Background This study tested the null hypothesis that type of injection (corticosteroid vs. placebo) is not a predictor of arm-specific disability as measured with the Disabilities of Arm, Shoulder and Hand questionnaire 1 to 3 months after injection of dexamethasone or placebo for treatment of trapeziometacarpal (TMC) arthrosis or de Quervain syndrome. Secondly, we tested if type of injection was a predictor of pain intensity. Methods Thirty-six English-speaking adults with TMC arthrosis or de Quervain syndrome were randomized for a dexamethasone or a placebo injection. At time of the injection, patients completed a demographic data sheet and validated questionnaires assessing arm-specific disability, pain intensity, depressive symptoms, pain catastrophizing, and patient's health-related beliefs. At an average of 1.4±0.42 months (range, 0.79–2.5 months) after the injection, patients completed questionnaires regarding arm-specific disability, pain, and treatment satisfaction. Grip and pinch strength measurements were measured at both time points. Bivariable and multivariable analyses assessed predictors of arm-specific disability and pain intensity at follow-up. Results Type of injection was not a predictor of arm-specific disability or pain intensity 1 to 3 months after injection. The best model both for arm-specific disability and pain intensity at follow-up included pain catastrophizing and explained 18 % and 33 % of the variability, respectively. Conclusions Catastrophic thinking was a better predictor of both of arm-specific disability and pain intensity than diagnosis or type of injection (steroid vs. placebo) 1 to 3 months after an injection.