Eve Wellbeing Weekly Scan — 2026-10-12

Standing policy: 2026-10-04-eve-wellbeing-reading-protocol. No diagnosis, no prescription, no medication advice. Alcohol: Young's agreed 1-2 weekend beers as wind-down stands — nothing here relitigates it. Modern lane only; the TCM/Kushi lane (2026-09-22-kushi-corpus-map) is never blended with this material.

1. New episodes since the 2026-10-11 scan

None this week. All six feed URLs re-verified via the iTunes lookup API (all six mints returned the skill-verified URL — no drift), all six RSS feeds re-fetched: 3,203 <item> blocks parsed (identical counts to last week: On Purpose 886, Rangan 682, Happiness Lab 278, WCDHT 646, Mel Robbins 444, Life Kit 267). The three newest items are all inside the 2026-10-11 scan, titles byte-identical. Nothing new to read. This is expected with a 1-day head start — the value this week is in the evidence stream, not the feed stream.

2. New primary evidence (each traced this pass)

2.1 Social isolation + loneliness cut ~2.2–2.6 years of total disease-free life — the mental-health share dominates

Wang X, Ma H, Heianza Y, Liang Z, Franco OH, Qi L. Associations of loneliness, social isolation and healthy lifestyle with life expectancy free of major physical disease and mental disorder. Nature Communications 17:6816, published 20 Aug 2026, 10.1038/s41467-026-74498-8, PMID 42624836. Cubic UK Biobank, n = 277,489 initially free of 7 major chronic diseases, median 13.8/13.4 yrs follow-up, 16,356 deaths.

• At age 50, vs neither-exposure, both-affected people live 2.4 (women) / 1.7 (men) fewer years free of the five physical diseases, and 3.7 (women) / 5.8 (men) fewer years free of the two mental disorders (depression, anxiety). Loneliness-only, SI-only, and both-only categories are all significant (all p < 0.001); both-together is not additive — the paper's note is the overlap between measures and residual confounding.
• Healthy lifestyle (5-factor composite) narrows the SI-associated loss but the loneliness-associated loss is NOT statistically significantly conditional on lifestyle. Meaning: lifestyle buffers the objective-isolation channel, but the felt-experience channel does not show the same statistical dependence.

Why it matters for the carer lane: a sole carer running a house with two teens is inside the isolation funnel by definition — opportunities to build new social ties are structurally scarce. This is the first fully quantified fission of the same single exposure into a physical-disease-free-LE channel and a mental-disorder-free-LE channel, with the mental side dominating for both sexes at every age (40/45/60/65 checked). Not a diagnosis — a structural pressure map.

2.2 Social-isolation + loneliness YLL, broken by lifestyle level

Gao J, Zhu Y, Yang S, Jiang J, Zhu J. Association of social isolation and loneliness with years of life lost across healthy lifestyle levels. BMC Public Health, published 11 Aug 2026, 10.1186/s12889-026-28852-6. UK Biobank, flexible parametric survival models. Complements 2.1 with the mortality (Years of Life Lost) analogue: SI alone → 1.02 YLL, loneliness alone → 0.81, SI+loneliness together → 2.21 YLL at age 40 (95% CIs exclude 0 for the total; the single-exposure CIs individually reachable across sub-analyses). Numbers reproduce the Yang et al. paper's shape — single channel, not additive, at population scale.

2.3 Loneliness/isolation → health, tested three independent ways — the physical-health leg gets weaker, not stronger

Nat Comms, published 15 Jul 2026, 10.1038/s41467-026-74758-7. Triangulation (observational + sibling-control + Mendelian randomisation) on up to 414,432 UK Biobank participants. Findings:

• Consistent evidence across all three methods: loneliness and social isolation causally affect mental health and wellbeing — and loneliness (the felt experience) also hits general health (multimorbidity, QALYs).
• NOT consistent evidence: effects on specific physical health outcomes (CVD, diabetes, stroke, blood pressure). The observational signal does not replicate under sibling-control or MR.
• Why this is the week's most important correction — and note carefully, it tightens rather than breaks the record: the fear in the "social connection is a mortality variable" reading is that the causality set is too broad. What survives triangulation is the felt channel (loneliness → mental health, → general health), not the physical-outcome channel. Loneliness does still appear on general health including death (multimorbidity, QALYs) — that should be stated alongside the null on specific physical endpoints.
• Strength: strong by triangulation standards — but the corollary is that interventions that improve felt connection are on better causal footing, than interventions that merely add social contact. Matches the "difficult ties" refinement in the 2026-10-11 scan: it is not number-of-ties, it is the quality-and-felt-experience of ties that clears the causal bar.

2.4 Chronic psychological stress and hair-cortisol — two independent methylation signatures, and they do not overlap

Miller M, Brown RD, Krautkramer C, Gogia R, Fritz H, Wylie D, Kinney K, Champagne FA. Association between chronic stress and the epigenome: Exploration of psychological and biological stress. PLOS One 21(4):e0346517, published 6 Apr 2026, 10.1371/journal.pone.0346517. UT Austin, longitudinal, 9 weeks, n = 69 healthy adults age 18–35, genome-wide salivary DNAm + averaged weekly EMA self-report stress + hair cortisol concentration (HCC).

• 956 CpGs associated with perceived chronic psychological stress; 10,335 CpGs associated with HCC — and the self-report and biomarker are not correlated with each other. Two largely non-overlapping methylation channels: perceived stress → hypomethylation; HCC → hypermethylation (including NR3C1, FKBP5, BDNF, DNMT1/3A — the canonical stress-pathway genes).
• Small n, small effect sizes, young healthy sample, both endpoints measured in saliva methylation which is an imperfect proxy for blood. Treat as mechanism mapping, not effect-size evidence.
• The counterintuitive result to flag: epigenetic age (PhenoAge) was negatively associated with both psychological stress and HCC in this small young cohort (sex- and ethnicity-dependent). This does NOT overturn the stress→accelerated-aging thread — but it shows clocks can run counter to direction-of-effect expectations at small sample sizes, which is worth remembering before any single-clock result is read as definitive.

2.5 Financial strain is the single most robust psychosocial predictor of accelerated epigenetic ageing across an age-diverse sample

Ciciurkaite G, Lee B, Peng S, Fekete M, Mitchell C, Perry BL. Psychosocial stressors, accelerated biological aging, and multiple morbidities: Evidence from an age-diverse sample. PLOS One 21(3):e0343987, published 6 Mar 2026, 10.1371/journal.pone.0343987. State-representative (Indiana) sample, n = 2,267 adults age 18–103, four psychosocial stressors (ACEs, stressful life events, chronic financial strains, everyday discrimination) + GrimAge2 + DunedinPACE clocks + six physical/mental outcomes.

• Alone, all four stressors associate with accelerated epigenetic ageing.
• Simultaneously, financial strain is the most robust predictor across all eight health outcomes (β = 0.09 on both clocks, p < 0.01/p < 0.001) — while ACEs and everyday discrimination drop below significance after mutual adjustment. The abstract's mutual-adjustment phrasing ("financial strains and discrimination emerged as more consistent") does NOT match the paper's own reported coefficients (discrimination non-significant on the clocks once financial strain is in the model) — read the coefficients, not the abstract.
• Note: Brea Perry is on this author list — the same group behind last week's PNAS "difficult ties" paper. Same Epidemiologic-Epigenetics pipeline, same clocks. Two independent papers, same mechanism family.

2.6 Inflammaging → epigenetic age acceleration, now with a causal (Mendelian-randomization) leg — mechanism under Duke-NUS two-wave immune-ageing

Ballan M, et al. (senior author Netea MG; author list assembled from the Human Functional Genomics Project citation, not read off the full paper this pass). Interferon-related inflammaging links epigenetic age acceleration to multimorbidity. Cell Genomics 6(6):101218, published 10 Jun 2026, 10.1016/j.xgen.2026.101218, PMID 41999740. Four cohorts of mixed age (BCG-PRIME n=384, iMed n=165, 300BCG n=283, 500FG n=212 — the last two mean-age 25, healthy; the first two ≥60), abstract and MR findings read from the publisher page.

• Circulating inflammatory protein panel (Olink) + DNA methylation clocks (Horvath, Hannum, GrimAge, PhenoAge) + two-sample Mendelian randomisation.
• Interferon-pathway cytokines (CXCL9, CXCL10, CCL11, IL-18) show causal MR evidence as drivers of epigenetic age acceleration and multimorbidity — not just correlation. Health-span clocks (GrimAge/PhenoAge) carry the association; lifespan clocks (Horvath/Hannum) do not.
• This is the mechanistic sentence underneath the Duke-NUS two-wave descriptive result (immune ageing at 40 and 60, PMID 42764338): the interferon path is one named causal candidate for how immune ageing accelerates. Corroborative, not independent.

3. Contradictions / refinements with existing notes

1. NOTHING falsified. No standing note, no prior scan finding, and no established evidence-check is contradicted this week. All six new papers are either new territory or refinements of threads already in the record.
2. Refinement (divides a previously-always-paired pair). "Social isolation/loneliness" has been treated as a single merged exposure in the standing notes. This week's triangulation + life-expectancy pair shows the two sub-exposures do NOT behave identically — (a) healthy lifestyle buffers the SI-associated loss but not the loneliness-associated one; (b) under sibling-control/MR the specific-physical-outcome leg vanishes for both, while loneliness (the felt side) keeps a causal route to mental health and general health. Wording for the standing note: "more social contact and feeling less lonely are not the same lever, and the evidence base for the second is currently easier to establish causally."
3. Refinement (the "financial strain of solo parenting" leg is now directly evidenced at the molecular level). Before this week, the family-load link ran through caregiver-status proxies. Ciciurkaite et al. (2026) directly measures chronic financial strain and reports it as the most robust predictor across both clocks and all six physical/mental outcomes — surviving mutual adjustment for ACEs, life events, and discrimination. Structural, not personal — worth holding as context, not as commentary on Young's specific finances.
4. Refinement ("perceived stress" is not "cortisol"). Miller et al. (2026) gives the first genome-wide longitudinal human data showing the two channels leave largely non-overlapping methylation signatures and are not even correlated. Any future reading that aggregates "→ stress" outcomes should now note which channel is being measured (self-report vs HCC).

4. Unverified / unresolved / not-yet-checkable

• Feeds parsed but no new audio this week — nothing to score for promotional framing, nothing to carry forward.
• Mel Robbins Eisenberg male-hormones follow-up — no fresh episode or primary-source trace surfaced this pass (searches returned the same general-audience secondary sources as before). Still untraced.
• Perimenopause → sleep → cortisol thread: this week's evidence pulls are (imperfect) mechanisms at the ageing end, not new perimenopause-endpoint RCTs. Unresolved; the RCT-infrastructure wait (NCT06497894, NCT06306404) stands as noted last week.
• Caregiver → grandchild cognition (ELSA grandparent study, PMID 41587188): result shows a dementia-status effect only for caregiving grandmothers, no effect for frequency, no effect for grandfathers on decline. Hard to place usefully in the carer-protocol lane (grandparent- specific, does not directly map to daily solo-carer load). Parked uninterpreted — needs a targeted pull before it is usable.
• All evidence this week is associative or MR-with-assumptions. Nothing here is a clinical recommendation or a diagnosis of any person.

5. Standing-related caveats

• All evidence is from modern lanes only — no Kushi/TCM framing is blended in, per the protocol's two-lane rule.
• All cohort studies carry healthy-worker / reverse-causation risks; UK Biobank specifically over-represents healthier, more-educated mid-life adults, so absolute numbers skew optimistic and do not transfer to younger adult carers.
• All statistical effects are population-scale. No effect here is large enough to be actionable for one household, and no effect here is small enough to dismiss relative to the structural load it describes.

Related

• 2026-10-04-wellness-podcast-evidence-base — corpus, frequency map, checks
• 2026-10-04-wellbeing-source-map — carer-under-load applied read
• 2026-10-04-eve-wellbeing-reading-protocol — standing policy
• 2026-09-22-kushi-corpus-map — traditional lane (kept separate)