Professional Documents
Culture Documents
12238 JEADV
ORIGINAL ARTICLE
Abstract
Background Differences in response rates of biologics for the treatment of moderate-to-severe plaque psoriasis have
been reported in several meta-analyses published to date. However, the usefulness of these meta-analyses is limited as
they do not reflect currently approved recommendations in the Summaries of Product Characteristics (SmPCs) and
clinical practice.
Objective To estimate the efficacy of biologics in the treatment of moderate-to-severe plaque psoriasis in the
currently approved conditions of use in the European Union (EU) at relevant time points for evaluation of response
in clinical practice (failure assessment as recommended in the SmPCs and/or at the end of the induction phase).
Methods Randomized placebo-controlled studies of biologics currently authorized in the EU in adult patients with a
diagnosis of moderate-to-severe plaque psoriasis were searched in several databases. A meta-analysis using fixed or
random-effects model depending on heterogeneity across and within studies was performed. The efficacy was estimated
using risk difference (RD) of Psoriasis Area and Severity Index (PASI) 50, PASI 75 and PASI 90 response rates at the end-
point in clinical trials, at the end of the induction phase (week 24) and at the time points recommended for evaluation of
primary failure in the approved SmPCs. Several sensitivity analyses were performed to assess for robustness.
Results Sixteen publications met the defined inclusion criteria. According to this meta-analysis at the primary endpoint
times, infliximab (at week 10) has the greatest probability of response with respect to placebo for all PASI-based efficacy
measures (PASI 50, PASI 75 and PASI 90). At the end of the induction phase (week 24), ustekinumab 45 mg has the
greatest probability of achieving PASI 75 response (RD 75.5%, 95%CI 71.5–79.4%], followed by ustekinumab 90 mg,
infliximab, adalimumab and etanercept. At the time points recommended for primary failure assessment according to the
approved SmPCs, ustekinumab 45 mg (at week 28) also has the greatest probability of achieving PASI 50 response [RD
80.7%, 95%CI 77.2–84.2%], followed by ustekinumab 90 mg, infliximab, adalimumab and etanercept.
Conclusion From a clinical practice perspective, and in terms of the most relevant efficacy measures (PASI 50 and
PASI 75) and time points (end of induction phase [week 24] and time to assess primary failure as per the SmPCs), in the
currently approved conditions of use ustekinumab is the most efficacious therapeutic alternative for moderate-to-severe
plaque psoriasis, followed by infliximab, adalimumab and etanercept.
Received: 21 March 2013; Accepted: 12 July 2013
Conflicts of interest
Lluís Puig has received consultancy and speaker’s honoraria from Abbott, Janssen, MSD, and Pfizer, and participated in
clinical trials sponsored by these laboratories. Ignacio García González is an employee of Janssen-Cilag, S.A and owns
stock in Johnson & Johnson.
Funding sources
Janssen-Cilag, S.A. provided funding for independent data analysis by Pertica, Spain.
JEADV 2014, 28, 1633–1653 © 2013 European Academy of Dermatology and Venereology
1634 Lluıs Puig et al.
failing to respond to, or having a contraindication to, or being maintenance treatment, combination therapy, dose
intolerant to other systemic therapies including cyclosporine, intensification or switching is to be done).13 Psoriasis Area and
methotrexate and PUVA are eligible for treatment with Severity Index (PASI) 75 and PASI 90 (namely, the percentage
biologics. of patients achieving at least a 75% or 90% improvement with
The differences in response rates of biologics are well respect to baseline PASI) are the relevant outcome measures for
established as demonstrated in the only head-to-head study efficacy, whereas PASI 50 is the lower threshold of efficacy, and
comparing two approved biologics published to date,5 and 100% minus the percentage of patients achieving PASI 50
several articles presenting meta-analyses comparing biologics response can be considered the percentage of failure.
indirectly.6–11 These meta-analyses are used in cost-efficacy The aim of this study was to estimate the efficacy of currently
analyses12 that may contribute to establishing treatment available biologics for the treatment of moderate-to-severe
recommendations for the use of biologics in clinical practice. plaque psoriasis under the currently approved conditions of use
When performing economic evaluations by combining costs in the EU at clinically relevant time points of evaluation by
with efficacy estimates it is necessary to apply clinically relevant means of risk difference (RD) with respect to placebo.
data based on available scientific evidence, to optimize treatment
decision-making.13 Clinically relevant time points to compare Methods
biologics currently approved in the EU for the treatment of
moderate-to-severe psoriasis are the primary endpoint (used for Identification of randomized controlled trials
regulatory evaluation of efficacy), time of failure assessment Studies published in the Medline, EMBASE, the Cochrane
[when a decision has to be made according to the approved Central Register of Controlled Trials databases were identified
Summaries of Product Characteristics (SmPCs)], and at the end using combinations of the following keywords: psoriasis; any of
of the induction phase (when a decision on continuing these treatments: adalimumab, etanercept, infliximab or
JEADV 2014, 28, 1633–1653 © 2013 European Academy of Dermatology and Venereology
Table 1 (a) Data abstracted from selected studies in the systematic review. (b) Data abstracted from non-placebo controlled studies included in the sensitivity analyses
Author, year Comparator and Patients (n) Week Number of Number of Number of
Dosage of evaluation patients patients patients
of response achieving PASI achieving PASI achieving
50 response 75 response PASI
90 response
(a) Data abstracted from selected studies in the systematic review
Chaudhari, 2001 Placebo 11 10 ND 2 ND
Author, year Comparator and Patients (n) Week Number of Number of Number of
Dosage of evaluation patients patients patients
of response achieving PASI achieving PASI achieving
50 response 75 response PASI
90 response
Leonardi, 2008 Placebo 255 12 26 8 5
Ustekinumab 45 mg (≤100 kg) 168 12 143 124 79
Author, year Comparator and Patients (n) Week Number of Number of Number of
Dosage of evaluation patients patients patients
of response achieving PASI achieving PASI achieving
50 response 75 response PASI
90 response
Papp, 2005 Placebo 193 12 18 6 1
Etanercept 25 mg biw 196 12 126 67 21
Placebo 410 12 41 15 3
Ustekinumab 45 mg 409 12 342 273 173
Ustekinumab 90 mg 411 12 367 311 209
Placebo (LOCF) 410 24 41 15 3
Ustekinumab 45 mg 409 24 366 292 202
Ustekinumab 90 mg 411 24 376 335 240
Placebo (LOCF) 410 28 41 15 3
Ustekinumab 45 mg 409 28 369 276 178
Ustekinumab 90 mg 411 28 380 314 217
Reich, 2005 Placebo 77 10 6 2 1
Infliximab 5 mg/kg 301 10 274 242 172
Placebo 77 24 6 3 1
Infliximab 5 mg/kg 301 24 248 227 161
Saurat, 2008 Placebo 53 12 14 8 4
Adalimumab 40 mg eow 108 12 98 83 53
Placebo 53 16 16 10 6
Adalimumab 40 mg eow 108 16 95 86 54
Strober, 2011 Placebo 72 12 ND 5 3
Etanercept 50 mg biw 139 12 ND 55 19
1637
Table 1 (Continued)
Author, year Comparator and Patients (n) Week Number of Number of Number of
Dosage of evaluation patients patients patients
of response achieving PASI achieving PASI achieving
PASI, Psoriasis Area and Severity Index; LOCF, Last Observation Carried Forward; ND, No data provided
Lluıs Puig et al.
Table 2 Pooled efficacy data at the primary endpoint times to plateau of PASI 75 response rate plots in clinical trials),13
(pooled risk difference vs. placebo) and at the recommended time point for assessing treatment
Treatment PASI 50 (%) PASI 75 (%) PASI 90 (%) failure according to the SmPCs.
[95% CI] [95% CI] [95% CI] 2 Only approved treatment regimens indicated for therapy of
Adalimumab† 66.4 (62.4–70.5) 63.0 (59.3–66.7) 36.5 (25.7–47.4)* plaque psoriasis according to the respective SmPC were
Etanercept 52.2 (47.1–57.3) 31.0 (26.6–35.4) 10.7 (7.8–13.6) included in the meta-analysis. Thus, for ustekinumab, only
25 mg
weight-based efficacy data (45 mg for patients with body-
biw or
50 mg qw‡ weight ≤100 kg, 90 mg for patients with bodyweight
Etanercept 62.0 (57.8–66.1) 43.5 (40.0–47.1) 19.3 (16.6–22.0) >100 kg) were considered. In the case of etanercept, different
50 mg biw‡ treatment regimens were considered separately, including
Infliximab§ 80.5 (74.4–86.5) 75.7 (72.1–79.3) 49.5 (45.6–53.4) etanercept 25 mg twice weekly (biw), etanercept 50 mg
Ustekinumab 76.4 (72.5–80.2) 70.1 (65.8–74.3) 47.2 (42.6–51.8) weekly (qw), etanercept 50 mg biw for the first 12 weeks fol-
45 mg‡,¶
lowed by 50 mg qw and finally, etanercept 50 mg biw.
Ustekinumab 76.8 (71.7–81.9) 66.5 (60.2–72.9) 35.5 (29.0–42.0)
90 mg‡,¶ 3 All trials used in the meta-analysis compared the respective
biologics vs. placebo. Evidence on placebo response beyond
*DerSimonian-Laird pooled risk difference.
12 weeks is limited, as clinical trials are usually designed with
†Evaluation of response at week 16.
‡Evaluation of response at week 12. a cross-over phase from placebo to active treatment after the
§Evaluation of response at week 10. primary efficacy endpoint has been measured. Thus, a last
¶Efficacy (PASI response rates) based on recommended posology as per observation carried forward (LOCF) approach was used for
the approved SmPC: ustekinumab 45 mg for patients with bodyweight the placebo group at week 24. This approach uses the last
≤100 kg, and ustekinumab 90 mg for patients with bodyweight >100 kg. observed efficacy data in the clinical trial (which usually coin-
SmPC, Summaries of Product Characteristics; PASI, Psoriasis Area and
cides with the endpoint). In addition, the available clinical
Severity Index; CI, confidence intervals
evidence shows that placebo efficacy does not increase
significantly from the endpoint until week 24 assessment.14,15
ustekinumab; randomized trial; English language and published Using LOCF data for the placebo group qualifies as an accept-
from 1 January 2001 to 1 May 2012. able approach, as it is consistent with current clinical evidence
and allows for minimizing population effects on a study basis.
Inclusion and exclusion criteria 4 In the case of etanercept the response rates at week 12
To be considered for inclusion in the meta-analysis, the studies corresponding to the measure at the endpoint and SmPC
had to meet the following inclusion criteria: Patients (human) were pooled for arms where patients received 50 mg biw for
with a diagnosis of moderate-to-severe plaque psoriasis; the first 12 weeks.
placebo-controlled studies; PASI 75 as main outcome; 5 In case of a missing outcome value, we have contacted the
assessment of primary efficacy endpoint at 10, 12 or 16 weeks; authors to retrieve the data.
random and blinded allocation of participants in the treatment
and control group, and finally, trial publication as part of full Statistical methods
articles in peer-reviewed journals. Meta-analyses were done using the fixed (Mantel–Haenszel)
Only biologics currently authorized in the EU as of May 2012 or random-effects (DerSimonian & Laird) model in the meta-
were included in the analysis. Likewise, only authorized treat- analytic statistical package in STATA version 10.0 (StataCorp
ment regimens were included (i.e. combined therapy was not LP, College Station, TX, USA) depending on heterogeneity
included); observational, retrospective studies, case reports/ser- across individual studies. These estimations take into consid-
ies and review articles were also not included. eration the within-study comparison as well as differences
between studies. Heterogeneity was studied by the I2 statistic
Data extraction proposed by Higgins and Thompson.16
The inclusion criteria for the analysis were applied by one For the comparative models, the weighted RD was calculated
researcher and verified by another. There were no discrepancies based on weighting of individual results by the inverse variance;
between the researchers. Data extracted from each trial included this limits the impact of studies with wide confidence intervals
the main study and patient characteristics as well as results from (CIs). The study-specific standard errors for the estimated RD
outcome parameters and time points. The following criteria were used to account for within-study variation.
were applied for data extraction: To address the optimal reporting of meta-analyses, we used
1 To consider consistent time points for the analysis, data were the recent revision of the PRISMA (Preferred Reporting Items
extracted at three different time points: at the primary for Systematic reviews and Meta-Analyses) and we adopted the
endpoint time, at week 24 (end of induction phase according definitions proposed by the Cochrane Collaboration.17
JEADV 2014, 28, 1633–1653 © 2013 European Academy of Dermatology and Venereology
1640 Lluıs Puig et al.
%
Study Weight
ID RD (95% CI) (M-H)
–0.9 0 0.9
Figure 2a Forest plot depicting meta-analysis of PASI 50 efficacy results at the primary endpoint. RD, risk difference; M-H, Mantel-
Haenszel; D+L, DerSimonian & Laird.
JEADV 2014, 28, 1633–1653 © 2013 European Academy of Dermatology and Venereology
Biologics in psoriasis: a meta-analysis at relevant time points 1641
%
Study Weight
ID RD (95% CI) (M-H)
–0.959 0 0.959
Figure 2b Forest plot depicting meta-analysis of PASI 75 efficacy results at the primary endpoint. RD, risk difference; M-H, Mantel-
Haenszel; D+L, DerSimonian & Laird.
Sixteen publications of placebo-controlled studies with seven studies of etanercept14,18–23 (n = 2.367), four of inflix
biologics including 6.905 patients (4.483 with biologic treatment imab15,24–26 (n = 1.072), three of adalimumab27–29 (n = 1.470)
and 2.422 with placebo) met the predefined inclusion criteria: and two of ustekinumab30,31 (n = 1.996). All studies reported
JEADV 2014, 28, 1633–1653 © 2013 European Academy of Dermatology and Venereology
1642 Lluıs Puig et al.
%
Study Weight
ID RD (95% CI) (M-H)
–0.661 0 0.661
Figure 2c Forest plot depicting meta-analysis of PASI 90 efficacy results at the primary endpoint. RD, risk difference; M-H, Mantel-
Haenszel; D+L, DerSimonian & Laird.
PASI 75 as the primary efficacy measure at different time points biologic treatments showed superior responses compared to
and most of them also reported PASI 50 and PASI 90 measures. placebo (Table 1a). Data abstracted from non-placebo con-
Baseline demographic characteristics (age, gender or disease trolled studies included in the sensitivity analyses are detailed in
duration) were similar between trials and consistently all Table 1b.
JEADV 2014, 28, 1633–1653 © 2013 European Academy of Dermatology and Venereology
Biologics in psoriasis: a meta-analysis at relevant time points 1643
Efficacy results at the primary endpoint At week 24, in terms of PASI 75 response (a relevant efficacy
According to this meta-analysis at the primary endpoint times measure in clinical practice), ustekinumab 45 mg has the
(Table 2, Fig. 2), infliximab (at week 10) has the greatest greatest probability of achieving this outcome (RD 75.5%,
probability of response with respect to placebo for all PASI- 71.5%–79.4%), followed by ustekinumab 90 mg (RD 75.0%,
based efficacy measures (PASI 50, PASI 75 and PASI 90). As 69.3%–80.7%), infliximab (RD 69.2%; 64.1%–74.4%),
regards PASI 75, the relevant efficacy measure at the endpoint adalimumab (RD 63.3%, 59.4%–67.1%), etanercept 50 mg
in clinical trials, infliximab (at week 10) has the greatest proba- biweekly (RD 55.3%, 50.6%–60.1%), 50 mg biweekly (RD
bility of achieving PASI 75 response (RD 75.7%, 95% CI 51.0%, 43.6%–58.4%) and etanercept 25 mg biweekly or 50 mg
72.1%–79.3%) followed by ustekinumab 45 mg (at week 12) weekly (RD 48.8%, 37.7%–59.9%).
(RD 70.1%, 65.8%–74.3%), ustekinumab 90 mg (at week 12)
(RD 66.5%, 60.2%–72.9%), adalimumab (at week 16) (RD Sensitivity analysis
63.0%, 59.3%–66.7%), etanercept 50 mg biweekly (at week 12) To incorporate the maximum amount of clinical evidence into
(RD 43.5%, 40.0%–47.1%) and etanercept 25 mg biweekly or the meta-analysis a sensitivity analysis, including results from
50 mg weekly (at week 12) (RD 31.0%, 26.6%–35.4%). studies without a placebo control arm, was carried out
(Table 1b). Five clinical trials on biologics without a placebo-
Efficacy results at the recommended time point for controlled arm accounting for 2.947 patients were included in
assessing failure as per the approved SmPCs the sensitivity analysis: three studies of etanercept5,32,33
According to this meta-analysis at the recommended time point (n = 1.372), and one of infliximab34 (n = 653), adalimumab35
as per the approved SmPCs (Table 3, Fig. 3), ustekinumab (at (n = 366) and ustekinumab5 (n = 556), respectively. In this sen-
week 28) rendered the greatest probabilities of response in all sitivity analysis, for those trials without placebo arm a pooled
PASI measures (PASI 50, PASI 75 and PASI 90). placebo response rate derived from all placebo arms was used. In
According to the SmPC-based recommendation regarding this case, for the relevant parameters from a clinical practice per-
withdrawal (and switching) of biologics, PASI 50 response is the spective (PASI 50 and PASI 75), the results obtained in the three
threshold for efficacy and therefore not achieving this outcome
is a relevant measure of failure in clinical practice. In this regard, Table 3 Pooled efficacy data at the recommended time point for
ustekinumab 45 mg (at week 28) has the greatest probability of assessing failure as per approved SmPCs (pooled risk difference
achieving PASI 50 response (RD 80.7%, 77.2%–84.2%), followed vs. placebo)
by ustekinumab 90 mg (at week 28) (RD 80.9%, 76.5%–85.4%), Treatment PASI 50 (%) PASI 75 (%) PASI 90 (%)
adalimumab (at week 16) (RD 66.4%, 62.4%–70.5%), etanercept [95% CI] [95% CI] [95% CI]
50 mg biweekly (at week 12) (RD 62.0%, 57.8%–66.1%) and e- Adalimumab† 66.4 (62.4–70.5) 63.0 (59.3–66.7) 36.5 (25.7–47.4)*
tanercept 25 mg biweekly or 50 mg weekly (at week 12) (RD Etanercept 52.2 (47.1–57.3) 31.0 (26.6–35.4) 10.7 (7.8–13.6)
25 mg biw
52.2%, 47.1%–57.3%). No PASI 50 data are available for
or 50mg qw‡
infliximab. Etanercept 62.0 (57.8–66.1) 43.5 (40.0–47.1) 19.3 ( 16.6–22.0)
Psoriasis Area and Severity Index 75 is also a relevant efficacy 50 mg biw§
measure at the moment of evaluation of response in clinical Infliximab¶ ND 67.1 (59.6–74.5) ND
practice as it defines treatment success and achieving this Ustekinumab 80.7 (77.2–84.2) 71.2 (67.0–75.4) 52.0 (47.4–56.6)
outcome goal justifies the continuation of treatment without 45 mg**,††
changes. In this regard, ustekinumab 45 mg (at week 28) has the Ustekinumab 80.9 (76.5–85.4) 69.3 (63.2–75.5) 39.1 (32.5–45.8)
90 mg**,††
greatest probability of achieving PASI 75 response (RD 71.2%,
67.0%–75.4%), followed by ustekinumab 90 mg (at week 28) *DerSimonian-Laird pooled risk difference.
†Evaluation of response at week 16 (after 10 SC doses of adalimumab
(RD 69.3%, 63.2%–75.5%), infliximab (at week 22) (RD 67.1%,
40 mg).
59.6%–74.5%), adalimumab (at week 16) (RD 63.0%,
‡Evaluation of response at week 12 (after 24 SC doses of etanercept
59.3%–66.7%), etanercept 50 mg biweekly (at week 12) (RD 25 mg or 12 SC doses of etanercept 50 mg).
43.5%, 40.0%–47.1%) and etanercept 25 mg biweekly or 50 mg §Evaluation of response at week 12 (after 24 SC doses of etanercept
weekly (at week 12) (RD 31.0%, 26.6%–35.4%). 50 mg).
¶Evaluation of response at week 22 (after 4 doses of infliximab (sic)).
Efficacy results at the end of the induction phase (week 24) **Evaluation of response at week 28 (after 3 SC doses of ustekinumab
45 mg or 90 mg in week 0. 4 and 16).
According to this meta-analysis, ustekinumab (at week 24),
††Efficacy (PASI) based on recommended posology as per the approved
rendered the greatest probabilities of response in all PASI-
SmPC: ustekinumab 45 mg for patients with bodyweight ≤100 kg. and
based outcome measures (PASI 50, PASI 75 and PASI 90 ustekinumab 90 mg for patients with bodyweight >100 kg.
response RD) at the end of the induction phase (week 24) SmPC, Summaries of Product Characteristics; PASI, Psoriasis Area and
(Table 4, Fig. 4). Severity Index ; CI, confidence intervals
JEADV 2014, 28, 1633–1653 © 2013 European Academy of Dermatology and Venereology
1644 Lluıs Puig et al.
%
Study Weight
ID RD (95% CI) (M-H)
–0.895 0 0.895
Figure 3a Forest plot depicting meta-analysis of PASI 50 efficacy results at the recommended time point for assessing failure as per the
approved SmPCs. RD, risk difference; M-H, Mantel-Haenszel; D+L, DerSimonian & Laird.
meta-analyses (Tables 5, 6 and 7) were similar. For infliximab, excluded from the analysis: one each for adalimumab,27 inflix-
this sensitivity analysis provides an estimate of the PASI 50 RD imab24 and etanercept,14 respectively. In this second sensitivity
at the time point for assessing failure as per the approved SmPC analysis, the results obtained in the three meta-analyses
(30% primary failure rate, with 95% CIs non overlapping with (Tables 5, 6 and 7) were also similar across all PASI measures
those corresponding to ustekinumab), which was not available (PASI 50, PASI 75 and PASI 90).
in the initial meta-analysis. In addition, a sensitivity analysis was carried out using
A second sensitivity analysis was carried out excluding phase efficacy results regardless of the weight-based dosing recom-
II clinical trials, which allowed for minimization of publication mended in the SmPC. In this case, the results obtained in
bias (Begg P-value = 0.119). In this case, three studies were the three meta-analyses (Tables 5, 6 and 7) were similar for
JEADV 2014, 28, 1633–1653 © 2013 European Academy of Dermatology and Venereology
Biologics in psoriasis: a meta-analysis at relevant time points 1645
%
Study Weight
ID RD (95% CI) (M-H)
–0.809 0 0.809
Figure 3b Forest plot depicting meta-analysis of PASI 75 efficacy results at the recommended time point for assessing failure as per the
approved SmPCs. RD, risk difference; M-H, Mantel-Haenszel; D+L, DerSimonian & Laird.
JEADV 2014, 28, 1633–1653 © 2013 European Academy of Dermatology and Venereology
1646 Lluıs Puig et al.
%
Study Weight
ID RD (95% CI) (M-H)
–0.623 0 0.623
Figure 3c Forest plot depicting meta-analysis of PASI 90 efficacy results at the recommended time point for assessing failure as per the
approved SmPCs. RD, risk difference; M-H, Mantel-Haenszel; D+L, DerSimonian & Laird.
the relevant parameters from a clinical practice perspective ustekinumab, to capture conditions where the recommended
(PASI 50 and PASI 75) for all biologics except for the two evaluation of response differs from the SmPC36 (namely, when
approved doses of ustekinumab, resulting in ustekinumab ustekinumab treatment should be stopped in case of inadequate
90 mg yielding a higher probability of response than us- response by 16 weeks after starting treatment) instead of efficacy
tekinumab 45 mg. at week 28 (after three doses), results at week 16 (after two
Finally, a fourth sensitivity analysis was carried out to capture doses) were explored. In the case of infliximab, to capture a
different time points for assessing response. In the case of probably incidental mistake in the SmPC13, results after three
JEADV 2014, 28, 1633–1653 © 2013 European Academy of Dermatology and Venereology
Biologics in psoriasis: a meta-analysis at relevant time points 1647
Table 4 Pooled efficacy data at the end of induction phase (week 24) (pooled risk difference vs. placebo)
Treatment PASI 50 (%) PASI 75 (%) PASI 90 (%)
[95% CI] [95% CI] [95% CI]
Adalimumab 64.3 (49.0–79.6) 63.3 (59.4–67.1) 45.7 (42.1–49.3)
Etanercept 62.3 48.8 25.4 (14.5–36.2)*
25 mg biw (53.1–71.5)* (37.7–59.9)*
or 50 mg qw
Etanercept ND 51.0 (43.6–58.4) ND
50 mg biw/
50 mg qw
Etanercept 68.1 (63.5–72.8) 55.3 (50.6–60.1) 27.8 (23.6–31.9)
50 mg biw‡
Infliximab 74.2 (69.1–79.4) 69.2 (64.1–74.4) 50.6 (45.3–55.9)
Ustekinumab 80.7 (77.2–84.2) 75.5 (71.5–79.4) 58.2 (53.7–62.8)
45 mg†
Ustekinumab 79.6 (74.9–84.2) 75.0 (69.3–80.7) 48.5 (41.7–55.3)
90 mg†
doses (week 14) were also explored instead of efficacy after four in which a decision should be made regarding withdrawal (and
doses (week 22). switching) of the biologic according to the approved SmPCs
In this case, the results obtained in the meta-analyses at the (week 12, 16, 14–22 or 28 for etanercept, adalimumab,
recommended time point for assessing primary failure with infliximab and ustekinumab, respectively), depending on
weight-based dosing as per the SmPC (Table 6) were similar for whether the threshold for efficacy is reached (namely, PASI 50
the clinically relevant parameters (PASI 50 and PASI 75) for us- response).13,37 Calculation of failure RDs is important for
tekinumab 45 mg, having the greatest probability of achieving cost-efficacy estimations, as both dose intensification and
PASI 50 and PASI 75 response. In terms of PASI 50, us- switching to another biologic imply cost increases (the loading
tekinumab 45 mg has the greatest probability of achieving phase of treatment is more expensive than maintenance
response, followed by ustekinumab 90 mg, infliximab, ada- treatment for any given biologic).
limumab and etanercept; the corresponding approximate failure Secondly, the doses in the clinical trials of ustekinumab did
rates are 22%, 26%, 30%, 35%, and 36%–43%, respectively, with not correspond in all patients to the approved weight-based
marked overlap of 95%CIs. In terms of PASI 75, ustekinumab dosing recommendations and therefore response rates should be
45 mg yielded the greatest probability (approximate RD 71%) of analysed by baseline weight groups (≤100 kg or >100 kg) to
achieving a satisfactory response, followed by infliximab, properly inform decision making by reflecting the approved
adalimumab, ustekinumab 90 mg, and etanercept, but the 95% conditions of use in clinical practice. As shown in our analysis,
CIs of both doses of ustekinumab, adalimumab and infliximab response rates to ustekinumab 90 mg when only patients
overlapped. weighing >100 kg are considered are lower than in the
original publications of ustekinumab trials, where baseline
Discussion randomisation was stratified according to weight (≤90 kg or
From a regulatory point of view, the approved SmPCs establish >90 kg),5,30,31 but doses were not weight-based.
the recommended framework of use in clinical practice, and in In addition, when it comes to the available meta-analyses,
this regard, the meta-analyses comparing the clinical efficacy of some of them do not include the most recently approved
biologics in the treatment of moderate-to-severe psoriasis that alternatives (i.e. ustekinumab), and the most recent clinical evi-
have been published to date usually present two important dence (i.e. comparative studies of briakinumab vs. etanercept) is
limitations. generally lacking.
First, comparative efficacy has been only estimated at the The relevance of this study, which seeks to minimize the
moment in which primary efficacy endpoint was measured in above mentioned limitations, lies in its potential application
the randomized clinical trials (week 10, 12 or 16 for infliximab, in clinical practice. The present analysis is not limited to
etanercept/ustekinumab and adalimumab respectively) or at the the effects of treatment on PASI 75 response rate at the pri-
end of the induction phase (week 24), and not at the time point mary endpoint of the clinical trials, with limited implications
JEADV 2014, 28, 1633–1653 © 2013 European Academy of Dermatology and Venereology
1648 Lluıs Puig et al.
%
Study Weight
ID RD (95% CI) (M-H)
–0.88 0 0.88
Figure 4a Forest plot depicting meta-analysis of PASI 50 efficacy results at the end of the induction phase (week 24). RD, risk differ-
ence; M-H, Mantel-Haenszel; D+L, DerSimonian & Laird.
in daily clinical practice, but includes other clinically rele- the first meta-analysis comparing efficacy of biologics at all three
vant time points and efficacy outcomes that guide therapeu- time points.
tic decisions such as withdrawal or optimization of the The meta-analysis at the end of induction phase (week 24)
biologic treatment (at time points defined by the SmPCs or offers the possibility to compare all biologics using the same
at the end of induction phase). time point. This represents an advantage with respect to previ-
Using one extensive data set provides robust insight of the ously published meta-analyses limited to primary endpoints and
comparative efficacy at three different time points, and this is allows for comparison of all biologics in the same conditions,
JEADV 2014, 28, 1633–1653 © 2013 European Academy of Dermatology and Venereology
Biologics in psoriasis: a meta-analysis at relevant time points 1649
%
Study Weight
ID RD (95% CI) (M-H)
–0.847 0 0.847
Figure 4b Forest plot depicting meta-analysis of PASI 75 efficacy results at the end of the induction phase (week 24). RD, risk differ-
ence; M-H, Mantel-Haenszel; D+L, DerSimonian & Laird.
when the maximal clinical efficacy is achieved regardless of the biologics (24 weeks, after the induction phase of treatment) may
speed of action of each agent. Choosing week 12 as a primary serve as basis for eventual economic analyses.
endpoint for head-to-head comparative trials including etaner- The main differences between the present and a previously
cept5,22,23 has been criticized on account that 12 weeks is not an published meta-analysis10 of efficacy outcomes at the end of
adequate time frame to assess the efficacy of etanercept,38 and induction phase (week 24) reside in our including the weight-
providing a similar time frame for efficacy comparison of all based dosing efficacy data for ustekinumab as well as the placebo
JEADV 2014, 28, 1633–1653 © 2013 European Academy of Dermatology and Venereology
1650 Lluıs Puig et al.
%
Study Weight
ID RD (95% CI) (M-H)
–0.686 0 0.686
Figure 4c Forest plot depicting meta-analysis of PASI 90 efficacy results at the end of the induction phase (week 24). RD, risk differ-
ence; M-H, Mantel-Haenszel; D+L, DerSimonian & Laird.
response (in this study a LOCF placebo response has been used), the clinical relevance of these results, especially taking into
and our wider selection of studies. account that not all these time points are those when treatment
In the short term (at the primary endpoint), infliximab is the failure or efficacy should be assessed according to the SmPCs or
biologic with the highest RDs as regards both PASI 75 and PASI clinical guidelines.
90 responses, followed by ustekinumab 45 mg, ustekinumab The comprehensive set of sensitivity analyses we have carried
90 mg, adalimumab and etanercept. Nevertheless, using differ- out confirms the robustness of our results and highlights
ent time points for different biologics (10, 12 or 16 weeks) limits the need for efficacy comparison according to the weight-based
JEADV 2014, 28, 1633–1653 © 2013 European Academy of Dermatology and Venereology
Biologics in psoriasis: a meta-analysis at relevant time points 1651
Table 5 Pooled efficacy data at the primary endpoint times. Sensitivity Analysis (pooled risk difference vs. placebo)
Treatment PASI 50 (%) PASI 75 (%) PASI 90 (%)
[95% CI] [95% CI] [95% CI]
I. inclusion of studies without placebo control arm
Adalimumab 66.4 (62.4–70.5) 64.3 (61.3–67.3) 40.7 (33.2–48.3)*
Etanercept 25 mg biw or 50 mg qw 58.3 (47.2–69.5)* 35.5 (27.8–43.2)* 13.8 (7.9–19.6)*
Etanercept 50 mg biw 63.9 (58.5–69.2)* 44.5 (39.7–49.4)* 17.9 (14.0–21.7)*
Infliximab 78.5 (75.5–81.5) 72.5 (69.9–75.1) 49.1 (42.3–56.0)*
Ustekinumab 45 mg 77.0 (73.7–80.4) 69.7 (66.0–73.4) 45.6 (41.6–49.6)
Ustekinumab 90 mg 77.5 (73.1–81.8) 65.2 (59.9–70.6) 34.4 (29.9–39.8)
II. phase II studies excluded
Adalimumab 66.8 (62.6–70.9) 64.0 (60.2–67.8) 42.5 (38.9–46.0)
Etanercept 25 mg biw or 50 mg qw 51.3 (45.8–56.7) 31.4 (26.7–36.2) 10.7 (7.6–13.8)
III. efficacy regardless of patient’s weight
Ustekinumab 45 mg [all weights] 73.5 (69.9–77.1) 63.4 (59.6–67.2) 40.8 (37.0–44.7)
Ustekinumab 90 mg [all weights] 77.9 (74.6–81.3) 68.7 (65.0–72.3) 44.2 (40.4–48.1)
dosing recommendations for ustekinumab, especially as regards mately 19%, much lower than those of infliximab (30%) ada-
primary endpoint calculations. In this regard, the efficacy results limumab (34%) or etanercept (38% to 48%), with no
of ustekinumab differ significantly from those in other published overlapping of the 95%CIs, except for infliximab (Table 6) and
meta-analyses10 and cost-efficacy calculations12 because of their adalimumab. This has potentially important economic implica-
usage of efficacy data not derived from weight-based dosing, tions, taking into account the cost of switching to another bio-
which limits their usefulness or applicability to daily clinical logic therapy (higher than that of maintenance), which should
practice (to date, weight-based dosing of ustekinumab has been be considered in this significant proportion of patients.
taken into account in only one meta-analysis).9 As regards pooled efficacy data at the end of the induction
Pooled efficacy data at the recommended time points to assess phase (week 24), the PASI 75 RDs (Table 4) are highest for
primary failure (Table 3) show that the probability of primary ustekinumab 45 mg (76%) and ustekinumab 90 (75%), followed
failure with ustekinumab, regardless of the dose, is approxi- by infliximab (69%, with 95% CI values overlapping those of
Table 6 Pooled efficacy data at the recommended time point for assessing failure as per approved SmPCs. Sensitivity Analysis (pooled
risk difference vs. placebo)
Treatment PASI 50 (%) PASI 75 (%) PASI 90 (%)
[95% CI] [95% CI] [95% CI]
I. inclusion of studies without placebo control arm
Adalimumab 66.4 (62.4–70.5) 64.2 (61.2–67.2) 40.7 (33.2–48.3) *
Etanercept 25 mg biw or 50 mg qw 58.3 (47.2–69.5)* 35.5 (27.8–43.2)* 13.8 (7.9–19.6) *
Etanercept 50 mg biw 63.9 (58.5–69.2)* 44.5 (39.7–49.4)* 17.9 (14.0–21.7)*
Infliximab [week 22] 70.4 (66.6–74.3) 68.0 (64.7–71.4) 45.6 (41.7–49.6)
Infliximab [week 14] 75.3 (71.8–78.9) 68.9 (65.5–72.2) 46.2 (42.3–50.2)
II. phase II studies excluded
Adalimumab 66.8 (62.6–70.9) 64.0 (60.2–67.8) 42.5 (38.9–46.0)
Etanercept 25 mg biw or 50 mg qw 51.3 (45.8–56.7) 31.4 (26.7–36.2) 10.7 (7.6–13.8)
III. efficacy regardless of patient’s weight
Ustekinumab 45 mg [all weights] 79.8 (76.6–83.1) 64.9 (61.1–68.7) 44.1 (40.3–48.0)
Ustekinumab 90 mg [all weights] 82.0 (78.9–85.1) 72.3 (68.7–75.8) 51.6 (47.7–55.4)
IV. different time point for evaluating response
Infliximab [week 14] ND 70.9 (63.7–78.0) ND
Ustekinumab 45 mg [week 16] 77.5 (73.7–81.2) 70.9 (66.7–75.1) 49.6 (45.0–54.2)
Ustekinumab 90 mg [week 16] 74.0 (68.6–79.4) 63.2 (56.8–69.7) 38.7 (32.0–45.3)
JEADV 2014, 28, 1633–1653 © 2013 European Academy of Dermatology and Venereology
1652 Lluıs Puig et al.
Table 7 Pooled efficacy data at the end of induction phase (week 24). Sensitivity Analysis (pooled risk difference vs. placebo)
Treatment PASI 50 (%) [95% CI] PASI 75 (%) [95% CI] PASI 90 (%) [95% CI]
I. inclusion of studies without placebo control arm
Adalimumab 64.3 (49.0–79.6) 63.3 (59.4–67.1) 40.5 (25.1–55.8)*
Etanercept 25 mg biw or 50mg qw 68.0 (53.9–82.1)* 54.6 (44.2–65.1)* 31.2 (16.6–45.8)*
Etanercept 50 mg biw/50 mg qw 73.5 (65.6–81.5) 58.2 (48.5–67.9)* 32.4 (24.3–40.4)
Etanercept 50 mg biw† 68.1 (63.5–72.8) 55.3 (50.6–60.1) 27.8 (23.6–31.9)
Infliximab 73.6 (70.7–76.6) 68.8 (65.8–71.8) 47.3 (44.2–50.5)
II. phase II studies excluded
Adalimumab ND 63.5 (59.5–67.5) 47.0 (43.3–50.7)
Etanercept 25 mg biw or 50 mg qw 59.2 (52.0–66.4) 48.4 (34.5–62.4) 24.2 (18.8–29.6)
III. efficacy regardless of patient’s weight
Ustekinumab 45 mg [all weights] 79.1 (75.8–82.4) 69.3 (65.6–72.9) 50.3 (46.4–54.2)
Ustekinumab 90 mg [all weights] 81.5 (78.4–84.6) 78.1 (74.8–81.4) 58.2 (54.3–62.0)
JEADV 2014, 28, 1633–1653 © 2013 European Academy of Dermatology and Venereology
Biologics in psoriasis: a meta-analysis at relevant time points 1653
13 Puig L. Induction phase, primary endpoint, time to decide on primary 27 Gordon K, Langley R, Leonardi C et al. Clinical response to adalimumab
failure, and therapeutic goals in biologic treatment of psoriasis. J Eur Acad treatment in patients with moderate to severe psoriasis: double-blind,
Dermatol Venereol 2013; 27: e257–e260. randomized controlled trial and open- label extension study. J Am Acad
14 Gottlieb A, Matheson R, Lowe N et al. A randomized trial of etanercept Dermatol 2006; 55: 598–606.
as monotherapy for psoriasis. Arch Dermatol 2003; 139: 1627–1632. 28 Saurat JH, Stingl G, Dubertret L et al. Efficacy and safety results from the
15 Reich K, Nestle F, Papp K et al. Infliximab induction and maintenance randomized controlled comparative study of adalimumab vs
therapy for moderate-to-severe psoriasis: a phase III, multicentre, methotrexate vs placebo in patients with psoriasis (CHAMPION). Br J
double-blind trial. Lancet 2005; 366(9494): 1367–1374. Dermatol 2008; 158(3): 558–566.
16 Higgins J, Thompson S. Quantifying heterogeneity in a meta-analysis. 29 Menter A, Tyring S, Gordon K et al. Adalimumab therapy for moderate
Stat Med 2002; 21:1539e58. to severe psoriasis: A randomized, controlled phase III trial. J Am Acad
17 Higgins JPT, Green S, eds. Cochrane handbook for systematic reviews of Dermatol 2008; 58(1): 106–115.
interventions. Version 5.1.0 [updated March 2011]. The Cochrane Colla- 30 Leonardi C, Kimball A, Papp K et al. Efficacy and safety of ustekinumab,
boration, 2011. URL www.cochrane-handbook.org (last accessed: 22 a human interleukin-12/23 monoclonal antibody, in patients with
December 2012). psoriasis: 76-week results from a randomised, double-blind, placebo-con-
18 Leonardi CL, Powers JL, Matheson RT et al. Etanercept as monotherapy trolled trial PHOENIX 1. Lancet 2008; 371: 1665–1674.
in patients with psoriasis. N Engl J Med 2003; 349: 2014–2022. 31 Papp K, Langley R, Lebwohl M et al. Efficacy and safety of ustekinumab,
19 Papp KA, Tyring S, Lahfa M et al. A global phase III randomized con- a human interleukin-12/23 monoclonal antibody, in patients with
trolled trial of etanercept in psoriasis: safety, efficacy, and effect of dose psoriasis: 52-week results from a randomised, double-blind, placebo-con-
reduction. Br J Dermatol 2005; 152: 1304–1312. trolled trial PHOENIX 2. Lancet 2008; 371: 1675–1684.
20 Tyring S, Gottlieb A, Papp K et al. Etanercept and clinical outcomes, 32 Sterry W, Ortonne JP, Kirkham B et al. Comparison of two etanercept
fatigue, and depression in psoriasis: double-blind placebo-controlled regimens for treatment of psoriasis and psoriatic arthritis: PRESTA
randomised phase III trial. Lancet 2006; 36: 29–35. randomised double blind multicentre trial. BMJ 2010 Feb; 2(340): c147.
21 van de Kerkhof PC, Segaert S, Lahfa M et al. Once weekly administration 33 Strohal R, Puig L, Chouela E et al. The efficacy and safety of etanercept
of etanercept 50 mg is efficacious and well tolerated in patients with when used with as-needed adjunctive topical therapy in a randomised,
moderate-to-severe plaque psoriasis: a randomized controlled trial with double-blind study in subjects with moderate-to-severe psoriasis (the
open-label extension. Br J Dermatol 2008; 159: 1177–1185. PRISTINE trial). J Dermatolog Treat 2013; 24: 169–178.
22 Strober BE, Crowley JJ, Yamauchi PS, Olds M, Williams DA. Efficacy and 34 Barker J, Hoffmann M, Wozel G et al. Efficacy and safety of infliximab vs.
safety results from a phase III, randomized controlled trial comparing the methotrexate in patients with moderate to severe plaque psoriasis: results
safety and efficacy of briakinumab with etanercept and placebo in patients of an open-label, active-controlled, randomized trial (RESTORE 1). Br J
with moderate to severe chronic plaque psoriasis. Br J Dermatol 2011; Dermatol 2011; 165: 1109–1117.
165: 661–668. 35 Thaci D, Ortonne P, Chimenti S et al. A phase IIIb, multicentre,
23 Gottlieb AB, Leonardi C, Kerdel F et al. Efficacy and safety of bria- randomized, double-blind, vehicle-controlled study of the efficacy
kinumab vs. etanercept and placebo in patients with moderate to severe and safety of adalimumab with and without calcipotriol/beta-
chronic plaque psoriasis. Br J Dermatol 2011; 165: 652–660. methasone topical treatment in patients with moderate to severe
24 Chaudhari U, Romano P, Mulcahy LD, Dooley LT, Baker DG, Gott- psoriasis: the BELIEVE study. Br J Dermatol 2010; 163:
lieb AB. Efficacy and safety of infliximab monotherapy for plaque- 402–411.
type psoriasis: a randomised trial. Lancet 2001; 357(9271): 1842– 36 NICE Technology appraisal, Psoriasis - ustekinumab (TA180).
1847. Ustekinumab for the treatment of adults with moderate to severe
25 Gottlieb AB, Evans R, Li S et al. Infliximab induction therapy for psoriasis. 2009. reviewed on January 2010. URL http://www.nice.org.
patients with severe plaque-type psoriasis: a randomized, double- uk/nicemedia/pdf/TA180Guidance.pdf (last accessed: 22 December
blind, placebo-controlled trial. J Am Acad Dermatol 2004; 51: 534– 2012).
542. 37 Puig L, Dauden E, Carrascosa JM. Commentary on European and British
26 Menter A, Feldman S, Weinstein G et al. A randomized comparison of guidelines for the treatment of psoriasis. Actas Dermosifiliogr 2010; 101:
continuous vs intermittent infliximab maintenance regimens over 1 year 285–290.
in the treatment of moderate-to- severe plaque psoriasis. J Am Acad 38 Kuhn A, Luger TA. Psoriasis: is ustekinumab superior to etanercept for
Dermatol. 2007; 56:31.e1–31.e15. psoriasis? Nat Rev Rheumatol 2010; 6: 500–501.
JEADV 2014, 28, 1633–1653 © 2013 European Academy of Dermatology and Venereology