A nxiety D isorders can Network MetaAnalysis inform guidelines April Slee PRIMENT 1 Coauthors Irwin Nazareth Paulina Bondaronek Yifeng Liu and Zhihang Cheng and Nick Freemantle ID: 932225
Download Presentation The PPT/PDF document "Pharmacotherapy for Generalised" is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.
Slide1
Pharmacotherapy for Generalised Anxiety Disorders – can Network Meta-Analysis inform guidelines?April Slee, PRIMENT
1
Co-authors: Irwin Nazareth, Paulina Bondaronek, Yifeng Liu and Zhihang Cheng and Nick Freemantle
1
Slide2DisclosuresICH department of neurology
Spirox, Inc.
Reflow Medical, Inc.Johnson & Johnson
Providence
2
2
Slide3BackgroundCentral problem in comparative effectiveness research is a lack of active comparison trials6 times as many trials funded by industry vs. government on clinicaltrials.gov1Placebo-controlled trials require smallest sample size
Strategic considerations impact choice of comparisons for active-controlled trialsMixed treatment comparisons can estimate relative treatment effects across trials
1Ehrhardt S, Appel LJ, Meinert CL. JAMA. 2015 Dec 15;314(23):2566-7.
3
Slide4What is a Mixed Treatment Comparison (MTC)
Conventional
meta analyses
combine trials making similar comparisons
Mixed Treatment Comparisons
(AKA Network Meta Analyses) join up the links to provide
fully conditional estimates of effect
4
Slide5B-P Effect
Direct and Indirect Comparisons
Drug APlacebo
Drug B
Placebo
Improvement
A-P
Effect
Direct Comparison
Direct Comparison
A-B
Effect
Indirect Comparison
Trial 1
Trial 2
5
Slide6Mixed Treatment Comparison
Direct
Comparison
Direct
Comparison
Indirect
Comparison
Drug A
Drug B
Placebo
Placebo
Direct
Comparison
Direct
Comparison
(common reference)
Drug B
Drug A
Placebo
Drug D
Drug E
Drug H
Drug
F
Drug G
Indirect analysis
Head-to-head study
6
Slide7Mixed treatment comparisons for generalised anxiety7
Slide8Main subtypes of anxiety disorders
Generalised Anxiety
Long-lasting, not specific
Obsessive Compulsive
Repetitive obsessions and compulsions
Panic Disorder
Terror, shaking, trouble breathing
Post-Traumatic Stress
Anxiety from traumatic experience
Social Anxiety
Intense fear of social interactions
Specific Phobia
Specific animals, objects, situations
8
Slide9Brief Introduction to GAD (DSM IV/V)Excessive worry for at least 6 months that is:Difficult to control
Causing clinically significant distressNot due to a substance
Not due to another mental disorderSymptoms for at least 6 months:At least 3 of the following: restlessness, fatigue, difficulty concentrating, irritability, muscle tension, sleep disturbance
Association, American Psychiatric (2013). Diagnostic and statistical manual of mental disorders : DSM-5. (5th ed.). Washington, D.C.: American Psychiatric Association. p. 222.
ISBN
978-0-89042-554-1
.
20
9
Slide1010
Slide11Records identified through other industry registries, clinicaltrials.gov (n = 107)
Records identified through database searching
(n = 1,885)
Potential studies (n= 1,992)
Records excluded (n=421)
Prior to 1994 (n=380)
Duplicates (n=41)
Records screened (n=1,571)
Records excluded by title and abstract review (n=1,066)
Records excluded (n=416
)*
Duplicates
(n=99)
Records
not meeting inclusion
criteria
(n=416
)
Studies Included (n=89)
*Not
DSM-IV, IV-TR, V, ICD-10 or CCMD-3 GAD (
n=33), Not
randomised trial (n=215
), <
2 pharmacologic treatments (
n=17),
Major
comorbidities/refractory (
n=22), Relapse prevention/discontinuation design (
n=10), Other (n=20)
Full-text review (n=505)
11
Slide12AgomelatineBenzodiazepine
Bupropion
BuspironeCitalopram
Duloxetine
Escitalopram
Fluoxetine
Hydroxyzine
Imipramine
Maprotiline
Mirtazapine
Ocinaplon
Opipramol
Paroxetine
Pregabalin
Quetiapine
Sertraline
Tiagabine
Venlafaxine
Vilazodone
Vortioxetine
Placebo
12
Slide13Outcomes: Anxiety (HAM-A), % CompleteMTC Analysis with WINBUGS
Prior belief,
dist’n
of
a
(specify)
Posterior belief,
dist’n of
a
given X
Data = x
(study estimates)
Model L (
a
; x)
Likelihood of
a
given X
13
Slide14Mean Difference (95%
CrI
)
-10
-9
-8
-7
-6
-5
-4
-3
-2
-1
0
1
2
3
4
Drug
Mean (
95%
CrI
)
Favours Placebo
Favours Active Drug
Duloxetine (T:8, P:1355)
Escitalopram (T:13, P:1581)
Fluoxetine (T:8, P:264)
Hydroxyzine (T:2, P:187)
Imipramine (T:1, P:26)
Maprotiline
(T:1, P:30)
Citalopram (T:2, P:37)
Bupropion (T:2, P:41)
Benzodiazepine (T:15, P:1019)
Agomelatine
(T:2, P:202)
Buspirone
(T:6, P:311)
Opipramol
(T:2, P:144)
Ocinaplon
(T:1, P:31)
Mirtazapine (T:10, P:318)
Paroxetine (T:17, P:1862)
Pregabalin
(T:11, P:1957)
Quetiapine (T:4, P:1804)
Sertraline (T:6, P:485)
Tiagabine
(T:5, P:1292)
Venlafaxine (T:14, P:2275)
Vilazodone
(T:3, P:866)
Vortioxetine
(T:4, P:1074)
3
11
(
2
11, 4
11
)
2
38
(
1
57, 3
19)
2
36
(
1
08, 3
64
)
3
03 (0
99, 5
04)
0
58
(-
2
34
,
3
49)
-
0
35
(-
3
78
,
3
11)
2
05
(-
0
09
,
4
19
)
4
59 (1
68, 7
54)
2
39
(
1
46, 3
31
)
3
52
(
1
26, 5
81)
2
38
(
0
91, 3
83
)
2
02
(-
0
23, 4
27)
7
81
(
0
85, 14
68)
3
16
(
1
81
,
4
52)
2
29
(
1
50
,
3
09
)
2
87
(
1
95
,
3
80)
3
60
(
2
41, 4
80)
2
69
(
1
34
,
4
03)
0
79
(-
0
53, 2
14
)
2
67
(
1
87, 3
47
)
1
45
(-
0
09
,
2
99)
1
11
(-
0
21, 2
46
)
Efficacy (
D
HAM-A vs. Placebo)
14
Slide15Favours Placebo
Favours Active Drug
Opipramol (T:2, P:144)
0
30
(
0
07, 1
10
)
0.50
0.71
1.0
1.41
Odds Ratio (95%
CrI
)
Drug
OR (95
%
CrI
)
Agomelatine
(T:2, P:202)
0
67
(
0
38, 1
17
)
Benzodiazepine (T:15, P:1019)
1
44
(
1
12
,
1
86)
Bupropion (T:2, P:41)
1
01
(
0
10, 12
77)
Buspirone
(T:6, P:311)
0
76
(
0
46, 1
23)
Citalopram (T:2, P:37)
3
38
(
0
71, 23
84)
Duloxetine (T:8, P:1355)
1
09
(
0
88, 1
34
)
Escitalopram (T:13, P:1581)
0
94
(
0
77
,
1
14
)
Fluoxetine (T:8, P:264)
1
28
(
0
55, 2
90)
Hydroxyzine (T:2, P:187)
0
98
(
0
56
,
1
69
)
Mirtazapine (T:10, P:318)
3
19
(
0
59, 22
43)
Maprotiline (T:1, P:30)
2
34
(
0
18, 33
71)
Imipramine (T:1, P:26)
2
78
(
0
70, 12
24
)
Ocinaplon (T:1, P:31)
0
73
(
0
24
,
2
18)
Sertraline (T:6, P:485)
0
94
(
0
65
,
1
36)
Quetiapine (T:4, P:1804)
1
43
(
1
15
,
1
80)
Pregabalin (T:11, P:1957)
0
82
(
0
66
,
1
00
)
Paroxetine (T:17, P:1862)
1
24
(
1
02
,
1
49)
Venlafaxine (T:14, P:2275)
0
98
(
0
83
,
1
15
)
Tiagabine (T:5, P:1292)
1
37
(
1
09
,
1
74
)
Vilazodone (T:3, P:866)
1
60
(
1
20, 2
13)
Vortioxetine (T:4, P:1074)
0
88
(
0
68, 1
15
)
Safety (Discontinuation vs. Placebo)
15
Slide16Efficacy vs. Vortioxetine
Drug
Mean (95%CI)
-0.5
0
0.5
1
1.5
2
2.5
3
3.5
4
Agomelatine
2.40 (-0.22, 5.03)
Benzodiazepine
1.28 (-0.36, 2.88)
Bupropion
3.48 (0.27, 6.71)
Buspirone
1.27 (-0.71, 3.22)
Duloxetine
2.00 (0.41, 3.55)
Escitalopram
1.27 (-0.31, 2.82)
Fluoxetine
1.25 (-0.58, 3.08)
Hydroxyzine
1.91 (-0.52, 4.32)
Mirtazapine
2.04 (0.15, 3.94)
Ocinaplon
6.69 (-0.35, 13.7)
Paroxetine
1.18 (-0.37, 2.73)
Pregabalin
1.75 (0.13, 3.37)
Quetiapine
2.48 (0.68, 4.27)
Sertraline
1.58 (-0.32, 3.46)
Venlafaxine
1.56 (0.00, 3.09)
Favours
Comparator
Favours
Vortioxetine
16
Slide17Summary: Pros and Cons of MTC
Pros
Makes use of all evidence (direct and indirect)
Useful when there are many available therapies so unlikely all pairwise trials performed
Useful when new therapies tested against placebo
Can use a range of endpoints (differences, ratios)
Cons
Bayesian model, requires specification of priors
Standard meta analysis assumption of similarity and homogeneity
Additional assumption: direct and indirect evidence is estimating the same effect
17
Slide18Thanks for your attention!Questions?18
Slide19There are 3 assumptions central to the validity of MTC:
Similarity
(transitivity):
Trials are similar concerning moderators of the relevant treatment effect. Ideally, any subject who is eligible for 1 trial should be eligible for any trial. Factors such as disease severity and study duration should not differ in ways that impact the treatment effect. The similarity assumption is required to make use of indirect comparisons.
Homogeneity:
Trials are sufficiently homogeneous to be quantitatively combined, that is, estimates of treatment effects across studies are estimating the same effect.
Consistency:
Indirect evidence is consistent with direct evidence, that is, direct and indirect evidence are estimating the same effect
.
Song F1,
Loke
YK, Walsh T, Glenny AM, Eastwood AJ, Altman
DG.Methodological
problems in the use of indirect comparisons for evaluating healthcare interventions: survey of published systematic reviews. BMJ. 2009 Apr 3;338:b1147.
19
Slide20Model detailsProgramming as in the NICE DSU Technical Support Document 2 was used with modifications to the priors. 100,000 sample iterations with an initial burn-in period of 100,000 iterations,
Univariate random effects to allow for heterogeneity between studies in the treatment comparison effects.
Diagnostics: autocorrelation plots for convergence and examination of Markov traces. Pairwise meta-analysis to check consistency of direct and indirect evidence
20