outside the predefined visit window was used for the ITT analysis
but not for the PP analysis.
The study enrolled 2 groups of patients treated with oral study drug or oral and topical study drug, respectively. The randomization
process assigned equal numbers of patients to each treatment group.
There were 2 centers in the study. Blinded study medication was identified using the patient randomization number.
Secondary and Ancillary Efficacy Variables and Their Statistical Analyses
Lesion counts: For analysis purposes, we assume that separate papule, pustule, and nodule counts are normally distributed. Papule, pustule, and nodule counts were summarized
for each treatment group and at each visit using summary statistics. Change from baseline for papule, pustule,
nodule and total inflammatory lesions were analyzed at each postrandomization visit using the same analysis of variance (ANOVA) model as the primary efficacy variable.
IGA score and CEA score: IGA and CEA ranged from 0 to 5. The primary analysis of IGA and CEA was based on changes from baseline scores at end point and was summarized for each treatment group and at each visit. For each visit, these changes from baseline scores were analyzed using Cochran–Mantel–Haenszel (CMH) tests stratified by center.
Responders: The analysis of the percentage of patients in the ITT population achieving at least a 50% reduction in inflammatory lesions at the week 12 visit (end point) was conducted using ANOVA test stratified by center.
Demographics and Baseline Characteristics
Demographic variables were age, gender, and race. Baseline characteristics included a complete medical history, lesion score (total number of facial lesions and number of papules, pustules,
nodules) and IGA/CEA scores. IGA and CEA scores were summarized using frequency distributions and were tested for baseline comparability of treatment groups using ANOVA. Total number of facial lesions, total number of inflammatory lesions, and numbers of papules, pustules, and nodules were summarized
for each treatment group using summary statistics (mean, standard deviation, median, minimum, and maximum).
Adverse Clinical Experiences
AEs will be coded using the MedDRA Dictionary (Medical Dictionary for Regulatory Activities, version 4.1). All AEs and study drug–related AEs will be tabulated by seriousness, death, and discontinuation because of adverse experiences for each treatment group. Study drug–related AEs were defined as AEs considered likely or definitely related to study drug or with missing relation.
Any AEs, including both observed or volunteered problems, complaints, or symptoms, were to be recorded. AEs resulting from concurrent illnesses or reactions to concurrent medications
are also to be recorded. Each AE is to be evaluated for duration, intensity, and relationship with the study medication or other causes. The intensity of the event was characterized as mild, moderate, or severe as follows:
Mild: Events are usually transient, requiring no special treatment, and do not interfere with the subject’s daily activities.
Moderate: Events traditionally introduce a low level of inconvenience
or concern to the subject and may interfere with daily activities, but are usually ameliorated by simple therapeutic measures.
Severe: Events interrupt subject’s usual daily activity and traditionally
require systemic drug therapy or other treatment.
One of the following determinations will then be used to document
the relationship of the AE to the study drug: not related, possible, probable.
Discontinuation and Replacement of Patients
Any patient found to have entered the study in violation of this protocol was withdrawn from the study. Any female subject who became pregnant during the study was withdrawn from the study. Any subject who required the use of an unacceptable
concomitant medication was withdrawn from the study. If a patient was withdrawn from the study, regardless of the cause, all evaluations required at the scheduled end of the study were performed. Patients discontinued from the study were not replaced.
All study variables—vital signs data, outcome variables (total lesion count, IGA, CEA—were summarized at each study visit with means and standard deviations, medians, and extrema. The primary outcome variable was the 12-week change in the total lesion count (from baseline to week 12). Additional outcomes
include 12-week changes in CEA score and IGA score, as well as changes in IGA, CEA, and lesion counts from baseline
to interim weeks (4 and 8) and the follow-up visit (week 16). Comparisons of changes in outcome variables within each treatment group (eg, tests of the efficacy of each treatment) were Wilcoxon signed-rank tests. Comparisons of changes in outcome variables between treatment groups were Wilcoxon rank-sum tests, stratified by study site. Two categorical outcomes
were examined—subjects having a week 12 IGA less