Study design and ethical approval
The study design is a 26-week, parallel-group, triple-blind, randomized, placebo-controlled study in patients with a first episode of psychosis, allocated to receive either 2g/qd active NAC or matched placebo as adjunctive to TAU in a specialised early intervention service. Equal numbers of active and placebo groups will be recruited (ratio 1:1). The primary endpoint is end of treatment at 26-weeks; with a post-discontinuation follow-up at 52 weeks.
Institutional ethics and governance approval was obtained from the Melbourne Health Human Research Ethics Committee (HREC 2017.145). The study Sponsor is Orygen, The National Centre of Excellence in Youth Mental Health (referred to as Orygen). The Sponsor will not be involved in study design; collection, analysis, or interpretation of data; writing of the report; and will not participate in the decision to submit the report for publication. A steering committee comprising the Principal Investigator, Coordinating Chief Investigator, Associate Investigators, Project Manager, Study Monitor and other team members will meet on a regular basis to oversee the conduct of the trial.
The trial will be conducted in accordance with Good Clinical Practice (GCP; 44) guidelines and the Declaration of Helsinki (45). The protocol has been developed in line with GCP and the Standard Protocol Items: Recommendations for Interventional Trials (SPIRIT) guidelines (46). A SPIRIT checklist is provided as an Additional File. The study is registered with the Australian and New Zealand Clinical Trials Registry (ACTRN12618000413224). Any amendments to the protocol will result in notifications to relevant parties including ethics, the Sponsor and trial registry.
Study setting
The study will take place at Orygen Youth Health (OYH) the Victorian State Government-funded public youth mental health service for western metropolitan Melbourne. The Early Psychosis Prevention and Intervention Centre (EPPIC), which is part of OYH, was the first Australian specialist early intervention service for FEP, established in 1992. EPPIC is located over two regions; Parkville and Sunshine. In 2002, international guidelines for early intervention were developed, in part, guided by the EPPIC model (47). The core treatment elements include 18 months of: assertive case management; as necessary low dose atypical antipsychotic medications; and evidence-based psychotherapy (47).
All participants will be recruited from the two EPPIC locations. Approximately 250-290 new clients are admitted to EPPIC each year. Previous studies targeting this population at OYH were able to achieve similar recruitment targets, with more stringent eligibility criteria (e.g., 48). The recruitment forecast allows for the recruitment of 1 participant per week (average of 4 to 5 participants per month).
Individuals with psychotic disorders are notoriously difficult to recruit and retain in research, therefore flexibility is essential. The study team have developed skills in engaging and retaining youth with psychotic disorders in research. A team-based approach is utilised, facilitated by tracking notes relating to all participant interactions and action plans in the RPMS, to ensure clear and consistent messages/information is provided to participants.
Eligibility criteria
A participant will be considered eligible for inclusion in this study only if all of the following criteria apply: (i) aged 15 to 25 years (inclusive, at time of enrolment); (ii) within their first three months of admission to EPPIC (defined as having signed consent within three months from date of first face to face outpatient visit with treating EPPIC clinician); (iii) at least two weeks of stable use of their primary medication (e.g., antipsychotic medications); (iv) if female: using effective contraception if sexually active; (v) capacity to consent to the study and comply with study procedures as ascertained by the research interviewer and treating team.
Participants who meet any of the any of following criteria will not be eligible for participation in this study: (i) previous distinct episode of psychosis with inter-episode recovery and recurrence, with the period of recovery lasting more than 6 months; (ii) known or suspected clinically relevant systemic medical disorder, and/or recent gastrointestinal ulcers or renal stones; (iii) females who are pregnant or lactating; (iv) prior sensitivity or allergy to NAC; (v) currently taking >250mg of NAC/day; or >200ug of selenium/day (a one month washout period will be required if individuals who are currently taking these nutraceuticals would like to participate); (vi) inability to comply with either the requirements of informed consent or the treatment protocol; and/or (vii) non-fluency in English.
There are additional exclusion criteria for the MRS component of the study. Participants will be excluded from MRS for the following reasons: (i) serious head injury; (ii) seizures or history of epilepsy; (iii) thyroid or neurological disorder; (iv) claustrophobia or inability to tolerate length of time in the scanner; (v) any implanted metal (including certain types of cerebral clips); (vi) any implanted electronic device such as pacemaker, implantable cardioverter/defibrillator, insulin pump, implanted hearing device and/or neurostimulator.
A participant will be withdrawn if they: (i) cease taking their trial medication for seven consecutive days; (ii) cease effective contraception or become pregnant; and/or (iii) withdraw consent or develop serious adverse events associated with the study drug. Discontinuation due to adverse events could be either at the request of the participant or the discretion of the investigator.
Study medication
Previous studies demonstrate that a dose of 2 g/day of NAC generally appears to be effective and well tolerated (49), and the majority of studies showing efficacy of adjunctive NAC treatment for psychotic and mood symptoms have utilised a daily dose of 2g with no significant difference from placebo in side-effect profile (37, 38, 50-54). While several studies have administered NAC in a 1g dose twice daily (37, 50, 51), NAC will be administered as a once daily dose of 2000mg (2 x 1g capsules to be taken with their usual medication at the same time each day) in the current study in order to improve treatment adherence. Two previous trials by our group utilised a 24-week treatment period (37, 50). A similar treatment period will be used in the current study though the treatment period has been extended to 26 weeks. A 26-week post treatment follow-up will be conducted to determine whether NAC improves outcomes over the longer term. This is especially salient since the trial of NAC in schizophrenia by Brier et al. (55) showed benefits that were most robust at 9 and 12 months.
To ensure blinding, placebo capsules will be carefully matched in appearance, flavour and packaging with the active treatment. The placebo will comprise encapsulated microcrystalline cellulose. As NAC has a distinctive odour, placebo capsules will be lightly dusted with NAC powder to help prevent unblinding.
The nature of the study population is such that participants will already be taking prescribed medications. Typical treatment at EPPIC comprises low-dose SGAs. Other medications may also be prescribed in this population, e.g., mood stabilisers, anticonvulsants, antidepressants, benzodiazepines, benztropine, propranolol and non-benzodiazepine sedatives such as zopiclone and zolpidem, will be allowed. Antipsychotic medication dosing may change during the study period and this data will be captured. Antipsychotic dosing (risperidone equivalents) will be a covariate in the analysis of impact of NAC on the primary outcome.
Study medication will be dispensed at baseline, and follow-up visits at weeks 4, 8, 12 and 18.
Measures
The schedule of assessments and endpoint measures are detailed in Table 1.
Primary outcome
The Positive and Negative Symptom Scale (PANSS) will be used to measure overall symptom severity PANSST, and severity of positive (PANSS+), negative (PANSS-) and general (PANSSG) psychopathology symptoms (56). The PANSS has sound psychometric ratings in terms of both validity and reliability (57). It has been widely used as an outcome measure in clinical trials for psychotic disorders. In the current study, change from baseline to week 26 on the PANSST score is the primary outcome measure.
Secondary outcomes
Secondary outcomes will include symptomatology, functioning and QoL, neurocognition, neuroimaging, substance use and biological correlates.
Symptomatology: The PANSS+, PANSS-, and PANSSG, subscales will be used to assess positive and negative psychotic symptoms, and general psychopathology, respectively. Depressive symptoms will be assessed using the Montgomery Åsberg Depression Rating Scale (MADRS; 58). The Clinical Global Impressions – Severity of Illness (CGI-S; 59) is a brief interviewer-rated scale that measures illness severity. The Clinical Global Impressions – Improvement (CGI-I; 59) is also an interviewer-rated scale and indexes the patient’s improvement or worsening relative to the baseline. The Patient Global Impressions – Improvement (PGI-I; 60), which is commensurate with the CGI-I, will be used as a self-report measure of perceived change since commencing treatment.
Functioning and Quality of Life: The Social and Occupational Functional Assessment Scale (SOFAS; 61) will be used to assess level of social and occupational functioning and is not directly influenced by the overall severity of the individual's psychological symptoms. The Sheehan Disability Scale (SDS; 62) is a brief self-report tool developed to assess functional impairment in three inter-related domains; work/school, social and family life.
The Simple Physical Activity Questionnaire (SIMPAQ; 63) will be used to measure physical activity over a 7 day period around various assessment time pints. The SIMPAQ is a brief 8-item interview covering time spent in bed, sitting or lying down, walking, exercise, sport and other activities.
Social inclusion will be measured using the self-report questionnaire Social Inclusion in Mental Illness – Long Edition (SIMI-LE; 64). The SIMI-LE covers five inter-related domains: (i) housing, neighbourhood and services, (ii) relationships, activities and setbacks, (iii) employment and education, (iv) finances and (v) health and wellbeing.
The Assessment of Quality of Life – 8 dimensions (AQoL-8D; 65) will be used as a measure of health-related QoL. The AQoL-8D was selected to assess units of quality-adjusted life years, enabling the quantification of the economic benefit of treatment.
Neurocognition: Neurocognition will be assessed using a variety of standardised computer-based and pencil and paper tasks. Intellectual functioning will be assessed using the two subtest version Wechsler Abbreviated Scale of Intelligence – Second Edition (WASI-II; 66). The Cogstate computerised assessment battery will include Detection, Identification and n-back (1-back and 2-back) tasks. The Detection test measures processing speed using a simple reaction time paradigm, the Identification test measures attention and processing speed using a choice reaction time paradigm, and the n-back tests measure visual working memory. Cogstate has demonstrated test-retest reliability and responsiveness, small practice effects, no significant floor or ceiling effects, and is an ideal measure to assess changes in neurocognition over time. Pencil and paper tests include: Symbol Digit Modalities Test (SDMT; 67) as a measure of visual attention and processing speed; Digit Span (68), a subtest of the Wechsler Adult Intelligence Scale designed to assess auditory-verbal attention span and working memory; and, the Rey Auditory Verbal Learning Test (RAVLT; 69) for the assessment of verbal learning and memory.
Neuroimaging: Participants will complete a 1-hour magnetic resonance imaging MRI session on a 3T Siemens Magnetom Skyra scanner at The Melbourne Brain Centre, Heidelberg, at 2 visits: (a) baseline, and (b) end week 26. We will acquire structural, spectroscopy (glutathione) and functional resting state, as follows:
High-resolution structural data will be acquired with a T1-weighted magnetisation-prepared rapid acquisition gradient echo (MPRAGE) volumetric sequence.
Brain glutathione concentration will be measured using a single-voxel MEscher-GArwood Point-RESolved Spectroscopy (MEGA-PRESS) spectral editing sequence. A single voxel will be placed in the left anterior cingulate. This region has the advantage of showing glutathione reductions in schizophrenia (70), being a major hub region for brain connectivity, and being relatively easy to shim.
Functional resting state MRI data will be acquired in the axial plane using a T2*-weighted echo-planar imaging sequence. Data will be acquired for approximately 10 mins, during which time the participant will be asked to keep their eyes open and focused on a fixation cross in the middle of the screen, but to allow their minds to wander.
For all sequences except the functional resting state participants will be asked to lie still and relax, and will have the option of listening to music or watching a DVD. Following the scans, a radiologist will review all images to assess whether there is any abnormal pathology. They will complete a report for each scan session and send this to the study’s designated doctor/registrar. In the rare event that further investigation is required, the treating doctor/registrar will contact the participant and discuss the radiological findings and refer on as needed.
Biological correlates: Measurement of inflammatory, oxidative and nitrosative stress markers will provide evidence of the drivers of neuroprogression and clarify the neuroprotective mechanism of NAC (71). Approximately 25 mL of blood will be taken at baseline, at the 26-week treatment endpoint, and at 52-weeks follow-up. Cytokines associated with acute and chronic inflammation will be assessed. Markers of oxidative and nitrosative stress will also be analysed. Other markers may also be investigated if evidence of their involvement in these processes is demonstrated.
Participants will also be invited to provide a further 2.5 mL sample of blood (using the PaxGeneTM tube) for analysis of genetic variation, and to identify the relationship between gene expression and treatment response, or medication side-effects.
Other measures
Cohort characteristics: Demographic data (sex, gender, age, living arrangements, educational level and employment status), medical and physical history, and medications, will be collected at screening with items sourced from Orygen’s harmonized questionnaire battery (72).
Diagnosis: Diagnostic information will be obtained using a number of measures. The Structured Clinical Interview for DSM-5 Research Version (SCID-5-RV; 73) modules pertaining to mood, psychosis, substance use and anxiety will be administered (73). Because of the potential phenomenological overlap and difficulties of differential diagnosis of personality disorders (especially borderline personality disorder, BPD) and FEP (74), all the BPD items of the Structured Clinical Interview for DSM-IV Axis II Disorders Personality Questionnaire (SCID-II PQ; 75) will be administered. The 25-item Personality Inventory for DSM-5 brief form (PID-5-BF; 76) will be administered for a dimensional assessment of personality.
Medication adherence: The 10-item Medication Adherence Rating Scale (MARS, 77) will be used to determine compliance with antipsychotic and trial medications. Compliance with the trial medications will also be determined by using a monthly pill count and can also be delineated from MRS and blood biomarkers.
Substance use: The Alcohol, Smoking and Substance Involvement Screening Test (ASSIST; 78) is a self-report measure for detecting and managing substance use and related problems in primary and general medical care settings. An overview of recent consumption and associated problems is obtained.
Physical health measures: Physical health measurements include weight, height, waist circumference, blood pressure and heart rate, and will be assessed at baseline, 4-weeks, 8-weeks, 12-weeks, 26-weeks and 52-weeks. These measurements will be completed by trained study team members.
Determination of sample size
Previously, moderate effects have been observed (Cohen’s d=0.57) in those who received NAC on global symptom outcomes over a 24 week period (37). We have elected to use a conservative estimate of effect of 0.45 (Cohen’s d) for between group differences on the PANSST score. With power (1-β) set at 0.80, alpha (α) set at .05, and a one-tailed test we would need a total sample size of 124 (62 per group). Allowing for attrition rate of 30%, we calculate that we would need a total sample size of 162 (81 per group).
Assignment of interventions
Randomisation
Participant eligibility will be established before randomisation. Participants will be assigned randomly and consecutively in a 1:1 ratio to one of two interventions, NAC or placebo. The randomisation sequence will be based on a computer-generated permutated block randomisation scheme with stratification for site. The randomisation sequence will be computer-generated by a statistician who is independent of the day-to-day conduct of the trial. The allocation sequence will be concealed within the Research Project Management System (RPMS).
Blinding
This study involves a triple-blind protocol. The investigators, the Sponsor monitor, the study’s biostatistician, research assistants, the clinical treating teams, the laboratory team and the participant will be blinded to whether the participant is receiving NAC or placebo for the duration of the trial. Unblinding will only be permitted in the case of a medical emergency when the appropriate management of the patient necessitates knowledge of the treatment randomisation. All cases of unblinding will be documented.
Enrolment, consent and retention
Informed, written signed consent will be obtained by a trained research assistant from all participants aged 15 to 25 years inclusive. For those individuals less than 18 years of age, written informed consent will also be required from a parent or legal guardian. Optional consent will be sought for the use of de-identified information collected during this study for future research projects, and the storage and use of blood samples for this and associated studies. The Participant Information and Consent Form will be made available by the corresponding author upon request. Once consent is acquired, the initial screening assessment will be conducted to confirm eligibility. Prior to completion of the baseline assessment randomisation will occur. The participant flow through the study is detailed in Figure 1. At the end of the study period, clinical care will continue to be provided to participants either through EPPIC (as determined by the EPPIC eligibility criteria) or the health service to which they have been discharged. Post-trial access to NAC will not be directly provided; however, as NAC is readily available as an over-the-counter supplement, a participant may continue with NAC treatment at the end of the trial treatment period under the guidance of their treating clinician, should they wish to do so.
Data collection, management and analysis
The majority of assessments will be administered to participants in person via semi-structured interview; however, assessments may also be conducted over the phone, or via online video services if required. Flexibility with timing and location of assessments (e.g. home visits, after school hours etc.) will be essential to aid in participant retention. The study team will undergo extensive training prior to completing assessments, and inter-rater reliability between raters will be established. Consent will be sought from participants for audio recordings to be made of research visits to assist in the regular assessment of inter-rater reliability of various measures over the course of the study.
Each case report file (CRF) will be regularly monitored to ensure data quality, and de-identified data will be entered into an electronic CRF embedded within Orygen’s web-based Research Project Management System. The electronic CRF data will be audited to ensure data entry accuracy.
Study protocol, statistical analysis plan, informed consent templates and de-identified participant data will be made available to investigators whose proposed use of the data has been approved by an independent review committee, for up to 3 years following publication.
Study findings will be presented at relevant conferences and published in peer-reviewed journals. A summary of the findings will be provided to participants on request.
Statistical methods
Statistical analyses will follow the International Conference on Harmonization E9 statistical principles. Reporting of results will be done in accordance with Consolidated Standards of Reporting Trials (CONSORT) guidelines (79). All participants who were randomised and who have at least one post-baseline observation will be included in a modified intent-to-treat analysis (80). All attempts will be made to ensure that all participants have follow-up data even if they discontinue from the treatment. The primary efficacy analysis will assess average treatment group differences for the primary outcome measure (PANSST) over the entire study period and will use a likelihood based mixed-effects model, repeated measures approach (MMRM). The MMRM model includes the fixed, categorical effects of treatment, visit, treatment-by-visit interaction, and Region (Parkville and Sunshine) will also be included as a factor in the model. Antipsychotic dosing may be also used as the covariate in the model. The MMRM includes all available data at each time point and is the preferred method of analysing clinical trial data in psychiatry as compared to more traditional repeated measures analysis of variance and analysis of covariance models. Planned comparisons will be done with the MMRM models to determine between group differences in change in symptom measures from baseline to week 26. Differences between treatment groups in terms of secondary outcome measures will be examined using MMRM as per the primary outcome. Correlational analyses as well as multilevel modelling will be used to develop a better understanding of how changes in blood- and brain-derived biomarkers relate to symptoms, functioning, and neurocognition. No interim analyses will be conducted.
Dynamic MEGA-PRESS MRS data will be corrected for frequency drift and edit-off and edit-on spectra combined to produce a glutathione edited spectrum. Edited data will be pre-processed and fitted using the TARQUIN analysis algorithm, with the resonance at 2.9ppm used to determine the level of glutathione. Matching water reference data will also be collected and used to determine the absolute concentration of glutathione. Resting-state functional imaging data will be analysed using the GraphVar toolbox and the Automated Anatomical Labelling atlas for parcellation. To delineate the default mode network (DMN) for each participant, functional connectivity (i.e., Pearson correlation) will be computed between the signal averaged over a spherical region-of-interest positioned at the medial prefrontal cortex and the signal at all other brain voxels. These functional connectivity maps will be r-to-z transformed and carried forward to a second-level statistical analysis to identify significant within- and between-group differences in DMN connectivity. Factorial models similar to the MMRM approach described above will be adopted to test for main effects and interactions using the appropriate combination of F and t-tests. A cluster-based permutation approach will be used to identify significant differences satisfying a family wise error rate of .05. Lastly, betweenness-centrality will be calculated for the medial prefrontal cortex nodes and correlated with glutathione concentration.
Monitoring
A Data Safety and Monitoring Committee (DSMC) will be established prior to commencement of recruitment. The DSMC will include a sponsor representative, clinician, and researcher with expertise in first episode psychosis, and an independent biostatistician. Data monitoring will be conducted by a Clinical Research Associate.
Safety
A record will be made of any adverse event that arises during the trial. An adverse event will be defined as any unfavourable medical change that is accompanied by functional or clinical impairment, which may or may not be related to the study treatment. Any undesirable medical condition occurring from the time of signing consent (even if no study treatment or pharmaceutical product has been administered) will be considered to constitute an adverse event. In our NAC trials with schizophrenia, bipolar, and depression populations, there have been no statistically significant differences between adverse events reported in NAC or placebo groups, although numerically some gastrointestinal side effects have been noted in the NAC groups.