Subtyping patients with heroin addiction at treatment entry: factor derived from the Self-Report Symptom Inventory (SCL-90)

Inclusion criteria comprised a diagnosis of heroin addiction according to DSM-IV criteria and duration of illness of at least 1 year.

The sample consisted of 1,055 subjects, evaluated at their treatment entry. Data came from the Pisa addiction dataset: a database including anonymous individual information originally collected for clinical or other research purposes. Mean age was 30 ± 7 years (range 16 to 59), 884 (83.8%) were male, 133 (12.6%) had a low educational level (less than 8 years), 691 (65.5%) had never been married, 483 (45.8%) were unemployed and 25 (2.4%) were unable to work due to health impairment. Among those employed, 295 (28.0%) had a 'white collar' job and 276 (26.2%) a 'blue collar' job. Mean duration of addiction was 7.20 ± 6.0 years. A total of 502 (47.6%) had been addicted for less than 5 years, 272 (25.8%) between 5 and 10 years, 152 (14.4%) between 10 and 15 years, 100 (9.5%) between 15 and 20 years, 29 (2.7%) between 20 and 28 years. All these patients were Italians, and were only included once in the sample. In all, 170 (16.1%) began treatment for the first time.

Developed by Derogatis and colleagues [13], the SCL-90 is made up of 90 items, each rated on a 5-point scale of distress. These items are clustered in nine dimensions. 'Somatisation' reflects distress arising from perceptions of bodily dysfunction. Complaints focused on cardiovascular, gastrointestinal, respiratory and other systems with strong autonomic mediation have been included. Headaches, backaches, and pain and discomfort localised in the gross musculature are additional components, as are other somatic equivalents of anxiety. 'Obsessive-Compulsive' reflects behaviours that are closely identified with the clinical syndrome of the same name. The focus of this criterion is on thoughts, impulses and actions that are experienced as unremitting and irresistible by the individual but are of an ego-alien or unwanted nature. Behaviours indicative of a more general cognitive difficulty (for example, 'mind going blank', 'trouble remembering') also load on this dimension. 'Interpersonal Sensitivity' focuses on feelings of personal inadequacy and inferiority, particularly by comparison with other individuals. Self-deprecation, feelings of uneasiness, and marked discomfort during interpersonal interactions are characteristics of people showing high levels for this dimension. Feelings of self-consciousness and negative expectations regarding interpersonal communications are further typical sources of distress. 'Depression' reflects a broad range of the concomitants of the clinical depressive syndrome. Symptoms of dysphoric affect and mood are represented, as are signs of withdrawal of interest in life events, lack of motivation, and loss of vital energy. This dimension mirrors feelings of hopelessness and futility, as well as other cognitive and somatic correlates of depression. Several of the items included have to do with thoughts of death and suicidal ideation. 'Anxiety' subsumes a set of symptoms and experiences usually associated clinically with a high degree of manifest anxiety. General indicators such as restlessness, nervousness, and tension are included here, as are additional somatic c signs (for example, 'trembling'). Scales measuring free-floating anxiety and panic attacks are an integral aspect of this dimension, and an item on feelings of dissociation is included. 'Hostility' is organised around three categories of hostile behaviour: thoughts, feelings, and actions. Items range from feelings of annoyance and urges to break things, to arguments and uncontrollable temper outbursts. 'Phobic Anxiety' reflects symptoms that have been observed with a high incidence in conditions termed phobic anxiety state or agoraphobia. Fears of a phobic nature oriented towards travel away from home, open spaces, crowds, or public places and means of transport are represented by this parameter. In addition, several scales representing social phobic behaviour have been included. 'Paranoid Ideation' derives from the notion that paranoid behaviour is best considered from a syndromal point of view. Projective ideation of hostility, suspiciousness, centrality, delusions, loss of autonomy, and grandiosity as cardinal paranoid characteristics are assessed within the limitations imposed by a self-report format. 'Psychoticism' represents florid, acute symptomatology, as well as behaviours typically viewed as more oblique, less definitive, indicators of psychotic processes. Four items reflect Schneiderian first rank symptoms of schizophrenia: auditory hallucinations, thought broadcasting, external thought control, and external thought insertion. In addition, secondary signs of psychotic behaviour, as well as indications of a schizoid life style, are represented too. Global scores for SCL-90 items are Total SCL-90 score (sum of all items), the number of items rated positively (PST), and the positive symptom distress index (PSDI), which is calculated by dividing the sum of all items by the score for PST.

An exploratory factor analysis was performed on the 90 SCL items. The ratio of patients/items (11:1) is high enough to authorise this analysis because it is higher than the recommended 10:1 ratio. Factors were extracted by using a principal component analysis (PCA; type 2) and then rotating this orthogonally to achieve a simple structure.

This simplification is equivalent to maximising the variance of the squared loading in each column. To limit the factor number, the criterion used was an eigenvalue >1.5. Items loading with absolute values >0.40 were used to describe the factors. This procedure makes it possible to minimise the crossloadings of items on factors. In order to make factor scores comparable, they were standardised into z scores. All the subjects were assigned to one of five different subtypes on the basis of the highest factor score achieved (dominant SCL-90 factor). This procedure gives the opportunity to classify subjects on the basis of the highest symptomatological cluster. In this way it is possible to solve the problem of identifying a cut-off point for the inclusion of patients in the different clusters identified.

In order to verify how distinct the subtypes are, we analysed the mean z scores and 95% CI across the factors for each dominant group. We also performed a discriminant analysis by utilising the scores of the five factors to predict membership in each dominant group. Lastly, we compared age, sex and duration of dependence between the various dominant SCL-90 factor groups. Continuous variables were compared between groups by means of one-way ANOVA followed by post hoc Student-Newman-Keuls F test or by Kruskal-Wallis test when appropriate, and categorical ones by means of χ² analysis. All statistical analyses were carried out using SPSS v. 4.0 (SPSS, Chicago, IL, USA).

We did not analyse age and gender correlations with SCL-90 before the factor analysis because SCL-90 is a symptom scale and not a psychological test. As a result, the scale response is not affected by age and gender but by the level of severity of psychiatric disorders. Factor analysis is used to summarise the empirical correlations of SCL-90 items into psychopathological dimensions. Therefore, age and gender do not enter into factor analysis. However, exploring the relationship between the psychopathological dimensions derived from factor analysis and age and gender is of clinical interest, because it may provide useful hints about gender-specific and age-specific psychopathological profiles in patients with addiction.

Additionally, SCL-90 global scores were left unanalysed. Our choice was to focus on differential psychopathological profiles, and these tend to be obscured when global scores are used.

Using an exploratory PCA of the 90 items of the SCL-90, a 5-factor solution was identified (Table 1). A total of 77 items with a loading >0.40 were retained. We named factors on the basis of items with the highest loadings. The first factor reflected a depressive 'worthlessness and being trapped' dimension; this accounted for 29.9% of the variance. The second factor, accounting for 4.2% of the variance, picked out a 'somatisation' dimension. The third factor identified a 'sensitivity-psychoticism' dimension; this accounted for 3.0% of the total variance. Panic symptoms loaded on the fourth factor, 'panic anxiety', accounted for 2.15% of the total variance. The last, fifth factor singled out a 'violence-suicide' dimension, which accounted for 2.0% of the total variance. Overall, the five factors accounted for 37.8% of the variance of the items.

On the basis of the highest z scores obtained on the five SCL-90 factors (dominant SCL-90 factor) subjects were assigned to five mutually exclusive groups. The group whose dominant was 'worthlessness and being trapped' comprised 150 subjects (14.2%), the group with 'somatisation' as its dominant gathered 257 subjects (24.4%), the group showing 'sensitivity-psychoticism' as its dominant included 205 subjects (19.4%), the group identified by 'panic anxiety' as its dominant numbered 235 subjects (22.3%), and the group whose dominant was 'violence-suicide' group profiled a cluster of 208 subjects (19.7%). These five groups were sufficiently distinct, and failed to reveal any significant overlap. All these patients showed positive scores in their dominant factors only, alongside negative scores in all the others, the only exception being a small number of patients whose dominant was 'worthlessness and being trapped', who recorded a positive score for the 'sensitivity psychoticism' factor (mean = 0.07; 95% CI = -0.06 to 0.19) This finding was confirmed by the discriminant analysis, which indicated a percentage of correctly classified 'grouped' cases as high as 95.26%.

The main factor (worthlessness, feeling trapped) brings together depressive, obsessive-compulsive and psychotic symptoms. Treatment-seeking addicts who display depressed mood usually report feelings of uselessness and the feeling of being trapped in a corner. These patients feel abandoned, sad, with no goal or interest; they are excessively preoccupied with difficulties, and report feelings of guilt, while experiencing a low or zero sexual drive, too. Obsessive-compulsive symptoms include difficulties in making decisions, completing a task and concentrating, along with worries about one's ineptitude, an 'empty mind' sensation and an incapacity to dominate one's thoughts. Other symptoms, such as the need to check out actions several times or act slowly so as to avoid making mistakes, are not featured. Compulsions and memory impairment do not appear in any factor. Thought disorders consist of feeling alone even when with other people, the thought that one's mind is not working properly, while never feeling really close to others. Lastly, these subjects report a feeling of inferiority, are easily hurt (interpersonal sensitivity), do not like being alone (phobic anxiety) and often feel nervous and upset ('free' anxiety). On the whole, this factor is essentially made up of depressive, obsessive and psychotic features, dominated by feelings of uselessness and of being trapped in a corner.

The second factor (somatisation) is distinguished by a number of somatic and anxious elements, which are usually a feature of opiate withdrawal. The patient complains of muscle aches, back pain, heavy legs and arms, weakness and tiredness, loss of sensitivity and paraesthesia somewhere in the body. Hot flushes and cold shivers are possible too, as well as nausea and stomach ache. Sleep is disturbed and broken up, while getting to sleep is difficult. Patients wake up early at dawn and cannot get back to sleep. They report a sensation of choking, or of being breathless; they may tremble, are aware of their heart beating, or even of chest pain. Appetite is low. Interpersonal sensitivity is heightened, so that they are easily annoyed and irritated.

The third factor features sensitivity and psychoticism. Patients have the impression that others stare at them and speak about them, may do something against them or exploit them with unpredictable consequences. They think they are not respected by their workmates or are disapproved of because of their own views. They get the impression that others do not sympathise with them or approve of their behaviour, or even show explicit hostility towards them. They feel uneasy when they find other people staring at them or simply in speaking with acquaintances, or may even feel threatened when others are there in the same room. They feel uncomfortable in open or crowded spaces, or when doing things in a group (for example, eating). These behaviours may be defined as psychotic as long as the patient is convinced that others control or influence their thoughts, in some cases actually being identified as imposed from outside that individual's mind. Obsessive-compulsive features of a checking type, or taking a lot of time in doing things out of a fear of making mistakes, may also be part of the picture. Lastly, there may be feelings of estrangement and detachment from reality, with the impression that common and familiar things no longer belong to them.

The fourth factor (panic anxiety) can be summed up as agoraphobia, a fear of going around alone, episodes of critical anxiety, fear of travelling by bus, train or subway, sensations of fainting, dizziness or fear of feeling sick or upset in front of other people. Generalised fear is a feature, with the need to avoid certain things, places or activities in order to prevent panicking.

The fifth factor (violence-suicide) includes violent acting outs and features of self-directed aggressiveness. Patients have moments when they cry or throw objects with the aim of breaking them or smashing them into pieces, and suffer from outbursts of rage. They often get into arguments and feel the urge to push, hurt or beat up others. Side by side with all this, they have suicidal thoughts, or longings for death, are upset, excited or restless, and find it hard to stay seated or lie down for any length of time.

The female/male ratio was 1:4.5 for patients in the 'worthlessness and being trapped' group (group 1), 1:6.4 for 'somatisation' (group 2), 1:7.1 for 'sensitivity-psychoticism' ones (group 3), 1:5 for 'panic anxiety' (group 4) and 1:3.7 for 'violence-suicide' (group 5). These differences were not statistically significant (χ² = 6.83 P = not significant).

The presence of a depressive dimension factor in opioid addicts at treatment entry is not surprising. It can be justified by psychological/psychiatric conditions preceding or following substance abuse and dependence. Precursors such as sensation-seeking, impulsiveness, behavioural disinhibition, hyperthymic and cyclothymic temperaments, typically framed in the bipolar mood spectrum, have all been considered predictive of subsequent addictive behaviour [14‐21]: they are all candidates ranking as possible facilitators of substance encounter and escalation to addiction [12, 17, 22, 23]. Moreover, mood alterations can follow substance abuse and dependence. Besides depression, anxiety, and dysphoric mood accompanying opioid, stimulant, and alcohol or sedative withdrawal, the persistence of a depressive state related to repeated substance use has been observed with alcohol and other substances [24‐28] and one hypothesis put forward is that of a reward deficiency syndrome, with anhedonia and a difficulty in deriving pleasure from non-drug related stimuli both prominent [29].

As stated in the Introduction, the association between mood disorders and addiction may involve such a close interaction at neurobiological levels between predisposing factors, addictive processes and addictive consequences, that the attempt to clinically distinguish between addictive-related or independent depression may turn out to be little more than an inconclusive theoretical exercise [11]. The depressive condition experienced by opioid addicts when asking for treatment may originate in a multifactorial interaction which gives rise to the particularities of clinical presentation, marked out by several depressive features, the most prominent of which are feelings of uselessness and of being trapped in a corner.

The second psychological/psychiatric dimension, shown by opioid addicts on entering treatment can be recognised from somatic symptoms. These are consistent with those that are observed within the opiate withdrawal syndrome and are associated with anxiety. Anxiety is again a major feature in the fourth dimension resulting from factorial analysis, in the form of panic anxiety-related symptoms. Anxiety and panic anxiety may be linked with the withdrawal syndrome. The pathophysiology of withdrawal actually overlaps with that of panic disorder, as noradrenergic circuitry around the locus coeruleus is involved in both cases: the cognitive aspect (substance deficiency vs. fear of dying or losing control) usually makes the difference, but most addicts often mistake panic symptoms for withdrawal, however unlikely this may be in given circumstances, or develop the conviction that substance use during withdrawal will prevent them from undergoing potentially dangerous arousal, in the context of a panic-related cognitive conditioning.

However, anxiety is not peculiar to opiate withdrawal, and other determinants of anxiety disorders in addicts cannot be skipped. In fact, a substance-associated nature has been indicated in 20% of panic cases, 25% of social phobias, 40% of obsessive-compulsive disorders, and 50% of agoraphobia [24, 30].

The third psychological/psychiatric dimension shown by opioid addicts on entering treatment is characterised by sensitivity and psychoticism. The relationship between psychotic symptoms and substance use has been widely investigated. The most typical case of psychosis in addicted persons is the appearance of a schizophrenia-like syndrome in chronic stimulant abusers. Psychotic symptoms are reported in 40% of stimulant abusers [31, 32], and in half of the more chronic cocaine users [33, 34]. Such symptoms have been observed even in the absence of psychiatric substance-related vulnerability [35] and can, in fact, be induced experimentally [36‐41]. Psychotic symptoms may also be associated with cannabinoid use, where the symptoms usually produced by cannabis, such as anxiety, depersonalisation, and derealisation [42], increase and are found associated with hallucinations and delusions [43‐46]. Alcohol too has been associated with the appearance of psychotic symptoms both during intoxication and withdrawal [47, 48]. However, the prevalence of psychotic disorders in opioid addicts is low. This may depend on a variety of factors, such as the difficulty chronic psychotics have in going through the regular environmental interactions that are made unavoidable by their need for drug supplies, or on the glossing over that derives from the general antidysphoric effects of opioids and long-acting opioids such as methadone in masking the proneness to psychosis of some patients. This last explanation is consistent with the existence of an opiate withdrawal syndrome [49]. In fact, the majority of psychotic subjects who develop opioid addiction are more likely to be diagnosed as suffering from borderline pictures, intermittent psychotic disorders such as bipolar I, or atypical pictures including substance-induced psychosis. Also, given the high rate of current polyabuse of psychotomimetic drugs, such as cannabis, mild psychotic syndromes may be frequent on psychometric grounds, even when underrated on clinical grounds [42‐46].

Lastly, the fifth psychological/psychiatric dimension shown by opioid addicts on treatment entry is most easily identified through by violent acting outs and features of self-directed aggressiveness. Aggressiveness and self-injurious behaviour are far from being incompatible, and usually run parallel, as both are supported by impulsiveness, often reflecting the severity of opiate intoxication [50]. The form usually taken by impulsiveness in addicts is connected with their extreme proneness to drug-related stimuli [51‐57], but a more general reduction of inhibitory control over impulsiveness in areas of behaviour not directly linked with drug use can be observed. The performance of smokers, alcoholics, cocaine users, and opiate addicts in carrying out behavioural tasks designed to measure impulsiveness, such as the Iowa Gambling Task, Stroop test, and other behavioural inhibition tasks, indicates an increase in the level of impulsiveness [58‐65]. The altered response to these tests may easily depend on an underlying, previously active mental disorder or earlier psychic conditions [66‐72]. Moreover, data consistent with the direct action of drugs in inducing impulsiveness have been reported for cases of nicotine, alcohol, heroin and cocaine use [58, 73‐77].

Subjects with impulsive personality structures and earlier involvement in drug use are those who seem to develop the most severe withdrawal syndromes, suggesting that opiate balance and control over aggressiveness share the same roots. Before the onset of addiction, impulsive subjects display proneness to aggression, but also a disposition towards risk taking, drug use included. In the context of drug use, these subjects show a tendency to move more quickly towards quicker transition to tolerance and regular drug use. Once addiction has developed, the two kinds of damage run parallel and mirror the severity of addiction itself, together with the disruptive behaviour associated with drug seeking. Even in the case of impulsiveness, rage and violence, it is often impossible to disentangle earlier psychological/psychiatric conditions from those that follow the effects of substances, addictive processes, psychiatric consequences and their interactions

Among the sociodemographic variables investigated, the only one that significantly differentiated the groups of patients identified by factor analysis was age: sensitive-psychotic, violent/suicidal and panic addicts proved to be younger. Psychopathological dimensions seem to be unrelated to gender, since the sex ratio does not vary to a significant degree across dimensions. Even the duration of dependence did not differ between dimensions, so that the contribution to the quality of symptoms can be considered similar, and subtypes stand as distinct psychopathological profiles.