Table 1. Properties of studies included for analysis.

Antidepressant (including dose range, duration, adherence)	Authors	Psychiatric diagnoses and comorbidities (including method of identification)	Medical comorbidities and concurrent treatments (including method of identification)	Outcome Measures (including times of measurement)	Number of Subjects (including at commencement, completion, and analysis)	Results of Outcome Measures (including relevant confidence intervals, standard deviations, and/or p values)
• Citalopram (20-50mg, 2 months) • Paroxetine (20-50mg, 2 months)	Perna, Bertani [19]	• Panic disorder +/- agoraphobia (DSM-IV; clinical interview + MINI) • Nil psychiatric comorbidities (physical examination and medical history assessment)	• Nil significant physical illnesses (history and examination) • Nil concurrent psychotropic meds or psychotherapy	• FQ (Baseline, days 7 & 60) Others: • Panic Associated Symptom Scale • Sheehan Disability Scale (SDS)	• Citalopram: 30 started, 27 completed; 27 analysed • Paroxetine: 28 started, 25 completed; 25 analysed	FQ Agoraphobia • Significant reductions from baseline to endpoint according to post-hoc comparisons (p not stated): • Citalopram: baseline 15.7 +/- 12.3; endpoint 9.9 +/- 9.1 • paroxetine baseline 15.6 +/- 11.9; endpoint 8.4 +/- 10.4 • Significant time effect found for antidepressant treatment (F 3.8, p<0.05) FQ Blood injury • Significant reductions from baseline to endpoint according to post-hoc comparisons (p not stated): • Citalopram: baseline 21.1 +/- 10.8; endpoint 15.6 +/- 9.9 • Paroxetine baseline 17.5 +/- 9.5; endpoint 12.4 +/- 10.1 • Significant time effect found for antidepressant treatment (F 9.0, p<0.001)
• Desvenlafaxine (100mg or 150mg, 12 weeks)	Cheng, DuPont [20]	• Nil GAD, mood & psychotic disorders (DSM version not stated; method not stated)	• Vasomotor symptoms post-menopause • Nil concurrent physical illnesses • Nil concurrent psychotropic or hormonal meds	• POMS (Baseline, weeks 4 & 12) Others: • Greene Climacteric Scale • Menopause Symptoms–Treatment Satisfaction Questionnaire	• 100mg: 153 started, 134 completed; 120 analysed • 150mg: 152 started, 137 completed; 117 analysed • Placebo: 153 started, 138 completed; 123 analysed	Significant reductions in POMS Tension and POMS Depression vs placebo (p<0.05):POMS Tension • 100mg: Baseline 9.3 (SD = 6.7), change -4.1 +/- 0.4 • 150mg: Baseline 9.0 (SD = 7.2), change -3.9 +/- 0.4 • Placebo: Baseline 8.6 (SD = 6.4), change -2.4 +/- 0.4 • No statistically significant difference between doses (p value not stated) POMS Depression • 100mg: Baseline 8.3 (SD = 8.6), change -5.5 +/- 0.6 • 150mg: Baseline 8.2 (SD = 11.1), change -4.5 +/- 0.6 • Placebo: Baseline 7.4 (SD = 9.6), change -2.3 +/- 0.6 • No statistically significant difference between doses (p value not stated)
• Escitalopram 10-20mg, 12 weeks	Alamy, Wei [21]	• Specific phobia (DSM-IV as per MINI) • Nil comorbid MDD, GAD, Social Phobia, OCD, PD, PTSD (method not stated) • Nil recent history of substance use/dependence (method not stated) • Nil lifetime BD, SCZ, organic brain syndrome (method not stated)	• Nil significant abnormalities as per haematology, chemistry, serum pregnancy (for women) & ECG • General direction for subjects to make efforts to expose themselves to avoided situation	• FQ (Screening, baseline, weeks 1,2,4,8 and 12) Others: • Main Phobia Scale (MPS) • CGI-I • HAM-A • MINI phobia module • Symptom Occurrence Scale	• Escitalopram: 6 started, 5 completed; 5 analysed • Placebo: 7 started, 6 completed; 6 analysed	No statistically significant reduction in FQ found (p not stated):FQ Agoraphobia: • Escitalopram: Baseline 3.0 (SD = 2.4), endpoint: 1.2 (SD = 1.1) • Placebo: baseline 2.9 (SD = 4.3), endpoint: 0.4 (SD = 0.8) FQ Blood Injury: • Escitalopram: Baseline 4.8 (SD = 4.1), endpoint: 1.6 (SD = 1.3) • Placebo: Baseline 6.9 (SD = 8.3), endpoint: 4.0 (SD = 7.2)
• Fluoxetine (20mg, 6 months)	Wood, Mortola [22]	• Late Luteal Phase Dysphoric Disorder (SCID for DSM-III-R) • Comorbidities excluded (SCID for DSM-III-R and psychometric abnormalities during follicular phase:     - BDI > = 12     - STAI > = 45     - MMPI T-score > = 70     - POMS total T-score > = 70)	• Premenstrual Syndrome (history, examination, Calendar of Premenstrual Experiences) • Nil significant medical or gynecologic disorders (history and examination) • Nil comments on concurrent meds	• POMS (Baseline, months 3 & 6) Others: • Calendar of Premenstrual Experiences • Beck Depression Inventory (BDI) • State-Trait Anxiety Inventory (STAI) • Minnesota Multiphasic Personality Inventory	• 8 started, 8 completed, 8 analysed	Significant differences between luteal phase and follicular phase when baseline T-scores of fluoxetine and placebo groups pooled (p<0.01) (individual group baselines not stated, nor was the sample from which the T-scores were derived described): • POMS Depression T-score: Luteal: 59.6 +/- 4.6 Follicular: 44.8 +/- 5.2 • POMS Tension T-score: Luteal: 56.6 +/- 3.3 Follicular: 40.5 +/- 3.6 Luteal phase: significant differences between fluoxetine and placebo (p<0.005) • POMS Depression T-score: Fluoxetine: 39.7 +/- 2.9 Placebo: 50.9 +/- 4.5 • POMS Tension T-score: Fluoxetine: 39.1 +/- 3.4 Placebo: 50.4 +/- 3.8 Follicular phase: no significant differences between placebo & fluoxetine (p not stated): • POMS Depression T-score: Fluoxetine: 39.7 +/- 2.2 Placebo: 43.4 +/- 2.9 • POMS Tension T-score: Fluoxetine: 37.8 +/- 2.9 Placebo: 43.1 +/- 3.7
• Fluvoxamine (150mg, 12 weeks)	Sharp, Power [23]	• Panic disorder +/- agoraphobia (DSM-III-R; clinical interview + HAM-A > = 15) • Nil depressive disorder (MADRS > = 21) • Nil OCD, paranoid PD, psychotic, manic or substance disorders (method not stated)	• Nil severe illnesses (as per GP referral) • Nil concurrent or recent psychotropics • Nil past psychotherapy	• FQ (baseline, weeks 7 & 12, 6 months) Others: • Hamilton Anxiety Scale • Montgomery-Asberg Depression Rating Scale (MADRS) • Panic diaries (created by investigators)	• Fluvoxamine: 36 started, 24 completed; 29 analysed • Placebo: 37 started, 20 completed; 28 analysed	FQ Agoraphobia • Fluvoxamine: Significant difference found vs placebo at day 84 (p<0.05) • Fluvoxamine: Significant reductions between baseline and endpoint found, t = 3.44, df = 28, p<0.002 (but baseline and end-point scores not stated) • Fluvoxamine: Significant time and interaction effects found
• Fluvoxamine (50-200mg*, 4 weeks; *study stated a maximum range of “209mg” but this was likely a typographic error)	Itil, Shrivastava [24]	• Major affective disorders (RDC; clinical interview + HAMD > = 15) • Nil substance dependence (method not stated)	• Nil significant organic disease • Nil concurrent or recent psychotropics	• POMS (baseline, weeks 1,2,3 and 4) Others: • Clinical Global Impression [25] • Hamilton Depression Scale (HAM-D) • SCL-90 • BDI • Dosage record • Treatment Emergent Signs & Symptoms	• Fluvoxamine: 22 started, 10 completed; 9 analysed • Imipramine: 25 started, 13 completed; 14 analysed • Placebo: 22 started, 11 completed; 14 analysed	POMS Depression • Significant difference between fluvoxamine vs placebo at week 3 (p = 0.028) but not at week 4 (p = 0.06)
• Nefazodone (300-600mg, 12 weeks; although results indicated final dosage range was 200-600mg; disparity not discussed)	Van Ameringen, Mancini [25]	• Social phobia–generalised (DSM-IV; SCID-I/P) • Nil exclusions (sample included MDD, dysthymia, PD with agoraphobia, OCD, alcohol abuse/dependence)	• Nil recent or concurrent psychotropics • Nil current psychotherapy • Concurrent organic illnesses not discussed	• FQ (baseline, weeks 4, 8 & 12) Others • BDI • STAI • SDS • Social Adjustment Scale Self-Report (SAS-SR) • Fear of Negative Evaluation Scale (FNES) • Social Avoidance and Distress Scale (SADS) • Social Anxiety Thoughts Questionnaire (SATQ) • Brief Social Phobia Scale (BSPS) • Liebowitz Panic and Social Phobic Disorders rating form (LPSPD)	• 23 started, 21 completed, all evaluated	FQ-Agoraphobia • No significant difference between baseline (9.0, SD = 7.8) and week 12 (6.4, SD = 6.5), F2.3, df 3,51, p = 0.09.
• Paroxetine (10-60mg, 12 weeks)	Mancini and Ameringen [26]	• Social phobia–generalised (DSM-III-R; SCID-I/P) • Nil exclusions (sample included MDD, dysthymia, PD, agoraphobia, PD with agoraphobia, simple phobia, OCD, GAD, alcohol abuse/dependence)	• Nil recent or concurrent psychotropics • Nil current psychotherapy • Concurrent organic illnesses not discussed	• FQ (baseline, weeks 4, 8 & 12) Others • BDI • STAI • SDS • SAS-SR • FNES • SADS • SATQ • LPSPD	• 18 started, all completed	FQ-Agoraphobia • Significant difference between baseline (11.8 +/- 10.8) and week 12 (5.8 +/- 7.1), F3.8, df 3,39, p = 0.017
• Sertraline (50-150mg, 12 weeks)	Alpert, Silva [27]	• MDD (DSM-III-R; semi-structure interview including DSM-III-R checklist; severity HAMD > = 18) • MMSE > = 23 • Nil other Axis I diagnoses (semi-structured interview)	• Nil acute or unstable physical conditions • Concomitant temazepam or chloral hydrate PRN allowed • Nil other recent or concurrent psychotropics	• POMS (Baseline, week 12) Others • Folate • HAM-D	• Sertraline: 12 started, 12 completed, 12 analysed • Nortriptylline: 10 started, 10 completed, 10 analysed	POMS Depression • Sertraline: Change from baseline (1.5, SD = 0.67) to endpoint (1.2, SD = 1.13) not statistically significant (p value not stated)
• Sertraline (50mg, 100mg, 200mg; 6 weeks)	Fabre, Abuzzahab [28]	• MDD (DSM-III; unclear diagnostic method; severity HAMD > = 22) • Nil history of psychotic disorders or current substance abuse (method not stated)	• Nil history of significant medical disease • PRN chloral hydrate allowed except for “nights before psychiatric scale testing” • Nil recent or other concurrent psychotropics	• POMS (Baseline, weeks 1 to 6) Others • HAM-D • CGI • ECG • Vitals • FBE • Biochemistry • Urinalysis	• 50mg: 95 started, 59 completed; 90 “all-patients group” (those who took medication on or after 11th day), 82 “evaluable group” (those who took 1+ dose of medication) • 100mg: 92 started, 47 completed; 89 “all-patients group”, 75 “evaluable group” • 200mg: 91 started, 39 completed; 82 “all-patients group”, 56 “evaluable group” • Placebo: 91 started, 86 “all-patients group”, 46 completed; 76 “evaluable group”	Results for “all-patients group”:POMS Depression • Sertraline 50mg: baseline 2.8 +/- 1.1; change vs endpoint -0.5 +/- 1.1 significantly different vs placebo, p = < 0.01 • Sertraline 100mg: baseline 2.7 +/- 1.0; change vs endpoint -0.5 +/- 1.1 statistically different vs placebo, p = <0.05 • Sertraline 200mg: baseline 2.5 +/- 0.7; change vs endpoint -0.4 +/- 1.0 statistically different vs placebo, p = <0.05 • Combined: baseline 2.7 +/- 1.1; change vs endpoint -0.5 +/- 1.0 statistically different vs placebo, p = <0.01 • Placebo: baseline 2.6 +/- 1.1; change vs endpoint -0.1 +/- 0.7 • No statistically significant differences amongst different dosages (p value not stated).
• Sertraline (50-200mg, 8 weeks)	Lydiard, Stahl [29]	• MDD (DSM III-R; Clinical interview; severity HAMD > = 18) • Nil concurrent dysthymia, bipolar, severe GAD, OCD, PTSD, psychotic disorders, severe PDs, substance dependence (Clinical interview)	• Nil concurrent significant medical illness (history, examination, ECG & blood tests) • Temazepam or chloral hydrate PRN allowed • Nil concurrent psychotropics	• POMS (baseline, weeks 1 to 8) Others • BDI • MADRS • Global Assessment Scale • CGI • Quality of Life Enjoyment and Satisfaction Questionnaire • Health-Related Quality of Life battery, v2	• Sertraline: 132 started, 96 completed, 128 analysed • Amitriptyline: 131 started, 81 completed, 127 analysed • Placebo: 129 started, 92 completed, 125 analysed	POMS Depression • Sertraline: change from baseline 1.9 (SE = 0.08) to endpoint by -1.0 (SE = 0.08) statistically different, p<0.005 • Placebo: change from baseline 1.9 (SE = 0.08) to endpoint by -0.5 (SE = 0.08) statistically different, p<0.005 • Change in POMS-D between sertraline and placebo was significantly different, p<0.001
• Sertraline (50,150mg, 12 weeks)	Finkel, Richter [30]	• MDD (semi-structured interview that included DSM-III-R checklist; severity HAMD24 > = 18) • Excluded: Nil acute or chronic organic mental disorder, MMSE <24, “clinically significant psychiatric illness, active suicidality”	• Hypnotics allowed • Nil acute or unstable medical condition or Ix abnormalities (physical exam, history, ECG, lab investigations) • Nil concurrent anticoagulants (except aspirin) • Nil current or past history of neurological disease • Nil past non-response to 6+ weeks of 2+ adequate doses of antidepressants	• POMS Others • HAM-D • HAM-A • CGI-S/I • Q-LES-Q • MMSE • DSST • SLT • Plasma drug concentrations	• Sertraline: 39 started, 26 finished; 7 quit due to SE; none due to lack of efficacy mean dose 100 +/- 40mg; serum levels not stated • Nortriptylline: 37 started, 19 finished; 11 quit due to SE; none due to lack of efficacy; mean dose 70+/-30mg; 29% had serum levels <50ng	No significant difference in concomitant meds or baseline valuesPOMS Tension • baselines: not stated • change from baseline: -5.9, significantly different vs nortriptyline • time effect not factored POMS depression results not stated
• Venlafaxine (week 1: 75mg, weeks 2–8: 150mg, 8 weeks)	Ozdemir, Boysan [31]	• 1st episode MDD (DSM IV-TR; SCID-I) • Nil current or past history of bipolar disorder or substance dependence (method not stated)	• Concurrent illnesses or medications not discussed	• POMS (baseline, weeks 1,2,4 & 8) Others • HAMD • BDI	• Venlafaxine: 25 started, all completed • Venlafaxine + Bright Light Therapy: 25 started, all completed	POMS Depression • Venlafaxine: baseline 37.16 +/- 11.36; endpoint 11.12 +/- 9.11; statistically significant reduction over time, p<0.001 POMS Tension • Venlafaxine: baseline 23.68 +/- 4.88; endpoint 9.60 +/- 3.93; statistically significant reduction over time, p<0.001