Statistical analysis
The primary outcome of this study was the change in the endoscopic nasal
polyp score at 20 weeks treatment compared to baseline for both placebo
and benralizumab treatment groups. This was analyzed using a two-tailed
paired t-test and with 0.05 determined to be significant a
priori . Lund-Mackay score as determined by CT scan, nasal blockage
score (NBS), SNOT-22 and UPSIT smell test were evaluated by the same
statistical methods. Baseline characteristics and laboratory values were
compared by either an unpaired two-tailed t-test or Fisher’s exact test
when required. Among the benralizumab treated subjects, blood eosinophil
count and skin prick test sensitivity were also compared against
endoscopic NP score reduction or CT scan by using a one-tailed Spearman
rank correlation and then plotted with a four-parameter logistic curve
interpolation.
An a priori power analysis was performed based on prior studies
using anti-IL5 monoclonal antibody therapy where there was an average of
32% reduction in nasal polyp score yielding an effect size of 0.535 and
standard deviation of 1.5. To achieve 80% power (alpha 0.05) using the
same effect size, a total sample size of 22 was determined to be
adequate for our study (11 per arm). To accommodate for potential
dropout, we randomized 24 total subjects. An intention to treat analysis
was performed.
Randomization and data monitoring was performed by an independent team.
An independent and blinded group was also responsible for data
monitoring and receipt of locked study data prior to unblinding. Two
separate groups then analyzed the data. Data analysis was performed
using Prism version 8.4.1 for Mac (GraphPad Software, San Diego, CA).