Br J Dermatol 2024; 190:184–190

https://doi.org/10.1093/bjd/ljad393
Advance access publication date: 13 October 2023 Systematic Review

Comparing binary efficacy outcomes for systemic

immunomodulatory treatments for atopic dermatitis in a
living systematic review and network meta-analysis
Aaron M. Drucker,1 Megan Lam,1 Rawaan Elsawi, 2 David Prieto-Merino,3 Rayka Malek,4
Alexandra G. Ellis,5 Zenas Z.N. Yiu ,6 Bram Rochwerg,7 Sonya Di Giorgio,8 Bernd W.M. Arents ,9
Tim Burton,10 Phyllis I. Spuls,11 Jochen Schmitt12 and Carsten Flohr 13

Downloaded from https://academic.oup.com/bjd/article/190/2/184/7311111 by guest on 07 June 2025

1Division of Dermatology, Department of Medicine, University of Toronto and Department of Medicine and Women’s College Research
Institute, Women’s College Hospital, Toronto, ON, Canada
2 Faculty of Medicine, University of Ottawa, Ottawa, ON, Canada
3 Faculty of Medicine, Universidad de Alcalá, Alcalá de Henares, Spain
4 School of Life Course and Population Health Sciences, King’s College London, London, UK
5 Brown University, Providence, RI, USA
6 Dermatology Centre, Salford Royal NHS Foundation Trust, The University of Manchester, Manchester Academic Health Science Centre,

NIHR Manchester Biomedical Research Centre, Manchester, UK

7 Departments of Medicine and Health Research Methods, Evidence and Impact, McMaster University, Hamilton, ON, Canada
8 Libraries and Collections, King’s College London, London, UK
9 Dutch Association for People with Atopic Dermatitis (VMCE), Nijkerk, the Netherlands
10 Patient Representative (independent), Nottingham, UK
11Department of Dermatology, Amsterdam Public Health/Infection and Immunology, Amsterdam, the Netherlands
12 Center for Evidence-Based Healthcare, Faculty of Medicine Carl Gustav Carus, Technische Universität (TU) Dresden, Dresden, Germany
13 Unit for Paediatric and Population-Based Dermatology Research, St John’s Institute of Dermatology, King’s College London, London, UK

Correspondence: Aaron M. Drucker. Email: aaron.drucker@wchospital.ca

Abstract
Background Systemic treatments for atopic dermatitis (AD) are evaluated primarily in placebo-controlled trials with binary efficacy out-
comes. In a living systematic review and network meta-analysis (NMA), we previously analysed continuous efficacy measures.
Objectives To compare binary efficacy outcomes of systemic treatments for AD.
Methods We searched the Cochrane Central Register of Controlled Trials (CENTRAL), MEDLINE, Embase, Latin American and Caribbean
Health Science Information (LILACS) database, Global Resource for Eczema Trials (GREAT) database and trial registries up to 1 March 2023.
We included randomized trials examining ≥ 8 weeks of treatment with systemic immunomodulatory medications for moderate-to-severe AD.
We screened titles, abstracts and full texts and abstracted data independently, in duplicate. Outcomes included the proportion of patients
achieving at least 50%, 75% and 90% improvements in Eczema Area and Severity Index (EASI 50, EASI 75 and EASI 90, respectively) and
Investigator Global Assessment (IGA) success. We performed random-effects Bayesian NMAs to calculate odds ratios (OR) and 95% credible
intervals (CrIs) between each intervention for each outcome.
Results Eighty-three trials with 22 122 participants were included in the systematic review. In analyses limited to trials of 8–16 weeks’ duration
with predominantly adult populations, abrocitinib 200 mg daily (OR 1.5, 95% CrI 1.1–2.2) and upadacitinib 15 mg daily (OR 1.7, 95% CrI 0.9–3.3)
and 30 mg daily (OR 2.5, 95% CrI 1.3–5.0) were associated with higher odds of achieving EASI 50 vs. dupilumab. Abrocitinib 100 mg daily (OR
0.7, 95% CrI 0.5–1.0), baricitinib 2 mg daily (OR 0.4, 95% CrI 0.3–0.5) and 4 mg daily (OR 0.5, 95% CrI 0.3–0.7), and tralokinumab (OR 0.4, 95%
CrI 0.3–0.6) were associated with lower odds of achieving EASI 50 vs. dupilumab. Results were similar for EASI 75, EASI 90 and IGA success.
Conclusions Supporting results for continuous outcome measures, upadacitinib 30 mg daily and abrocitinib 200 mg daily are the most ef-
ficacious with regard to binary efficacy endpoints up to 16 weeks in adults with moderate-to-severe AD, followed by upadacitinib 15 mg daily,
dupilumab and abrocitinib 100 mg daily. Dupilumab and both doses of upadacitinib and abrocitinib are more efficacious than baricitinib 4 and
2 mg daily and tralokinumab.

What is already known about this topic?

• A living systematic review and network meta-analysis (NMA) of systemic treatments for atopic dermatitis (AD) demonstrated that high-
dose upadacitinib and abrocitinib were more effective than dupilumab, which was more effective than tralokinumab and baricitinib.
• That analysis used continuous outcome measures (e.g. change from baseline in EASI score), and differences between medications –
although often statistically significant – were generally within published minimal important difference thresholds.

Accepted: 4 October 2023

© The Author(s) 2023. Published by Oxford University Press on behalf of British Association of Dermatologists. This is an Open Access article distrib-
uted under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted reuse,
distribution, and reproduction in any medium, provided the original work is properly cited.
Relative efficacy of systemic treatments for eczema, A.M. Drucker et al. 185

What does this study add?

• In this NMA, upadacitinib 30 mg daily and abrocitinib 200 mg daily were the most efficacious with regard to a least 50%, 75%
and 90% improvements in Eczema Area and Severity Index (EASI 50, EASI 75 and EASI 90, respectively) and Investigator Global
Assessment success, followed by upadacitinib 15 mg daily, dupilumab and abrocitinib 100 mg daily. All three agents are more effica-
cious than baricitinib 4 mg and 2 mg daily and tralokinumab.
• Binary outcomes of systemic treatments for AD among adults support previous findings from continuous outcomes.

Systemic immunomodulatory medications for atopic der- marked as relevant by a single reviewer was advanced to
matitis (AD) are primarily studied in placebo-controlled full-text screening and discrepancies on full-text screening,
trials, with few head-to-head comparisons. To facilitate data abstraction or risk-of-bias assessment were resolved
comparisons between treatment options, with rapid by discussion between the reviewers, with adjudication by
incorporation of new evidence as it becomes available, a senior team member (C.F.), if needed. We emailed the

Downloaded from https://academic.oup.com/bjd/article/190/2/184/7311111 by guest on 07 June 2025

we established a living systematic review and network corresponding authors of primary trial publications to obtain
meta-analysis (NMA) of systemic immunomodulatory data for outcomes not included in journal publications, con-
treatments for AD.1–3 In analyses to date, we have assessed ference abstracts or trial registries.
and reported relative efficacy with regard to change in In addition to the continuous outcome measures reported
continuous measures of signs, symptoms and quality of previously, we abstracted data reporting on the proportion
life, in line with recommendations from the Harmonizing of patients in each trial arm with EASI 50, EASI 75 and EASI
Outcome Measures for Eczema initiative. 4 This enables 90 and IGA success. We retroactively abstracted this data
robust comparisons with reference to minimal clinically for all previously included trials. Studies utilize different IGAs
important differences for established outcome measures and report IGA success differently;8 when available, we pref-
but has led to the exclusion of studies that do not report erentially abstracted data on the proportion of patients in
adequate information on continuous outcomes. 5 Instead, each arm with an IGA score of 0 or 1 and at least a 2-point
most trials for new medications focus their reporting improvement from baseline, as this definition of success is
on binary outcomes, including primary outcomes man- recommended by the FDA and is the most widely reported
dated by the US Food and Drug Administration (FDA) and in clinical trials.
European Medicines Agency. We performed random-effects NMAs for each binary
The objective of this study was to compare the relative outcome within a Bayesian framework using the ‘gemtc’
efficacy of systemic immunomodulatory medications for package in r [R Foundation for Statistical Computing,
AD with regard to the proportion of patients with at least Vienna, Austria; the statistical code is provided in Appendix
50%, 75% and 90% improvements in the Eczema Area and S2 (see Supporting Information)]. We generated network
Severity Index (EASI 50, EASI 75 and EASI 90, respectively) plots with the nodes being each treatment-dose regimen in
and Investigator Global Assessment (IGA) success in rand- the trials. We examined network plots and described their
omized clinical trials (RCTs). overall geometry. We calculated the odds ratios (OR) and
95% credible intervals (CrIs) between each pair of nodes
in the network. We used a normal prior distribution on the
Materials and methods log ORs such that the 95% coverage included log(1/30) to
log(30), based on the assumption that ORs are very unlikely
This is an update of a living systematic review, iterations of to be > 30 or < 1/30.9,10 We assessed the convergence of
which have been previously published,2,3 with the addition four chains using the Gelman–Rubin–Brooks plot of the
of results for newly added outcomes not included in the Gelman–Rubin shrink factor and by visual inspection of trace
original protocol (PROSPERO CRD42018088112).1 We used plots and posterior density distributions. We summarized
the PRISMA extension for NMA reporting guidance.6 treatment rankings using surface under the cumulative rank-
We searched the Cochrane Central Register of Controlled ing (SUCRA) curves.11
Trials (CENTRAL), MEDLINE, Embase, the Latin American We separated trials with adult participants (≥ 18 years)
and Caribbean Health Science Information (LILACS) data- from those with children, but we included trials with pre-
base, the Global Resource of Eczema Trials (GREAT) data- dominantly adult populations in the adult analysis even if
base and trial registries from inception, with updates every they included some adolescents. We calculated ORs for
4 months. This update includes trials found up to 1 March all pairwise comparisons in each network. As our main fig-
2023. The complete search strategy is provided in Appendix ures, we used forest plots with relative effect estimates
S1 (see Supporting Information). We included RCTs eval- comparing abrocitinib, baricitinib, upadacitinib and traloki-
uating systemic immunomodulatory treatments given for numab against dupilumab as the main referent; these are
at least 8 weeks for AD. We excluded studies terminated the approved targeted systemic immunomodulatory agents
early (new criteria implemented in most recent search). currently in widespread use in Europe and North America,
Abstract and full-text screening, data abstraction and risk- and dupilumab was the first targeted agent approved for
of-bias assessments were done by two investigators inde- AD. League tables are provided for all pairwise ­comparisons
pendently, in duplicate (A.M.D. along with either M.L. or between dupilumab, abrocitinib, baricitinib, u ­ padacitinib,
R.E.). We conducted risk-of-bias assessment at the trial level tralokinumab and placebo, as well as l­ebrikizumab,
using the Cochrane Risk of Bias tool.7 Any title or abstract which may be approved soon (Tables S1–S4; see Supporting
186 Relative efficacy of systemic treatments for eczema, A.M. Drucker et al.

Information). Complete league tables for all interventions in Results

each network are available in Tables S5–S9 (see Supporting
Information). As of 1 March 2023, there were 83 trials with 22 122 partici-
We assessed coherence using node-splitting to com- pants included in our systematic review (Figure 1). Up-to-date
pare direct and indirect evidence.12 If incoherence was trial characteristics for the living systematic review, including
present, we described it qualitatively. We conducted sub- extracted outcomes and risk of bias assessments, can be
group analyses separating trials that allowed and those downloaded from www.eczematherapies.com/research.
that did not allow concomitant topical anti-inflammatory Most trials included in analyses for EASI 50, EASI 75, EASI
therapy (e.g. topical corticosteroids). We conducted sen- 90 and IGA success are for targeted systemic agents (includ-
sitivity analyses including only trials with a low risk of bias ing those approved for use in moderate-to-severe AD, as
(no items scored as having an unclear or high risk of bias well as other experimental agents) compared with placebo.
on the Cochrane Risk of Bias tool).7 As we considered Most trials require a baseline IGA score of 3/4 or 4/4 and
binary efficacy outcomes complimentary to our main con- an EASI score of 16 for participants to be eligible, although
tinuous outcomes, we did not create GRADE informative some use slightly different EASI criteria (e.g. EASI of 12).
statements for them. We did not perform statistical tests There were insufficient data for children or for treatment >16

Downloaded from https://academic.oup.com/bjd/article/190/2/184/7311111 by guest on 07 June 2025

for publication bias. weeks duration to conduct separate NMAs for those groups.

2085 references imported for screening 126 duplicates removed

1 Nov, 2021 (n=444) 1 Nov, 2021 (n=27)
3 March, 2022 (n=414) 3 March, 2022 (n=16)
5 July, 2022 (n=393) 5 July, 2022 (n=30)
7 Nov, 2022 (n=444) 7 Nov, 2022 (n=17)
1 March, 2023 (n=390) 1 March, 2023 (n=36)

1959 tles and abstracts screened 1371 records excluded

1 Nov, 2021 (n=417) 1 Nov, 2021 (n=270)
3 March, 2022 (n=398) 3 March, 2022 (n=302)
5 July, 2022 (n=363) 5 July, 2022 (n=256)
7 Nov, 2022 (n=427) 7 Nov, 2022 (n=310)
1 March, 2023 (n=354) 1 March, 2023 (n=233)

566 records excluded or already included under a separate record

1 Nov, 2021 (n=143)
3 March, 2022 (n=93)
588 full-text items assessed for eligibility 5 July, 2022 (n=101)
1 Nov, 2021 (n=147) 7 Nov, 2022 (n=112)
3 March, 2022 (n=96) 1 March, 2023 (n=117)
5 July, 2022 (n=107) Reasons:
7 Nov, 2022 (n=117) 1 Wrong paent populaon
1 March, 2023 (n=121) 4 Wrong outcomes
10 Wrong study design
1 Wrong intervenon
116 Study already included under a separate publicaon or duplicate
15 Trial registry - Wrong paent populaon
23 new studies included 2 Trial registry – Wrong outcomes
1 Nov, 2021 (n=4) 144 Trial registry - Wrong study design
3 March, 2022 (n=3) 36 Trial registry - Wrong intervenon
5 July, 2022 (n=7a) 37 Trial registry - Study already included under a separate publicaon or duplicate
7 Nov, 2022 (n=5) 200 Trial registry - No results
1 March, 2023 (n=4)

astudies included in a single registry entry 0 Addional records idenfied through other sources

60 studies from last publicaonb

83 studies included (Search up to June 15, 2021)

Figure 1 PRISMA flowchart of the study screening and selection process for the living systematic review of systemic immunomodulatory
treatments for atopic dermatitis, including results up to 1 March 2023. This diagram includes screening, inclusion and exclusion for the overall
systematic review up to 1 March 2023, not just studies assessed for binary outcomes. aOne ClinicalTrials.gov entry (NCT03568162) includes two
separate clinical trials that met the inclusion criteria. bSearch up to 15 June 2021.
Relative efficacy of systemic treatments for eczema, A.M. Drucker et al. 187

Network plots for EASI 50, EASI 75, EASI 90 and IGA 30 mg daily (OR 2.5, 95% CrI 1.3–5.0) were associated with
success demonstrated a predominantly spoke-and-wheel higher odds of achieving EASI 50 vs. dupilumab 600 mg
pattern, with most active interventions connected to the then 300 mg every 2 weeks (Figures 2a, 3a). Abrocitinib
network through placebo (Figure 2). There are few head-to- 100 mg daily (OR 0.7, 95% CrI 0.5–1.0), baricitinib 2 mg
head studies comparing different active interventions. daily (OR 0.4, 95% CrI 0.3–0.5) and 4 mg daily (OR 0.5,
Abrocitinib 200 mg daily (OR 1.5, 95% CrI 1.1–2.2) and 95% CrI 0.3–0.7) and tralokinumab 600 mg then 300 mg
upadacitinib 15 mg daily (OR 1.7, 95% CrI 0.9–3.3) and every 2 weeks (OR 0.4, 95% CrI 0.3–0.6) were associated

(a) (b)

e weekly
Dupilumab 400
eekly
Q2W

Dupilumab 400 mg ×1 then 100 mg SC Q4W

Dupilumab

Dupilumab 600 mg then 300 mg Q4W

W
Dupilu

Dupilum

mg SC onc
SC Q4

ce w

Dup
then 200 mg SC

y
SC on

Dup

weekl
mab 60

ilum
0 mg

ab 600
600 mg ×1

ilum
ly

ab 60

mg ×1
eek
0 mg

once
mg then 200
0 mg

then 10

ab 6
ce w
GB

0 mg

2W
mg the
Etras

en 20

0 mg
0

then 200 mg SC
R8

×1 then

0m
g on

gQ
mg ×1
then 300 mg
Etr

×1 th
mg ×1
30

imod

2W
mg th

mab 30
g×

00 m
00 m

n 300 mg
asim
IS

10

Etr

OD
Q

1 th

ab 400
B

en 30
300 m
kg

Dupilumab 400

asim
1m

mg

ab 3
83

D
ab 3

GB
o

ab 400

400

O
en 3
d2

gO
m
01

Dupilu

gP
Etr
R
00
g OD
g

od
D

ilum
0 mg

OD
ilum
–1
20

83

Dupilum

72
mg

mab

asim
IS

4m
gO

00 m
g once

b3

Q4W
OD

1
Q2W
0m

0
IV
B

Q2W

O
Dupilum

10
10

mg
IS

ma

Dup
Dup
83

OD

gP
72

ib
od
O
Dupilu

Q2W
on

B
g th

g on

DS
D
0

kg
IS

gP

itin
u
83

m
OD
10
O

2m
15

B
OD

pil
day

weekly
en

m
01

m
PO

b2
83

Downloaded from https://academic.oup.com/bjd/article/190/2/184/7311111 by guest on 07 June 2025

DS
0

ric
g

ce w
Du

–1

gO
m

PO 20

4
06
6

ni
Ba
1
00
g

ib
0m

IV
00 g m

iti
and
th

D
n
IS 1

eek
mg

iti

ric
mg
en

gt

on
ISB 2 B i b
OD D

ric
83 he

Ba
tin
29

b
75

the O

da
Q

ni

ly
83 01 n

Ba
iti O i mg
4W

ric

y
n3
m

06
ric gP y 50 60

1
80
g

00 00
1m dail m 0m Ba

an
Q

mg mg Ba ice
g th 02
4W

gQ

d
ISB
inib N0 OD

29
the Q2 tw 830 en
n3 it
gP
O 75 4W AS mg
KH 00 W aric 600 m 60
K4 mg B 0m y mg gQ 02
083 Q4 st 4 dail the 4W N0
150 W ila ice n3 AS D
mg rem O tw ISB 00 gO
Q4 Ap m gP 830
600
m gQ 40
m
KH W 30 mg 2W 002
K40 ast N
83 3 mil then AS
00 m Apre D 300 OD
gQ O mg g PO
2W
30 mg PO KHK
Q 4W ib 30 m
KHK
4083 itinib 4083 ocitin
600 m Abroc 150 m Abr
g Q2W g Q4W OD
OD
0 mg PO 0 mg PO
itinib 20 inib 20
KHK4083 Abroc KHK4083 Abrocit
600 mg 300 mg
Q4W Q2W
OD
100 mg PO OD Abrocitinib
100 mg PO
Abrocitinib
Lebrikizumab 125 mg SC KHK4083 600 mg Q2W
1×
Abrocitinib 10 mg PO OD Abrocitinib 10 mg PO OD

Lebrikizumab 125 mg SC
Q4W KHK4083 600 mg Q4W
ZPL-3893787 ZPL-3893787
30 mg PO OD 30 mg PO OD
SC 1× mg SC 1×
ab 250 mg Ustekinu mab 125 Ustek
Lebrikizum mab 90 Lebrikizu inuma
Q4W mg SC b 90 mg
g SC at week SC at
125 m Ustek s 0 and SC Q4W Ustek weeks
0 and
then inum 4 5 mg inum 4
SC 1× Q4W ab 45 ab 12
0 mg g SC mg S ikizum
ab 45
ab 25 Lebr 1× mg
izum 50 m W C at SC SC at
Lebrik then
2
CQ
2 Up wee g Up wee
1× gS
ada
c itin
ks 0
and 250 m Q4
W ada ks 0
SC m ib 7 4 umab C citin and
mg 250 rikiz gS ib 7 4
500 en Up .5 m
Leb 25
m W Up .5 m
umab 2 th ada gP
OO n1 Q4 ada gP
rikiz 0a
nd
Q4
W cit
inib D the SC W cit OO
Leb 1× D
eek SC Up 30 SC mg Q2 Up
inib
Cw –1 Tr ad mg mg 50 C Tr 30
mg
S mg W a lo ac 50 n2 gS ad mg
kg Q4 ki itin PO
b2 the
alo ac
b 500 0.1 SC
nu ib OD ma 1× 5 0m ki
nu
itin PO
OD
ma m u ib
u ab –1 ab 15 kiz SC n2 m
4W

g bri e 15
kiz um mg

W
m 60 Le mg th ab
ri liz mg
Tr
Q

Leb kg

Q4
PO
mo 0 00 d2 60
4W

Te
a
SC

m PO
Ne b5 an 0
lo

0.5 OD

ze
SC
g

W
gQ
g –1

m
ki

ma OD
Tr

ab k0

pe
th
4W

g –1 g
nu

Q4
en u
alo

lum
m th
m

0m

zu ikiz we
m
gQ

en

W
m

Sp
30
Tra
kg

oli br

SC
kin

ab

4W

ab
n1

0 Le C
kg

30

Q4

eso
2.0

m gS
0m

1
45

m
4W

lok

um

g Q4W
Ne
Tez

mg – 0

28
the

0.1

gQ

SH
Risank
m

lim
SC
0m
ab

0
Q2
ab
4W

inu

Risankizum
n3

mg Q4W
SHR030
gQ

and 4

ab
um

gS
ep
C

Risankizumab 300 mg at weeks 0 and 4

mg
50
SHR0302 4 mg PO

R0
W

0m

ab
mg –1
kg

SHR
gS

30

m
the

ma
Spes

2W
elum
gQ
iz

ab
C
90 m

izu

SC
302

60
0m
ol

.5

30 m
0m

Q2

en 1
b1

izumab
W
em

SC

b0

zu
ol
m

0m
0302
at weeks 0

Q2
kg
ki
olim

W
gS

m
ab 2
b2

mg Q4
N

8m
50

ri
ma
60

2 8 mg
mg

Ne

C th

W
then

gI
2.0

then
C

mg SC then 90
ma

Le
mg

u
o
b

Placebo
80 m

gP
ab 60

4m

V
liz

ab 150 mg
Q2
60

Placeb
ma
u

gS
ab

300 m
liz

Q4
g SC

mo
SC

Od
W

mg SC
ab

um
lizu

umab 60

g PO
mo

0m
gS
PO dai

W
Ne
um

0 mg

Q2
b 150 mg

liz

aily
Ne

90 m
mo

CQ

g at w
ab 2
liz

mo
W

dail
daily
Ne
mo

ab 60
IVQ

Ne

at weeks
2W
ly

um

Nemoliz
umab

y
Ne

eeks
Nemolizumab 90
oliz
4W

olizum
Risankizuma
oliz

Nem

0 an
0 and 4
Nem

Nem

d4
(c) (d)

e weekly
Dupilumab 400 mg ×1 then 200

Dupilumab 400 mg ×1 then

y

100 mg SC Q4W
mg SC Q4W

eekl

Dupiluma
Dupilu
Dupil

mg SC onc
ce w

Dup

y
Dup

weekl
SC on
umab

ilum

mab 60
ilum

b 600 mg
then 100

ab 60

g once
600 mg

0 mg

then 200
ab 6
Du

0 mg

Dupilumab 400 mg ×1 then

0m
pil

00 m

en 20

300 m
um

OD
Fe

then 300
×1 th
400 mg ×1

g ×1
then 30

D
Etr
za
ab

g×

OO

b 400 mg

gO
mg th

PO
kin

200 mg SC Q2W

OD
asim
6

1 th

en 30
then
00

mab
gP

um

Etr

07 2

mg
mg SC Q2W

O
0 mg
mg

od
en 3

4m

asim
400

mg Q2W
bI

gP
GB

Dupilu

ib 4
300

OD
OD

DS1
0 mg
×1

1m
V
Dupilumab

Dupiluma
R

2m
00 m

inib

od
ab

60

citin

O
Q4W

83
the

mg
gO

gP
gP

0m
ilum

2m
0

inib
ricit

Q2W

IS
n

10
gQ

Bari

1m
2m

B
once
D
Et

30

g th

gO

D
it
83
kg
Ba
Dup
ra

O
0m

ric
2W

01
ib
ib
sim

en
m

OD

g
D

itin
itin

Ba
wee

m
20
g
–1
go

30
od

O 0m D

80
ric
ric

gP
IV

0m

gO
1

nc

IS
kly

2
gt
Ba
Ba

B
m

1m 00
on
e

83 0m
gQ

he
g

Etr
SN
we

01
da

26
OD

asim ib n6
tin
2W
y

50 OD
ek

i 00 0
N0
1

od ric mg
ly

m mg
an

ISB y
2m Ba OD th gQ AS 40 dail
d

830 en
gO PO ice
29

600 75 4W 02
D mg mg N0 O tw
KH 30 mg
the Q4 AS gP
inib m
K4
083 cit ISB n3 W t 40 daily
150 A bro 830
600
00
mg ilas ice
mg Q2 rem O tw
Q4 OD
mg
th W Ap gP
W g PO
en 3 30 m
0m 00 m ast
20 gQ mil
itinib
KHK
4083 4W Apre OD
KHK Abr
oc g PO
4083 150 30 m
300 m
mg
Q4W itinib
g Q2W Abroc
KHK4 OD
PO OD
083 30 mg PO
0 mg 0 mg Q2 ib 200
itinib 10 W Ab rocitin
Abroc
KHK4083 600 KHK4083 600 mg PO OD
mg Q2W mg Q2W Abrocitinib 100

Abrocitinib 10 mg PO OD KHK4083 600 mg Q4W Abrocitinib 10 mg PO OD

mg Q4W ZPL-3893787
KHK4083 600 125 mg SC 1× 30 mg PO OD
Lebrikizumab
Upadacit Ustekinu
inib 7.5
mg PO Q4W mab 90
4W OD 5 mg SC mg SC at we
mg SC
Q umab 12 eks 0 an
12 5 Lebrikiz Ustek
inum d 4
then 1× ab 45
g SC 1× mg SC mg SC
250 m
Upa
daci 250 4W
umab mab CQ at w
ikiz 4W tinib kizu S Upa eeks
Lebr CQ 30 m Lebri mg dac 0 an
gS g PO 1 25 4 W itin
ib 7
d4
50m OD × then CQ .5 m
n2 C1 gS W Up gP
the W Up gS m Q2 ada OO
SC
1× Q2 ada 50 m 50 SC citin D
C cit ab 2 n2 ib
mg gS inib um the mg Up 30
500 0m
Tr 15 rikiz 1× 50 Tr ad mg
um
ab 25
a lo
ki mg Leb g SC e n2 alo
ki
ac
itin PO
kiz the
n nu PO 0m 2 th W nu ib OD
L ebri 2 m O D b 50 n d Q4 m
ab 15
nd ab ma 0a SC mg
W

0a 60 u ek –1 60
Q4

kiz PO
Tr
W

ek 0m bri we g 0
a
Q4

Le m m OD
lo
g

we SC kg gt
m

gt
ki
Tr
SC

C mg he
nu
10

he .1
4W

gS
alo

0 n3
Q4
g –1

n3
SH

4W

00
m
en

m ab
Tra

b5
ab

00
kin
gQ

00 00
th

m
R0

SC

um
gQ

ma
4W

5 m
45

m
SC

lok

um

iz
kg

ab
30

gQ
0m

–1

izu
Tez

gQ ol
m
0m

um
0.5
28
g

a
mg
Risan

inu

rik 2W
SH

m
g
b3
m

2W
g Q4W

Ne
n3

ikiz eb
epe
mg

SC
ab

n1
20

Risankizum

mg

L
kg

at weeks 0 and 4

br
R0

Risankizumab 300 mg at

00
4W

Spe
mg Q4
the

Le
ab

Q2
lum
ab

the

30
kizu

.0
izu

4W
302

m
PO
m

SHR0

W
gQ
SC

15
so

gS
ol

90 m
en
izu

a
SHR0302
SC
ab
da

gQ
m

b2

0m
limab
mab

4m

C
mg

C th
0m
ol

um
Ne
ily
then 90

mg
m

Q2
80
60 m

gS
302 8
liz
gP
Ne

60

then
ab 6

gS

W
Placebo

30
ab 150 mg

20

mg
mo

Placebo

CQ
600
ab

Od
0 mg

0m
ab
Ne
um

umab
um

SC
g SC

4 mg PO

mg PO

2W
mg SC

aily

um

mg
b6

Risankizumab 150 mg
liz

oliz

Q2
liz
at w
mo

oliz

IV
90 m
ma
Nem

mo

W
Ne

at weeks

lizu

Q4W
eeks

daily
ab 90

Nem
Ne

daily
ab

weeks 0
mo

izum
0 an
lizum

Ne
0 and 4

d4

Nemol
Nemo

and 4

Figure 2 Network plots of studies included in the network meta-analysis of adults treated between 8 and 16 weeks for the proportion of patients
achieving (a) at least a 50% improvement in Eczema Area and Severity Index (EASI-50), (b) at least a 75% improvement in EASI (EASI 75), (c) at least
a 90% improvement in EASI (EASI 90) and (d) success on Investigator Global Assessment (IGA). The width of each line connecting two treatments
(nodes) is proportional to the number of head-to-head trials for that comparison. Some studies included in the analyses of trials of adults included a
minority proportion of adolescent participants (12–17 years old). IV, intravenous; OD, once daily; PO, per os; Q1W, once weekly; Q2W, every 2 weeks;
Q4W, every 4 weeks; SC, subcutaneous.
188 Relative efficacy of systemic treatments for eczema, A.M. Drucker et al.

(a)
Treatment OR (95% CrI)
EASI 50 Favors Favors
dupliumab comparator
Abrocitinib, 100 mg, once daily 0.7 (0.5 to 1.0)
Abrocitinib, 200 mg, once daily 1.5 (1.1 to 2.2)
Baricitinib, 2 mg, once daily 0.4 (0.3 to 0.5)
Baricitinib, 4 mg, once daily 0.5 (0.3 to 0.7)
Tralokinumab, 600 mg then 300 mg, every 2 wk 0.4 (0.3 to 0.6)
Upadacitinib, 15 mg, once daily 1.7 (0.9 to 3.3)
Upadacitinib, 30 mg, once daily 2.5 (1.3 to 5.0)
0.3 0.5 1.0 2.0 5.0
EASI 50 vs dupilumab (OR)

(b)

Downloaded from https://academic.oup.com/bjd/article/190/2/184/7311111 by guest on 07 June 2025

Treatment OR (95% CrI)
EASI 75 Favors Favors
dupliumab comparator
Abrocitinib, 100 mg, once daily 0.7 (0.5 to 1.0)
Abrocitinib, 200 mg, once daily 1.6 (1.2 to 2.1)
Baricitinib, 2 mg, once daily 0.4 (0.3 to 0.6)
Baricitinib, 4 mg, once daily 0.5 (0.3 to 0.7)
Tralokinumab, 600 mg then 300 mg, every 2 wk 0.4 (0.3 to 0.5)
Upadacitinib, 15 mg, once daily 1.2 (0.9 to 1.7)
Upadacitinib, 30 mg, once daily 2.1 (1.6 to 2.7)
0.3 0.5 1.0 2.0
EASI 75 vs dupilumab (OR)

(c)
Treatment OR (95% CrI)
EASI 90 Favors Favors
dupliumab comparator
Abrocitinib, 100 mg, once daily 0.9 (0.6 to 1.2)
Abrocitinib, 200 mg, once daily 1.7 (1.3 to 2.2)
Baricitinib, 2 mg, once daily 0.4 (0.3 to 0.6)
Baricitinib, 4 mg, once daily 0.5 (0.3 to 0.8)
Tralokinumab, 600 mg then 300 mg, every 2 wk 0.4 (0.3 to 0.5)
Upadacitinib, 15 mg, once daily 1.4 (1.0 to 1.9)
Upadacitinib, 30 mg, once daily 2.6 (1.9 to 3.4)
0.3 0.5 1.0 2.0 3.0
EASI 90 vs dupilumab (OR)

(d)
Treatment OR (95% CrI)
IGA Favors Favors
dupliumab comparator
Abrocitinib, 100 mg, once daily 0.8 (0.6 to 1.1)
Abrocitinib, 200 mg, once daily 1.6 (1.3 to 2.0)
Baricitinib, 2 mg, once daily 0.5 (0.3 to 0.7)
Baricitinib, 4 mg, once daily 0.6 (0.4 to 0.9)
Tralokinumab, 600 mg then 300 mg, every 2 wk 0.4 (0.3 to 0.5)
Upadacitinib, 15 mg, once daily 1.6 (1.1 to 2.3)
Upadacitinib, 30 mg, once daily 2.9 (2.0 to 4.1)
0.3 0.5 1.0 2.0 4.0
IGA response vs dupilumab (OR)

Figure 3 Forest plots of network meta-analysis results of adults treated between 8 and 16 weeks for the proportion of patients achieving (a) at least
a 50% improvement in Eczema Area and Severity Index (EASI 50), (b) at least a 75% improvement in EASI (EASI 75), (c) at least a 90% improvement
in EASI (EASI 90) and (d) success on Investigator Global Assessment (IGA). Some studies included in the analyses of trials of adults included a
minority proportion of adolescent participants (12–17 years old). Results are presented as odds ratios (OR) with 95% credible intervals (CrIs) for
different medications vs. dupilumab. ORs > 1 indicate that the comparator is associated with higher odds of achieving the efficacy outcome than
dupilumab. ORs < 1 indicate that the comparator is associated with lower odds of achieving the efficacy outcome than dupilumab. OD, once daily;
Q2W, every 2 weeks.
Relative efficacy of systemic treatments for eczema, A.M. Drucker et al. 189

with lower odds of achieving EASI 50 than dupilumab trials included exclusively adult populations or also included
600 mg then 300 mg every 2 weeks (Figure 3a, Table S5). adolescents aged ≥ 12 years. Node splitting did not suggest
Each active treatment was associated with higher odds of incoherence between direct and indirect estimates, but the
achieving EASI 50 vs. placebo, with ORs ranging from 2.6 network was sparse, making it difficult to check properly for
(baricitinib 2 mg daily) to 16.6 (upadacitinib 30 mg daily). incoherence and other assumptions underlying NMA. AD is
A similar pattern of results was seen for EASI 75 (Figure 3b, a chronic condition, with systemic immunomodulation often
Table S6), EASI 90 (Figure 3c, Table S7) and IGA success required over many years, but our analyses were limited to
(Figure 3d, Table S8). There were no important differences trials up to 16 weeks’ duration. Randomized trial populations
in effect estimates between medications of interest in anal- may have important differences from people treated with
yses limited to trials permitting vs. not permitting topical systemic immunomodulation in real-world practice, so clini-
anti-inflammatory treatments (Table S9). Findings were sim- cians and patients should be aware that the results do not
ilar in sensitivity analyses limited to studies at low risk of generalize to all adults with AD considering systemic ther-
bias. Node splitting did not suggest incoherence between apy. Azathioprine, ciclosporin and methotrexate were not
direct and indirect evidence, including visual inspection connected to networks that included newer treatments, so
of forest plots and checking for statistical differences [all we were unable to compare those medications with regard

Downloaded from https://academic.oup.com/bjd/article/190/2/184/7311111 by guest on 07 June 2025

P > 0.05, Figure S1 (see Supporting Information)]. to the included binary outcomes.
The Gelman–Rubin–Brooks plots showed good con- In conclusion, relative efficacy estimates in an NMA of
vergence of the models. The trace plots showed a good systemic treatments for AD in adults using binary outcomes
mix between chains and the posterior density distribution of clinical signs are consistent with analyses for continu-
showed approximately normal shapes for the posterior ous change in clinical signs.3 Upadacitinib 30 mg daily and
log(OR). These good convergence features were consistent abrocitinib 200 mg daily are the most efficacious with regard
for parameters across all models. to achieving binary efficacy endpoints, followed by upad-
acitinib 15 mg daily, dupilumab 600 mg then 300 mg every
2 weeks and abrocitinib 100 mg daily. These are more effica-
Discussion cious than baricitinib 4 mg and 2 mg daily and tralokinumab
600 mg then 300 mg every 2 weeks.
In this study, higher doses of abrocitinib and upadacitinib
were associated with higher proportions of treatment Funding sources
responders for binary outcomes of clinical signs of AD
in adults than dupilumab, with lower doses of abrocitinib This work was supported through a UK National Institute
and upadacitinib being closer in efficacy to dupilumab. for Health Research Career Development Fellowship held
Dupilumab was superior to baricitinib and tralokinumab. The by C.F. (CDF-2014-07-037). Knowledge translation compo-
pattern of results is similar to and supports previously pub- nents of this work were funded by the Eczema Society of
lished results from NMA of continuous outcomes.3 These Canada and the Innovation Fund of the Alternate Funding
results may be helpful for patients, clinicians and health Plan for the Academic Health Sciences Centres of Ontario.
technology assessors seeking to understand the relative The funders had no role in the planning or conduct of the
efficacy of systemic treatments for AD. research, or the decision to publish these results.
Binary transformations of continuous outcomes lead to
the loss of important information because they fail to cap- Conflicts of interests
ture clinically meaningful improvements for many AD trial
participants.13 Reporting guidelines specific to AD and for A.M.D. has received compensation from the British Journal
trials in general suggest reporting continuous measures.4,14 of Dermatology (Section Editor), American Academy
Nevertheless, binary outcomes are recommended as pri- of Dermatology (guidelines writer), National Eczema
mary outcomes by regulatory agencies and so are included Association (grant reviewer) and Canadian Agency for
and reported in most AD trials. It is reassuring that the NMA Drugs and Technologies in Health (consultant); and has
results for binary outcomes were consistent with those for received research grants to his institution from the National
continuous outcomes.3 It is also reassuring that relative effect Eczema Association, Eczema Society of Canada, Canadian
estimates were consistent for binary outcomes, regardless Dermatology Foundation, Canadian Institutes for Health
of whether the included trials allow or do not allow adjunc- Research, US National Institutes of Health and Physicians
tive topical treatment, as topical anti-inflammatory treatment Services Incorporated Foundation. A.G.E. is an employee
may be an important effect modifier for AD trials.15 of Stratevi, a healthcare consultancy that receives finan-
This study had several limitations. The included trials cial compensation from numerous pharmaceutical compa-
were generally similar in design, but may have differences, nies; Stratevi was not involved with the submitted work.
including whether they allowed topical anti-inflammatory Z.Z.N.Y. is an Associate Editor for the British Journal of
therapy, potentially affecting the transitivity assumption of Dermatology. B.W.M.A. is a Patient Associate Editor for
NMA. We were reassured by similar findings in our second- the British Journal of Dermatology. P.I.S. has consulted
ary analyses limited to trials allowing and not allowing topi- for Sanofi and AbbVie (unpaid), was Principal Investigator
cal anti-inflammatory therapy, respectively. Other potential (PI) of the Methotrexate vs. Azathioprine for severe Atopic
differences between trials that could have acted as effect Dermatitis (MAcAD) study, received departmental inde-
modifiers and led to violations of the transitivity assump- pendent research grants for the TREAT NL registry, for
tion of NMA included different run-in periods off topical which she is Chief Investigator (CI), from pharma compa-
treatment, varying baseline disease severity and whether nies since December 2019, and is involved in performing
190 Relative efficacy of systemic treatments for eczema, A.M. Drucker et al.

clinical trials with Sanofi, Pfizer, AbbVie, Novartis, Leo 3 Drucker AM, Morra DE, Prieto-Merino D et al. Systemic immu-
Pharma and Galderma, for which financial compensation nomodulatory treatments for atopic dermatitis: update of a living
is paid to the department/hospital. J.S. received institu- systematic review and network meta-analysis. JAMA Dermatol
2022; 158:523–32.
tional funding for investigator-initiated trials from Novartis,
4 Grinich EE, Schmitt J, Küster D et al. Standardized reporting
Sanofi, Pfizer and ALK; has received fees for consulting from
of the Eczema Area and Severity Index (EASI) and the Patient-
Novartis and Pfizer; and is co-PI of the German national Oriented Eczema Measure (POEM): a recommendation by the
AD registry TREATgermany, which is funded by Sanofi Harmonising Outcome Measures for Eczema (HOME) Initiative.
Aventis Deutschland, Galderma, LEO Pharma and Lilly Br J Dermatol 2018; 179:540–1.
Deutschland. C.F. is CI of the UK National Institute for Health 5 Lam M, Spuls PI, Leshem YA et al. Reporting of Harmonising
Research-funded TREAT (ISRCTN15837754) and SOFTER Outcome Measures for Eczema (HOME) core outcome set
(NCT03270566) trials, as well as the UK–Irish Atopic eczema instruments in randomized clinical trials for systemic treatments
Systemic Therapy Register (A-STAR; ISRCTN11210918) in atopic dermatitis. Br J Dermatol 2023; 189:494–6.
and a PI in the European Union (EU) Horizon 2020-funded 6 Hutton B, Salanti G, Caldwell DM et al. The PRISMA extension
statement for reporting of systematic reviews incorporating net-
BIOMAP Consortium (http://www.biomap-imi.eu/). He
work meta-analyses of health care interventions: checklist and
also leads the EU Trans-Foods consortium. His department
explanations. Ann Intern Med 2015; 162:777–84.

Downloaded from https://academic.oup.com/bjd/article/190/2/184/7311111 by guest on 07 June 2025

has received funding from Sanofi-Genzyme and Pfizer for 7 Higgins JPT, Altman DG, Gøtzsche PC et al. The Cochrane
skin microbiome work. He has also received compensation Collaboration’s tool for assessing risk of bias in randomised trials.
from the British Journal of Dermatology (Section Editor) and BMJ 2011; 343:d5928.
EuroGuiDerm (guidelines lead). M.L., R.E., D.P.-M., R.M., 8 Futamura M, Leshem YA, Thomas KS et al. A systematic review
T.B. and B.R. declare no conflicts of interest. of Investigator Global Assessment (IGA) in atopic dermatitis (AD)
trials: many options, no standards. J Am Acad Dermatol 2016;
74:288–94.
Data availability 9 Turner RM, Jackson D, Wei Y et al. Predictive distributions
for between-study heterogeneity and simple methods for
A complete data file is available to download from www.
their application in Bayesian meta-analysis. Stat Med 2015;
eczematherapies.com/research.
34:984–98.
10 Golder S, Loke YK, Wright K et al. Reporting of adverse events in
Ethics statement published and unpublished studies of health care interventions: a
systematic review. PLOS MED 2016; 13:e1002127.
Not applicable. 11 Salanti G, Ades AE, Ioannidis JP. Graphical methods and numer-
ical summaries for presenting results from multiple-treatment
meta-analysis: an overview and tutorial. J Clin Epidemiol 2011;
64:163–71.
Supporting Information
12 Dias S, Welton NJ, Caldwell DM et al. Checking consistency
in mixed treatment comparison meta-analysis. Stat Med 2010;
Additional Supporting Information may be found in the
29:932–44.
online version of this article at the publisher’s website.
13 Silverberg JI, Simpson EL, Ardeleanu M et al. Dupilumab pro-
vides important clinical benefits to patients with atopic dermatitis
who do not achieve clear or almost clear skin according to the
References Investigator’s Global Assessment: a pooled analysis of data from
two phase III trials. Br J Dermatol 2019; 181:80–7.
1 Drucker AM, Ellis A, Jabbar-Lopez Z et al. Systemic immuno- 14 Schulz KF, Altman DG, Moher D. CONSORT 2010 Statement:
modulatory treatments for atopic dermatitis: protocol for a sys- updated guidelines for reporting parallel group randomised trials.
tematic review with network meta-analysis. BMJ Open 2018; BMJ 2010; 340:c332.
8:e023061. 15 Leshem YA, Bissonnette R, Paul C et al. Optimization of pla-
2 Drucker AM, Ellis AG, Bohdanowicz M et al. Systemic immu- cebo use in clinical trials with systemic treatments for atopic
nomodulatory treatments for patients with atopic dermatitis: a dermatitis: an International Eczema Council survey-based
systematic review and network meta-analysis. JAMA Dermatol position statement. J Eur Acad Dermatol Venereol 2019;
2020; 156:659–67. 33:807–15.