DAY 4

Longevity: Metabolic Health
Glucose, Insulin, ApoB, Visceral Fat

2026-05-25 · BigCat's Vitality Protocol

Evidence base for this issue: Mendelian randomization + large prospective cohorts + intervention RCTs; thresholds drawn from ADA / EAS / Attia "Medicine 3.0" consensus

SUB · Glucose Screening / Early Detection

Fasting Glucose & HbA1c — Beyond the "Reference Range"

ONE-LINE TAKEAWAY

Fasting glucose and HbA1c alone miss most early insulin resistance: by the time HbA1c reaches 5.7% ("pre-diabetes"), β-cell function has typically already lost ≥50%. Optimal ≠ "within range."

EVIDENCE

Large cohorts: ARIC, Framingham, UK Biobank all show CV events rise linearly across HbA1c 5.0–5.6%, well below ADA's 5.7% cutoff. Selvin 2010 (NEJM): HbA1c independently predicts CV events and all-cause mortality in non-diabetic populations. UKPDS: average β-cell function at type-2 diabetes diagnosis ≈ 50% residual.

SCIENCE + MECHANISM

HbA1c reflects the prior 8–12 weeks of average glucose (glycated RBC product) — cheap and stable, but with three blind spots: (1) variability blindness — the same 5.6% can come from steady 5.6 or from oscillations between 4.5 and 7.5; the latter is more vasculotoxic; (2) RBC kinetics blindness — iron-deficiency anemia, thalassemia, recent blood loss skew HbA1c falsely low or high; (3) compensation blindness — for 10–15 years insulin rises to hold glucose normal (hyperinsulinemic phase), invisible to a glucose-only panel. Fasting glucose itself reflects only nocturnal hepatic output and basal insulin, missing postprandial decompensation.

PROTOCOL

Marker	"Normal" (reference)	"Optimal" (longevity)	Alert
HbA1c	< 5.7%	4.9–5.4%	≥ 5.5% → check insulin
Fasting glucose	< 100 mg/dL	75–90 mg/dL	≥ 95 + HbA1c ≥ 5.5%
OGTT 2-hour	< 140 mg/dL	< 120 mg/dL	≥ 140 = IGT
Fructosamine	< 285 µmol/L	< 250 µmol/L	Backup when HbA1c unreliable

• Annual labs: HbA1c + fasting glucose + fasting insulin (frequently omitted, essential)
• HbA1c 5.5–6.4% or glucose ≥ 95: add 75 g OGTT with insulin at 0/30/60/120 min (Kraft test)
• Heavy menstrual bleeding or iron deficiency: HbA1c may read falsely low — cross-check with fructosamine or OGTT
• Wearing a CGM for 2 weeks reveals hidden "normal-average, high-variability" patients

FEMALE-SPECIFIC NOTE

Heavy menstrual bleeding or chronic low iron can falsely depress HbA1c, masking real glycemic status — fructosamine or OGTT cross-check is wise. A history of gestational diabetes raises lifetime type-2 risk ~7× (Bellamy 2009 Lancet) and warrants lifelong annual screening. PCOS often presents insulin resistance even at normal BMI (up to 70%) — glucose-only panels miss it; insulin and HOMA-IR are mandatory. During the perimenopausal transition, falling estrogen produces a silent upward drift in fasting glucose — a common window for unnoticed dysregulation.

COMMON MYTHS

Myth 1: "HbA1c 5.6% is normal" — reference ranges were built to diagnose disease, not to define health. Above 5.4% already sits well above the all-cause-mortality nadir.
Myth 2: "Normal fasting glucose = safe" — hyperinsulinemic compensation can hide dysglycemia for a decade; pair with insulin and postprandial data.
Myth 3: "CGMs are only for diabetics" — for non-diabetic users, a 2-week CGM is a high-leverage tool to personalize which foods/timing trigger postprandial spikes.

THIS WEEK + REFLECTION

THIS WEEK

Run the "metabolic triad" labs: HbA1c + fasting glucose + fasting insulin (the third is often skipped — request it explicitly). Compute your own HOMA-IR.

Reflection: Compare your HbA1c with the optimal range (4.9–5.4%). If yours sits at 5.5%+, is that "normal" or the earliest observable metabolic signal? Would you treat it as a red line on your 30-year longevity curve?

SUB · Insulin Resistance / Metabolic Early Screen

Insulin Resistance — The 10-to-15-Year Window Before Diabetes

ONE-LINE TAKEAWAY

Fasting insulin + HOMA-IR + triglyceride/HDL ratio is the annual "metabolic vitals" panel everyone should run. Insulin resistance sits upstream of almost every chronic disease — CV, Alzheimer's, several cancers.

EVIDENCE

Mechanism + large cohorts: Reaven 1988 (Diabetes) first unified the metabolic syndrome around hyperinsulinemia. Kraft (14,000+ OGTT insulin curves) showed 75% had abnormal insulin curves 10–20 years before diabetes diagnosis. Petersen & Shulman's mechanistic work (Cell Metab) maps IR → hepatic DNL → ectopic lipid → systemic inflammation. McLaughlin 2003 (Ann Intern Med): TG/HDL ratio is a validated IR proxy in non-diabetic populations.

SCIENCE + MECHANISM

Insulin resistance = same glucose requires more insulin to maintain. Early on, β-cells over-secrete to compensate; glucose looks normal, but the hyperinsulinemia itself does damage: more VLDL output, suppressed lipolysis, mTOR activation, lower SHBG. A Kraft test (OGTT with insulin at 0/30/60/120 min) detects elevated and delayed insulin peaks long before glucose moves — "pre-diabetes with normal glucose." Only when β-cells fail to ~50% does fasting glucose climb, by which point the system is in late decompensation.

VISUAL: Insulin vs Glucose Timing Mismatch

The silent 15 years before diabetes: insulin rises first, β-cells quietly decline, glucose moves last. Watching glucose alone = missing the entire early window.

PROTOCOL

Marker	Method	Normal	Optimal	Red line
Fasting insulin	Blood draw	< 25 µIU/mL	< 6 µIU/mL	≥ 10
HOMA-IR	(glucose × insulin) / 405	< 2.5	< 1.0	≥ 2.0
TG / HDL ratio	Standard lipid panel (mg/dL)	< 3.0	< 1.5	≥ 2.5
Waist / height	Tape measure	< 0.55	< 0.50	≥ 0.55

• Many labs report TG/HDL in mmol/L; convert: mg/dL = mmol/L × 88.5 (TG) or × 38.7 (HDL)
• Any marker above "optimal" → intervene: 150+ min/wk Zone 2 + 2× resistance training; protein 1.6–2.0 g/kg; halve refined carbs
• Recheck at 12 weeks — HOMA-IR typically drops 30–50%
• Stubborn high IR with rising weight: discuss metformin / GLP-1 agonists with your physician (case-by-case)

FEMALE-SPECIFIC NOTE

PCOS is the most common cause of IR in reproductive-age women, can exist independent of weight, and is missed without an insulin panel. Pregnancy IR peaks at gestational weeks 24–28; after gestational diabetes, annual screening should continue for ≥10 years. During the perimenopausal transition, estrogen withdrawal reduces skeletal-muscle insulin sensitivity ~10–15% (Carr 2003), so HOMA-IR drifts up even when lifestyle is unchanged — resistance training + adequate protein is the strongest reversal lever (Stacy Sims emphasizes repeatedly). Luteal-phase TG can be physiologically 15–20% higher, so retest TG/HDL during the follicular phase to reduce noise.

COMMON MYTHS

Myth 1: "I'm lean, so I can't be insulin-resistant" — TOFI (Thin Outside, Fat Inside) is common in East Asian populations; lean body with high visceral fat, hidden IR.
Myth 2: "IR is just diabetes" — it is the upstream driver of much more: ApoB, Lp(a), hepatic fat, endometrial pathology, neurodegeneration.
Myth 3: "Cutting sugar is enough" — refined carbs matter, but without resistance training + adequate protein, muscle glucose uptake stays poor.

THIS WEEK + REFLECTION

THIS WEEK

Compute two cheap IR proxies: (1) TG/HDL ratio in mg/dL (convert from your latest lipid panel); (2) waist-to-height ratio. If either exceeds the "optimal" threshold, run a 12-week resistance + Zone 2 program and recheck HOMA-IR.

Reflection: To drop HOMA-IR by 1.0, losing 5 kg of body weight vs gaining 2 kg of lean muscle — which is more sustainable on a 30-year metabolic risk curve, and why?

SUB · Lipoproteins / CV Causality

ApoB — The Causal Particle in Atherosclerosis

ONE-LINE TAKEAWAY

LDL-C measures "how much cargo is in the cars." ApoB measures "how many cars are on the road." What crashes into the artery wall is the car (particles), not the cargo (cholesterol mass). Longevity target: ApoB < 60–80 mg/dL — well below most "normal" cutoffs.

EVIDENCE

Mendelian randomization (highest causal evidence): Ference 2018 (JAMA) and Sniderman 2019 (JAMA Cardiol) established ApoB as the direct causal driver of atherosclerosis; at matched ApoB, LDL-C and non-HDL-C carry no independent predictive value. UK Biobank analysis (Marston 2022 JAMA Cardiol, ~390,000 subjects): ApoB outperforms LDL-C, especially in subjects with elevated TG, diabetes, or metabolic syndrome. Lp(a) is an additional independent genetic risk (elevated in ~20% of the population — measure once in a lifetime).

SCIENCE + MECHANISM

ApoB-100 is the surface marker of LDL, VLDL, IDL, and Lp(a) — one particle, one ApoB. These particles are small enough (<70 nm) to cross the endothelium, get oxidized, and be engulfed by macrophages to form foam cells — the start of atherosclerosis. LDL-C only counts the cholesterol inside those particles, but cholesterol-per-particle varies (metabolic syndrome typically produces small-dense LDL: many particles, less cholesterol each → LDL-C looks fine but ApoB is high and the risk is worse). Lp(a) = LDL + covalently bound apo(a), genetically determined, diet-resistant, an independent early-CV risk marker.

VISUAL: ApoB vs LDL-C — Particle Count vs Cholesterol Mass

Same LDL-C of 120: on the left, 3 big tankers (low risk); on the right, 15 speedboats (high risk). Watching LDL-C alone is failing to count the boats.

PROTOCOL

Must-test (Lp(a) at least once in lifetime; ApoB annually)
• ApoB (electrophoresis / immunoturbidimetry, cheap)
• Lp(a) once at baseline (lifetime stability ±20%)
• Carotid ultrasound + CAC score (imaging catches plaque earlier than blood markers; consider at age 35+ or with family history)

ApoB target (risk-stratified)
• General population (longevity optimization): < 80 mg/dL
• Primary prevention + any extra risk (family Hx / high Lp(a) / diabetes / smoking): < 60 mg/dL
• Established ASCVD / secondary prevention: < 50 mg/dL

Intervention ladder: diet (saturated fat <10% energy + soluble fiber 25 g) → exercise → statin (first-line, 30+ years of safety data) → ezetimibe → PCSK9 inhibitor (for high Lp(a) or refractory ApoB)

FEMALE-SPECIFIC NOTE

Premenopausal estrogen suppresses ApoB production and upregulates LDL-receptor clearance — the natural CV protection that delays ASCVD in women ~10 years on average (Mendelsohn 2005). Across perimenopause, ApoB and LDL-C can drift up 10–20% within 12 months, often dismissed as "just aging" — missing a key early-intervention window. SWAN data shows the most pronounced 12-month ApoB increase clusters around the final menstrual period (FMP). MHT improves lipids modestly and variably and shouldn't be the sole lever; CTT meta-analyses confirm statin primary-prevention benefits in women are comparable to men (overturning the older "women don't benefit" misreading).

COMMON MYTHS

Myth 1: "Normal LDL-C = safe" — 20–30% of people have LDL-C / ApoB discordance; ApoB is more accurate.
Myth 2: "Statins cause diabetes, avoid them" — RR ≈ 1.1, absolute increase ~0.2 per 100 person-years, while CV events drop 25–35%; net benefit overwhelming.
Myth 3: "Supplements (red yeast rice, omega-3, plant sterols) replace statins" — typical 5–15% reductions are insufficient for secondary prevention. Red yeast's Monacolin K is literally low-dose lovastatin but unstandardized.

THIS WEEK + REFLECTION

THIS WEEK

At your next blood draw, explicitly add ApoB + Lp(a) (often omitted from standard panels; add-on cost is low, and Lp(a) only needs measuring once in a lifetime). Log them in your long-term health record.

Reflection: If your Lp(a) ≥ 50 mg/dL (and cannot be moved by lifestyle), would you proactively lower your ApoB target to < 60 mg/dL? Is this "treating a disease that doesn't exist," or intercepting plaque before it forms?

SUB · Body Composition / Visceral Fat

Visceral Fat — The Metabolic Battleground BMI Cannot See

ONE-LINE TAKEAWAY

People with a "normal" BMI can carry substantial visceral fat (VAT) and a metabolic risk approaching obesity. Waist circumference and waist-to-height ratio + DEXA are 10× more informative than the bathroom scale. For Asian women, waist ≥ 80 cm already flags metabolic syndrome.

EVIDENCE

Large cohorts + imaging: Després 2012 (Nat Rev Cardiol) established the causal chain between VAT and CV/metabolic risk. Lovejoy 2008 SWAN (Int J Obes): during the menopausal transition, VAT increases on average 20–30% even without weight gain. INTERHEART (Yusuf 2005 Lancet), 52 countries, 27,000 subjects: waist-to-hip ratio outperforms BMI for MI prediction, adjusted population-attributable risk ~20%. CARDIA 25-year follow-up: VAT/SAT ratio independently predicts CAC progression.

SCIENCE + MECHANISM

VAT and subcutaneous fat (SAT) are functionally different endocrine organs. VAT is metabolically active: secretes IL-6, TNF-α, resistin, PAI-1, and drains directly into the portal vein → hepatic DNL ↑ → ApoB / TG ↑ and insulin resistance ↑. SAT is relatively inert; part of it (lower-body SAT) is even protective. Fat distribution matters more than total mass — this is the TOFI phenotype (lean but high VAT), especially common in East Asians whose subcutaneous storage capacity is smaller, so excess calories spill into visceral and hepatic depots (ectopic lipid). "Normal-weight metabolically obese" individuals carry CV mortality comparable to obese subjects with metabolic syndrome (Stefan 2017).

PROTOCOL

Marker	Method / threshold	Notes
Waist (WC)	F < 80 cm / M < 90 cm (Asian IDF cutoff)	1 cm above navel, end-expiration, relaxed
Waist-to-height (WHtR)	< 0.50 ("keep your waist ≤ half your height")	Robust across sex and ethnicity
DEXA body composition	VAT < ~450 g / body-fat % 18–28% (female reference)	Gold standard; every 3 years is enough
Metabolic syndrome (IDF)	Elevated WC + any 2: TG ≥ 150 / HDL F<50 M<40 / BP ≥130/85 / FBG ≥100	3/5 → full metabolic workup

• Reducing VAT beats reducing total weight — for the same 5% loss, VAT can drop 25–35% (Ross 2015 RCT)
• Strongest combination: Zone 2 aerobic 150+ min/wk + resistance 2×/wk + protein 1.6–2.0 g/kg + cut alcohol and liquid sugar
• <6 h sleep accumulates VAT even at energy balance (Hairston 2010) — the four metabolic pillars are non-substitutable
• GLP-1 agonists (semaglutide) preferentially reduce VAT vs SAT, but rebound is rapid on discontinuation

FEMALE-SPECIFIC NOTE

Premenopausal estrogen routes fat to subcutaneous depots (pear-shaped) — relatively protective. The menopausal transition drives a rapid VAT increase that is the central physiological event behind midlife metabolic deterioration in women (SWAN). Even when the scale is unchanged, a waist circumference rising 1–2 cm/year warrants attention. The 3–5 years around FMP is a golden intervention window; resistance training + adequate protein + sleep meaningfully blunts VAT accumulation (Mary Claire Haver, The New Menopause). MHT shows a moderate VAT-reducing effect; discuss individualized benefit/risk with a clinician.

COMMON MYTHS

Myth 1: "Normal BMI = healthy" — BMI does not distinguish muscle from fat nor visceral from subcutaneous; using it alone is a 1980s tool.
Myth 2: "Crunches reduce belly fat" — spot reduction does not exist; waist shrinks via whole-body energy balance and the body's preferential VAT mobilization.
Myth 3: "Low-fat diets best for visceral fat" — long-term RCTs (DIRECT, PREDIMED) show Mediterranean or low-carb diets are equal or better; adherence matters more than macronutrient ratio.

THIS WEEK + REFLECTION

THIS WEEK

Measure your waist with a tape (1 cm above the navel, end-expiration, fasted morning) and compute waist-to-height ratio. Re-measure monthly at the same time and track a 6-month trend. If you have access to DEXA, get a baseline within 3 years.

Reflection: If the scale stays flat but your waist quietly grows 1 cm/year, which number do you trust more? How does the "silent drift" of body composition fit into a 30-year health-decision framework?

📚 Deep Dive — Resources

📕 Peter Attia. Outlive: The Science and Art of Longevity. 2023. — Chapters 6–8 develop the metabolic-health, ApoB, Lp(a), and ASCVD "Medicine 3.0" frame
🎙 The Drive (Peter Attia) #224 with Allan Sniderman — ApoB causal evidence; #308 with Tom Dayspring — lipoprotein deep series
📕 Mary Claire Haver. The New Menopause. 2024. — Female-lens integration of perimenopausal VAT / IR / CV risk
📕 Robert Lustig. Metabolical. 2021. — Fructose, hepatic lipid metabolism, and the metabolic-syndrome mechanism
📄 Ference BA, et al. ApoB and CV disease. JAMA. 2019;321(4):364-373. — Key Mendelian randomization paper
📄 Sniderman AD, et al. ApoB vs LDL-C concordance. JAMA Cardiol. 2019;4(12):1287-1295.
📄 Marston NA, et al. UK Biobank ApoB analysis. JAMA Cardiol. 2022;7(3):250-256.
📄 Lovejoy JC, et al. SWAN VAT changes. Int J Obes. 2008;32(6):949-958.

🧠 Reflections

① If fasting insulin could only add one bit of information beyond HOMA-IR, what would it be?

HOMA-IR is the product of fasting insulin and glucose — a "steady-state average." But the same HOMA-IR can arise from high insulin × low-normal glucose (classic compensatory IR) or low insulin × high-normal glucose (early β-cell decline). The first calls for reduced mTOR drive and improved muscle uptake; the second is closer to β-cell stress and may need to be watched for LADA / MODY trajectories. HOMA-IR alone cannot distinguish these "same score, different type" cases — so the absolute insulin value still carries independent signal.

② Can ApoB be too low? Is there a harmful lower threshold?

Mendelian randomization suggests lifetime ApoB-LDL-C exposure and ASCVD risk are monotonically positively correlated with no clear floor. PCSK9-inhibitor trials (FOURIER, ODYSSEY) reduced LDL-C to 20–30 mg/dL with no clinically relevant cognitive or hormonal harm. But "lower is better" assumes the same mechanism extends to very low ranges, and 30+ year longevity cohorts are still maturing. Practical stance: once targets are met, chasing extremely low ApoB likely has diminishing returns, but fear of "too low" should not derail reaching the target.

③ Can lifestyle target visceral or subcutaneous fat selectively?

Not strictly, but there are preferences. Aerobic exercise (especially Zone 2) reduces VAT more efficiently than pure resistance training (Vissers 2013 meta), likely through hepatic lipid mobilization and systemic catecholamine response. Low-carb / low-fructose diets preferentially reduce hepatic ectopic lipid (Schwarz 2017). But "spot-burn your belly" is marketing — every intervention mobilizes the most metabolically available depots first, and VAT happens to be the most metabolically active, not because you trained your abs.

④ When CGM shows postprandial spikes but HbA1c is normal, should you intervene?

It depends on the person. Short-term glycemic variability (GV) independently associates with oxidative stress and endothelial dysfunction (Monnier 2006), but its causal link to hard endpoints (CV events, mortality) in non-diabetic populations is unproven. Recurrent postprandial spikes >180 mg/dL plus family history or IR markers → adjust meal composition (protein first, fiber, post-meal walk). Occasional spikes with overall GMI matching HbA1c → don't over-worry. In healthy users, CGMs can amplify "data anxiety"; the tool should serve decisions, not the reverse.

⑤ How do you spot "normal but unhealthy" on a routine panel?

Three heuristics: (1) reframe "reference range" as a population percentile — you want an optimal percentile, not just "not hospital-bound"; (2) trend > single value — HbA1c 5.2 → 5.5 → 5.7 across three years tells more than a single 5.7; (3) interpret in pairs — fasting insulin with glucose, ApoB with LDL-C, TG with HDL, waist with BMI. Pairs are more robust than any single number. Healthy longevity is not "disease-free" but "lowest-risk trajectory" — these tools translate "in range" into "how far am I from optimal."

← Back to Index