Results
Results are presented following the GRADE informative statements (35).
The systematic search retrieved 4377 citations. After excluding
duplicates and screening the title and abstract, 29 full text papers
were retrieved for the evaluation of dupilumab’s efficacy and safety
(figure 1A). Twenty-two studies were excluded due to lack of abstract,
dose not approved by the regulatory authorities, and duplicate data.
Nine additional articles were suggested by the GDG group but excluded
due to dose not approved by the regulatory authorities, non-randomised
double-blind study design, not reporting outcomes of interests, or
duplicate data (Table S3). The SR for the efficacy and safety included
seven RCTs (36, 37, 38, 39, 40, 41) (figure 1A). For the economic
evidence, after screening 1552 hits, five studies were considered
suitable for inclusion (51,52, 53, 54,55) (figure 1B).
Characteristic of included
studies
The main characteristics of the studies included are detailed in Tables
S4 and S5. The RCTs included in the SR evaluated 1678 adults and 167
adolescents with moderate-to-severe AD inadequately controlled by
topical treatment. Follow-up under treatment ranged from 16 weeks
(36,37,39, 40) to one year (38). One RCT recruited responders from SOLO
trials and continued the intervention for another 36 weeks (41). In all
trials evaluated only regulatory-approved doses were considered.
Evidence of efficacy and safety
The summary of findings and certainty of evidence per
outcome are reported in Tables 3, 4, 5A and
5B.
SCORAD index
Six RCTs included in the SR reported the percentage change from baseline
in SCORAD index assessed at 16 weeks (36,37,39,40). Dupilumab reduced
with high certainty of evidence the SCORAD value compared to standard of
care in adults (MD -30,72%; 95% CI -34,65% to -26,79%) and in
adolescents (MD -34%; 95% CI -43.74% to -24.26%). One study reported
on SCORAD reduction at 52 weeks (MD -32.1%; 95%CI, -39.27% to
-24.93%) (38). Another RCT reported a small to no effect in the 36
weeks follow-up of SOLO trials (MD + 0.97%; 95%CI +0.69% to +1.25%)
(41).
Eczema Area and Severity Index
(EASI)
Six RCTs reported the proportion of patients achieving 75% improvement
(EASI-75) at 16 weeks (36,37,38,39,40). Dupilumab treatment in adult
patients with standard of care versus placebo resulted with high
certainty in a significant increase in the number of patients who
achieved EASI-75 (RR 3.09; 95%CI 2.45 to 3.89; absolute increase +383
per 1000 patients, 95%CI from +266 to +530). Similar results were
reported for adolescents (RR 5.03; 95%CI 2.37 to 10.71; absolute
increase +332 per 1000 patients, 95%CI from +113 to +800). All six RCTs
also reported the proportion of patients with 50% improvement (EASI-50)
at 16 weeks. Their results showed a significant increase in EASI-50
responders compared to standard of care in adults (RR 2.43; 95%CI 2.04
to 2.89) and adolescents (RR 4.71; 95%CI 2.64 to 8.40). A comparable
increase was reported at 52 weeks for EASI-75 (RR 3.02; 95%CI 2.29 to
3.98) and EASI-50 (RR 2.63; 95% CI 2.12 to 3.26) in one RCT (38). The
impact on EASI was maintained during the 36 weeks of follow-up in the
SOLO trials (EASI-75 RR 2.36, 95%CI 1.66 to 3.34; EASI-50 RR 1.85,
95%CI 1.39 to 2.44) (41).
Pruritus
Six RCTs measured the effect of dupilumab treatment on pruritus through
the proportion of patients with an improvement of ≥ 4 points in the
numerical rating scale (NRS) at 16 weeks (36,37,38,39,40). Dupilumab
significantly reduced pruritus with high certainty of evidence, both for
adults (RR 2.96; 95% CI; 2.37 to 3.70; absolute effect + 311 per 1.000
patients, 95%CI from +217 to +429) and for adolescents (RR 7.68; 95%
CI; 2.83 to 20.84; absolute effect +318 per 1.000 patients, 95%CI from
+ 87 to +945). One study (38) reported a significant reduction of
pruritus at 24 weeks (RR 3.98; 95%CI 2.71 to 5.84) and at 52 weeks (RR
3.36; 95% CI 2.45 to 4.60), and the effect was maintained during the 36
weeks follow-up (SOLO trials; RR 3.83, 95%CI 2.10 to 6.97) (41). The
effect on pruritus was also quantified by percent change from baseline
of peak pruritus NRS score (36,37,38,39,40,41). The pooled analysis
illustrated a significant dupilumab-induced NRS score improvement at 16
weeks for adults (MD -28.04%; 95% CI -32.65% to -23.43%) and for
adolescents (MD -28.90%; 95% CI -39.34% to -18.46%). In the 36 weeks
follow-up period of SOLO trials no improvement in the NRS score was
reported (MD -0.53%, 95%CI -0.79 to -0.26) (41).
Safety
Dupilumab may increase (low certainty of evidence) treatment-related AE
at 16 weeks (RR 1.29; 95%CI 0.62 to 2.72; absolute increase +118 per
1000 patients; 95% CI from -155 to +702) (36,39). Most of the
treatment-related AEs were eye inflammation (conjunctivitis). Dupilumab
was safe in adolescents: the analysis of the potential increase in
dupilumab-related AE showed little to no difference with moderate
certainty (RR 1.04; 95%CI 0.85 to 1.26; absolute increase +28 per 1000
patients; 95%CI from - 104 to + 180) (40). Dupilumab-related severe AE
were reduced for adults (RR 0.50; 95%CI 0.09 to 2.70; absolute increase
-12 per 1.000 patients; 95%CI from -22 to +40) and for adolescents (RR
0.35; 95%CI 0.01 to 8.36, absolute increase -8 per 1000 patients;
95%CI from -12 to +87). The evidence is very uncertain both for adults
and adolescents. The SOLO trials reported decreased treatment-related AE
(RR 0.86, 95%CI 0.75 to 1.00) and increase treatment-related severe AE
(RR 2.95, 95%CI 0.36 to 24.07) (41).
Investigator’s Global Assessment (IGA) score
Four RCTs defined the primary outcome as the proportion of patients who
achieved both a score of 0/1 (0=clear or 1=almost clear) on the
investigator’s global assessment and a reduction of ≥ 2 points from
baseline at 16 weeks (37,38,39). Dupilumab significantly increased the
proportion of patients’ achieving both end-points with high certainty of
evidence (RR 3.46; 95% CI 2.79 to 4.3; absolute effect + 270 per 1.000
patients, 95%CI from + 197 to +363). Two other RCTs (36,40) defined the
primary outcome as the proportion of patients with an IGA response 0/1
at 16 weeks, showing a significant effect both for adults (RR 18.11;
95%CI 2.50 to 131.17) and for adolescents (RR 10.37; 95%CI 2.50 to
42.95). The effect was maintained in the 36 weeks follow-up of SOLO
trials (RR 2.47, 95%CI 1.65 to 3.71) (41).
Use of rescue medication
Five RCTs reported on this outcome at 16 weeks (37,38,39,40). The pooled
analysis showed that dupilumab significantly reduces with high certainty
of evidence the proportion of the patients who use any rescue
medication, both for adults (RR 0.36; 95%CI 0.28 to 0.46, absolute
effect - 270 per 1.000 patients, 95%CI from - 304 to - 228 fewer) and
for adolescents (RR 0.35; 95%CI 0.22 to 0.56, absolute effect - 382 per
1.000 patients, 95%CI from -45 to -259). The effect was maintained in
the 36 weeks follow-up of SOLO trials (MD 0.69, 95%CI 0.43 to 0.96)
(41).
Pain
One RCT included in the SR measured the effect of dupilumab on pain
trough the proportion of patients with no complains in the item 4
(pain/discomfort) of the EQ-5D questionnaire. For the adult population,
dupilumab significantly reduced the number of patients with pain and
discomfort, with high certainty of evidence (RR 1.89; 95%CI 1.44 to
2.49; absolute effect + 330 more per 1,000 patients; 95%CI +163 to
+552) (39).
Sleep disturbance
Six RCTs measured the impact on sleep disturbance with the change in the
POEM score at 16 weeks. Dupilumab significantly reduced the severity of
sleep disturbance (MID=4) with high certainty of evidence in adults (MD
-7.29; 95%CI -8.23 to -6.35) (36,37,38,39) and with moderate certainty
of evidence in adolescents (MD -6.30; 95%CI -8.81 to -3.79) (40). One
RCT evaluated POEM at 52 weeks and showed a similar effect (MD is -8.4;
95%CI -10.12 to -6.68) (38). In the 36 weeks follow-up of SOLO trials,
an opposite effect was reported (MD 0.96, 95%CI 0.68 to 1.23) (41).
Anxiety and depression
Six RCTs reported this outcome considering the change from baseline of
HADS at 16 weeks. The pooled analysis showed that dupilumab reduces
symptoms of anxiety and depression with high certainty of evidence in
adults (MD -3.08; 95%CI -4.41 to -1.75) and with moderate certainty in
adolescents (MD -1.30; 95%CI -3.38 to +0.78). In the 36 weeks follow-up
of SOLO trials an opposite effect was reported (MD 0.31, 95%CI 0.04 to
0.57) (41). Two RCTs (38,39)) evaluated this outcome as the proportion
of the patients with no clinically relevant symptoms of anxiety and
depression at 16 weeks (RR 1.78; 95% CI 1.35 to 2.33) and one of the
two RCTs reported a better effect at 52 weeks (RR 2.40; 95%CI 1.5 to
3.87) (38).
Quality of life
QoL outcome was measured by DLQI for adults (36,37,38,39) and by CDLQI
for adolescents (40). The pooled analysis showed a significant
improvement (above the MID of 4 and 6, respectively) in the QoL with
high certainty of evidence both for adults (MD -4.80; 95%CI -5.55 to
-4.06) and for adolescents (MD -13.60; 95%CI -15.13 to -12.07). Four
RCTs reported on the QoL in adults with AD measured by Global Individual
Sign Score (GISS) (37, 38, 39) and showed that dupilumab improves QoL
(MD -26.39%; 95% CI -30.62% to -22.15%). Dupilumab improves
QoL at 52 weeks: DLQI (MD -4.40; 95%CI -5.7 to -3.05) and GISS (MD
-29.10%; 95%CI from -36.67% to -21.53%) (38). In the 36 weeks
follow-up of SOLO trials, deterioration of QoL was reported (MD 0.74,
95%CI 0.47 to 1.01) (41).
Cost-effectiveness
Four Markov model-based evaluations assessing dupilumab versus standard
of care (42, 43, 44, 45), and one comparing dupilumab with best
supportive care (education, psychological support, emollients, topical
corticosteroids, bandages, and hospitalisation) (46) were included into
the analysis. Three evaluations were conducted from the perspective of
the United States’ healthcare system (42,43,44), one was performed in
Canada (45), and one in the UK (46). The annual dupilumab related cost
per patient was highest in the US studies (up to 37,000 US$) followed
by the Canadian study (25,918 C$) and by the UK study (16,500 £). The
costs of medication were lower in the UK (632.45 £) compared with Canada
(959.94 C$) and the US (up to 1,300.00 US$). The ICER per QALY of
dupilumab added to the standard of care was 100,000 US$ or higher
(42,43,44,45). The sensitivity analyses showed variations in the ICER
from 78,300 US$ in patients with severe AD to 159,988 US$ in those
with moderate AD. The Canadian Agency for Drugs and Technologies in
Health (CADTH) undertook an analysis for patients, refractory to, or
ineligible for systemic immunosuppressant therapies obtaining an ICER of
133,877 C$, which is within the range of ICERs found by the US studies
(Table 5A). The National Institute for Clinical Excellence undertook an
analysis of dupilumab as fifth-line treatment, after topical therapies
and systemic immunosuppressant have failed. In this scenario the ICER
value was 28,495.00 £, which is lower than the previous ones (Table 5B).
There was moderate certainty of the evidence for the studies assessing
the economic impact of dupilumab versus standard of care due to concerns
for indirectness, as unitary costs were provided by studies performed in
the US (42,43,44) and Canada (45). These results may not apply outside
high-income countries. There was low certainty for the study that
compared dupilumab with the best supportive care due to serious concerns
for indirectness in the comparator (including therapies beyond topical
treatment) and the population (patients receiving dupilumab as
fifth-line treatment, after systemic immunosuppressant therapies).
Discussion
Main findings
The current systematic review showed that dupilumab as add-on treatment
for moderate-to-severe AD in adults and adolescents significantly
reduces short-term (16 weeks) AD symptoms, severity, use of rescue
medication, and improves quality of life. For adults there is good
evidence for long-term efficacy (52 weeks). Dupilumab may increase
short-term drug-related AE. The evidence for severe drug-related AE is
very uncertain. All RCTs were mainly powered for efficacy and less
powered to show rare adverse events which are now frequently reported in
the post-marketing literature.
The ICER per QALY of dupilumab versus standard of care was above 100,000
US$, considered as threshold for the willingness to pay in several
high-income countries (47). Drug-related costs were the key driver of
this ICER in all studies. The CADTH analysis recommended a price
reduction of at least 54% to obtain an ICER value below the threshold
of 50,000 C$. Another important factor impacting the ICER was the
profile of patients included in the analysis. In the NICE analysis,
dupilumab plus topical corticosteroids was found to be cost-effective
for treating atopic dermatitis not responding to other systemic
therapies, such as ciclosporin, methotrexate, azathioprine and
mycophenolate mofetil, or when these options were contraindicated or not
tolerate. Dupilumab improved their quality of life compared to best
supportive care and it was key for generating an acceptable ICER value
of 28,500 £, which is in line with previous authors suggesting that the
high cost of dupilumab for severe AD is offset by the quality of life
improvement (48).
The main reason to downgrade the certainty of the evidence for the
efficacy and safety outcomes was imprecision and inconsistency, and the
indirectness for economic data. All the studies were funded by the same
company and reported positive effects, which might raise concerns for a
potential sponsorship bias. Moderate certainty of evidence for economic
impact was available from three studies with low risk of bias but with
important indirectness. All economic analyses were performed in
high-income countries in line with their health system perspectives,
thus their results may not be applicable to other countries.
Current report in the context of previous
research
This SR is the most up to date review on the effectiveness, safety and
economic impact on dupilumab in AD. Similar to previous SRs the current
analysis reinforces the short-term (16 weeks) efficacy of dupilumab in
improving SCORAD, EASI, IGA, pruritus, and quality of life (49,50,51).
In addition, the current SR provides evidence for long-term (52 weeks)
benefit in adults.
According to the current SR in adults with AD dupilumab may increase
treatment-related AE (conjunctivitis/ injection side reactions/
eosinophilia), although there is low certainty of evidence. The evidence
for treatment-related severe AE is very uncertain both for adults and
for adolescents. The pooled analysis from laboratory findings from three
randomized, double-blinded, placebo-controlled phase 3 trials showed no
clinically important changes in routine laboratory parameters that could
be attributed to dupilumab, thus supporting the use of dupilumab as a
systemic treatment for moderate-to-severe AD that does not require
laboratory monitoring (52).
The use of variable outcomes limited the conclusions of previous SRs
(50). For the current SR the GDG predefined and prioritised AD-related
outcomes.
Previous SRs included all dupilumab doses for AD, while the current SR
only included FDA/EMA approved doses, which are more informative for
issuing recommendations for clinical practice.
Although a recent SR included the efficacy of dupilumab for adolescent
AD population (51), they did not report separately for this population.
Furthermore, the current SR followed the GRADE approach for rating of
the certainty of evidence. In contrast to previous SRs (49,50) that
assessed only the risk of bias, the current SR considered all relevant
aspects related with the certainty of evidence like heterogeneity,
indirectness or imprecision of the results.
Finally, an evaluation of cost-effectiveness was included, thus
providing additional support for the GDG in formulating recommendations.
Limitations and
strengths
The current SR has several strengths. First, a comprehensive systematic
search was conducted on three databases, checking for efficacy, safety
and cost-effectiveness. Second, rigorous evaluation methods were
employed, including the use of the GRADE approach to rate the certainty
of the evidence. The outcomes included were prioritised beforehand and
the minimal important difference was considered when available for all
AD-related outcomes.
Optimal presentation of results into tabulated format (summary of
findings) is provided aiming to improve communication to patients,
clinicians, and other stakeholders.
There are some limitations as well. Only studies published in English
were included. However, the studies included in previous systematic
reviews were thoroughly evaluated and additional studies were suggested
by the GDG, which decreases the possibility of missing studies. No
observational studies were included, which could inform better on
outcomes with low quality of evidence (i.e, serious adverse events).
However, they will be considered when formulating recommendations. A
‘de novo ’ economic analysis was not conducted; however, a
rigorous and explicit critical appraisal and transferability assessment
of cost-effectiveness data is provided.
Conclusion
Dupilumab demonstrated a significant short-term benefit for the adults
and adolescents with uncontrolled moderate-to-severe atopic dermatitis,
by improving symptoms and disease severity, reducing the use of rescue
medications and improving the quality of life. For adults there is
evidence for long-term benefit. Thresholds for cost-effectiveness are
probably acceptable for some high-income countries, however dupilumab
might not be equally cost-effective in countries with limited resources.
Although short term safety data showed no visible increase of AE, more
accurate AE reporting is warranted in RCTs for both adult and adolescent
population, combined with long-term safety evaluation using
observational and effectiveness studies and registries. There are
several ongoing open-label studies (53,54) and registries (55)
evaluating the long-term safety and efficacy of dupilumab in atopic
dermatitis that are likely to be informative in formulating
recommendations.