The purpose of this study was to evaluate the immune dysfunction hypothesis of chronic fatigue syndrome (CFS) by comparing immunologic data from patients with CFS with data from patients with other fatiguing illnesses--major depression and multiple sclerosis (MS)--and with data from healthy sedentary controls. The subjects were 65 healthy sedentary controls, 71 CFS patients (41 with no axis-I diagnosis), 23 patients with mild MS, and 21 patients with major depression. Blood was sampled and assayed for the following: (1) immunologic serologic variables--circulating immune complexes (i.e., Raji cell and C1q binding), immunoglobulins A, E, G, and M, and IgG subclasses; (2) cell surface activation markers--the proportion of CD4+ cells expressing CD45RA+ and CD45RO+ and the proportion of CD8+ cells expressing CD38+, CD11b-, HLA-DR+ and CD28+; and (3) natural killer (NK) total cell count as well as the proportion of lymphocytes expressing NK cell surface markers (i.e., CD3-/CD16+ and CD56+. Of the 18 variables studied, differences between CFS patients and controls were found only for IgG1 and IgG3. When CFS patients were stratified by the presence or absence of concurrent axis-I disease, it was the group with axis-I disorder that had the lowest IgG1 values-contrary to expectation. When data from patients with MS and major depression were also evaluated, the subclass deficiency was no longer significant. The one group to show evidence for immune activation (i.e., an elevated proportion of CD4+ cells expressing the CD45RA+ activation marker) was the group with mild MS. These data support neither immune dysfunction nor immune activation in CFS or in major depression, for the variables studied. The reductions in IgG subclasses may be an epiphenomenon of patient or control subject composition. In contrast, MS, even in the mild and early stages, as in the patients studied here, is associated with immune activation.