P090 Clinical prediction models for methotrexate treatment outcomes in rheumatoid arthritis patients: a review of existing models and summary of their limitations. (23rd April 2022)

• Record Type: Journal Article
• Authors: Gehringer, Celina; Martin, Glen; Hyrich, Kimme; Verstappen, Suzanne; Sergeant, Jamie
• Is Part Of: Rheumatology. Volume 61(2022)Supplement 1
• Publication Date: 2022-04-23
• Subjects: Rheumatism -- Periodicals; Rheumatology -- Periodicals; 616.723005
• DOI: 10.1093/rheumatology/keac133.089
• Languages: English
• ISSNs: 1462-0324

Abstract

Abstract: Background/Aims: Methotrexate (MTX) is the preferred first line therapy for rheumatoid arthritis (RA), according to NICE guidelines. MTX has several advantages over other treatments including effectiveness and low cost; however, around 40% of patients are classed as non-responders after 6 months. Therefore, there is a clinical need to identify patients at high-risk of poor outcomes, such that patients could potentially be fast tracked onto alternative therapies to improve their clinical outcomes and quality of life. Such risk stratification is possible through prognostic prediction models, although models which have previously been developed appear to have had little impact on practice. This may be in part due to methodological features of their development and validation but, to date, no review has collated the evidence in this field. This review therefore aimed to (i) identify and summarise multivariable prediction models of MTX treatment outcomes in biologic-naïve adult RA patients, and (ii) appraise their methodological properties. Methods: A systematic search was carried out using Medline Ovid to identify studies developing or validating prediction models of MTX outcomes in the population of interest, including demographic, disease-specific or treatment-related covariates, published between 2005 and 2020. Models were stratified by outcome definition, and information on predictors, predictor associations with the outcome, model performance, handling of missing data and model validation were extracted. Results: Twenty-two studies (14 (64%) using data from observational studies and 8 (36%) from randomised controlled trials) were identified. Of these, 15 (68%) based their outcome on a state of disease activity, such as low disease activity or remission, 4 (18%) used the EULAR response criteria, and 3 (14%) predicted discontinuation due to adverse events (AEs). AEs were also incorporated into the composite outcome with disease activity in 3 (14%) studies, 1 (5%) investigated both outcomes in separate models, and only 1 (5%) accounted for potential competing risks to their primary outcome. Internal validation using cross sampling techniques, which is critical for reducing overfitting, was completed in only 5 (23%) studies. Only 4 (18%) studies carried out external validation in new data. Missing data was appropriately handled using multiple imputation in 5 (23%) studies, whilst others used single imputation (n = 1, 4%) or complete case analysis (n = 13, 59%), resulting in potentially biased risk estimates, or did not report how they handled missing data (n = 3, 14%). Conclusion: This review summarises current prediction models of MTX treatment outcomes in RA. It highlights several methodological shortcomings that should be addressed in future model development and validation to improve accuracy of predictions. Without tackling these issues, prediction of MTX treatment outcomes will remain at high risk of bias and should not be recommended for informing risk stratification for RA treatment decisions. Disclosure: C. Gehringer: None. G. Martin: None. K. Hyrich: None. S. Verstappen: None. J. Sergeant: None.