This network meta-analysis included 113 trials, 34 of which focused on fatigue. The trials were either randomized controlled trials (RCTs) or quasi-RCTs; they involved exercise interventions, behavioral interventions, and combinations of both.

There were 6 types of exercise intervention: aerobic (25 trials), balance (9), resistance (8), flexibility (1), combined exercise (24), and general (28).
There were 6 types of behavioral intervention: cognitive behavioral therapy (CBT, 14), energy conservation (12), neurocognitive rehabilitation (7), relaxation and biofeedback (5), emotional expression therapy (3), and education or information (mostly as active controls, 8).
There were 2 types of combined intervention: behavioral techniques integrated with exercise (11) and physical rehabilitation (i.e., physiotherapy, 6).
There were 2 types of control: medication (2) and “treatment as usual” (82).

The outcomes were subjective fatigue ≤3 months (primary outcome) and >3 to 6 months (secondary outcome) after treatment.

There were three tiers of analysis:

The first was a pairwise meta-analysis (82 trials).
The second involved additional trials that included subgroup analyses (113 trials total).
The third excluded 10 trials (out of the total 113) because they were too dissimilar from the other trials.

The treatments were ranked based on estimates of the surface under the cumulative ranking curve (SUCRA), where 1 indicates a 100% likelihood of a treatment being more effective than other treatments. The quality of the evidence was evaluated with GRADE (Grading of Recommendations, Assessment, Development, and Evaluations).

Direct comparisons in this meta-analysis

Best behavioral and/or exercise interventions for fatigue in MS Original paper