LDL:
To C or not to C
Edition: 2025-06-02
Just what does the “C” in LDL-C
stand for? I’m getting less certain every day.
Some sources report that it stands for
“Cholesterol”. Other sources
report that it stands for
“Calculated”. Even if that’s nailed
that down, it’s still a long way from the
whole character string being an
abbreviation of anything remotely useful.
LDL
These initials stand for Low Density
Lipoproteins. When this abbreviation is
used with no further qualification, it’s
completely useless. When I see it, I
assume it’s been spoken or written by
someone seriously careless or under-informed.
⟲Here’s lipidologist
Thomas Dayspring being pretty blunt about it:
| |
First of all,
the misinformation on labeling lipid
metrics is one of the things that it’s
a miracle hasn’t given me a stroke yet.
I do a lot of peer review. I’m one
of the associate editors of the
Journal of Clinical
Lipidology, and I will reject
a paper instantly that uses improper
lipid metrics. Don’t tell me the LDL
is this, because LDL is a low density
lipoprotein. It’s not a laboratory metric.
You want to tell me what the
LDL-cholesterol is, the LDL particle
number is, the lipidomics of an LDL, is
the LDL oxidized or not? Great. We do
have assays that will measure all of
those. Don’t identify yourself as an
ignoramus. And look, I’ve told this to
many of the top lipidologists in the
country, well actually, “Stop
telling people,
what’s your LDL.” |
|
⟲ As LDL-Cholesterol
On the standard lipid panel, is the LDL
estimated or measured, and in either case,
how? If it is known that “-C”
stands for “cholesterol”, the number
is still a complete mystery.
On the standard panel, of course, the TG and HDL
numbers are quite useful. The TC not so much.
VLDL-C, if present, may be a whole
different form of SoC
handwaving: (TG÷5)*.
The “LDL” number is why this
page is here.
⟲ As LDL-Calculated
If it is known the author/speaker means
“calculated”, matter are likewise
not much further along. As LDL-Calculated
or calculated LDL-C, being a guess is
at least known, based on a standard lipid
panel (usually CPT Code 80061), but which
guess? The top forms of calculated LDL
estimate are:
LDL-C (NIH), herein LDL-CNIH
LDL-C (Friedewald),
herein LDL-CFriedewald
LDL-C (Martin-Hopkins),
herein LDL-CM-H
LDL-C (other: Iranian, Hatta or Puavilai)
LDL-C is often not identified as
such, and needs to be. What these
calculations all have in common is that
they are trying to torture 3 actual
measurements (TC, TG, HDL) into confessing
something that they don’t really know,
namely: do your LDL particles include any
that are atherogenic.
⟲ LDL-CNIH
is the new idol.
Any lipid panel you get in 2021 or later
may declare “LDL Chol Calc (NIH)”
. As of 2021-02-01, per Stanford Lab, NIH has finally trashed
Friedewald, but since NIH also stands for the
well-known institutional syndrome of
“Not Invented Here", apparently
NIH figured they could also improve on
Martin-Hopkins. So they’ve come up with
yet another way to get the actually-measured
lipid markers (HDL,TC,TG) to disclose
something they cannot.
I’ve replaced dashes with underscores
here, to avoid confounding with minus signs.
The new contortion (for mg/dL, presumably,
and not ISO UoM) is:
LDL_CNIH = TC÷0.948 - HDL_C÷0.971 - (TG÷8.56 + [TG×NonHDL_C]÷2140 - TG²÷16100) - 9.44
Lest you assume that there is some new
measurement here, probably not:
NonHDL_C = TC - HDL_C
per Dayspring
And the ambiguous “C” in HDL_C apparently
means “cholesterol” and not
yet-another “calculated”.
If so, the final new impressive fiction is then:
LDL_CNIH=TC÷0.948-HDL_C÷0.971-(TG÷8.56+[TG×(TC-HDL_C)]÷2140-TG²÷16100)-9.44
And on the actually useful measure TG,
the Stanford summary concludes with the destructive:
The NIH equation performs equally
well in both fasting and non-fasting states.
Translation: equally useless, as the
important TG number can be significantly
distorted in fed state.
NIH is rapidly replacing Friedewald, and
appears to have completely deprecated
all the other pretenders to the throne of:
most impressive
fake LDL number to
scare patients with.
Note, however, that if LDL-C is
completely missing from a lab report, it
may be due to LDL-CNIH not
being reported if TG is over
800 mg/dL (yikes!).
⟲ LDL-CFriedewald
is now out to pasture, but any
“LDL-C” values you have from
2020 or earlier are apt to be based on
its much cruder fiction:
LDL = TC - HDL - (TG ÷ 5)
(in mg/dL)
Dr. Davis has written about this folly
many times. Here’s one: Leprechauns, nymphs, high cholesterol,
and other fanciful notions
⟲ LDL-CM-H
is now fading.
Martin-Hopkins was
gradually replacing Friedewald until NIH
arrived. You may still encounter it in
recent reports.
In the Friedewald approximation, VLDL is
presumed to be TG÷5. M-H used the
same approach, but replaced the 5 with a variable factor from a table.
⟲ Converting between NIH,
Friedewald & M-H is possible
just by re-running the calculations from
the actual measurements. But doing so is
a complete waste of time.
If low density lipoproteins matter, and
they do, they need to be actually measured,
which leads us to the next mess sometimes
encountered with standard lipid panels.
⟲
As Measured “Direct”
LDL[-C]
This (CPT Code 83721) is also known as
Direct LDL-C, Direct LDL, DLDL or
LDL-D. It actually measures LDL, but still
lumps all the subfractions together.
When TG is above 400 mg/dL (also yikes),
the lab doing the lipid panel may silently
perform a DLDL and report that.
⟲
LDL-P: What Is Really Sought
Low Density Lipoproteins do matter, and
deserve to be actually measured, not guessed
at. Lipoprotein subfractions can be measured,
and assays for this have been available for
‘only’ about a quarter century
now. Why they aren’t used more often, if
not routinely, says a lot about the supposed
specialty of cardiology.
There are three major assay methods (all
still based on a simple blood draw):
NMR (Nuclear Magnetic
Resonance), CPT Code 83704
IMS-MS (Ion Mobility, such
as CardioIQ®), also CPT Code 83704*
VAP (Vertical Auto
Profile), CPT Code 83701
Electrophoresis:
CPT Code 83700
BHD’s sdLDL-C test,
assay method unclear
* So be sure to check the test
description text.
NMR is the preferred test on this site.
In addition to being the assay that
Dr. Davis recommends, it’s the one
most familiar to the readership. The
key value is:
NMR Small LDL-P (target: < 200 mmol/L)
and this is a number that
cannot be teased
out of a standard lipid panel (although
when you get an NMR, you’ll often get
another lipid panel anyway).
IMS-MS, VAP and electro can be useful, if
you have a skilled practitioner who is
familiar with them. VAP was off the market
for a while, and is now back, but the
program hasn’t yet re-evaluated it for
suitability.
The total LDL-P Particle Number is no longer a key focus.
Dr. Davis: “…high or highish LDL particle number
in the absence of small LDL particles is a
gray zone: We do not know how atherogenic
(plaque-causing) this situation is. My
suspicion is that it is not plaque-causing
or is minimally so, as we have had plenty
of people with high LDL particle number,
all large, with zero heart scan
scores.”
As a curio, LDL-P {total} Particle Number is
is an actually measured LDL, ⅒ of that number is a much more
accurate statement of ‘LDL” than
any calculated LDL-C.
⟲
Confounders and Closing Remark
Since it’s almost always the case that at
least something has been
measured, it’s further necessary to know
if the test was done fasting, and not in
the context of any weight loss in the
preceding 30 days, both of which
distort lipoproteins generally, and TG
in particular. If these criteria aren’t
met, even an actual LDL-P measurement
is not useful.
Needless to say, if you’ve ever heard…
“I’m
prescribing a statin for you
because your ‘LDL’ is
too high”
…you might need to be looking
for a real doctor.
___________
Bob Niland
[⎆disclosures] [⎆topics]
[⎆abbreviations]
* Calculated VLDL-C allows
cookie-cutter MDs to concede that TG
matters, without actually admitting
that TG matters.
Tags: LDL-C,LDL-P,NIH