Viral Reactivation and Persistence
Many ME/CFS cases follow acute infections, and evidence suggests ongoing viral activity in some patients .
1 Herpesviruses
Human herpesviruses establish lifelong latent infections with potential for reactivation.
1.1 Epstein-Barr Virus (EBV)
EBV infects B cells and establishes latency (Munir et al. 2025):
- Acute infection: Infectious mononucleosis is a common ME/CFS trigger (Bu et al. 2024)
- Reactivation markers: Elevated early antigen (EA) antibodies, viral load
- Prevalence: 10–20% of ME/CFS patients show evidence of reactivation
- Mechanism: May drive chronic B cell activation and autoantibody production (Walitt et al. 2024)
EBV-Infected B Cells and CNS Demyelination Recent research has demonstrated a direct mechanism by which EBV-infected B cells can cause neurological damage . Autoreactive B cells identified in healthy human blood can cross the blood–brain barrier following viral infection of the cerebrum. When these B cells express EBV Latent Membrane Protein 1 (LMP1), they can infiltrate the brain and induce demyelinating lesions through direct myelin antigen capture followed by complement activation and microglial activation. While this research focused on multiple sclerosis pathogenesis, the mechanism has potential relevance for ME/CFS given the documented role of EBV as a disease trigger, the neuroinflammation observed in ME/CFS patients, and the overlap between ME/CFS and MS symptomatology. This finding provides a concrete pathway by which post-infectious immune dysregulation could lead to CNS involvement.
EBV Infection During Adolescent Immune Development
Certainty: 0.50. EBV infection during adolescence may create unique risk for persistent autoantibody-mediated ME/CFS due to coincidence of viral B cell infection with pubertal immune maturation. The certainty level reflects: (1) strong epidemiological association between EBV-triggered mononucleosis and ME/CFS onset, particularly in adolescents; (2) well-characterized immune maturation during puberty; (3) documented mechanisms for EBV-driven autoimmunity; (4) however, EBV is ubiquitous and most infected adolescents do not develop ME/CFS, suggesting additional required factors; (5) the specific contribution of infection timing versus other variables (genetic susceptibility, viral strain) remains uncertain.
We hypothesize that EBV infection during adolescence creates a unique risk for persistent autoantibody-mediated ME/CFS, because viral B cell infection coincides with pubertal immune maturation when tolerance mechanisms are being reorganized.
Epidemiological context: Infectious mononucleosis (primary EBV infection) is a common ME/CFS trigger, particularly in adolescents. While young children typically experience asymptomatic primary EBV infection, delayed first exposure in adolescence produces symptomatic mononucleosis in 35–50% of cases (Munir et al. 2025). This age-dependent presentation reflects developmental differences in immune response. When infection occurs during adolescence or young adulthood, symptoms can be more severe than in younger children, and the infection is associated with increased risk for subsequent autoimmune disease development (Bu et al. 2024).
Immunological timing hypothesis: EBV preferentially infects B cells, establishing lifelong latency. During adolescence, the immune system undergoes substantial reorganization: thymic output is declining, peripheral tolerance mechanisms are maturing, and the B cell repertoire is being shaped. EBV infection during this critical window may:
- Infect B cells during active repertoire selection, potentially immortalizing autoreactive clones that would otherwise be deleted
- Disrupt tolerance checkpoint establishment, allowing autoreactive B cells to persist
- Drive aberrant germinal center reactions producing GPCR autoantibodies (Section autoantibodies)
- Create long-lived plasma cells secreting autoantibodies that persist for decades
Age-dependent outcome predictions:
- Young children (\(<\) 10 years): Immune system still highly plastic; ongoing development may clear aberrant B cell clones through mechanisms described in Hypothesis Immune Memory Pruning in Development. Higher recovery probability.
- Adolescents (10–18 years): Infection at the edge of immune maturation; some patients clear aberrant clones, others do not. Variable outcomes, overall high recovery rates.
- Young adults (18–25 years): Tolerance mechanisms largely established; aberrant B cell populations persist indefinitely. Lower recovery probability.
- Adults (\(>\) 25 years): No developmental clearance mechanism; autoantibody-producing cells become permanent. Recovery rare without intervention.
Treatment implications: If this hypothesis is correct, B cell depletion therapy (rituximab) might be particularly effective in adolescents and young adults with recent EBV-triggered ME/CFS, before long-lived plasma cells establish permanent autoantibody production. The timing of intervention relative to disease onset may be critical—early B cell depletion could prevent establishment of pathogenic plasma cell populations.
Research directions:
- Compare GPCR autoantibody titers by age at ME/CFS onset and EBV status
- Track autoantibody trajectories in EBV-triggered versus non-EBV-triggered cases
- Assess whether EBV-triggered cases show different B cell subset distributions
- Trial of early rituximab in adolescents with recent EBV-triggered ME/CFS
Limitations: This hypothesis is speculative. EBV is ubiquitous (95% adult seropositivity), so most ME/CFS patients will have been infected regardless of trigger. The specific role of infection timing versus other factors (genetic susceptibility, viral strain, co-infections) is unknown. Additionally, many adolescents with EBV-triggered ME/CFS do recover, suggesting protective factors beyond simple timing. See Section Pediatric-Adult ME/CFS Comparison Study for a proposed study design that could inform this hypothesis.
Longitudinal data from the first German study of adolescent and young adult ME/CFS following EBV-triggered infectious mononucleosis provide partial support for the age-dependent hypothesis. Pricoco et al. (2024) followed 25 patients (12 adolescents, 13 young adults) for 12 months after EBV-confirmed onset . The results revealed a stark age-dependent recovery pattern: 45% of adolescents no longer met ME/CFS diagnostic criteria at 12 months and showed improvement in fatigue and health-related quality of life, whereas 100% of young adults retained their diagnosis with minimal symptom or quality-of-life improvement. Patients averaged 27 distinct symptoms each, with exhaustion, daily-life limitations, rest requirements, and PEM being the most persistent. While the sample size is small and the study lacked a non-EBV control group, the differential recovery trajectory is consistent with the hypothesis that post-pubertal immune maturation reduces the capacity for spontaneous resolution of virus-triggered autoimmune processes.
Infectious mononucleosis (IM), caused by primary EBV infection, may trigger a form of ME/CFS with greater immunological burden than non-IM post-infectious ME/CFS, involving three measurable dimensions of immune dysregulation. First, massive polyclonal B cell activation during acute IM likely establishes a larger latent EBV reservoir in CD27+ memory B cells than subclinical EBV seroconversion, though direct reservoir-size comparison between IM and subclinical seroconversion has not been performed in ME/CFS. Second, NK cells, which expand dramatically during acute IM, may show impaired long-term EBV surveillance after the acute response resolves (note: NK “exhaustion” is a less established paradigm than T cell exhaustion; post-activation recovery kinetics differ, and the analogy to T cell exhaustion borrows credibility that the NK literature does not yet support). Third, CD8+ T cells targeting EBV epitopes develop exhausted phenotypes (PD-1+, TIM-3+, LAG-3+) from chronic antigen exposure, mirroring the T cell exhaustion already documented in ME/CFS. These three dimensions may all be downstream of a single core mechanism — overwhelming EBV antigen load during acute IM → global immune dysregulation — rather than three independent compartment failures. The consequence may be a subtype with larger viral reservoir, impaired antiviral surveillance, and enrichment of autoantibodies cross-reactive with EBV antigens. Whether IM-triggered patients differ from non-IM post-infectious ME/CFS in treatment response remains an open question. The IM association is disproportionately concentrated in early-onset ME/CFS (OR 2.32) (McGrath et al. 2026), consistent with the thymic-EBV synchrony model (Section Cross-Disease Extrapolation: Shared Symptoms Do Not Establish Shared Mechanisms). However, the 87% non-conversion rate after IM (Katz et al. 2009) equally supports a host-susceptibility model where trigger identity matters less than individual vulnerability factors, and the rituximab Phase III trial (Fluge 2019, n=151) showed no overall benefit across the ME/CFS population despite the EBV-mediated B cell reservoir hypothesis — demonstrating the gap between B cell dysregulation models and clinical response. Certainty: 0.40 (IM→ME/CFS conversion rate ~13% in prospective adolescent studies (Katz et al. 2009); differential adolescent recovery documented (Pricoco et al. 2024); three-compartment model based on known IM immunology but no head-to-head immunophenotyping of IM-ME/CFS vs non-IM ME/CFS). Not yet replicated: no study has directly compared IM-triggered to non-IM post-infectious ME/CFS with deep immune profiling. Falsifiable prediction: Head-to-head immunophenotyping of IM-triggered vs non-IM ME/CFS will show: (a) larger EBV reservoir in CD27+ memory B cells; (b) lower NK cytotoxicity (K562 assay); (c) higher exhausted CD8+ T cell frequency (PD-1+TIM-3+LAG-3+). All three dimensions are directional predictions with effect sizes drawn from general IM biology — not from ME/CFS-specific data. If any show no meaningful difference between groups, the three-dimension excess-burden model is not supported. Note: this is a compound hypothesis (three conjoint predictions); the certainty of 0.40 reflects the compound probability discount.
1.2 Human Herpesvirus 6 (HHV-6)
HHV-6 infects T cells and can integrate into chromosomes :
- Two species: HHV-6A and HHV-6B
- Evidence for active infection in some ME/CFS patients
- Can affect mitochondrial function
- Neurotropic (infects brain tissue)
HHV-6 miRNA-Mediated Mitochondrial Fragmentation
A key molecular mechanism linking HHV-6 reactivation to ME/CFS mitochondrial dysfunction was identified by Hennig, Prusty et al. (Hennig et al. 2022). HHV-6A encodes a viral microRNA, miR-aU14, which selectively inhibits host miR-30 family processing by binding pri-miRNA hairpin loops. Loss of miR-30 activates the miR-30–p53–DRP1 axis, producing profound mitochondrial fragmentation. The same mechanism impairs the type I interferon response, enabling productive HHV-6A reactivation from latency. miR-aU14 was characterized as a “readily druggable master regulator of the herpesvirus lytic–latent switch” (Hennig et al. 2022). The DRP1-driven fragmentation downstream of miR-aU14 is the same pathway that ISR kinase activation protects against, placing HHV-6 reactivation directly upstream of the mitochondrial architecture disruption discussed in Section Replication Status of Chapter Energy Metabolism and Mitochondrial Function.
HHV-6 Neuroinvasion in ME/CFS Post-Mortem Tissue
Kasimir, Prusty et al. examined post-mortem brain tissue from ME/CFS patients (n=3) vs controls (n=24) . HHV-6 viral miRNA (miR-aU14) was detected in multiple CNS regions — choroid plexus, hippocampus, amygdala, and dorsal root ganglia — exclusively in ME/CFS patients. EBV dUTPase was also found in all three ME/CFS brains but absent in controls. These findings confirm HHV-6 reactivation with neuroinvasion in ME/CFS CNS tissue, though the extremely small sample (n=3) requires replication before conclusions can be drawn.
1.3 Cytomegalovirus (CMV)
CMV establishes latency in monocytes and other cells :
- Reactivation documented in some ME/CFS patients
- Can cause significant inflammation upon reactivation
- Associated with T cell exhaustion (Iu et al. 2024)
1.4 Varicella-Zoster Virus (VZV)
VZV establishes latency in sensory ganglia following primary infection (chickenpox) and reactivates as herpes zoster (shingles) (Ariza, Mena Palomo, and Williams 2025):
- Elevated anti-VZV dUTPase antibodies detected in ME/CFS patients as part of the poly-herpesvirus co-reactivation pattern (M. Palomo et al. 2026)
- Encodes a dUTPase homolog (ORF8) with immunomodulatory properties (Ariza, Mena Palomo, and Williams 2025)
- Neurotropism in sensory ganglia may contribute to sensory and autonomic symptoms, though this has not been directly studied in ME/CFS
- Included in the poly-herpesvirus co-reactivation pattern described below (Section dUTPase Evidence Base)
1.5 Herpes Simplex Virus 1 (HSV-1) and Encephalitis
HSV-1 infects 57% of American adults and establishes latency in trigeminal ganglia. While most carriers remain asymptomatic, HSV-1 can cause herpes simplex encephalitis (HSE), an acute neurotropic infection affecting approximately 1 in 500,000 individuals per year. HSE presents with decreased consciousness, fatigue, confusion, and personality changes, and can mimic stroke presentation.
HSE as a Prospective Model for Post-Viral ME/CFS
The Uppsala University ME/CFS Collaborative Research Center (directed by Jonas Bergquist, OMF-funded) has studied HSE as a uniquely informative model for post-viral fatigue and ME/CFS onset, since HSE patients can be followed prospectively from the acute infection . A series of studies in approximately 50 HSE patients with serial CSF and blood sampling has revealed:
- NMDA receptor autoimmunity: Anti-NMDAR IgG antibodies developed in 24.5% of HSE patients, never present at disease onset but appearing after 3 months . 56% had anti-NMDAR antibodies of any isotype class (IgG 25%, IgM 29%, IgA 27%), though only IgG correlated with cognitive outcome (Westman et al. 2018).
- Impaired neurocognitive recovery: NMDAR-seropositive patients had dramatically worse neurocognitive recovery — median MDRS score increase of 1.5 points versus 10 points at 24 months (p=0.018) .
- Proposed causative chain: Neurofilament light chain (NFL, a marker of neuronal damage) correlated with both impaired cognition (rho=-0.36, p=0.020) and subsequent NMDAR autoimmunization (p=0.006), suggesting: brain tissue damage → NMDAR antigen release → autoimmunization → prolonged CSF inflammation → persistent neurocognitive dysfunction .
- CSF proteomics: Temporal analysis of 890 CSF proteins (LC-MS) revealed an acute multi-pathway response at days 0–9 (acute phase, antimicrobial pattern recognition, glycolysis/gluconeogenesis), which resolved by 2 weeks. Six proteins were significantly reduced in NMDAR-seropositive patients, including apolipoprotein A1 — previously linked to NMDAR encephalitis — and complement factor I .
- Predictive viral antibody signature: PhIP-Seq analysis identified HSV-1 UL42 and UL48 antibody signatures that predict which HSE patients develop secondary NMDAR encephalitis (75% sensitivity, \(>\) 99% specificity, PPV 90%, OR 209) (Westman, Nääs, et al. 2025).
Certainty: 0.45. The HSE → NMDAR autoimmunity → persistent neurocognitive dysfunction sequence, documented prospectively in a cohort with known viral onset, may represent a generalizable mechanism for post-infectious ME/CFS. The certainty level reflects: (1) strong prospective evidence for the causative chain in HSE (NFL → NMDAR autoimmunization → cognitive decline); (2) 70–80% of ME/CFS patients report infection-triggered onset; (3) GPCR autoantibodies are documented in ME/CFS (Section GPCR Autoantibody-Driven Dysfunction), and receptor internalization by autoantibodies is established for NMDAR (Section Autoantibodies as Monocyte Programmers); (4) however, HSE causes severe, focal brain damage with high NFL, whereas most ME/CFS-triggering infections (EBV, enteroviruses) do not typically produce this degree of tissue destruction; (5) extrapolation from HSV-1 to other viral triggers remains unconfirmed.
If viral-induced brain tissue damage is a necessary precursor to autoantibody-mediated ME/CFS, then: (1) ME/CFS patients with higher acute-phase NFL levels should have worse long-term outcomes; (2) subclinical neuroinflammation markers (CSF neopterin, TSPO-PET signal) should precede autoantibody development; (3) early immunomodulatory therapy post-infection should reduce NMDAR autoantibody incidence and improve neurocognitive outcomes; (4) ME/CFS triggered by non-neurotropic infections should show lower autoantibody prevalence than ME/CFS triggered by neurotropic viruses.
Treatment implications: This model suggests a therapeutic window in the acute and early recovery phases of triggering infections. If autoantibody development follows neuronal damage with a 3-month lag (as in HSE), early anti-inflammatory or immunosuppressive intervention during this window could prevent the establishment of chronic autoimmunity. Bergquist’s group is conducting a 160-patient clinical trial at Harvard testing low-dose naltrexone, pyridostigmine, and their combination, which may provide relevant data.
Limitations: HSE causes catastrophic, often fatal brain injury; the degree of neuronal damage in typical ME/CFS-triggering infections is orders of magnitude lower. The relevance of this severe model to mild post-viral fatigue requires demonstration that the same mechanism operates at subclinical tissue damage levels. Additionally, only a subset of ME/CFS patients report post-infectious onset, limiting the generalizability of any virus-triggered model.
1.6 HSV-1/2 as a Visible, Patient-Reportable Viral Reactivation Probe
Unlike EBV and HHV-6 — whose reactivation is invisible, requires serological or molecular testing to detect, and cannot be timestamped by the patient — HSV-1/2 reactivation produces visible mucocutaneous lesions (cold sores, genital herpes) that the patient can self-document. The advantage of HSV over the other herpesviruses in the panel is recurrence frequency: VZV reactivation (shingles) is also visible but typically occurs once in a lifetime; HSV reactivates often enough for within-patient temporal analysis, making it the only herpesvirus in the panel suited as a longitudinal probe.
Mechanistic Rationale for HSV as a PEM Probe
The pathway linking exertion to HSV reactivation is mechanistically coherent at several steps, though critically none has been prospectively tested in ME/CFS:
- PEM involves sympathetic nervous system activation — documented by CPET studies showing elevated norepinephrine, heart rate, and blood lactate during and after exercise challenge in ME/CFS patients.
- Sympathetic activation drives catecholamine release (norepinephrine, epinephrine), which acts on beta-adrenergic and alpha-adrenergic receptors on immune cells — including Tregs — altering immune surveillance of latent viral reservoirs.
- Catecholamine-mediated stress shows a significant but weak association with HSV recurrence in the general population. A meta-analysis of 11 prospective studies found a positive association between psychosocial stress and symptomatic HSV recurrence (\(r\) = 0.083, 95% CI 0.025–0.141, \(p\) = 0.005), explaining less than 1% of variance, with some indication of publication bias (Chida and Mao 2009). Notably, psychological distress was more strongly associated than stress stimuli per se — and PEM is primarily a physiological stimulus — meaning the meta-analysis’s strongest finding is not the one most relevant to PEM.
- The molecular mechanism has been partially characterized in animal models: host stress leads to increased Treg-mediated suppression of CD8+ T cell immune surveillance at the trigeminal ganglion, permitting HSV-1 to exit latency (Yu et al. 2018).
- HSV-1/2 latency–reactivation kinetics are among the best-characterized in virology. Reactivation follows a predictable cascade: immediate-early gene expression → early gene expression → viral DNA replication → virion assembly and anterograde transport → visible lesion. Only productive (not abortive) replication produces visible lesions; subclinical shedding is more common and is detected only by PCR.
- The complete pathway, with exertion as the temporal anchor: exertion exceeding the anaerobic threshold → sympathetic activation → catecholamine surge → Treg-mediated CD8+ T cell suppression at ganglia → HSV exit from latency → anterograde transport (1–3 days) → viral replication in epithelium (1–2 days) → visible cold sore. The cold sore provides information beyond what the patient’s symptoms or autonomic monitoring already convey because it is an independent biological readout: productive viral replication at a ganglion anatomically distant from the sites generating PEM symptoms (muscle, vasculature, brain), and the cold sore is not a PEM criterion symptom. If catecholamines drive PEM symptoms AND HSV reactivation but the two are mechanistically unlinked, the cold sore adds timing precision and biological specificity that HRV/norepinephrine cannot provide, because it confirms that catecholamine levels reached a threshold sufficient to reactivate dormant virus at a distant anatomical site.
Known Triggers and Base Rates
HSV-1 global prevalence is approximately 67% in adults under 50 years (~57% in the US), with 20–40% of seropositive individuals experiencing recurrent outbreaks (Gopinath et al. 2023). This means approximately 33–43% of adults are seronegative and excluded from the probe, and among the seropositive, only a minority experience regular recurrences. Known triggers — UV radiation, psychological stress, fever, menstruation, physical trauma, sleep disruption — are all established in the general population but have never been prospectively tracked against PEM episodes in ME/CFS. A patient seropositive for HSV-1 who experiences 2–3 outbreaks per year is in the normal range; a patient whose outbreaks cluster temporally with PEM episodes provides a different signal (Gopinath et al. 2023). Whether the seropositive probe-eligible subset differs clinically from the excluded seronegative subset is unknown; probe findings will not generalize to approximately 33–43% of patients.
The Probe Framework
Certainty: 0.20. The mechanistic rationale is coherent but every ME/CFS-specific link is untested. The stress→HSV reactivation link is established in general-population studies (Chida 2009 meta-analysis, 11 studies, 0.53 discounted certainty) (Chida and Mao 2009). The PEM→catecholamine pathway is established in ME/CFS CPET literature (documented elevated norepinephrine during and after exercise challenge). However: (1) no prospective study has tracked HSV outbreak timing relative to PEM episodes in ME/CFS; (2) the Treg→CD8+ T cell→HSV latency mechanism has been demonstrated in mice (Yu et al. 2018) (discounted certainty 0.30 due to cross-species translation gap) but not in humans with ME/CFS; (3) the null serology literature (Buchwald 1996, n=548; Koelle 2002, 22 twin pairs) found no differences in HSV-1/2 positivity or titre between ME/CFS patients and controls (Buchwald et al. 1996) (Koelle et al. 2002), suggesting that HSV infection per se is not the driver — the temporal pattern of reactivation (frequency relative to PEM) may be informative where static serology is not; (4) HSV recurrence is multifactorial (UV exposure, fever, menstruation, sleep disruption, psychological stress) and PEM-independent triggers must be distinguished from PEM-associated triggers; (5) baseline HSV recurrence in the general population is 20–40%, meaning many reported outbreaks may be normal population-variation. Origin: brainstorm.
If HSV outbreaks systematically cluster around exertion episodes (within 24–72 hours of exertion exceeding the individual anaerobic threshold), this would provide: (a) direct evidence of endogenous viral reactivation triggered by PEM — closing the gap between immune dysfunction models and the failure to detect reactivation in cross-sectional blood sampling; (b) an individualized, patient-reportable biomarker of viral reactivation requiring no laboratory testing; (c) a secondary endpoint for antiviral trials. The cold sore is an independent biological readout (productive viral replication at a ganglion anatomically distant from symptom-generation sites; not a PEM criterion symptom), not simply one more PEM-correlated symptom counted against others. If HSV outbreaks show no temporal clustering with exertion but are randomly distributed, the probe framework is falsified. If HSV outbreaks cluster with exertion but antiviral suppression of HSV does not reduce PEM severity, HSV reactivation is a downstream consequence of PEM rather than a driver — still useful as a real-time viral reactivation probe, but not as a therapeutic target. If catecholamines drive both outbreak clustering and PEM through independent pathways, concordance would still establish the cold sore as a catecholamine-saturation biosensor complementary to HRV/norepinephrine — confirming that systemic catecholamine levels reached a threshold sufficient to reactivate dormant virus, which autonomic monitoring alone cannot establish.
Falsifiable prediction: A prospective diary study (\(n\) ≥ 100 HSV-1 seropositive ME/CFS patients, 12 months — powered for the expected 15–25 outbreak events over 6 months in a 20–40% recurrence population) tracking daily exertion, PEM onset, HSV prodromal symptoms, and visible cold sore onset with photographic documentation PLUS home lesion-swab PCR (self-collected swab kits) to confirm HSV etiology vs aphthous stomatitis will show: (1) ≥60% of all HSV-PCR-confirmed outbreaks occur within the 24–96 hour window post-exertion exceeding the anaerobic threshold, versus the fraction expected by chance given the proportion of observation days falling within post-exertion windows; (2) per-patient outbreak frequency is higher during months with ≥2 PEM episodes than during PEM-light months; (3) self-reported stress scores (PSS), sleep disruption, menstrual cycle phase, and UV exposure do not independently predict outbreak timing after controlling for exertion. Bidirectional lag analysis (outbreak→PEM and PEM→outbreak) must be included to exclude the alternative that HSV prodrome precipitates PEM-like flares. Falsified if PCR-confirmed outbreaks are equally temporally distributed around exertion and non-exertion windows, OR if stress/sleep/UV scores predict outbreak timing better than exertion.
Consequence: If validated, a patient’s cold sore diary could function as a real-time, patient-owned viral reactivation biomarker — inexpensive and accessible, requiring no laboratory infrastructure. This would give patients and clinicians a tool to document what is currently invisible: the connection between exertion and viral recrudescence. However, this is entirely untested and may be a dead end — HSV recurrence in the general population is common and multifactorial.
Severity applicability: unknown — study populations not stratified by ME/CFS severity. Evidence source: general-population stress–HSV link (Chida 2009) + murine Treg–CD8–HSV mechanism (Yu 2018). Translation gap: animal model → human; general-population → ME/CFS. Competing explanations: UV exposure, fever, menstruation, sleep disruption, psychological stress independent of PEM, random variation. Diagnostic criteria for cited studies: Buchwald 1996 and Koelle 2002 used CDC 1994 (Fukuda) criteria — broader, less severe CFS population than contemporary IOM/ICC definitions.
The Treg-HSV Latency Connection
Certainty: N/A — mechanistic hypothesis with zero direct empirical support in ME/CFS. Synthesis of two independent lines: (1) the Treg→CD8+ T cell→HSV latency axis established in mice (Yu et al. 2018) (discounted certainty 0.30), and (2) the Treg-herpesvirus-immune hyperactivation model proposed for ME/CFS (Sepúlveda et al. 2019) (certainty 0.50, computational model only). If both lines are correct, ME/CFS patients — particularly those with elevated Treg percentages — have HSV outbreaks that are more frequent and more tightly coupled to PEM episodes. However, the Treg model in ME/CFS is unvalidated (computational), the Treg-to-HSV reactivation mechanism has not been tested in humans, and the Treg phenotype in ME/CFS is itself contested. Origin: brainstorm.
Falsifiable prediction: In HSV-1 seropositive ME/CFS patients with prospective diary: Treg percentage (CD4+CD25+CD127low) correlates with monthly HSV outbreak frequency (\(r\) ≥ 0.3); PEM-associated outbreaks show higher pre-outbreak Treg percentages; CD8+ T cell functionality (IFN-\(\gamma\) ELISpot to HSV-1 peptides) is lower in patients with more frequent outbreaks. Falsified if no correlation between Treg percentage and outbreak frequency.
Consequence: If confirmed, this would link previously disconnected ME/CFS research areas — Treg biology and herpesvirus reactivation — through a specific molecular mechanism. However, this is entirely speculative and decades from clinical application.
Analogizing the LSR to HSV
The LSR framework was developed for EBV and is EBV-specific: the ratio of anti-BZLF1 IgG to anti-VCA-p18 IgG distinguishes abortive lytic reactivation from LLPC antibody maintenance (Lytic-to-Structural IgG Ratio (LSR) as a Diagnostic Biomarker). HSV-1/2 have different latency biology (neuronal, not B cell), different reactivation triggers, and different antibody targets.
Certainty: N/A — research question. An HSV analog would compare anti-ICP0 (immediate-early, expressed at reactivation initiation) to anti-gB/gD (late structural). If HSV reactivation in ME/CFS follows the abortive lytic replication pattern (Section HSV-LSR Analog Requires De Novo Assay Development), the ICP0-to-structural IgG ratio should be elevated in patients with frequent PEM-associated outbreaks. If the HSV-LSR is elevated but EBV-LSR is normal, the signal is compartmentalized (trigeminal ganglion-specific, not systemic B cell-driven). However, anti-ICP0 IgG has never been measured in any disease context. Origin: brainstorm.
Consequence: If validated, an HSV-LSR would provide an objective serological correlate of visible reactivation. But this requires anti-ICP0 serology development that does not currently exist, placing it years from implementation.
Critique: structurally unfalsifiable — the prediction depends on an anti-ICP0 IgG ELISA that does not exist. No commercial or research-grade ICP0 serology assay has been developed, and ICP0 immunogenicity in natural human infection has never been characterized. Until such an assay is created and validated, this question cannot be empirically tested.
Structural Limitations of the HSV Probe
Certainty: N/A — methodological assessment. The HSV probe faces several structural barriers. Baseline HSV recurrence is common (20–40% of seropositive adults) — the probe’s discriminant value depends on demonstrating PEM-specific temporal clustering above background rates. Approximately 33–43% of adults are HSV-1 seronegative and excluded; whether the probe-eligible seropositive subset differs clinically from the excluded seronegative subset is unknown — Buchwald 1996 compared ME/CFS vs controls, not HSV-seropositive vs seronegative within ME/CFS. Patient self-report of cold sores is unreliable without PCR confirmation — aphthous stomatitis (canker sores) is immune-mediated, not viral, and clinically indistinguishable from herpes labialis. Any validation study must include home lesion-swab PCR (commercially available self-collection kits are standard in HSV clinical trials) to confirm HSV etiology. HSV-1 (trigeminal/oral, ~57–67% adult prevalence) and HSV-2 (sacral/genital, ~12% US adult prevalence) have different latency sites and triggers; conflation dilutes the signal. The most parsimonious null hypothesis — that HSV outbreaks and PEM share a catecholamine trigger but are mechanistically independent, making the HSV probe redundant with existing autonomic monitoring (HRV, norepinephrine) — has not been falsified. The probe’s argument against this null is that a cold sore confirms catecholamine levels reached a threshold sufficient to reactivate dormant virus at a distant anatomical site, which HRV alone cannot establish; this argument itself is untested. Concordance between elevated EBV-LSR and frequent HSV outbreaks is equally consistent with confounding by illness severity and cumulative stress exposure — sicker patients have more PEM, more distress, and plausibly more of both readouts. Origin: brainstorm — categories 10–12 consolidation.
Consequence: none — this is a methodological quality assessment enumerating structural barriers the probe must overcome. The prospective diary study (HSV Outbreak Frequency × PEM Timing as Endogenous Viral Reactivation Probe) is the critical first step.
1.7 Why Herpesvirus Antibodies Persist for Decades
Herpesvirus antibodies are commonly cited as evidence of viral reactivation in ME/CFS — but antibody titre interpretation is constrained by a body of basic immunology that is often overlooked in clinical reasoning. The persistence of antiviral antibodies decades after infection is largely the work of long-lived plasma cells (LLPCs), which maintain antibody output independently of ongoing antigenic stimulation. This does not mean that herpesvirus reactivation does not occur in ME/CFS — the evidence for abortive lytic replication and poly-herpesvirus co-reactivation is substantial (Sections HSV-LSR Analog Requires De Novo Assay Development, dUTPase Evidence Base). It means that structural antigen IgG titre alone — the most commonly used clinical test — cannot distinguish between memory of a past infection and evidence of current viral activity. The implication is not that all herpesvirus serology is useless, but that its interpretation requires the right antigen targets and the right clinical context.
1.7.1 The Immunology of Lifelong Antibody Maintenance
After an acute infection, short-lived plasma cells in secondary lymphoid organs produce the initial antibody burst. A fraction of activated B cells differentiate into long-lived plasma cells that migrate to survival niches in the bone marrow, where they constitutively secrete antibody for the lifetime of the host — without requiring re-exposure to antigen, without dividing, and independently of memory B cell replenishment (Amanna, Carlson, and Slifka 2007) (Hammarlund et al. 2017) (Robinson et al. 2022).
The Amanna–Slifka longitudinal cohort (n=45, followed up to 26 years) established the quantitative foundation (Amanna, Carlson, and Slifka 2007):
- Antiviral antibody half-lives: VZV ~50 years; EBV, measles, mumps >200 years
- By contrast, vaccine-induced antibodies to non-replicating antigens (tetanus: 11 years; diphtheria: 19 years) wane much faster
- Memory B cell numbers did NOT correlate with serum antibody titres for 5/8 antigens tested — demonstrating that antibody levels and B cell memory are independently regulated
The Hammarlund primate model provided direct experimental confirmation (Hammarlund et al. 2017): bone marrow LLPCs survived for more than a decade after sustained depletion of memory B cells. BrdU pulse-chase experiments detected plasma cells labelled at the time of vaccination still alive 10 years later without any evidence of cell division. The LLPC pool alone — without memory B cell input and without antigenic stimulation — was sufficient to maintain stable serum antibody titres.
A critical structural determinant of antibody longevity is the physical form of the antigen that drove the original response (Slifka and Amanna 2019) (Chackerian and Peabody 2020). Highly multivalent, repetitive antigen arrays — such as viral capsids and envelope glycoprotein lattices — cross-link B cell receptors with high avidity, lowering the activation threshold and preferentially driving LLPC differentiation. Herpesviruses present precisely this type of antigen: icosahedral capsids with repeating structural subunits, dense glycoprotein arrays on the viral envelope, and tegument proteins packaged at high copy number. They are structurally optimised to generate exceptionally durable antibody responses.
Robinson et al. demonstrated that LLPCs accumulate in bone marrow at a rate of approximately one cell per hour after immunisation, and that the final number scales with the duration of the germinal centre reaction (Robinson et al. 2022). This implies that the magnitude of the initial infection — the antigen load and the duration of the germinal centre response — determines long-term antibody levels, not the frequency of subsequent reactivation events.
1.7.2 What This Means for Herpesvirus Serology
The persistence of herpesvirus antibodies decades after infection is therefore expected and normal — but the picture is not binary. LLPCs provide a substantial antigen-independent baseline that persists for the host’s lifetime. Against this high baseline, the contribution of any individual reactivation event to total structural antigen IgG may be modest and difficult to detect. However — and this is a critical caveat — structural antigen IgG CAN rise measurably during clinical herpesvirus reactivation: VZV IgG rises during shingles, CMV IgG rises during CMV disease, and EBV VCA/EBNA-1 IgG can rise during EBV-driven post-transplant lymphoproliferative disorder. The LLPC baseline dominates the total signal, but reactivation can contribute to it. The correct interpretive principle is therefore not “structural IgG is pure LLPC output and never indicates viral activity” — it is “structural IgG has a large antigen-independent LLPC component, and elevated titres alone, without corroborating evidence from viral DNA, viral mRNA, or lytic-cycle antibody data, are insufficient to diagnose active reactivation.”
This does not mean herpesvirus reactivation does not occur in ME/CFS — the dUTPase, ALR, and poly-herpesvirus co-reactivation evidence is discussed below (Sections HSV-LSR Analog Requires De Novo Assay Development, dUTPase Evidence Base). It means that antibody titre alone — particularly against structural viral antigens (VCA, gB) and latency proteins (EBNA-1, which is not a structural antigen but is functionally in the same interpretive category because it is constitutively expressed during latency) — cannot distinguish between:
- Normal LLPC maintenance without viral activity
- Low-level abortive lytic replication in the absence of productive infection
- Overt lytic reactivation with infectious virion production
1.7.4 Evidence in ME/CFS: Is Antibody Elevation Real?
The question of whether ME/CFS patients have elevated herpesvirus antibodies above the healthy seropositive baseline is not settled:
Evidence of elevation:
- Palomo et al. (n=40 ME/CFS, 16 controls): elevated anti-dUTPase IgG to EBV, HHV-6, VZV; 72.5% co-expression of multiple herpesvirus antibodies vs 31% controls (I. M. Palomo et al. 2026)
- Apostolou et al. (n=95 ME/CFS, 110 controls): elevated salivary EBNA-1 IgG and EBV-VCA IgG in ME/CFS — detected locally in saliva, not systemically in plasma (Apostolou et al. 2022)
- Loebel et al. (n=92 CFS, 50 controls): EBNA-6 repeat region IgG enhanced in CFS, though overall EBV IgG pattern “quite similar” between groups (Loebel et al. 2017)
Evidence of no elevation:
- Cliff et al. (n=251 ME/CFS, 107 controls): UK ME/CFS Biobank — no differences in seroprevalence for six herpesviruses between ME/CFS and controls (Cliff et al. 2019)
- Blomberg et al. (Swedish cohort, Canada criteria): suspension multiplex immunoassay for HHV-1-7 — overall IgG anti-herpes-viral reactivities did NOT show significant differences between ME/CFS and blood donor controls (Blomberg et al. 2019)
Certainty: contested. Two large seroprevalence studies (Cliff n=251, Blomberg) found no overall elevation; three smaller studies (Palomo, Apostolou, Loebel) found antigen-specific elevations. The contradiction is likely resolvable by antigen specificity — dUTPase (an early lytic protein) and EBNA-6 (a latency protein with human homology) may be selectively elevated while structural antigen antibodies (VCA, EBNA-1, gB) are not. If this is correct, ME/CFS is characterised by a qualitative shift in antibody fine specificity (toward lytic-cycle and cross-reactive epitopes) rather than a quantitative elevation in total herpesvirus IgG. This would explain why seroprevalence studies (which measure presence/absence of IgG against structural antigens) find no difference, while studies using lytic-cycle antigens find elevations.
Falsifiable prediction 1: A study measuring IgG against structural antigens (VCA, gB, gE) AND lytic-cycle antigens (dUTPase, EA-D) AND latency-protein repeat regions (EBNA-6) simultaneously in the same ME/CFS cohort should find elevations only in the lytic and repeat-region categories, not in structural antigen IgG. Falsified if structural antigen IgG is equally elevated.
Falsifiable prediction 2: ME/CFS patients with elevated anti-dUTPase antibodies should show higher viral DNA loads (qPCR) than patients with normal anti-dUTPase levels. Falsified if dUTPase antibody elevation is independent of viral DNA detection.
Consequence: If the antigen-specificity model is correct, herpesvirus serology in ME/CFS diagnosis and research should shift from measuring total IgG seroprevalence to targeted measurement of lytic-cycle antibodies (dUTPase, EA-D). Total IgG panels — the most commonly used clinical test — are uninformative.
Certainty: 0.45. In normal LLPC biology, the ratio of IgG against lytic-cycle herpesvirus antigens (EBV BZLF1, EA-D, dUTPase) to structural antigens (VCA-p18 — the immunodominant capsid protein measured by standard VCA IgG — EBNA-1, gB) should remain stable — both antibody types are produced by the same LLPC pool established during primary infection. If ME/CFS involves abortive lytic reactivation (preferentially boosting lytic-antigen responses via short-lived plasmablasts) or epitope-specific autoimmune drift (affecting specific lytic epitopes), the LSR will be elevated. This ratio uses each patient as their own internal control, eliminating confounding by infection timing, age, and initial antigen load. The concept directly explains the contradictory literature: structural-antigen IgG is normal (Cliff, Blomberg null findings), while lytic-antigen IgG is elevated (Palomo dUTPase, Loebel EBNA-6 repeat). A partial independent replication of dUTPase antibody elevation now exists outside the Ariza/Williams group — Maes et al. (2025) found EBV and HHV-6 dUTPase antibodies accounted for 63.7% of fatigue-depression-anxiety variance in RRMS patients, though this is a different disease population and requires confirmation in ME/CFS (Maes et al. 2025). Origin: brainstorm.
The precedent from multiple sclerosis is instructive: EBNA-1 antibodies cross-react with GlialCAM specifically in MS patients, not in seropositive healthy controls (Lanz et al. 2022). The pathogenic feature is the qualitative shift in antibody fine specificity — the same principle may apply in ME/CFS, where cross-reactive antibodies could target autonomic neurons, skeletal muscle, or mitochondrial antigens. The elevated titre is a byproduct of autoimmune drift, not the primary pathology.
Falsifiable prediction: The LSR (anti-BZLF1 IgG / anti-VCA-p18 IgG ratio) will discriminate ME/CFS from healthy seropositive controls (AUC ≥ 0.75) and correlate with symptom severity (r ≥ 0.3 with fatigue score), while neither anti-BZLF1 nor anti-VCA alone will show significant discrimination. IgG avidity (urea-wash ELISA) for anti-herpesvirus antibodies will be uniformly high (avidity index >0.6, consistent with remote past infection) — low-avidity IgG (\(<\) 0.4) would refute the LLPC model and support recent abortive reactivation.
Consequence: A blood test comparing the ratio of two antibody types could distinguish ME/CFS patients with an ongoing immune problem from those with normal antibodies from past infections — providing a low-cost diagnostic tool using existing clinical assays.
The evidence base for elevated anti-herpesvirus antibodies in ME/CFS suffers from methodological threats on both sides. The positive studies (Palomo, Apostolou, Loebel) are small, use different antigen preparations with no standardisation, and have not been independently replicated — the dUTPase findings derive from a single research group. The null studies have their own limitations: Cliff 2019 measured seroprevalence (binary ±) rather than quantitative titres, and in a population where >90% of adults are seropositive for EBV and VZV, seroprevalence cannot detect titre differences — Cliff’s null result for structural antigen seroprevalence confirms universal prior infection, not absence of titre elevation. Blomberg 2019 used a suspension multiplex immunoassay whose sensitivity for detecting subtle titre differences compared to standard ELISA has not been established, and compared ME/CFS patients to blood donor controls who are systematically healthier than the general population. The question of whether herpesvirus antibody elevation exists in ME/CFS is not scientifically settled for any antigen class — the literature is too methodologically heterogeneous on both sides to support a definitive conclusion. Researchers designing studies should pre-register protocols, use ≥3 distinct antigen preparations spanning structural and lytic-cycle targets, match controls on age/sex/infection timing, and apply correction for multiple comparisons. Until a study meeting these criteria is published, the question remains open. Origin: brainstorm.
Consequence: The herpesvirus serology literature in ME/CFS is too methodologically heterogeneous to support a definitive conclusion in either direction. Researchers designing studies should pre-register their protocols, use ≥3 distinct antigen preparations spanning structural and lytic-cycle targets, match controls on age/sex/infection timing, and apply Bonferroni correction.
Certainty: 0.30. An alternative to viral reactivation as the driver of elevated herpesvirus antibodies is intrinsic B cell dysregulation. Sun et al. (2024) performed scRNA-seq on peripheral blood from ME/CFS patients (n=4) and found that memory B cells in ME/CFS show a unique subtype early in pseudotime, with increased trajectory toward plasma cell differentiation — suggesting B cell overactivity (Sun et al. 2024). If this finding replicates in larger cohorts, it would mean that elevated herpesvirus antibodies in ME/CFS could reflect dysregulated B cell → plasma cell differentiation rather than antigen-driven responses to viral reactivation. The certainty is low (0.30) because the finding comes from a single study with n=4 patients and n=4 controls, and the link between increased plasma cell differentiation and herpesvirus antibody titres has not been directly tested.
Falsifiable prediction: ME/CFS patients with elevated herpesvirus antibodies will show increased frequencies of antibody-secreting cells (CD19^low CD27^high CD38^high) in peripheral blood AND increased plasma cell numbers in bone marrow aspirates, independent of viral DNA load. Falsified if antibody elevation always correlates with detectable viral DNA.
Consequence: If this mechanism is confirmed, herpesvirus antibody elevation in ME/CFS would be a readout of B cell dysregulation rather than viral reactivation — changing the diagnostic and therapeutic implications entirely. Antiviral therapy would not be expected to reduce antibody titres, and the focus would shift to B cell-targeted interventions. However, this is early-stage; the finding may turn out to be a dead end.
Total IgG against structural herpesvirus antigens (VCA, CMV gB, VZV gE) and latency proteins (EBNA-1) has a substantial antigen-independent component maintained by long-lived plasma cells established at primary infection — not a pure readout of the current state of the virus. LLPCs maintain antibody titres for decades without antigenic stimulation ((Amanna, Carlson, and Slifka 2007), half-life >200 yr for EBV; (Hammarlund et al. 2017), LLPC survival >10 yr without memory B cells). However, structural antigen IgG CAN be boosted by clinical reactivation (e.g., VZV IgG rises during shingles), so elevated titres against a high LLPC baseline are ambiguous. An elevated titre may indicate: (a) greater LLPC numbers from a more intense primary infection; (b) individual variation in bone marrow niche capacity; (c) ongoing antigenic stimulation from viral reactivation; or (d) intrinsic B cell dysregulation driving excess plasma cell differentiation ((Sun et al. 2024), preliminary, n=4). The critical distinction is between detectable antibodies (universal in seropositive adults, expected from LLPCs) and significantly elevated titres above age/sex-matched norms (could reflect any of a–d). Without viral DNA (qPCR), viral mRNA (in situ hybridisation), or lytic-cycle antibody data, herpesvirus IgG titre alone is uninformative about viral activity — but when confirmatory testing is negative or unavailable, a time-limited antiviral trial with defined stopping rules remains a defensible clinical option for the post-infectious subset. This principle applies to all herpesviruses (EBV, HHV-6, CMV, VZV, HSV-1/2) and is not specific to ME/CFS — it is a general constraint on serological interpretation.
Consequence: Clinicians should not diagnose “chronic EBV” or “chronic herpesvirus reactivation” based on IgG serology alone, and should not initiate indefinite antiviral therapy on the basis of isolated structural antigen IgG titres. A time-limited antiviral trial with predefined stopping rules and symptom tracking is a reasonable shared-decision option when clinical suspicion is high and confirmatory testing is unavailable or negative (acknowledging that qPCR may be negative during abortive lytic replication). This balances patient safety — protecting against unnecessary long-term antiviral treatment — with access to a therapeutic trial in appropriate cases.
The evidence collectively argues that herpesvirus serology in ME/CFS has been systematically misinterpreted: the large antigen-independent LLPC baseline means that isolated structural-antigen or latency-protein IgG elevation is insufficient to diagnose active viral reactivation, while the contradictory literature on antibody elevation has not been resolved by any study to date — the evidence is too methodologically heterogeneous on both sides to support a definitive conclusion (Is Herpesvirus Antibody Elevation a Universal Feature of ME/CFS?, Methodological Weaknesses in the Herpesvirus Serology Literature). The strongest practical implication is that clinical decisions should not rest on structural antigen IgG alone (Herpesvirus Antibody Titres Cannot Be Interpreted as Evidence of Active Infection) — but neither should the absence of accessible confirmatory testing preclude a time-limited therapeutic trial in appropriate patients. The dUTPase, ALR, and poly-herpesvirus evidence (Sections HSV-LSR Analog Requires De Novo Assay Development, dUTPase Evidence Base) establishes that viral reactivation occurs in a subset of ME/CFS patients; what remains unresolved is how to identify that subset with available tests.
Consequence: The existing clinical practice of diagnosing “viral reactivation” from herpesvirus IgG panels alone is not supported by basic immunology. The path forward is: (1) use viral DNA (qPCR), viral mRNA, and lytic-cycle antibody testing where available — with the caveat that qPCR may be negative during abortive lytic replication; (2) where these are unavailable, a time-limited antiviral trial with predefined stopping rules remains a reasonable clinical option for the post-infectious subset; (3) research should focus on validating antigen-specific biomarkers (including dUTPase antibodies, pending independent replication) rather than continuing to measure total IgG seroprevalence.
1.8 What Antiviral Response Reveals About the Elevated LSR
#clinical-caution()
1.8.1 Two Candidate Mechanisms
An elevated LSR (anti-BZLF1 IgG / anti-VCA-p18 IgG) could arise from two fundamentally different mechanisms, with different therapeutic implications:
Mechanism 1 — Abortive lytic reactivation (ALR): Stress, inflammation, or immune suppression triggers EBV’s immediate-early genes (BZLF1, BRLF1). The lytic cascade initiates through early gene expression (dUTPase, EA-D), preferentially boosting short-lived plasmablast (SLPB) responses to lytic antigens. Because the reactivation is abortive — the viral DNA replication step is either blocked or insufficient for productive replication — late structural genes (VCA, gB) are NOT expressed, and no infectious virions are produced. The LSR rises because lytic IgG (from both LLPC baseline AND SLPB output) increases while structural IgG (LLPC-only) stays stable.
Mechanism 2 — LLPC autoimmune drift (no ongoing viral activity): ME/CFS-associated B cell dysregulation (increased plasma cell differentiation trajectory, (Sun et al. 2024)) drives preferential survival and clonal expansion of bone marrow LLPCs producing cross-reactive lytic-epitope antibodies. Anti-BZLF1/dUTPase IgG is selectively enriched relative to anti-VCA-p18 IgG through niche competition for APRIL/BAFF/IL-6 survival signals. The LSR rises — but there is NO active viral replication (no viral DNA, no viral mRNA, no ALR). The elevated lytic antibodies are a readout of B cell dysregulation, not viral activity.
1.8.2 Interpreting Valacyclovir Response Through the LSR
Certainty: 0.15. Valacyclovir (acyclovir prodrug) inhibits EBV DNA polymerase (BALF5) — it blocks viral DNA replication but does NOT block BZLF1 or dUTPase expression (immediate-early and early genes are transcribed before DNA replication). It does NOT deplete existing LLPCs or SLPBs. Because valacyclovir acts selectively on DNA replication (a mid-cascade node in ALR) and has no effect on LLPC maintenance, its clinical response pattern can help distinguish between the two candidate mechanisms — but imperfectly.
If valacyclovir produces clinical improvement AND LSR declines after 3–6 months, Mechanism 1 (ALR-driven) is supported but not proven: valacyclovir is also active against VZV and HSV, and clinical improvement could reflect suppression of non-EBV herpesviruses. Inference certainty is low (0.10) due to poly-herpesvirus antiviral activity and the inability to exclude off-target mechanisms.
If valacyclovir produces zero clinical response AND LSR remains elevated, the most likely interpretation is Mechanism 2 (LLPC autoimmune drift): the elevated LSR reflects B cell dysregulation and clonal LLPC selection, not ongoing viral replication. Valacyclovir has no target in this cascade — it neither blocks LLPC antibody secretion nor depletes cross-reactive LLPC clones. However, an alternative cannot be excluded: if ALR is stuck at the immediate-early/early gene expression stage (BZLF1 and dUTPase are expressed, but the cascade never reaches DNA replication), valacyclovir would have NO target in the cascade because the DNA polymerase step is never reached. In this scenario, lytic antigens continue to be produced and the LSR stays elevated despite successful DNA polymerase blockade. The distinction between “LLPC-driven LSR” and “early-gene-only ALR with valacyclovir-irrelevant bottleneck” cannot be made from valacyclovir response alone — it requires direct measurement of viral gene expression (EA-D IgG, dUTPase IgG kinetics, or ideally viral mRNA in PBMCs).
If valacyclovir fails but valganciclovir succeeds, the dominant virus is likely HHV-6 or CMV (broader-spectrum ganciclovir covers both, while acyclovir is EBV/VZV/HSV-selective), OR the virus is in a tissue sanctuary site (CNS, DRG) that ganciclovir accesses better than acyclovir. If both valacyclovir and valganciclovir fail, and LSR is confirmed elevated, this is the strongest available negative evidence that the elevated LSR is NOT driven by ongoing viral replication — Mechanism 2 (LLPC autoimmune drift) becomes the leading explanation.
Falsifiable prediction: In a cohort of ME/CFS patients with elevated LSR, pre- and post-valacyclovir (6 months at therapeutic dose) LSR measurements will show: (a) decline in LSR only in patients who clinically respond; (b) no decline in LSR in non-responders. Falsified if LSR declines in non-responders (would suggest ALR continues despite absent clinical benefit — the antibody signal is more sensitive than symptom change) OR if LSR stays unchanged in responders (would suggest valacyclovir’s clinical benefit is independent of ALR suppression).
Consequence: If confirmed, an elevated LSR that does not budge after valacyclovir would strongly indicate that the antibody abnormality is LLPC-driven — directing treatment toward B cell-targeted interventions (currently research-stage only) rather than antiviral escalation. This would spare patients from years of empirically escalated antiviral therapy (valacyclovir → valganciclovir → repeat cycles) when the underlying mechanism is not viral. Origin: brainstorm — cascade trace.
Certainty: N/A — research question. The cascade-to-drug tracing above describes what valacyclovir response and non-response would mean for the LSR. But there is a third, more ambiguous pattern: valacyclovir produces clinical improvement but the LSR remains elevated. This would suggest viral reactivation is rate-limiting for symptoms but the LSR is maintained by a mechanism valacyclovir does not touch — either (a) LLPCs produce the lytic IgG baseline while SLPB-derived lytic IgG from ALR events is too small a fraction to move the ratio, or (b) the clinical benefit comes from suppressing a different herpesvirus (VZV, HSV) that is contributing to symptoms without affecting EBV LSR. This question cannot be answered without a prospective study measuring LSR, viral DNA (qPCR for each herpesvirus), and clinical symptoms simultaneously before, during, and after antiviral therapy.
Consequence: none — this is a methodological question about measurement synchronicity. The resolution depends on data that does not yet exist.
The LSR can function as a bridge between serological immunology and diagnostic pharmacology — the same ratio that resolves the contradictory herpesvirus serology literature (Lytic-to-Structural IgG Ratio (LSR) as a Diagnostic Biomarker) also yields differential predictions about antiviral drug responses (What Valacyclovir Response Tells Us About the Mechanism Driving Elevated LSR). If validated, a single pre-treatment blood test would do double duty: it would stratify patients by mechanism (ALR-driven vs LLPC-driven), AND predict whether valacyclovir is likely to provide benefit. This closes a gap in the current clinical algorithm: the LSR provides objective pre-treatment stratification, replacing the “time-limited trial with stopping rules” heuristic with a biomarker-informed decision. An elevated LSR that declines with valacyclovir confirms ALR at the individual-patient level without requiring qPCR (which may be negative during ALR). An elevated LSR that does not budge after valacyclovir provides strong negative evidence for LLPC-driven antibody abnormality, directing treatment away from antiviral escalation and toward B cell-targeted research. The LSR validation study (Lytic-to-Structural IgG Ratio (LSR) Diagnostic Biomarker Validation) is the most direct path from this convergent hypothesis to clinical deployment — it requires no new assay development and could be executed on existing clinical ELISA platforms within 2–3 years. The central unanswered question is whether LSR elevation exists at all in ME/CFS: the ratio has never been computed in any disease population, and the normal range is entirely unknown. Until the first measurement is made, the entire convergent framework — serological interpretation, pharmacodiagnostic inference, and mechanism-based treatment stratification — is a reasoning scaffold awaiting its first data point.
Consequence: If the LSR is validated, a single inexpensive blood test could simultaneously diagnose a mechanism AND predict an antiviral response AND guide treatment decisions — replacing the current clinical reality of diagnosing “viral reactivation” from structural antigen IgG alone and empirically cycling through antivirals without knowing whether they will work. This is the clinical promise of the LSR framework, but it requires the foundational step no one has taken: measuring the ratio for the first time.
1.9 Extending the LSR to HSV-1/2
The LSR framework was developed for EBV, but its logic — compare lytic-cycle antibody output to structural-antigen LLPC baseline — should in principle apply to any herpesvirus with a defined lytic gene cascade. HSV-1/2 offer a unique cross-viral validation opportunity: HSV reactivation produces visible lesions, providing an external clinical anchor against which the serological ratio can be calibrated.
Certainty: N/A — research question with zero direct evidence. The LSR has never been measured in any disease population. This question requires answering two prior unknowns: (1) does the LSR discriminate ME/CFS from controls (the foundational LSR study, Lytic-to-Structural IgG Ratio (LSR) Diagnostic Biomarker Validation), and (2) do HSV outbreaks cluster with PEM (the HSV dormancy-undormancy probe framework, Viral Encephalitis as Mechanistic Template for Post-Infectious ME/CFS). Only if both are affirmative does the cross-viral intersection become testable. If validated, the implication would be that the LSR reflects systemic immune dysregulation affecting multiple herpesvirus reservoirs — not just EBV — and that visible HSV outbreaks are the peripheral readout of a process operating silently in B cells (EBV) and T cells (HHV-6). If the LSR is elevated but HSV outbreak patterns are unremarkable, the LSR reflects an EBV-specific B cell compartment process. If HSV outbreak patterns are PEM-correlated but the LSR is normal, viral reactivation is compartmentalized (trigeminal ganglion-specific). Origin: brainstorm.
Consequence: The combination of a blood test (LSR) and a patient diary (HSV outbreak frequency) could triangulate viral reactivation across two distinct latency compartments without invasive tissue sampling. But this requires two prior validations that have not been done.
Falsifiable prediction (conditional): In a nested analysis within the LSR validation study, adjusted for PEM frequency, disease severity, and perceived stress: HSV-1 seropositive patients with elevated EBV-LSR will report higher HSV outbreak frequency than those with normal LSR. Falsified if LSR and HSV outbreak frequency are independent after adjusting for severity and stress confounders. Concordance between elevated LSR and frequent HSV outbreaks is equally consistent with confounding by illness severity and cumulative stress exposure — sicker patients have more PEM, more distress, and plausibly more of both readouts — and this must be ruled out before interpreting concordance as mechanistic. Critique: structurally unfalsifiable pending two prior validations — (1) LSR discrimination of ME/CFS from controls, and (2) HSV outbreak–PEM temporal clustering. This is a hypothesis about hypotheses.
The EBV LSR (anti-BZLF1 / anti-VCA-p18) is achievable with existing clinical ELISA platforms. An HSV analog (anti-ICP0 / anti-gB or anti-gD) would require anti-ICP0 IgG ELISA development — ICP0 is not a clinical antigen, has never been measured in serum, and its immunogenicity profile in natural infection is unknown. The HSV outbreak diary approach is the pragmatic near-term alternative. Consequence: none — methodological observation about feasibility differences. Origin: brainstorm.
1.10 Abortive Lytic Replication and dUTPase
A key challenge in ME/CFS virology is that standard viral load measurements frequently show no difference between patients and controls, despite serological evidence of immune activation against herpesviruses. Ariza et al. propose that abortive lytic replication (ALR) resolves this paradox (Ariza, Mena Palomo, and Williams 2025).
In ALR, herpesviruses initiate the lytic cycle but do not complete it: immediate-early and early lytic genes are expressed, producing viral proteins, but no infectious virions are assembled (Ariza, Mena Palomo, and Williams 2025). Because no new virions are produced, conventional viral load assays (qRT-PCR for viral DNA in plasma) remain negative, while the host immune system responds to the expressed viral proteins.
dUTPase as Immunomodulatory Driver Among the early gene products expressed during ALR, herpesvirus-encoded deoxyuridine triphosphate nucleotidohydrolases (d UTPases) have emerged as particularly significant (Ariza, Mena Palomo, and Williams 2025) (M. Palomo et al. 2026):
- dUTPases are conserved across EBV (BLLF3), HHV-6 (U45), and VZV (ORF8)
- These enzymes possess novel immunomodulatory and neuromodulatory functions beyond their catalytic role
- They can directly activate innate immune signaling, potentially driving chronic immune activation without productive infection
- Anti-dUTPase IgG antibodies are elevated in ME/CFS: across cohorts, 30.91% (single-sample analysis) to 52.7% (longitudinal sampling) of patients show simultaneous antibodies against multiple herpesvirus dUTPases versus 17.21% of controls (Ariza, Mena Palomo, and Williams 2025)
- Anti-dUTPase antibody levels correlate directly with fatigue and pain severity (M. Palomo et al. 2026)
- EBV dUTPase protein has been detected in postmortem ME/CFS brain tissue but not in non-ME/CFS controls, suggesting ALR occurs in the central nervous system (Ariza, Mena Palomo, and Williams 2025)
The dUTPase findings derive primarily from one research group (Ariza, Williams, and collaborators). While results are internally consistent across multiple cohorts and include longitudinal data, independent replication by other groups is needed. The postmortem brain tissue finding is from a small sample. Additionally, elevated anti-dUTPase antibodies establish an association with ME/CFS but do not prove that ALR-derived dUTPase is causally driving symptoms rather than serving as a marker of broader immune dysregulation.
1.11 Poly-Herpesvirus Co-Reactivation
Rather than a single herpesvirus driving ME/CFS, accumulating evidence points to simultaneous reactivation of multiple herpesviruses as a hallmark of post-infectious ME/CFS (Ariza, Mena Palomo, and Williams 2025) (M. Palomo et al. 2026):
- 72.5% of ME/CFS patients co-express antibodies to EBV, HHV-6, and VZV simultaneously, versus 31% of controls (M. Palomo et al. 2026)
- This pattern suggests systemic immune surveillance failure rather than reactivation of a single pathogen
- The proposed cascade: initial herpesvirus undergoes ALR → chronic dUTPase production drives T cell exhaustion → CD8+ T cell dysfunction permits sequential reactivation of additional herpesviruses
- Disease severity may correlate with the number of co-reactivated viruses and the duration of T cell exhaustion
Certainty: 0.35. The poly-herpesvirus ALR model is mechanistically coherent and consistent with the dUTPase antibody data (multiple cohorts showing 30–72% co-reactivation in ME/CFS versus 17–31% in controls). However, the hypothesis has not been directly tested: no study has prospectively tracked the temporal sequence of herpesvirus reactivation events, and the proposed T cell exhaustion cascade remains inferred from cross-sectional serological data. Independent replication of the dUTPase findings is pending.
If poly-herpesvirus ALR is a primary driver of post-infectious ME/CFS, then: (1) patients with antibodies against more herpesvirus dUTPases should have more severe symptoms; (2) single-agent antiviral therapy should produce only partial benefit, while broad-spectrum or combination antiviral approaches should be more effective; (3) prospective longitudinal sampling should reveal sequential herpesvirus reactivation events following the index infection; (4) dUTPase-targeted interventions should reduce immune activation markers independently of viral load.
Treatment implications: This model predicts that antiviral selection should target all reactivated herpesviruses rather than a single agent. It also suggests that dUTPase itself could be a therapeutic target, and that immune checkpoint modulation to reverse T cell exhaustion might break the reactivation cycle.
Limitations: The ALR concept, while explaining the negative viral load paradox, is difficult to measure directly in clinical settings. The proposed temporal cascade (initial ALR → dUTPase → T cell exhaustion → further reactivation) has not been captured in longitudinal data. The framework currently applies primarily to post-infectious ME/CFS; its relevance to non-infectious-onset cases is unknown.
1.12 Reactivation Patterns and Causal Relationships
The relationship between herpesvirus reactivation and ME/CFS immune dysfunction remains incompletely understood. Three mechanistic hypotheses can be distinguished by their testable predictions:
Certainty: 0.45. Consistent with documented NK cell dysfunction impairing viral clearance; however, the directionality of causation between NK dysfunction and viral reactivation has not been experimentally established in ME/CFS.
If reactivation is primarily a consequence of impaired immune control (particularly NK cell dysfunction), then: (1) improving NK cell function should reduce viral titers without affecting other ME/CFS symptoms; (2) viral reactivation markers should correlate with NK cell cytotoxicity but not independently predict symptom severity; (3) antiviral therapy alone should have minimal clinical benefit.
Certainty: 0.30. Limited antiviral trial data (Lerner, Montoya) show some benefit in subgroups, but controlled trials have not consistently demonstrated that viral suppression produces sustained clinical improvement across the ME/CFS population, lowering confidence in this model as a universal driver.
If reactivation is a primary driver of ongoing immune activation, then: (1) antiviral therapy should reduce both viral titers and immune activation markers (cytokines, immune cell activation); (2) viral load should independently predict symptom severity after controlling for immune markers; (3) successful viral suppression should produce sustained clinical improvement.
Certainty: 0.50. This model is the most mechanistically plausible given the documented bidirectional interactions between immune dysfunction and viral reactivation; it is consistent with the partial and heterogeneous response to antiviral monotherapy, though direct experimental evidence for a self-sustaining cycle in ME/CFS remains limited.
If reactivation and immune dysfunction form a self-sustaining cycle, then: (1) interventions targeting either viral replication or immune dysfunction should produce partial but incomplete benefit; (2) combined antiviral and immune-modulating therapy should be synergistic; (3) breaking the cycle at any point should eventually normalize both viral titers and immune function, though with temporal lag.
Current evidence does not definitively distinguish these mechanisms, though the limited efficacy of antiviral monotherapy in most ME/CFS patients suggests reactivation is unlikely to be solely causal. Longitudinal studies tracking viral titers, immune markers, and symptom severity following targeted interventions are needed to resolve this question.
While herpesvirus reactivation is documented in a subset of ME/CFS patients, no study has established the causal direction. Elevated viral titers and reactivation markers could be: (a) a cause of immune dysfunction driving symptoms, (b) a consequence of pre-existing NK cell/T cell dysfunction permitting reactivation, or (c) an epiphenomenon of general immune dysregulation with no independent causal role. The partial and inconsistent response to antiviral monotherapy does not resolve this ambiguity. Longitudinal intervention studies with simultaneous tracking of viral load, immune function, and symptoms are required.
1.13 Long COVID Evidence on EBV: Null and Epitope-Specific Findings
Long COVID provides a controlled, post-infectious setting in which to test the EBV-reactivation model, and the results temper a causal interpretation. In a prospective cohort of SARS-CoV-2-seropositive blood donors with asymptomatic or mild COVID, EBV was not reactivated in individuals reporting post-COVID syndrome (Salzburg cohort) (Hoeggerl et al. 2023). A retrospective peptide-microarray study of post-COVID syndrome found that overall EBNA-1 IgG binding was similar to convalescents, but that patients showed epitope-specific enhanced reactivity — to the glycine-alanine repeat region and a central domain distinct from the multiple-sclerosis-associated epitope — with no increase in autoantibodies (Lorenz et al. 2026). Together these findings suggest that EBV-associated antibody elevation in post-infectious illness is (a) not invariably a marker of active reactivation, and (b) epitope-specific rather than a generalised reactivation or autoimmunity signal. The caveat is that both cohorts are mild long-COVID populations; reactivation may be more prominent in the severe, ME/CFS-phenotype subgroup that the paper’s dUTPase and co-reactivation findings describe. The long-COVID nulls therefore argue for caution, not for discarding the reactivation model: they support a severity- and epitope-contingent framing rather than a universal causal EBV-reactivation claim.
2 Other Implicated Viruses
2.1 Enteroviruses
Enteroviruses (Coxsackieviruses, Echoviruses) have been implicated:
- Detection of viral RNA in muscle and gut biopsies
- Elevated antibodies in some patients
- Possible persistent low-level infection
- Historical associations with epidemic ME/CFS outbreaks
2.2 Parvovirus B19
Parvovirus B19 can cause chronic arthritis and fatigue:
- Associated with ME/CFS onset in some patients
- Viral DNA detectable in tissues years after infection
- May persist in bone marrow and synovium
2.3 SARS-CoV-2 and Long COVID
The COVID-19 pandemic highlighted viral triggers for ME/CFS-like illness:
- Long COVID (post-acute sequelae of SARS-CoV-2 infection, PASC) shares many features with ME/CFS
- Viral persistence documented in some patients
- Similar immune abnormalities observed
- Provides opportunity to study post-infectious ME/CFS from known onset
Persistent Immune Activation Independent of Viral Replication A proteomic study of 92 non-hospitalized PASC patients (mean 34 months post-infection) versus 73 matched recovered controls identified 26 differentially expressed plasma proteins—23 upregulated and 3 downregulated—revealing a distinct low-grade inflammatory signature (Fineschi et al. 2026). The most significantly elevated proteins included oncostatin M (OSM), IL-1 receptor antagonist (IL-1RN), IL-6, IL-12B, IL-2, CCL22, CSF3, CSF1, and HLA-DRA. Gene set enrichment analysis identified five activated immune pathways: inflammatory response, TNF-\(\alpha\)/NF-\(\kappa\)B signaling, IL-6/JAK/STAT3, IL-2/STAT5, and allograft rejection.
Critically, plasma spike protein levels did not differ between PASC patients and controls, indicating that this persistent immune activation occurs independently of ongoing viral replication (Fineschi et al. 2026). This shares a conceptual parallel with the herpesvirus abortive lytic replication model (Section HSV-LSR Analog Requires De Novo Assay Development): in both cases, chronic immune activation persists without evidence of productive viral infection. The mechanisms differ—ALR involves ongoing viral protein expression from incompletely reactivated herpesviruses, whereas the PASC finding may reflect a self-sustaining immune dysregulation initiated but no longer maintained by viral activity—but the clinical implication is the same: viral load negativity does not exclude virus-initiated pathology.
The activated pathways overlap substantially with those documented in ME/CFS: TNF-\(\alpha\) elevation (Section Cytokines and Inflammatory Mediators), IL-6/JAK/STAT3 signaling, and IL-2/STAT5 dysregulation (see Hypothesis IL-2 Pathway in ME/CFS Pathophysiology). This convergence supports the hypothesis that post-infectious ME/CFS and PASC share a common immunological endpoint regardless of the triggering pathogen. The study found no evidence for distinct biological subtypes within PASC, suggesting a unitary inflammatory state rather than the heterogeneous subtype structure sometimes proposed.
Using total-body PET-CT imaging and rectosigmoid biopsies, Peluso et al. (Peluso et al. 2024) detected SARS-CoV-2 spike-encoding RNA—both single- and double-stranded—in gut lamina propria tissue up to 676 days after initial infection. Double-stranded RNA present in tissue would constitutively activate innate immune sensors (RIG-I, MDA5), generating a perpetual interferon signature and sustaining T cell activation without active viral replication. Total-body PET imaging revealed widespread lymphoid tissue involvement, supporting the gastrointestinal tract as a viral reservoir maintaining systemic immune dysregulation in long COVID and post-COVID ME/CFS. Study: (tissue biopsy + PET imaging, \(n=5\) biopsy participants, Science Translational Medicine, certainty: 0.70, consistent with multiple independent viral persistence studies).
SARS-CoV-2 PKR-ISR Antagonism and Long COVID Implications
SARS-CoV-2 actively suppresses the integrated stress response (ISR) during active infection: the nucleocapsid protein domain N2b sequesters double-stranded RNA (dsRNA), preventing PKR (protein kinase R) and RIG-I receptor activation, abolishing stress granule formation and ISR-induced translational arrest . This ISR suppression is a fundamental viral immune evasion strategy conserved across coronavirus species. In the context of viral persistence (documented in gut tissue by Peluso et al. above), ongoing N2b-mediated PKR suppression could chronically disable the ISR response that would otherwise limit viral spread and trigger cellular repair programs.
Variant-specific differences in ISR antagonism may partly explain differential long COVID rates. Christ et al. found that Omicron nucleocapsid protein less efficiently suppresses stress granule formation than ancestral or Delta variants — Omicron-infected cells formed stress granules in ~60% of cells versus ~5% for ancestral SARS-CoV-2 (Christ et al. 2023). If higher ISR activation during acute Omicron infection reduces viral establishment in tissue reservoirs, this could contribute to the lower long COVID incidence observed after Omicron waves, though this connection remains speculative and unmeasured.
Certainty: 0.40. Different ME/CFS subgroups may have opposite ISR states depending on their triggering pathogen, explaining heterogeneous trial responses (Hennig et al. 2022) (P. Wang et al. 2023). Post-herpesviruses (HHV-6, EBV) subgroup: ISR chronically activated — PERK elevated, ATF4 active, WASF3 upregulated, mitochondrial supercomplex dysfunction. Post-SARS-CoV-2 (long COVID) subgroup: ISR suppressed during acute phase via N2b-mediated PKR inhibition , then dysregulated in recovery — potentially a rebound state or permanently impaired ISR dynamics.
This bidirectional model predicts that ISR suppression (ISRIB, eIF2B activators) will benefit the herpesvirus-triggered subgroup (excessive ISR) but harm or provide no benefit to the post-SARS-CoV-2 subgroup (already suppressed ISR). Conversely, ISR-prolonging agents (salubrinal, guanabenz-class GADD34 inhibitors) would be predicted to benefit only the ISR-suppressed subgroup (Costa-Mattioli and Walter 2020). The absence of pathogen-stratified subgrouping in ME/CFS clinical trials means this divergence has not been tested.
Testable prediction: Stratifying ME/CFS patients by triggering pathogen (HHV-6 dUTPase IgG positive vs. post-COVID onset) will reveal opposite directions of phospho-eIF2α and ATF4 expression in skeletal muscle biopsy. In a crossover trial, ISRIB will improve fatigue in the HHV-6+ subgroup while worsening or showing no effect in the post-COVID subgroup. If ISR markers are uncorrelated with trigger → hypothesis falsified.
Limitation: Most ME/CFS patients cannot identify their triggering pathogen with certainty. Pathogen-stratified muscle biopsy studies do not yet exist. The ISR activation finding (Wang 2023) is from an NIH cohort with mixed trigger history (n=14); the SARS-CoV-2 ISR suppression finding is from cell models, not patient tissue .
Certainty: 0.55. ISR activation in ME/CFS is restricted to specific tissue compartments — skeletal muscle (P. Wang et al. 2023), CNS tissue (choroid plexus, hippocampus, amygdala) , possibly dorsal root ganglion sensory neurons and endothelium — and is not detectable in peripheral blood mononuclear cells. The null result from Omdal et al. (n=48, peripheral blood stress biomarkers, long COVID) is therefore not a refutation of ISR involvement but evidence of compartmentalization. This predicts that PBMC-based ME/CFS biomarker studies systematically underestimate ISR involvement, and explains the consistent failure of blood-based biomarker panels to discriminate ME/CFS from healthy controls with high sensitivity.
The tissue specificity is mechanistically plausible: PERK is preferentially activated in cells with high secretory/biosynthetic demand (muscle, neurons, endothelium) rather than circulating immune cells. HHV-6 neuroinvasion places viral ISR triggers directly in CNS tissue. Muscle WASF3 elevation (P. Wang et al. 2023) confirms the skeletal muscle as a primary affected compartment.
Testable prediction: Paired sampling in the same ME/CFS patients (PBMC + skeletal muscle biopsy + skin biopsy [DRG-adjacent fibers] + saliva [salivary glands as accessible secretory tissue]) will show ISR markers elevated in muscle and skin but not PBMC. Effect sizes in tissue will substantially exceed those in blood. If muscle and PBMC ISR markers correlate strongly → compartmentalization hypothesis falsified.
Limitation: Only skeletal muscle (Wang 2023, n=14) and CNS post-mortem (Kasimir 2022, n=3) have been examined. No skin biopsy ISR study exists in ME/CFS. The comparison between tissue compartments in the same patients has not been done.
Certainty: 0.25. ME/CFS affects females at approximately four times the rate of males. Estrogen modulates the unfolded protein response and ER stress pathways: 17β-estradiol influences PERK sensitivity through estrogen receptor interactions at the ER membrane and XBP1 splicing efficiency via ERα-mediated co-regulation (Costa-Mattioli and Walter 2020). Female cells may therefore have a lower threshold for ISR activation following viral insult, predisposing to chronic ATF4 persistence after the same viral trigger that would produce only a transient ISR response in male cells.
This estrogen-modulated ISR threshold model predicts: (a) in premenopausal females, symptom severity fluctuates with estradiol levels — phases of lower estradiol (follicular nadir, premenstrual) may shift the ISR threshold lower, triggering episodes of enhanced ATF4 activity and worsened PEM; (b) postmenopausal women initiating hormone replacement therapy may show altered ISR dynamics in muscle; (c) males who develop ME/CFS may show a distinct ISR signature (shifted toward the GCN2 arm rather than the PERK arm) compared with females with equivalent disease severity.
Testable prediction: Sex-stratified phospho-eIF2α and ATF4 levels in ME/CFS skeletal muscle will show greater elevation in females than in age- and severity-matched males. In premenopausal females, serial measurement will show cycle-phase correlation with ATF4-target metabolites (GDF15, FGF21) in plasma. If ISR markers show no sex difference at matched severity → hypothesis falsified.
Limitation: No sex-stratified ISR study exists in ME/CFS muscle. The estradiol–PERK interaction is established in vitro and in animal models; its operation in human skeletal muscle under chronic viral stress has not been tested. Sex-ratio differences in ME/CFS may arise from multiple non-ISR mechanisms (immune susceptibility, health-seeking behavior, diagnostic criteria sensitivity). (Not yet replicated.)
Certainty: 0.20 — mechanistic inference from non-MCAS mast cell literature; no direct ISR measurement in MCAS patient samples.
Approximately 25% of ME/CFS patients have concurrent MCAS . The mechanistic relationship between MCAS and the ISR pathway is bidirectional and mutually reinforcing in theory, though not yet directly tested.
ISR acting on mast cells: Mast cells are secretorily demanding cells whose function depends on intact unfolded protein response (UPR) machinery (Wilhelm et al. 2017). In mast cells, the eIF2α-phosphorylation arm of ISR suppresses XBP1 (a key UPR effector) (Fan and Zhang 2024), which would reduce IgE-mediated degranulation efficiency. Simultaneously, chronic eIF2α-P would suppress the mTORC1-driven protein synthesis required for granule reconstitution after degranulation (Iskarpatyoti et al. 2022). The expected consequence: granule compositional changes over time and altered degranulation dynamics — potentially explaining why MCAS in ME/CFS can present differently from classical allergic MCAS.
Mast cells acting on ISR: Mast cell degranulation requires ER calcium mobilization (STIM1/ORAI1 SOCE). Chronic repeated degranulation, as in MCAS, produces sustained low-level ER calcium depletion. PERK is the canonical ER stress/calcium sensor; the PERK–eIF2α–ATF4 axis is functional in mast cells (S. Wang and Zhang 2017). However, a key constraint from Soboloff and Berger (2002) (Soboloff and Berger 2002): eIF2α phosphorylation in mast cells requires severe sustained ER calcium depletion, not moderate single-event depletion. Single degranulation events are unlikely to trigger significant ISR; only the cumulative chronic pattern of MCAS-pattern degranulation is a plausible trigger.
Therapeutic implication (speculative): In the MCAS subgroup, mast cell stabilizers (ketotifen, cromolyn, Section Multi-Domain Co-Occurrence Model) that reduce degranulation frequency could secondarily reduce ER calcium stress, attenuating PERK input and potentially easing the ISR burden. This would predict that MCAS-targeted treatment in the ME/CFS MCAS subgroup produces ISR biomarker improvement (GDF15 reduction, ATF4 target metabolite normalization) in addition to symptom improvement — a testable pharmacodynamic endpoint.
Falsifiable prediction: ME/CFS patients with MCAS overlap (≥2 MCAS criteria) will show higher ATF4-target gene expression (ASNS, CHOP, SLC7A11) in skin mast cells at 3mm punch biopsy compared to ME/CFS patients without MCAS. eIF2α-P will be elevated in PBMC-derived in vitro mast cell differentiation assays from ME/CFS+MCAS patients compared to ME/CFS-only controls. If ISR markers show no MCAS-dependent difference → bidirectional crosstalk hypothesis not supported.
Limitation: No study has measured ISR markers in MCAS patient samples of any kind. The mechanistic chain is assembled from non-MCAS mast cell models and non-mast cell ISR studies. HRI kinase data in mast cells are absent. The IgE-suppression finding from Fan 2024 may not generalize to non-IgE MCAS triggers (MRGPRX2, complement, substance P). The gain-of-function direction (MCAS → ISR) requires repeated/chronic mast cell activation, not single acute events. (Not yet replicated in ME/CFS or MCAS.)
A multi-institutional Lancet Infectious Diseases position paper (Proal et al. 2025) by leading long COVID and ME/CFS researchers established the conceptual and trial-design framework for treating long COVID as a viral reservoir disease. The authors synthesized evidence that SARS-CoV-2 persists for months to years in tissue reservoirs in affected individuals, propose that this persistence drives ongoing immune activation and symptoms, and identify critical design considerations for reservoir-targeting trials: mechanism of action of candidate therapeutics, participant selection by reservoir biomarker burden, treatment duration, standardized measurables, and potential combination approaches. The failure of a 15-day nirmatrelvir/ritonavir RCT to improve symptoms is reinterpreted as evidence that active replication is not the primary target—stable viral reservoirs require a different therapeutic strategy, drawing lessons from HIV and hepatitis C functional cure models. Study: (consensus review, Lancet Infectious Diseases 2025, certainty: 0.75).
The Fineschi et al. study (Fineschi et al. 2026) was restricted to non-hospitalized cases and included participants infected across different pandemic waves (pre-Omicron unvaccinated and Omicron vaccinated), introducing heterogeneity. The cross-sectional design at a single time point cannot capture the temporal evolution of immune activation documented in ME/CFS (see Section Cytokines and Inflammatory Mediators). Additionally, the absence of an ME/CFS comparator arm limits direct mechanistic comparison, though the study originated from the ME/CFS Collaborative Research Centre at Uppsala University, suggesting awareness of this overlap.
A multicenter case–control study by Briese et al. (Briese et al. 2023) screened blood, feces, and saliva of ME/CFS patients using polymerase chain reaction and high-throughput sequencing (391 ME/CFS patients, 292 controls). The study found no consistent group-specific differences in viral nucleic acid between cases and healthy controls. Its one significant finding was a lower prevalence of anelloviruses in cases (30.5% vs 54.1%, \(p < 0.001\); see Constitutive Immune Activation Without Secretory Output). The authors recommend that future investigations into viral infection in ME/CFS focus on adaptive immune responses rather than surveillance for viral gene products — a recommendation, not a measured outcome.
This null result constrains the productive-replication arm of the viral-persistence model for the non-tissue compartments sampled (blood, feces, saliva): it argues against an ongoing, actively replicating, nucleic-acid-detectable viral infection in the periphery as the proximate driver of ME/CFS. It does not bear on tissue-resident virus (gut lamina propria, CNS, dorsal root ganglia) or on abortive lytic replication — both of which the paper addresses elsewhere (Section HSV-LSR Analog Requires De Novo Assay Development; Peluso et al. detected spike RNA in gut tissue by biopsy, not blood surveillance). It is consistent with the PASC finding that persistent immune activation can occur independently of ongoing peripheral viral replication (Section above).
Consequence: This finding narrows where peripheral viral studies should look in ME/CFS — viral-gene-product surveillance of blood/feces/saliva does not identify an ongoing productive infection. It does not disprove a viral trigger or a tissue-resident or abortive-lytic reservoir, which is precisely why tissue biopsy, not surveillance, remains the lever.
Severity applicability: unknown — study population not stratified by severity in the detail available to this integration. Evidence source: human ME/CFS cohort (multicenter, US). Diagnostic criteria for cited study: IOM.
Certainty: 0.25.
Two studies (one read at abstract level) are jointly compatible with — but do not establish — the interpretation that anellovirus burden reflects immune-dysregulation state rather than productive viral infection in ME/CFS and its post-infectious relatives. Briese et al. (Briese et al. 2023) found a lower prevalence of anelloviruses in ME/CFS cases than healthy controls (30.5% vs 54.1%, \(p < 0.001\)). Maguire et al. (Maguire et al. 2026) (IMPACC, \(n=1154\)) found elevated Anelloviridae transcript reads in the long-COVID physical-disability patient-reported-outcome group—characterized by fatigue, cognitive dysfunction, and post-exertional malaise overlapping ME/CFS symptoms—after controlling for age, sex, and acute COVID-19 severity, and noted that Anelloviridae had previously been linked to chronic fatigue syndrome and multiple sclerosis.
Why this must remain a speculation, not a finding: Neither study measured anellovirus burden and immune markers in the same ME/CFS patients. Briese reported virome only (no paired immune data); Maguire’s immune/metabolomic associations are within a hospitalized COVID-19 cohort (acute-phase immunosuppression and steroid exposure are an uneliminated confound). The immune-state-marker reading is therefore an inference requiring a harmonized quantitative TTV assay plus paired immune phenotyping in one ME/CFS cohort.
Directional reading from the existing sibling hypothesis: The document already interprets the identical Briese depletion directionally (Section Constitutive Immune Activation Without Secretory Output): constitutive innate-immune hyperactivation clearing commensal viruses would predict low TTV, whereas adaptive-immune exhaustion would predict high TTV. This directional map (innate-hyperactivation-dominant → low; adaptive-exhaustion-dominant → high) is the testable form of the hypothesis; the present environment holds the non-directional form and defers to Constitutive Immune Activation Without Secretory Output for the mechanistic commitment.
Critique — the direction difference may be a method artifact: The apparent opposite directions across the two cohorts must be treated cautiously, because the studies measured different quantities: Briese reported binary PCR-based prevalence, whereas Maguire reported abundance of transcriptomic sequencing reads. Until a single harmonized assay is applied to ME/CFS, long-COVID, and healthy controls within one study, the divergence could be a synthesis artifact. A second caution: anellovirus burden may be a redundant bystander that adds little beyond directly measured immune markers.
Testable prediction: Using the directional map above: a harmonized TTV assay in one cohort will show that innate-hyperactivation-dominant subgroups have lower TTV and adaptive-exhaustion-dominant subgroups higher TTV than controls, and that TTV tracks an immune marker rather than symptom severity alone. If TTV burden is a stable, immune-independent, disease-defining feature across all subgroups, the hypothesis is falsified; if harmonization eliminates the ME/CFS-versus-long-COVID direction difference (both non-significant), the “direction differs by cohort” reading is rejected.
Consequence: If confirmed — and only if anellovirus burden adds incremental value beyond directly measured immune markers, which is untested — it could provide an inexpensive immune-state biomarker to stratify and monitor ME/CFS and long-COVID patients. This is early-stage and may turn out to be a dead end.
Severity applicability: unknown — Briese cohort not stratified in available detail; Maguire is a hospitalized COVID-19 cohort. Evidence source: human ME/CFS cohort (Briese) + human long-COVID cohort (Maguire). Translation gap: long-COVID → ME/CFS (post-infectious cousin, population-discounted). Competing explanations: method incommensurability (binary PCR prevalence vs transcript abundance); redundant-bystander (adds little beyond direct immune markers); sequencing-platform or PCR primer bias; sampling compartment; medication effects; hospitalization and acute-phase immunosuppression/steroid exposure (Maguire); disease-stage differences; cohort-specific immune composition; post-hoc subgroup/multiple-testing concerns in Maguire; causal direction unestablished (immune \(arrow\) anellovirus, not reverse). Diagnostic criteria for cited studies: IOM (both).
Two studies are jointly compatible with — but do not establish — the interpretation that productive viral infection is not a detectable, group-defining feature of ME/CFS in the peripheral compartments sampled, and that anellovirus burden may reflect immune state rather than viral cause. Briese et al. (Briese et al. 2023) found no group-specific viral-nucleic-acid differences in blood, feces, and saliva of ME/CFS, with lower anelloviruses as the one significant finding (No Productive Viral Infection Detected by Nucleic-Acid Surveillance of Non-Tissue Compartments in ME/CFS); Maguire et al. (Maguire et al. 2026) found Anelloviridae associated with the long-COVID physical-disability group, consistent with (though not proving) an immune-dysregulation reading (Anellovirus Burden as an Immune-State Signature Rather Than a Viral Cause in ME/CFS). Together they argue that peripheral viral-gene-product surveillance is not a productive lever in ME/CFS and that the adaptive immune response merits greater research focus.
What the evidence supports: peripheral (blood/feces/saliva) viral-gene-product surveillance does not identify ongoing productive infection in ME/CFS, and the null does not bear on tissue-resident or abortive-lytic virus. What remains open: whether anellovirus burden is a genuine immune-state marker at all, whether the apparent directional difference between cohorts is biological or a method artifact of incommensurate assays, and whether it adds predictive value beyond directly measured immune markers. The decisive next step — a single harmonized quantitative TTV assay applied to ME/CFS, long-COVID, and healthy controls within one cohort, with paired immune phenotyping — will resolve these open questions.
Consequence: If this reading survives harmonization, it redirects the viral research agenda in ME/CFS away from “is a virus replicating in the periphery?” toward “how is the host immune response dysregulated?” and could provide an inexpensive biomarker of immune state for stratification and treatment monitoring. Until the method-artifact caveat is resolved, the anellovirus angle could collapse once measured consistently across cohorts.
Cross-reference: See Constitutive Immune Activation Without Secretory Output for the directional (innate-hyperactivation) reading of the Briese depletion; the present synthesis holds the weaker, non-directional form.
3 Tick-Borne Infections
Tick-borne infections represent an important and often underdiagnosed trigger for ME/CFS-like illness. The clinical overlap between post-treatment Lyme disease syndrome (PTLDS), ME/CFS, and chronic tick-borne infections creates significant diagnostic and therapeutic challenges.
3.1 Lyme Disease and Post-Treatment Lyme Disease Syndrome
Acute Lyme Disease. Lyme disease, caused by Borrelia burgdorferi (North America) or Borrelia afzelii/garinii (Europe), is transmitted by Ixodes ticks and represents the most common vector-borne infection in temperate regions (Paul M. Lantos et al. 2021):
- Incidence: \(>\) 470,000 cases annually in the United States (Paul M. Lantos et al. 2021)
- Geographic expansion: Endemic areas expanding due to climate change and deer population increases
- Characteristic presentation: Erythema migrans (bulls-eye rash) in 70–80% of cases; flu-like illness, arthralgia, neurological symptoms
- Standard treatment: 2–4 weeks of oral doxycycline or amoxicillin for early localized disease
Post-Treatment Lyme Disease Syndrome (PTLDS). Approximately 10–20% of patients treated for Lyme disease develop persistent symptoms despite standard antibiotic therapy (Rebman et al. 2023):
- Defining features: Fatigue, cognitive dysfunction (“brain fog”), musculoskeletal pain persisting \(\geq\) 6 months post-treatment
- Symptom overlap with ME/CFS: 26 of 29 core ME/CFS symptoms are present in PTLDS patients (Rebman et al. 2023); however, the proportion meeting formal ME/CFS diagnostic criteria has not been systematically determined
- PEM consideration: Some PTLDS patients report post-exertional worsening, though this has not been systematically studied with ME/CFS-specific methodology
- Biomarker studies: Shared immune abnormalities including altered cytokine profiles and T cell exhaustion markers
Systematic comparison of symptom profiles between PTLDS and ME/CFS cohorts reveals striking overlap (Rebman et al. 2023). Of the 29 symptoms assessed using the DePaul Symptom Questionnaire, 26 (90%) showed comparable prevalence and severity between conditions. Both groups exhibited: fatigue (100% prevalence), unrefreshing sleep (\(>\) 90%), cognitive impairment (\(>\) 85%), post-exertional malaise (\(>\) 80%), and widespread pain (\(>\) 75%). This overlap suggests either shared pathophysiology or that PTLDS represents a subset of post-infectious ME/CFS. Mechanistic Hypotheses for Persistent Symptoms. Several mechanisms may explain symptom persistence after antibiotic treatment:
- Immune dysregulation: Persistent inflammation and autoimmunity triggered by infection; molecular mimicry between borrelial antigens and host tissues (Steere et al. 2016)
- Microbial persistence: Controversy exists regarding whether Borrelia can persist in tissue reservoirs (synovium, nervous system) despite negative blood tests; biofilm formation may protect organisms
- Tissue damage: Irreversible damage to neural, articular, or cardiac tissues during acute infection
- Microbiome disruption: Prolonged antibiotic courses may cause persistent gut dysbiosis contributing to symptom chronicity
- Microglial activation: B. burgdorferi directly activates human microglia through TLR1/TLR2/MyD88 pathways, inducing NF-\(\kappa\)B-driven release of TNF-\(\alpha\), IL-1\(\beta\), IL-6, and chemokines — and microglia are necessary intermediaries for B. burgdorferi-induced neuronal apoptosis in a transwell co-culture model, as the spirochete alone does not directly kill neurons in vitro (Myers, Kaushal, and Philipp 2009) (Parthasarathy and Philipp 2015) (Cassiani-Ingoni et al. 2006). Critically, non-viable B. burgdorferi remnants (antibiotic-killed debris) still activate microglia and sustain neuroinflammation ex vivo (Parthasarathy and Gadila 2022), providing a mechanistic bridge between treated infection and persistent neurological symptoms. In vivo human PET imaging provides evidence consistent with elevated glial activation years after antibiotic treatment in PTLDS patients — though TSPO-PET specificity limitations (detailed in Section Sleep EEG Delta/Alpha Ratio as a Non-Invasive Thalamic Calcium Proxy below) constrain this interpretation (Coughlin et al. 2018). Microglial activation is thus a convergent neuroimmunological endpoint for both viral (post-infectious) and bacterial (post-Lyme) triggers of ME/CFS-like illness — the paper’s existing viral microglial coverage (Sleep EEG Delta/Alpha Ratio as a Non-Invasive Thalamic Calcium Proxy) represents only half the post-infectious picture.
3.2 Bartonella Species
Bartonella species are intracellular bacteria transmitted by various vectors including ticks, fleas, lice, and sand flies.
Species and Transmission.
- Bartonella henselae: Cat scratch disease; cats are primary reservoir
- Bartonella quintana: Trench fever; transmitted by body lice
- Bartonella bacilliformis: Carrión’s disease; sand fly transmission in South America
- Tick transmission: Multiple Bartonella species have been identified in Ixodes ticks, suggesting co-transmission with Borrelia
Chronic Bartonellosis and ME/CFS-Like Symptoms. Chronic Bartonella infection can present with neuropsychiatric and systemic symptoms overlapping with ME/CFS (Edward B. Breitschwerdt et al. 2025a):
- Neurological: Encephalopathy, cognitive dysfunction, peripheral neuropathy, neuroretinitis
- Systemic: Chronic fatigue, lymphadenopathy, low-grade fever, sweats
- Dermatological: Striae-like lesions (characteristic “Bartonella striae”), papular eruptions
- Vascular: Endothelial dysfunction, vasculitis-like presentations
Breitschwerdt et al. used specialized enrichment culture techniques to detect Bartonella DNA in blood samples from patients with chronic fatigue and neurological symptoms (Edward B. Breitschwerdt et al. 2025a). Of patients tested, 26% were positive for Bartonella species DNA; the same cohort yielded 24% Babesia DNA positivity (see updated Babesia limitation below, (Edward B. Breitschwerdt et al. 2025b)). Without healthy control data in this report, the clinical significance of detection remains uncertain—Bartonella DNA may represent active infection, past exposure, or subclinical carriage. These findings require replication in larger cohorts with appropriate controls to determine whether detection rates exceed background prevalence. Diagnostic Challenges. Bartonella diagnosis is notoriously difficult:
- Serology limitations: Sensitivity 40–60%; cross-reactivity between species; seronegative chronic infection documented
- Culture requirements: Specialized enrichment culture (BAPGM) over 2–3 weeks; not widely available
- PCR sensitivity: Standard PCR may miss low-level bacteremia; requires specialized laboratories
- Clinical diagnosis: Often made on clinical grounds with therapeutic trial
3.3 Babesia Species
Babesia are intraerythrocytic parasites transmitted by Ixodes ticks, frequently co-transmitted with Borrelia.
Epidemiology and Presentation.
- Primary species: B. microti (North America), B. divergens (Europe), B. duncani (Western US)
- Clinical syndrome: Fever, hemolytic anemia, thrombocytopenia, splenomegaly; can be asymptomatic
- Chronic infection: May persist for months to years, particularly in immunocompromised hosts
- Co-infection impact: Babesiosis with concurrent Lyme disease produces more severe illness and longer symptom duration (Krause et al. 1996)
ME/CFS Relevance.
- Chronic fatigue: Persistent infection causes ongoing hemolysis, cytokine activation, and profound fatigue
- Co-infection complexity: Patients with ME/CFS-like symptoms after tick exposure may have undiagnosed Babesia as sole or co-pathogen
- Treatment complexity: Requires different antimicrobial regimen than Lyme disease; may explain antibiotic treatment failures
3.4 Other Tick-Borne Pathogens
Additional tick-borne infections may trigger or contribute to ME/CFS-like illness:
Anaplasmosis and Ehrlichiosis.
- Pathogens: Anaplasma phagocytophilum, Ehrlichia chaffeensis, E. ewingii
- Clinical features: Fever, leukopenia, thrombocytopenia, elevated transaminases
- Chronic sequelae: Less well-characterized than PTLDS, but persistent symptoms reported
Rickettsia Species.
- Rocky Mountain spotted fever (R. rickettsii), other spotted fever groups
- Can cause severe acute illness with potential for chronic neurological sequelae
Tick-Borne Relapsing Fever.
- Borrelia hermsii, B. turicatae, and related species
- Characterized by recurring febrile episodes
- May be confused with Lyme disease due to genus similarity
3.5 Symptom-Domain-Specific Findings (2024–2026)
Recent work has moved beyond the observation that tick-borne illness resembles ME/CFS toward mapping which pathogen mechanisms drive which symptom domains, and toward identifying why the same individual might develop chronic illness after a tick-borne or a viral trigger.
Fatigue domain — the domain that persists. A CDC claims-based analysis of 24,503 Lyme disease patients matched to 122,095 controls found that pain, cognitive, and fatigue codes occurred roughly 5% more frequently in Lyme patients, but that most excess symptom coding declined toward control levels by 6–12 months (Nawrocki et al. 2025). The exception was fatigue, which persisted beyond one year — the single symptom domain that did not normalize. This population-level result complements the frequently cited 10–20% PTLDS figure (which derives from smaller, referral-based symptomatic cohorts (Rebman et al. 2023)) with claims-based data: at the population scale the absolute excess risk is modest, yet the domain that lingers is precisely the one that defines ME/CFS.
The two estimates are not contradictory. The CDC figure measures ICD-coded symptom burden across an unselected Lyme population; the 10–20% PTLDS figure measures symptom-questionnaire-defined persistence in patients who present for ongoing care. They describe different denominators. Together they suggest that post-Lyme fatigue is common enough to matter at the population level and severe enough in a subset to meet chronic-illness thresholds.
Immune domain — T-cell signatures parallel ME/CFS.
Immunophenotyping of 272 post-treatment Lyme disease (PTLD) patients against 28 healthy controls (cross-sectional flow-cytometry study) identified reduced circulating CXCR5+ CD4+ naïve T cells (5.2% vs 8.3%) and expanded CXCR3+CCR4−CCR6− CD8 T cells (43.1% vs 25.7%); an elastic-net classifier separated PTLD from controls at AUC 0.83, though on only 28 controls and without an external validation set (Girgis et al. 2025). Cytokine panels alone did not distinguish the groups. Factor analysis linked specific immune-cell frequencies to specific symptom severities: a female-specific central-memory CD8 expansion tracked the high-fatigue subgroup, and CXCR5+ CD4+ naïve abundance tracked musculoskeletal pain.
Evidence type: observational cross-sectional immunophenotyping study (single-center, small control arm, no external classifier validation, no within-study ME/CFS comparator). Certainty: 0.55.
Consequence: This gives the post-Lyme fatigue overlap with ME/CFS a candidate immunological signature rather than only a symptom-checklist resemblance — if the same T-cell subset shifts replicate in ME/CFS cohorts, it would suggest different infectious triggers share an immune signature, and that a blood immunophenotype (not a cytokine panel) is the place to look for it. Replication in larger, externally validated cohorts is required before the signature can be relied upon.
Shared susceptibility — a genetic bridge between triggers.
An in-silico binding-affinity analysis reported that HLA alleles associated with ME/CFS susceptibility (C*07:04, DQB1*03:03) bind Borrelia burgdorferi antigens weakly, whereas ME/CFS-protective alleles (B*08:01, DPB1*02:01) bind strongly — the same pattern the authors report for herpesvirus and SARS-CoV-2 antigens (Georgopoulos, James, and Peterson 2025). (That the pattern recurs across unrelated pathogens is itself a caution: it could reflect a property of the prediction algorithm — these alleles being general weak binders in silico — rather than a shared biological mechanism, and needs orthogonal wet-lab validation.) The proposed mechanism: weak HLA presentation → inefficient antigen clearance → antigen persistence → chronic immune activation, offering — if confirmed functionally — one reason the same host could develop ME/CFS, PTLDS, or Long COVID depending on which pathogen arrives. Certainty: 0.30 (in-silico prediction only; no wet-lab or functional validation; few alleles tested; binding affinity does not guarantee a functional immune outcome).
Falsifiable prediction: In a cohort genotyped for these alleles, individuals carrying the “weak-binding” susceptibility alleles will show higher rates of chronic post-infectious illness after documented tick-borne or viral infection than carriers of the “strong-binding” protective alleles. Refuted if chronic-illness incidence is independent of HLA-binding class, or if measured (wet-lab) binding does not match the in-silico predictions.
Consequence: If validated, this would suggest HLA typing could contribute to risk stratification across post-infectious illnesses — but the evidence would still be a few alleles with predicted binding differences, not a validated screening tool, and it is not usable for individual risk prediction at present.
Autonomic domain — direct comparison confirms overlap. A head-to-head tilt-table study compared hemodynamic responses in late-stage Lyme disease, CFS, and post-COVID groups against controls; all three patient groups showed autonomic dysfunction, and the Lyme and CFS groups displayed similar hemodynamic patterns (Milovanovic et al. 2025). This provides the first controlled head-to-head comparison for the dysautonomia previously described in post-Lyme illness only through case series and uncontrolled reports (Strong Mechanistic Overlap (Tier 2)).
Neuro-Bartonella — an established neurotropic pathogen. Three 2024–2026 sources deepen the Bartonella–cognition link beyond the single chronic-fatigue culture study cited above (Edward B. Breitschwerdt et al. 2025a), and they contribute different kinds of evidence. A comprehensive review (Bush et al. 2024) synthesizes Bartonella as an established neurotropic genus producing cognitive impairment, neuropsychiatric symptoms, neuropathy, and movement disorders, proposing “neurobartonelloses” as a clinical entity — synthesis, not new primary data; a national retrospective cohort of B. henselae encephalitis (Yakubovsky et al. 2026) establishes CNS neurotropism and documents persistent cognitive sequelae, though from acute brain infection rather than chronic ME/CFS-like illness; and a small pilot (Guirguis et al. 2024) directly relevant to chronic cognition detected Bartonella DNA at higher rates in mild-cognitive-impairment patients than controls, though it was too small (n=16 vs 30) to resolve the unanswered question of Bartonella DNA prevalence in healthy populations.
Diagnostics — staging active from resolved infection. Two developments address the diagnostic limitations noted above. VlsE-specific antibody class and isotype can differentiate Lyme disease stage (early vs late vs post-treatment) (Nair et al. 2025), and paired CSF–serum epitope profiling identifies immunoreactive patterns that distinguish PTLDS from resolved Lyme (Marques et al. 2026). Both offer a potential route to separate patients with ongoing antigenic stimulation from those with resolved infection — the distinction at the heart of the chronic-Lyme controversy — but both require validation in the disputed populations.
Treatment — a non-antibiotic probe.
An open-label pilot (n=15) tested disulfiram, an anti-persister agent acting outside the antibiotic framework, in patients with persistent symptoms after prior antibiotics; some improved, but significant adverse effects including neuropathy occurred (Kuvaldina et al. 2025). Certainty: 0.35 (open-label, no control arm, small n). This is not a treatment recommendation: there is no controlled evidence of efficacy, the harm signal (neuropathy) is material, and neuropathy would be especially difficult to distinguish from — and could worsen — the small-fiber neuropathy already documented in this population.
Consequence: For patients, this means disulfiram is not a validated option and carries a real risk of nerve damage; for researchers, it is a mechanistic probe worth a controlled trial, not a clinical tool.
Breitschwerdt et al. (2025) found Babesia DNA in 24% (12/50, 95% CI 12–36%) of chronic fatigue patients via PCR and enrichment blood culture, with the combined Babesia+Bartonella positivity reaching 46% (23/50) (Edward B. Breitschwerdt et al. 2025b). However, the study had no healthy controls — without regional matched controls, even the Breitschwerdt detection rate cannot be interpreted as elevated, because the 95% CI (12–36%) overlaps substantially with background seroprevalence in endemic regions (1–20%). The enrichment culture method (BAPGM) has unresolved specificity questions. A 1996 case-control study found zero Babesia seropositivity in 47 CFS cases (vs 2/47 controls — not a significant difference, but the only controlled comparison points toward no association) (MacDonald et al. 1996). CDC surveillance confirms Babesia can produce chronic neurologic symptoms including fatigue (Locke et al. 2023), but whether detection rates in fatigue patients exceed background seroprevalence in endemic regions is unknown. Certainty: n/a (documents an unresolved evidence gap; the sole controlled study is directionally neutral).
The “50% Babesia in ME patients” figure circulating in patient communities conflates the combined Babesia+Bartonella positivity rate with Babesia alone — the actual Babesia-only rate was 24%. Without controlled data, the clinical significance of Babesia DNA detection in a fatigue population cannot be determined: it may represent active infection driving symptoms, past exposure unrelated to current illness, or detection of non-viable organisms by an enrichment method more sensitive than clinically validated tests.
Consequence: For patients, this means positive Babesia testing alone cannot confirm causation; for clinicians, Babesia screening in tick-exposed ME/CFS patients remains reasonable clinical practice but the evidence base for interpreting results is substantially weaker than for Borrelia — before screening you should know what you will do with a positive result. A well-controlled prevalence study with regional healthy controls is the single step that would most resolve the current uncertainty.
3.5.1 Mechanistic Bridges to Existing ME/CFS Pathways
The tick-borne findings above need not sit in isolation; several plausibly connect to mechanisms this book develops elsewhere. The following are exploratory bridges (Origin: brainstorm), offered as testable hypotheses rather than established links.
Chronic or recently-cleared tick-borne infection sustains interferon-γ signalling, which activates indoleamine-2,3-dioxygenase (IDO) and diverts tryptophan down the kynurenine pathway — the same enzymatic shunt discussed for ME/CFS (Tryptophan/Kynurenine Trap) and mechanistically upstream of the peripheral-serotonin-depletion model (Peripheral Serotonin Depletion as Multi-System Convergence Point). If IDO activity is sustained after the pathogen is controlled, tryptophan diversion could lower peripheral serotonin and NAD⁺ substrate availability — one candidate biochemical route by which post-Lyme fatigue could persist, the domain that population data show is the most persistent after Lyme disease (Nawrocki et al. 2025). Certainty: 0.28 (Origin: brainstorm) — IDO activation in Lyme is established and the ME/CFS serotonin/kynurenine machinery is documented, but the specific post-Lyme tryptophan → serotonin → fatigue chain has not been measured directly, and a key quantitative uncertainty remains: only a small fraction of tryptophan is normally committed to serotonin synthesis, so whether IDO upregulation depletes that fraction enough to affect serotonin-dependent function is unestablished.
Falsifiable prediction: Post-Lyme patients with persistent fatigue will show elevated kynurenine:tryptophan ratios and lower peripheral (platelet/serum) serotonin than recovered post-Lyme controls, with magnitude tracking fatigue severity. Refuted if KYN:TRP and serotonin do not differ by fatigue status, or if any difference is fully explained by acute-phase inflammation rather than sustained IDO activity.
Consequence: If it holds, an inexpensive metabolic ratio could point to a candidate neurochemical mechanism behind lingering post-Lyme fatigue — but this currently links two separately-documented mechanisms, not a measured pathway. (This speculation is contingent on the peripheral-serotonin-convergence hypothesis Peripheral Serotonin Depletion as Multi-System Convergence Point remaining well-supported: if that upstream model weakens, this bridge weakens with it.)
The PTLD immunophenotyping study separated patients from controls only with a multi-parameter classifier, not by any single cell frequency (Girgis et al. 2025) — echoing the long-standing difficulty of finding one reproducible immune marker in ME/CFS (CD8+ T-cell Exhaustion as a Downstream Consequence of Exosomal EV Cargo Rather than Antigen-Driven). If the discriminating signal in both conditions is inherently combinatorial, single-marker studies would be expected to disagree across cohorts, and reanalysis of existing ME/CFS flow-cytometry datasets with multivariate classifiers could recover a signal that univariate tests miss. Certainty: 0.35 (Origin: brainstorm) — the combinatorial finding is solid for PTLD, but its transfer to ME/CFS is inference, not measurement.
Falsifiable prediction: A multivariate (e.g. elastic-net) classifier applied to published ME/CFS immunophenotyping datasets will separate patients from controls with materially higher AUC than the best single marker in the same data. Refuted if multivariate and univariate performance are equivalent, or if any multivariate signal fails to replicate across independent cohorts.
Consequence: If true, this reframes a decade of “inconsistent” ME/CFS immune results as an analysis-method artifact rather than absence of signal, and points to how existing data should be re-mined — a low-cost, high-leverage research direction.
Bartonella species characteristically infect and dysregulate vascular endothelium, and the tick-borne section above notes their vasculitis-like presentations. Endothelial activation is also central to the microclot and endothelial-dysfunction pathology this book develops for ME/CFS and Long COVID (Treatment Safety: Coagulation Interventions). A chronic endotheliotropic infection could therefore contribute to microclot formation through direct endothelial injury rather than the autoantibody or fibrin(ogen)-conformation routes emphasised elsewhere. Certainty: 0.15 (Origin: brainstorm) — a double extrapolation: it assumes both that Bartonella is present and active in the patient and that its endotheliotropism produces the microclot phenotype, neither established in ME/CFS. The certainty is conditional on Bartonella presence being confirmed.
Falsifiable prediction: Post-tick-exposure ME/CFS patients with laboratory or PCR evidence of Bartonella will show higher circulating endothelial-activation markers (von Willebrand factor, soluble thrombomodulin) and microclot burden than tick-exposed Bartonella-negative patients. Refuted if microclot burden is independent of Bartonella status, or if endothelial markers do not differ.
Consequence: If confirmed, it would identify a subset in whom treating a chronic bacterial infection — not just anticoagulating — addresses the vascular pathology; at present it is a mechanistic conjecture, not grounds for antibiotic use.
Babesia parasitises and lyses erythrocytes, releasing free hemoglobin into plasma. Free hemoglobin is cleared by haptoglobin, but sustained low-grade hemolysis can deplete haptoglobin reserves — and ME/CFS patients show reduced haptoglobin after exertion, with haptoglobin phenotype tracking cognitive dysfunction severity (Moezzi et al. 2025). If chronic Babesia infection drives persistent low-level hemolysis, it would deplete the same haptoglobin capacity already documented as compromised in ME/CFS, creating a biochemical convergence independent of the immune and neurological mechanisms discussed above. Free hemoglobin is itself a pro-oxidant (Fenton chemistry), and heme released from lysed cells upregulates heme oxygenase-1 via Nrf2 — a pathway also implicated in fibromyalgia-associated fatigue (Luo et al. 2025). Certainty: 0.18 (Origin: brainstorm) — each component is documented (Babesia hemolysis, haptoglobin depletion in ME/CFS, HO-1 in fatigue) but the specific Babesia-to-haptoglobin-to-fatigue chain has never been measured in a single cohort; the connection is inferential chemistry, not measured biology.
Falsifiable prediction: In a tick-exposed ME/CFS cohort, plasma haptoglobin will be inversely proportional to Babesia parasitemia, and free hemoglobin / heme will be proportional to parasitemia after controlling for known confounders (recent exertion, hemolysis from other causes). Refuted if no relationship exists, or if haptoglobin is suppressed by an Babesia-independent mechanism.
Consequence: If confirmed, this would provide a specific, testable biochemical pathway by which Babesia infection directly worsens ME/CFS — distinct from the non-specific “chronic infection causes fatigue” — and could rationalise haptoglobin as a monitoring marker during anti-Babesia treatment. At present it is a theoretical chain of individually-documented links, not a measured pathway, and does not justify treatment on this basis alone.
The evidence for Babesia specifically driving ME/CFS symptoms rests on a single uncontrolled pilot study of 50 patients from a tick-borne disease referral practice (Edward B. Breitschwerdt et al. 2025b). Several features of the broader Babesia–ME/CFS question remain unresolved and point to concrete research that would clarify the landscape:
- Null hypothesis: Babesia prevalence in ME/CFS does not exceed the geographic background rate (1–20% depending on region). Current evidence is insufficient to reject this null, because no study has compared Babesia detection rates in ME/CFS patients against matched controls from the same endemic area.
- Detection vs disease: Babesia microti typically causes acute self-limited hemolytic illness in immunocompetent adults and is cleared by splenic and adaptive immunity; persistent PCR-positive parasitemia beyond 6 months is uncommon outside of asplenia or immunosuppression (though low-level carriage has been documented in immunocompetent individuals, the frequency and clinical significance are unresolved). The Breitschwerdt findings may reflect transient subclinical exposure detected by highly-sensitive enrichment culture (BAPGM) rather than chronic active infection driving ME/CFS symptoms. Notably, even if Babesia acts only as an acute trigger that clears — the model the paper endorses for other infections — the persistent-driver framework tested below would miss it; a separate trigger-then-clear mechanism should be tested independently.
- Methodological confounds: The enrichment culture method used (BAPGM) requires 7–14 days of culture — prolonged incubation increases contamination risk. The study cohort was enriched for patients who already suspected tick-borne illness (Berkson’s bias). No independent replication exists; the positive Babesia–fatigue data come from one research group, while the sole independent attempt (MacDonald 1996, serology-based) was null.
- What would reject the null: An independent study showing (a) Babesia detection (PCR + serology) in materially more ME/CFS patients than matched endemic controls, (b) evidence of active infection currently associated with symptoms (whether via persistent parasitemia or acute-trigger-then-clear with post-infectious sequelae), and (c) haptoglobin/LDH abnormalities correlating with Babesia positivity — tested by a lab without financial ties to the diagnostic assay. (The specific numerical thresholds from the brainstorm are illustrative, not prescriptive: the crucial criterion is a statistically significant difference from controls, not passing any particular percentage.) Certainty: n/a (epistemic framing, not a claim; Origin: brainstorm categories 10–12).
Consequence: For researchers, the three most impactful studies are (1) a Babesia serosurvey with physically-matched controls, (2) independent PCR replication of the Breitschwerdt finding, and (3) a haptoglobin/LDH ratio screening study to determine whether biochemical evidence of hemolysis can serve as a cost-effective triage for Babesia PCR in ME/CFS (Babesia Hemolysis → Haptoglobin Depletion as a Fatigue-Amplifying Pathway; Origin: brainstorm).
If the null hypothesis framed above is to be tested rigorously, several concrete investigations — drawn from the brainstorm phase (Origin: brainstorm) — would provide the highest information gain:
- Controlled serosurvey with 2-day CPET: A three-arm study in Babesia-endemic US regions comparing ME/CFS patients, physically-matched healthy controls, and post-treatment Lyme disease patients (n ≥ 200 per arm). All tested by Babesia IFA + PCR + hemolytic panel (haptoglobin, LDH, reticulocyte count, indirect bilirubin, free hemoglobin, RBC deformability by ektacytometry). A subset (n=60, 20 per arm) undergoes 2-day CPET to test whether Babesia seropositivity predicts greater day-2 VO₂max decline and post-exercise haptoglobin drop. Certainty: n/a (research proposal — the study design is what needs validation, not the hypothesis). The predicted finding: Babesia detection rate exceeds matched controls in the ME/CFS arm, and seropositive patients show greater PEM severity by CPET metrics.
- Haptoglobin/LDH ratio as a Babesia screening triage tool: Two cheap, standard blood tests (haptoglobin and LDH, combined as the Hp/LDH ratio) could serve as a cost-effective gate: patients with Hp/LDH ratio below 0.5 would proceed to Babesia PCR, while those above would not — potentially reducing the number needed to PCR-test by 4× compared to universal testing. The screening performance (sensitivity, specificity, PPV) needs validation against Babesia PCR in an endemic ME/CFS cohort (n ≥ 500). Certainty: n/a (proposal — the ratio’s screening performance has never been measured). If validated, this would give clinicians a practical, low-cost way to decide which ME/CFS patients warrant Babesia testing.
- ODE model of Babesia parasitemia → haptoglobin depletion → PEM threshold: A compartment ODE model (Babesia parasitemia, RBC count, plasma free hemoglobin, haptoglobin) could predict the haptoglobin-depletion threshold at which subpatent-level parasitemia (0.3% infected RBCs, undetectable by blood smear) depletes haptoglobin reserves enough to amplify post-exertional hemoglobin toxicity. The prediction — that even very low-level infection shifts the exercise free-hemoglobin exposure curve — is testable against haptoglobin measurements in asymptomatic PCR-positive blood donors and controlled Babesia-infected animal models. Certainty: n/a (modeling proposal — the model parameters and predictions need calibration against measured kinetics).
- RBC deformability as a functional biomarker: If Babesia-infected red blood cells are mechanically rigid (established in parasitology), then ektacytometry-measured RBC deformability (elongation index at shear stress 3 Pa) could serve as a treatment-responsive biomarker for Babesia-associated microvascular dysfunction in ME/CFS. The prediction: Babesia-PCR-positive ME/CFS patients show lower deformability indices than PCR-negative patients, and the index improves after successful Babesia treatment. Certainty: n/a (proposal — no RBC deformability data exist for ME/CFS, Babesia-associated or otherwise).
- Diagnostic equity: Babesia PCR testing costs approximately USD 200–400 out-of-pocket and is often denied by insurance outside of acute illness. A program subsidizing Babesia PCR for ME/CFS patients in endemic areas would simultaneously identify treatable cases and generate the first unbiased prevalence data — serving clinical and research goals in a single initiative.
These proposals collectively cost less than a single large RCT and could resolve the Babesia–ME/CFS question within 2–3 years if funded in parallel. They are presented as research directions, not clinical recommendations: none of the tests described should be ordered outside a research protocol until validated. Certainty: n/a (aggregate research proposals).
Consequence: For funders and researchers, the Babesia question is resolvable with modest investment — the studies above, run in parallel, could transform “unknown” to “known or ruled out” within a few years. For patients, this means the current uncertainty has a concrete path to resolution, and testing outside a research context should be approached with full understanding of the interpretive caveats described above.
The frequently-cited 26-of-29 symptom overlap between PTLDS and ME/CFS is striking, but part of it may be definitional rather than mechanistic: both conditions are defined largely by the same core features — fatigue, cognitive dysfunction, unrefreshing sleep, and pain — so a high symptom concordance is partly guaranteed by how the two syndromes are constructed. Shared symptom checklists cannot, on their own, establish shared biology. Certainty: n/a (methodological caveat, not a claim).
Consequence: Readers should not treat symptom-overlap statistics as evidence of a common cause; the mechanistic bridges (immune signatures, autonomic testing, HLA) — not the symptom counts — are what would establish genuine shared pathophysiology. That said, the counterpoint deserves weight: post-exertional malaise and the specific severity and pattern of overlap may exceed what shared diagnostic criteria alone would produce, so the definitional critique bounds but does not dissolve the overlap.
Three new axes of evidence refine the picture of tick-borne illness as an underrecognised ME/CFS trigger. The overlap is now domain-specific: most post-Lyme symptoms resolve, but fatigue persists beyond 12 months in population data — precisely the defining ME/CFS domain (Nawrocki et al. 2025). The mechanistic picture is sharpening: a combinatorial T-cell signature separates post-Lyme illness from resolved infection (Girgis et al. 2025), autonomic testing confirms overlap with ME/CFS patterns (Milovanovic et al. 2025), and three bridge speculations connect tick-borne mechanisms to the paper’s existing pathways — kynurenine→serotonin, combinatorial immune classifiers, and Bartonella→endothelial→microclot (Infection-Driven Kynurenine Shunt as a Route to Post-Lyme Fatigue, Combinatorial Immune Signatures, Not Single Markers, Behind Cross-Study Inconsistency, Bartonella Endotheliotropism as a Non-Autoantibody Route to Microclots). The hardest constraint is honest: the 26/29 symptom overlap is partly definitional — shared diagnostic criteria guarantee symptom concordance — and the Babesia link remains the most inferential of the three pathogens (Babesia Remains the Least-Studied and Most Contentious of the Three in ME/CFS, How Much of the PTLDS–ME/CFS Overlap Is Definitional?). What survives: post-Lyme illness and ME/CFS share a fatigue-dominant phenotype with overlapping immune and autonomic abnormalities, and the bridges connecting them are framed to be experimentally testable — which patients fit the pattern, which mechanism drives it, and whether treating the infection alters the outcome all remain open.
Consequence: The clinical task shifts from “does this patient have ME/CFS or post-Lyme?” — a question grounded in symptom overlap — to “among patients with ME/CFS, which ones have a tick-borne trigger that a mechanism-informed workup could identify?” For researchers, the task is to test the specific bridges rather than document further phenomenological resemblance.
3.6 Glycolytic Reprogramming of Host Cells by Tick-Borne Pathogens
The paper’s treatment of infection-driven metabolic reprogramming has centred on viral pathogens: neurotropic viruses reprogram glia from oxidative phosphorylation toward glycolysis, a shift that stabilises epigenetically and sustains chronic neuroinflammation (Post-Viral CNS Reprogramming Hypothesis, Brainstem Glial Senescence as a Self-Reinforcing Autonomic Trap). An emerging, parallel literature shows that the non-viral tick-borne pathogens — Borrelia, Bartonella, and Babesia — also drive glycolytic reprogramming of host cells, through pathogen-specific routes. This matters because such a shift converges on the same immunometabolic endpoints already implicated in ME/CFS: lactate accumulation, HIF-1\(\alpha\) signalling, mitochondrial downregulation, and trained-immunity memory.
Borrelia burgdorferi. The evidence is strongest for Borrelia. In human THP-1 monocytes, B. burgdorferi upregulates glycolysis-related genes — notably lactate dehydrogenase A (LDHA) and thioredoxin (TXN) — with measurable glucose depletion and lactate accumulation in culture supernatant (Dong et al. 2026). More consequentially, B. burgdorferi induces a persistent trained-immunity phenotype in macrophages characterised by augmented glycolytic output and profound downregulation of mitochondrial components; in a mouse model this reprogramming persisted after the initial stimulus, spread tissue-wide (heart), and was reversed in vivo by glycolysis inhibition, which restored mitochondrial components and reduced both inflammation and spirochetal burden (Barriales et al. 2021). Borrelia-induced trained immunity is cell-type specific (murine synoviocytes but not skin fibroblasts), which may contribute to symptom heterogeneity (Bernard and Hu 2020). Borrelia itself is obligately glycolytic — its reduced genome depends on glycolysis for carbon metabolism, and lactate-dehydrogenase inhibitors suppress its growth (Lynch et al. 2023) — so a glycolytically reprogrammed host cell may also provide a permissive niche for pathogen persistence.
Bartonella. Bartonella henselae stabilises hypoxia-inducible factor 1\(\alpha\) (HIF-1\(\alpha\)) in infected cells (shown in HeLa and large-vessel endothelial cell lines) — the first demonstration of HIF-1\(\alpha\) activation in any bacterial infection — driving hypoxia-response gene transcription including VEGF, increasing cellular oxygen consumption, and producing cellular hypoxia with decreased ATP (Kempf et al. 2005). HIF-1\(\alpha\) stabilisation is the canonical switch toward glycolytic metabolism and is the same node implicated in viral glial reprogramming and in the ME/CFS HIF hypotheses (HIF Pathway Inertia: Failure of Hypoxic Adaptation in ME/CFS). Given Bartonella’s endothelial tropism, this would link tick-borne infection to both the glycolytic-shift and the endothelial-dysfunction threads of ME/CFS pathophysiology — though the evidence is a single, unreplicated, 20-year-old cell-line study.
Babesia. Evidence is weakest for the intra-erythrocytic protozoan Babesia. Infected murine red blood cells develop increased, apparently de novo glucose-uptake systems relative to uninfected cells (Ohmori et al. 2004) — an increase in erythrocyte glucose consumption, though (since mature red cells lack nuclei and mitochondria) this is a transporter-level metabolic shift, not transcriptional “reprogramming” in the sense used for nucleated immune cells. It rests on a single small animal study with no human or mechanistic follow-up.
Certainty: 0.35. This certainty reflects a well-supported mechanism in non-ME/CFS systems but zero direct ME/CFS evidence, and the closest human data cut against it: PTLDS serum metabolomics found no glycolytic signature (Fitzgerald et al. 2021) (consistent either with cell-type-restricted reprogramming invisible in serum, or with its absence). The evidence is also markedly asymmetric across the three pathogens — this is better read as one moderately-evidenced pathogen plus two single-study analogies of different mechanistic kind, not three co-equal converging lines. Borrelia (the moderately-evidenced case) drives a Warburg-like glycolytic shift via LDHA induction and a trained-immunity macrophage memory with mitochondrial downregulation, persisting at least 7 days in a mouse model (Dong et al. 2026) (Barriales et al. 2021). Bartonella stabilises HIF-1\(\alpha\) with host ATP depletion (Kempf et al. 2005) — a programme consistent with a glycolytic shift, though glycolysis itself was not measured (single unreplicated cell-line study). Babesia increases glucose uptake in infected erythrocytes (Ohmori et al. 2004) — a transporter-level effect in anucleate cells, mechanistically distinct from the transcriptional/epigenetic reprogramming of the other two and not “reprogramming” in that sense. In a tick-borne-triggered or co-infected ME/CFS subset, a Borrelia-type glycolytic shift may compound the pre-existing immunometabolic dysfunction documented in ME/CFS — the failed CD8+ T-cell Warburg switch (Mandarano et al. 2020) and the cell-danger-response glycolytic reprogramming (Naviaux et al. 2016) — and, if it stabilises epigenetically as trained immunity does, could persist after pathogen clearance to help sustain chronic symptoms, mirroring the post-viral glial mechanism (Post-Viral CNS Reprogramming Hypothesis).
Direction of effect is unresolved. Two Borrelia readouts point opposite ways: raised supernatant lactate in THP-1 monocytes at 24h (Dong et al. 2026) versus decreased baseline glycolytic rate in primary human monocytes (Kerstholt et al. 2022). These are orthogonal measurements (post-infection lactate output versus pre-stimulation basal rate, in different cell substrates and timepoints) rather than a strict head-to-head contradiction — but precisely because they measure different things, the net direction of host glycolysis in vivo remains genuinely undetermined.
Falsifiable prediction. In a tick-borne-triggered ME/CFS subset with documented (serology/PCR-confirmed) infection, isolated monocytes/macrophages — assayed both at rest and after ex-vivo restimulation (e.g., LPS or Pam3CSK4, since trained immunity is a primed state that may be silent until re-triggered) — will show a trained-immunity glycolytic signature (elevated LDHA/lactate, downregulated mitochondrial transcripts, enhanced cytokine responses) exceeding that of non-tick-borne ME/CFS patients and healthy controls; and this signature will be absent or attenuated in patients whose infection was eradicated early. Falsified if restimulated tick-borne-triggered ME/CFS monocytes show no glycolytic/mitochondrial difference from other ME/CFS patients (a resting-only null would be consistent with a primed-but-unexpressed state and would not by itself refute the hypothesis).
Limitations. No study has measured tick-borne pathogen-driven glycolytic reprogramming in any ME/CFS patient — the ME/CFS relevance is entirely a mechanistic parallel, not a direct observation. The strongest evidence (Borrelia trained immunity) is from mouse and human-cell models (Barriales et al. 2021), measured only 7 days post-stimulation with no test of persistence after pathogen clearance and no ME/CFS-relevant clinical endpoint; whether such a glycolytic memory lasts the months-to-years of ME/CFS in humans is unknown. Trained immunity is classically a primed state requiring a second stimulus to manifest (Wendeln et al. 2018), so any resting metabolic signature may be absent unless re-triggered (e.g., by exertion). The Bartonella evidence rests on a single, unreplicated, 20-year-old study in cancer and large-vessel-endothelial cell lines that measured HIF-1\(\alpha\) and ATP but not glycolysis directly (Kempf et al. 2005); the Babesia evidence is a single small animal study in a rodent-specific parasite species (Ohmori et al. 2004) — so the “three-pathogen” framing is better read as one moderately-evidenced pathogen (Borrelia, with a direction-of-effect contradiction) plus two single-study analogies. PTLDS serum metabolomics did not surface a glycolytic signature (Fitzgerald et al. 2021), consistent with the reprogramming being cell-type-restricted (tissue/monocyte) rather than systemically detectable in serum — or with its absence. This is a hypothesis about a subset, not all ME/CFS; the size of any such subset is unknown, and severity applicability is unknown. Consequence: If confirmed, this would identify a tick-borne-triggered ME/CFS subgroup whose persistent illness is driven partly by a pathogen-imprinted metabolic memory in immune cells — a mechanism that, unlike the infection itself, might be addressable with metabolic (glycolysis-modulating) rather than purely antimicrobial approaches; at present it is a research direction, not a basis for any treatment.
This is a mechanistic hypothesis, explicitly not a treatment recommendation: there is zero human data for this indication, no trial has tested it, and patients should not take metformin for ME/CFS on this basis. Certainty: 0.22. (Origin: brainstorm.) With that stated: if the tick-borne glycolytic-reprogramming hypothesis (Tick-Borne Pathogen-Driven Glycolytic Reprogramming as a Compounding Immunometabolic Insult in a Post-Infectious ME/CFS Subset) holds, its most translatable therapeutic corollary is that the trained-immunity glycolytic memory — not the pathogen — is the target. Barriales et al. showed that glycolysis inhibition in vivo reversed the Borrelia-trained macrophage phenotype, restoring mitochondrial gene expression and reducing inflammation (Barriales et al. 2021). Metformin, via AMPK activation and reduced acetyl-CoA availability for histone acetylation, opposes trained-immunity establishment in other models and is a generic, well-characterised drug — making it a plausible metabolic (not antimicrobial) probe for a tick-borne-triggered subset in whom the infection is documented and treated, distinct from and not in tension with the paper’s rejection of prolonged antimicrobials for “seronegative chronic Lyme” (“Seronegative Chronic Lyme” as a Claimed Cause of ME/CFS — Not Supported).
Harms — including a mechanism-specific one. Metformin carries the usual risks: gastrointestinal intolerance, vitamin B12 depletion, and, rarely, lactic acidosis (a relative contraindication in renal impairment, e.g. eGFR <30 — and an irony given the lactate-centred rationale). More fundamentally, metformin is a mitochondrial complex I inhibitor: in a disease this book argues is characterised by oxidative-phosphorylation deficiency (Chapter Energy Metabolism and Mitochondrial Function), suppressing complex I could worsen the OXPHOS deficit even while it dampens glycolytic trained immunity — so the net effect could be harmful rather than therapeutic. This tension is itself a reason the idea must be tested, not assumed beneficial.
Falsifiable prediction. In an open-label pilot in tick-borne-triggered ME/CFS with a documented, treated prior infection, metformin will reduce ex-vivo monocyte glycolytic output (ECAR/OCR) relative to baseline after 12 weeks; falsified if monocyte ECAR/OCR does not change (mitochondrial-transcript expression and fatigue are secondary exploratory endpoints, and any worsening of fatigue or OXPHOS markers would count against the idea).
Consequence: If a future trial confirmed net benefit, an inexpensive, widely available drug could offer a way to unwind the metabolic aftermath of a tick infection in the subset it applies to — but the complex I concern means it could equally make patients worse, so today this is only a research direction and patients should not act on it.
The tick-borne glycolytic-reprogramming hypothesis (Tick-Borne Pathogen-Driven Glycolytic Reprogramming as a Compounding Immunometabolic Insult in a Post-Infectious ME/CFS Subset) is testable now with existing tools, and its central empirical gaps are well defined. Priority studies: (1) Ex-vivo monocyte metabolic phenotyping (Seahorse ECAR/OCR) comparing tick-borne-triggered ME/CFS, viral-triggered ME/CFS, and controls — the single most informative experiment, capable of confirming or refuting a pathogen-specific glycolytic signature and of serving as a stratification biomarker. (2) Resolving the direction-of-effect contradiction (Dong (Dong et al. 2026) ↑ vs Kerstholt (Kerstholt et al. 2022) ↓) via a systematic replication across cell substrates (primary monocytes vs THP-1) and timepoints — determining whether Borrelia raises or lowers host glycolysis is a precondition for the whole model. (3) CSF (not serum) metabolomics in PTLDS, since the negative serum finding (Fitzgerald et al. 2021) may reflect cell-type/compartment restriction that a CNS-compartment measurement would capture. (4) Post-hoc stratification of existing ME/CFS trial datasets by documented tick-exposure history (respecting “Seronegative Chronic Lyme” as a Claimed Cause of ME/CFS — Not Supported — documented exposure, not “chronic Lyme” labels) to look for hidden responder subgroups at zero marginal cost.
Consequence: These are concrete, mostly inexpensive experiments that could within a few years either establish tick-borne-triggered ME/CFS as a metabolically distinct, treatable endotype or close the hypothesis — turning a currently untested mechanistic parallel into evidence one way or the other.
The environments in this section and their neurological extensions assemble a single argument, resting on markedly asymmetric evidence: one moderately-evidenced pathogen plus weaker analogies. Borrelia (the moderately-evidenced case) drives a Warburg-like glycolytic shift via LDHA induction and a reversible trained-immunity macrophage memory with mitochondrial downregulation (persisting ≥7 days in mice); Bartonella stabilises HIF-1\(\alpha\) with host ATP depletion — a programme consistent with but not shown to be a glycolytic shift (single cell-line study); Babesia raises erythrocyte glucose uptake — a transporter-level effect in anucleate cells, mechanistically distinct and not “reprogramming” in the transcriptional sense (Tick-Borne Pathogen-Driven Glycolytic Reprogramming as a Compounding Immunometabolic Insult in a Post-Infectious ME/CFS Subset). The organising idea is that the Borrelia case, if it holds, reaches the same metabolic endpoint the paper documents for viral glial reprogramming, so the “post-viral” framing of chronic post-infectious metabolic memory (Post-Viral CNS Reprogramming Hypothesis) may be too narrow — a broader post-infectious glial reprogramming model would predict the same self-sustaining metabolic-epigenetic cycle from bacterial (not protozoal) triggers (Non-Viral Tick-Borne Pathogens May Drive Comparable Glial and Immune Glycolytic Reprogramming), with candidate downstream consequences reaching the astrocyte–neuron lactate shuttle (Borrelia-Trained Immune Glycolysis May Reverse the Astrocyte–Neuron Lactate Gradient) and a metabolic — not antimicrobial — therapeutic corollary (Metformin as a Hypothetical Trained-Immunity Metabolic Reset — Research-Stage Only). The “post-infectious memory” label applies only where persistence after pathogen clearance is plausible (Borrelia trained immunity, viral glial reprogramming); it does not extend to Babesia, whose effect requires ongoing parasitaemia.
Two constraints keep this firmly hypothetical. First, no study has measured any of this in a single ME/CFS patient: every link is a mechanistic parallel from in-vitro, animal, or other-disease data, and the strongest single dataset (Borrelia trained immunity) is a 7-day mouse experiment. Second, the foundational direction-of-effect is itself unresolved — Borrelia has been reported both to raise and to lower host glycolysis — and the closest human data (PTLDS serum metabolomics) found no glycolytic signature at all. The central open question is therefore not “how does this cause ME/CFS” but “does it occur in ME/CFS at all,” answerable by the concrete experiments in Testing Tick-Borne Glycolytic Reprogramming in ME/CFS: A Research Programme.
Consequence: If future measurement confirms even part of this convergent picture, tick-borne-triggered ME/CFS would become a metabolically defined, potentially treatable endotype unified with post-viral ME/CFS under one “post-infectious metabolic memory” framework; if it does not, the tick-borne trigger collapses into just one more entry point to the common immunometabolic pathway the paper already describes — either outcome sharpens the model, and neither is settled today.
Certainty: 0.20. The established microglial-activation axis for B. burgdorferi (certainty 0.70, supported by three independent laboratories using human and primate microglia (Myers, Kaushal, and Philipp 2009) (Parthasarathy and Philipp 2015) (Cassiani-Ingoni et al. 2006) (Kuhlow et al. 2005)) carries an unstudied downstream consequence: sustained TLR→NF-\(\kappa\)B→iNOS activation is a known trigger of mitochondrial complex inhibition through nitric-oxide-mediated respiratory-chain damage (established biochemistry; reviewed in (Peacock et al. 2015)), and B. burgdorferi drives M1 microglial polarisation (Akinlusi et al. 2025) — a state metabolically characterised in macrophages (but not yet measured in microglia) by a shift from oxidative phosphorylation to aerobic glycolysis, as documented in B. burgdorferi-trained macrophages (Barriales et al. 2021). If microglial mitochondria are similarly damaged or reprogrammed, three downstream effects become plausible: (1) mtDNA release into the cytosol activating NLRP3 and/or cGAS-STING to sustain IL-1\(\beta\) output — the terminal effect already demonstrated in ME/CFS where exercise-released exosomal mtDNA activates human microglia to produce IL-1\(\beta\) (Tsilioni, Natelson, and Theoharides 2022); (2) impaired mitophagy leading to accumulation of damaged mitochondrial mass and energy failure in the cell type responsible for CNS immune surveillance; and (3) a Warburg-shifted microglial metabolism that locks the cell in a pro-inflammatory M1 phenotype epigenetically — the same trained-immunity mechanism documented in peripheral macrophages (Tick-Borne Pathogen-Driven Glycolytic Reprogramming as a Compounding Immunometabolic Insult in a Post-Infectious ME/CFS Subset) extended to the CNS compartment.
The complete chain is untested. No study has measured mitochondrial function (respiration, membrane potential, mtDNA integrity, mitophagy) in microglia after B. burgdorferi or Bartonella exposure — the central mechanistic link between microbial activation and sustained microglial dysfunction is a blank in the literature. For Bartonella henselae, the gap is even wider: the single existing study demonstrating microglial infection is a 2001 feline in vitro experiment with no cytokine readouts and no mitochondrial data (Munana et al. 2001). The evidence for the terminal step (mtDNA→microglia→IL-1\(\beta\)) is ME/CFS-specific (Tsilioni, Natelson, and Theoharides 2022), the evidence for the initial step (Bb→microglia→TLR→inflammation) is robust, and the evidence for the middle (microglial mitochondrial dysfunction following bacterial infection) is absent.
Falsifiable prediction. In primary human microglia or iPSC-derived microglia exposed to live and non-viable B. burgdorferi in vitro, mitochondrial respiration (Seahorse OCR) will be reduced and extracellular acidification rate (ECAR) will be elevated relative to unstimulated controls beginning 6–24 hours post-exposure, and mtDNA release into the supernatant will be detectable by qPCR. Falsified if microglia maintain normal mitochondrial respiration and produce no mtDNA after 48 hours of B. burgdorferi exposure at any MOI. This is experimentally straightforward and has never been done.
Limitations. This is a molecular gap analysis, not a finding. The 0.20 certainty reflects that the mechanism is deductive — assembled from separately-established upstream and downstream components — with the central link unmeasured. The Bartonella microglial infection evidence is in feline cells only; human relevance unestablished. The B. burgdorferi-induced oxidative stress evidence (Peacock et al. 2015) (Wawrzeniak et al. 2020) derives from non-microglial cell types; direct demonstration in microglia is absent. Replication status: the upstream Bb→microglia axis is well-replicated (6+ independent studies, human and primate cells); the downstream ME/CFS mtDNA→microglia step is a single study; the middle mitochondrial step is not replicated at all because it has never been measured.
Consequence: For researchers, this is a tractable experiment — primary human microglia exposed to B. burgdorferi and assessed for mitochondrial function — that would fill a mechanism-defining gap in under 6 months. For patients and clinicians, this is pure basic science with no immediate clinical application; the gap-identification may help direct research funding toward a specific, testable question rather than an undifferentiated “neuroinflammation” construct.
Certainty: 0.30. (Origin: brainstorm.) The mechanistic distinction between B. burgdorferi priming microglia (which it does — TLR1/TLR2 upregulation, M1 polarisation with IL-1\(\beta\) induction — documented in human and primate microglia (Cassiani-Ingoni et al. 2006) (Myers, Kaushal, and Philipp 2009) (Akinlusi et al. 2025)) and B. burgdorferi triggering full-blown IL-1\(\beta\) secretion (which requires a second signal for NLRP3 inflammasome assembly — established in macrophage biology; inferred, not demonstrated, in Bb-exposed human microglia) provides a candidate molecular framework for PEM in tick-borne-triggered ME/CFS. In this two-hit model, Bb primes microglia but does not, by itself, drive sustained IL-1\(\beta\) release — the priming event upregulates inflammatory machinery without the activating trigger. The second hit is provided by exercise-induced exosomal mtDNA release (demonstrated in ME/CFS patients (Tsilioni, Natelson, and Theoharides 2022)), which binds TLR9 and/or enters the cytosol to activate NLRP3, producing mature IL-1\(\beta\) and the characteristic post-exertional neuroinflammatory crash. This model — the central link of which (Bb-primed microglia + mtDNA trigger) has never been tested directly — provides a specific molecular interpretation of clinical PEM: exercise → mtDNA release → second hit on primed microglia → IL-1\(\beta\) spike → neuroinflammatory symptom exacerbation.
Falsifiable prediction. In Bb-exposed primary human microglia, NLRP3 and pro-IL-1\(\beta\) mRNA will be elevated at 24 h but mature IL-1\(\beta\) secretion will be minimal. Addition of cell-free mtDNA (10 \(\mu\)g/mL) at 24 h will increase IL-1\(\beta\) secretion 5–10-fold relative to mtDNA alone or Bb alone, and this amplification will be blocked by MCC950 (NLRP3 inhibitor) and DNase. Falsified if Bb alone drives full IL-1\(\beta\) secretion (single-hit), or if mtDNA does not amplify the response in primed microglia.
Limitations. NLRP3 two-hit biology is established in macrophages but has not been specifically demonstrated in Bb-exposed human microglia; the mtDNA second-hit has been shown in ME/CFS serum exosomes ((Tsilioni, Natelson, and Theoharides 2022)) but the priming-by-Bb component is inferred from TLR upregulation studies, not from a direct Bb-primed + mtDNA-triggered experiment. The model predicts PEM specifically in tick-borne-triggered ME/CFS and does not speak to viral-triggered PEM (where the priming signal may be different). Certainty 0.30 reflects the inferential assembly — each component is separately plausible but the full two-hit chain in microglia has never been tested.
Consequence: If correct, this gives PEM a molecular mechanism specific to tick-borne-triggered disease — primed brain immune cells, not just “energy depletion,” explain why exercise crashes are so severe. Pacing would work because it prevents the mtDNA second-hit from reaching primed microglia, not because it conserves energy generally. For researchers, the two-hit model generates a clean, testable prediction about when and why IL-1\(\beta\) appears after exercise.
Certainty: 0.35. (Origin: brainstorm.) If tick-borne-triggered ME/CFS involves a distinct microglial mitochondrial damage pathway (Bb→TLR→iNOS→NO→complex inhibition→mtDNA release→feed-forward), then the combination of two measurements — serum exosomal mtDNA (post-exercise, per Tsilioni2022 protocol (Tsilioni, Natelson, and Theoharides 2022)) and tick serology (B. burgdorferi, Bartonella, Babesia IgG/IgM/IgA) — might identify the subgroup in whom this mechanism is active: seropositive patients with elevated post-exercise exosomal mtDNA would have both the trigger history and the active effector signal. A panel of three measures — exosomal mtDNA post-exercise (research-grade assay, not a standard clinical test), Borrelia C6 peptide IgG, and Bartonella IgG — could serve as a stratification tool for trials testing microglial-targeted interventions (e.g., minocycline, NLRP3 inhibitors) specifically in tick-borne-triggered disease.
The key uncertainty is specificity, not feasibility. Exosomal mtDNA elevation post-exercise has been demonstrated in ME/CFS (n=~20, Tsilioni2022) but not stratified by trigger type — whether viral-triggered and tick-borne-triggered ME/CFS differ in their exosomal mtDNA response to exercise is completely unknown. Borrelia serology has known cross-reactivity and sensitivity limitations (seronegative chronic infection is documented, though rare). The panel’s positive predictive value for identifying patients likely to respond to microglial-targeted therapy is nil — untested. Serial measurement (pre- and post-exercise mtDNA, before and after a microglial-targeted intervention) could serve as a pharmacodynamic biomarker, but this is a research proposal, not a clinical recommendation.
Consequence: A simple, relatively inexpensive blood test panel — mitochondrial DNA fragments measured after exertion plus tick-exposure antibodies — could help identify the ME/CFS patients most likely to benefit from treatments targeting brain immune cell inflammation. But the validation gap is large: the panel could be investigated now in any ME/CFS cohort with documented tick-exposure history, ideally alongside a candidate intervention, but should not currently be used for treatment decisions.
Three lines of reasoning bear on whether tick-borne bacterial pathogens activate the same neuroimmunological substrate — microglia — that the paper’s existing viral-triggered coverage describes in detail (Sleep EEG Delta/Alpha Ratio as a Non-Invasive Thalamic Calcium Proxy). First, B. burgdorferi activates human microglia directly through TLR1/TLR2 pathways with a consensus certainty of 0.70 (3 independent laboratories using human and primate microglia (Myers, Kaushal, and Philipp 2009) (Parthasarathy and Philipp 2015) (Cassiani-Ingoni et al. 2006) (Kuhlow et al. 2005)), and non-viable bacterial debris suffices to sustain this activation — providing a mechanistic explanation for post-antibiotic neuroinflammation in PTLDS (Parthasarathy and Gadila 2022). Second, the downstream consequence of this activation — mitochondrial dysfunction in microglia — is a deductive gap: the terminal effector (ME/CFS mtDNA→microglia→IL-1\(\beta\)) is established (Tsilioni, Natelson, and Theoharides 2022), the upstream trigger (Bb→microglia) is robust, but the middle step (microglial mitochondrial function after bacterial exposure) has never been measured (Microglial Mitochondrial Dysfunction After Bacterial Infection — Untested Mechanistic Gap). Third, the two-hit priming model (Two-Hit Microglial Priming: Bb as the Primer, Exercise as the Trigger) proposes — as a hypothesis whose central link has never been tested — that Bb primes microglia through TLR/M1 polarisation, and exercise-induced exosomal mtDNA provides the second hit that triggers IL-1\(\beta\)-mediated neuroinflammatory crashes. The evidence constraining these proposals is as important as the proposals themselves: Bartonella microglial infection has only a single 2001 feline study with zero mitochondrial or cytokine data (Munana et al. 2001); the microglial mitochondrial step is a complete blank; and the two-hit model has not been tested in any Bb-exposed human microglial preparation. What survives: the paper’s microglial activation coverage now accommodates both viral and bacterial triggers, and the key experimental gaps (measure mitochondrial function in Bb-exposed microglia; test the two-hit model; measure CSF mtDNA in tick-borne ME/CFS) are specific, tractable, and should be prioritised.
Consequence: The clinical task shifts from “tick-borne infection is a historical footnote in ME/CFS” to “tick-borne pathogens can drive exactly the same glial pathology as viral triggers, through pathways that are now mapped in sufficient detail to generate specific, falsifiable hypotheses.” For researchers, the most informative single experiment — measuring mitochondrial function in microglia after Borrelia exposure — has never been done and would define the mechanism within months. For patients and clinicians, this integration changes nothing about today’s management but clarifies why post-Lyme neurological symptoms can persist after antibiotic treatment: the problem may not be the ongoing infection but rather the microglial activation it leaves behind — and possibly, if the mitochondrial damage hypothesized above is confirmed, a self-sustaining inflammatory cycle.
3.7 Clinical Implications for ME/CFS Evaluation
When to Consider Tick-Borne Infections. Tick-borne infection evaluation should be considered in ME/CFS patients with:
- Geographic residence or travel to endemic areas
- Known tick exposure or recall of erythema migrans rash
- Onset following outdoor activities in wooded/grassy areas
- Symptoms suggesting disseminated Lyme: migratory arthralgias, facial palsy, heart block
- Marked sweats, air hunger, or hemolytic laboratory abnormalities (suggesting Babesia)
- Neuropsychiatric predominance with striae-like skin lesions (suggesting Bartonella)
- Previous inadequately treated or seronegative Lyme disease
Diagnostic Approach.
- Lyme disease: Two-tier testing (EIA/IFA followed by Western blot); consider C6 peptide ELISA; PCR on synovial fluid for Lyme arthritis
- Babesiosis: Blood smear, Babesia PCR, antibody testing; repeat testing during symptomatic episodes
- Bartonellosis: Serology (IgG, IgM), enrichment culture (specialized laboratories), PCR
- Co-infection panels: Given frequent co-transmission, comprehensive tick-borne disease panels are warranted
Tick-borne infection diagnosis remains controversial, with significant disagreement between IDSA/AAN/ACR guidelines and organizations like ILADS. Key issues include:
- Sensitivity of standard two-tier testing (estimated 30–40% in early disease, though estimates vary by study (Steere et al. 2016))
- Interpretation of “indeterminate” Western blots
- Validity of clinical diagnosis in seronegative patients
- Role of prolonged antibiotic therapy (not supported by controlled trials, but advocated by some practitioners)
Patients and clinicians should be aware of these controversies and the current limitations of evidence for chronic tick-borne infection treatment.
A recurring proposition holds that a large fraction of internationally case-defined CFS is in fact undiagnosed “seronegative chronic Lyme disease” that responds to prolonged antimicrobial therapy. The most-cited primary source is a retrospective cohort reporting that 99% of 210 CFS patients in a Lyme-endemic practice met criteria for “seronegative Lyme” and that 62–88% improved on antibiotics (Shor 2011). The available high-quality evidence does not support this claim, on two independent grounds.
Diagnostic premise. The “seronegative Lyme” label in that cohort was assigned when a patient had any one of several non-specific findings — a single highly-specific Western-blot band, any co-infection serology (Babesia, Bartonella, Ehrlichia), a low CD57 count, an elevated C4a, or an elevated C6 peptide — none of which is a validated test for active Borrelia infection (Paul M. Lantos 2015). Genuine seronegative Lyme is an established but narrow entity, described only in patients with documented prior acute Lyme treated early enough to blunt seroconversion (Dattwyler et al. 1988); it does not license extrapolation to undifferentiated chronic fatigue. When patients labelled “chronic Lyme” by such alternative methods are compared prospectively with CFS patients, the phenotypes are indistinguishable and neither group shows objective evidence of active Borrelia — i.e., “chronic Lyme” so diagnosed is CFS (Patrick et al. 2015) (Feder et al. 2007).
Therapeutic claim. Three independent double-blind randomised controlled trials of prolonged/repeated antibiotics for persistent symptoms attributed to Lyme found no benefit over placebo (Berende et al. 2016) (Klempner et al. 2001) (Fallon et al. 2008); the uncontrolled 62–88% “response” in the retrospective series is fully explained by placebo effect, regression to the mean (highest-vs-lowest symptom-score comparison), co-intervention for comorbidities, and observer bias in an unblinded single-author chart review by a declared advocacy-position clinician. Prolonged antibiotic courses for this indication are not benign: surveillance data document septic shock, Clostridioides difficile colitis, and fatal outcomes (Marzec et al. 2017). This limitation concerns the seronegative-chronic-Lyme-drives-CFS proposition specifically; it does not bear on genuine, serologically or PCR-confirmed tick-borne infection, whose overlap with ME/CFS is discussed above (Viral-Gene-Product Surveillance of Peripheral Compartments Is Null; Anellovirus Burden May Be an Immune-State Marker, Tick-Borne Illness as an ME/CFS Trigger — Convergent Model). Replication status: the null therapeutic result is independently replicated across three RCTs; the diagnostic critique is consistent across multiple reviews and one prospective comparison.
Consequence: For patients, this guards against an expensive, months-long antibiotic regimen that high-quality trials show does not work and that has caused documented harm, including deaths — a “treatable Lyme” label offered for ME/CFS should prompt scrutiny of whether any validated test actually documents active infection. For clinicians, it clarifies that seronegative-Lyme criteria built from non-specific markers cannot establish active Borrelia, and that confirmed tick-borne infection (not inferred “chronic Lyme”) is the appropriate threshold for antimicrobial treatment.
Treatment Considerations.
- Acute Lyme: Standard 2–4 week doxycycline course is well-established
- PTLDS: No treatment proven effective in controlled trials; extended antibiotic courses not recommended by IDSA (Paul M. Lantos et al. 2021); symptomatic management similar to ME/CFS
- Babesiosis: Atovaquone plus azithromycin (7–10 days, longer for immunocompromised); clindamycin plus quinine for severe cases
- Bartonellosis: Prolonged antibiotic courses (weeks to months) often required; regimens include doxycycline, azithromycin, rifampin combinations
- Co-infections: Require treatment of all identified pathogens; single-agent therapy may be inadequate
What proportion of ME/CFS cases have an undiagnosed tick-borne infection as the inciting event or ongoing driver? Given the symptom overlap between PTLDS and ME/CFS, improved diagnostic tools for chronic Borrelia, Bartonella, and Babesia infections could identify a treatable subset. Key research needs include: development of more sensitive diagnostic assays; prospective studies of tick-borne infection cohorts for ME/CFS development; controlled treatment trials in patients with documented chronic infection.
4 Intestinal Parasites as Post-Infectious Triggers
While viral and bacterial triggers of ME/CFS have received substantial research attention, intestinal parasites represent an underexplored but potentially important class of triggering agents. The evidence base is heterogeneous: one large, well-controlled natural experiment provides strong population-level data, while several smaller studies document symptom overlap with contested pathogenicity.
4.1 The Bergen Giardia Outbreak: A Natural Experiment
The 2004 waterborne outbreak of Giardia duodenalis in Bergen, Norway—affecting 1,300 laboratory-confirmed cases from contaminated drinking water—provides the most rigorous available evidence linking a parasitic infection to ME/CFS onset. Because the exposure was sudden, uniform, and documented, the cohort offers unusually clean causal inference compared to typical retrospective studies. Importantly, the entire evidence base derives from one research group at the University of Bergen; no independent replication in a separate Giardia cohort has been published. The Bergen group’s series of longitudinal studies collectively traces the natural history of post-parasitic ME/CFS from the acute phase to a decade later.
Three-year follow-up. At three years, chronic fatigue prevalence in exposed individuals substantially exceeded background rates . A clinical case series from the same cohort found that at least 5% of all laboratory-confirmed cases developed CFS meeting Fukuda criteria, with a progressive worsening course in the majority (57%) at specialist referral (Naess et al. 2012). The SF-36 profile showed disproportionate impairment in physical functioning, vitality, and social functioning with relatively preserved mental health—a pattern consistent with ME/CFS and distinct from primary depression.
Five-year follow-up. A structured multi-specialist evaluation at five years (psychiatry, neurology, and infectious disease) found 41.5% of persistently fatigued patients still met CFS criteria, with a further 13.2% having idiopathic chronic fatigue (Mørch et al. 2013). Sleep apnoea and depressive disorders were identified as differential diagnoses in a minority, underscoring that post-parasitic fatigue is not simply depression. Partial natural recovery was observed: 20.8% had resolved at five years, and fatigue scores improved significantly from the three-year mark.
Six-year controlled prospective follow-up. The methodologically strongest study in the series—a controlled prospective cohort with 748 exposed individuals and 878 matched population controls—found that at six years, 30.8% of Giardia-exposed reported chronic fatigue symptoms on a postal questionnaire versus controls (relative risk 2.9, 95% CI 2.3–3.4), and IBS was present in 39.4% (RR 3.4, 95% CI 2.9–3.9) (Hanevik et al. 2014). Formal ME/CFS diagnostic criteria were not applied to the full cohort; the outcome measured was self-reported chronic fatigue, not clinically diagnosed ME/CFS. Conditions declined slowly from the three-year peak but remained highly elevated: a nearly threefold excess of chronic fatigue symptom burden six years after documented parasitic infection in a population-matched controlled cohort.
Ten-year follow-up. At ten years, 26% of exposed still reported chronic fatigue versus 11% of unexposed controls (OR 3.0); IBS remained 43% versus 14% . As with the six-year data, these are self-reported symptom prevalences from a postal questionnaire, not clinical ME/CFS diagnoses. The persistence of excess fatigue burden a full decade after parasite eradication cannot be attributed to ongoing infection.
Immune mechanism. Immunological substudies of the Bergen cohort, all from the same Bergen research group, have identified two abnormalities. At five years post-infection, post-infectious CFS patients showed significantly lower NK-cell counts versus non-fatigued Giardia-exposed controls, with NK-cell levels correlating negatively with both fatigue and abdominal symptom severity (Hanevik et al. 2012). This parallels the NK cell count reductions documented in ME/CFS from other causes (Innate Immunity), though those studies primarily measured cytotoxic function rather than counts. A separate study found elevated soluble CD40 ligand (sCD40L)—a marker of platelet and immune activation—in post-infectious CFS patients versus non-fatigued controls, correlating with current fatigue severity (Hanevik et al. 2017); notably, that same study found no difference in antigen-specific CD4 T-cell responses between groups. Long-term Giardia-specific cellular immune responses are also detectable at five years (Bolstad et al. 2017). These findings are consistent with immune dysregulation persisting after parasite clearance, though causality has not been demonstrated and all immunological substudies originate from the same Bergen research group (n=15–19 per study), limiting independent replication.
The Bergen waterborne Giardia outbreak constitutes the primary evidence base for parasitic ME/CFS triggers: a defined exposure event, laboratory-confirmed pathogen, matched controls, and longitudinal follow-up to ten years — all from one research group at the University of Bergen, with no independent replication in a separate Giardia cohort. A controlled prospective study found a threefold excess risk of chronic fatigue symptoms (self-reported, not formally diagnosed ME/CFS) at six years (Hanevik et al. 2014); at ten years, 26% of exposed still had chronic fatigue versus 11% controls . At least 5% of all laboratory-confirmed cases developed clinically diagnosed ME/CFS (Naess et al. 2012). Immunological substudies (n=15–19 per study) found reduced NK-cell counts (Hanevik et al. 2012) and elevated sCD40L (Hanevik et al. 2017) persisting years after parasite clearance, consistent with ongoing immune dysregulation (Bolstad et al. 2017).
4.2 Other Intestinal Parasites with Symptom Overlap
Cryptosporidium parvum. A 1-year follow-up of an adult outbreak-associated cohort after acute Cryptosporidium parvum infection found persistent fatigue in 22%, joint pain in 33%, and IBS-consistent symptoms in 28% (Stiff et al. 2017). The multi-system symptom cluster—extending beyond GI symptoms to joint and fatigue phenotypes—is consistent with post-infectious ME/CFS, though formal CFS diagnostic criteria were not applied and follow-up was limited to 12 months. This extends the post-infectious trigger spectrum beyond Giardia to a second waterborne protozoan, though the evidence requires replication with longer follow-up and formal diagnostic ascertainment.
Blastocystis spp. Blastocystis is a highly prevalent intestinal protozoan of contested pathogenicity. Some carriers remain entirely asymptomatic, while symptomatic individuals report gastrointestinal and fatigue symptoms. A 2013 clinical letter reported an in-house comparison of ME/CFS patients positive for Blastocystis or Dientamoeba fragilis (Dunwell 2013), but no controlled trial data exist in ME/CFS cohorts. The pathogenicity of Blastocystis remains genuinely disputed in the scientific literature, with some evidence suggesting it may be commensal in many hosts, making individual-level attribution difficult.
Dientamoeba fragilis. Dientamoeba fragilis follows a similar pattern: associated with gastrointestinal symptoms and fatigue in symptomatic carriers, with disputed pathogenicity in immunocompetent hosts. No ME/CFS-specific controlled studies exist. Diagnosis is complicated by the limited sensitivity of standard stool microscopy compared to PCR-based multiplex testing, though the latter is not routinely ordered during ME/CFS clinical evaluation.
Toxoplasma gondii. Latent toxoplasmosis is associated with neuropsychiatric sequelae and should be considered in the differential diagnosis before an ME/CFS diagnosis is established in patients with relevant exposure history. It is not considered an established ME/CFS co-trigger in the current literature.
Trigger spectrum context. Parasitic infections represent a minority of ME/CFS triggers. A large multi-country survey of 1,773 ME/CFS patients found that 60.3% reported a preceding infectious illness, with mononucleosis (EBV) the single most common trigger (30% of infections), distributed across more than 100 distinct pathogens (Jason, Yoo, and Bhatia 2022); no single pathogen dominates the trigger landscape. The Bergen cohort data illustrate that even relatively uncommon parasitic triggers can generate post-infectious ME/CFS in a substantial minority of exposed individuals.
4.3 The Inciting Trigger vs. Sustaining Mechanism Distinction
A conceptual distinction is essential for interpreting the parasite literature and for understanding why conventional medicine may appear to “ignore” this field.
Medical management of chronic illness appropriately focuses on sustaining mechanisms—what maintains disease activity in an established patient today. A patient with ME/CFS of ten years’ duration who tests positive for Blastocystis is unlikely to recover from ME/CFS after eradication, because by this stage the sustaining mechanisms (immune dysregulation, autonomic dysfunction, mitochondrial damage) are hypothesized to be operating independently of the original trigger. From the clinician’s perspective, treating the parasite in this context has low expected benefit, and the Bergen data support this: ME/CFS persists a decade after confirmed parasite eradication.
This reasoning is clinically sound for established, longstanding disease. However, it creates a systematic blind spot for two important questions:
- Parasites as inciting triggers: A parasite may initiate the post-infectious immune dysregulation cascade that eventually becomes self-sustaining ME/CFS, even if the parasite itself is subsequently cleared. In this model, the parasite is causally essential (it started the process) but therapeutically irrelevant by the time ME/CFS is diagnosed.
- The early treatment window: If parasites act as inciting triggers, aggressive treatment at the time of acute infection—before immune dysregulation becomes self-sustaining—might prevent ME/CFS from establishing. This early therapeutic window is unstudied.
4.4 Diagnostic Gap and Clinical Implications
A practical diagnostic gap exists: standard stool ova and parasite (O&P) microscopy examination has limited sensitivity for Dientamoeba fragilis and Blastocystis. PCR-based multiplex stool testing is substantially more sensitive but is not routinely ordered during ME/CFS clinical evaluation. This creates a systematic underdetection of potentially relevant findings in patients whose ME/CFS onset followed gastrointestinal symptoms.
If intestinal parasites act as inciting triggers for post-infectious ME/CFS via immune dysregulation, does aggressive early treatment of acute parasitic infection reduce subsequent ME/CFS incidence? The Bergen Giardia cohort provides the ideal test case—what proportion of the CFS cases in that cohort received timely, confirmed eradication therapy, and does treatment timing correlate with chronic outcome? A prospective cohort study tracking ME/CFS development after documented gastrointestinal parasitic infection, with early vs. delayed treatment arms, could directly address this therapeutic window hypothesis. No such study has been conducted.
What is the prevalence of Blastocystis, Dientamoeba fragilis, and other intestinal protozoa in ME/CFS patients compared to healthy controls, when assessed by sensitive PCR-based multiplex testing rather than standard microscopy? If prevalence is elevated, does treatment of symptomatic carriers with gastrointestinal phenotype improve ME/CFS outcomes? The absence of such a study is a gap in the differential diagnosis literature, particularly given the well-documented gut-brain axis dysfunction in ME/CFS (Gut Microbiome Alterations).
4.5 Emerging Hypotheses from the Parasite–ME/CFS Interface
The Bergen data and the trigger-vs-sustaining framework generate several mechanistically grounded hypotheses that connect intestinal parasitic infection to known ME/CFS pathways.
Certainty: 0.30. Mechanistically coherent and consistent with the established autoimmune hypothesis for ME/CFS (autoantibodies), but no epitope homology data or autoantibody profiling in Bergen cohort patients has been published. Certainty is limited by absence of direct evidence.
Giardia duodenalis (syn. G. lamblia) expresses a large family of variant-specific surface proteins (VSPs) that undergo rapid antigenic variation to evade immune clearance. We speculate that certain VSP epitopes may share structural homology with human G-protein-coupled receptor epitopes — particularly muscarinic and beta-2 adrenergic receptors already implicated in ME/CFS autoimmunity . Acute Giardia infection could prime autoreactive B-cell clones via molecular mimicry; after parasite clearance, these clones might persist and produce receptor-blocking autoantibodies, driving the autonomic and vascular dysfunction of ME/CFS. This model would explain why post-Giardia ME/CFS clinically resembles post-viral ME/CFS: both would converge on the same GPCR autoimmune endpoint via different antigenic routes.
Testable predictions: (1) Anti-muscarinic and anti-β2-adrenergic receptor autoantibodies are elevated in Bergen cohort ME/CFS patients vs non-ME/CFS exposed and unexposed controls, replicating the Loebel 2016 pattern in a parasitically triggered cohort. (2) VSP structural homology to human GPCR epitopes is detectable by computational proteomics (e.g., BLAST, HHpred) at an E-value threshold ≤ 1×10⁻⁴; failure to detect homology above this threshold at both sequence and predicted-structure levels constitutes disconfirmation of this specific mechanism. (3) Immunoadsorption produces clinical improvement in post-Giardia ME/CFS patients with confirmed elevated GPCR autoantibodies, as has been shown for other ME/CFS autoimmune subgroups .
Certainty: 0.35. The IDO1–kynurenine axis has been implicated in ME/CFS neuroinflammation in preliminary studies, and IDO1 upregulation during intestinal parasite infection has been proposed mechanistically, but no data directly link persistent IDO1 upregulation to post-Giardia ME/CFS specifically.
Intestinal parasite infection, including Giardia, is hypothesized to upregulate indoleamine 2,3-dioxygenase (IDO1) in intestinal macrophages—the rate-limiting enzyme diverting tryptophan into the kynurenine pathway rather than toward serotonin synthesis—though direct demonstration in the Bergen cohort is lacking. The kynurenine pathway is already implicated in ME/CFS: excess quinolinic acid (a neurotoxic NMDA receptor agonist) and deficient serotonin precursor availability are proposed contributors to cognitive dysfunction, fatigue, and disordered sleep (Kynurenine Pathway and Quinolinic Acid Excitotoxicity: The “Fog Machine”). We speculate that parasitic infection may epigenetically prime IDO1 expression in gut-associated macrophages into a durably upregulated state, by analogy with the H3K4me3-mediated trained innate immunity mechanism described for other macrophage-activating stimuli (Netea, Quintin, and Meer 2011); the specific H3K4me3/IDO1 connection has not been demonstrated in the context of Giardia infection. After parasite clearance, persistently elevated IDO1 activity maintains a chronic kynurenine shunt — explaining ME/CFS fatigue, cognitive dysfunction, and neuroinflammation as downstream sequelae of a gut enzyme reprogrammed by an infection that is long gone.
Testable predictions: (1) Serum kynurenine/tryptophan ratio is elevated in Bergen cohort ME/CFS patients vs non-ME/CFS exposed at 5 and 10 years. (2) H3K4me3 enrichment at the IDO1 promoter in intestinal macrophage biopsies is greater in post-Giardia ME/CFS patients than controls. (3) IDO1 inhibitors (e.g., epacadostat) reduce kynurenine/tryptophan ratio and improve cognitive symptoms in post-parasitic ME/CFS.
Selective or partial secretory IgA (sIgA) deficiency predisposes to protracted and recurrent Giardia infection. If sub-clinical mucosal IgA insufficiency — not reaching the threshold for clinical diagnosis of selective IgA deficiency, but functionally reduced at intestinal mucosal surfaces — is more prevalent in individuals who converted to ME/CFS after Bergen Giardia exposure, this would: (1) identify a measurable immunogenetic risk factor for parasite-triggered ME/CFS, (2) explain the ~5% conversion rate, and (3) identify a potential prevention target. Serum IgA and secretory IgA measurement in the Bergen cohort ME/CFS subgroup vs non-converters would directly test this at low cost.
Certainty: 0.20. Mechanistically coherent but entirely inferential; no direct evidence from ME/CFS or post-Giardia studies. Speculative integration of Bergen NK findings with GALT immunotolerance biology.
Giardia colonizes the duodenum and proximal jejunum — the anatomical site of the intestinal immune checkpoint where NK cells in the gut-associated lymphoid tissue (GALT) normally suppress autoreactive lymphocyte clones. The Bergen immunophenotyping data show chronically reduced peripheral NK cell counts in post-infectious ME/CFS patients (Hanevik et al. 2012). We speculate that GALT-resident NK cell depletion during or after Giardia infection may impair this local tolerogenic checkpoint. Giardia cysteine proteases are known to degrade secretory IgA, reducing a key mucosal barrier; combined with NK depletion, autoreactive clones primed by parasite antigens (potentially via molecular mimicry as in Mucosal Checkpoint Failure: GALT NK Depletion Permits Autoreactive Escape) could escape into systemic circulation, driving GPCR autoimmunity or other autoimmune mechanisms observed in ME/CFS (autoantibodies). This “mucosal checkpoint failure” model would predict that post-Giardia ME/CFS patients show evidence of a broader autoimmune diathesis, not limited to GPCR antibodies.
Testable predictions: (1) GALT NK cell counts (intestinal biopsy) are lower in post-Giardia ME/CFS patients vs recovered Giardia-exposed controls. (2) A broader panel of autoantibodies (ANA, anti-thyroid, anti-GPCR) is elevated in post-Giardia ME/CFS vs post-Giardia non-CFS. (3) Reconstitution of mucosal NK cell function (mechanism to be defined) reduces autoantibody titres in a relevant animal model.
Certainty: 0.20. Based on known Giardia–mast cell biology and the documented MCAS–ME/CFS comorbidity, but no study has examined mast cell phenotype or tryptase levels in the Bergen cohort or any post-parasitic ME/CFS population.
Acute Giardia infection induces significant intestinal mast cell degranulation as part of the innate immune response to luminal parasites. We speculate that in a subset of susceptible individuals, this acute mast cell activation permanently primes intestinal mast cell populations toward hyper-reactivity: a lowered degranulation threshold and elevated baseline tryptase. This post-parasitic mast cell sensitization would clinically manifest as Mast Cell Activation Syndrome (MCAS, Connections to Allergies and Mast Cell Activation), which carries substantial phenotypic overlap with ME/CFS (fatigue, cognitive dysfunction, orthostatic intolerance, chemical sensitivities). Under this model, a proportion of post-Giardia ME/CFS patients may be post-parasitic MCAS cases in which mast cell mediators — histamine, prostaglandins, PAF — drive autonomic dysfunction, neuroinflammation, and immune dysregulation. The treatment implication is testable: mast cell stabilizers (cromolyn sodium, ketotifen) and combined H1/H2 antihistamine blockade represent low-risk interventions that could be trialled in post-parasitic ME/CFS patients with clinical MCAS features.
Testable predictions: (1) Serum tryptase and urinary prostaglandin D₂ metabolites are elevated in post-Giardia ME/CFS patients vs post-Giardia non-CFS and unexposed controls. (2) The proportion of ME/CFS patients meeting MCAS criteria (AAAAI 2020 consensus) is higher among those with post-parasitic onset than among those with post-viral onset. (3) H1/H2 antihistamine + cromolyn sodium trial produces symptom improvement specifically in post-parasitic ME/CFS patients with elevated baseline tryptase.
Giardia selectively colonizes the proximal small intestine — a region with normally sparse, carefully regulated microbiota that is distinct from the colonic microbiome typically studied in ME/CFS research. Wensaas et al. 2018 documented persistent duodenal histological changes (villous atrophy, intraepithelial lymphocytes) in a subset of Bergen cohort patients at ten years . We hypothesize that Giardia infection may permanently shift the small intestinal (duodenal) microbiome composition — reducing butyrate-producing species and increasing LPS-presenting gram-negative bacteria — resulting in chronic low-grade LPS translocation, TLR4 activation, and systemic neuroinflammation as a sustaining ME/CFS mechanism. This would connect the parasite trigger mechanistically to the gut–brain axis and microbiome abnormalities documented in ME/CFS (Gut Microbiome Alterations). Unlike colonic microbiome studies (which have been conducted in ME/CFS, see Gut Microbiome Alterations), duodenal microbiome profiling via duodenal aspirate or proximal small-bowel biopsy has not been performed in any post-parasitic ME/CFS cohort.
The Bergen Giardia cohort represents one of the highest-value underutilized research resources in ME/CFS science: a cohort with a known exposure date, laboratory-confirmed pathogen, longitudinal clinical follow-up to 10 years, and a well-characterized PI-CFS subgroup. Existing studies have measured cytokine profiles, NK counts, sCD40L, and cellular immune responses — but the cohort has not been systematically interrogated for: (1) GPCR autoantibody panels (anti-muscarinic, anti-β2-adrenergic) as in Mucosal Checkpoint Failure: GALT NK Depletion Permits Autoreactive Escape; (2) duodenal biopsy for epigenetic profiling (H3K4me3 at IDO1, IL-6, and TNFα promoters); (3) small intestinal microbiome composition by shotgun metagenomic sequencing; (4) serum tryptase and urinary mast cell mediators; (5) secretory IgA levels. Recontacting surviving PI-CFS patients and matched non-CFS exposed controls for a cross-sectional biomarker panel study would simultaneously test multiple mechanistic hypotheses at low marginal cost, given that the cohort is already characterized. This is likely the highest return-on-investment study available in post-parasitic ME/CFS research.
The Stiff 2017 data establish that 22% of Cryptosporidium parvum outbreak survivors report persistent fatigue at 12 months (Stiff et al. 2017), but follow-up was insufficient and formal ME/CFS criteria were not applied. A Bergen-style prospective controlled cohort study following a documented Cryptosporidium outbreak for 3–10 years with formal ME/CFS diagnostic ascertainment would determine: (1) whether the Giardia findings generalize to a second apicomplexan protozoan; (2) whether the ME/CFS conversion rate differs between parasite species; (3) whether the immunological profile (NK cells, sCD40L, antigen-specific responses) recapitulates the Bergen findings or differs systematically. Given that Cryptosporidium outbreaks occur recurrently in Europe and North America, an opportunistic prospective cohort design is feasible. No such study has been initiated.
5 Infection-Induced Cumulative Damage and Disease Progression
The progressive, often step-wise deterioration seen in many ME/CFS patients following repeated infections or viral reactivation suggests that each infectious event produces cumulative, irreversible damage rather than merely triggering reversible inflammation.
Certainty: 0.40. This model is mechanistically plausible and consistent with clinical observations of step-wise deterioration after infections, and is supported by preliminary Long COVID epidemiological data (Eckey et al. 2025). However, the irreversibility claim is difficult to test prospectively, and alternative explanations (e.g., deconditioning, psychological factors in rating, or regression to the mean) have not been excluded. The certainty is limited by the absence of controlled longitudinal data directly testing the ratchet prediction.
We propose that each infection in ME/CFS patients produces cumulative, irreversible damage that progressively worsens disease through multiple interconnected mechanisms (Eckey et al. 2025).
Cumulative damage mechanisms
Viral reactivation and persistent viral load: ME/CFS patients frequently experience reactivation of latent viruses (EBV, HHV-6, CMV) or recurrent infections with new pathogens. Each reactivation adds to the total viral antigenic load. Unlike an immunocompetent host who clears viruses completely, ME/CFS patients with impaired immune function may never fully clear these reactivations. The viral genome and viral proteins (which are inherently immunogenic and inflammatory) persist or accumulate. This creates a ratchet effect—viral burden goes up with each reactivation and rarely returns to baseline. A specific cellular mechanism for this persistence—virus residing within long-lived lymphocytes themselves, below conventional detection thresholds—is proposed in Speculation Lymphocyte Reservoir Ratchet, supported by comparative virology evidence from feline coronavirus .
Additional microglial priming events: As described in the neuroinflammatory cascade model (Hypothesis Neuroinflammatory Cascade: From CNS to Peripheral Symptoms), each infection represents a major microglial priming event. Acute infections trigger intense microglial activation, and subsequent viral reactivations produce additional priming. Since primed microglia show exaggerated responses to subsequent stimuli (as discussed in the PEM kindling hypothesis, Hypothesis Post-Exertional Malaise Kindling and Progressive Sensitization), each infectious episode not only causes direct damage but increases the microglial response to future infections. This creates a positive feedback: infection → microglial priming → exaggerated response to next infection → more priming.
Critical note on model interdependence: This ratchet model shares core mechanistic assumptions with the Kindling Hypothesis (Hypothesis Post-Exertional Malaise Kindling and Progressive Sensitization) and the neuroinflammatory cascade model (Hypothesis Neuroinflammatory Cascade: From CNS to Peripheral Symptoms); they should be interpreted as complementary components of a unified explanatory framework rather than independent corroboration of each other. Specifically: the Ratchet Model predicts irreversible step-wise decline with each infection, while the Kindling Model predicts progressive threshold reduction from exertion triggers. Both could be simultaneously true (infections cause larger priming steps; exertion causes smaller priming steps), but they make distinct testable predictions. The Ratchet Model uniquely predicts that baseline functioning follows a ratchet pattern (asymmetric: high damage but minimal recovery), whereas pure Kindling predicts threshold reduction independent of infection status. Distinguishing between these predictions empirically requires longitudinal threshold tracking with separate quantification of crash frequency from infections versus non-infectious triggers.
Further depletion of metabolic reserves: During acute infection, energy expenditure increases substantially due to fever, immune activation, and metabolic stress. In ME/CFS patients, metabolic reserves are already depleted. Each infection represents a major metabolic stress that exhausts remaining reserves. Unlike immunocompetent hosts who recover metabolically after infection, ME/CFS patients may never fully restore their metabolic baseline before the next infection occurs. The metabolic nadir becomes progressively lower with each infection.
Immune exhaustion from repeated activation: The adaptive immune system responds to each infection by activating clones of T cells and B cells specific to the infection. In the context of persistent and recurrent infections, these same clones are repeatedly activated. Repeated activation produces immune exhaustion—T cell exhaustion markers increase, B cell function declines. Additionally, the repeated need to generate immune responses may accelerate telomere shortening and cellular senescence, reducing the lifespan of immune cells.
Clinical manifestation: Step-wise baseline deterioration
The combination of these mechanisms produces a characteristic clinical pattern: each infection is followed by a step-wise decline in baseline functioning that does not fully resolve before the next infection. A patient might experience:
- Baseline functioning: Level A (e.g., able to work 4 hours daily)
- Infection 1 → acute illness → recovery to baseline attempt, but only reaches Level B (3 hours daily) due to incomplete metabolic recovery and persistent microglial priming
- Infection 2 → acute illness → recovery to attempted baseline, but only reaches Level C (2 hours daily)
- Infection 3 → acute illness → recovery to attempted baseline, but only reaches Level D (bedbound)
This step-wise progression differs from other chronic conditions where infections represent temporary setbacks from which full recovery to baseline is expected. In the ratchet model, each infection represents a permanent downward step in baseline capacity. Over years, repeated infections can convert a mildly-affected patient into a severely-affected patient, even if individual infections are not severe.
Key clinical implications
Infection prevention is disease-modifying: In the ratchet model, preventing infections is not merely symptomatic management but disease-modifying therapy. Each prevented infection preserves baseline functioning and prevents another step-wise decline. A patient with effective infection prevention can potentially avoid progressive deterioration that would occur with repeated infections.
Prophylactic interventions are justified: Standard infection prevention approaches (masking during high-transmission periods, hand hygiene, limiting exposure to ill contacts) might be expected in moderately or severely affected patients. More aggressive approaches—such as FFP2 masking in community settings during respiratory season, or prophylactic antiviral therapy during high-risk periods if safe options become available—could potentially have substantial long-term benefit by preventing cumulative damage.
Rapid infection treatment is critical: Early, aggressive treatment of identified infections (rapid antiviral therapy for herpesvirus reactivation, prompt antibiotic therapy for bacterial infections) might minimize the damage window and reduce the microglial priming response by shortening infection duration.
Immunological intervention may not restore lost function: In conditions where immune deficiency is reversed (e.g., HIV treatment restoring CD4 counts), patients often improve dramatically because the deficit was reversible. In the ratchet model, infection-induced damage is largely irreversible. Therefore, immunological interventions (such as immune modulation or restoration) might prevent future decline but would not restore previously-lost baseline functioning. This suggests that prevention is substantially more important than treatment—once damage is done, it persists.
Relationship to baseline deterioration in Long COVID
The ratchet model of cumulative infection-induced damage provides a mechanistic framework that explains the step-wise baseline deterioration observed in Long COVID patients experiencing recurrent COVID-19 infections. Preliminary epidemiological data and clinical observations suggest that each COVID reinfection produces additional baseline functional loss beyond what would be expected from reinfection alone. This pattern aligns with the infection ratchet hypothesis and suggests that similar mechanisms may apply to ME/CFS (Eckey et al. 2025).
The ratchet model’s central claim—that each infection produces irreversible baseline decline—has not been tested prospectively. Key data gaps:
- No longitudinal study has tracked ME/CFS patients through sequential infections with pre- and post-infection functional measurement to quantify stepwise decline versus recovery.
- The model shares core assumptions with the Kindling Hypothesis and the neuroinflammatory cascade model; these cannot serve as independent corroboration of each other.
- Alternative explanations for observed step-wise decline (deconditioning during acute illness, psychological reporting bias, regression to the mean after flare) have not been excluded.
- The claim that “immunological intervention may not restore lost function” is unfalsifiable without first identifying which damage is structural versus functional.
Certainty: 0.40. Each component of this pathway is separately documented in ME/CFS, but the unified causal chain has not been directly tested; the certainty reflects strong biochemical plausibility offset by the absence of any study linking pathogen burden to CoQ10 levels.
Plasma coenzyme Q10 (CoQ10) has been reported to be reduced in ME/CFS, with lower levels correlating with fatigue, autonomic, and neurocognitive symptom severity (Maes et al. 2009). CoQ10 functions not only as an electron-transport-chain carrier but as a regenerable lipophilic antioxidant that is oxidatively consumed when it quenches lipid peroxyl radicals; it is normally regenerated by the electron transport chain and by other antioxidants such as glutathione and vitamin E (Morris et al. 2013). We propose that persistent or reactivating infection contributes to the documented CoQ10 deficiency by imposing a sustained reactive-oxygen-species (ROS) load that consumes CoQ10 faster than it can be regenerated.
The mechanistic ingredients are individually established: ME/CFS shows elevated lipid peroxidation (malondialdehyde-modified adducts and related oxidative-stress markers) (Maes and Leunis 2014) (Maes et al. 2021); inflammation and ROS damage the electron transport chain in a self-amplifying feed-forward loop that further increases oxidative consumption (Morris and Maes 2014); and viral reactivation (e.g. HHV-6) drives mitochondrial fragmentation and increased ROS in cell models (Schreiner et al. 2020). A mechanistically distinct but broadly analogous precedent comes from Long COVID, where infection severity predicts sustained antioxidant depletion months later Al-Hakeim et al. (2023) — a single-trigger model rather than the persistent/reactivation model proposed here, but consistent with the principle that a viral trigger can induce lasting antioxidant drawdown. Each citation supports its own link; none establishes the conjunction, which remains untested. Within the infection-ratchet framework above (Speculation Infection-Induced Irreversible Damage: The Ratchet Model), accelerated CoQ10 consumption is one candidate mechanism by which each infectious episode depletes metabolic reserves.
Falsifiable prediction: In ME/CFS patients, markers of active pathogen burden (e.g. EBV or HHV-6 viral load, or a validated composite infection-activity index) will correlate inversely with plasma and, where measurable, tissue CoQ10 levels, and positively with lipid-peroxidation markers; and effective suppression of viral reactivation will be followed by a measurable rise in CoQ10 or fall in lipid-peroxidation markers. The hypothesis is refuted if CoQ10 deficiency is present without elevated oxidative-consumption markers, or if pathogen burden and CoQ10 levels are uncorrelated after adjustment for confounders.
Competing explanation: The deficiency may instead arise from impaired CoQ10 biosynthesis — primary genetic CoQ deficiencies from COQ-pathway mutations produce a static, non-consumption-driven low state (Laredj, Licitra, and Puccio 2014) — or from generalized downregulation of mitochondrial biogenesis, statin use, or gastrointestinal malabsorption. These are not mutually exclusive with oxidative consumption and have never been tested against it in the same ME/CFS cohort. This mechanism is developed with equal seriousness in Open Question Is ME/CFS CoQ10 Deficiency Driven by Oxidative Consumption or Impaired Biosynthesis?, where the two are treated as co-equal candidates.
Replication status: Not yet replicated as a unified causal chain; individual components (low CoQ10, elevated lipid peroxidation) are supported by more than one study but the causal link to infection is untested.
Consequence: If infection-driven consumption is confirmed to be a real contributor, then simply supplementing CoQ10 without addressing the underlying infectious/oxidative source may be like refilling a leaking tank — controlling the infection (if an active infection is present) or the oxidative load could matter as much as replacing the depleted CoQ10; if instead the deficiency is a biosynthesis problem, supplementation alone would be the more logical fix. This does not mean supplementation is useless — mixed trial results leave that question open, and for an individual patient a trial of CoQ10 may still be warranted regardless of mechanism. Distinguishing the two would change how clinicians approach the underlying drivers (antiviral therapy, if indicated, versus antioxidant strategies), though the question is currently unresolved.
The documented reduction in CoQ10 in ME/CFS (Maes et al. 2009) is consistent with at least two distinct upstream mechanisms that current data cannot distinguish: (a) accelerated oxidative consumption driven by ongoing infection and ROS (Speculation Infection-Driven Oxidative Consumption as a Contributor to CoQ10 Depletion), and (b) impaired biosynthesis — whether from COQ-pathway genetic variation (Laredj, Licitra, and Puccio 2014), general mitochondrial-biogenesis downregulation, or secondary factors such as statin exposure or malabsorption. The two make different clinical predictions but have never been measured against each other in the same cohort.
What would resolve it: a study measuring, in the same ME/CFS patients, (1) CoQ10 consumption/turnover kinetics rather than a single static plasma level, (2) pathogen-activity biomarkers, (3) expression of COQ-biosynthesis genes, and (4) tissue (not only plasma) CoQ10. A consumption-dominant picture predicts high turnover with elevated oxidative markers and intact biosynthesis genes; a biosynthesis-dominant picture predicts low turnover with reduced biosynthetic-gene expression. Mixed pictures — concurrent consumption and biosynthesis impairment — are likely, so the study design should model both mechanisms simultaneously with quantitative partitioning rather than treating them as mutually exclusive alternatives.
Consequence: Answering this determines whether the priority for a CoQ10-deficient patient is controlling an infection/oxidative source versus straightforward repletion — a fork that current evidence leaves open and that a single well-designed study could settle.
(Certainty: 0.45. Origin: brainstorm.) If CoQ10 deficiency in ME/CFS is driven by oxidative consumption rather than impaired biosynthesis, the two mechanisms should leave different signatures in the redox state of the CoQ10 pool, not merely its total quantity. CoQ10 cycles between an oxidized form (ubiquinone) and a reduced antioxidant form (ubiquinol); a sustained oxidative load shifts the pool toward the oxidized form as reduced ubiquinol is spent quenching lipid peroxyl radicals faster than the electron transport chain and glutathione/vitamin-E can regenerate it (Morris et al. 2013). A biosynthesis defect, by contrast, lowers total CoQ10 while leaving the reduced:oxidized ratio comparatively preserved (Laredj, Licitra, and Puccio 2014).
Caveat (important): this discriminating logic assumes intact regeneration machinery — a functional electron transport chain, glutathione, and vitamin E. In ME/CFS, electron-transport-chain dysfunction is itself documented, and impaired regeneration would depress the ubiquinol:ubiquinone ratio independently of oxidative consumption. A depressed ratio could therefore reflect a biosynthesis-or-biogenesis defect coupled to ETC damage rather than active consumption, blurring precisely the distinction the biomarker is meant to draw. Interpreting the ratio consequently requires an independent measure of ETC/regeneration capacity; the biomarker is discriminating only when regeneration is shown to be intact.
Falsifiable prediction: Measured against controls, ME/CFS patients whose deficiency is consumption-driven will show a depressed ubiquinol:ubiquinone (reduced:oxidized) ratio that co-varies with lipid-peroxidation markers (e.g. malondialdehyde adducts (Maes and Leunis 2014)) and with pathogen-activity markers, whereas biosynthesis-driven cases will show low total CoQ10 with a near-normal ratio. A paired oxidized/reduced CoQ10 assay plus an MDA/CoQ10 “consumption index” would separate the groups. The prediction is refuted if the redox ratio is normal despite elevated lipid peroxidation, or if it does not track oxidative-stress markers.
Replication status: Not yet tested in ME/CFS — most published CoQ10 measurements report total plasma CoQ10 only, not the redox ratio.
Consequence: If it works, a paired blood test measuring the ratio of CoQ10’s two forms (rather than just how much is present), interpreted alongside a measure of electron-transport-chain function, could help tell a clinician whether a patient’s low CoQ10 comes from an infection burning through it or from the body making too little — pointing to different treatments — but this is a research proposal, not a validated test, and the redox ratio is not interpretable on its own.
(Origin: brainstorm.) Symmetry requires stating the case against this whole line of reasoning. Several features weaken the claim that infection-driven CoQ10 depletion is mechanistically central to ME/CFS:
- The reported plasma CoQ10 reduction rests largely on a single cohort (Maes et al. 2009) (n=58 vs 22, lower-tier journal) and has not been widely independently replicated with tissue-level measurement.
- Lowered antioxidants and raised lipid peroxidation are a generic consequence of chronic inflammation seen across many conditions (including depression (Maes 2011)); they may be nonspecific markers of illness rather than ME/CFS-specific drivers.
- Even if CoQ10 is genuinely consumed by infection-driven ROS, the deficit could be quantitatively minor relative to other bioenergetic defects, making it an epiphenomenon rather than a lever.
- The HHV-6→ROS link is cell-line evidence (Schreiner et al. 2020), and the strongest infection→antioxidant-depletion data come from Long COVID Al-Hakeim et al. (2023), not ME/CFS directly.
- CoQ10 supplementation trials in ME/CFS have yielded mixed results; if repletion does not reliably reverse symptoms despite correcting plasma CoQ10, the deficiency may be epiphenomenal regardless of whether it arises from consumption or biosynthesis.
If the null holds — that CoQ10 consumption has no meaningful causal role — then the infection-driven-consumption speculation (Infection-Driven Oxidative Consumption as a Contributor to CoQ10 Depletion) and the diagnostic value of the redox-ratio biomarker (The Ubiquinol:Ubiquinone Redox Ratio as a Consumption-Specific Biomarker) would both need to be retired, and CoQ10’s appearance in ME/CFS would be reframed as a downstream marker rather than a target.
Consequence: This is the honest counterweight — it warns patients and clinicians not to over-invest in CoQ10 as a cause or cure until the deficiency is shown to be more than a common, nonspecific footprint of chronic illness.
6 Infection Susceptibility in ME/CFS
The immune dysfunction documented in ME/CFS—NK cell cytotoxicity deficits (Caligiuri et al. 1987) , CD8+ T cell exhaustion (Iu et al. 2024) (Gil et al. 2023), complement consumption (Sorensen et al. 2003), and neutrophil/monocyte abnormalities (Section TRPM3 Channelopathy: Single Research Group)—converges on a profile of impaired pathogen defense. Each of these deficits independently reduces the capacity to clear infections; their combination creates compounding vulnerability.
Clinical significance for bacterial infections. While viral reactivation is the most studied consequence of ME/CFS immune dysfunction (Section Viral Reactivation and Persistence), the implications for bacterial infection susceptibility are underexplored. NK cells contribute to early bacterial defense through direct killing and cytokine-mediated neutrophil recruitment; their functional impairment in ME/CFS (Caligiuri et al. 1987) may delay initial containment of bacterial pathogens. CD8+ T cell exhaustion—with elevated SLAMF6, SLAMF7, EOMES, and TOX markers and altered metabolic programming (Iu et al. 2024)—may reduce the adaptive immune system’s capacity to clear intracellular bacteria and coordinate immune responses to extracellular pathogens. Complement consumption (reduced C3/C4 with elevated activation fragments ) may transiently impair opsonization, though whether the terminal pathway (C5–C9) critical for killing encapsulated bacteria such as Neisseria meningitidis is compromised has not been specifically assessed (see Section Complement Dynamics During PEM: Reconciling Sorensen 2003 vs Nunes 2024).
Infection as a disease-modifying event. Within the ratchet model of cumulative infection-induced damage (Speculation Infection-Induced Irreversible Damage: The Ratchet Model), each infection in an ME/CFS patient carries dual risk: the acute infection itself and the potential for permanent baseline deterioration. The Q Fever Fatigue Syndrome literature provides the strongest evidence for this in the bacterial domain: 98.9% of 368 QFS patients experienced PEM at 10-year follow-up, with severely compromised quality of life . This provides evidence that bacterial infections can produce persistent post-infectious syndromes resembling ME/CFS and supports the clinical imperative of infection prevention as disease-modifying therapy.
7 Vaccination in ME/CFS: Efficacy and Safety
Given the immune dysfunction documented in ME/CFS, a critical clinical question is whether patients mount adequate vaccine responses and whether vaccination is safe.
7.1 Vaccine Immunogenicity
The only study directly measuring vaccine immunogenicity in CFS patients found reassuring results: Prinsen et al. showed that CFS patients mounted antibody responses to influenza vaccination identical to healthy controls, with similar seroprotection rates across all three influenza strains (Prinsen et al. 2012). CFS patients actually showed increased cellular proliferation post-vaccination. The authors concluded that “standard seasonal influenza vaccination is thus justified and, when indicated, should be recommended for patients suffering from CFS.”
The Prinsen et al. finding (Prinsen et al. 2012) is the only published study of vaccine immunogenicity in ME/CFS. It assessed influenza vaccination only; responses to other vaccines (meningococcal, pneumococcal, COVID-19) remain unstudied. Given that ME/CFS immune dysfunction varies substantially across patients and that the Iu et al. CD8+ T cell exhaustion findings (Iu et al. 2024) postdate this study, vaccine responses in ME/CFS patients with documented severe immune dysfunction may differ from those with milder phenotypes.
7.2 Vaccine Safety: Infection vs. Antigen Exposure
Population-level data from Norway provide strong evidence distinguishing infection risk from vaccination risk:
- Meningococcal vaccine: No association between meningococcal B vaccine and CFS in a case-control study (adjusted OR 1.06, CI 0.67–1.66) (Magnus et al. 2009)
- Pandemic influenza: Actual influenza infection doubled ME/CFS risk (HR 2.04, CI 1.78–2.33), while the adjuvanted pandemic vaccine showed no increased risk (HR 0.97) (Magnus et al. 2015)
These findings provide strong population-level evidence that, at least for pandemic influenza, it is infection rather than antigen exposure or adjuvant challenge that triggers ME/CFS. This has direct clinical implications: vaccination reduces infection risk without adding ME/CFS risk, making it a net protective intervention. The Magnus et al. influenza data (Magnus et al. 2015) are particularly powerful, as they tracked the entire Norwegian population from 2009 to 2012, providing statistical power unavailable to smaller studies.
COVID-19 causes persistent T cell exhaustion (elevated PD-1, TIM-3, CTLA-4) (Gil et al. 2023) and complement consumption (Klein et al. 2024). However, studies tracking pre-existing vaccine antibodies have found that COVID-19 does not cause measles-like immune amnesia (Klein et al. 2024). Whether the combination of pre-existing ME/CFS immune dysfunction with post-COVID immune damage compounds to reduce vaccine efficacy for future vaccinations is unstudied and clinically relevant: if it does, ME/CFS patients with post-COVID worsening may require additional booster doses or antibody titre monitoring.