Lead Investigator: Neeraj Narula, Hamilton Health Sciences
Title of Proposal Research: Vedolizumab vs. Adalimumab in Ulcerative Colitis: Validation of Indirect Comparisons
Vivli Data Request: 7542
Funding Source: None
Potential Conflicts of Interest: Neeraj Narula holds a McMaster University Department of Medicine Internal Career Award. Neeraj Narula has received honoraria from Janssen, Abbvie, Takeda, Pfizer, Merck, Lupin and Ferring. None of the listed conflicts of interest are aware of or are involved in this research
Summary of the Proposed Research:
Ulcerative colitis (UC) affects over 120,000 Canadians and is an inflammatory bowel disease that affects the large intestine and is characterized by diarrhea, rectal bleeding, abdominal pain, urgency, and tenesmus. Patients with severe UC and/or disease that cannot be adequately controlled with corticosteroids are candidates for biologic monoclonal antibody treatments such as vedolizumab and adalimumab.
Vedolizumab and adalimumab are two biologic therapies that have demonstrated efficacy in achieving and maintaining remission through several pivotal placebo-controlled trials, including ULTRA 1 (NCT00385736), ULTRA 2 (NCT00408629), and GEMINI 1 (NCT00783718). Recently, the VARSITY trial (NCT02497469) demonstrated the superiority of vedolizumab to adalimumab among patients with moderate to severe UC. While studies comparing provide robust evidence to inform positioning of therapies, trials are costly and time-consuming. Therefore, there is increasing interest in performing indirect comparisons of biologics using patient-level data from different placebo-controlled clinical trials. However, the reliability of these comparisons relative to trials comparing active therapies remains unclear.
The Mayo Score is a tool that is used to determine UC disease activity and is comprised of four patient-reported and endoscopic assessments: stool frequency, rectal bleeding, endoscopic findings, and physician’s global assessment. Each parameter is scored from 0 to 3, with higher scores indicating greater disease severity. The total Mayo Score ranges from 0 to 12, while the partial Mayo Score excludes the endoscopic subscore and thus ranges from 0 to 9. While the Mayo Score has been widely used in clinical trials for UC as primary endpoints, mucosal healing (determined by endoscopy) remains an important target of treatment as well.
The primary objective of this study is to compare the efficacy of vedolizumab and adalimumab to achieve post-induction (week 6/8) mucosal healing as determined by the Mayo Score. The secondary objectives of this study include clinical remission at one-year and mucosal healing at one-year. Additionally, clinical response, reduction in markers of disease activity (e.g. fecal calprotectin, C-reactive protein), and reduction in histologic measures of disease activity at post-induction (week 6/8) and one-year will be analyzed.
Data from GEMINI 1 and VARSITY, which is being requested from Vivli, will be pooled to obtain a cohort of biologic-naïve UC patients treated with adalimumab. Biologic-naïve patients treated with adalimumab in ULTRA 1 and ULTRA 2, which is also being requested from Vivli, will be pooled to comprise the adalimumab cohort.
Statistical Analysis Plan:
In GEMINI 1, 746 patients were randomized to vedolizumab, of which 388 were biologic-naïve. Although GEMINI 1 re-randomized patients at the end of induction therapy (week 6) based on response status, those who did not respond at week 6 were randomized to continue receiving vedolizumab 300mg every 4 weeks during maintenance, thus mitigating bias favouring responders in this analysis. All patients who received adalimumab in ULTRA 1 (n=248) were biologic-naïve. Therefore, the eligible study population for the aggregated comparison includes 248 patients from ULTRA 1 and ULTRA 2 and 388 from GEMINI 1. In VARSITY, 385 patients were randomized to vedolizumab, of which 305 were biologic-naïve, and of the 386 patients randomized to adalimumab and 307 were biologic-naïve. Only biologic-naïve patients from VARSITY will be compared. Data from these trials are being requested as there were common outcome timepoints. Mayo Scores were captured at two common time points of interest (week 6/8 and 52) in GEMINI 1, ULTRA 1, ULTRA 2, and VARSITY. An indicator variable will be used to identify patients from the trials analyzed.
The primary analysis will be conducted as intention-to-treat, where patients with missing data will be assumed to not have achieved the outcomes of interest. A separate case analysis will be conducted where patients with missing outcome data (e.g. one year Mayo Score) will be excluded from the primary analysis.
Descriptive statistics will be calculated to compare the rate of achievement of outcomes between groups. In addition, logistic regression will used to assess the treatment effect on the outcome of interest. Univariate analyses will be conducted to identify associations between covariates and the outcome of interest, and any variables with a p-value < 0.10 will be included in the multivariate model.
In addition, propensity score matching on a one-to-one basis using k-nearest neighbour without replacement will be performed if significant baseline differences between groups are observed. The propensity score represents the conditional probability of receiving vedolizumab or infliximab, given the observed co-variates. The propensity scores will be estimated using a nonparsimonious logistic regression model that will consider baseline covariates. The treatment group will be regressed on these baseline covariates. Subsequently, graphs of the propensity scores for participants treated in both studies will be compared to assess the region of common support. This assesses the propensity score distributions between both groups and identifies the region of common support (overlap).
Continuous variables will be presented as means (and standard deviations [SD] or as medians and interquartile ranges [IQR]), if the data is skewed. Binary variables will be presented as proportions or percentages. Descriptive statistics will be used to summarize baseline demographics, disease characteristics and outcome parameters of included patients. Differences between groups will be compared using the Mann-Whitney U test or chi-squared test. Data will be analyzed using Stata, which is available on the Vivli secure platform.
A Multicenter, Randomized, Double-blind Placebo-controlled Study of the Human Anti-TNF Monoclonal Antibody Adalimumab for the Induction of Clinical Remission in Subjects With Moderately to Severely Active Ulcerative Colitis
Data Contributor: AbbVie
Study ID: NCT00385736
Sponsor ID: M06-826
A Multicenter, Randomized, Double-blind, Placebo-controlled Study of the Human Anti-TNF Monoclonal Antibody Adalimumab for the Induction and Maintenance of Clinical Remission in Subjects With Moderately to Severely Active Ulcerative Colitis
Data Contributor: AbbVie
Study ID: NCT00408629
Sponsor ID: M06-827
A Phase 3, Randomized, Placebo-Controlled, Blinded, Multicenter Study of the Induction and Maintenance of Clinical Response and Remission by MLN0002 in Patients With Moderate to Severe Ulcerative Colitis
Data Contributor: Takeda
Study ID: NCT00783718
Sponsor ID: C13006
A Randomized, Double-Blind, Double-Dummy, Multicenter, Active-Controlled Study to Evaluate the Efficacy and Safety of Vedolizumab IV Compared to Adalimumab SC in Subjects With Ulcerative Colitis
Data Contributor: Takeda
Study ID: NCT02497469
Sponsor ID: MLN0002-3026
- Narula, N., Wong, E.C.L., Dulai, P.S. et al. Vedolizumab and Adalimumab in Biologic-Naïve Ulcerative Colitis: Comparison of Patient-Level Clinical Trial Data and VARSITY for Week 6 Clinical Remission. Dig Dis Sci (2023). doi: 10.1007/s10620-023-07825-4
- Wong, E.C., Dulai, P.S., Marshall, J.K., Jairath, V., Reinisch, W. and Narula, N., 2023. Matching-adjusted Indirect Comparisons vs Propensity Score Matching with Individual Patient-level Data to Estimate Treatment Efficacy. Inflammatory Bowel Diseases, p.izad077. doi: 10.1093/ibd/izad077