Six RCTs (910 patients, range 46 to 311) were included in the review.
Depression severity: Modafinil treatment statistically significantly improved depression scores compared to placebo (Hedge's g -0.35, 95% CI -0.61 to -0.10; six RCTs). However, there was evidence of substantial statistical heterogeneity (I²=67%). Sensitivity analysis excluding one trial did not significantly alter the results, but did explain the statistical heterogeneity (I²=0%).
Subgroup analyses showed that modafinil improved depression scores in patients with unipolar depression but this was not quite statistically significant (Hedge's g -0.41, 95% CI -0.84 to 0.01; four RCTs). However, the improvement became statistically significant when the one trial that showed very different results was removed. Modafinil statistically significantly improved depression scores in patients with bipolar depression (Hedge's g -0.30, 95% CI -0.52 to -0.09; two RCTs).
Secondary outcomes: Modafinil statistically significantly increased rate of remission compared to placebo (OR 1.61, 95% CI 1.04 to 2.49; NNT 10; five RCTs; I²=32%). Response rates were similar between participants who received modafinil or placebo (five RCTs). Modafinil appeared to be generally safe and well tolerated. Modafinil had a statistically significant benefit on fatigue scores compared to placebo (Hedge's g -0.15, 95% CI -0.28 to -0.02; six RCTs) but had no significant effect on sleepiness.
Compared to placebo, modafinil had a statistically significant early treatment beneficial effect on total depression scores at one week (four RCTs) and on fatigue (two RCTs) and sleepiness (three RCTs). There were no differences between modafinil and placebo on sad mood scores (results were partly reported in the review).
Results for meta-regression analyses were reported in the review.