BackgroundThe upper extremity plays an important role in daily functioning of patients with Multiple Sclerosis (MS) and strongly influences their quality of life. However, an explicit overview of arm-hand training programs is lacking. The present review aims to investigate the training components and the outcome of motor training programs for arm and hand in MS.MethodsA computerized systematic literature search in 5 databases (PubMed, CINAHL, EMBASE, PEDro and Cochrane) was performed using the following Mesh terms: Multiple Sclerosis, Rehabilitation, Physical Education and Training, Exercise, Patient-Centered Care, Upper Extremity, Activities of Daily Living, Motor Skills, Motor Activity, Intervention Studies and Clinical Trial. The methodological quality of the selected articles was scored with the Van Tulder Checklist. A descriptive analyses was performed using the PICO principle, including scoring of training components with the calculation of Hedges’g effect sizes.ResultsEleven studies were eligible (mean Van Tulder-score = 10.82(SD2.96)). Most studies reported a specific improvement in arm hand performance at the ICF level that was trained at. The mean number of training components was 5.5(SD2.8) and a significant correlation (r = 0.67; p < 0.05) between the number of training components and effect sizes was found. The components ‘client-centered’ and ‘functional movement’ were most frequently used, whereas ‘distribution based practice’, ‘feedback’ and ‘random practice’ were never used. The component ‘exercise progression’ was only used in studies with single ICF body function training, with the exception of 1 study with activity level training. Studies including the component ‘client-centred’ demonstrated moderate to high effect sizes.ConclusionMotor training programs (both at the ICF body function and activity level) have shown to improve arm and hand performance in MS in which the value of the training specificity was emphasized. To optimize upper extremity training in MS the component ‘client-centred’ and ‘exercise progression’ may be important. Furthermore, given the importance attributed to the components ‘distribution based practice’, ‘feedback’ and ‘random practice’ in previous research in stroke patients, the use of these components in arm hand training should be explored in future research.