Thread by @Lpa_Doc

Ok, moving on the new one:

Saturday Morning Class #2: How is Lp(a) measured and why does it matter?

Time to try to clear up the confusion: about 15 yr after Lp(a) was discovered in 1963 by Kare Berg, Prof John Albers at UWash created an assay to measure Lp(a).
1/19

He based this on the whole “mass” of Lp(a) and all its components (apoB, apo(a), cholesterol and cholesteryl esters, phospholipids, triglyceride and carbohydrate on Lp(a). (ncbi.nlm.nih.gov/pubmed/405443)
2/19

What is mass: it is a property of a physical object that quantifies the amount of matter it contains. Unlike weight, the mass of something stays the same regardless of location. Fun fact: in Einstein’s E=MC2, M is the mass of an object. 3/19

Amazing that 5 characters could explain how most things in the Universe work.

How is Lp(a) measured: you need 3 things: 1- an antibody to detect apo(a) and 2- reference samples of absolute known values based on another method, 3- a detecting system. 4/19

For example, an Lp(a) value of 10 mg/dL will give 10 arbitrary units (color, light, radioactivity, etc depending on method), a value of 100 mg/dL will give 100 units. The patient’s sample is then taken, if it has 23 units, the value give is 23 mg/dL, etc. 5/19

This method, and John, became a shooting star (“Johnny made a record, went up to number 1”: , great song…). 6/19

Problem #1: patients have different proportions of each component and the antibody in the assay only binds one component (apo(a)), but the standards are based on 7 components, so error of measurement could be significant 7/19

Problem #2: the mass assay, in mg/dL, gets entrenched clinically so physicians understand it and in labs measure it. Inertia sets in not to innovate. Unlike Johnny’s untimely demise in the Bad Company song, the Lp(a) mass assay had many more #1s. 8/19

Problem #3: unlike most other proteins, Lp(a) has multiple “kringles”, they look like a string of identical pearls. (Fun fact; kringle comes from a Danish pastry folded over twice. Pic from on recent trip to Solvang CA). 9/19

Most antibodies recognize all these kringles, thus bind more than once which is a problem. They can have an artifact due to this, higher signal or lower signal, relative to the reference standards, adding a second layer of error 10/19

So, 15 years after this comes along Dr Santica Marcovina, also at UWash, who created an apo(a) particle concentration method that only measures apo(a) component and only binds once per particle, and validates the standards to this, so no over or under-reading 11/19

This measures the number of circulating apo(a) particles only, and not the other components, thus true reflector of Lp(a) ncbi.nlm.nih.gov/pubmed/7533064. (Fun fact: there is a lovely story behind this, where the understudy overtakes the Professor, but it ends well for both). 12/19

This is reported as nanomoles apo(a) per liter of plasma (nmol/L or nM). This has the least bias in accurately reading values. There are other contributors to this story but not enough space to describe, hope no one is offended 13/19

Problem #4: this assay is only available at UWash, but is used to certify commercial systems.

A third measure of “Lp(a)-cholesterol” is reported in mg/dL, only estimating the cholesterol content of Lp(a), but this particular methods is highly inaccurate. 14/19

ncbi.nlm.nih.gov/pubmed/27919356 15/19

Take home messages:

1-For your results, check if in mg/dL, nmol/L or Lp(a) cholesterol. The normal levels are <30 mg/dL, <75 nmo/L and <10 mg/dL, respectively. I had a patient that had all 3 measured and neither he or his referring physician could make sense of it 16/19

2-Because the mg/dL and nmol/L measure different things, there is no “conversion factor” that is accurate. On average its about 2.5 for mg/dL to nmol/L (based on numbers above), but could range from 1.5-4 or even more, one could use a range of 1. ncbi.nlm.nih.gov/pubmed/30100157 17/19

3-There is natural variability in Lp(a), up to 25%, just like with other lipids. For example, TGs are even more variable. This variability is around a preset genetic threshold, so a patient with level of 100 can never go to 20 without Rx 18/19 ncbi.nlm.nih.gov/pubmed/29174389

4- The field is moving away from mg/dL assays and these should disappear in about 5-10 yrs, as rec by an NHLBI Working group.
5-In meantime, we just need to be aware of these issues so that we can properly interpret the data 19/19

Quiz:

Q1- 5- points
Choose the correct answer. The best method to measure Lp(a) to minimize error and bias of multiple identical kringles is:

Q2- 5- points
Choose the correct answer. One can convert mg/dL to nmol/L by:

Bonus questions: 1 point.
What is best sports team in the universe, choose the correct answer:
Hint: its one of these 🏈⚾️🏀🏒- it’s not the Dallas Cowboys