Fichier PDF

Partage, hébergement, conversion et archivage facile de documents au format PDF

Partager un fichier Mes fichiers Convertir un fichier Boite à outils PDF Recherche PDF Aide Contact



LancetAP.pdf


Aperçu du fichier PDF lancetap.pdf

Page 1 2 3 4 5 6 7 8 9 10 11 12

Aperçu texte


Articles

were not available, we used change in Brief Psychiatric
Rating Scale26 from baseline to endpoint, and then values
at study endpoint of these scales. Intention-to-treat
datasets were used whenever available. Secondary outcomes were all-cause discontinuation, weight gain, use
of antiparkinson drugs as a measure of extrapyramidal
side-effects, prolactin increase, QTc prolongation, and
sedation. Studies in which antiparkinson drugs were
given prophylactically were excluded from the analysis of
extrapyramidal side-effects. Because multiple-treatments
meta-analysis requires reasonable homogeneity we
focused on acute treatment, which we defined as 6-weeks
duration. If 6-week data were not available, we used data
from between 4 and 12 weeks (the datapoint closest to
6 weeks was given preference).
Study selection and data extraction were done independently by at least two of eight reviewers (FR, DÖ, SL,
LS, AC, MS, MPS, and BL). Data extraction forms were
sent to original authors of trial reports when necessary
with a request to provide missing data and the option to

CLO
–0·22
(–0·41 to
–0·04)
–0·29
(–0·44 to
–0·14)
–0·32
(–0·47 to
–0·16)
–0·38
(–0·57 to
–0·20)
–0·39
(–0·60 to
–0·19)
–0·43
(–0·58 to
–0·28)
–0·44
(–0·61 to
–0·28)
–0·45
(–0·62 to
–0·28)
–0·49
(–0·68 to
–0·30)
–0·49
(–0·66 to
–0·31)
–0·50
(–0·67 to
–0·33)
–0·50
(–0·69 to
–0·30)
–0·55
(–0·74 to
–0·36)
–0·55
(–0·73 to
–0·38)
–0·88
(–1·03 to
–0·73)

make corrections. Missing standard deviations were
estimated from p values or with the mean standard
deviation of the other included studies.27

Statistical analysis
Multiple-treatments meta-analysis combines direct and
indirect evidence for all relative treatment effects and
provides estimates with maximum power.28–31 The model
was fitted into a Bayesian context with hierarchical
models (appendix pp 66–69). A common heterogeneity
parameter was assumed for all comparisons. For
continuous outcomes, the relative effect sizes were
calculated as standardised mean differences (Hedges’ g).
For binary outcomes, relative effect sizes were calculated
as odds ratios (ORs). Both types of effect sizes are
reported with their 95% credible intervals (CrIs). To rank
the treatments we used the surface under the cumulative
ranking (SUCRA) probabilities.31 SUCRAs expressed as
percentages compare each intervention to an imaginary
intervention that is always the best without uncertainty.

1·10
0·57
0·87
0·97
0·70
0·76
0·76
0·60
0·65
0·71
0·68
0·61
0·67
0·46
1·00
(0·69 to 1·69) (0·68 to 1·43) (0·59 to 1·22) (0·63 to 1·42) (0·39 to 1·16) (0·40 to 0·82) (0·50 to 1·10) (0·51 to 1·09) (0·38 to 0·89) (0·43 to 0·95) (0·48 to 1·01) (0·43 to 1·01) (0·39 to 0·90) (0·45 to 0·99) (0·32 to 0·65)
0·93
0·81
0·90
0·66
0·53
0·70
0·71
0·56
0·60
0·67
0·63
0·56
0·63
0·43
(0·69 to 1·22) (0·60 to 1·08) (0·62 to 1·24) (0·37 to 1·10) (0·40 to 0·70) (0·51 to 0·95) (0·51 to 0·96) (0·38 to 0·78) (0·43 to 0·83) (0·44 to 0·95) (0·43 to 0·89) (0·39 to 0·79) (0·44 to 0·87) (0·32 to 0·57)

AMI
–0·07
(–0·19 to 0·05)

OLA

–0·09
–0·03
(–0·21 to 0·03) (–0·10 to 0·04)
–0·16
(–0·32 to
–0·00)

0·87
0·97
0·71
0·58
0·76
0·76
0·60
0·65
0·72
0·68
0·61
0·68
0·46
(0·76 to 1·01) (0·78 to 1·20) (0·43 to 1·13) (0·50 to 0·66) (0·63 to 0·91) (0·64 to 0·90) (0·47 to 0·76) (0·53 to 0·79) (0·54 to 0·94) (0·53 to 0·86) (0·47 to 0·77) (0·54 to 0·84) (0·41 to 0·52)
RIS

–0·09
–0·07
(–0·21 to 0·02) (–0·19 to 0·06)

1·12
0·82
0·66
0·87
0·88
0·69
0·75
0·83
0·78
0·70
0·78
0·53
(0·88 to 1·40) (0·49 to 1·29) (0·58 to 0·76) (0·73 to 1·04) (0·72 to 1·06) (0·53 to 0·88) (0·61 to 0·91) (0·61 to 1·08) (0·60 to 1·01) (0·53 to 0·89) (0·62 to 0·96) (0·46 to 0·60)
PAL

–0·17
–0·10
–0·08
0·01
(–0·38 to 0·04) (–0·29 to 0·08) (–0·26 to 0·11) (–0·22 to 0·20)

Treatment

–0·21
(–0·32 to
–0·09)
–0·22
(–0·36 to
–0·08)
–0·23
(–0·37 to
–0·08)
–0·27
(–0·43 to
–0·10)
–0·26
(–0·41 to
–0·12)
–0·27
(–0·47 to
–0·08)
–0·27
(–0·45 to
–0·10)
–0·33
(–0·50 to
–0·16)
–0·33
(–0·48 to
–0·18)
–0·66
(–0·78 to
–0·53)

–0·14
(–0·21 to
–0·08)
–0·15
(–0·25 to
–0·06)
–0·16
(–0·25 to
–0·07)
–0·20
(–0·33 to
–0·06)
–0·20
(–0·29 to
–0·10)
–0·21
(–0·37 to
–0·05)
–0·21
(–0·34 to
–0·08)
–0·26
(–0·39 to
–0·13)
–0·26
(–0·38 to
–0·15)
–0·59
(–0·65 to
–0·53)

–0·11
(–0·18 to
–0·05)
–0·13
(–0·22 to
–0·03)
–0·13
(–0·23 to
–0·03)
–0·17
(–0·31 to
–0·04)
–0·17
(–0·27 to
0·07)
–0·18
(–0·34 to
–0·02)
–0·18
(–0·32 to
–0·04)
–0·23
(–0·37 to
–0·10)
–0·24
(–0·35 to
–0·12)
–0·56
(–0·63 to
–0·50)

Efficacy (SMD with 95% Crl)

0·74
0·60
0·79
0·79
0·63
0·68
0·75
0·71
0·63
0·70
0·48
(0·43 to 1·20) (0·48 to 0·75) (0·61 to 1·01) (0·61 to 1·02) (0·46 to 0·85) (0·52 to 0·88) (0·53 to 1·02) (0·52 to 0·95) (0·47 to 0·85) (0·53 to 0·93) (0·39 to 0·58)
ZOT

–0·05
–0·04
(–0·16 to 0·08) (–0·21 to 0·14)

0·86
1·13
1·14
0·90
0·97
1·07
1·02
0·91
1·01
0·69
(0·51 to 1·32) (0·66 to 1·78) (0·67 to 1·81) (0·51 to 1·46) (0·56 to 1·55) (0·61 to 1·71) (0·58 to 1·65) (0·51 to 1·47) (0·58 to 1·61) (0·41 to 1·07)
HAL

–0·06
–0·05
–0·01
(–0·19 to 0·08) (–0·24 to 0·14) (–0·10 to 0·08)

1·32
1·33
1·05
1·13
1·25
1·19
1·06
1·17
0·80
(1·11 to 1·57) (1·11 to 1·57) (0·82 to 1·31) (0·93 to 1·35) (0·93 to 1·63) (0·92 to 1·50) (0·82 to 1·34) (0·95 to 1·43) (0·71 to 0·90)
QUE

–0·07
–0·06
–0·02
–0·01
(–0·20 to 0·08) (–0·25 to 0·14) (–0·12 to 0·08) (–0·12 to 0·11)

1·01
0·80
0·86
0·95
0·90
0·81
0·61
0·89
(0·80 to 1·25) (0·60 to 1·04) (0·68 to 1·07) (0·69 to 1·26) (0·68 to 1·19) (0·61 to 1·03) (0·70 to 1·13) (0·52 to 0·71)
ARI

–0·10
–0·09
–0·06
–0·04
–0·04
(–0·27 to 0·07) (–0·31 to 0·12) (–0·19 to 0·07) (–0·19 to 0·10) (–0·19 to 0·11)

0·80
0·86
0·95
0·90
0·80
0·61
0·89
(0·59 to 1·04) (0·68 to 1·07) (0·69 to 1·27) (0·68 to 1·18) (0·6 to 1·05) (0·69 to 1·14) (0·51 to 0·72)
SER

–0·10
–0·09
–0·05
–0·04
–0·04
0·00
(–0·24 to 0·04) (–0·29 to 0·11) (–0·15 to 0·04) (–0·16 to 0·08) (–0·16 to 0·09) (–0·15 to 0·16)

1·09
1·21
1·14
1·02
0·78
1·13
(0·81 to 1·45) (0·84 to 1·69) (0·81 to 1·56) (0·73 to 1·39) (0·83 to 1·52) (0·61 to 0·98)
ZIP

–0·11
–0·10
–0·07
–0·05
–0·05
–0·01
–0·01
(–0·30 to 0·08) (–0·32 to 0·11) (–0·22 to 0·09) (–0·22 to 0·11) (–0·22 to 0·13) (–0·21 to 0·19) (–0·19 to 0·16)

1·11
1·06
0·94
0·72
1·05
(0·80 to 1·50) (0·78 to 1·41) (0·70 to 1·24) (0·81 to 1·33) (0·59 to 0·86)
CPZ

–0·11
–0·10
–0·07
–0·05
–0·05
–0·01
–0·01
0·00
(–0·28 to 0·05) (–0·32 to 0·11) (–0·20 to 0·07) (–0·20 to 0·09) (–0·20 to 0·10) (–0·19 to 0·17) (–0·17 to 0·14) (–0·20 to 0·20)
–0·17
(–0·33 to
–0·00)
–0·17
(–0·32 to
–0·02)
–0·50
(–0·60 to
–0·39)

0·96
0·86
0·65
0·96
(0·66 to 1·34) (0·61 to 1·19) (0·68 to 1·32) (0·50 to 0·84)
ASE

–0·16
–0·12
–0·11
–0·10
–0·06
–0·07
–0·05
–0·05
(–0·37 to 0·06) (–0·25 to 0·01) (–0·25 to 0·03) (–0·25 to 0·05) (–0·24 to 0·11) (–0·22 to 0·09) (–0·25 to 0·14) (–0·23 to 0·12)
–0·16
(–0·36 to 0·04)
–0·49
(–0·66 to
–0·31)

–0·12
(–0·23 to
–0·02)
–0·45
(–0·51 to
–0·39)

0·91
0·69
1·01
(0·64 to 1·22) (0·73 to 1·36) (0·54 to 0·86)
LUR

–0·11
–0·10
–0·07
–0·07
–0·06
–0·06
0·00
(–0·24 to 0·02) (–0·24 to 0·03) (–0·23 to 0·10) (–0·20 to 0·06) (–0·24 to 0·13) (–0·22 to 0·11) (–0·16 to 0·16)
–0·44
(–0·52 to
–0·35)

–0·43
(–0·52 to
–0·34)

–0·39
(–0·52 to
–0·26)

–0·39
(–0·49 to
–0·30)

–0·38
(–0·54 to
–0·23)

–0·38
(–0·51 to
–0·25)

–0·33
(–0·45 to
–0·21)

1·12
0·77
(0·83 to 1·50) (0·61 to 0·96)
ILO

0·69
(0·56 to 0·84)

–0·33
(–0·43 to
–0·22)

PBO

All cause discontinuation (OR with 95% Crl)

Figure 2: Efficacy and all-cause discontinuation of antipsychotic drugs
Drugs are reported in order of efficacy ranking. Comparisons between treatments should be read from left to right and the estimate is in the cell in common between the column-defining treatment and
the row-defining treatment. For efficacy, standard mean differences (SMDs) lower than 0 favour the column-defining treatment. For all-cause discontinuation, odds ratios (ORs) higher than 1 favour the
column-defining treatment. To obtain SMDs for comparisons in the opposite direction, negative values should be converted into positive values, and vice versa. To obtain ORs for comparisons in the
opposite direction, reciprocals should be taken. Significant results are in bold and underlined. CLO=clozapine. AMI=amisulpride. OLA=olanzapine. RIS=risperidone. PAL=paliperidone. ZOT=zotepine.
HAL=haloperidol. QUE=quetiapine. ARI=aripiprazole. SER=sertindole. ZIP=ziprasidone. CPZ=chlorpromazine. ASE=asenapine. LUR=lurasidone. ILO=iloperidone. PBO=placebo.

www.thelancet.com Published online June 27, 2013 http://dx.doi.org/10.1016/S0140-6736(13)60733-3

3