The Lp(a) test: levels, units and what guidelines say

Lipoprotein(a), written Lp(a), is a lipoprotein particle that carries one apoB-100 molecule, like LDL, plus a second protein called apolipoprotein(a). Its cholesterol is reported as part of a standard LDL-C result, so a routine lipid panel does not show it separately. An Lp(a) test measures it directly.

The 2022 European Atherosclerosis Society (EAS) consensus statement describes Lp(a) as predominantly determined by genetics (more than 90%), more than any other lipoprotein. Adult levels are usually reached by about age 5 and may rise until adulthood. In men they stay relatively constant. In women they tend to rise at menopause.

Levels also differ by ancestry. The EAS statement cites UK Biobank medians of 16, 19, 31 and 75 nmol/L in people of Chinese, White, South Asian and Black ancestry respectively. The 2024 National Lipid Association (NLA) update estimates that about 1 in 5 people have high Lp(a), and notes that measurement rates are low.

• EAS 2022: Lp(a) should be measured at least once in adults. Repeat measurement is not required because it does not improve risk prediction, with exceptions that include kidney or liver disease and acute infections.
• NLA 2024: the evidence supports measuring Lp(a) at least once in every adult for risk stratification, and cascade screening of first-degree relatives of patients with high Lp(a).
• AHA/ACC 2018: a family history of premature ASCVD is a relative indication for measuring Lp(a).

Because the result is usually a one-time number, it is easy to lose. A value from years ago sitting in a scanned PDF does nobody any good at the next visit. That is the practical case for keeping it in a structured field rather than an attachment.

The EAS stresses that the link between Lp(a) and cardiovascular risk is continuous. The cut-offs below are pragmatic categories, and each source asks clinicians to read them alongside the patient's overall risk.

The EAS adds a higher marker: levels above 180 mg/dL or above 430 nmol/L identify people whose lifetime ASCVD risk is equivalent to untreated heterozygous familial hypercholesterolemia.

Lp(a) can be reported as a mass (mg/dL) or as a particle concentration (nmol/L). The two do not map onto each other with a fixed factor, because apolipoprotein(a) comes in many sizes. Its structure varies in the number of repeated segments, called kringle IV repeats, so particles of different sizes carry different mass.

The EAS panel states it does not recommend using a standard factor to convert between mg/dL and nmol/L, because assays vary extensively. It calls a factor of 2 to 2.5 a best guess where both units are in use. It also notes that assays based on polyclonal antibodies cannot report molar units and may overestimate or underestimate Lp(a) depending on isoform size. Its recommendation is to measure in molar units when available, with assays traceable to official reference materials.

What this means in practice

1. Record the value in the unit the laboratory reported it in.
2. Do not convert a patient's historical result to compare it with a new one in a different unit.
3. Read the paired guideline cut-offs (for example 50 mg/dL and 125 nmol/L) as two parallel categories, not as proof that 50 mg/dL equals 125 nmol/L for a given patient.
4. If a patient's results switch units or laboratories, treat the new value as the start of a separate series.

Every Lp(a) particle carries one apoB molecule, so Lp(a) particles are counted inside an ApoB test. Their cholesterol is also included in reported LDL-C. The EAS panel does not recommend routinely correcting LDL-C for Lp(a) cholesterol.

For a clinic, the practical point is that Lp(a) answers a different question from the rest of the lipid panel. ApoB and LDL-C are followed across draws. Lp(a) is usually measured once and carried forward as context.

LongevityEHR is an EHR for cash-pay longevity, functional and concierge clinics. It does not interpret Lp(a) or recommend treatment. It is built to keep the number, its unit and the clinic's reading of it intact.

• Units are never converted. When AI lab extraction reads an Lp(a) value whose unit does not match the unit your clinic tracks, it flags the mismatch. Release is blocked until a provider confirms, and the value releases unclassified rather than graded against the wrong band.
• A unit change splits the trend. If a patient's Lp(a) arrives in a different unit later, the trend shows a visible split with the unit labeled on each segment. Values are never mixed on one axis.
• Your clinic's bands. The clinic owner sets optimal, borderline and critical bands for Lp(a), per sex if needed. See biomarker tracking.
• Review before release. Extracted results wait for your team. Extracted values reach the patient portal only after a provider releases them.
• History stays as it was. Each released value keeps a snapshot of the range that applied that day, even if you change the band later.

If you are deciding where to draw your own bands, our guide to optimal vs reference ranges covers how to document them and who in the clinic should own them.