240 ‒ The confusion around HDL and its link to cardiovascular disease | Dan Rader, M.D.

KEYPOINTS

• 01

  HDL biology is far more complex than LDL biology, which is why it is so poorly understood

  Dan Rader, a lipidologist at the University of Pennsylvania, explains that apoB stays with its particle for life. ApoB100 is made in the liver and apoB48 in the gut. HDL's main protein, apoA1, behaves differently. It is secreted lipid-poor and picks up cholesterol and phospholipids from cells through the transporter ABCA1, and its other proteins swap on and off between particles. People who lack ABCA1 make plenty of apoA1 but have almost undetectable HDL, because the apoA1 can't acquire lipid and is cleared quickly.

• 02

  Humans are unusual in carrying most cholesterol on apoB particles

  In a normal mouse, more than 90% of blood cholesterol is in HDL. Rader and Peter note that if humans had primate-like or mouse-like apoB levels, around 20 mg/dL instead of about 100, that five-fold difference would roughly separate developing atherosclerosis in a lifetime from never developing it. This is one reason mouse HDL research translates poorly to people.

• 03

  HDL subfraction labels are mostly noise for clinicians

  HDL2 and HDL3 refer to particle size and number of apoA1 molecules (about four in HDL2 and three in HDL3). They have nothing to do with apoA2 or apoA4. Rader considers fractionation interesting biologically but relatively unimportant clinically. Peter says his practice stopped paying attention to HDL fractionation around 2016, and Rader agrees that was reasonable.

• 04

  High HDL cholesterol is not a free pass

  The inverse link between HDL cholesterol and heart disease goes back to early Framingham data from the late 1970s and early 1980s. But a high HDL-C does not offset a high LDL-C or apoB. Peter calls the idea that a high HDL means you can ignore heart disease risk 'bunk' and cites two Mendelian randomization studies that fail to support a causal protective role. Rader says he would not change the treatment he'd otherwise give just because HDL is high. Both say HDL-C above about 90 deserves a closer look rather than reassurance.

• 05

  Raising HDL-C with CETP inhibitors has repeatedly failed

  CETP moves cholesterol ester from HDL to apoB particles, and people deficient in it can have HDL-C above 100 mg/dL. Pfizer's first inhibitor, torcetrapib, was tested on top of Lipitor, and in September 2006 the combination did slightly worse, which Peter calls a bombshell. After three failures, Rader's reading is that raising HDL-C with CETP inhibition neither helps nor hurts. A CETP inhibitor might still be useful for lowering LDL and apoB, and that is being tested.

• 06

  Niacin raised HDL and lowered apoB, but outcome trials found little benefit

  Niacin modestly lowers triglycerides, apoB, LDL and Lp(a) (roughly 15–20%) and could raise HDL-C from about 50 to 90 mg/dL at strong doses. Rader once prescribed it widely and says he has had to 'eat crow'. AIM-HIGH showed little benefit, and a large, well-powered Merck trial was also disappointing; together they 'killed niacin'. He has had to tell patients it probably wasn't helping them.

• 07

  HDL-C tracks triglyceride metabolism and insulin resistance

  Rader calls HDL-C an 'integrator' of triglyceride handling, a little like HbA1c for glucose. After a fatty meal, triglycerides might rise from 100 to 400 mg/dL before clearing. Each excursion lowers HDL-C slightly (for example, 40 to 37–38), and repeated meals push it to a lower steady state. That is why a high triglyceride-to-HDL ratio (both in mg/dL) is reasonably associated with insulin resistance, though not in African-American patients, where it has long been found unhelpful.

• 08

  What matters may be HDL function, especially cholesterol efflux

  Macrophages in the artery wall use ABCA1 to unload cholesterol onto apoA1. The cholesterol eventually returns to the liver through SR-B1, a process called reverse cholesterol transport. Rader's garbage-truck analogy: HDL collects cholesterol where you don't want it and dumps it at the liver. Two people with the same HDL-C of 50 can have very different flow through this pathway. A static concentration tells you nothing about velocity.

• 09

  Efflux capacity predicts events, and a trial is testing it

  Rader's lab measures efflux capacity in macrophage-like cells (J774 cells and human macrophages). Using many thousands of baseline samples from EPIC-Norfolk with 10+ years of follow-up, they found efflux capacity predicted incident cardiovascular events. That is still association, not proof of causation. CSL112, an infusion of apoA1 complexed with lipid, was being tested in patients after acute coronary syndrome: four weekly infusions, with events measured at 90 days. If positive, it would be the first evidence that promoting efflux reduces events.