A comparison of thought and perception disorders in borderline personality disorder and schizophrenia: psychotic experiences as a reaction to impaired social functioning

According to the first Stern's definition [1], Borderline Personality Disorder (BPD) is a nosographic entity on the border between neurosis and psychosis. This definition received support from several authors who worked on pseudo-neurotic disorders which were too severe to benefit from classical psychoanalytic treatment, or psychotic-like disorders which were insufficiently severe to be considered psychotic disorders [2]-[5]. According to psychodynamic perspectives, patients with BPD are usually capable of reality testing, but may lose this capacity and suffer transient psychotic episodes under severe stress (especially when they are exposed to real or imagined abandonment [6]), during psychotherapy (so called `transference psychosis’) [5],[7],[8], or under the influence of alcohol or other drugs [7],[9]-[12].

This was consistent with both DSM-III (Diagnostic and Statistical Manual of Mental Disorders, 3rd ed.) [13] and DSM-III-R [14], which reported that in BPD transient psychotic symptoms may occur during periods of extreme stress; however, none of the eight criteria for DSM-III/ DSM-III-R BPD referred to psychotic features and instead the classification of concurrent psychotic symptoms was covered by Axis I psychotic disorders. In contrast with DSM-III and DSM-III-R, some authors reported high rates of some peculiar psychotic-like features when assessing DSM-III and DSM-III-R BPD [15]-[20]. Zanarini and colleagues [21] conducted an in-depth evaluation of the cognitive features of BPD comparing three clinical groups (BPD, n = 50; other personality disorders, n = 55; schizophrenia, n = 32) with healthy controls (n = 46). They used the Diagnostic Interview for Borderline Revised (DIB-R) [22] to assess psychopathology over the previous two years and the Structured Clinical Interview for DSM-III (SCID) [23] to evaluate lifetime psychotic episodes. Using the DIB-R, quasi-psychotic thought characterized by transient (lasting less than two days), circumscribed (affecting not more than two areas of life ) or atypical psychotic symptoms (based on reality or totally fantastic), were observed in 20 patients (40%) with BPD, one patient (1.8%) with other personality disorder (PD), and none of the patients with schizophrenia (SC), whereas only patients with SC reported true psychotic thought, defined as prolonged, widespread, bizarre and stereotypical psychotic symptoms, i.e. Schneiderian first-rank symptoms. Using SCID all 32 patients with SC, seven patients (14%) with BPD and two patients (2.6%) with other PD reported one or more lifetime psychotic episodes. In BPD patients these occurred only in the context of either co-morbid substance use disorder or co-morbid affective disorder as previously reported by Pope and colleagues [17]. These findings [21] confirmed the predicted specificity of true psychotic thought for SC and extended previous reports of the prevalence of transient psychotic-like symptoms in BPD, suggesting that quasi-psychotic thought was a highly discriminant sign and therefore “virtually pathognonomic” for BPD.

In DSM-IV [24] the section on BPD was supplemented: a ninth criterion for a categorical diagnosis of BPD, the presence of “transient stress-related paranoid ideation or severe dissociative symptoms” was added; it was stated that the severity and duration of psychotic symptoms should not be such as to warrant an additional diagnosis of psychotic disorder, and the development of psychotic-like symptoms (e.g. hallucinations, body-image distortion, ideas of references, and hypnagogic phenomena) during times of stress was recognized as an associated feature supporting diagnosis of BPD. However no explicit reference to “regression proneness” or the loss of reality testing was included in DSM-IV and this drew a certain amount of criticism [11].

Since the introduction of DSM-IV, some studies have reported high rates of psychotic symptoms in BPD [25]-[30], but few of these studies used a structured interview capable of discriminating between quasi-psychotic thoughts and true psychotic thoughts in the manner of DIB-R. For instance, Yee and colleagues [29] found that 29.2% of 171 patients being treated for DSM-IV BPD reported `hearing voices’ on Symptom-Checklist-90 (SCL-90) [31], but an in-depth evaluation (covering the frequency, duration and type of perception disorder) was only conducted in a non-random subsample of ten cases. On this basis the authors suggested that hallucinatory phenomena were an ongoing, pervasive feature of BPD and were not atypical for this patient group.

Two other studies [27],[28] compared patients with BPD and SC using SCID-II to diagnose DSM-IV BPD and the Psychotic Symptom Rating Scale (PSYRATS) [32] to assess the subjective characteristics of hallucinations and delusions (e.g. frequency, duration, amount of distress etc.). Kingdon and colleagues [27] found that 29% of 33 patients with BPD experienced paranoid delusions compared with 61% of 59 patients with SC and 65% of 19 patients with a diagnosis of both BPD and SC. Only 50% of patients with BPD experienced auditory hallucinations compared with 66% of SC patients and 90% of patients with SC and BPD; there were no significant differences in the frequency, duration or location of hallucinations; however substance use was not taken into account in this study. Slotema and colleagues [28] explicitly excluded patients with alcohol abuse, substance use or current delusions; they recruited a sample of 38 women with BPD, 51 women with SC and 66 healthy women and confirmed that the frequency, duration, and location of auditory verbal hallucinations were similar in patients with BPD and SC.

More recently, Pearse and colleagues [30] found highest lifetime prevalence of psychotic symptoms in DSM-IV BPD (80%), using the Present State Examination (PSE) [33], though patients with affective and substance use disorders were not excluded and psychotic symptoms met DSM-IV SC criteria in 10% of the sample.

In summary, although previous studies suggest so much high frequency of psychotic symptoms in DSM-IV BPD than it has been recently proposed to avoid terms like pseudo-psychotic or quasi-psychotic [34], there is currently no consensus on their prevalence and characteristics (type, frequency, duration, location etc.). Since the introduction of the ninth criterion by the DSM-IV task force - which remained unchanged in DSM-IV-TR [35] and DSM-5 [36] - no study has explicitly estimated the prevalence of quasi-psychotic thought in BPD.

Similarly, there are few papers addressing psychotic reactivity, the crucial aspect of BPD included in the ninth criterion for DSM-IV BPD. Suzuki and colleagues [37] conducted an in-depth evaluation of five patients with BPD and psychotic symptoms that provided support for the psychodynamic perspective [3],[5],[7], which assigns interpersonal functioning the crucial role in the development of both delusion and hallucinations. These authors also distinguished between different types of symptoms: they suggested that paranoid delusions were due to the projection of rage and hostile feelings onto other people, whereas hallucinations tended to occur as a result of avoidance of an interpersonal relationship, and the presence of both delusions and hallucinations was a consequence of interpersonal problems in patients with a passive attitude.

More recently, Glaser and colleagues [38] explicitly investigated stress-related psychotic responses in a sample of 65 patients with BPD and, although they did not explicitly refer to the psychodynamic framework, they reported that enhanced psychotic reactivity to the stresses of daily life was more common in BPD than in psychotic disorders or Cluster C personality disorders. They also found a wide range of stress-related symptoms, which they termed reality disturbances, that seem to fit better with the broader definition included in the DSM-IV-TR description of BPD (psychotic-like symptoms such as hallucinations, body-image distortion, ideas of references, hypnagogic phenomena) than with the narrowly defined ninth diagnostic criterion for DSM-IV-TR BPD (i.e. paranoid and dissociative ideation).

In order to assess whether the transition from DSM-III to DSM-IV led to changes in reports of thought and perception disorders in BPD and whether this psychopathology is a reaction to interpersonal functioning, the aims of the present study were (1) to describe perception and thought disorders in a group of outpatients with DSM-IV BPD, comparing their symptoms with those of outpatients with DSM-IV SC; and (2) to explore relationship between these symptoms and function searching for possible correlations between perception and thought disorders and personal and social functioning in both groups and subsequently estimating the extent to which level of personal and social functioning predicts perception and thought disorders in both groups.

A total of 64 patients gave their informed consent and fulfilled the eligibility criteria; eight (12.5%) patients were excluded because they had a current DSM-IV substance use disorder (n = 4, 6.2%) or a current DSM-IV mood disorder (n = 2, 3.1%) or a diagnosis of both BPD and SC (n = 2, 3.1%).

Following screening with SCID-I and SCID-II 28 (50%) patients were assigned to the BPD group and 28 patients (50%) were assigned to the SC group. All patients in the SC group were taking antipsychotic drugs in accordance with evidence-based recommendations provided by American Psychiatric Association Practice Guidelines [41], but none of the BPD patients were taking antipsychotic drugs.

There were no significant group differences in gender, marital status, education or employment (Table 1). However, the mean age of the SC group was significantly higher than that of the BPD group.

As expected, patients with BPD showed less severe symptoms than those with SC, as indicated by higher scores on both DIB-R cognitive section and DIB-R psychotic experience subsection.

The overall cognitive pattern shown by BPD patients in our investigation seems to be roughly consistent with that reported by previous authors who used the same tool to assess DSM-III BPD [17],[20],[21]: the BPD group presented (a) ubiquitous odd thinking, unusual perceptual experiences and non-delusional paranoia; (b) frequent transient, circumscribed and atypical psychotic experiences (i.e. quasi-psychotic thought), and (c) some ongoing, bizarre, Schneiderian first-rank symptoms (i.e. true psychotic thought).

Zanarini and colleagues [21] suggested that quasi-psychotic thought was “virtually pathognonomic” for BPD because of its high prevalence in their BPD group (40%) and its absence from their SC group (0%) whereas in the present study this type of thought, although more frequent in the BPD group (82.1%), was also encountered in around half the SC group and therefore does not seem to discriminate between BPD and SC. The high prevalence of quasi-psychotic thought in our SC group compared to that of Zanarini and colleagues [21] might be explained by the difference in setting (outpatients only vs. outpatients and community inpatients) or by differences in psychopharmacological treatments (e.g. first- vs. second- generation antipsychotics, oral vs. long- acting injectable agents etc.), specifically antipsychotic agents might have controlled and altered the clinical expression of some psychotic episodes in our SC patients.

The prevalence of true psychotic thought in DSM-III BPD reported in previous studies which used DIB-R ranged from 0% [21] to 3.8% [46], whereas in the present study almost half the DSM-IV BPD group (46.4%) reported at least one true psychotic experience during the previous two years. This finding seems to be consistent with some recent studies of psychopathology in DSM-IV BPD, which reported a noteworthy prevalence of persistent psychotic symptoms, even if a careful comparison is complicated by differences in samples and methods. For instance, Kingdom and colleagues [27] reported that 29% of BPD patients had paranoid delusions according to SCID-I and 50% had auditory hallucinations according to PSYRATS, but this study did not exclude or take into account substance use disorder; they failed to find significant differences in the persistence of hallucinations in BPD and SC. Yee and colleagues [29] reported that 29.2% of 171 patients with BPD assessed by SCL-90 heard voices, but their qualitative assessment of perception disorders was carried out son a non-random subsample of ten cases. Pearse and colleagues [30] found highest lifetime prevalence of psychotic symptoms in DSM-IV BPD (80%), but 75% of the sample had a history of alcohol and/or illicit drug use, and affective disorders were not excluded.

The prevalence of quasi-psychotic thought in our sample of DSM-IV BPD was also higher than that found by Zanarini and colleagues [21] in a DSM-III BPD sample using the same assessment tool (82.1% vs. 40%).

Taken together, these findings suggest that changes in the DSM definition of BPD (i.e. the introduction of the ninth diagnostic criterion and the more complete description of BPD) may have increased its sensitivity to psychotic and psychotic-like symptoms leading to the inclusion of patients who have had psychotic episodes that were not severe enough to meet criteria for either brief psychotic disorder or another Axis I psychotic disorder (e.g. the criteria proposed by DSM-5 for Attenuated Psychosis Syndrome). Given that in the present study the mean psychotic experiences score for the BPD group was one third of the score for the SC group, we suggest that a severity scale for psychotic symptoms, such as the Clinician-Rated Dimensions of Psychosis Symptom Severity introduced in DSM-5, might be used to improve the assessment of psychotic symptoms in BPD as it would allow a more careful distinction between the psychotic features of BPD and SC. However, placing this dimensional scale in DSM-5 Section III is unlikely to be the most effective way to encourage its use, as Maj [47] has already noted.

Overall functioning was similar in the two groups; however aggressive acts and socially disturbing behavior were more frequent in the BPD group than the SC group, even considering that antipsychotic therapy may have had a crucial role in decreasing violent behavior in SC patients [48].

Our most important finding about the relationship between thought disorders and functioning concerns non-delusional paranoia. Not only were symptoms of non-delusional paranoia more severe and more frequent in the BPD group than SC group, consistent with the results of Zanarini and colleagues [21], but their severity was correlated with and predicted by the level of personal and social functioning in the BPD group, but not the SC group. These results are consistent with the most influential psychodynamic perspectives [7],[9],[10],[12] in suggesting that severe difficulties in interpersonal functioning induce the development of undue suspiciousness, ideas of references and other symptoms of non-delusional paranoia in BPD patients.

Despite the differences in methods, our results appear to be consistent with those of Suzuki and colleagues [37] and Glaser and colleagues [38] and support the concept of stress-related thought disorder. However in the present study there was neither a correlation between non-delusional paranoia and psychotic experiences nor any functioning areas worked as predictor for psychotic experience, so low interpersonal functioning did not appear to induce psychotic experiences in BPD patients and thus, unlike Glaser and colleagues [38], we found no evidence in our BPD group of either true psychotic reactivity or quasi-psychotic reactivity, only stress-related non-delusional paranoia.

Our findings are therefore consistent with the reference to “stress-related paranoid ideation” included in the ninth diagnostic criterion for DSM-IV BPD and maintained in DSM-5, and suggest that it is still a valid and essential discriminating feature for BPD.

The lack of a prospective longitudinal assessment of functioning and psychopathology is an important limitation of the present study; this made it impossible to infer a true causal relationship between the interpersonal functioning and non-delusional paranoia; we can only suggest that level of interpersonal functioning is associated with the severity of non-delusional paranoia over the last two years, we could not determine whether the development of non-delusional paranoia in BPD patients was due to a worsening of personal or social functioning.

In view of the strong association between psychotic disorders and substances use disorder [17],[21], the administration of a urine drug test alongside SCID-I would have enhanced the accuracy of sample selection and thus the characterization of thought and perception disorders in the two diagnostic groups.

Furthermore, the limited sample size and the lack of information about non-antipsychotic pharmacological treatment assumed by patients should be considered as two other limitations of the present study since they both could affect the accuracy of results.

Future research on BPD might investigate the relationship between personal and social functioning, non-delusional paranoia, and thought and perception disorders prospectively in order to determine whether non-delusional paranoia is due to changes in interpersonal functioning and whether and how it may evolve into quasi- or true psychotic experience.

BPD patients reported less severe psychotic experiences with less frequent true psychotic thought, more frequent quasi-psychotic thought, and more severe non-delusional paranoia than SC patients. Interpersonal functioning seems to predict non-delusional paranoia in BPD, which would validate the “stress-related paranoid ideation”, included in the ninth diagnostic criterion for DSM-IV and DSM-5 BPD. PBD patients had higher scores on the psychotic experiences subscale that support the use of a dimensional assessment of the severity of thought and perception disorders, for example the Clinician-Rated Dimensions of Psychosis Symptom Severity introduced in DSM-5, Section III.

These results are consistent with the most influential psychodynamic perspectives in suggesting that severe difficulties in interpersonal functioning induce the development of undue suspiciousness, ideas of references and other symptoms of non-delusional paranoia in BPD patients.