Arthritis & Rheumatism, Volume 62,
November 2010 Abstract Supplement
Abstracts of the American College of
Rheumatology/Association of Rheumatology Health Professionals
Annual Scientific Meeting
Atlanta, Georgia November 6-11, 2010.
Meta-Analysis and BiologicsComparing Apples to Apples: An Example Using Radiographic Outcomes in Rheumatoid Arthritis.
Ingham1, Mike, Carter2, Chureen
Many meta-analyses exist for biologics in rheumatoid arthritis (RA). Most have focused on signs and symptoms. Few have attempted to look at radiographic outcomes of disease progression.
This study used meta-analytic techniques to compare adalimumab (ADA), certolizumab (CER), etanercept (ETA) and infliximab (IFX). Randomized controlled trials (RCTs) of biologics in active RA, added to methotrexate, reporting 1 year radiographic outcomes, Erosion scores (ES), Joint Space Narrowing scores (JSNS) and modifications of Total Sharp scores (mTSS) were included. Sources for trials were drawn from recent Cochrane and Agency for Health Care Research and Quality publications. Moderator variables related to outcome were created from trial baseline patient population results, with dichotomization of continuous variables to allow for simplified stratification of trials. Trial data were split into higher or lower dose intensity, based on trial arms. Primary outcomes included differences in mean change scores from baseline of mTSS between active and control groups (MD). Due to varying versions of mTSS used in trials, standardized mean differences (SMD) were also analyzed (standardized by change scores' standard deviation). Negative results favor the biologic.
Four trials met inclusion criteria (one for each product). For higher dose intensity, mTSS MD [95% confidence limits] were ADA: -2.60 [-3.82 to -1.38]; CER: -2.60 [-2.74 to -2.46]; ETA: -3.34 [-3.46 to -3.22]; IFX: -6.80 [-9.26 to -4.34]. SMD were: -0.44; -3.10; -5.24; -0.92 respectively. Simultaneous inclusion of all four trials however was complicated by observed differences in baseline moderator variables, confirmed by an I2 heterogeneity score of 99.5 out of 100. Stratifying the ADA/IFX separately from the ETA/CER trials optimized reduction in the I2 statistic vs. all four trials together. Within these strata, for SMD comparisons, IFX demonstrated twice the effect size of ADA, while ETA was 1.7 times that of CER.
Meta-analytic comparisons that do not consider more in depth RCT inclusion criteria risk ignoring important baseline population differences that might mask clinically meaningful results or lead to inaccurate conclusions. In the absence of access to patient level data, this type of stratification may prove useful for reducing heterogeneity. Confirmation using other outcomes and larger numbers of trials, while keeping treatment constant, is warranted.
Baseline moderator variable dichotomization of RCTs in active RA for radiographic outcomes (DMARD=Disease Modifying Anti-Rheumatic Drug, MTX=Methotrexate, CRP=c-Reactive protein, EU=European union, NA=North America)
|High dose cohorts||ADA||CTZ||ETA||IFX|
|Mean disease duration <10 yrs=0||1||0||0||1|
|Inclusion min tender/swollen joint count<9=0||0||1||1||0|
|Mean age<54 yrs=0||1||0||0||1|
|Rheumatoid Factor Positive <80%=0||1||1||0||1|
|# of prior failed DMARDs <2=0||1||0||1||1|
|MTX dose <15mg/wk=0||1||0||0||1|
To cite this abstract, please use the following information:
Ingham, Mike, Carter, Chureen; Meta-Analysis and BiologicsComparing Apples to Apples: An Example Using Radiographic Outcomes in Rheumatoid Arthritis. [abstract]. Arthritis Rheum 2010;62 Suppl 10 :76