The primary objective was updated and clarified in the synopsis and section 10.1 to include the wording “and assess the mood-rating score difference between arms from before the first infusion to 24 hours after the final infusion, supplemented by a 95% confidence interval. There will also be a 24-week follow-up after the final infusion session.” The Exclusion criteria was updated in the synopsis and section 11.2.3 to include “Currently taking any of the contraindicated medications listed in section 12.7.2” The period of clinical supervision was updated to be consistent throughout the protocol to include “from the beginning of the infusion” in sections 9.2.2, 11.3, 11.3.2.5, 11.3.4, 11.3.9 and section 11.5.

Pharmacovigilance contact information has been updated throughout the protocol from CRF-UCC to SJH-CRF Observer’s Assessment of Alertness/Sedation Scale - Responsiveness Subscale included in assessments section 11.3 and 11.3.4

Correction of error to clarify that Concomitant medications two months prior to visit 0 will be documented. Section 11.2.1 and 11.3.9 updated to include; “Where possible, patients taking any regular benzodiazepines (every day for the past five days) should omit their dose on the morning of infusion sessions (see section 12.7.1). It is appreciated that omission of benzodiazepines may not be possible for all patients. This will be as per the Investigators discretion and documented in the patient notes”. Section 12.7.1 updated to include: “However where possible, patients taking any benzodiazepines should omit their dose on the morning of infusion sessions. As per the SmPc for Ketalar, Diazepam is known to increase the half life of ketamine and prolongs its pharmacodynamic effects. Concurrent use of diazepam or other benzodiazepines will increase plasma levels and reduce the clearance rate of ketamine. Section 11.3.1 and 11.3.10.2 updated to include sponsor e-mail address clinicaltrialsponsorship@tcd.ie Table 1 Schedule of Assessments updated to include, fasting >6 hours before dosing and the omission of diazepines on the morning of dosing. Sections 9.2.2, 11.3.9 and 11.4 updated to include video conference instead of home visits. Section 13.2.2 updated to include Unlikely related: A clinical event, including laboratory test abnormality, with a temporal relationship to drug administration which makes a causal relationship improbable, and in which other drugs, chemicals or underlying disease provide plausible explanations.

• updated to clarify that participants will receive a course of up to eight infusions over four weeks • updated to include fasting for at least 3 hours before dosing but should take their cardiovascular medications with a sip of water. This has been reduced from six to three hours for patient comfort. Patients are not undergoing general anaesthesia, hence a prolonged fasting time is unnecessary. • updated to clarify that scores relating to sleep, appetite and medication adherence are carried over during visits. These include items 4 and 5 on the MADRS, and items 1,2,3,4,6,7,8,9 & 17 on the QIDS-SR. These scores will be carried over from -40 minutes before each infusion to +60 minutes, +120 minutes, and +24 hours after each infusion. • Not all patients have an ECG done on admission. For logistical reasons, timing of pre-infusion ECG has been extended to include period from admission up to first infusion and review by the trial anaesthetist. • To minimise patient burden, we are removing the NART assessment from the protocol as it is not essential. • Clarification of exclusion criteria regarding comorbid Axis 1 diagnoses. • We have broadened the time for completion of pre-infusion MADRS (-40 +/- 20 mins) to facilitate clinic logistics. • We have clarified exclusion criteria to be in line with SmPC’s for both ketamine and midazolam. Patients can be randomised to either midazolam or ketamine. Bradycardia is a known adverse effect of midazolam and ketamine. Therefore, patients with pre-existing bradycardia will be excluded. • We have removed “or 20% increase” referring to blood pressure increase lasting more than 15 minutes that would render a patient ineligible to continue in the trial. People with low-normal blood pressure can have a 20% increase yet still be within the normal range of blood pressure. • corrected collection of concomitant medication at Visit 0 from two months prior to baseline to one month

• Sections 7 and 11.2.3. Exclusion criteria with regards to psychiatric comorbidities have been clarified and amended to bring in line with recent major ketamine trials. • Data entry procedure has been changed from single to double entry to help minimise data entry errors.

• Typographical and grammatical errors have been corrected throughout the text. • The former name (NUIG) for the University of Galway has been replaced with the new name throughout the text. • The millilitre amount of intravenous fluid has been removed throughout the text since the amount of fluid infused depends on the IMP and the weight of the patient. • In Section 17, data retention period has been updated to 25 years in accordance with Clinical Trial Regulation (CTR) (EU Regulation 536/2014) as this trial will soon be transitioning to CTR. • Statistical analysis plan in Sections 7 and 14 have been updated in line with the Statistical Analysis Plan V2.0 approved by the Data Monitoring Committee and the Trial Steering Committee. • In Section 11.3.2.3 and Table 1, the CTQ and the SAPAS have been removed to reduce participant burden as they are non-essential and the study will be underpowered for these analyses. • In Sections 7 and 10.3, exploratory objective 6 has been removed as the study will be underpowered to address this research question. • The descriptive term “severe” depression has been removed throughout the text to make the text consistent with the MADRS rating scale cut-off of 20 for entry which technically also includes moderate depression. • Concomitant medications are recorded primarily for safety reasons before and during the randomised treatment period (Visits 0-8). To ensure accuracy and full visibility to the monitoring staff, medications for Visits 0-8 are now being exported directly from the participant’s electronic health record instead of manually recorded on a handwritten medication log. At long-term follow-up (Visits 9, 10 and 11), concomitant medications will continue to be recorded on another scale (CSRI). To avoid duplication, the concomitant medication log is no longer recording medications at these long-term follow-up timepoints. This has been clarified in Section 11.3.2.9. and Table 1.

• There is no item 17 on the QIDS-SR16. This error has been rectified throughout the text. • It has been clarified in Table 1 footer that BMI is a derived variable from height and weight which are recorded variables. • In Section 11.3.2.1. and Table 1, the assessment of treatment-resistant depression has been simplified from the dimensional Maudsley Staging Method (MSM) to the categorical Antidepressant Treatment Response Questionnaire (ATRQ). This is to bring the methodology of the trial in line with other major trials in this area. There is no additional participant burden with the change from MSM to ATRQ since all the necessary information to rate the ATRQ has already been collected for all existing participants as part of the MSM. • In Section 9.2.2., it has been clarified that “low” doses refer to subanaesthetic doses. • In Section 11.3.4. (iv), recent references have been added to support the statement that there is no evidence of a withdrawal syndrome in the published ketamine literature to date. The risk of withdrawal remains a theoretical concern. The literature on ketamine safety is monitored on an ongoing basis by the investigators. Should any new concerns arise in the future, the Protocol and the PIL will be updated as appropriate. • In Section 11.3.6., the term “Withdrawal eCRF” was replaced with the correct term “End of Study Form”. • In Section 13.2.5., the location of the displaying of the Emergency unblinding procedure has been updated. • In Section 16.1., it has been clarified that information exported from the electronic health record cannot be pseudonymised. These data are filed separately from the pseudonymised source data documents.

• It has been clarified throughout the text that AEs are followed up until resolution or final visit. • Section 23 has been updated to remove the section about trial newsletter. • The role of the sponsor in publication activities was removed in Section 23. • The Department of Psychiatry, Trinity College Dublin affiliation has been added to Section 23 • The definition of relapse (Section 11.3.3) has been updated to bring in line with recent major trials of antidepressants in relapse prevention.

• It has been clarified throughout the document that depression can be “moderate or severe” at trial entry. • Population figure for Ireland has been updated with a 2023 estimate on Page 17.