Peter Attia MD402 ‒ NMR blood analysis: how mortality risk and more can be assessed from a single blood sample
At a glance
WHAT IT’S REALLY ABOUT
How NMR blood testing evolved into particle counts, inflammation, and mortality risk
- Jim Otvos recounts how an irreproducible 1986 NMR ‘cancer test’ led him to discover that key plasma NMR signals originate from lipoproteins, enabling quantitative NMR-based lipoprotein particle testing.
- The episode contrasts standard lipid panels (chemical assays and often-calculated LDL-C) with NMR deconvolution methods that estimate particle concentrations and subclasses (LDL-P, HDL-P) and explains why particle number better reflects atherogenic burden than particle size.
- They detail LDL-C vs LDL-P discordance findings (e.g., MESA/Framingham), emphasizing that risk tracks with particle number in discordant cases and that LDL-P/ApoB are most valuable for LDL-related risk management and therapy intensification decisions.
- Otvos describes development of LP-IR to detect insulin resistance and predict diabetes earlier than glucose-based thresholds, framing insulin resistance as the causal driver analogous to LDL exposure in atherosclerosis.
- The discussion culminates in MVX and GlycA: NMR-derived composite markers of systemic inflammation and ‘metabolic vulnerability’ that show striking associations with mortality—including signals appearing in young, ostensibly healthy adults—along with barriers to wider clinical deployment due to commercialization and hardware challenges.
IDEAS WORTH REMEMBERING
5 ideasA debunked “cancer test” accidentally launched clinical NMR lipoprotein testing.
Otvos traced NMR lipoprotein testing back to an attempt to replicate a highly publicized but flawed 1986 “cancer NMR” claim. The reproducible signal differences turned out to be driven by lipoproteins (and TG/HDL patterns), not cancer—leading to the insight that NMR could quantify lipoprotein particle concentrations and subclasses from a single plasma spectrum.
NMR measures particle concentration; routine panels mostly measure (and often estimate) cholesterol content.
Standard lipid panels chemically measure total cholesterol, HDL-C, and triglycerides; LDL-C is commonly calculated (historically via Friedewald TG/5), not directly measured, and can be inaccurate—especially as TG rise or in metabolic dysfunction. NMR instead infers particle counts (e.g., LDL-P) by deconvolving composite spectral signals linked to particle size-dependent frequency shifts.
LDL particle number (LDL-P/ApoB) matters more than LDL size; “pattern A” doesn’t confer safety.
Small, dense LDL appears higher-risk only when risk is indexed to LDL-C; once indexed to LDL particle number, particle size adds little or no incremental risk signal. Clinically, “large/fluffy LDL is benign” is a misconception; familial hypercholesterolemia is a key counterexample (high LDL-P with large particles yet extreme risk).
When LDL-C and LDL-P disagree, risk follows particle number—most useful for treatment targeting, not broad risk scoring.
In discordant cases (LDL-C percentile doesn’t match LDL-P percentile), outcomes track more closely with LDL-P than LDL-C. Otvos emphasized that LDL-P often adds limited improvement to global risk calculators (because correlated markers like HDL-C capture some of the same variance), but it can meaningfully improve LDL-related risk management—i.e., deciding whether LDL-lowering is actually sufficient.
LP-IR reframes diabetes prevention around measuring insulin resistance before glucose becomes abnormal.
LP-IR (0–100) combines six NMR lipoprotein size/subclass parameters to detect insulin resistance earlier than glucose rises and to predict incident type 2 diabetes. Otvos argues diabetes prevention parallels ASCVD prevention: insulin resistance is the upstream causal driver, while hyperglycemia is a downstream manifestation that appears after substantial beta-cell loss.
WORDS WORTH SAVING
5 quotesSo this was nothing to do with cancer per se. It was to do with the l-lipoproteins, the lipid levels of people with cancer.
— Jim Otvos
So an alternate explanation f-for the extra risk that small, dense LDL seem to confer is that there are simply more particles rather than the size of the particles being the determining, uh, you know, characteristic of, of the atherogenicity of these particles.
— Jim Otvos
I would say that's, uh, a fallacious idea.
— Jim Otvos
Once that happens, you've lost about 50% of your beta cell function.
— Jim Otvos
MVX scores are virtually unassociated with age.
— Jim Otvos
High quality AI-generated summary created from speaker-labeled transcript.
