A similar pattern of associations between pro-inflammatory markers and food groups/physical activity was observed within each group, although the association between TNF-a and red meat consumption failed to reach significance, probably due to reduced statistical power ( p > 0.4).
Finally, the associations between diet, physical activity and inflammation were significant in the entire group and remained significant or borderline significant in both subgroups, i.e., cognitively intact or patients with MCI.
This effect reached significance after correcting for multiple comparisons in the subgroup of persons with MCI and approached significance in the analyses on the total sample.