Low-Dose Naltrexone in Long COVID: Navigating Paradoxical Responses, Dose Sensitivity, and the Case for Ultra-Low-Dose Initiation

A clinical commentary on individualized titration, interpreting worsening symptoms, and the emerging science of LDN in post-acute sequelae of SARS-CoV-2

Yoon Hang Kim MD MPH  |  Board Certified in Preventive Medicine | Integrative and Functional Medicine Physician| www.directintegrativecare.com

Among the most challenging clinical problems I encounter is the long COVID patient who presents with profound, pervasive fatigue — a fatigue that does not yield to rest, that accumulates with minimal exertion, and that has stripped away not just productivity but identity. Many of these patients arrive having already tried multiple interventions, and more than a few arrive specifically to discuss low-dose naltrexone. They have read, researched, and often participated in patient communities where LDN is discussed with remarkable sophistication. And yet, the conversations that matter most are not the ones about mechanism or evidence — they are the ones about what happens when LDN does not behave the way a patient expected, or when worsening symptoms after a dose change leave them uncertain about what they are actually experiencing.

This article addresses those conversations directly. It explores the pharmacokinetics of LDN in the context of long COVID, the challenge of interpreting symptom worsening on or after the medication, the rationale for ultra-low-dose initiation in a population that is far more sensitive than conventional dosing protocols anticipate, the mechanistic basis for LDN's neuroimmune effects, and the persistent practical barriers around compounding accuracy. It is written from the perspective of a clinician who has prescribed LDN for years and who approaches it not as a panacea but as a carefully individualized tool in a complex therapeutic landscape.

The Scope of Long COVID as a Clinical Problem

Long COVID — formally designated as post-acute sequelae of SARS-CoV-2 infection (PASC) — is estimated to affect a substantial proportion of individuals who contract COVID-19, with some analyses suggesting that at least 10% develop persistent symptoms meeting diagnostic criteria. The clinical presentation is strikingly heterogeneous, encompassing post-exertional malaise, unrefreshing sleep, cognitive impairment (sometimes called brain fog), dysautonomia, cardiopulmonary symptoms, gastrointestinal dysfunction, and a fatigue that shares many features with myalgic encephalomyelitis and chronic fatigue syndrome (ME/CFS). Indeed, the symptomatic and pathophysiological overlap between these conditions is substantial enough that clinicians familiar with ME/CFS have effectively served as the advance scouts for long COVID management — adapting frameworks that were already established, if underutilized, in conventional medicine.

What makes long COVID particularly difficult to treat is not merely the breadth of its symptom profile but the severity of functional impairment it produces. Even patients who eventually achieve meaningful improvement often describe a ceiling — a level of function at perhaps 60 to 70% of their prior baseline — that persists despite sustained therapeutic effort. Many cannot return to full-time work. Others must carefully pace every activity to avoid post-exertional crashes that may set them back days or weeks. This is not a population in which therapeutic expectations should be set optimistically, and part of any honest conversation about LDN must begin with an acknowledgment that even a beneficial response may be partial and hard-won.

What Low-Dose Naltrexone Is — and Is Not

Naltrexone is an FDA-approved opioid receptor antagonist indicated at doses of 50 to 100 mg daily for the management of opioid use disorder and alcohol use disorder. At this dosage range, its primary action is sustained, near-complete blockade of mu-opioid receptors, effectively preventing opioid-mediated euphoria and thereby reducing reinforcement. Low-dose naltrexone (LDN) refers to a pharmacologically distinct regime typically ranging from 1 to 5 mg daily — roughly one-tenth to one-twentieth of the conventional therapeutic dose — in which the drug's behavioral profile shifts substantially. Rather than sustained receptor antagonism, LDN produces brief, intermittent receptor blockade that results in a compensatory upregulation of endogenous opioid production. This "endorphin rebound" effect, combined with mechanisms that appear entirely independent of opioid receptor signaling, is the conceptual foundation on which LDN's anti-inflammatory and neuroimmune effects rest.

It is important to recognize that LDN does not directly modulate serotonin or dopamine in any established pharmacological sense. While there are indirect downstream effects on the limbic system mediated through opioidergic pathways — and while the endogenous opioid system interacts broadly with reward and mood circuitry — LDN should not be conceptualized as an antidepressant or as a monoaminergic agent. Its primary relevant effects are neuroimmune: modulation of microglial activation, reduction of pro-inflammatory cytokine cascades, and restoration of ion channel function in immune cells. This distinction matters clinically because some patients arrive expecting LDN to address mood, motivation, or hedonic tone directly, when the more accurate frame is that any improvement in those domains would be downstream of reduced neuroinflammation and improved immune regulation.

Pharmacokinetics: Why Timing and Dose Matter More Than Most Clinicians Appreciate

Understanding the clinical behavior of LDN requires a working grasp of naltrexone pharmacokinetics, and specifically the distinction between the parent compound and its primary active metabolite. After oral administration, naltrexone undergoes extensive first-pass hepatic metabolism, with bioavailability ranging from approximately 5 to 60% depending on individual hepatic enzyme activity. The parent compound has a fast elimination half-life of approximately 4 hours, though published ranges extend from 3.9 to 10.3 hours across studies, reflecting considerable interindividual variation. The major active metabolite, 6-beta-naltrexol, has a mean elimination half-life of approximately 13 hours and is present in plasma at concentrations ten to thirty times higher than the parent drug following oral administration. Because 6-beta-naltrexol is itself a competitive opioid receptor antagonist — albeit with reduced CNS penetration relative to the parent compound — its prolonged half-life means that receptor effects extend meaningfully beyond what the naltrexone half-life alone would predict.

The practical clinical implication of this pharmacokinetic profile is that the drug's biological activity is effectively cleared within approximately two to four days following cessation of a low-dose regimen in most patients — generally within the range of seven to thirty days for those with slower metabolism, older age, or reduced renal clearance. This timeline is essential for interpreting what patients report during dose changes or after discontinuation.

Interpreting Symptom Worsening: One of the Most Consequential Clinical Decisions

One of the most clinically nuanced and frequently underappreciated challenges in prescribing LDN for long COVID is the proper interpretation of symptom worsening. Consider a patient who begins LDN at 1.5 mg and reports a meaningful decline in functional capacity over the ensuing weeks — dropping, by their own assessment, from approximately 40% of baseline function to 20%. They stop the medication, and their function deteriorates further to roughly 10%. How should the clinician — and the patient — understand this trajectory?

Worsening symptoms during or after LDN should not be reflexively attributed to the drug without systematically considering the alternatives: dose sensitivity, natural disease fluctuation, and the retrospective unmasking of benefit that was only recognized once it was gone.

There are at least three distinct possibilities that must be carefully considered. The first is true pharmacological toxicity — that LDN is itself producing harm, either through dose-dependent side effects (sleep disturbance, vivid dreaming, nausea, and paradoxical worsening are all described) or through an idiosyncratic mechanism in this particular patient. If this is the case, symptoms should improve after stopping, typically within seven to thirty days given the pharmacokinetics described above. The second possibility is that the symptom worsening reflects natural disease fluctuation — something endemic to long COVID regardless of any intervention. Long COVID does not follow a steady trajectory; it is characterized by fluctuations, crashes, and partial recoveries that can easily be mistakenly correlated with medication changes. The third, and perhaps most counterintuitive possibility, is that LDN was providing subtle benefit that was not consciously recognized while the drug was being taken, and that the further worsening after discontinuation represents not drug withdrawal in the classical pharmacological sense but rather the return of the underlying pathophysiology that was being partially ameliorated. This phenomenon — retrospective recognition of benefit only after it is lost — is described by patients across many chronic conditions and requires explicit discussion before any trial of LDN begins.

Distinguishing among these possibilities is genuinely difficult in individual patients without a controlled rechallenge. The practical clinical approach is to ensure that any trial is conducted at the lowest feasible starting dose, over a sufficiently long washout period if discontinuation becomes necessary, and with a clear pre-specified plan for how worsening will be attributed and interpreted before the patient begins the medication — rather than in the emotionally charged context of a decline.

The Case for Ultra-Low-Dose Initiation in Long COVID

Standard LDN dosing protocols, which typically begin at 1 to 1.5 mg and titrate upward over weeks to a target of 3 to 4.5 mg daily, were developed largely in the context of fibromyalgia, multiple sclerosis, Crohn's disease, and ME/CFS — conditions in which patients, while sensitive, generally tolerate incremental dose escalation reasonably well. The long COVID population appears to be meaningfully different. Clinical experience — my own and that reported in emerging literature — suggests that a significant proportion of long COVID patients cannot tolerate even the standard starting doses without pronounced side effects or functional worsening that, regardless of its true cause, undermines adherence and therapeutic confidence.

The appropriate response to this reality is not to abandon LDN but to adapt the dosing strategy with far greater individualization than conventional protocols suggest. For a subset of long COVID patients — particularly those with severe baseline fatigue, evidence of significant autonomic dysregulation, or a prior history of pronounced medication sensitivity — initiating LDN at doses as low as 0.001 mg (one microgram) may be warranted. This represents not simply a cautious titration but a fundamentally different pharmacological approach, sometimes referred to as ultra-low-dose naltrexone (ULDN). At microgram-level dosing, the opioid receptor dynamics differ further still from standard LDN, and the clinical rationale shifts toward receptor sensitization and tolerance-avoidance rather than the endorphin rebound paradigm that governs conventional LDN. From that nanogram or microgram foundation, dose escalation can be conducted in extremely small increments — doubling perhaps every two to four weeks — allowing the patient's nervous system to accommodate gradually rather than encountering an abrupt pharmacological intervention in an already dysregulated neuroimmune milieu.

Mechanisms of Action: Why LDN Is Biologically Plausible for Long COVID

The mechanistic rationale for LDN in long COVID has grown substantially more sophisticated in recent years, and is now grounded in at least three distinct and convergent biological pathways that map directly onto the known pathophysiology of the condition.

Microglial Modulation and TLR4 Antagonism

At doses below 5 mg, naltrexone functions as a glial cell modulator — a pharmacological classification that is distinct from its conventional identity as an opioid antagonist. Specifically, LDN exerts antagonist effects at Toll-like receptor 4 (TLR4), a non-opioid receptor expressed at high density on microglial cells, the resident immune cells of the central nervous system. TLR4 activation is a key upstream trigger for the production of pro-inflammatory cytokines including IL-6, TNF-alpha, and IL-1 beta. In chronic neuroinflammatory states — including those associated with long COVID, ME/CFS, and fibromyalgia — microglia are maintained in a pathologically activated state that sustains neuroinflammation independently of ongoing infectious stimulus. By antagonizing TLR4 signaling, LDN attenuates this downstream cytokine cascade and effectively shifts microglia from a neurotoxic, pro-inflammatory state toward a more quiescent, neuroprotective phenotype. This is the mechanism most directly relevant to the brain fog, cognitive impairment, and central sensitization that characterize long COVID's neurological manifestations.

The Endorphin Rebound Pathway

The second major mechanism operates through the opioidergic system itself. When LDN is administered, it produces a brief, partial blockade of mu-opioid receptors. The brain interprets this as a relative deficit in endogenous opioid tone and responds by upregulating the production of endorphins — principally beta-endorphin — in a compensatory fashion. This transient receptor blockade, followed by an endorphin surge, has downstream consequences for immune regulation: endogenous opioids modulate natural killer cell activity, T-regulatory cell function, and cytokine balance in ways that may help restore the immune homeostasis disrupted by SARS-CoV-2. Notably, decreased peripheral blood concentrations of beta-endorphin have been observed in ME/CFS and fibromyalgia, conditions that share overlapping pathophysiology with long COVID, suggesting that endorphin depletion may be a feature common to post-viral neuroimmune syndromes. LDN's endorphin rebound effect offers a mechanism directly addressing this deficit.

TRPM3 Ion Channel Restoration

Perhaps the most mechanistically specific finding to emerge in recent years concerns the Transient Receptor Potential Melastatin 3 (TRPM3) ion channel — a transmembrane calcium-permeable channel expressed on natural killer cells and involved in cellular signaling and immune surveillance. Research has demonstrated that TRPM3 function is significantly impaired in both ME/CFS and long COVID patients, with reduced calcium influx disrupting the normal operation of NK cells and impairing their capacity for pathogen recognition and clearance. A study published in 2025 in Frontiers in Molecular Biosciences investigated TRPM3 function in NK cells from long COVID patients treated with LDN at doses of 3 to 4.5 mg daily and found that LDN restored TRPM3 ion channel activity to levels comparable to those observed in healthy controls. This restoration of calcium influx provides a concrete molecular explanation for one pathway through which LDN may alleviate fatigue and immune dysregulation in post-viral syndromes, and it positions TRPM3 as a meaningful therapeutic target in this population.

Cortisol and the HPA Axis: What the Evidence Actually Shows

One area of genuine clinical uncertainty concerns LDN's effects on cortisol and the hypothalamic-pituitary-adrenal (HPA) axis. Patients with long COVID frequently present with symptoms suggestive of HPA dysregulation — including morning fatigue, impaired stress response, and abnormal cortisol rhythms. It is therefore clinically meaningful to understand whether LDN itself might perturb cortisol levels, and if so, in what direction.

The available evidence, while primarily derived from standard-dose naltrexone studies rather than LDN-specific research, indicates that acute administration of naltrexone at conventional doses (50 mg) does disinhibit the HPA axis, producing transient increases in ACTH and cortisol. This occurs because endogenous opioids normally exert tonic inhibitory control over corticotropin-releasing factor neurons in the hypothalamus; blocking opioid receptors removes this inhibition and transiently activates the stress axis. However, important qualifications apply. These effects are demonstrably dose-dependent, have been studied almost exclusively at 50 mg rather than at the 1 to 5 mg LDN range, show marked sex differences (women exhibit substantially larger cortisol responses to opioid blockade than men), and are subject to considerable interindividual variability that likely reflects baseline differences in central opioid tone. Studies have also yielded conflicting results depending on the population studied — alcohol-dependent patients, for instance, may show blunted or absent cortisol responses to naltrexone challenge. For patients with suspected endorphin depletion — a plausible feature of long COVID — the cortisol response to LDN may differ further still from what has been documented in healthier populations. In short, the assertion that LDN meaningfully elevates cortisol in long COVID patients should be regarded as biologically plausible but not established by existing evidence.

The Emerging Clinical Evidence Base

The evidence base for LDN in long COVID remains in an early phase, but it is accumulating with meaningful momentum. A 2023 retrospective cohort study by Bonilla and colleagues at Stanford University evaluated 59 patients with PASC who received LDN off-label and found improvements in fatigue, brain fog, post-exertional malaise, sleep quality, and functional status — among the most clinically relevant outcomes in this population. A 2024 retrospective study published in Clinical Therapeutics by Tamariz and colleagues evaluated 108 patients in a post-COVID clinic and found that LDN was associated with a relative hazard of clinical improvement of 5.04 (95% CI 1.22 to 20.77, p = 0.02) compared to physical therapy alone — a striking signal from a relatively modest sample. A 2024 pilot study published in Brain, Behavior, and Immunity — Health evaluated LDN combined with NAD+ supplementation in 36 patients with persistent post-COVID fatigue and demonstrated significant improvement in SF-36 quality of life scores after 12 weeks of treatment.

A 2025 systematic review and meta-analysis (preprint, medRxiv) synthesized data from four observational pre-post studies totaling 155 patients and reported a pooled moderate effect size for fatigue reduction (Hedges' g = -0.74; 95% CI -1.11 to -0.37; p < 0.001) — comparable to or exceeding effect sizes observed for pharmacological agents used in ME/CFS and fibromyalgia. No randomized controlled trials have yet been completed in this space, though multiple trials are registered and ongoing, including a well-designed phase II double-blind, placebo-controlled trial at the University of British Columbia evaluating LDN specifically for post-COVID fatigue syndrome. The effect sizes reported across observational studies are encouraging, but the absence of controlled trial data means that placebo response, regression to the mean, and selection bias cannot yet be ruled out as contributors to the observed benefit.

Compounding Accuracy: A Practical Barrier That Demands Attention

There is one practical challenge in delivering ultra-low-dose LDN that receives insufficient attention in both clinical discussions and patient communities: compounding accuracy. LDN is not commercially available as a finished pharmaceutical product in doses below 50 mg, which means that all LDN prescriptions must be filled by compounding pharmacies. For doses in the standard LDN range of 1 to 4.5 mg, reputable compounding pharmacies can generally produce capsules of acceptable potency. However, for microgram-level dosing — where one is working in the range of 0.001 to 0.1 mg — the technical demands on the compounding process escalate considerably. Small errors in geometric dilution of the active ingredient can produce doses that deviate substantially from the prescribed amount, and very few compounding pharmacies routinely perform potency assay validation for preparations in this concentration range.

The practical implication is that when prescribing ultra-low-dose LDN, clinicians must be deliberate in their selection of compounding pharmacy. Pharmacies that specialize in LDN, that can document their dilution methodology, and that have access to analytical validation are preferable to general compounding operations that may lack the technical infrastructure to produce accurate microgram-level preparations. This is not a theoretical concern: clinical variability in LDN response may in some cases reflect not true interindividual pharmacological differences but inconsistencies in what the patient is actually receiving.

Clinical Takeaway: Individualized, Patient-Centered Caution

Long COVID represents one of the defining clinical challenges of this era — a condition that is simultaneously debilitating, poorly understood, and inadequately served by conventional therapeutic approaches. LDN occupies an important and growing place in its management, not because it reliably produces dramatic recoveries, but because it is mechanistically rational, has an established safety profile, and continues to generate encouraging clinical signals in a population with few options.

But the medication demands respect — specifically, respect for the extraordinary sensitivity that long COVID patients can exhibit to pharmacological stimuli of all kinds. Standard LDN protocols, designed for populations that are already quite sensitive by ordinary clinical standards, may be far too aggressive for a meaningful subset of long COVID patients. Ultra-low-dose initiation in the microgram range, combined with extremely gradual titration over months rather than weeks, is not simply a cautious variant of standard practice — in some patients, it may be the only approach that permits a therapeutic trial at all without triggering a cascade of worsening symptoms that derails the entire effort.

Symptom worsening during or after LDN should always be interpreted in the full context of the disease's natural variability, the pharmacokinetics of clearance, and the counterintuitive possibility that a perceived worsening after stopping LDN reflects the loss of a benefit that was too subtle to register consciously while the drug was active. These nuances require time, explanation, and genuine partnership between clinician and patient — the kind of relationship that direct integrative care is specifically designed to support.

For patients considering LDN as part of their long COVID management, the conversation should begin not with mechanism or miracle stories, but with an honest discussion of what realistic improvement looks like in this condition, how the trial will be structured, and how both benefit and harm will be recognized, measured, and acted upon. With that foundation in place, LDN remains one of the most promising tools available — imperfect, individualized, and worth exploring with appropriate care.

References

1. Bonilla H, Tian L, Marconi VC, Shafer R, McComsey GA, Miglis M, et al. Low-dose naltrexone use for the management of post-acute sequelae of COVID-19. Int Immunopharmacol. 2023;124:110966. doi:10.1016/j.intimp.2023.110966. PMID 37703724
2. Tamariz L, Bast E, Klimas N, Palacio A. Low-dose naltrexone improves post-COVID-19 condition symptoms. Clin Ther. 2024;46(3):e101–e106. doi:10.1016/j.clinthera.2023.12.009. PMID 38267326
3. Isman A, Nyquist A, Strecker B, Harinath G, Lee V, Zhang X, Zalzala S. Low-dose naltrexone and NAD+ for the treatment of patients with persistent fatigue symptoms after COVID-19. Brain Behav Immun Health. 2024;36:100733. doi:10.1016/j.bbih.2024.100733. PMID 38356686
4. Naik H, Cooke E, Boulter T, Dyer R, Bone JN, Tsai M, et al. Low-dose naltrexone for post-COVID fatigue syndrome: a study protocol for a double-blind, randomised trial in British Columbia. BMJ Open. 2024;14(5):e085272. doi:10.1136/bmjopen-2024-085272. PMID 38740499
5. Sasso EM, Muraki K, Eaton-Fitch N, et al. Low-dose naltrexone restored TRPM3 ion channel function in natural killer cells from long COVID patients. Front Mol Biosci. 2025;12:1582967. doi:10.3389/fmolb.2025.1582967. PMID 40458265
6. Tamariz L, et al. Effect of low dose naltrexone for long covid: a systematic review [preprint]. medRxiv. 2025. doi:10.1101/2025.09.09.25335451
7. Younger J, Parkitny L, McLain D. The use of low-dose naltrexone (LDN) as a novel anti-inflammatory treatment for chronic pain. Clin Rheumatol. 2014;33(4):451–459. doi:10.1007/s10067-014-2517-2. PMC3962576
8. Löhn M, Wirth KJ. Potential pathophysiological role of the ion channel TRPM3 in ME/CFS and the therapeutic effect of low-dose naltrexone. J Transl Med. 2024;22:630. doi:10.1186/s12967-024-05412-3. PMID 38972971
9. Crabtree BL. Review of naltrexone, a long-acting opiate antagonist. Clin Pharm. 1984;3(3):273–280. PMID 6329589
10. StatPearls [Internet]. Naltrexone. Treasure Island (FL): StatPearls Publishing; 2023. NCBI Bookshelf NBK534811
11. King AC, Schluger J, Gunduz M, et al. Hypothalamic-pituitary-adrenocortical axis response and biotransformation of oral naltrexone: preliminary examination of relationship to family history of alcoholism. Neuropsychopharmacology. 2002;26(6):778–788. PMID 11988154
12. Lovallo WR, King AC, Farag NH, et al. Naltrexone effects on cortisol secretion in women and men in relation to a family history of alcoholism. Psychoneuroendocrinology. 2012;37(12):1922–1928. PMC3449011
13. Panerai AE, Vecchiet J, Panzeri P, et al. Peripheral blood mononuclear cell beta-endorphin concentration is decreased in chronic fatigue syndrome and fibromyalgia but not in depression: preliminary report. Clin J Pain. 2002;18(4):270–273. PMID 12131067