We examined efficacy and safety of one specific cranial electrical stimulator (CES) device at a fixed setting in subjects with treatment-resistant major depressive disorder (MDD). Thirty subjects (57% female, mean age 48.1 ± 12.3 years) with MDD and inadequate response to standard antidepressants were randomized to 3 weeks of treatment with CES (15/500/15,000 Hz, symmetrical rectangular biphasic current of 1-4 mAmp, 40 V) or sham CES (device off) for 20 min, 5 days per week. The primary outcome measure was improvement in the 17-item Hamilton Depression Rating Scale (HAM-D-17). Adverse effects (AEs) were assessed using the Patient Related Inventory of Side Effects (PRISE). Completion rates were 88% for CES, 100% for sham. Both treatment groups demonstrated improvement of about 3-5 points in HAM-D-17 scores (p < 0.05 for both), and no significant differences were observed between groups. Remission rates were 12% for CES, and 15% for sham, a nonsignificant difference. CES was deemed safe, with good tolerability; poor concentration and malaise were the only distressing AEs that differed significantly between CES and sham (p = 0.019 and p = 0.043, respectively). Limitations include a small sample and lack of an active comparator therapy. Although both treatment groups improved significantly, this CES at the setting chosen did not separate from sham in this sample. Thus we cannot rule out that the benefit from this setting used in this particular form of CES was due to placebo effects. Since this form of CES has other settings, future studies should test these settings and compare it against other CES devices. Clinicaltrials.gov ID: NCT01325532.