Rheumatoid arthritis (RA) is a chronic, inflammatory, autoimmune disease that primarily affects the joints and is associated with autoantibodies that target various molecules including modified self-epitopes. The identification of novel autoantibodies has improved diagnostic accuracy, and newly developed classification criteria facilitate the recognition and study of the disease early in its course. New clinical assessment tools are able to better characterize disease activity states, which are correlated with progression of damage and disability, and permit improved follow-up. In addition, better understanding of the pathogenesis of RA through recognition of key cells and cytokines has led to the development of targeted disease-modifying antirheumatic drugs. Altogether, the improved understanding of the pathogenetic processes involved, rational use of established drugs and development of new drugs and reliable assessment tools have drastically altered the lives of individuals with RA over the past 2 decades. Current strategies strive for early referral, early diagnosis and early start of effective therapy aimed at remission or, at the least, low disease activity, with rapid adaptation of treatment if this target is not reached. This treat-to-target approach prevents progression of joint damage and optimizes physical functioning, work and social participation. In this Primer, we discuss the epidemiology, pathophysiology, diagnosis and management of RA.

Crohn Disease (CD) is a complex genetic disorder for which more than 140 genes have been identified using genome wide association studies (GWAS). However, the genetic architecture of the trait remains largely unknown. The recent development of machine learning (ML) approaches incited us to apply them to classify healthy and diseased people according to their genomic information. The Immunochip dataset containing 18,227 CD patients and 34,050 healthy controls enrolled and genotyped by the international Inflammatory Bowel Disease genetic consortium (IIBDGC) has been re-analyzed using a set of ML methods: penalized logistic regression (LR), gradient boosted trees (GBT) and artificial neural networks (NN). The main score used to compare the methods was the Area Under the ROC Curve (AUC) statistics. The impact of quality control (QC), imputing and coding methods on LR results showed that QC methods and imputation of missing genotypes may artificially increase the scores. At the opposite, neither the patient/control ratio nor marker preselection or coding strategies significantly affected the results. LR methods, including Lasso, Ridge and ElasticNet provided similar results with a maximum AUC of 0.80. GBT methods like XGBoost, LightGBM and CatBoost, together with dense NN with one or more hidden layers, provided similar AUC values, suggesting limited epistatic effects in the genetic architecture of the trait. ML methods detected near all the genetic variants previously identified by GWAS among the best predictors plus additional predictors with lower effects. The robustness and complementarity of the different methods are also studied. Compared to LR, non-linear models such as GBT or NN may provide robust complementary approaches to identify and classify genetic markers.

BackgroundMethotrexate (MTX) remains the disease-modifying anti-rheumatic drug of first choice in rheumatoid arthritis (RA) but response varies. Predicting non-response to MTX could enable earlier access to alternative or additional medications and control of disease progression. We aimed to identify baseline predictors of non-response to MTX and combine these into a prediction algorithm.MethodsThis study included patients recruited to the Rheumatoid Arthritis Medication Study (RAMS), a UK multi-centre prospective observational study of patients with RA or undifferentiated polyarthritis, commencing MTX for the first time. Non-response to MTX at 6 months was defined as “no response” using the European League Against Rheumatism (EULAR) response criteria, discontinuation of MTX due to inefficacy or starting biologic therapy. The association of baseline demographic, clinical and psychosocial predictors with non-response was assessed using logistic regression. Predictive performance was assessed using the area under the receiver operating characteristic curve (AUC) and calibration plots.ResultsOf 1050 patients, 449 (43%) were classified as non-responders. Independent multivariable predictors of MTX non-response (OR (95% CI)) were rheumatoid factor (RF) negativity (0.62 (0.45, 0.86) for RF positivity versus negativity), higher Health Assessment Questionnaire score (1.64 (1.25, 2.15)), higher tender joint count (1.06 (1.02, 1.10)), lower Disease Activity score in 28 joints (0.29 (0.23, 0.39)) and higher Hospital Anxiety and Depression Scale anxiety score (1.07 (1.03, 1.12)). The optimism-corrected AUC was 0.74.ConclusionsThis is the first model for MTX non-response to be developed in a large contemporary study of patients commencing MTX in which demographic, clinical and psychosocial predictors were considered. Patient anxiety was a predictor of non-response and could be addressed at treatment commencement.Electronic supplementary materialThe online version of this article (10.1186/s13075-018-1645-5) contains supplementary material, which is available to authorized users.