Low-dose naltrexone for post-COVID fatigue syndrome: a study protocol for a double-blind, randomised trial in British Columbia


Background and rationale

Approximately 15–20% of adults with a confirmed or suspected SARS-CoV-2 infection experience long-term symptoms lasting over 3 months.1–4 The presence of new or persistent symptoms following acute COVID-19 disease is now referred to as post COVID-19 condition (PCC) or ‘long COVID’.5–11 Among the hundreds of symptoms reported by people with PCC, fatigue is one of the most common and may have the greatest impact on functioning.5 12–20 Given that millions of individuals may be currently affected by PCC worldwide, it has become a priority to investigate the potential treatments in randomised controlled trials (RCTs).5 21 22

However, it has been challenging to identify candidate treatments for PCC as it is a heterogeneous illness, and the underlying pathobiology is poorly understood. It is suspected that different groups of people with PCC may have distinct underlying disease processes, such that the ideal pharmacological therapy may not be the same for all.5 21 Increasingly, studies have suggested that PCC may not represent a single disease but rather a collection of different conditions or subtypes.15 23 24

For example, clinical experience and patient-centred studies have indicated that a proportion of people with PCC present with a symptom profile indistinguishable from myalgic encephalomyelitis/chronic fatigue syndrome (ME/CFS).25–30 ME/CFS is characterised by persistent disabling fatigue accompanied by other symptoms including non-restorative sleep and post-exertional malaise.31–33 While the precise pathogenesis of ME/CFS also remains unresolved, it usually follows acute infections.34 When provoked by a viral infection, ME/CFS is often referred to as a post-viral fatigue syndrome (PVFS).34–37 It is believed that some PCC patients have developed a PVFS from SARS-CoV-2, and we will refer to this subset of PCC patients as having ‘post-COVID fatigue syndrome’ (PCFS).35–38

A low dose of the medication naltrexone is a potential treatment for PCFS.39 40 Naltrexone is an opiate antagonist that is approved for treatment for alcohol and opiate use disorders.41 For these indications, it is typically prescribed at 25–50 mg.42 At lower doses (≤5 mg), it has been used off-label for chronic pain, multiple sclerosis, Crohn’s disease, recurrent depression, fibromyalgia (FM) and ME/CFS.43–57 Although evidence supporting the use of low-dose naltrexone (LDN) in ME/CFS has been limited to case series and chart reviews,46 54 it has been investigated in clinical trials for related conditions such as FM.50 51 54 57 In these and other studies, LDN has been shown to be safe with a limited side-effect profile.49 51–53 57

Based on its hypothesised mechanism of action, it is plausible that LDN could be efficacious for ME/CFS and PCFS. LDN increases circulation of the endogenously produced opiate-like molecule beta-endorphin, which is reduced in ME/CFS.58 59 Furthermore, LDN has been found to antagonise toll-like receptors on neuroglia and peripheral blood mononuclear cells, resulting in reduced production of inflammatory cytokines such as interleukin-6 (IL-6) and tumour necrosis factor (TNF).60–63 Increased IL-6 and TNF signalling have been implicated in PCC,64 and studies have implicated increased neuroinflammation in ME/CFS and PCC pathogenesis.5 65–68

There is ongoing public interest in investigating LDN for PCC. Media outlets including Rolling Stone, National Geographic, Reuters and The New York Times Magazine have all touted LDN as a potential PCC treatment, citing the anecdotal experiences of people with PCC and physicians.69–76 However, published evidence for LDN in the post-COVID-19 context remains limited. In a single-centre study, 52 patients treated with LDN had, on average, overall improvement in activities of daily living, energy levels, pain, concentration and sleep disturbance.40 In a retrospective study, 37 of 59 (62.7%) patients treated with LDN reported improvement in at least one symptom.77

RCTs are required to determine whether LDN is an effective treatment for post-COVID-19 symptoms. Since there is no widely accepted pharmacological treatment for PCFS, the ideal comparator group is a placebo. Accordingly, we have designed a double-blinded placebo-controlled trial of daily LDN versus placebo for the treatment of fatigue severity in PCFS.


Study objectives are outlined in table 1. The primary objective is to determine whether LDN can reduce fatigue severity associated with PCFS, as measured by the Fatigue Severity Scale (FSS). The secondary objectives are to determine whether it can reduce pain, reduce symptom severity, improve health-related quality of life (HRQOL) and increase activity levels. We have developed additional exploratory objectives that examine other patient-reported outcome measures (PROMs), laboratory outcomes and physical measurements.

Table 1

Summary of study objectives and associated outcomes

Trial design

The development of this trial protocol followed the Standard Protocol Items: Recommendations for Interventional Trials (SPIRIT) guidelines.78 This trial is designed as a randomised, controlled, double-blind prospective trial with two parallel groups and a primary end point of fatigue severity at 16 weeks. The intervention group will receive LDN capsules dosed at 1.0 mg to 4.5 mg daily and the control group will receive placebo capsules. Randomisation will be stratified by sex and performed as permuted block randomisation with a 1:1 allocation. The trial will be conducted in British Columbia (BC), Canada.

Methods and analysis

Study setting

The trial will involve a collaboration between BC’s Post-COVID-19 Interdisciplinary Clinical Care Network (PC-ICCN) and the Complex Chronic Diseases Program (CCDP) at BC Children’s and Women’s Hospital and Health Centre (C&W) located in Vancouver. The PC-ICCN was founded as a learning health system for post-COVID-19 care and research in BC.79–82 The network previously comprised of five physical Post-COVID Recovery Clinics (PCRCs) but has now consolidated to a single virtual programme. Adults throughout BC may be referred to this programme by their primary care provider (PCP) if they have had COVID-19 and meet the criteria for PCC. The CCDP is an interdisciplinary programme that supports patients with ME/CFS and related conditions.83

Eligible participants will be recruited from throughout BC. Participants will have virtual or in-person study visits, may have their study product and pedometer delivered to them, complete questionnaires electronically and have blood tests done at their local laboratories. The collection of exploratory data during in-person visits will be optional.

Eligibility criteria

Inclusion and exclusion criteria are listed in table 2. To be included, participants must be aged 19–69 years, have significant fatigue and related symptoms that started after a SARS-CoV-2 infection and meet the criteria we have developed for PCFS. These criteria are adapted from the Institute of Medicine (IOM) ME/CFS standard clinical criteria,33 but with a duration of symptoms of 3 months rather than 6 months to be consistent with PCC definitions (box 1).11 Diagnosis for eligibility purposes will be determined from clinical assessment by a study physician and supported by laboratory data and responses to the screening and baseline questionnaires. If there is clinical uncertainty regarding the diagnosis, the case will be discussed with a second physician. Participants who do not have a documented positive PCR test will be eligible if they are determined by a physician through medical history to have had a positive rapid antigen test (RAT) or compatible symptoms. Individuals will be excluded if they have a history of ME/CFS prior to SARS-CoV-2 infection, have significant liver disease, have taken naltrexone within 30 days or have taken opioids within 15 days.

Table 2

Inclusion and exclusion criteria

Box 1

Study diagnostic criteria of post-COVID fatigue syndrome

Diagnosis requires that the patient have the following three symptoms after a SARS-CoV-2 infection:

  • A substantial reduction or impairment in the ability to engage in pre-illness levels of activity (occupational, educational, social or personal life) that:

  • Post-exertional malaise.

  • Unrefreshing sleep.

At least one of the two manifestations must be present:

  • Cognitive impairment.

  • Orthostatic intolerance.


Absence of other diseases or conditions that explain symptoms, based on differential diagnosis.

Table 3

Product supply timeline and titration schedule


Eligible participants will be randomised at a ratio of 1:1 (n=80 each) into either an active treatment group with LDN or a placebo. The treatment duration is 16 weeks. The LDN and placebo will be compounded by Macdonald’s Prescriptions Labs (Vancouver, BC) and dispensed at the C&W Pharmacy where the blinding will occur.

Macdonald’s Prescriptions Labs will compound the required doses of LDN from Naltrexone Hydrochloride USP supplied by MEDISCA in empty gelatin CONI-SNAP capsules.84 The compounded LDN capsules will be filled with CELLULOSE, NF/EP (Microcrystalline) (Flocel 101).85 Placebo capsules will look identical to the compounded LDN capsules and filled with the same diluent and food colouring. We will complete batch testing of the LDN and placebo compounds (online supplemental appendix 1).

Supplemental material

The dose-titration schedule from 1 mg to 4.5 mg is outlined in table 3. The drug will be dispensed to participants by certified courier, temperature-controlled shipping, in-person pick-up or delivered by staff. Participants will be able to adjust treatment doses by reverting to the previous well-tolerated dose if they experience persistent but minor side effects following any increase in dose. If a participant has reverted to a previous dose, that dosage will be maintained for the remainder of the study period. Changes in doses will be documented by the participant by completing a daily dosing diary, completed for the first 4 weeks and 7 days after any change in dose.

Figure 1
Figure 1

Flowchart of initial study procedures. This flowchart outlines the process for study recruitment, eligibility assessment and baseline assessments. A full study timeline is outlined in table 4. BC, British Columbia; LDN, low-dose naltrexone; PCC, post-COVID-19 condition; PC-ICCN, Post-COVID-19 Interdisciplinary Clinical Care Network; SAQ, Short Answer Questionnaire.

Table 4

Participant timeline and schedule of study procedures

By allowing participants to reduce doses if experiencing any potential side effects, we expect low rates of medication use interruption. In addition to diaries, participants may also have visits or contact with the study team where adherence can be discussed. Furthermore, there are treatment compliance questions asked with each series of questionnaires. The participants will be asked to return the unused study drug, empty containers and study drug diary sheet(s).

Participants will be asked to maintain any other regular medications at their current doses for the duration of the trial unless there is an essential need for a new medication or dose change. Participants can withdraw from the study at any time without giving reasons. Withdrawal criteria are described in online supplemental box S1.

Outcome measures

The primary outcome measure is fatigue severity, as measured by the FSS. The FSS is a 9-item PROM scored from 9 (least fatigue) to 63 (most fatigue).86 A score of >36 is consistent with clinically significant fatigue.87 88 The FSS has been validated in multiple diseases and has been used in randomised trials for ME/CFS.87 89–91 The FSS received the highest level of recommendation of any subjective fatigue measure for ME/CFS by the National Institute of Neurological Disorders and Stroke Common Data Elements (NINDS CDE) Project and was a recommended measure by the Post-COVID-19 Core Outcome Set (PC-COS) initiative.17 18 92–95 We have previously investigated the FSS in patients with PCC in BC and demonstrated that the instrument has strong acceptability, internal consistency and construct validity in this population.17

Secondary PROMs will include pain severity as measured by the pain Visual Analogue Scale (VAS); total symptom score on the Patient Phenotyping Questionnaire Short Form (PQSymp-12) and HRQOL captured by the EQ-5D-5L instrument. Pain is a common symptom in PCC, and studies have suggested that LDN may be an effective analgesic.5 16 43 49 52 53 55 The pain VAS is a single-item tool that has been shown to have strong psychometric properties among patients with chronic pain.96 The PQSymp-12 is a 12-item questionnaire that covers seven clusters of symptoms derived from the Canadian Consensus Criteria for ME/CFS; it has been recommended as a core assessment measure for ME/CFS by the European Network on ME/CFS (EUROMENE)97 and is included in the UK ME/CFS biobank.98 The EQ-5D-5L is a generic HRQOL instrument that was recommended by PC-COS.95 By applying Canadian preference weights, responses to the EQ-5D will be used to derive a health utility (HU) score from 0 (dead) to 1 (perfect health).99

An additional secondary outcome is the average step count. We will ask participants to wear a pedometer and document daily step counts for 7 days prior to starting the study drug and again in week 16. All participants will use the same brand and type of pedometer (OZO Fitness CS1 Easy Walk Pedometer). Step counts have been used previously in randomised trials to measure a change in activity levels among patients with ME/CFS.91 100–102

There will be several exploratory outcomes (online supplemental table S1), including PROMs (fatigue VAS, sleep, depression symptoms, anxiety symptoms, self-reported health and functional status), laboratory based (inflammatory markers, CK, thyroid profile, AM cortisol and ACTH level) and based on physical measurements (grip strength, sit and stand test, and orthostatic changes in vitals). Physical measurements will be limited to participants who choose to attend in-person visits.

MRI study

As a sub-study of the RCT, 25 participants of each study arm are planned to have brain MRI scans at baseline prior to the intervention/placebo and at 16 weeks. A multimodal functional and spectroscopy (fMRI/MRS) protocol piloted in an ME/CFS study (REB# H20-01804, unpublished) will be employed (online supplemental figure S1). MRI findings will be linked to the primary and other outcome measures.

Participant timeline

The participant timeline is detailed in figure 1 and table 4.

Sample size

The sample size was calculated based on the primary hypothesis of reduction in fatigue severity with treatment. To detect a 4.7-point difference (effect size (d)=0.5) in the Fatigue Severity Scale (FSS) (9–63) between arms, we estimate a sample size of 64 participants per arm assuming 80% power, 5% significance and a pooled SD of 9.4 (estimated from the CCDP Data Registry).83 To account for possible loss to follow-up of 20%, we estimate a final sample size of 80 per arm, for a total target sample size of 160 participants. We chose this method for sample size estimation (as opposed to the use of a minimal clinically important difference (MCID)) because we believed this to be a realistic treatment effect and there were no published MCID values available for the FSS in ME/CFS or PCC.103 104 In a sensitivity analysis, we calculated sample size using a published MCID for systemic lupus erythematosus and this yielded a similar estimate (online supplemental appendix 2).105

Recruitment and screening

New PC-ICCN patients will be contacted regarding study participation, and the PC-ICCN directory will be used to identify other candidate PCC patients to contact. Additionally, the trial will be accessible to individuals through REACH BC, a provincial online platform that facilitates connections between research studies and participants. All potential participants will be asked to complete an online pre-screening questionnaire, and those potentially eligible will meet with research staff to provide consent. Consented participants will complete baseline questionnaires and be assessed by a physician to confirm eligibility. Baseline laboratory studies for all participants will be done prior to initiation of the study drug and abnormal results will be reviewed by a study physician.


Participants will be randomised into either the LDN treatment group or the placebo at a ratio of 1:1 (n=80 each), as per a computer-generated randomisation schedule stratified by sex and using permuted blocks varying between two, four and six participants. A statistician who is not part of the study team will generate a randomisation sequence and corresponding randomisation codes. The randomisation codes will be used to maintain the blinding and will be uploaded to REDCap. The randomisation sequence will be provided to the unblinded CWH Research Pharmacy. After confirmation of eligibility, research staff will randomise participants by REDCap, which will provide the randomisation code. CWH Pharmacy staff will then dispense the study drug based on the randomisation allocation sequence.


All participants will be blinded to their treatment regimen. The placebo and intervention capsules will appear identical, and the C&W Pharmacy will distribute the study drug to study staff in identical containers. Participants, their healthcare providers and all study staff including research assistants, coordinators, statisticians, trial physicians and investigators will be blind to allocation. Unblinding will only occur when knowledge of the actual treatment is essential for further management of the patient or investigation of serious adverse events (SAEs). If unblinding is deemed necessary by the DSMB or investigator, the C&W Pharmacy will be contacted for release of treatment allocation.

Data collection and management

We will use the secure REDCap platform for the storage of study data.106 107 Participants will complete questionnaires electronically, with the links provided to the participant via email. Data from other sources will be entered manually and will include study physician assessments, laboratory results, dose diary information, step counts, physical assessment measures and adverse events (AEs). REDCap field validation tools will be used where possible to optimise data accuracy (eg, dates that are out of range and data that are missing). No new data will be collected from participants who withdraw, except for reason for withdrawal and details regarding AEs and SAEs.

Biological specimens

Leftover plasma will be stored at −80°C at the BC Children’s Hospital Biobank for up to 10 years to allow for additional sample testing related to this protocol that may be identified from the results of this study.

Monitoring and oversight

A Trial Steering Committee (TSC) will be formed with patient partners, investigators and other research team members. Additionally, we have formed a Data Safety and Monitoring Board (DSMB) that is comprised of peer researchers with expertise in clinical trials and ethics and independent from the study team. Lastly, an independent study monitor from the CWH Quality Assurance Office has been hired to verify participant rights and well-being, data collection and compliance with regulatory requirements. Roles of the TSC, DSMB and study monitor are outlined in online supplemental box 2.

Statistical methods

Primary and secondary outcomes will be analysed by intention-to-treat. The primary outcome (FSS score at 16 weeks) will be analysed using a linear mixed effects model adjusting for baseline level, sex (stratification factor) and other relevant prognostic factors identified a priori. The model will include interaction between treatment arm and time, treatment arm and baseline level and include all post-randomisation timepoints at which the FSS is captured. To assess the FSS at 16 weeks, we will calculate an estimated marginal mean difference between arms with a corresponding 95% CI, with statistical significance set at 0.05. Similar contrasts at each interim time point will be provided. Effect modification by baseline FSS level will be demonstrated graphically. Participants who are lost to follow-up will be compared descriptively with those who remain in the trial. If selection bias occurs, we will consider inverse probability weights for censored individuals.

For secondary and exploratory outcomes, questionnaire, laboratory and physical measure data will be analysed similarly with generalised linear models, adjusting for baseline level and other relevant prognostic factors and using link function based on the variable type from questionnaires (eg, logit for binary outcomes).

Effect modification by baseline factors will be considered by the inclusion of interaction terms with treatment arm in the above models. Possible effect modifiers include baseline fatigue severity, sex, gender, age, severity of and time of acute SARS-CoV-2 infection, pre-existing co-morbidities, COVID-19 vaccination, final dose and side effects. The significance of effect modification will be based on the likelihood ratio test comparing models with and without the interaction term.

Dose-response analyses will involve a comparison of various dosing levels as a covariate versus control in the primary linear model. Secondary analysis will look only at dose comparisons within the intervention arm. Dose will be included in these models as a non-linear effect via restricted cubic splines.

We will conduct a per-protocol analysis to assess the expected effect of adhering to the trial protocol using G-methods, which allow for adjustment for post-randomisation confounding.108

All analyses of primary and secondary outcomes will be pre-specified in detail in a Statistical Analysis Plan and signed off on by all investigators prior to data analysis.


Protocols to address particular AEs and SAEs are described in online supplemental appendix 3. We will implement REDCap alerts for AEs noted through the questionnaires. Additionally, participants will be asked if they have had any AEs at each study visit. All AEs will be assessed by a study physician. All SAEs will be reported to the DSMB. SAEs will be reported to the Research Ethics Board (REB) and Health Canada (HC) as per local regulations. For mild AEs, the patient may be reassured to continue taking the medication as per protocol. Previous studies and our clinical experience have suggested that LDN is generally well tolerated, and mild AEs will often ease with treatment continuation.40 48 49 51–53

Inspections and auditing

The trial will be subject to inspections or audits by HC, REB and the Canadian Institutes for Health Research.

Patient and public involvement

Patient partners will be included as part of the TSC.

Trial dates

This trial started recruitment in January 2024 and aims to complete follow-up by the end of 2024.


This report described the protocol for a 16-week, phase II RCT to investigate the efficacy of LDN for treating fatigue severity in patients with PCFS, an illness we have defined as ME/CFS symptoms persisting at least 3 months following SARS-CoV-2 infection.

Our study will build on prior and ongoing evaluations of LDN. In our review of studies listed on ClinicalTrials.gov, we identified one upcoming trial (NCT05946551) which also investigates LDN in PCC. However, this trial is smaller (expected n=36) and focused on feasibility outcomes. There are no trials listed that investigate LDN in ME/CFS.

A positive outcome in our trial would inspire greater confidence in LDN as a treatment for the millions of patients with PCFS symptoms and could prompt larger, multi-institutional phase III studies. Unlike other candidate PCC treatments such as Paxlovid, stellate ganglion blockade and hyperbaric oxygen,110–116 LDN is widely available, relatively inexpensive and generally safe. A negative outcome in this trial would also be a valuable contribution to the literature and would directly impact clinician decisions regarding prescribing LDN. The results of this trial may inform guidelines for PCC.

The trial has limitations. It is limited to English speakers and is based in a single province. We do not have a restriction on how long a participant may have had their symptoms since COVID-19. This may limit the treatment effect if LDN efficacy is greater earlier in the disease course. The decentralised nature of the trial also limits the number of objective outcomes that can be collected from all participants.

However, our decentralised strategy for this trial has several advantages. First, it will allow individuals who live outside Vancouver to participate, including those in communities who may not have access to off-label or investigational treatments.14 Second, it will permit the inclusion of more symptomatic individuals. Some individuals with PCFS have reported symptom exacerbation from even minimal cognitive and physical exertion,117 and remote participation may prevent flare-ups experienced from in-person visits. Third, it will encourage participation from patients who may be reluctant to attend in person given the risks of COVID-19 re-infection. Lastly, it will expedite study completion by broadening the pool of eligible applicants and reducing logistical barriers associated with in-person recruitment and enrolment.

This trial has other strengths. By using the provincial PC-ICCN and REACH BC directories, we will be able to efficiently identify and contact hundreds of potential participants by email. Our focus specifically on individuals with the ME/CFS phenotype distinguishes this trial from others for PCC and increases the likelihood that participants will have a similar underlying pathophysiology. Lastly, our trial includes multiple secondary and exploratory outcome measures that may be valuable for further hypothesis generation.

This is one of the first trials in Canada investigating a pharmacological treatment for PCC and will have a direct impact on how this illness is treated. We hope that it will also promote engagement, good faith and optimism among the PCC community—a group that has experienced stigma and has expressed frustration regarding the paucity of interventional studies for their illness.21 118–122 Furthermore, the trial has implications beyond COVID-19; we expect that the results will have applicability to ME/CFS and other post-infection fatigue syndromes, including those that could emerge from future pandemics.123


Following UBC REB guidelines, all study-related information will be stored in locked facilities at C&W, and all electronic material stored on secure network drives or servers. Participants will be allocated study identification (ID) numbers and a master file linking the study ID and personal information will be saved separately.

This post was originally published on https://bmjopen.bmj.com