Molecular Heterogeneity, Biomarker Discovery, and Targeted Therapy in Asthma - Genomic Biomarkers for Pharmaceutical Development - page 80

selected on the basis of ≥3% sputum eosinophils, and showed significant benefit in terms of
lung function, with trends toward effects on ACQ and exacerbation rates [66] . Interestingly,
patients in that study with concomitant nasal polyposis, a common comorbidity of eosino-
philic asthma particularly evident in patients with aspirin-intolerant asthma, exhibited the
greatest magnitude of clinical benefit. A third agent targeting IL5 is benralizumab, an anti-
body directed against IL5Rα that is cytotoxic to eosinophils. In a single dose study in severe
eosinophilic asthmatics presenting with exacerbations, the rate of exacerbations in the sub-
sequent 24 weeks was significantly reduced by benralizumab treatment [67] . Taken together,
these relatively small proof-of-concept studies demonstrate that targeting IL5 is capable of
partially reducing the rate of exacerbations in asthma patients that have evidence of eosino-
philic airway inflammation and a past history of asthma exacerbations.
To build on the small proof-of-concept studies showing that IL5 blockade could poten-
tially reduce the rate of exacerbations in moderate-severe asthmatics with evidence of eosino-
philic airway inflammation, a larger study ('DREAM') was conducted, in which 616 patients
were randomized (1:1:1:1) to receive placebo or one of three dose levels of mepolizumab for
52 weeks [68] . The primary outcome measure was the rate of clinically significant asthma
exacerbations during the treatment period. For study entry, patients were required to have
had a history of at least two exacerbations in the previous year and evidence of eosinophilic
airway inflammation by fitting at least one of the following criteria: sputum eosinophil count
≥3%, FeNO ≥50 ppb, blood eosinophils ≥0.3×10 9 /L, or 'prompt deterioration of asthma con-
trol after a 25% or less reduction in regular maintenance inhaled or oral corticosteroids'. With
this study design, statistically significant reductions in exacerbation rates were observed in
all dose arms, with the exacerbation rate in treated patients reduced by 48%, 39%, and 52%
in the three active arms relative to the placebo arm with no apparent dose ranging effect.
While blood and sputum eosinophils were significantly reduced in the mepolizumab-
treated patients as compared to placebo-treated patients, there were no significant effects on
FeNO, FEV1, or ACQ in this study. Overall, the DREAM study provided further compelling
evidence that IL5 blockade may significantly reduce asthma exacerbations provided that
patients with evidence of eosinophilic airway inflammation and a recent history of exacerba-
tions are selected.
However, it remains unclear how the patient selection criteria (i.e., either sputum or blood
eosinophilia, or elevated FeNO, or deterioration of control upon steroid reduction) will be
translated into practicable guidelines from the standpoints of regulatory approval and broad
clinical applicability. In particular, no single study has been described in which patients pre-
dicted to benefit from IL5 blockade (i.e., patients with elevated sputum or blood eosinophils
or FeNO) were compared directly to patients predicted not to benefit (i.e., patients below the
threshold for all three biomarkers). While it may be reasonable to assume that IL5 blockade
might benefit patients with 'eosinophilic asthma', however it may be defined, in order to prove
this hypothesis it is equally important to demonstrate that the 'diagnostic-negative' population
fails to show benefit as it is to demonstrate that the 'diagnostic-positive' population does show
benefit (provided that exposure to drug is deemed not to pose significant safety risks to the
'diagnostic-negative' population). A corollary of this point is the definition of a suitable cutoff
for the diagnostic biomarker: while 3% sputum eosinophils, 0.3×10 9 blood eosinophils/L, or
50ppb FeNO may sound like reasonable cutoffs based on observational studies, it is unclear
whether those cutoffs are optimal for selecting IL5 blocking treatment in individual patients.