NMR Blood Testing: One Blood Sample, Four Different Risk Scores

A standard lipid panel doesn’t measure LDL cholesterol — it estimates it, calculating the number from total cholesterol, HDL-C, and triglycerides. Jim Otvos, the biochemist who commercialized nuclear magnetic resonance (NMR) blood testing, tells Peter Attia that running the same blood sample through an NMR scanner instead of a chemistry analyzer produces something categorically different: a direct particle count, an insulin-resistance score, an inflammation marker, and an emerging composite mortality-risk index — all from one scan.

How It Works

Otvos’s path to NMR testing started with a mistake. In 1986, a New England Journal of Medicine paper claimed NMR signal patterns in blood plasma could detect cancer. When Otvos tested plasma from healthy people, including postpartum women, the supposed “cancer signal” showed up in samples with no cancer at all. Tracing the source led him to lipoprotein particles, not tumors — and, with roughly $100,000 in funding from Siemens Medical Systems, to the discovery that NMR could distinguish VLDL, LDL, and HDL particles by size and concentration in a single spectrum.

That distinction matters because particle size and count, not just cholesterol mass, drive cardiovascular risk. Research from Ronald Krauss at UC Berkeley had already shown that people with small, dense LDL particles (so-called Pattern B) carry substantially higher cardiovascular risk than people with large LDL particles at the same LDL cholesterol level.

Otvos’s NMR method built on that work, becoming the first FDA-cleared way to directly quantify LDL particle number (LDL-P) rather than infer it. From the same spectrum, the platform also derives an NMR-based insulin-resistance score (LP-IR) and GlycA, a marker of chronic low-grade inflammation that Otvos discusses as an alternative to standard hs-CRP testing.

What This Means for You

The clinical translation was never simple. NMR spectrometers exist in every chemistry department, but, as Otvos explains, none exist in clinical laboratories — commercializing the science required building a dedicated platform, the Vantera Clinical Analyzer, which eventually won FDA clearance. Labcorp acquired Otvos’s company, LipoScience, in 2014, and he continued directing NMR-based research there until 2022.

The episode’s most-discussed discordance is between LDL-C and LDL-P: a standard panel can report a normal LDL cholesterol number while particle count runs high, particularly in people with metabolic syndrome — a gap the source podcast and a peer-reviewed review of the NMR LipoProfile both describe as a blind spot for chemistry-based testing. Otvos also pushes back on the idea that large, “fluffy” LDL particles are harmless — a framing he says the data doesn’t support once particle number is accounted for.

The newest addition to the panel, per Otvos, is the Metabolic Vulnerability Index (MVX) — a composite score built from the same NMR spectrum that incorporates markers associated with inflammation, malnutrition, and muscle wasting. Otvos describes early work suggesting MVX may flag elevated mortality risk in young, apparently healthy adults, long before those risks would show up through conventional testing.

The Pattern

This is the third Stack Archive Biohacking post this year built around blood testing that extracts more from a single draw than a standard panel: an at-home finger-prick test for Alzheimer’s biomarkers, Alamar’s dried blood spot proteomics kit, and now NMR lipoprotein analysis. The throughline, across all three, is diagnostic density — one sample, several independent risk signals — rather than any single new biomarker.

Source: Peter Attia MD

Content created with AI assistance and reviewed for accuracy.