. Van Skike CE, Diaz-Granados JL, Matthews DB. Chronic intermittent to dependence on nicotine, alcohol, cannabis, and cocaine: results of ethanol exposure produces persistent anxiety in adolescent and adult the National Epidemiologic Survey on Alcohol and Related Conditions rats. Alcohol Clin Exp Res. 2015;39(2):262–71. (NESARC). Drug Alcohol Depend. 2011;115(1–2):120–30. 25. Roelofs SM. Hyperventilation, anxiety, craving for alcohol: a subacute alcohol withdrawal syndrome. Alcohol. 1985;2(3):501–5.

Submit your next manuscript to BioMed Central and we will help you at every step: • We accept pre-submission inquiries • Our selector tool helps you to ﬁnd the most relevant journal • We provide round the clock customer support • Convenient online submission • Thorough peer review • Inclusion in PubMed and all major indexing services • Maximum visibility for your research

Submit your manuscript at www.biomedcentral.com/submit