Van der Kolk 2007.

Study characteristics
Methods	Design: multi‐centre, randomised, placebo‐controlled, parallel, flexible‐dose, double‐blind, long‐term maintenance Duration of intervention: 5 weeks Placebo run‐in: No information Post‐treatment: 6 month follow‐up
Participants	Setting: Quote: “Multi‐centre, recruited via newspaper ads, the Internet, and solicitation from medical and mental health professionals" Sample size: 59 randomised to fluoxetine and placebo Mean age: 34.9 years (18 ‐ 65). Sex: 8 men and 51 women Diagnostic measure: DSM‐IV Inclusion criteria: Quote: “Individuals 18‐65 years old with current PTSD and mixed trauma exposure at least one year prior to intake were recruited via newspaper ads, the Internet, and solicitation from medical and mental health professionals" Exclusion criteria: Quote: "Exclusion criteria were unstable medical condition; contraindications to either treatment (i.e., pregnancy; glaucoma or detached retina; history of severe allergies or multiple adverse drug reactions); inability to be weaned off current psychotropic medications; psychotic or bipolar disorder; current alcohol or substance abuse/dependence; severe dissociation; active suicidality or life‐threatening mutilation; prior exposure to active study interventions; concurrent trauma‐focused treatment; unstable living situation; GAF<40; and disability compensation for PTSD or pending trauma‐related lawsuit. Initial telephone screening was used to assess likely presence/absence of inclusion and exclusion criteria; potential participants were invited for in‐person assessment" Comorbidity: Axis I and II disorders at baseline Dropout rates: 7/59 (4/30 in the fluoxetine and 3/29 in the placebo group)
Interventions	Pharmacological intervention: Quote: “Starting dosage was 10 mg/day of fluoxetine (or pill placebo equivalent). Dosage was increased in 10 mg increments per week to a maximum of 60 mg/day or until symptom remission was achieved. Increases or decreases in dosage were based on physician judgment of clinical response and presence/absence of dose‐limiting side effects"
Outcomes	Primary outcomes: CAPS Secondary outcomes: BDI Time points: CAPS: Rating was based on 1‐week interval for immediate pre‐ and posttreatment assessment; 1‐month interval was used at baseline and six‐month follow‐up Data estimation: LOCF
Notes	Dates of trial: Quote: “The study ran from July 2000 through July 2003” Industry‐funded: No. Quote: "This study was supported by grant R01MH58363 from the National Institutes of Mental Health" Medication provided by industry: No information Any of the authors work for industry: No
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Participants were randomly assigned to 1 of 3 treatment conditions: EMDR, fluoxetine, or pill placebo. Randomization was stratified by presence/absence of concurrent supportive psychotherapy. In order to ensure approximately equal numbers in each treatment condition, random assignment was blocked in groups of 12 consecutive participants, so that in each block, 4 participants were assigned to each condition. Participants in all 3 conditions received a total of 8 weekly treatment sessions. Following cessation of treatment, participants in the 2 active treatment conditions were asked to refrain from initiating new treatment during the 6‐month follow‐up period"
Allocation concealment (selection bias)	Unclear risk	There was insufficient information provided to determine if study medication allocation was concealed. Quote: "Starting dosage was 10 mg/day of fluoxetine (or pill placebo equivalent)"
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	The study was described as "double‐blind", although no information was provided on which parties were blinded and how blinding was achieved (re participants) Quote: "The fluoxetine and placebo interventions were administered in a double‐blind, fixed flexible‐dose design according to standard protocol for double‐blind pharmacologic interventions for PTSD. All raters were blind to treatment condition and were never assigned the same participant for both pretreatment and posttreatment evaluation. All EMDR sessions were videotaped, and an independent evaluator assessed treatment fidelity through videotape review from randomly sampled sessions"
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Quote: "All raters were blind to treatment condition, and were never assigned the same participant for both pre‐ and post‐treatment evaluation"
Incomplete outcome data (attrition bias) All outcomes	Low risk	A similar proportion of participants withdrew from the fluoxetine group (4/30, 13%) compared to the placebo group (3/29; 10%). The 2 groups did not differ by sample characteristics at baseline. Quote: "Significantly more patients discontinued the study due to an adverse event in the 40‐mg/d fluoxetine treatment group (13.1%) than in the 20‐mg/d fluoxetine treatment group (4.3%; P = 0.005); however, neither of these rates differed significantly from the placebo treatment group (8.0%, both P > 0.20). There were no other significant differences between treatment groups for other reasons of discontinuation, including discontinuation due to lack of efficacy (20 mg fluoxetine, 6.7%; 40 mg fluoxetine, 3.8%; placebo, 6.8%; P = 0.416) or discontinuation due to patient decision (20 mg fluoxetine, 11.0%; 40 mg fluoxetine, 6.9%; placebo, 13.6%; P = 0.182)". Twelve people dropped out during the 8‐week treatment phase, leaving 76 treatment completers. There were no significant differences in dropout rates on any baseline measure of psychopathology or across treatment conditions. Completers in each group were 24 of 29 (83%) for EMDR, 26 of 30 (87%) for fluoxetine, and 26 of 29 (90%) for pill placebo. Dropouts were significantly younger than completers (dropout mean age = 27.1 years, completer mean age = 37.6 years; F = 6.785, df = 1,86; p < .05). In addition, dropouts were more likely to have had child‐ versus adult‐onset trauma (χ2 = 6.175, df = 1,88; p = .013)Study hypotheses were tested on both treatment completer and intent‐to‐treat (ITT) samples. Intent‐to‐treat analyses were conducted using an early termination assessment, when available, or a last‐observation‐carried forward (LOCF) procedure to impute missing data. Long term effects of active‐treatment completers were assessed at 6 months posttreatment for both follow‐up completer and intent‐to‐follow (ITF) samples. For the ITF sample, missing follow‐up data were also estimated using LOCF. Participants in the 3 treatment conditions did not differ significantly on any demographic variable or on any baseline measure of psychopathology"
Selective reporting (reporting bias)	Unclear risk	The study protocol was not available for this study
Other bias	Low risk	No other sources of bias were identified