Quantification of upper limb dysfunction in the activities of the daily living in persons with multiple sclerosis.
Pisa M., Ruiz JA., DeLuca GC., de Andres Crespo M., DelMastro HM., Olson KM., Triche EW., Lo AC.
BACKGROUND: Dysfunction in upper limb (UL) function has been reported as an important indicator for disease progression in persons with multiple sclerosis (PwMS), thus a relevant outcome in clinical trials. However, standard assessment of UL function is limited to Nine-Hole Peg Test (NHPT) which assesses fine dexterity. This study aimed to deeply endophenotype UL involvement in PwMS and identify the most accurate set of measures needed to capture the complexity of UL dysfunction in the activities of daily living (ADL). METHODS: 257 PwMS underwent an extensive UL assessment using standardized measures of grip strength and endurance, coordination, vibratory and tactile sensation, dexterity, capacity and functionality. Limitation in ADL was defined from an objective perspective using a timed test (Test d'Evaluation de la performance des Membres Supérieurs des Personnes Âgées: TEMPA) and from a subjective perspective using a questionnaire (Disabilities of the Arm, Shoulder and Hand: DASH). Disease severity subgroups were compared utilizing the Kruskal-Wallis test and frequencies determined the prevalence of abnormal UL for each measure. The Jonckheere-Terpstra test compared tested variables with disease severity. Then Receiver operating characteristic (ROC) curve analysis was used to test the accuracy of each tested variable in defining abnormality in the TEMPA and DASH. Cut-off scores were calculated using the Youden index. The predictive value of various tests over TEMPA and DASH were tested using a linear regression analysis. RESULTS: UL dysfunction was highly prevalent in all the modalities tested, even in participants with no/mild disability. Box and Block Test (BBT), finger-nose test (FNT), and NHPT were independently selected with ROC analyses as the most accurate measures in detecting abnormalities in TEMPA and DASH. In multivariate regression models, BBT and FNT, and NHPT all contributed to predicting TEMPA (adj. R2 0.795, P