Effects of Erythropoietin for Cognitive Side-effects of ECT

The trial will include patients with a diagnosis of major depression (MDD) unipolar disorder (UD) or bipolar disorder (BD) with a current moderate to severe depressive episode symptoms (a score of \>17 on the Hamilton Depression Rating Scale 17-items (HRDS-17) scheduled for ECT treatment. Patients will be recruited from Psychiatric Centres in The Mental Health Services in the Capital Region of Denmark and will undergo an eligibility assessment prior to randomization to 4 intravenous infusions of either recombinant human EPO (40.000 IU/ml; Epoetin alpha; Eprex, Janssen-Cilag) or placebo (1 ml NaCl) diluted with 100 ml saline (0.9% NaCl). Cognitive functions, mood symptoms, and blood- and urine markers of inflammation, oxidative stress, and neuroplasticity will be assessed 3 times during the trial. First time at baseline, second time 3 days after ECT session 8 (patients skip one ECT session day after 8 ECTs to minimise the confounding effects of acute side-effects of ECT due to anaesthesia etc.), and the third time at a 3 month follow-up after ECT completion. In addition, the neuronal substrates for potential effects of EPO on cognition are investigated with structural and functional MRI after 8 ECT sessions (after 3 weekly EPO or saline infusions). Block randomization and power calculations have been conducted by the independent Pharma Consulting Group AB (www.pharmaconsultinggroup.com). Treatment groups are stratified for age (\>40 or \<40) and gender. The difference in cognitive change between EPO and saline-treated groups from baseline to post-treatment in our previous trial was 0.5 SD. Based on these findings, the sample size of N=52 (n=26 per group) in the current trial will reach a \>0.8 power to detect a clinically relevant difference in the primary outcome measure (the cognitive composite score) between the 2 groups at an alpha level of 5% (two-sided test). The study is also powered to investigate differences in functional magnetic resonance imaging (fMRI) blood-oxygen dependent level (BOLD) response in key neural networks based on previous fMRI studies from our group in which sample sizes of 30 age and gender matched participants (n=15 per group) had the power of \>0.8 to show drug-related effects on task-related neural response at an alpha level of p\<0.05. In the current trial, inclusion of 52 participants (n=26 per treatment group) therefore ensures sufficient statistical power to detect EPO-related effects on neural activity. Behavioural, mood, and biomarker data will be analysed using Mixed Models Design and Intention to Treat (ITT) approaches. Resting state and task-related fMRI data will be pre-processed and analyzed using FMRIB Expert Analysis Tool (FEAT) and the 'randomize' algorithm integrated in FSL, FMRIB Software Library (www.fmrib.ox.ac.uk/fsl).