A randomized, double-blind, placebo-controlled, hybrid parallel-arm study of low-dose naltrexone as an adjunctive anti-inflammatory treatment for major depressive disorder

The participants (n = 48) will be adults aged 18 to 55 years with moderate MDD currently receiving antidepressant medication. Healthy participants (n = 24) will be recruited to complete the baseline outcome measures to enable comparison with individuals without MDD. The complete inclusion criteria are outlined in Table 1. Strict exclusion criteria will be applied to limit the heterogeneity of the sample. The exclusion criteria are outlined in Table 2.

Participants will be recruited by the members of the study team from general practices within the greater Auckland area and via advertisements placed in local newspapers, noticeboards, and online using social media, allowing potential participants to make initial contact. Participants may also be recruited from an ethics-approved database of individuals who previously expressed interest in participating in studies on MDD (Auckland Health Research Ethics Committee number AH23223). Participants will be checked for eligibility at a screening visit and approved for inclusion in the trial by the study psychiatrist. The participant information sheet and informed consent form will be provided to participants prior to screening, allowing time for participants to seek independent advice. The participant information sheet and informed consent form provide details on the study including the type of trial, proposed involvement of participants, possible side effects, and risks of participation. Participants will have the opportunity to ask questions of the study investigators prior to and during the screening visit. If individuals choose to participate, verbal understanding of the information and written informed consent will be given to members of the study team at the screening.

A randomized, double-blind, placebo-controlled, hybrid parallel-arm, superiority study (Fig. 1) will be used to test the potential of LDN as an adjunctive treatment for MDD. Participants with MDD will be allocated into parallel groups of high and low inflammatory status in blocks of 12 in a 1:1 ratio. The study will take place primarily at the Clinical Research Centre at the University of Auckland. After an initial screening, participants with MDD who meet the inclusion criteria will be prospectively stratified into low/high inflammatory status (n = 24 per group) based on the levels of high-sensitivity C-reactive protein (hs-CRP). Hs-CRP is a pro-inflammatory acute phase protein. Serum chemistry and hematology will also be conducted at the initial screening, including measurements of complete blood count, liver function tests, hs-CRP, and a human chorionic gonadotrophin (hCG) pregnancy test. The laboratory variable stated normal ranges will be used to categorize the results, and an additional interpretation by the study’s clinicians will deem abnormal values as either clinically or not clinically significant. A second blood sample, measuring only hs-CRP, will be taken at least 1 week later and be used to confirm inflammatory status. A high inflammatory state will be characterized by hs-CRP ≥ 3 mg/L, and a low inflammatory state will be characterized by hs-CRP ≤ 1 mg/L. Once the inflammatory status is confirmed, participants will complete a urine drug test and their baseline assessments and be randomly assigned to receive either placebo or LDN.

The primary, secondary, and tertiary endpoints of the study will be measured at baseline, following 12 weeks of LDN or placebo and following a further 12 weeks of LDN only. Previous studies using LDN indicate that a minimum of 12 weeks is required to observe anti-inflammatory effects [26, 39]. The anti-depressant effects of LDN will be measured using the Montgomery-Åsberg Depression Rating Scale (MADRS) which assesses the severity of depressive symptoms [40].

The primary outcome for this trial is change in MADRS scores at 12 weeks relative to baseline. The secondary outcomes include change in MADRS scores at 12 weeks in the high inflammatory group versus the low inflammatory group, detection of central inflammation in MRI scans, and measurement of peripheral inflammatory markers.

MRI scans will be conducted using the 3-T Siemens Vida Fit scanner at the Centre for Advanced MRI at the University of Auckland. A T1-weighted image will be acquired for segmentation and anatomical reference using a magnetization prepared rapid gradient echo (MPRAGE) sequence: repetition time (TR) = 2000 ms; echo time (TE)= 2.85 ms; flip angle= 8^o; 208 slices; slice thickness= 1.00 mm, field of view (FOV) = 256 × 256 mm, matrix = 256 × 256, voxel size = 1.0 × 1.0 × 1.0 mm, and acquisition time (TA) = 4 min 56 s.

The qMT technique will be used to measure the magnetization exchange rate. To acquire the qMT data, a series of MT-weighted 3D fast low-angle shot (FLASH) sequences are applied: echo time (TE) = 3.78 ms, repetition time (TR) = 35 ms, voxel size = 2 × 2 × 5 mm, varying flip angles for T1 mapping, and acquisition time (TA) = 10 min. Four pulse powers of 1000 degrees, 900 degrees, 800 degrees, and 400 degrees are applied to acquire 2, 3, 1, and 2 measurements, respectively. Finally, FABBER CEST in FSL is used to model the MT parameters.

Brain temperature will be calculated from the whole-brain spectroscopy data obtained using the EPSI sequence: TR1 = 1710 ms, TR2 = 591 ms, TE = 17.6 ms, voxel size = 5.6 × 5.6 × 10 mm, and TA = 17 min. The Metabolite Imaging and Data Analysis System (MIDAS) package will be used to analyze the data. The formula T_brain =  − 102.61 × ∆_{water − CRE} + 206.1^°C will be used to calculate the absolute brain temperatures according to the distance between the creatine (CRE) and water peaks.

DWI scans will use a three-shell protocol to acquire 18 non-diffusion-weighted images (b = 0 s/mm²) and 180 diffusion-weighted images (30 at b = 500 s/mm²; 60 at b = 1000 s/mm²; 90 at b = 2000 s/mm²) using non-collinear diffusion-weighting directions. The following are the other imaging parameters: TE/TR = 78/12,500 ms, slice thickness 2 mm, FOV = 224 × 224 mm, matrix = 112 × 112, resulting in 2 mm³ isotropic voxels, and TA = 15 min.

Blood samples will be acquired to measure peripheral inflammatory markers: hs-CRP, interleukin (IL)-1β, IL-2, IL-6, and IL-8. IL-10, IL-12p70, tumor-necrosis factor (TNF)-α, interferon-inducible protein (IP)-10, monocyte chemoattractant protein (MCP)-1, vascular endothelial growth factor (VEGF), regulated upon activation, normal T cell expressed, secreted (RANTES), erythrocyte sedimentation rate (ESR), and glycoprotein nonmetastatic melanoma protein B (GPNMB) will be analyzed using multiplexed bead-based immunoassays. The blood samples will be analyzed for ESR 1 h after collection. Prior to analysis of the remaining peripheral inflammatory markers, the blood samples will be centrifuged, and the plasma will be frozen at − 80 °C.

Tertiary outcomes include electroencephalography (EEG) assessments and questionnaires compared between baseline, at 12 weeks, and at 24 weeks and the change in MADRS score between 12 and 24 weeks. Questionnaires for tertiary outcomes include the Beck Depression Inventory (BDI-II; [41]), Behavioural Activation for Depression Scale (BADS; [42]), Profile of Mood States (POMS; [43]), Lifetime Stress and Adversity Inventory (STRAIN; [44]), the SF-36 health survey [45], and the Sickness Questionnaire (SicknessQ; [46]). These questionnaires will enable a thorough understanding of each participant’s symptoms, stressors, and quality of life. The National Institute of Health (NIH) Toolbox will be used to assess the aspects of cognition including attention, executive function, psychomotor speed, and memory [47]. Resting-state and task-based electroencephalography (EEG) scans will be conducted with tasks including doors [48], long-term potentiation (LTP, [49]), and the attention network task (ANT; [50]) to assess responsiveness to positive and negative feedback, memory, and attention. In combination, these outcomes will be used to determine if a specific profile of symptoms is associated with inflammatory MDD.

Finally, the Generic Assessment of Side Effects (GASE, [51]) questionnaire will be used to monitor the adverse effects of LDN, and treatment expectancy effects will be measured using the Stanford Expectations of Treatment Scale (SETS, [52]). Pregnancy status will be reassessed at 12 weeks with a human chorionic gonadotropin (hCG) urine test.

A complete list of outcome measures is summarized in the Standard Protocol Items: Recommendations for Interventional Trials (SPIRIT) figure (Fig. 2), and the SPIRIT checklist is provided as Additional file 1.

Following stratification of the eligible participants into high inflammatory and low inflammatory status, participants will be randomly allocated to parallel LDN or placebo groups in blocks of 12 in a 1:1 ratio. Computer-generated randomization will be performed by one member of the research team. LDN or placebo will be prescribed by the study psychiatrist(s) and delivered to the participant by the study pharmacist(s). The randomizer and pharmacist will be the only members of the research team unblinded to the identity of the medication. To maintain triple-blind conditions, the participant and study team members conducting the measurements of primary and secondary outcomes will not be aware of the identity of the medication.

During patient debriefing, participants will be asked to identify the medication they think they received. When all participants in their randomization block have completed the trial, participants and study team members will be informed of the correct medication identity. If the participant experiences a reaction or acute deterioration of health during the study and requires medical intervention, the identity of the medication may be revealed. An on-call member of the study team will be available 24/7 to break blinding. While this member is blinded, the randomizer will have prepared code-break envelope which may be broken in the event unblinding is required.

Naltrexone will be supplied to CompoundLabs by a licensed pharmaceutical material supplier who will confirm the identity and potency of the naltrexone. LDN capsules will be prepared by blending 1.5 mg naltrexone hydrochloride with microcrystalline cellulose (MCC) and filling the powder into a gelatin capsule shell. Placebo will be MCC only in a gelatin capsule shell. An unblinded pharmacist independent of the study team will store and dispense the interventions to participants, in accordance with Medicine Regulations 1984 [53]. The intervention will be dispensed to participants once per month. Participants in the durability arm will take one capsule orally once daily at night for 1 week, then 2 capsules at night for 1 week, then 3 capsules (4.5 mg) at night for 10 weeks. This dosing regimen will be repeated at 12 weeks, at the beginning of the 12-week open-label extension. If participants experience adverse effects, the dose may be titrated down to either 2 capsules daily or 1 capsule twice daily.

There will be no special criteria for modifying or discontinuing the allocated medication.

Participants will be contacted by the study team after the day of medication delivery at 1 day, 3 days, 1 week, 2 weeks, 4 weeks, 6 weeks, 8 weeks, 10 weeks, 12 weeks, 13 weeks, 14 weeks, 16 weeks, and 20 weeks. The study team may contact participants via email, phone call, or text message to remind participants to adhere to study protocols (e.g., take medication daily) and to check for any adverse effects. At 4, 8, 16, and 20 weeks, participants will return to the University of Auckland for a treatment refill. At 12 and 24 weeks, participants will return to the University of Auckland for follow-up measurements of all outcomes. Participants will return their capsule bottles at the end of each 4-week period, and the number of capsules remaining in the bottle will be used to calculate intervention compliance.

There will not be any modification to the usual access to care pathways while participants are enrolled in the study. If a participant becomes ineligible for the study while enrolled, the participant may be discharged from the trial. If a participant is discharged due to physical health or psychiatric reasons, the participant will be referred to appropriate medical services. Upon completion of the study, participants will be reimbursed for their time in grocery vouchers. Individuals who complete initial blood tests but do not meet the inclusion criteria will be given a $20 grocery voucher. Healthy control participants who complete a single visit will be given $80 in grocery vouchers. Individuals with MDD who participate in the 24-week trial will receive a total of $320 in grocery vouchers: $80 at each study visit and $20 at each collection of their next treatment supply. There is no expectation of harm to participants caused by the current trial.

Linear mixed effects models will be used to assess the primary outcome (change in MADRS scores at 12 weeks). Time (baseline and at 12 weeks) and drug (naltrexone and placebo) will be treated as fixed effects, and participants will be treated as a random effect. The primary estimand of interest will be the time × drug interaction coefficient with associated confidence intervals. Satterthwaite’s method (two-tailed) will be used to obtain p-values with an alpha set at p < .05.

To inform the power calculations for the primary outcome, a sensitivity analysis of the primary outcome was conducted for the fixed sample size of 48 (12 participants per group). Monte Carlo simulations were conducted using the mixed effects model described above with 10,000 simulations per run. Data were simulated on each iteration using the following parameters: n = 48, 4 dropouts with data missing at random, α = 0.05, (1−β) = 0.8, baseline MADRS scores of 30, random effect variance = 6.32, and error variance = 4.89. Variance estimates were obtained from linear mixed effects models fit to data from a previous antidepressant trial [54]. The present study is sensitive to detect changes of ~ 6 MADRS points according to the Monte Carlo simulations. The previous pilot study of 12 participants with MDD showed an 18-point decrease in MADRS scores with LDN compared to an 8-point drop with placebo [37]. The current study of 48 participants is considerably better powered for the primary outcome measure.

Sub-group analyses, such as a comparison of the hs-CRP level between sexes, may be conducted if a significant effect on hs-CRP is observed. A modified intention-to-treat scheme will be used including all participants who received at least one dose of intervention. All missing data will be classified as “missing at random” or “not missing at random” prior to unblinding. Data imputation techniques will be used where necessary.

The GASE scale, a 36-item scale that assesses the most frequent side effects in clinical trials of medicines [51], will be used to evaluate the side effects at monthly visits. Any adverse events that occur during the trial will be recorded, and any serious adverse events will be reported within 24 h to the Centre for Adverse Reactions Monitoring consistent with the guidelines provided by the New Zealand Medicines and Medical Devices Safety Authority (“MedSafe”).

The study will be conducted to International Conference on Harmonisation (ICH) Good Clinical Practice (GCP) clinical trial standards. Two independent consultant psychiatrists and a biostatistician will form the Data and Safety Monitoring Committee for the trial. The study is considered low risk given the comprehensive inclusion/exclusion criteria; however, in the unlikely event of a serious adverse event, the Safety Monitoring Committee may decide to suspend the study or request suspension until study protocols are appropriately revised.

Separate paper-based files will be kept for each participant; however, the majority of data will be captured by the online Research Electronic Data Capture (REDCap) tools hosted at the University of Auckland [55]. REDCap is a secure web-based platform designed to capture data and manage databases online. Demographics, medical history, height, weight, current medications, MADRS, and GASE scores will be entered directly into REDCap. Data from MRI scans and blood sample analysis (including hs-CRP and peripheral cytokine analysis) will also be stored electronically on secure University of Auckland servers on password-protected files.

During the study, participants will be identified by a unique study number and/or code. The name and any other identifying detail will not be included in any trial data electronic file. On all study-specific documents, other than the signed consent form, the participant will be referred to by their unique number/code. All data will be held for a period of 10 years from the completion of the study.

The results from the present study will be published in selected academic journals and presented at academic conferences. The results may also be distributed through social media, community forums, or news outlets. Participants may request a summary of their individual results at the end of the trial.