The shifting sands of time and medical
research constantly reshape the
topography of our knowledge landscape.
For decades we have been trained in
the descriptions of “good” and “bad”
cholesterol. Observational data from the
Framingham Heart Study and others
convinced us that high-density lipoprotein
cholesterol (HDL-c) is “good cholesterol”
and protective to the cardiovascular
system. We have hypothesized that higher
the HDL-c level, the more protection is
conferred and that targeting and raising
the HDL-c therapeutically will lead to
primary prevention of cardiovascular
disease. It appears that it is time for us to
rethink this hypothesis.
The results of several recent, largescale investigations have called the HDL-c
hypothesis into serious doubt. Scrutiny
of the latest evidence, and a fresh look at
the older evidence, has prompted several
authorities to herald that “the HDL
hypothesis is on the ropes,” and that “this
is the death knell of niacin, for sure…”1,2
While these statements are broad and
over-reaching, our present understanding
of the HDL-c hypothesis certainly requires
some revising.
The inverse association between HDL-c
and coronary heart disease (CHD) has
been observed repeatedly throughout
large-scale studies. Castelli et al, 1988,
published a now infamous characterization
of HDL-c and risk from the Framingham
Heart Study (FHS), which demonstrated
that for men 50–70 years old, at any level
of low-density lipoprotein cholesterol
(LDL-c), HDL-c concentrations less than
25 mg/dL were associated with the highest
incidence of CHD. Conversely, at any level
of LDL-c, HDL-c concentrations greater
than 65 mg/dL were associated with the
lowest incidence of CHD.3
HDL-c is a strong predictor of CHD risk
in both men and women. The Coronary
Primary Prevention Trial (CPPT), Multiple
Risk Factor Intervention Trial (MRFIT),
Lipid Research Clinics Follow-up Study
(LRCF), Israeli Ischemic Heart Disease
Study (IIHDS), FHS, and others have
demonstrated that HDL-c concentrations
less than 40 mg/dL are associated with
approximately double the incidence of
CHD compared to HDL-c concentrations
over 50 mg/dL.4
We have learned from
these studies that low concentrations of
HDL-c are our best predictor of CHD.5
Inconsistent Studies
The problem with the HDL hypothesis
is that the medical literature contains
important inconsistencies with regard to
it. Extremely low HDL-c is not consistently
correlated with early CHD, and extremely
high levels are not consistently protective
against CHD. More importantly,
interventional trials repeatedly fail to show
a cardioprotective benefit from raising
HDL-c levels in humans. There is a big
difference between observing that low
HDL-c is correlated with CHD risk and the
converse, ie, demonstrating that raising the
HDL-c decreases the incidence of CHD.
The announcement of “The death of the
HDL hypothesis” comes on the heels of the
failures of 4 large-scale interventional trials
over recent years, all of which were focused
on increased HDL-c as a means to lower
cardiovascular disease (CVD) incidence.
Cholesteryl ester transfer protein
(CETP) inhibition has been a target of
recent pharmaceutical innovation for
novel lipid modification therapy. CETP
exchanges triglycerides from VLDL
and LDL molecules for cholesteryl
esters from HDL molecules; therefore,
CETP inhibition should raise HDL-c
levels and improve CVD morbidity and
mortality. The ILLUMINATE trial was
an investigation of the CETP inhibitor,
torcetrapib, in 15 067 patients at high
CVD risk.6
While there was a statistically
significant increase in HDL-c of 72% over
baseline levels and decreases in LDL-c
levels in participants taking torcetrapib,
the trial had to be stopped prematurely
due to a statistically significant increased
risk of cardiovascular events and all-cause
mortality in the group taking the drug.
Dalcetrapib, another CETP inhibitor,
underwent recent clinical investigation
in the dal-OUTCOMES trial7
; 15 871
patients who had a recent acute coronary
syndrome were randomized to take either
dalcetrapib or placebo, in addition to usual
care. Over 31 months of follow-up, the
HDL-c levels of patients taking dalcetrapib
increased 31-40% (compared to 4-11% in
the placebo group); however, there were
no significant differences in nonfatal
myocardial infarctions, ischemic strokes,
unstable angina, cardiac arrests, or deaths
from coronary heart disease between the
two groups.
The AIM HIGH trial was designed to
investigate whether using extended-release
niacin in addition to simvastatin to raise
low HDL-c was superior to simvastatin,
alone, in reducing CVD deaths, events,
hospitalizations, or revascularizations.8
A
total of 3414 participants were randomized
into 2 groups; all received simvastatin
40-80 mg/day (with or without ezetimibe,
10 mg/day), to ensure that their LDL-c
levels remained between 40-80 mg/dL.
Extended-release niacin, 1500-2000 mg/day,
was given to 1718 participants, while 1696
received placebo. Niacin significantly
increased HDL-c and lowered LDL-c and
triglycerides compared to placebo over
the 3-year follow-up. However, there was
no difference in outcomes between the 2
groups and the trial was stopped early due
to lack of efficacy.
The Heart Protection Study TwoTHRIVE (HPS2-THRIVE) was a very large,
multinational study which investigated
extended-release niacin (ERN) in primary
prevention of major cardiovascular
events.9
Participants were randomized
to take either 2000 mg/day of ERN plus
laropiprant, 40 mg, or placebo; all 25 673
patients also took simvastatin with or
without ezetimibe. After nearly 4 years
of follow-up, there were no significant
differences between the 2 groups
against the composite primary outcome:
prevention of first myocardial infarction,
stroke or revascularization. Serious adverse
events, including diabetic complications,
new-onset diabetes, infections, myopathies,
hemorrhages, and others were significantly
higher in the ERN group.
Each of the 4 research trials discussed
above was fraught with problems and
potential confounders. The problems with
these trials leave us to ask ourselves and
the pall bearers of the HDL hypothesis
whether that hypothesis was tested in these
trials in the first place. The ILLUMINATE
and dal-OUTCOMES trials were
investigations of novel lipid modification
agents. The expected efficacy of those
drugs was based on the HDL hypothesis,
but the trials were not designed to test
the HDL hypothesis. Their failures and
adverse events were likely to be the result
of their novel mechanism-of-action and
off-target effects, not some failure of the
HDL hypothesis.
In the AIM HIGH study, participants
started the study with baseline LDL-c
levels of 40-80 mg/dL; in HPS2-THRIVE,
mean baseline lipid levels were: total
cholesterol 128 mg/dL, LDL-c 63 mg/dL,
HDL-c 44 mg/dL. Therefore, we are left
to wonder whether we should expect to
see any differences in outcomes between 2
groups which are already at target LDL-c
levels, regardless of the type of add-on
medication. Finally, in HPS2-THRIVE,
laropiprant was added to extendedrelease niacin as an anti-flushing agent.
The problem is that laropiprant is a DP1
receptor antagonist, meaning that it blocks
the action of prostaglandin D2
(PGD2
) and
prevents PGD2
-dependent vasodilation.
Therefore, its use in the study is a plausible
confounder, especially as it pertains to
adverse events in the study.
Doubts About HDL-C as a
Therapeutic Target
The National Lipid Association (NLA) has
produced a consensus statement on HDL.
The statement was discussed at the annual
scientific sessions in May, 2013, and will
be published in late 2013. Among others,
these 3 will be prominent: 1) we can no
longer consider HDL-c a therapeutic target;
2) HDL cholesterol cannot be considered a
biomarker of HDL particles (HDL-p); 3) we
must redouble our efforts to understand
the HDL particle.10
We can no longer consider HDL-c to
be a target of therapy. Simply put, there
is no evidence behind it. While there are
important shortcomings and problems with
the trials that have been conducted, the
evidence we have is the evidence we have,
and little of it demonstrates that raising
HDL-c levels changes cardiovascular risks
or outcomes.
Beyond Cholesterol – The
HDL Particle
HDL cholesterol is not the same as HDL
particles. Cholesterol transport is just
one small part of the action of the HDL
particle in the body. Too often we refer to
HDL and LDL and others as “cholesterol,”
which they are not; they are transport
molecules which carry several types of
lipids, proteins and other components,
depending on their species. It is likely
that we have missed most of what is
important about HDL-p because we have
taken our eye off the ball and focused on
HDL-c rather than on understanding the
HDL particle.
The HDL particle is a complex transport
molecule in the human body, which we
have scarcely begun to understand. Peter
Toth, MD, the president of NLA, presented
the HDL consensus statement at the
May 2013 sessions and said, “…The fact
that there are plausible reasons for why
these trials failed suggests that the HDL
hypothesis has still not been tested. For this
reason, it is far too premature to abandon
the HDL hypothesis. On the contrary,
we need much more research in order to
understand the reason for the unexpected
results of these failed trials. It is premature
to abandon the research efforts to better
elucidate how the modulation of HDL
metabolism and functionality impacts risk
for CHD.”5
We are at the very beginning of our
understanding of the HDL particle. We
have incorrectly oversimplified it as
“good cholesterol” and the vehicle of
reverse cholesterol transport. The HDL
molecule is a complex transport molecule
in the human body. While it does
transport lipid species, it also transports
cytokines, messenger RNA, microRNA,
and others. While it does participate, at
times, in reverse cholesterol transport,
it also plays important roles in
inflammation, gene transcription, and
cell-to-cell communication that we do
not fully understand. Put simply, the
HDL hypothesis is not dead; it may not
yet have even been fully formed. We
stand not at the foot of the grave of the
HDL hypothesis, but on the precipice
of a journey through a new landscape
of HDL science and exploration, one on
which we should now embark with a
fresh perspective and great excitement