Atomoxetine and Executive Function in PTSD

Aim 1: To examine the impact of ATX in enhancing executive function. Based on preliminary data, the investigators hypothesize that ATX, compared with placebo, will be more effective in improving EF in Veterans with PTSD and AD. To test this hypothesis, 12-week, randomized, double-blind, placebo-controlled trial of ATX will be conducted among Veterans with PTSD. The outcome measure will be the analyses of changes in Behavior Rating Inventory of Executive Function-Adult (BRIEF-A), the response inhibition task, GoNogo (GNG), Trail making and digit span tests (TMT and DST). The rationale for this aim is that impaired EF may be a powerful risk for PTSD treatment resistance and poor QoL and psychosocial function (PSF), therefore, improving EF with ATX may ultimately improve Veterans QoL, PSF, and reducing disability rating. The investigators expect that the 12-week treatment with ATX will significantly improve attention, EF, inhibitory function in Veterans with PTSD. Aim 2: To examine the impact of ATX on improvement of QoL and daily psycho-social function. Based on literature and preliminary data, the investigators hypothesize that ATX, compared with placebo, will be more effective in improve Veterans' QoL, PSF, and disability rating. The outcome measure will be the analyses of changes in Adult ADHD Quality of Life-29 (AAQOL-29), WHO Disability Assessment Schedule 2.0 (WHODAS 2.0), The Inventory of Psychosocial Functioning (IPF). The investigators expect that the 12-week treatment with ATX will significantly improve QoL and psychosocial function, and significantly reduce EFD and AD related disability. Aim 3: To examine the impact of ATX on overall AD and PTSD treatment outcomes. The hypothesis for this aim is that successful treatment of attention and executive deficits will be associated with greater reduction of PTSD and AD symptoms. To achieve this goal, more clinical and psychological assessments including depression, anxiety, and other functional scales, such as CAPS-5 scores, BDI, BAI, and CARRS-S:S attention deficit scores will be performed to assess the PTSD overall improvement. The investigators expect that the significant improvement of attention and executive function will lead to a significant improvement of overall PTSD outcomes. Recruitment: Recruitment of subjects will take place in three sites, including the Ralph H. Johnson (RHJ) VAMC, Savannah VA Community Based Outpatient Clinic (CBOC), and Hinesville CBOC. The decision to use three sites were based on the calculation of study sample size, the availability of the target patient population, and the feasibility of recruiting the proposed study sample. The investigators will devote 2 months to study start-up; 40 months for active recruitment (10 months in year 01, 12 months in year 02 and year 03, 6 months in year 04); 3 months for the patients enrolled in the last recruitment month to complete the treatment and follow-up; 1 months for the study close-out; and 6 months for data analysis. The total length of study will be 4 years. The investigators plan to recruit 160 subjects with PTSD in 4 years, and allow 20% dropout with a final of 128 subjects (64 from the RHJ VAMC, 32 from the Savannah CBOC, and 32 from the Hinesville CBOC) entering to the final analysis. The investigators plan to recruit subjects seeking outpatient treatment in the VA PTSD Clinician Team (PCT) clinics and the Mental Health Service Line (MHSL) clinics at RHJ VAMC, and at Savannah and Hinesville CBOCs. The primary source of recruitment will be PCT clinic referrals by PCT intake team, which also referring patients to Savannah and Hinesville clinic. As the PI and coinvestigators in this research project are either directors or attending physicians in each of the three sites, the investigators will have direct contact with patients assigned to the PCT clinics and MHSL clinics. Therefore, the investigators anticipate no difficulty in recruiting 1 PTSD subject every 2 weeks from RHJ VAMC, and 1 PTSD subject/4weeks from each of the CBOCs. This estimate is very reasonable, as the investigators recruited over 44 subjects in 2 years in the Preliminary Study only at RHJ VAMC. In the investigators' previous studies, it was found about 50% of Veterans with PTSD meet the ADHD criteria (Adams et al, 2015). Flyers will also be posted in appropriate areas of VAMC facilities. Treatment Duration: The double-blind treatment will last 12 weeks. Study visits will be scheduled at weeks 2, 5, 9, and 13. Safety assessment, such as C-SSRS will be conducted at each visit. In addition, the C-SSRS will also be conducted by weekly phone call between in-person visits. Patients randomized after the initial assessment. Those patients will be followed until the end of the study. Following their final visit, subjects will be referred to their primary care physician for a clinical follow up treatment. The 4th treatment visit and final follow up visit can both be virtual if the subjects chooses to do so, this decision will be optional depending on the subject's preference. Data Collection: An operations manual and case report form will be developed. The site study coordinator will collect all data. After a patient consent to participate in the study, the site coordinator will create a patient casebook, which will contain the consent forms, all relevant source documents, and any other information pertinent to the study. The case report forms will be completed at each visit. This is a password-protected system that is accessed through the VA intranet. The study personnel, including the biostatistician, statistical programmer, database programmer, will have access to read the data. The information collected about the patients during the course of the trial will be stored at the Research Database Storage in the Charleston VAMC PTSD Program. The database will be shared among the researchers involved in this project in the future, but information that would disclose personal identity (name, address, telephone, and security number) will not be released and will always be kept separate from the research database. Primary Data Analysis: Baseline clinical and demographic characteristics will be collected, and association of baseline characteristics and clinical predictors of study outcomes will be performed utilizing appropriate regression models. Study outcomes will be collected longitudinally at study baseline, during study treatment (weeks 8, 12) and at study follow-up (week 16). Continuous study outcomes (BRIEF-A, TMT, and DST (Aim1); AAQoL, IPF, and WHODAS (Aim2); CAPS-5, and CARRS (Aim3)) will be assessed utilizing linear mixed effects regression models (LMM) including all post randomization time points. Transformations using restricted cubic splines or fractional polynomials will be considered for continuous model covariates if significant departure from linearity is detected. Normality of residuals will be assessed through Q-Q plots and when departure from normality is detected, appropriate transformations of non-parametric modeling will be utilized. Generalized linear mixed effects regression models (GLMM) will be utilized for assessment of count data (GoNoGo omission/commission errors, suggesting inattention and impulsivity). For count GoNoGo omission and commission error measures, overdispersion will be evaluated and alternative models (e.g., negative-binomial, zero-inflated) used as needed. Further, the GoNoGo task will record latency to response in the Go tasks. With the possibility of temporal trends in the longitudinal collection of data (learning) as well as non-response during the response window (fixed censoring), random effects survival models will be considered to examine study group differences in response time (RT) to Go tasks. Contrasts of interest (end of treatment outcomes) will be assessed with model based least square means and associated standard errors. Subgroup analyses will be pre-specified and considered exploratory. Treatment heterogeneity by clinic will be evaluated using appropriate models with clinic-by-treatment interaction. Both naïve and adjusted models will be presented; naïve models will contain design specific covariates only (study visit, baseline response, study clinic, and treatment assignment when appropriate) while covariate adjusted models will be developed using characteristics that were found to be significantly associated with study outcomes or are previously known confounders. In all longitudinal models, a treatment by time interaction will be included to assess differential response over the course of study treatment and when found significant, stratified analysis will be performed and results presented. Participants positively responding to all Go and NoGo tasks will be assessed for task compliance on a per participant basis. Sensitivity analysis with and without participant data will be conducted to assess the impact of such instances, although the investigators expect the occurrence rate to be minimal. To assess the synchronous relationship between primary study outcomes and cooccurring measures of depression and anxiety, time varying covariates (BAI, BDI) will be independently added to each model and assessed for significance. Further, moderation of outcomes by both baseline, and concurrent measures of depression and anxiety will be evaluated through model interaction terms (ATX x depressions/anxiety). When significant, stratified analysis will examine the relationship between depression and anxiety with each outcome and compared across treatment groups. Secondary Data Analysis: To assess if improvements in EF during study treatment are associated with co-occurring QoL, IPF, PTSD and AD symptom improvement, LMM will be developed including the time varying effect of EF variables (BRIEF-A, GNG) on QoL, WHODAS, IPF) and on PTSD (CAPS-5) as well as AD symptoms (CARRS). At each treatment measurement timepoint, change from study baseline will be calculated as a summary measure prior to analysis. Both concurrent and time lagged independent variables will be assessed for association. Following completion of the primary analysis described above, durability of treatment efficacy will be performed on measured collected at the study follow-up visit. Similar linear and generalized linear models will be developed to assess outcomes as described in the primary data analysis section. Missing Data: Missing data in longitudinal studies can be a problematic feature but can be mitigated through study design considerations. To minimize missing data and study attrition, study simplification and enhanced communication between study coordinators and participants will be emphasized. The investigators will make every effort to prevent attrition, e.g., telephone reminders prior to visits, participation compensation, flexible scheduling, and reinforcing adherence at each visit. However, these methods do not ensure that all data will be collected, and appropriate analytic methods will be employed to accommodate missing data (e.g. multiple imputation). In addition, in keeping with the ITT principal, the investigators will make every effort to continue assessments for the entire course of randomized treatment, even among those who stop participating in the study assigned intervention or fail to complete the full medication time course. Data Quality Assurance: REDCap facilitates the process of gathering and storing data, ensuring proper documentation of variables, checking data validity, and facilitating the processes of creating analytic databases for statistical analyses. For each variable, a set of range and consistency checks will be developed by the biostatistician in conjunction with the Principal Investigator and Research Coordinator. Data will be audited for correctness by two independent data-entry personnel periodically (quarterly) and discrepancies will be evaluated, and any errors identified will undergo review and resolution. Data ranges will be assessed, and examination of outliers performed during quarterly data checks. Extreme values that are plausible will be retained in the analysis. Implausible data will be flagged for additional checks (ie. a report of 500 cigarettes smoked per day). Data and Safety Monitoring: The data safety monitoring committee (DMC) will be charged with monitoring this proposed clinical trial. This DMC will be constituted specifically for this study. The DMC will provide ongoing independent evaluation of the study progress including participant accrual and retention, adverse events monitoring, and data analysis plan. The DMC will be comprised of a minimum of three experts in clinical research including an internal medicine physician, a psychiatrist, and another clinical scientist with statistical expertise. The DMC will initially meet before study start-up and at that time will determine the frequency of subsequent reviews. It is proposed that the DMC meet at least every 6 months. In addition to the above noted safety and accrual parameters, the DMC will also review completeness of follow-up, data quality, protocol deviations, and review protocol modifications. Within each DMC report, balance of important study characteristics will be assessed across blinded study groups (A/B); specifically race/ethnicity, baseline PTSD and AD measures. Participant age and sex are randomization stratum and should retain balance, however, these will also be assessed at each DMC meeting. Dosing Schedule: The medication (40mg to 80mg ATX or placebo) schedule is shown in Table 3. Dispensing of blinded medication will begin at randomization and will be dispensed at weeks 2, 4, 8, and 12. The initial dose of ATX is 40mg or placebo and will be given for one week, the dose will be titrated to 80mg or placebo daily if tolerated in the second week through week 12. Medication Packaging, Assignment and Dispensing Plan: Active and placebo ATX (40mg capsules) will be designed by Dr. David Shirley at the Plantation Pharmacy, Charleston SC, then labeled, packaged, and distributed by the VA Clinical Research Pharmacy Coordinators. Packaging will consist of boxes containing either ATX or placebo. For the active treatment periods, capsules will be boxed to contain a four-week supply. Each box will be uniquely numbered, and the box numbers will be tied to their contents (active or placebo) in a database at the Charleston VAMC. This database will also contain the randomization treatment assignment of each participant. At each dispensing visit site, personnel will access via a password - secured Kit Assignment Program on the study's SharePoint website to obtain box numbers for dispensing to a specific participant. To ensure that each patient is assigned boxes filled with capsule corresponding to their treatment assignment, all blinded box assignments will require the use of two unique patient identifiers. Records will be maintained of all doses dispensed, and Veterans will be asked to return all unused tablets and empty prescription vials to the clinic for assessment of treatment compliance. Assessment of Compliance and Methods to Maximize Compliance: Medication will be stored and dispensed from the VAMC investigational drug pharmacy. Records will be maintained of all doses dispensed, and Veterans will be asked to return all unused tablets and empty prescription vials to the clinic for assessment of treatment compliance. Compliance with medication will be assessed via returned capsule count in the presence of the patient at each visit, which is a traditional method of compliance assessment. Methods successfully utilized to help minimize study withdrawals, including follow-up phone calls, and availability of 24-pager coverage to discuss medication-related and other issues, will also help maximize medication compliance.