Prodromal Detection and Prevention Research Program

The retrospective prodromal signs described in:prodromal-recognition generate a research programme aimed at prospective validation, early detection, and preventive intervention.

1 Study 1: The Sibling Stress Test β€” Prospective Metabolic Phenotyping in First-Degree Relatives

Design: Prospective cohort. N=200 asymptomatic siblings of confirmed ME/CFS patients (ages 10–25). Baseline and annual assessments:

  • Cardiopulmonary exercise testing (CPET) with 2-day repeat protocol
  • 24-hour HRV analysis (Holter)
  • PBMC respirometry (Seahorse: spare respiratory capacity)
  • Quantitative sensory testing (small fiber function)
  • Home polysomnography (sleep architecture)
  • Salivary cortisol diurnal curve
  • Serum ferritin, BH4 metabolites, methylation panel Primary question: Do siblings show a gradient of subclinical abnormalities correlating with number of prodromal signs (Retrospective Prodromal Syndrome as Subclinical Metabolic Reserve Depletion)? Can a pre-symptomatic metabolic signature predict ME/CFS development within 5 years? Falsification: If siblings show no subclinical metabolic differences from unrelated controls despite sharing genetic background, the genetic-ceiling model of Architecture C is not supported. Power calculation: N=200 siblings (assuming 15% conversion rate over 5 years) provides 80% power to detect an OR of 3.0 for metabolic signature vs. conversion, at \(\alpha = 0.05\).

2 Study 2: Wearable Transition Detection β€” Machine Learning on Pre-Diagnostic Physiological Streams

Design: Retrospective analysis of consumer wearable data (Garmin, Apple Watch, Oura Ring) from ME/CFS patients who were wearing devices before diagnosis. N=1000+ feasible via patient registries and ME/CFS biobanks. Features extracted: Resting HR trend, HRV trajectory (RMSSD), sleep duration and efficiency, step count variability, recovery time from exercise, respiratory rate during sleep. Primary question: Is the prodromal-to-disease transition abrupt (supporting phase-transition model,:cognitive-cliff) or gradual (supporting linear decline model)? Can machine learning identify the inflection point? Falsification: If no detectable signal precedes diagnosis by >3 months, the prodromal decompensation model has no wearable signature. If the transition is abrupt (within 2 weeks), this supports the hysteretic phase-transition model; if gradual (>6 months decline), this supports linear reserve erosion.

3 Study 3: 2-Day CPET Subclinical PEM β€” Prodromal Biomarker Validation

Design: Cross-sectional comparison with longitudinal follow-up. Three groups (N=50 each): (a) at-risk individuals (prodromal composite score >8, see below), (b) healthy controls matched for age/sex/activity, (c) established ME/CFS patients. All undergo 2-day CPET protocol. Hypothesis: Prodromal individuals will show a Day 2 VO2max decline of 5–10% β€” intermediate between controls (0%) and established ME/CFS (>15%). This β€œsubclinical PEM” would be an objective, quantitative prodromal biomarker. Longitudinal component: Follow at-risk group for 3 years. Primary outcome: Does Day 2 decline magnitude predict ME/CFS conversion? Expected: >5\(\\times\) conversion rate for those with 5–10% decline vs. those with 0% decline.

4 Study 4: HRV Recovery Kinetics as Population-Scale Screening Tool

Design: Cohort study using existing wearable data. Define β€œHRV recovery ratio” = (time to return to resting HRV baseline after standardised exercise) / (expected recovery time for that fitness level, derived from population norms). Ratio >1.5 = prodromal flag. Key advantage: Measurable with consumer wearables at population scale, enabling screening without clinical visits. Validation: In a cohort wearing HRV-capable devices for >12 months, test whether progressive lengthening of HRV recovery time in the 6–12 months preceding ME/CFS diagnosis is detectable. Critical discriminator: recovery ratio should distinguish pre-ME/CFS from simple deconditioning (deconditioned individuals improve recovery time with training; pre-ME/CFS individuals do not)

5 Study 5: Exercise Non-Responder Genomics

Design: Cross-reference existing exercise intervention GWAS data (HERITAGE, DREW, Gene SMART cohorts) with ME/CFS GWAS hits (DecodeME). In large exercise cohorts, approximately 20% are β€œnon-responders” to aerobic training despite adherence. Primary question: Do exercise non-responders and ME/CFS patients share genetic variants in mitochondrial adaptation pathways (PGC-1\(\alpha\) regulators, AMPK signalling, mitophagy genes)? Falsification: If non-responder genetics are entirely unrelated to ME/CFS susceptibility loci, the shared mitochondrial biogenesis failure hypothesis (Adaptation Debt β€” Training Without Supercompensation as Mitochondrial Biogenesis Failure) is not supported.

6 Proposed Prodromal Composite Score for Risk Stratification

A weighted clinical checklist for identifying at-risk individuals, pending validation:

Proposed prodromal composite score. Components weighted by specificity and mechanistic plausibility. Total range 0–18. Proposed risk tiers: 0–4 low, 5–8 moderate (Tier 1 prevention), 9–12 high (Tier 2), >12 very high (Tier 3 + urgent referral). Requires prospective validation.
Sign Weight Rationale
Training:fitness ratio >3:1 (high volume, poor gains) 2 Strongest specific signal for failed adaptation
Late-game cognitive fade (duration-dependent) 1 Common, moderate specificity
Morning fatigue despite adequate sleep duration 1 Very common, low specificity alone
Fine action tremor (no neurological Dx) 2 Unusual in healthy youth, high specificity
Cold extremities + orthostatic symptoms 1 Common, moderate specificity
Post-exertional fog >24h recovery 2 Highly specific for energy deficit
Holiday/rest crashes (paradoxical) 2 Paradoxical pattern, high specificity
Family history ME/CFS or POTS 2 Genetic loading (Architecture C)
Neurodivergence (ADHD/ASD) 1 Architecture C risk factor
Hypermobility (Beighton β‰₯5) 1 Architecture C risk factor
Frequent minor infections (>6/year) 1 Immune surveillance deficit
Unrefreshing sleep (validated scale) 1 Prospectively validated