Depression affects up to 20% of people in their lifetime and can be a severe debilitating illness. Indeed, the World Health Organisation has estimated that depression will soon be the second leading contributor to the burden of disease worldwide. One of the big problems for patients and doctors is that currently available antidepressant drugs and psychotherapies do not work for 30% of people. However, about 60% of such treatment-resistant patients will recover fully with electroconvulsive therapy (ECT). Even though it was developed over 75 years ago, ECT continues to be the most powerful treatment for severe, often life-threatening, depression. Despite that, we have recently reported that severe depression symptoms return (called a "relapse") in nearly 40% of such responders within six months of completing a course of ECT. Actually, such high relapse rates are seen for all patients with treatment-resistant depression, irrespective of what treatment they have received. There is thus an urgent need for better treatments to prevent relapse and one such possibility is an old drug called ketamine. Ketamine blocks the activity of glutamate, one of the major chemical messenger systems in the brain. Because of this effect it is sometimes used as an anaesthetic but it can also make you feel a bit "high" and so is sometimes abused as a recreational drug. Fortunately, in small doses it is quite safe. Recently, it has been found that ketamine has a remarkably rapid, but brief, antidepressant effect, including reducing suicidal thoughts. We wish to evaluate ketamine as a way to reduce relapse rates in people who have just been treated successfully with ECT for severe depression. Developing such a new treatment, and understanding how it works, would be of tremendous benefit to persons with severe depression, their families, and the wider society.

DESIGN: This pilot trial has two phases. Phase 1 is a prospective open study that involves recruiting a cohort of patients with major depressive disorder (DSM-IV criteria) referred for ECT. Patients subsequently identified as being responders will be invited to participate in Phase 2, which is a two-group parallel-design randomised controlled pilot trial. These results describe the participants who completed full informed consent to randomisation. Consented ECT-responders were randomly allocated in a 1:1 ratio to a four-week course of either once-weekly ketamine or midazolam infusions, continuing regular care. Participants were followed-up over six-months following ECT to identify if and when relapse occured.

RANDOMISATION AND BLINDING: In the randomised pilot trial, ECT responders were randomised after baseline data collection. Patients and raters were blind to treatment allocation. The anaesthetist administering ketamine/midazolam infusions will not be blinded but will not be involved in assessments or data analysis. Computerised random allocation, using randomly permuted blocks, will be done independently.

PARTICIPANTS:

Inclusion criteria: Patients ≥18 years with unipolar major depressive disorder (DSM-IV), a 24-item Hamilton Rating Scale for Depression (HRSD-24) score of ≥21, and referred for ECT will be invited to participate in the open Phase 1. Patients will have a physical examination, routine haematology and biochemistry investigations, and an ECG to screen for any medical conditions that might affect ability to be treated with ECT or ketamine. For the randomised Phase 2, patients must have (i) received a substantial course of ECT in Phase 1 (i.e. at least five sessions), (ii) achieved at least response criteria (i.e. ≥60% decrease from baseline HRSD-24 score and score ≤16 on two consecutive weekly ratings), (iii) have a nominated adult who can stay with them for 24-hours on out-patient treatment days, (iv) have a Mini-Mental State Examination (MMSE) score of ≥24, and (v) be able to provide informed consent. Patients will continue on regular psychotropic and other medications as prescribed by their consultant psychiatrist. This reflects the conditions under which adjunctive ketamine would be used in routine clinical practice, thereby enhancing the generalisability and external validity of a future definitive trial.

Exclusion criteria are: any condition rendering patients medically unfit for general anaesthesia, ECT or ketamine; active suicidal intention; dementia; lifetime history of bipolar disorder, post traumatic stress disorder, or other Axis 1 diagnosis; ECT in the previous six-months; alcohol/substance abuse in previous six-months; pregnancy; inability/refusal to consent.

Patients will be given verbal and full written information and be asked to provide written informed consent.

INTERVENTIONS: Please see description

OUTCOMES Process outcomes

The focus of a pilot trial's outcomes is on trial process with assessment of the primary clinical outcome being secondary because the pilot itself is not designed to measure efficacy. Process outcomes that will inform a future definitive ketamine relapse prevention trial include information on the following:

• recruitment methods and rate
• willingness of participants to be randomised
• willingness of participants to complete assessments
• randomisation
• success of blinding
• ability to administer a course of ketamine infusions
• medical safety and acceptability of ketamine infusions in an ECT responder population
• rates of adverse dissociative and psychiatric events
• adherence to allocated treatment
• adherence to follow-up
• reasons for drop-out from treatment
• reasons for drop-out from follow-up
• a 95% confidence interval for the difference between the ketamine and midazolam groups in six-month relapse rates to help with power calculations for a future definitive trial

Clinical outcomes:

The following will be used to obtain baseline, intra-treatment, end-of-treatment and follow-up data. Most are standard measures used by ourselves and others in depression, ECT or ketamine studies. Raters will be trained on all measures before recruitment begins.

(i) Diagnosis and treatment history: Diagnosis of major depressive disorder will be confirmed using the mood episodes module of the Structured Clinical Interview for DSM-IV Axis I Disorders (SCID). The Columbia Antidepressant Treatment History Form (ATHF) will be used to obtain index episode treatment details and provide measures of treatment-resistance. Handedness will be recorded with the Edinburgh Handedness Questionnaire. The National Adult Reading Test (NART) will measure premorbid ability.

Additional baseline data from patient interview and case-note review will include age, sex, weight, height, occupation, educational attainment, duration of index depressive episode, number of previous depressive episodes, previous ECT, history of medical illness and surgical treatments, personal and family history of alcohol/substance dependency, presence of psychotic symptoms (detected by SCID), and current medications and other therapies. Changes in medications during Phases 1 and 2 will be documented at follow-up interviews.

(ii) Depression outcomes: The baseline clinical measure is absence of depression on the Hamilton Rating Scale for Depression. To enter Phase 1 patients must score ≥21. Response to ECT is defined as achieving ≥60% decrease from baseline HRSD-24 and score ≤16 on two consecutive weekly ratings. Remission criteria are ≥60% decrease in HRSD from baseline and score ≤10 on two consecutive weekly ratings. Patients must achieve at least response to participate in the Phase 2 pilot trial and be described here.

The primary clinical outcome measure is the relapse rate at six months as measured using the objectively-rated 24-item Hamilton Rating Scale for Depression (HRSD-24). Criteria for relapse are ≥10 point increase in HRSD-24 compared to baseline Phase 2 score plus HRSD ≥16; in addition, increase in the HRSD should be maintained one week later (if indicated, additional follow-ups will be arranged). Hospital admission, further ECT, and deliberate self-harm/suicide also constitute relapse. Timing of these events will be recorded.

Baseline and weekly intra-treatment course HRSD-24 scores will be obtained during the treatment periods of Phases 1 and 2. During the infusion sessions in Phase 2 HRSD-24 scores will be obtained 60 minutes before the infusion begins and at +120 and +240 minutes afterwards. Baseline scores on sleep and appetite items will be maintained for repeated measures within one day. The +240 HRSD-24 scores will serve as the weekly post-ECT scores up to follow-up-week 4. Depression measures will be repeated at weeks 6, 8, 12, 20 and 26 during the six-month follow-up. Subjective mood ratings will be measured at the above timepoints using the Quick Inventory of Depressive Symptoms, self-report version (QIDS-SR16).

(iii) Cognitive outcomes: Cognitive testing can be challenging for severely depressed patients. It is vital that patients be encouraged to remain in the trial throughout the whole study period and a great deal of tact is required of raters. Priority will be given to obtaining scores on the HRSD-24. In a recent meta-analysis we identified that ECT has the most pronounced subacute (i.e. 0-3 days post course) adverse effects upon verbal memory (delayed word list recall) and frontal executive function. Subanaesthetic doses of ketamine can cause problems with orientation, concentration, working and episodic memory but these resolve within two hours of beginning an infusion. There are no published data on effects of ketamine on cognition in ECT responders. We will use the following battery in Phase 1 (pre and post ECT course; the latter will serve as baseline for Phase 2) and Phase 2 (one day after the first and fourth infusions and at six-months). Parallel versions will be used to reduce practice effects.

1. Global cognition will be assessed with the revised Addenbrooke's Cognitive Examination (ACE-R; three parallel versions) which also generates Mini-Mental State Examination (MMSE) and verbal fluency scores. The ACE-R provides a total score (maximum=100) plus subscale scores for different aspects of cognition and has been used to study cognition in depression and by ourselves during ECT.
2. Forward and Backward Digit Spans - immediate short-term memory, attention and working memory
3. Trail Making Test (Part A) - motor and psychomotor speed
4. Frontal-executive function will be rated by Trail Making Test (Part B) plus letter and category verbal fluencies.
5. Anterograde verbal memory will be tested using the Free and Cued Selective Reminding Test variant (immediate and delayed recall) of the Buschke Selective Reminding Test.
6. To measure retrograde amnesia for autobiographical information we will use the Kopelman Autobiographical Memory Interview (K-AMI), The K-AMI will be administered pre and post ECT, after the fourth ketamine infusion, and at six months.

(iv) Ketamine psychotomimetic effects and adverse events: Acute psychotomimetic effects of ketamine are usually short-lived and restricted to the infusion period, resolving within one hour. They include dissociative and psychiatric symptoms. In line with previous ketamine trials we will measure dissociative effects with the Clinician-Administered Dissociative States Scale (CADSS), psychotomimetic effects with the Brief Psychiatric Rating Scale (BPRS; four-item positive symptom subscale), and mood elevation with the Young Mania Rating Scale (YMRS; mood item). These effects will be measured before, during (+35-40 mins) and after (+240 mins) the ketamine infusions. The Patient-Rated Inventory of Side Effects will be used to document other general adverse events by patients at the same time points.

All adverse medical, psychotomimetic and general events will be reported to the Trial Steering and Data Monitoring Committees.

STATISTICAL METHODS A formal sample size calculation is not appropriate for this pilot trial. Pilot trial data will be analysed on an intention-to-treat basis for all patients who completed at least one infusion. Data analyses will be performed blinded to allocation.

Inclusion Criteria:

• Patients ≥18 years with unipolar major depressive disorder (DSM-IV)
• 24-item Hamilton Rating Scale for Depression (HRSD-24) score of ≥21
• Referred for ECT

For the randomised Phase 2, patients must have

• received a substantial course of ECT in Phase 1 (i.e. at least five sessions)
• achieved at least response criteria (i.e. ≥60% decrease from baseline HRSD-24 score and score ≤16 on two consecutive weekly ratings)
• have a nominated adult who can stay with them for 24-hours on out-patient treatment days
• Mini-Mental State Examination (MMSE) score of ≥24
• able to provide informed consent

Exclusion Criteria:

• Any condition rendering patient medically unfit for ECT; general anaesthesia, ketamine or midazolam - assessed by physical examination, routine haematology and biochemistry investigations prior to enrolment in Phase I (routine care)
• Active suicidal intention
• Dementia, intellectual disability, or MMSE <24
• Lifetime history of bipolar affective disorder
• Current history of post-traumatic stress disorder
• Other Axis I diagnosis (DSM-IV)
• ECT in the six months prior to recruitment
• Alcohol dependence or substance misuse in the six months prior to recruitment
• Pregnancy or breast-feeding
• Residing in a nursing home
• Prisoner
• Diagnosis of terminal illness
• Inability or refusal to provide valid informed consent

n=6 participants were randomised

8 months (8 week treatment period and 6-month follow-up assessment period)

Tolerability of the trial agents was assessed at multiple points before, during and after treatment sessions over the 8 week treatment period using a battery of assessments(CADSS, BPRS, PRISE and YMRS) as well as assessment of physical health parameters before, during and after assessments. In addition, follow up assessments were performed at set intervals over the six-month follow-up period.

Recruitment and randomisation rates in this pilot trial were low, leading to small participant numbers analysed. Due to the potential for breach of confidentiality, limited participant-level information has been reported.