When last we left our intrepid hero, he was mighty confused.  Two follow-up lipid tests had yielded markedly different results.

Nov 2018 May 2019
LabCorp EverlyWell Quest
total 262 245 279
HDL 48 50 57
LDL 186 132* 204†
triglycerides 142 63 70

all measurements in mg/dL
* measured directly
† estimated via Martin-Hopkins method

Measure Twice

The receptionist relayed my GP's advice with the results, that he "strongly encourages you to start taking the prescribed atorvastatin."  I, as politely as possible, explained I had an independent result that differed and scheduled an appointment to discuss the situation.  I immediately send an email to EverlyWell requesting clarification, to which they promptly replied:

Our Cholesterol and Lipids panel measures direct values rather than  calculated ones. I reviewed your inquiry with our Executive Medical  Director and she advised that while the Martin-Hopkins calculation is  better than the Friederwald, there is still the chance there will be an  overestimation of LDL when the triglycerides are below 70. Directly  measuring LDL is the preferred method. Results variance is common and  expected between different labs, sample types and testing methods, in  these cases it is helpful to look at patterns. Both results indicated that your LDL levels are elevated.

My GP didn't know that direct LDL measurement with only a finger prick was even possible, and expressed skepticism at the EverlyWell results.

Could all these measurements be wrong?  Could they all be correct?  Which, if any, should I trust?  One seemed to validate my hypothesis, while the other didn't.  My doubt extended to the baseline measurement as well.

X-Ray Specs

The essential problem with all of these proxy measurements of CAD risk is that they don't actually measure damage to coronary arteries.  They can only measure the presence, prevalence, or balance of factors that may contribute to CAD if other conditions are present.

My GP had a suggestion to break the stalemate: a Coronary Artery Calcification assessment via CT scan, often simply called a "CAC score".  He said that the result of this test would trump the blood lipid panel results entirely, whether good or bad.  The only catch is that the test isn't covered by health insurance and costs $75.  I agreed, and silently vowed to go home and learn what I could about this test.  After my adventure so far, I have to admit I was skeptical that this was going to be as definitive as my GP made it sound.

The test uses a CT scan to inspect the calcium build-up within the coronary arteries, and scores the build-up.  Lower scores are better; zero indicates no calcification is apparent.

The test took less than 10 minutes.  I didn't need to remove my clothes, and nobody stuck me with a needle.  My results were waiting for me on my doctor's patient website within a couple hours.  I scored a zero; my doctor left a note that I needn't take statins.

A friend of mine, a few years older with familial heart disease risk who had been taking statins for about five years happened to have the heart CT scan done the same day.  He also scored a zero, and has since stopped taking statins.

(CAC isn't a perfect crystal ball, however.  It only detects coronary artery damage that has progressed significantly.)

Incentives

Why isn't a CAC score a prerequisite to drug therapy?  This is at least partially due to the high level of trust that general practitioners place in blood lipid panels.  These days, that particularly means the single, often estimated, LDL cholesterol number in the results.  The heart CT scan isn't covered by insurance, as yearly blood tests are, and certainly doesn't provide the same diagnostic breadth as those blood tests.

Why don't insurance companies pay for the test?  The documentary The Widowmaker claims that drug companies don't want to pay for the test because it is unlikely to reveal health problems that will materialize within the 5 year average time a health insurance company will typically cover a single person.  This is a confusing explanation to me, since statin therapy itself is almost certainly an attempt to address long term health problems, and invasive interventions (e.g., stents) are far more expensive.

Why do health insurance companies have customers for such short time-spans?  In the USA, health insurance is typically part of an employee's compensation from their employer, so any time an employee switches jobs, they very likely change health insurers as well.  This stands in contrast to life, homeowner, renter, and automotive/driver insurance in which an insurer and insured often have relationships that span decades, because the relationship is direct rather than dictated or mediated by an employer.

I can't resist editorializing: I don't know about you, but I'm much more satisfied with my life insurance, homeowners insurance, and car insurance than I've ever been with my health insurance.

Why is health insurance so commonly an employer benefit?  Today, health insurance premiums may be drawn from an employee's "pre-tax", gross income rather than post-tax income (as would be the case with all other self-funded, self-selected insurance policies I mentioned).  Looking back further, the origin of this unusual arrangement are federally mandated wage controls during World War II.

Conclusions

I'm neither a doctor nor a trained researcher, but my conclusions for my own personal health maintenance follow:

  • Blood lipid panel results need error bars; they are far from error free.
  • Blood lipid values are much more dynamic than popular wisdom would have you believe.
  • Total cholesterol is a poor proxy for all-cause mortality risk.
  • Dietary cholesterol is irrelevant to me.
  • I prefer saturated fats to vegetable oils.
  • I avoid added sugar and added sweeteners, by any name (sugar, fructose, fruit juice concentrate, brown sugar, honey, malt syrup, maple syrup, molasses, corn syrup, etc.).
  • I consider LDL cholesterol is a poor proxy for my all-cause mortality risk, because I don't have other significant risk factors.
  • I consider small-dense LDL particle count is a better proxy for CAD risk; I settle for triglyceride/HDL ratio when particle size measurement isn't available.
  • A CAC score provides useful, complementary information.

These are all conclusions that I've drawn for myself.  This next bit is some unsolicited advice:

You must be an active participant in your own health care. Nobody is more interested in or motivated by your good health than you are.  Pay your doctor for diagnostic skill, advice, and education.  If your doctor gives orders, change doctors.