Lead Investigator: Jose Antonio Pereira da Silva, Universidade de Coimbra
Title of Proposed Research: Long-term predictive value of patient global assessment regarding radiographic damage and physical function in patients with Rheumatoid Arthritis individual patient data meta-analysis
Vivli Data Request: 5327
Funding Source: None
Potential Conflicts of Interest: JAPS reports a research grant from Pfizer and AbbVie, and speaker fees from Pfizer, AbbVie, Roche, Lilly, Novartis. These potential conflicts will be disclosed in all papers resulting from this project.
Summary of the Proposed Research:
The treatment of Rheumatoid Arthritis (RA) has improved remarkably over recent years, due to not only the development of new therapies but also novel treatment strategies. Among these, the Treat-to-Target (T2T) recommendation epitomizes the consensual concept that disease treatment should aim at achieving, as early and consistently as possible, a target of level of remission, or at least low disease activity. To assess if the target is achieved or not physicians use composite indices that include different variables. The most common variables used in these indices (or remission definitions) are: number of tender (TJC28) and swollen joint counts (SJC28), a inflammatory parameter (blood analysis) and a patient reported outcome (PRO) designed by “Patient Global Assessment” (PGA). In this PGA the patients is questioned “Considering all the ways your arthritis has affected you, how do you feel your arthritis is today?” and asked to rate it from 0 to 100.
This study is designed to clarify whether PGA should or not be used to guide immunosuppressive therapy and evaluate its results in RA.
To this purpose we will test the relationship between two different definitions of disease remission and long-term outcomes in terms of structural damage (X-Rays) and function. The states of remission will be categorized as follows:
-“4v-Remission”: the American College of Rheumatology (ACR)/European League Against Rheumatism (EULAR) Boolean-based definition: TJC28<=1, SJC28<=1, C-reactive protein (CRP) in mg/dl<=1, and PGA<=1/10.
-“3v-Remission”: as above, excluding PGA.
Analyses will also specifically address the outcomes of patients that achieve 3v but not 4v-Remission (i.e. distinguished by only PGA). To achieve this goal on the most robust way we are trying to pool together individual level patient data from as many randomized clinical trials (RCT) as possible, including studies from different pharmaceutical companies.
A number of secondary outcomes will also be addressed including whether the relationship between remission states (i.e. 3v and 4v-remission) and long-term outcome is influenced by treatment arm and other factors (e.g. age, gender, co-medication, disease duration, other), over time.
Expected result include the demonstration that the value and efficacy of available biologic therapies is actually higher than previously acknowledged and also that PGA should not be included in the definition of the target for immunosuppressive therapy. This does not mean that the patient perspective should be disregarded. By the contrary, we support that, once disease control is achieved, adjunctive therapy of a different nature (pharmacological and/or nonpharmacological) should be guided by patients’ perspective, but maybe using more informative tools than PGA.
Statistical Analysis Plan:
A.6.1. Sample size calculations
To determine the minimal sample size required to this study we used the formula proposed by Hajian-Tilaki to compare two independent proportions regarding sensitivity and/or specificity of two test of unpaired design.
Assuming 95% confidence and 80% power to detect a difference of 7% from a specificity of 98% (i.e. P1=98% and P2=91%) based on the results of a previous study, it would be required a n=165 for each group. This means at least 165 patients in 4v-remission and 165 patients in 3v-remission. If we assume a more strict difference of 5% and specificities of P1=95% and P2=90%, the sample required will be n=433 for each group. Considering 15% as the average remission rate in clinical trials assessed by ACR/EULAR Boolean-based definition, the total number of patients required for pooled analysis is n=2887.
A.6.2. Data analysis
To guarantee privacy and security of IPD, the platforms in which the data is available require that statistics be performed via remote and secure online platforms, which impedes data download. However, all platforms commonly allow the use of R or SAS software, which we will use within each platform. For data synthesis we will use Stata software, version 14. Thus the same procedures will need to be performed in each platform. All significance tests will be two tailed and conducted at the 0.05 significance level.
Means and standard deviations (SD) will be used to describe normally distributed continuous data, medians and interquartile ranges will be used to describe continuous data that are not normally distributed, and frequencies and percentages will be used for categorical data. Data will be described using 95% confidence intervals (95% CI).
The number of true positive (TP), true negative (TN), false negative (FN) and false positive (FP) statistics will be determined for each platform. Then, sensitivity, specificity, positive and negative predictive value (PPV and NPV) will be determined for the ability of sustained remission (by 3v and 4v-remission) to predict good radiographic outcome at 2-years after baseline. The LR+ and LR- will then be calculated by the formula: sensitivity/(1-specificity). All analyses will be repeated for secondary outcomes: good function, and combined good radiographic and good function. To rank LR+ and LR- for 3v and for 4v-remission groups chi- square analysis using logistic regression will be used, in which the independent variable will be remission (3v and 4v-remission) and the dependent variable will be the outcome of interest (e.g. radiographic stability).
Missing data will not be submitted to any method of data imputation. The number and percentage of patients with missing values for each variable will be reported per trial.
A.6.3. Measures to adjust for confounders
In order to take into the time variable consideration (i.e. the evolution within individual between visits), and to adjust for important covariates (gender, age at baseline, disease duration at baseline, RF, anti-ACPA, radiographic damage at baseline, treatment arm), binomial Generalised Estimating Equations (bGEE) will be used, considering all visit time points common to the included trials. This will provide an Odd Ratio (OR) for 3v and 4v-remission groups.
A.6.4. Sensitivity analyses
In recent years a statistically significant reduction in radiographic progression during clinical studies in patients with RA has become difficult to detect due to early-escape study designs and to declining rates of progression in control-group patients has lead to difficulties in assessing differences in this outcome. Because of that we defined a strict definition of good outcome, i.e, a change =0, and we will consider a combined good radiographic and function stability. Although it has been shown that TNF inhibitor monotherapy had similar results to MTX monotherapy in terms of radiographic damage we will perform a sensitivity analysis considering the different treatment arms (i.e. bDMARD monotherapy, cDMARD monotheraphy and combined therapy – cbDMARD).
It is also argued that the number of years of follow-up could affect the results of radiographic outcomes. Thus, in addition to the main analysis of the 2-years outcome (the most frequently reported) we will assess outcomes also for 5 and 10-years after baseline, i.e, 4 and 9 years’ stability after 12 months.
Other factors that might influence radiographic progression are: patients with early versus established disease; patients naive to MTX or patients that failed MTX before bDMARD initiation. If the number of studies (and patients) allow, sensitivity analyses will be performed for these groups independently.
A.6.5. Data synthesis
The TP, TN, FN, and FP statistics obtained for each trial in the previous step will be used to synthetize the data.
To test heterogeneity among the studies, the I2 of Higgins and Thompson will be calculated. An I2 value close to 0% indicates no heterogeneity between studies, close to 25% indicates low heterogeneity, close to 50% indicates moderate heterogeneity, and close to 75% indicates high heterogeneity.
A.6.6. Exploratory analyses
The definition of “sustained” remission (and non-remission) based on only 2 time points (at 6 and 12 months) may not fully capture all relevant information. Thus it will be described both the duration of remission (3v and 4v) as well as the interruptions in this desired state. It will be also explored whether the multiple remission and relapse periods are related to the long- term radiographic progression and compare the performance of 3v and 4v-remission definitions. For this purpose, the “Continuity Reward” (ConRew) score and patient vector graphs proposed by Boers et al will be used.
A Prospective Multi-Centre Randomised, Double-Blind, Active Comparator-Controlled, Parallel-Groups Study Comparing the Fully Human Monoclonal Anti-TNFα Antibody Adalimumab Given Every Second Week With Methotrexate Given Weekly and the Combination of Adalimumab and Methotrexate Administered Over 2 Years in Patients With Early Rheumatoid Arthritis (PREMIER).
Study ID: NCT00195663
Sponsor ID: DE013
A Multicenter, Randomized, Double-Blind, Placebo-Controlled Study of the Human Anti-TNF Monoclonal Antibody Adalimumab in Rheumatoid Arthritis Patients Currently Receiving Treatment With Methotrexate
Study ID: NCT00195702
Sponsor ID: DE019
A Phase III Multi-center, Double-blind, Placebo-controlled, Parallel Group 24-Week Study to Assess the Efficacy and Safety of Two Dose Regimens of Liquid Certolizumab Pegol as Additional Medication to Methotrexate in the Treatment of Signs and Symptoms of Rheumatoid Arthritis and in Prevention of Joint Damage in Patients With Active Rheumatoid Arthritis Who Have an Incomplete Response to Methotrexate.
Study ID: NCT00160602
Sponsor ID: C87050
A Phase III Multicentre, Double Blind, Placebo-controlled, Parallel Group 52-week Study to Assess the Efficacy and Safety of 2 Dose Regimens of Lyophilised CDP870 as Additional Medication to Methotrexate in the Treatment of Signs and Symptoms and Preventing Structural Damage in Patients With Active Rheumatoid Arthritis Who Have an Incomplete Response to Methotrexate
Study ID: NCT00152386
Sponsor ID: C87027
A Phase III Multi-centre, Open-label, follow-on Study to CDP870-027, to Assess the Efficacy and Safety of Lyophilized CDP870 an Engineered Human Anti-TNF PEG Conjugate, as Additional Medication to Methotrexate, in the Treatment of Signs and Symptoms and Preventing Structural Damage in Patients With Active Rheumatoid Arthritis
Study ID: NCT00175877
Sponsor ID: C87028
A Phase III Multi-center, Open-label, Follow-up Study, to Assess the Safety and Efficacy of Liquid Certolizumab Pegol as Additional Medication to Methotrexate, in the Treatment of Signs and Symptoms and in the Prevention of Joint Damage in Patients With Active Rheumatoid Arthritis Who Participated in Study CDP870-050
Study ID: NCT00160641
Sponsor ID: C87051
A Double-Blind Study Evaluating the Efficacy and Safety of the Combination of Etanercept and Methotrexate in Comparison to Etanercept Alone or Methotrexate Alone in Rheumatoid Arthritis Patients.
Data Contributor: Pfizer Inc.
Study ID: NCT00393471
A 24-Month,Randomized,Double-Blind,Two-Period Study to Evaluate the Efficacy and Safety of the Combination of Etanercept and Methotrexate and Methotrexate Alone in Subjects With Early Rheumatoid Arthritis
Data Contributor: Pfizer Inc.
Study ID: NCT00195494
Ferreira RJO, Welsing PMJ, Jacobs JWG, et al. Revisiting the use of remission criteria for rheumatoid arthritis by excluding patient global assessment: an individual meta-analysis of 5792 patients. Annals of the Rheumatic Diseases Published Online First: 06 October 2020.