| Summary: | Background: Co-occurring substance use disorders (SUDs) occur in as many as half to three quarters of patients with severe mental disorders such as schizophrenia and bipolar disorder, a phenomenon also described as dual diagnosis. Past research from high income countries have shown that the presence of co-occurring SUDs are associated with a number of demographic and clinical factors, including worse clinical outcomes. Although the clinical profile and negative impact on prognosis has been well characterised in research on dual diagnosis samples from high income countries, little is known in the low- and middle-income country (LAMIC) context about the prevalence and factors associated with SUDs in patients with SMI, its clinical identification and diagnosis, as well as factors associated with its treatment. Objectives: In part-I of the thesis I investigate the clinical epidemiology of dual diagnosis by determining the prevalence of substance use disorders (SUDs) and its association with various clinical and demographic factors in two datasets, the first dataset containing data from a homogeneous sample of Xhosa South African patients with schizophrenia and the second dataset with data from a clinically more heterogeneous sample of patients with major affective and non-affective psychotic disorders. In part-II of the thesis I determine the psychometric properties of two Xhosa language versions of brief substance screening instruments, the Alcohol, Smoking and Substance Involvement Screening Test (WHO-ASSIST version 3) and the Severity of Dependence Scale (SDS), against a gold standard diagnostic instrument, the Structured Clinical Interview for DSM-IV-TR Axis I disorders (SCID-I for DSM-IV). In part-III of the thesis I examine aspects in the pharmacotherapy of dual diagnosis, firstly the association between methamphetamine use and extra-pyramidal medication side-effects (EPS) in patients treated with antipsychotics, and second in a systematic review, the efficacy of risperidone versus other antipsychotics for people with a dual diagnosis. Methods: This thesis consists of secondary analyses of two datasets (datasets#1 and dataset#2) and a Cochrane systematic review of existing randomised trials. Dataset#1 contains data from case-control study with the original aim of investigating the genomics of schizophrenia in the South African Xhosa population. Dataset#2 contains data collected in a cross-sectional manner from 3 studies; one study that investigating the presentation and psychobiology of psychosis, another which was a pilot randomised controlled trial of a text message treatment partner intervention and a third which was a neuroimaging and electrophysiological study of psychotic disorders. Across dataset #1 and #2, demographic information, clinical variables and psychotic and substance use disorder diagnoses were determined using the SCID-I for DSM-IV. In addition, for dataset #2 (conducted in a multi-cultural and ethnic population) we recorded self-identified ethnicity. For self-identified ethnic groups, the terms “Coloured” “Black” and “Caucasian” and “Other” (Asian), were not intended to reify sociocultural constructs but were instead used to study ongoing health disparities. Psychosocial Axis-IV problems were determined using a checklist from the SCID-I and legal involvement was determined using the legal section of the Addiction Severity Index (ASI). Functioning was determined with the global assessment of function, or GAF scale. The ASSIST and SDS instruments were used in a study based on data from dataset #1 investigating the psychometric properties of these screening tools. In turn, I investigated the relationship between methamphetamine use and medication related extra-pyramidal side effects (EPS) in a heterogeneous sample of patients with major affective and non-affective psychotic disorders (from dataset #2). For this study the Simpson Angus Scale for Parkinsonism (SAS), the Barnes Akathisia Rating Scale (BARS) and the Abnormal Involuntary Movement Scale (AIMS) were used to measure EPS. For the first two objectives, logistic regression modelling was used to determine factors associated with having a cooccurring SUD. For the validation study of the Xhosa language versions of the ASSIST and SDS, I determined the internal consistency, concurrent validity, and discriminant validity of these instruments and compared the sensitivity, specificity and the receiver operating characteristics (ROC) for the ASSIST and SDS. Logistic and ROC regression was used for the comparison of ROC area-under-the-curves for ASSIST versus SDS and for determining the impact of co-variates on ROC respectively. In part III, I determined the association between methamphetamine use and the presence of EPS using logistic regression. Finally, I conducted a Cochrane systematic review of randomised trials after a comprehensive literature search of several databases and duplicate study selection and data extraction. Where possible outcomes were pooled, and meta-analyses conducted using random effects models. Primary outcomes were changes in substance use and mental state. Secondary outcomes included substance craving, subjective-wellbeing, adherence, adverse effects, study retention, quality of life and mortality. Study selection, data extraction and quality appraisals were independently conducted. Random-effects meta-analysis was conducted, and the Cochrane risk of bias and GRADE approach used to assess evidence quality. Results: In both datasets#1 and #2, consisting of a total sample size of N=1420 (dataset#1) and N=248 (dataset#2) we found a high prevalence of lifetime SUDs (47.8% and 55.6%) in patients with schizophrenia and major affective or non-affective psychosis, respectively. In multivariable logistic models younger age, male gender, and legal involvement were significantly associated with co-occurring SUDs. Multiple substances were often used together, and SUDs were significantly correlated with one another. Methamphetamine use disorders were significantly more prevalent in the Western Cape province of South Africa and ethnic differences were also apparent with Coloured participants significantly more likely to use methamphetamine compared to Black participants. In addition, we found significant associations between post-traumatic stress symptoms, anxiety symptoms and suicidality and alcohol use disorders. Inpatient status and higher levels of prior admissions were associated with cannabis and methamphetamine use disorders. In a sample of 351 participants from dataset#1 who completed either the ASSIST (N=190), SDS (N=299), or both (n=138), good internal reliability was obtained for both the ASSIST-TSI (total substance involvement score) (Cronbach α= 0.77) and SDS (Cronbach α=0.80). The ASSIST and SDS demonstrated good concurrent validity (rs= 0.50, p 3 years were significantly more likely to have EPS. We found a significant interaction effect between MA use disorders and standardised antipsychotic dose on the occurrence of EPS (ORadj = 1.01, 95% CI= 1.00-1.01, p=0.042). There were no significant associations with EPS with comorbid alcohol, cannabis, or methaqualone use disorders. In the Cochrane systematic review of randomised controlled trials involving risperidone versus other antipsychotics for dual diagnosis, we identified eight trials containing a total of 1073 participants with SMI and co-occurring SUD. Seven trials contributed useable data to the review. Risperidone was compared to clozapine, olanzapine, perphenazine, quetiapine and ziprasidone. Only one study contained data comparing risperidone with a first-generation antipsychotic (perphenazine). Quality of the included studies varied from low to very low. Outcome data were frequently missing and little, or no data was reported in most studies for craving, subjective-wellbeing, metabolic disturbances, global impression of illness severity, quality of life or mortality. For risperidone versus clozapine we found no clear differences between these two antipsychotics in the reduction of positive psychotic symptoms or reduction in cannabis use, improvement in subjective well-being, numbers discontinuing medication, extrapyramidal side-effects or leaving the study early. Clozapine was associated with lower levels of craving for cannabis. For risperidone versus olanzapine we found no clear differences in the reduction of positive psychotic symptoms, reduction in cannabis use, craving for cannabis, parkinsonism, or leaving the study early. For risperidone versus perphenazine, quetiapine and ziprasidone respectively, we found no clear differences in the number of participants leaving the study early. Conclusion: In the context of a dearth of research into dual diagnosis in LAMIC countries such as South Africa, this thesis contains a large-scale epidemiological investigation into substance use comorbidity in patients with schizophrenia and other psychotic disorders. In addition, for the first time we determine the psychometric properties and validity of the Xhosa language ASSIST and SDS and systematically investigate the relationship between methamphetamine use and extra-pyramidal side effects in people with psychotic disorders treated with antipsychotics. For the first time we also conduct a systematic review of the efficacy of risperidone compared to other antipsychotics for a number of primary and secondary outcomes. Our findings have a number of implications for clinicians and services planners. Firstly, SUDs occur in at least half of the people schizophrenia and other psychotic disorders. Accurate detection of SUDs is possible using brief screening tools such as the ASSIST and SDS. Some populations such a younger male population may need particular attention. Moreover, assessment of patients will need to include current and past legal involvement as well as careful assessment of associated comorbid anxiety, post-traumatic symptomology and suicide risk. In patients with co-occurring methamphetamine use disorders clinicians should regularly assess for the development of EPS and carefully titrate antipsychotic dosage from lower to higher doses to avoid EPS. Currently there is not sufficient high-quality evidence favouring the superiority of risperidone over any other antipsychotic in people with SMI and co-occurring SUDs. Results of ongoing trials are awaited, and future trials need to use consistent methodologies and adhere to CONSORT reporting guidelines.
|
|---|
