2.1 Design, Subjects, and Power Analysis
We are planning a double-blind sham stimulation-controlled study with two randomized study groups. Ninety patients meeting the DSM-V [31] criteria for schizophrenia will be enrolled in the study. Based on a recent meta-analysis on rTMS of the frontal cortex for improving negative symptoms [10], we assumed effect size of 0.64. Using this assumption and incorporating the baseline value as a covariate, we calculated that the probability is equal to or greater than 90% (beta = 0.90) to find a significant (alpha = 0.05) difference between study groups in negative symptom improvement with this sample size (SAS PROC GLMPOWER). Patients will be assigned to the active and the sham group in a 1:1 ratio by an adaptive randomization algorithm implemented in R [32]. The algorithm took age, sex, education and negative PANSS score into account. Patient enrolled into the study will receive sequential patient codes, while treatment (active vs. sham) will be assigned to these codes. Patients from the outpatient care of the Department of Psychiatry and Psychotherapy, Semmelweis University will be enrolled who met all the inclusion and none of the exclusion criteria. An enrollment period of 30 months is planned, therefore approximately three subjects need to be included per month to reach target sample size. Based on the patient flow in the outpatient unit we can screen 4–5 subjects and enroll three subjects monthly. Inclusion criteria are (1) diagnosis of schizophrenia or schizoaffective disorder; (2) clinically stabilized on antipsychotic: a stable dose of antipsychotic medication for > 4 weeks; (3) age 18–55 years, and (4) presence of negative symptoms (based on PANSS): a negative subscore ≥ 16 points and [one of items N1–N7 scoring ≥ 4 or two items N1-N7 scoring ≥ 3]. The exclusion criteria are (1) any significant neurological illness; (2) intellectual disability, (3) history of head injury with loss of consciousness for more than 1 hour (4) history of epileptic seizures or epileptic activity on the baseline EEG (evaluated by an expert in clinical EEG and epilepsy), (5) alcohol or drug abuse within the past 3 months, (6) depressive episode or antidepressant treatment in the past 4 weeks. Severe positive symptoms inferfere with cognitive tests, (7) ECT in the medical history, (8) Implanted pacemaker, implanted drug pump, cochlear implant, implanted defibrillator, implanted neurostimulator or any other TMS incompatible implanted metal device, (9) Skin surface is severly injured in the stimulated region of the skull, (10) Sclerosis multiplex, (11) Pregnancy, (12) Severe sleep deprivation, (13) Severe heart failure, (14) Increased intracranial pressure, and (15) Untreated migrain.
Cognitive assessments and PANSS rating by an independent (blind to treatment condition) rater will be performed on all three visits: baseline, visit 2 (day 15) and visit 3 (day 30). A follow-up visit will be scheduled three months after visit 3 to assess PANSS and cognitive test battery (Fig. 1).
2.2 TMS treatment protocol
The TBS sessions will be delivered using the Magstim Rapid2 Plus1 stimulator (Magstim Company, Ltd). The TBS parameters we adopted follow the standard TBS protocols, with 3-pulse 50-Hz bursts given every 200 ms (at 5 Hz) and an intensity of 100% active motor threshold, as measured from the right first dorsal interosseous muscle by a 700-mm figure-of-eight coil [16, 35]. An identical-looking D70 Air Film Sham Coil will be used for sham stimulation. The sham coil also generates a magnetic field, that can be sensed by the participants, but this field does not penetrate the skull. We plan to deliver 1800 stimuli to the vermis and 1800 stimuli to the left DLPFC daily in two 9.5 minutes long blocks (separated by 50 minutes inter stimulation intervals /ISI/) for four weeks (altogether 20x2x1800 = 72000 stimuli) with a 100% motor threshold [17]. We do not intent to apply strategies to improve adherence.
2.3 Procedures to improve the blinding process.
Patients, care providers and raters will be blinded to treatment assignment, only study nurses, who generate the treatment allocation and deliver the pulses will be unblinded. Study nurses do not participate in any other activities in the study such as efficacy assessments. All patients will be instructed that they will be treated with TBS but will be blind to the individual group assignment. The study nurse who will deliver TBS will not take part in any assessments. All efficacy outcome measures will be assessed by blinded study personnel (raters), who will not be permitted access to the treatment sessions. Patients will be instructed not to disclose any details of the treatment session with the raters, and a research assistant will monitor the whole rating period to ensure that the procedure will be blinded. We will question all patients about the group assignment at visit 3 (end of treatment visit).
2.4 Clinical measures
The Positive and Negative Syndrome Scale (PANSS) will be administered to all patients to assess positive, negative, and general symptom severity. Everyday functioning will be measured by the Personal and Social Performance Scale (PSP) [36], while depressive symptoms will be assessed by the Calgary Depression Scale for Schizophrenia [37, 38]. Furthermore, the following demographic data will be collected from all participants: age, gender, education, job status, accommodation, family status, medications, smoking status, illness duration, schizophrenia subtype, and handedness. Since patients will be on antipsychotic medication, the mean Chlorpromazine equivalent dose will be calculated [39].
2.4 Measures of cognition and social cognition
The following tests will be performed to assess cognitive and social cognitive functioning: (1) Reading the Mind in the Eyes Test (RMET) [34]; the (2) 'Faux pas' test [33]; the (3) Wisconsin Card Sorting Test (WCST); the (4) Digit Span Forward and Backward tests and the (5) Karolinska Directed Emotional Face set of facial emotion recogniton .
In order to assess the capacity of mental state discrimination, the Revised Version of the RMET will be used [34]. The RMET presents participants with 36 black-and-white photographs of the eye region of the face, one at a time. Each photo shows the eye region of a different actor or actress. Pictures are of equal size and depict an equal number of male and female faces. Participants will be asked to choose which of four words (one target and three foils), displayed on the screen, best describes the mental state of the actor/actress. Although the RMET seems to be an emotion recognition paradigm, results from functional neuroimaging studies revealed test-related activation in brain areas related to ToM (dorsomedial prefrontal cortex and superior temporal cortex) [40]. This unique feature of the task may be due to direct instructions to attribute mental states, the application of complex social emotions, and only the presence of eye regions. The latter two factors lead to considerable ambiguity of social information that can be solved by active mentalization. A further advantage of the RMET is that numerous schizophrenia studies have applied this measure so far and confirmed that RMET is a reliable tool to detect differences in ToM between patients with schizophrenia and healthy controls [2, 23].
The Faux-pas test assesses the ability to recognition of 'Faux pas': someone mistakenly saying something they shouldn't have [33]. The test is considered an advanced test of Theory of Mind ability as it requires subtle social reasoning: one must be able to appreciate that two protagonists might have different knowledge states and also the emotional impact the statement can have on the listener. It is a well-known instrument used to evaluate theory of mind (ToM) in Autism Spectrum Disorders or Schizophrenia. The test includes 20 short stories containing incidents of faux pas. Each story is read to the individual, who is then asked questions to determine whether or not they recognized the faux pas. Understanding the mental states behind the kinds of actions presented in the Faux Pas task can be broken down into several distinct subtasks to make clear where the respondent is having trouble. The subject gets one point for each question answered correctly. As a result, the proportion of correct answers is calculated, where the maximum score ratio is 1.0. The test indicates a deficit below 0.75.
The WCST is a broadly used tool to measure executive functioning, such as concept formation, set-shifting, and flexibility [41]. In this study, a computerized, 100-card version will be used. Number of persevarative errors is the major outcome variable of the test, its minimum is 0, while it has no theroretical maximum value. Higher values indicate worse outcome. Punishment sensitivity (P) from the reinforcement learning model of WCST will also be used as outcome variable [26]. Its minimum value is 0 and it has no theoretical maximum value. Lower values indicate worse outcome.
The Digit Span Forward and Backward tests require subjects to remember and rearrange short lists of numbers. These tests assess short-term memory span [41]. The outcome variable is the number of items (numbers) the participant can recall, and lower values indicate worse outcome.
The Karolinska Directed Emotional Face set is an emotion recognition task [42], where subjects have to identify emotional expressions from photographs of 8 male and 8 female subjects. The pictures are chosen from Karolinska Directed Emotional Face set. There are 3 photographs from each faces (happy, neutral, and sad). Hit rate is the primary outcome measure, which ranges from 0–100%, and lower values indicate worse outcome.
2.5 Laboratory
Plasma levels of antipsychotic medications and their metabolites will be measured by the Department of Laboratory Medicine (Semmelweis University) as part of routine clinical care.
2.6 Statistical analysis plan
The primary endpoint will be the difference in negative symptom score (sum of PANSS items N1-N7) from baseline. The difference between study groups (active and sham) will be analyzed by a linear mixed model analysis (PROC MIXED in SAS) with the difference (relative to baseline) in negative symptom score as dependent variable and treatment group, time (visit), and treatment-by-visit interaction as predicting (independent) variables, while baseline negative symptom score will serve as covariate [43]. An unstructured covariance matrix will be used to model within-subject effects. If the model fails to converge using the unstructured covariance matrix, the following covariance structures will be modeled in the order given: heterogeneous Toeplitz, heterogeneous compound symmetry, heterogeneous autoregressive(1), Toeplitz, compound symmetry, autoregressive(1), variance components. The first covariance structure that allows for convergence will be selected for the final model. The effects of the same predictor variables on cognitive and social cognitive outcome variables will be analyzed in the same mixed model (separately for all cognitive and social cognitive tests).
Summary statistics for the PANSS negative score (observed and change from baseline) will be presented for all visits from Baseline through visit 3. For change from Baseline values at each post-Baseline visit, LS means, and standard errors (SE), the between-group difference in LS means with the corresponding 95% confidence interval, p-value, and effect size will also be presented. In addition, LS mean ± SE over time for the change from Baseline values by treatment group will also be presented in line plots.
No interim analysis and study audits are planned for this investigation.
2.7 Analysis sets
The following analyses sets will be used:
Randomized Analysis Set: The Randomized Analysis Set will consist of all unique subjects who were randomized.
Safety Analysis Set: The Safety Analysis Set will consist of a subset of subjects in the Randomized Analysis Set who received at least one theta-burst stimulation in one location.
Full Efficacy Analysis Set: The Full Analysis Set will consist of a subset of subjects in the Safety Analysis Set who have both a Baseline value and at least one post-Baseline value for the PANSS negative score.
Subjects will be classified according to the randomized treatment assignment.
Per-protocol Analysis Set: The Per-protocol Analysis Set will consist of a subset of subjects in the Full Efficacy Analysis Set who are at least 80% compliant (received the 80% of the planned stimulation) and do not have any protocol deviations, which is considered to have a substantial impact on primary efficacy outcome. Before the clinical database lock, the precise reasons for excluding subjects from the Per-protocol Analysis Set will be fully defined and documented a priori.
2.8 Patient and Public Involvement statement
Patients or the public were not involved in the design, or conduct, or reporting, or dissemination plans of our research
2.9 Prohibited Concomitant Medications
If any antidepressant medication or a new antipsychotic medication should be given to a subject, the patient should be withdrawn from treatment.
2.10 Data Management
Personal data of participants and data concerning their health and possible illness will be used by the investigators for the administration, conduct, scientific and statistical analysis of the study, taking into account the Data Protection Act, which of course implies that their name will not appear anywhere. They also take responsibility for ensuring that any personal data do not fall into the hands of unauthorized persons. Clinical variables will be analyzed in a reversible anonymized format using codes. Access to the codes is restricted to the investigators and study nurses.
Clinical data will be collected utilizing OpenClinica Community Version: 3.12.2. OpenClinica EDC (hereinafter referred to as Openclinica), an Electronic Data Capture (EDC) Clinical Data Management system developed by Akaza, Inc. for collecting eCRFs. OpenClinica is an open source EDC system that is compliant with 21 CFR Part 11, GCP, and HIPAA. User access to the application will be implemented via secure internet connection. The application interface will be accessible through an URL. The EDC system is hosted by Semmelweis University, Department of Psychiatry and Psychotherapy. All users are required to complete EDC training related to their project role. After training each user will sign and return the meeting training record to document the completion of their training and OpenClinica access will subsequently be granted via e-mail.
This is a single center investigator-initiated study, therefore no data monitoring committee is needed.
2.11 Withdrawal of Patient from Study
Patients may withdraw from the study at any time and for any reason without prejudice to their future medical care by the investigator or at the study site. Every effort should be made to keep patients in the study. The reasons for patients not completing treatment and/or the reasons for patients not completing the study will be recorded. A patient may be withdrawn from the study for any of the following reasons:
- Noncompliance with the protocol or significant protocol violation.
- A serious or intolerable AE(s)
- Lost to follow-up.
- The patient withdraws consent.
When a patient withdraws from active participation in the study, the reason(s) for discontinuation shall be recorded by the investigator on the relevant page of the eCRF. Whenever possible, all patients who discontinue treatment or withdraw from the study prematurely will undergo all assessments at the early withdrawal visit. Patients who fail to return for final assessments will be contacted by the site in an attempt to collect final data. The investigator should show due diligence and explore all possible options to reach a patient who fails to attend a visit. The investigator must document all attempts to contact the patient in the medical records/source documents (at least 3 documented approach, via phone, e-mail or regular mail). It is vital to obtain follow-up data on any patient withdrawn because of an AE. In every case, efforts must be made to undertake protocol-specified, safety, follow-up procedures. If patients are unable or unwilling to return for this follow-up visit, the site will document their efforts to bring the patients in through two documented telephone calls and a registered letter.
2.12 Adverse Event Reporting
Reports of adverse events, accidents, serious and unexpected adverse reactions, and device malfunctions will be sent immediately to ‘National Institute of Pharmacy and Nutrition’ at '[email protected]' (with the file number of the decision authorising the clinical trial).