Predicting toxicity: biomarkers and the value of the patient's opinion.

Safety is usually a secondary endpoint in clinical trials, and most observational studies on predictors focus on response, not toxicity. Research is clearly biased towards efficacy rather than safety, probably because of the difficulty in measuring toxicity arising from lack of standardization, low statistical power, missing data, misinterpretation, and confounding variables. An extensive search revealed very few studies on toxicity markers for the drugs used to treat rheumatoid arthritis. The literature contains many studies on predictors of the safety of methotrexate. Most analyze genetic markers, which are difficult to apply in daily practice. We found few studies on markers of toxicity for biologics. As many adverse events are related to patient characteristics, we recommend that clinicians understand the risk factors for the most common adverse events, especially infections and cancer, and manage risk on an individual basis. Moreover, patient involvement in monitoring safety is crucial.