Table 1.

Description of included studies

Study/Country	Total # of Patients	Study Design	Duration (week)	Patient population included in analysis	Definition of Relapse-Related Outcome	Mean Age (year)	% Male	% White	% SzAD	# of Pts per Arm	Mean Dose (range/fixed (mg/day)
Tamminga et al. '94(17)/USA	32	Rater Masked	52	OPs with TD stabilized for 1–6 months before randomization	Relapsee: study discontinuation due to decompensation	35.5	62.5	62.5	0	CLO: 19	293.8
										HAL: 13	28.5
Essock et al. '96(19)a/USA	124	OL	104	IPs in state hospital with FDA criteria for CLO use who were discharged	Relapsee: rehospitalization	41.2	60.8	NR	NR	CLO: 76	496
										FGA: 48	1386g
Speller et al. '97(33)/UK	60	DB	52	IPs on rehabilitation wards with moderate to severe negative symptoms, with combined score of ≥4 on the negative subscale of Manchester Scale	Relapse: when psychotic exacerbation (increase of 3≥on the combined score for the thought disturbance and paranoia items of BPRS) could not controlled with dose increase	63f	76.6	NR	0	AMI: 29	NR (100–800)
										HAL: 31	NR (3–20)
Tran et al. '98(12)b -study 1/North America	55	DB	46	Responders to acute phase therapy (BPRS-T decreased ≥40% from baseline or ≤18.) and have been OPs	Relapse: hospitalization for psychopathology	37.0	64.8	NR	11.5	OLA: 5	12.1 (12)
										HAL: 10	14.0 (14.0)
Tran et al. '98(12)b -study 2/International	62	DB	46	Responders to acute phase therapy (BPRS-T decreased ≥40% from baseline or ≤18.) and have been OPs	Relapse: hospitalization for psychopathology	37.0	64.8	NR	11.5	OLA: 48	11.5 (12)
										HAL: 14	16.4 (16)
Tran et al. '98(12)b -study 3/International	690	DB	22–84d	Responders to acute phase therapy (BPRS-T decreased ≥40% from baseline or ≤18.) and have been OPs	Relapse: hospitalization for psychopathology	37.0	64.8	NR	11.5	OLA: 534	13.9 (5–20)
										HAL: 156	13.2 (5–20)
Daniel et al. '98(18)/USA	203	DB	52	Medication-responsive OPs stable for ≥3 months but had hospitalization or decompensation within last 5 years, CGI-S≤4	Relapsee: hospitalization Treatment failure: 1) hospitalization; 2) 20% deterioration in BPRS-T, 3) discontinuation due to lack of efficacy or noncompliance, 4) use of other antipsychotics	37.0	75.4	60.1	0	SER: 94	(24)
										HAL: 109	(10)
Rosenheck et al. '99(23)a/USA	49	DB	52	Treatment refractory patients whose PANSS-T decreased ≥20% in the initial 6 week treatment	Relapsee: failed to maintain improvement	43.9	99.2	70.2	0	CLO: 35	628 (100–800)
										HAL: 14	28.2 (5–30)
Colonna et al. '00(42)a/France	322	OL	52	Responders after 1 month of acute phase treatment (BPRS-T decrease ≥20%)	Treatment failure: all discontinuation + cannot maintain response (BPRS-T ≥20% decrease)	37.5	67.0	97	0	AMI: 253	626 (200–800)
										HAL: 69	15.1 (5–20)
Csernansky et al. '02(32)/USA	361	DB	52	OPs judged stable by the principal investigater; stable dose of antipsychotics and same residence for 30 days	Relapse: 1) psychiatric hospitalization; 2) psychiatric care increase and 25% increase in PANSS-T, including ≥10 points increase; 3) self-injury, suicidal, homicidal ideation, violence; 4) CGI-C≥6	40.2	69.9	47.7	17.8	RIS: 177	4.9 (2–8)
										HAL: 184	11.7 (5–20)
de Sena et al. '03(43)c/Brazil	33	OL	52	Hospitalized due to an acute exacerbation	Relapse: first rehospitalization after discharge	27.7	18.0	18.9	0	RIS: 20	4.0f (flexible)
										HAL: 13	10f (flexible)
Kasper et al. '03(21)a study 1,2/USA (study 1), International (study 2)	633	DB	52	Acute phase patients who responded (≥30% reduction in PANSS-T) and not having any of 1) CGI-I≥6, 2) Adverse event of worsening schizophrenia, 3) PANSS psychotic subscale≥5	Relapsee: fail to maintain response	37.1	58.6	NR	0	ARI: 444	29.0 (30)
										HAL: 189	8.9 (10)
Lieberman et al. '03(20)c/China	143	DB	40	FEPs who discharged from 12 weeks of hospitalization for acute phase treatment	Relapsee: rehospitalization after week 12	28.7	52.0	0	0	CLO: 71	600f (flexible)
										CPZ: 72	400f (flexible)
Marder et al. '03(22)/USA	63	DB	104	Treated as OPs for ≥1 month but had ≥2 episodes of acute schizophrenic illness or having ≥2 years of continuing psychotic symptoms	Relapsee: psychotic exacerbation [1) ≥4points increase on the sum of BPRS cluster scores for thought disturbance and hostile-suspiciousness; 2) ≥3points increase on one of these clusters with one item ≥4	43.5	92.1	44.4	0	RIS: 33	5.7 (6)
										HAL: 30	4.5 (6)
Schooler et al. '05(44)a/International	400	DB	104	FEPs who achieved clinical improvement (≥20% decrease in PANSS total score)	Relapse: 1) ≥25% increase in PANSS-T, including ≥10 points increase; 2) CGI-C≥6; 3) deliberate self-injury; 4) suicidal or homicidal ideation or suicide; 5) violent behavior	25.5	71.4	74.4	7.6	RIS: 197	3.3 (up to 8)
										HAL: 203	2.9 (up to 8)
Lieberman et al. '03(45)/Green et al. '06(46)a/USA, Europe	133	DB	104	FEPs who remitted (PANSS P1, 2, 3, 5,6≤3, CGI-S≤3 for 4-week)	Relapse: failed to maintain remission	23.8	81.8	52.9	9.9	OLA: 75	10.2 (5–20)
										HAL: 58	4.82 (2–20)
Gaebel et al. '07(24)/Germany	151	DB	52	Successfully completed acute therapy (CGI-C≤3) in the first illness episode	Relapse: ≥10 increase of PANSS positive + CGI-C≥6 + GAF≥20 decrease; hMarked clinical deterioration: 1)single fulfillment of relapse criteria, 2) ≥7 increase in PANSS positive + ≥15 decrease in GAF	31.6	58.3	NR	0	RIS: 77	4.2 (2–4)
										HAL: 74	4.1 (2–4)
Kane et al. '08(34) study 1 – 3 pooled/International	473	DB	46	Completed initial 6 wks with ≥20% decrease of PANSS-total score and CGI-C≤4	Relapse: 1) ≥25% increase in PANSS-T, including ≥10 points increase; 2) discontinuation due to lack of efficacy; 3) aggravated psychosis with hospitalization; 4) ≥2 increase in CGI-S	34.7	63.4	45.6	6.3	ILO: 359	11.8 (4–16)
										HAL: 114	13.2 (5–20)
Kahn et al. '08(10)/Europe and Israel	351	OL	52	FEPs within 2 years since the onset of positive symptoms and had ≤14days of antipsychotic exposure	Relapsee: admitted to hospital after randomization	26.0	60.0	94.0	7.0	AMI: 88	450.8 (200–800)
										OLA: 89	12.6 (5–20)
										QUE: 60	498.6 (250–700)
										ZIP: 60	107.2 (40–160)
										HAL: 64	3.0 (1–4)
Crespo-Facorro etal. '10(47)/Spain	166	OL	52	FEPs who have improved by study medication to CGI-S≤4, 30%≥ decrease of BPRS-T, all BPRS item≤3 for ≥4 weeks	Relapse: 1) any key BPRS item≥5, 2) CGI-S≥6, CGI-G≥6, 3) psychotic hospitalization, 4) complete suicide	27.4	62.0	NR	2.4	OLA: 54	10.4 (5–20)
										RIS: 58	3.4 (3–6)
										HAL: 54	2.9 (3–9)

^aSubpopulation of responded or remitted patients used in this meta-analysis, but demographic data was obtained from study original total population

^bDemographic data was obtained from study 1,2,3 pooled data.

^c100% of patients discharged after randomization.

^dSubjects completed between 22 and 84 weeks of double blind therapy.

^eOriginal study didn't have relapse definition, mentioned outcome utilized as relapse in the analysis

^fReported median

^gChlorpromazine equivalent dose

^hUtilized marked clinical deterioration as relapse in analysis

Abbreviations: AMI=amisulpride, ARI=aripiprazole, BPRS=Brief Psychiatric Rating Scale, BPRS-T=BPRS total score, CGI-C=Clinical Global Impressions scale-change score, CGI-S=Clinical Global Impressions scale-severity score, CLO=clozapine, CPZ=chlorpromazine, DB =double blind, FEPs=first-episode patients, FGA=frrst-generation antipsychotics, GAF=Global Assessment of Functioning scale, HAL=haloperidol, ILO=iloperidone, IPs=inpatients, NR=not reported, OL=open label, OLA=olanzapine, OPs=outpatients, PANSS=Positive and Negative Syndrome Scale, PANSS-T=PANSS total score QUE=quetiapine, RIS=risperidone, SER=sertindole, SzAD=schizoaffective disorder, ZIP=ziprasidone