TY - JOUR
T1 - Placebo Rates in Crohn's Disease Randomized Clinical Trials
T2 - An Individual Patient Data Meta-Analysis
AU - Solitano, Virginia
AU - Hogan, Malcolm
AU - Singh, Siddharth
AU - Danese, Silvio
AU - Peyrin-Biroulet, Laurent
AU - Zou, Guangyong
AU - Yuan, Yuhong
AU - Sands, Bruce E.
AU - Feagan, Brian G.
AU - Dulai, Parambir S.
AU - Narula, Neeraj
AU - Ma, Christopher
AU - Jairath, Vipul
N1 - Publisher Copyright:
© 2025 The Authors
PY - 2024
Y1 - 2024
N2 - Background & Aims: Understanding placebo rates is critical for efficient clinical trial design. We assessed placebo rates and associated factors using individual patient data from Crohn's disease trials. Methods: We conducted a meta-analysis of phase 2/3 placebo-controlled trials evaluating advanced therapies in moderate to severe Crohn's disease (2010–2021). Deidentified individual patient data were obtained through Vivli Inc and the Yale University Open Data Access Project. Primary outcomes were clinical response and remission. Pooled placebo rates and 95% confidence intervals (CIs) were estimated using 1- and 2-stage meta-analytic approaches. Regression analyses identified patient-level factors associated with placebo rates. Results: Analysis of individual patient data from 8 induction (n = 1147) and 4 maintenance (n = 524) trials showed overall placebo clinical response and remission rates for induction were 27% (95% CI, 23%–32%) and 10% (95% CI, 8%–14%), respectively, and 32% (95% CI, 23%–42%) and 22% (95% CI, 14%–33%) for maintenance, respectively. Among biologic (bio)-naïve patients, placebo response and remission rates during induction were 29% (95% CI, 24%–35%) and 11% (95% CI, 8%–15%) respectively, and 26% (95% CI, 20%–33%) and 10% (95% CI, 8%–14%) for biologic (bio)-exposed patients, respectively. During maintenance, biologic-naïve response and remission rates were 41% (95% CI, 34%–48%) and 32% (95% CI, 24%–40%), respectively, and 29% (95% CI, 24%–34%) and 16% (95% CI, 13%–21%) for bio-exposed, respectively. Higher baseline C-reactive protein concentration predicted lower placebo rates, whereas higher baseline albumin levels and body mass index increased the odds of placebo outcomes. Increased baseline Crohn's Disease Activity Index and 2-item patient-reported outcome scores predicted higher response rates in induction, lower response rates in maintenance, and lower remission rates in induction and maintenance. Conclusions: Patient- and trial-level characteristics influence placebo rates in Crohn's disease trials. Careful implementation of eligibility criteria, outcome definitions, and patient stratification may reduce placebo rates.
AB - Background & Aims: Understanding placebo rates is critical for efficient clinical trial design. We assessed placebo rates and associated factors using individual patient data from Crohn's disease trials. Methods: We conducted a meta-analysis of phase 2/3 placebo-controlled trials evaluating advanced therapies in moderate to severe Crohn's disease (2010–2021). Deidentified individual patient data were obtained through Vivli Inc and the Yale University Open Data Access Project. Primary outcomes were clinical response and remission. Pooled placebo rates and 95% confidence intervals (CIs) were estimated using 1- and 2-stage meta-analytic approaches. Regression analyses identified patient-level factors associated with placebo rates. Results: Analysis of individual patient data from 8 induction (n = 1147) and 4 maintenance (n = 524) trials showed overall placebo clinical response and remission rates for induction were 27% (95% CI, 23%–32%) and 10% (95% CI, 8%–14%), respectively, and 32% (95% CI, 23%–42%) and 22% (95% CI, 14%–33%) for maintenance, respectively. Among biologic (bio)-naïve patients, placebo response and remission rates during induction were 29% (95% CI, 24%–35%) and 11% (95% CI, 8%–15%) respectively, and 26% (95% CI, 20%–33%) and 10% (95% CI, 8%–14%) for biologic (bio)-exposed patients, respectively. During maintenance, biologic-naïve response and remission rates were 41% (95% CI, 34%–48%) and 32% (95% CI, 24%–40%), respectively, and 29% (95% CI, 24%–34%) and 16% (95% CI, 13%–21%) for bio-exposed, respectively. Higher baseline C-reactive protein concentration predicted lower placebo rates, whereas higher baseline albumin levels and body mass index increased the odds of placebo outcomes. Increased baseline Crohn's Disease Activity Index and 2-item patient-reported outcome scores predicted higher response rates in induction, lower response rates in maintenance, and lower remission rates in induction and maintenance. Conclusions: Patient- and trial-level characteristics influence placebo rates in Crohn's disease trials. Careful implementation of eligibility criteria, outcome definitions, and patient stratification may reduce placebo rates.
KW - Crohn's Disease
KW - Meta-analysis
KW - Placebo Response
KW - Trial Design
UR - http://www.scopus.com/inward/record.url?scp=85213259374&partnerID=8YFLogxK
U2 - 10.1053/j.gastro.2024.10.009
DO - 10.1053/j.gastro.2024.10.009
M3 - Article
C2 - 39414161
AN - SCOPUS:85213259374
SN - 0016-5085
VL - 168
SP - 344
EP - 356
JO - Gastroenterology
JF - Gastroenterology
IS - 2
ER -