Abstract
Purpose :
Randomized, vehicle-controlled, and double-masked conjunctival allergen challenge (CAC) studies have been successfully used to support marketing applications for treatments of allergic conjunctivitis. These studies have been conducted using standardized challenge methodologies and the endpoints have been evaluated with consistent approach, therefore it may be feasible to use hybrid control arms in future CAC trials. The purpose of hybrid control is to pool the trial control subjects and certain historical control subjects based on adequate statistical matching, such as propensity score (PS). The advantage of using a hybrid control is to expose fewer subjects to allergens for whom there will be no benefit from the vehicle treatment.
Methods :
Among three completed Phase II and III CAC studies for the allergic conjunctivitis indication, two are used as historical control and one is selected as a reference. We compare the outcomes between the historical control and the reference trial control, after controlling PS, to evaluate the similarity in vehicle treatment responses. The PS matching incorporates prognostic factors such as age, sex, and baseline post-challenge itchiness and redness. We construct a hybrid control arm for the reference trial using PS matching between the trial treatment group and the historical control. The results of the treatment effect using the hybrid control are compared with the results using the trial control.
Results :
The reference study was randomized in a 2:1 ratio to treatment (91 subjects) or control (47 subjects). The combined historical control includes 75 subjects from the two studies. Figure 1 shows the post-challenge mean scores of itchiness and redness for the baseline, 16 hours post treatment time points, and 15-minutes post treatment, displayed by individual study and matched groups using PS matching. The mean differences between the two matched groups are reduced to less than or equal to 0.24. Table 1 shows that the conclusions on treatment effect are unchanged using the hybrid control in the reference trial. The least squares mean differences are less than or equal to 0.17 between the trial and hybrid controls.
Conclusions :
Introducing hybrid control arms in CAC studies maintains study integrity and validity. The use of a hybrid control could potentially reduce costs and improve subject safety.
This abstract was presented at the 2022 ARVO Annual Meeting, held in Denver, CO, May 1-4, 2022, and virtually.