We studied multiple sclerosis fatigue (MSF) and its relationship to depression and disability. Seventy-one patients [50 relapsing-remitting, 21 secondary progressive] were grouped by Fatigue Severity Scale (FSS) into MS-fatigue (MSF) (FSS>/=5; n=46) or MS-nonfatigue (MSNF) (FSS</=4; n=20). Forty-one patients were grouped into MS-depression (MSD) (n=15) or MS-nondepression (MSND) (n=26) by interview. Higher expanded disability status scale (EDSS) scores were noted in MSF than MSNF patients (P=0.0003); EDSS scores correlated with FSS scores (rho=0.43, P=0.003). However, fatigue was present in 58% (n=29) of relapsing-remitting patients and in 52% (n=26) of patients with mild physical disability (EDSS<3.5). Hamilton/Beck depression severity scores were higher in MSF than MSNF patients and correlated with FSS scores (P<0.05). MSD had higher FSS scores than MSND patients (P=0.008). After controlling for EDSS, depression severity continued to correlate with FSS scores (rho=0.37, P=0.02). After controlling for depression, FSS scores no longer correlated with EDSS scores (rho=0.27, P=0.09). Thus, MSF is independent of physical disability, but is associated with depression, suggesting that common mechanisms play a role in MSF and MSD including psychological factors or brain lesions in specific neuroanatomic pathways. Further study is warranted to determine if antidepressant medications improve fatigue in MS.