We explored the relationship between a set of immunological variables and a set of cognitive and functional status measures and a diagnosis of chronic fatigue syndrome (CFS) in civilians and veterans using various regression and factor analytic methods. Our approach emphasized the extraction of a few distinct factors in order to limit statistical problems associated with doing large numbers of multiple comparisons. This approach led to our finding cytokine data grouping into type 1 and type 2 clusters. A type 2 cluster plus a T and B cell factor predicted CFS caseness for Gulf War veterans but not for civilians with CFS. When a cognitive variable, reaction time, was added into the model, both immunological factors lost statistical significance; this indicates that the cognitive variable reaction time moderated the effects of the immunological factors in predicting patient status. We did a similar analysis on the roles of the immunological and cognitive variables in functional status using SF-36 data. Higher levels of these same two immunological factors predicted poorer general health as well as poorer physical and social functioning in Gulf War veterans but not in civilians with CFS. When the reaction time factor was added, only the lymphocyte factor remained significant. This implies that lymphocytes are directly related to functional status in Gulf War veterans with CFS, but the Th2 factor produces its effect on functional status via changes in cognitive abilities.
Copyright 2002 S. Karger AG, Basel