Melatonin and Circadian Rhythms
Circadian rhythm disruption represents a pervasive feature of ME/CFS that intersects with nearly every other aspect of pathophysiology discussed in this chapter. The circadian system coordinates temporal organization of physiological processes including the HPA axis cortisol rhythm, immune function cycling between pro- and anti-inflammatory states, metabolic switching between anabolic and catabolic metabolism, body temperature regulation and thermoregulation, and sleep-wake cycles and alertness patterns. Disruption of circadian timing thus has cascading effects across multiple systems already compromised in ME/CFS.
1 Objective Documentation of Circadian Disruption
Cambras et al. (2018) employed rigorous actigraphy monitoring to objectively document circadian rhythm abnormalities in ME/CFS patients (Cambras et al. 2018). This case-control study of 10 women with ME/CFS and 10 matched controls revealed that daily activity levels were significantly lower in ME/CFS patients, relative amplitude of the activity rhythm (difference between peak and nadir) was reduced, indicating flattened circadian variation, and stability of the activity rhythm across days was decreased, showing less consistent day-to-day patterns. Additionally, distal skin temperature showed lower nocturnal values in winter, suggesting impaired circadian regulation of peripheral blood flow and thermoregulation.
Subsequent work by the same group (Cambras et al., 2023) demonstrated that skin temperature circadian rhythm alterations in ME/CFS are significantly associated with endothelin-1 (ET-1) levels (\(p < 0.01\)), linking thermoregulatory circadian disruption to vascular tone dysregulation (Cambras et al. 2023). ET-1 is a potent vasoconstrictor peptide; its association with temperature rhythm stability and amplitude suggests that the circadian control of vascular tone — not just central hypothalamic clock output — is disrupted in ME/CFS. Self-reported symptom severity correlated with both ET-1 levels and temperature rhythm disruption, providing a direct link between vascular biomarker, circadian physiology, and patient-experienced symptoms. These findings demonstrate that circadian disruption in ME/CFS is not merely subjective patient reports of “feeling tired at the wrong times” but rather represents measurable alterations in the fundamental 24-hour organization of physiological functions. The reduced amplitude of activity rhythms parallels the flattened cortisol rhythm discussed earlier, suggesting a coordinated loss of circadian regulation across multiple output systems.
2 Sleep Architecture Versus Circadian Timing
An important conceptual distinction must be maintained between sleep architecture abnormalities (changes in sleep stage distribution, fragmentation, sleep efficiency) and circadian timing disruption (shifts in the phase or amplitude of 24-hour rhythms). ME/CFS patients exhibit both types of abnormalities, but they reflect different underlying mechanisms and require different therapeutic approaches.
Sleep architecture studies consistently document that ME/CFS patients experience longer sleep latency (time to fall asleep), more frequent awakenings during the night, more time in bed relative to total sleep time (reduced sleep efficiency), later and more variable wake times, irregular sleep patterns across days, and the paradox of unrefreshing sleep despite adequate or even prolonged total sleep duration. Children with ME/CFS often show continuous sleep exceeding 10 hours yet wake unrefreshed, indicating profound sleep dysfunction.
The circadian timing component involves altered phase relationships between sleep-wake cycles and other circadian outputs (body temperature, hormone secretion, immune function). Some ME/CFS patients show delayed sleep phase (natural sleep-wake times shifted later, resembling “night owl” patterns), while others exhibit irregular rhythms without clear 24-hour periodicity, and some maintain normal phase relationships but with reduced amplitude of rhythms.
3 Molecular Clock Dysfunction
Falsifiability: weakly — Falsified if repeat transcriptomic analysis fails to confirm NPAS2 or other clock gene alterations in independent ME/CFS cohorts with adequate sample size
Emerging evidence suggests disruption at the molecular level of circadian clock gene expression. Genome-wide association studies (GWAS) have reported nominally significant associations with NPAS2 (neuronal PAS domain protein 2), a core clock gene. Transcriptomic analysis of ME/CFS patient samples showed 10-fold higher NPAS2 expression compared to controls and elevated expression of other circadian rhythm genes in peripheral blood mononuclear cells (PBMCs). Enrichment of CLOCK gene variants in ME/CFS patients with comorbid fibromyalgia and epigenetic changes in “circadian entrainment” pathways suggest heritable and acquired alterations in clock gene function.
The molecular clock operates as a transcriptional-translational feedback loop involving core clock genes (CLOCK, BMAL1, PER1/2/3, CRY1/2) that regulate their own expression with approximately 24-hour periodicity. These clock genes also control thousands of downstream genes involved in metabolism, immune function, and cellular processes, creating temporal coordination across organ systems. Disruption of clock gene function could therefore produce pleiotropic effects consistent with the multi-system nature of ME/CFS.
Inflammatory cytokines, particularly IL-1\(\beta\) and TNF-\(\alpha\) elevated in ME/CFS (Chapter Immune System Dysfunction), directly disrupt clock gene expression and alter circadian rhythms. This creates bidirectional interactions where immune dysfunction disturbs circadian regulation, while circadian disruption impairs proper immune function, perpetuating a self-reinforcing cycle.
4 Melatonin and Its Therapeutic Potential
Melatonin serves as both a marker and mediator of circadian rhythms, secreted by the pineal gland predominantly at night in response to darkness signals from the suprachiasmatic nucleus (SCN). Melatonin synchronizes peripheral clocks throughout the body, exerts direct antioxidant and anti-inflammatory effects, modulates immune function and cytokine production, and facilitates sleep initiation though it is not primarily a sedative.
Limited evidence suggests altered melatonin production in ME/CFS, though findings have been inconsistent, likely reflecting the heterogeneity of circadian dysfunction patterns. Some patients show reduced melatonin amplitude, others exhibit phase shifts (melatonin rise at inappropriate times), while some maintain apparently normal melatonin profiles despite subjective circadian symptoms.
Castro-Marrero et al. (2021) conducted a 16-week randomized, double-blind, placebo-controlled trial of melatonin (1 mg) plus zinc (10 mg) daily in 50 ME/CFS patients (Castro-Marrero et al. 2021). The intervention significantly reduced physical fatigue perception (p<0.05) and improved the physical component summary score compared to placebo. Urinary 6-sulfatoxymelatonin (the primary melatonin metabolite) increased significantly in the treatment group (p<0.0001), confirming adequate absorption and metabolism. Importantly, the intervention was safe and well-tolerated with no significant adverse effects.
This represents the first rigorous randomized controlled trial evidence that melatonin supplementation may provide symptomatic benefit in ME/CFS. However, several important caveats apply: the effect size was modest (improvement but not remission), the mechanism of benefit remains unclear (improved sleep, circadian resynchronization, anti-inflammatory effects, or antioxidant actions), and individual responses varied substantially (some patients benefited greatly, others not at all), and long-term efficacy and optimal dosing require further study.
While melatonin supplementation showed benefits in the Castro-Marrero trial, clinicians and patients should recognize important limitations. Melatonin primarily aids sleep initiation but does not address sleep maintenance (frequent awakenings), may temporarily improve symptoms without addressing underlying circadian dysfunction, risks masking underlying sleep disorders requiring different treatments (sleep apnea, restless legs syndrome), and exhibits substantial individual variation in absorption, metabolism, and response. Additionally, optimal timing of melatonin administration depends on the specific circadian phase abnormality (delayed, advanced, irregular), which typically requires formal assessment.
5 Circadian Disruption as an Integrative Mechanism
The circadian rhythm abnormalities documented in ME/CFS should not be viewed as isolated sleep problems but rather as disruption of a master regulatory system that normally coordinates multi-system physiology. Loss of circadian organization contributes to HPA axis dysfunction (flattened cortisol rhythm discussed earlier in this chapter), immune dysfunction (loss of circadian immune regulation), metabolic dysfunction (disrupted glucose homeostasis and lipid metabolism), autonomic dysfunction (altered cardiovascular circadian patterns discussed in Chapter Cardiovascular Dysfunction), and thermoregulatory dysfunction (impaired circadian temperature variation).
Certainty: 0.40. Cambras et al. demonstrated that skin temperature circadian rhythms are disrupted in ME/CFS and correlate with endothelin-1 (ET-1) levels (Cambras et al. 2023). This suggests that circadian thermoregulation relies on two coupled oscillators: a central suprachiasmatic nucleus (SCN) oscillator driving melatonin and core body temperature rhythms, and a peripheral vascular oscillator (driven by ET-1, local endothelial clocks, and autonomic tone) controlling skin blood flow and heat dissipation. In ME/CFS, the central SCN oscillator may be relatively preserved (melatonin rhythm often normal), while the peripheral oscillator is dysregulated via ET-1 elevation and autonomic dysfunction. The resulting internal phase misalignment — the brain signaling “night” while the periphery signals metabolic activity — produces fatigue via constant error correction in hypothalamic integration centers. This model explains why melatonin alone often fails to normalize thermoregulation: it entrains the central clock but does not address the peripheral oscillator driving skin temperature dysregulation. Falsifiable: ME/CFS patients will show >4-hour phase shift between melatonin rhythm (saliva DLMO) and peripheral skin temperature rhythm minimum; melatonin administration will not re-entrain skin temperature, but ET-1 modulation (bosentan) will re-synchronize both oscillators.
Certainty: 0.30. Interventions targeting circadian resynchronization — whether through melatonin, light therapy, behavioural scheduling, or other chronotherapeutic approaches — might provide broader benefits than expected from improving sleep alone. By restoring temporal coordination across multiple systems, circadian interventions could theoretically address multiple aspects of ME/CFS pathophysiology simultaneously.
Testable prediction: A chronotherapy trial (timed light + melatonin + activity scheduling) in ME/CFS patients with documented circadian disruption would improve at least two non-sleep outcome domains (e.g. immune markers, cortisol rhythm amplitude, HRV) beyond what sleep improvement alone would predict. Falsified if chronotherapy normalises sleep metrics without measurable improvement in non-sleep circadian outputs.
Limitations: The Castro-Marrero 2021 melatonin trial (Castro-Marrero et al. 2021) is the only RCT supporting circadian intervention in ME/CFS, combined melatonin with zinc (individual contributions inseparable), and had a modest sample size. No study has tested multi-modal chronotherapy in ME/CFS.
Certainty: 0.30. (Raw certainty 0.55–0.75 in animal studies, from animal model → discounted to 0.28–0.38.) Preclinical work has mapped a molecular substrate by which the metabolic hormone leptin can influence the circadian system’s response to light. Leptin-receptor (LepR) neurons in the dorsomedial hypothalamus (DMH) regulate diurnal feeding and metabolism (Faber et al. 2021), and DMH LepR neurons project to the suprachiasmatic nucleus (SCN), where they can shift clock phase (Tang et al. 2023); the clock reciprocally gates leptin signalling (Osorio-Mendoza et al. 2025) — a bidirectional coupling. Building on this, an encapsulated-cell therapy delivering constitutively-produced leptin (engineered retinal pigment epithelium cells in subcutaneous alginate capsules) accelerated re-entrainment to both phase advances and phase delays in mice and diurnal cynomolgus macaques, reducing entrainment time by roughly a day in the primates without impairing sleep (increased NREM slow-wave energy) (Fleury et al. 2026). (Translation gap: animal → human. Not yet validated in ME/CFS patients.)
This suggests the principle that a metabolic hormone such as leptin, not only light or melatonin, can modulate how quickly the circadian system re-aligns to an external schedule — a potentially relevant lever for the circadian instability and unrefreshing sleep documented in ME/CFS. Two cautions temper this: the entrainment-acceleration effect rests on a single unreplicated animal intervention whose mechanism the authors themselves flag as unclear, and the Fleury platform delivers constitutive (around-the-clock, non-phasic) leptin — a lever that acts on overall level, not on restoring the shape or phase of a rhythmic signal. Falsifiable prediction: in an animal model, selective ablation of DMH LepR neurons, or of SCN LepR innervation, abolishes the entrainment-accelerating effect of systemic leptin elevation; and in humans, a timed metabolic challenge (e.g. a standardised meal or glucose load) timed against a light pulse measurably alters phase-shift magnitude.
Limitations: All supporting evidence is preclinical (mouse + nonhuman primate); the entrainment-speed endpoint does not translate directly to ME/CFS symptom relief (see Williams 2002 — melatonin and phototherapy produced no symptomatic improvement in CFS circadian disorder (Williams et al. 2002)); and the Fleury et al. mechanism of action on entrainment is explicitly acknowledged by the authors as unclear. (Evidence source: animal studies — Inference target: ME/CFS circadian system. Link is indirect.)
Consequence: If the coupling holds in humans, it would expand the toolbox for helping ME/CFS patients re-stabilise sleep-wake timing beyond light and melatonin alone — but this is an unreplicated preclinical lead, and whether metabolic signals can do so in the ME/CFS circadian system, let alone through a constitutive (non-phasic) lever, is entirely untested.
The preclinical entrainment result (Fleury et al. 2026) immediately raises the question of whether a metabolic/leptin lever could aid the circadian and sleep disruption of ME/CFS. The ME/CFS-specific data do not support assuming this direction of intervention, but neither do they settle it.
- Leptin is not low in ME/CFS, which weakens the rationale for raising it:: A cross-sectional cohort found higher circulating leptin in ME/CFS than controls (n = 31) (Domingo et al. 2024), and the only completed controlled measurement of plasma leptin in CFS found no significant baseline difference from matched controls — though its CFS mean was numerically ~35% higher (underpowered rather than a true null) (Cleare, O’Keane, and Miell 2001). On an amount axis, there is no evidence of a deficit to correct. It remains possible, however, that ME/CFS involves central leptin resistance in which more ligand is needed to achieve a normal signal (analogous to insulin resistance) — in which case raising levels could still be therapeutic — so the direction is not settled by the amount data alone.
- Light exposure is the strongest documented circadian correlate, not a proven “dominant” lever:: Actigraphy-based light exposure patterns are associated with multidimensional health outcomes in ME/CFS (Cambras et al. 2026) — an associational finding, not proof that light is the most modifiable or effective input. This asymmetry matters: the same caution used to dismiss the metabolic lever (no intervention evidence in ME/CFS) applies equally to light, whose own intervention RCT (phototherapy) showed no symptomatic benefit (Williams et al. 2002).
- Direct circadian interventions have shown no symptom benefit, cautioning that entrainment speed may not translate:: Melatonin and phototherapy produced no symptomatic improvement in CFS circadian disorders (Williams et al. 2002), so even a lever that speeds entrainment may not improve the sleep quality and fatigue complaints that define ME/CFS.
Consequence: The honest conclusion is that the metabolic-circadian coupling is worth studying in ME/CFS as a research and diagnostic direction, but there is currently no evidence that raising leptin — a hormone that is not low in this population — would help, and no circadian intervention (metabolic, light, or melatonin) has shown symptom benefit; the open question is whether the defect is a matter of amount, timing, or resistance.
The two direct human measurements of circulating leptin in ME/CFS are read as contradictory — no significant baseline difference from matched controls (Cleare, O’Keane, and Miell 2001) versus higher circulating leptin than controls (Domingo et al. 2024) — but both are single-timepoint snapshots of a hormone that is inherently circadian, and the two point numerically the same way (higher in CFS; Cleare is underpowered rather than a true null). Because a metabolic hormone such as leptin couples to the circadian system bidirectionally (Osorio-Mendoza et al. 2025), an open possibility is that the relevant abnormality is in the circadian phase or amplitude of the rhythm (flattened, phase-shifted, or desynchronized across tissues) — which a single spot sample can only partially detect, depending on shift magnitude and sampling time. (Origin: brainstorm.) A competing reading is that any mean elevation reflects central leptin resistance rather than a timing defect, which this timing hypothesis does not yet engage.
Testable prediction (Certainty: 0.15): A 24-hour serial plasma-leptin profile (e.g. hourly or 2-hourly sampling with DLMO or dim-light phase reference) in ME/CFS versus matched controls would reveal a phase-shifted, flattened, or amplitude-reduced rhythm even where single-timepoint means are indistinguishable — and would disambiguate the Cleare and Domingo readings as timing rather than amount. The hypothesis is falsified if serial profiling shows a normal circadian rhythm whose phase and amplitude are within a pre-specified effect-size margin (e.g. amplitude ratio ≥1.5-fold or phase offset ≥2 hours versus controls) — an explicit threshold, since a small underpowered cohort would otherwise default to “indistinguishable.”
Limitations: The reconciliatory hypothesis is not yet evidenced — no serial leptin profile has been published in ME/CFS (the only systematic review on the question remains a protocol without results (Musker et al. 2021)); both existing cohorts are small and single-timepoint, neither stratifies by severity, and neither was designed to resolve circadian timing; and a timing-only abnormality, if confirmed, would still not establish that correcting it is therapeutic, only that it is measurable. (Severity coverage: unknown — neither cohort stratifies by severity.)
Consequence: If confirmed, this would mean the field has been asking the wrong question about ME/CFS leptin — the relevant variable might be not how much but when — and it would provide a concrete, feasible measurement protocol to test it; but it is a research hypothesis with no current clinical application.
A coherent reading of the integrated evidence is that the metabolic hormone leptin can influence how quickly the circadian system re-aligns to an external schedule — the animal work identifies a molecular substrate (DMH LepR → SCN) and a delivery platform tested in mice and macaques (Metabolic-Circadian Entrainment Coupling: Leptin as a Modulator of Phase Re-alignment) — but applying this to ME/CFS is not straightforward, and the evidence is one research programme rather than independent confirmation. On an amount axis, there is no case for raising a hormone that is not low in ME/CFS (Could Raising Leptin Ever Help ME/CFS Circadian Disruption—or Would It Be the Wrong Direction?). On a timing/amplitude axis, the productive open question is whether ME/CFS leptin is abnormal in its rhythm rather than its mean, which can only be answered by a serial 24-hour profile and which would reconcile the two single-timepoint human measurements (Is ME/CFS Leptin Altered in Timing Rather Than Amount—the Cleare-vs-Domingo Reconciliation?). Critically, these two axes do not collapse into one intervention: the Fleury platform delivers constitutive (non-phasic) leptin — a lever on level, not on restoring rhythm shape or phase — so it is ill-suited to a timing defect even if one is found. What the evidence supports is that metabolic-circadian coupling is a genuine, research-relevant phenomenon for ME/CFS’s circadian instability; what remains open is whether any metabolic intervention would help, and on which axis.
Consequence: Clinicians should not expect a “leptin factory” to treat ME/CFS circadian disruption — it is a non-phasic lever, the hormone is not low in the population, and no circadian intervention has shown symptom benefit — and the productive research question is whether a serial profile would reveal a timing-not-amount abnormality worth measuring, not treating.