8
Before 1977, high circulating levels of cholesterol and the lipoproteins that carry it were generally considered to be bad actors in the causal chain of atherosclerosis and its adverse cardiovascular consequences, even if this hadn’t yet been proven in clinical trials. Thus, when the Framingham Heart Study published a report linking high levels of the previously overlooked small dense high-density lipoprotein (HDL) particles to reduced risk of heart attack, it created quite a stir.1 Throughout the late 1970s, 1980s, and early 1990s, studies, perspectives, and editorials touted HDL cholesterol as a major cardiovascular risk factor, on a par with LDL cholesterol (but in the opposite direction), and dubbed it the “good cholesterol,” although the cholesterol carried in HDL particles is of course the exact same molecule that is carried in LDL particles. HDL was thought to work by facilitating reverse cholesterol transport from the arterial wall to the liver, where it could be repurposed or cleared from the body, thereby undoing the mischief wrought by LDL.2 The ratio of LDL to HDL cholesterol was promoted as a superior all-purpose risk factor and is included in lab panels even today; a ratio below 3.5 is good, and a ratio above 5.0 means trouble. I personally contributed to the HDL hype in a 1989 analysis of four published observational epidemiologic studies showing that levels of HDL and LDL were associated in an opposite and proportionate way to cardiovascular risk.3 Thus arose the “HDL hypothesis” that raising HDL cholesterol levels would reduce the rate heart attack and other cardiovascular events.
However, as I also noted at the time, there were several major gaps in this tidy story.4
1. HDL is correlated with many other cardiovascular risk factors. Male sex (typically 40–50 mg/dL for men and 50–60 mg/dL for women), obesity, diabetes and pre-diabetes, and high triglyceride levels are associated with reduced levels of HDL cholesterol; exercise and alcohol intake are associated with increased levels of HDL cholesterol. Although epidemiologic studies point to the independence of HDL cholesterol as a predictive factor, one could not rule out that HDL is merely a marker for poorly quantifiable healthy behaviors like physical activity and diet.
2. We lack a good genetic model (comparable to familial hypercholesterolemia for LDL cholesterol) for HDL “deficiency” in animals or humans.5 For example, patients with Tangier disease (named for its discovery in Tangier Island in the Chesapeake Bay) have very low levels of HDL cholesterol, but their risk of heart attack is not elevated. Families with very high HDL cholesterol levels (> 100 mg/dL) and great longevity have been identified, but other families with high HDL cholesterol do not have increased longevity.
3. No practical pharmacological treatments to raise HDL cholesterol existed, with the possible exception of estrogen in post-menopausal women (we will see how that worked out in Chapter 9). The efficacy of weight loss and exercise to raise HDL are limited by compliance, and the pitfalls of prescribing alcohol are obvious.
4. Clinical trial evidence for the benefit of raising HDL cholesterol was completely lacking.
Fibrate and Niacin Trials
The lack of specific HDL-raising drugs did not stop investigators from trying to mount randomized clinical trials to establish the positive effects of raising HDL cholesterol in conjunction with lowering triglyceride levels, using two old classes of cholesterol-lowering drugs, the fibrates and niacin, in settings that minimized their impact on LDL cholesterol. Five such trials were conducted between 1980 and 2011:
• The Helsinki Heart Study (HHS) randomized 4081 asymptomatic 40- to 55-year-old men with non–HDL cholesterol > 220 mg/dL to receive either gemfibrozil or placebo.6 Gemfibrozil raised mean HDL cholesterol by 11%, lowered mean LDL cholesterol by 11%, and lowered mean triglycerides by 35%. Incidence of the primary composite cardiovascular outcome was reduced by 34% (P = 0.02), but mortality was not reduced.
• The Veteran Affairs HDL Intervention Trial (VA-HIT) randomized 2531 men with pre-existing atherosclerotic heart disease, HDL cholesterol < 40 mg/dL, LDL < 140 mg/dL, and triglyceride levels < 300 mg/dL to receive either gemfibrozil or a placebo.7 Gemfibrozil raised mean HDL cholesterol by 6% and lowered triglyceride levels by 31%, with no change in LDL cholesterol. The incidence of the primary outcome (fatal and nonfatal heart attacks) was reduced by a statistically significant 22%.
• The Fenofibrate Intervention and Event Lowering Trial in Diabetes (FIELD) trial randomized 9795 50- to 75-year-old patients with type 2 diabetes. and triglyceride levels < 443 mg/dL in 63 centers in Australia, Finland and New Zealand to receive fenofibrate or a placebo.8 Fenofibrate raised mean HDL cholesterol by 5%, lowered LDL cholesterol by 12%, and lowered triglycerides by 29%. The primary outcome (fatal and nonfatal heart attacks) was reduced by 11% in the fenofibrate group, but this reduction was not statistically significant. However, mortality was slightly (but not significantly) higher in the fenofibrate group.
• The Action to Control Cardiovascular Risk in Diabetes (ACCORD) trial, a 2×2 partial factorial NHLBI trial, randomized 5518 persons with type 2 diabetes, HDL cholesterol levels below 55 mg/dL for women and Blacks and below 50 mg/dL for everyone else, LDL cholesterol levels between 60 and 180 mg/dL, and serum triglyceride levels below750 mg/dL to receive fenofibrate or a placebo.9 Fenofibrate increased HDL cholesterol by about 1 mg/dL, decreased triglycerides by 25–30 mg/dL, and did not change mean LDL cholesterol levels. There was no significant difference in the primary outcome (major fatal or nonfatal cardiovascular events) or in any secondary outcome.
• The Atherothrombosis Intervention in Metabolic Syndrome with Low HDL/High Triglyceride and Impact on Global Health Outcomes (AIM-HIGH) trial10: This NHLBI trial randomized 3414 patients with metabolic syndrome and known cardiovascular disease, who were already taking high-dose simvastatin plus ezetimibe (as needed) to maintain LDL cholesterol levels between 80 and 100 mg/dL, to receive either slow-release niacin or a placebo. Metabolic syndrome was defined as HDL cholesterol below 40 mg/dl for men or below 50 mg/dl for women, triglyceride levels between 150 and 400 mg/dL, and LDL cholesterol below 180 mg/dL at entry. After two years, niacin produced a 4 mg/dL difference in mean HDL cholesterol, a -29 mg/dL difference in triglyceride, and a -4.5 difference in LDL cholesterol between the niacin and placebo group. However, there was no significant difference in the primary outcome (a composite of fatal and nonfatal major cardiovascular events) or any secondary cardiovascular outcome. There was a nonsignificant 2% increase in mortality in the niacin group.
The overall results of these studies were mixed at best. The two older gemfibrozil studies reported positive results, but the three later fenofibrate and niacin studies were all negative. None of the studies showed a beneficial effect on mortality. Only the HHS achieved more than a 6% increase in HDL cholesterol, but even in this best case, the reported improvement in cardiovascular outcomes could as easily have been attributed to the similar reduction in LDL cholesterol and/or the far larger reduction in triglyceride levels. Clearly, a drug that specifically produced substantially larger increases in HDL cholesterol was needed to truly test the HDL hypothesis.
Cholesteryl Ester Transfer Protein (CETP) Inhibitors
A new class of drugs, the CETP inhibitors, suddenly changed the landscape for the HDL hypothesis at the dawn of the new millennium. CETP is a protein that facilitates the movement of cholesteryl esters (cholesterol bound chemically to fatty acids) from HDL to LDL particles.10 Inhibiting this protein raises HDL cholesterol and lowers LDL cholesterol, thereby (hypothetically) lowering cardiovascular risk. However, the evidence to support this hypothesis from naturally occurring mutations in CETP in animals and humans is mixed. Some families with defective CETP actually have increased cardiovascular risk despite high HDL cholesterol levels. In any case, by 2005, four such drugs were undergoing clinical evaluation—torcetrapib, dalcetrapib, evacetrapib, and anacetrapib—which could raise HDL cholesterol as much as twofold while lowering LDL cholesterol by as much as a statin. What was not to like?! Pfizer, the developer of torcetrapib, the first of these drugs, saw it as a blockbuster and eagerly awaited the results of its 15,067-patient ILLUMINATE trial, which they would use to market torcetrapib in combination with atorvastatin. Hoffman-LaRoche (dalcetrapib), Eli Lilly (evactrapib), and Merck (anacetrapib) were equally bullish about the three flagship trials of their drugs—dal-OUTCOMES (15871 patients), ACCELERATE (12,092 patients), and REVEAL (30,449 patients), respectively.11
The results were an unmitigated disaster—at least from the perspective of their sponsors, who had invested hundreds of millions of dollars in the development of these drugs. In the ILLUMINATE trial, published in 2007, the torcetrapib group experienced a 25% increase in cardiovascular events and a 58% increase in mortality—both statistically significant—despite a 74% increase in mean HDL cholesterol and a 25% decrease in LDL cholesterol.12 Although the study investigators raised the possibility that the adverse result may have been due to “off-target” effects of torcetrapib on potassium and bicarbonate levels that were unrelated to CETP inhibition, all hopes that torcetrapib would be a breakthrough drug were dashed and hopes for CETP inhibition in general were cast into doubt. Five years later, the next domino to fall was dalcetrapib, a less potent CETP inhibitor without the off-target metabolic effects of torcetrapib. In the dal-OUTCOMES trial, adverse cardiovascular outcome rates did not differ between the dalcetrapib and placebo groups, despite a nearly 30% higher mean HDL cholesterol in the former group and no difference in LDL cholesterol.13 Five years later, evacetrapib and anacetrapib completed the gloomy picture. In the ACCELERATE trial, there was no difference in cardiovascular event rates between the evacetrapib and placebo group, despite a whopping 133% increase in mean HDL cholesterol and 31% decrease in LDL cholesterol in the group receiving the active drug.14 Finally, in the REVEAL trial, anacetrapib did modestly but significantly reduce the primary cardiovascular outcome (which included revascularizations as well as clinical events) by 9%, but this reduction was probably attributable more to the 17 mg/dL (18%) reduction in non–HDL cholesterol than the 43 mg/dL (104%) increase in HDL cholesterol,15 Although the results suggested that anacetrapib was safe and also had a favorable effect on diabetes, its further development was discontinued.
Like the CAST trial described in Chapter 3, these trials represent a cautionary tale against being overly facile about leaping from finding an association in observational studies into assuming the efficacy of clinical interventions based on that association. But let us stop short of concluding that the concept of low HDL cholesterol as a target for intervention is a complete washout. The CETP inhibitors have taught us only the unhelpfulness of this particular mechanism of raising HDL cholesterol—i.e., by interfering with the normal pathway by which HDL particles rid themselves of cholesteryl esters, thereby perhaps creating engorged HDL particles that cannot function normally. The fact that some natural CETP mutations are not associated with decreased cardiovascular risk ought in retrospect to have alerted us to this possibility. It is still possible that raising HDL cholesterol by other mechanisms, such as stimulating the production of the A1 and A2 apoproteins that comprise the structural base of HDL, might prove beneficial. Before completely dismissing the HDL hypothesis, one should consider what dire fate might have befallen the LDL cholesterol hypothesis if the CETP inhibitors (and not the statins) were first introduced as LDL cholesterol-lowering drugs in the 1980s, rather than as HDL cholesterol-raising drugs in the 2000s. But be that as it may, the statin trials have firmly established the benefit of LDL cholesterol-lowering, while the trail to discover a drug that lowers cardiovascular risk by raising HDL cholesterol level has for now gone cold.
Impact of HDL on Decline in Cardiovascular Mortality
Clearly, HDL has played no role whatsoever in the observed decline in mortality from heart attacks and strokes since 1960. We have no safe, practical, and effective treatment for raising HDL cholesterol on a population scale, and it is far from certain that this hypothetical treatment would actually be beneficial.
1. T Gordon, WP Castelli, MC Hjortland, WB Kannel, TR Dawber. High-density lipoprotein as a protective factor against coronary heart disease. The Framingham Study. Am J Med 1977; 62: 707–714.
2. M Krieger. The ‘best’ of cholesterols, the ‘worst’ of cholesterols: a tale of two receptors. Proc Nat Acad Sci, USA 1998; 95:4077–4080.
3. DJ Gordon, JL Probstfield, RJ Garrison, JD Neaton, WP Castelli, JD Knoke, DR Jacobs, S Bangdiwala, HA Tyroler. High-density lipoprotein and cardiovascular disease: Four American Studies. Circulation 1989; 79:8–15.
4. DJ Gordon, BM Rifkind. HDL—Clinical implications of recent studies. N Engl J Med 1989; 321:1311–1316.
DJ Gordon. Role of circulating HDL and triglycerides in coronary artery disease. End Metab Clin N Amer 1990; 19:299–310.
DJ Gordon. HDL and cardiovascular disease. Cardiology Board Review 1990; 7:29–40.
DJ Gordon. HDL and CHD—An epidemiologic perspective. J Drug Devel 1990, 3(suppl):11–17.
5. CE Kosmas, D Silverio, A Sourlas, F Garcia, PD Montan, E Guzman. Primary genetic disorders affecting high density lipoprotein (HDL). Drugs in Context 2018; 7: 212546. DOI: 10.7573/dic.212546.
6. MH Frick, O Elo, K Haapa, OP Heinonen, P Heinsalmi, P Helo, JK Huttunen, P Kaitaniemi, P Koskinen, V Manninen, H Maenpaa, M Malkonen, et al. Helsinki Heart Study: primary-prevention trial with gemfibrozil in middle-aged men with dyslipidemia. Safety of treatment, changes in risk factors, and incidence of coronary heart disease. New England Journal of Medicine 1987; 317: 1237–1245.
7. HB Rubins, SJ Robins, D Collins, CL Fye, JW Anderson, MB Elam, FH Faas, E Linares, EJ Schaefer, G Schectman, TJ Wilt, J Wittes. Gemfibrozil for the secondary prevention of coronary heart disease in men with low levels of high-density lipoprotein cholesterol. Veterans Affairs High-Density Lipoprotein Cholesterol Intervention Trial Study Group. N Engl J Med 1999; 341:410–418.
8. The ACCORD Study Group. The effect of combination lipid therapy in type 2 diabetes mellitus. N Engl J Med 2010; 362:1563–1574. DOI: 10.1056/NEJMoa1001282.
9. The AIM-HIGH Investigators. Niacin in patients with low HDL cholesterol levels receiving intensive statin therapy. N Engl J Med 2011; 365:2235–2267.
10. PJ Barter, HB Brewer, J Chapman, CH Hennekens, DJ Rader, AR Tall. Cholesteryl ester transfer protein. A novel target for raising HDL and inhibiting atherosclerosis. Arteriosclerosis, Thrombosis, and Vascular Biology 2003; 23:160–167.
11. AR Tall, DJ Rader. Trials and Tribulations of CETP inhibitors. Circulation Research 2018; 122:106–112. https://doi.org/10.1161/CIRCRESAHA.117.311978.
12. PJ Barter, M Caulfield, M Eriksson, SM Grundy, JJP Kastelein, M Kornajda, J Lopez-Sendon, L Mosca, JC Tardif, DD Waters, CL Shear, JH Revkin, KA Buhr, MR Fisher, AR Tall, HB Brewer, for the ILLUMINATE Investigators. Effects of torcetrapib in patients at high risk for coronary events. N Engl J Med 2007; 357:2109–2122. doi: 10.1056/NEJMoa0706628.
13. GG Schwartz, AG Olsson M Abt M, CM Ballantyne, PJ Barter, J Brumm, BR Chaitman, IM Holme, D Kallend, LA Leiter, E Leitersdorf, JJV McMurray, H Mundl, SJ Nicholls, PK Shah, JC Tardif, S Wright, dal-OUTCOMES Investigators. Effects of dalcetrapib in patients with a recent acute coronary syndrome. N Engl J Med 2012; 367:2089–2099. doi: 10.1056/NEJMoa1206797.
14. AM Lincoff, SJ Nicholls, JS Riesmeyer, PJ Barter, HB Brewer, KAA Fox, CM Gibson, C Granger, V Menon, G Montalescot, D Rader, AR Tall, E McErlean, K Wolski, G Ruatolo, B Vangerow, G Weerakkody, SG Goodman, D Conde, DK McGuire, JC Nicolau, JL Leiva-Pons, Y Pesant, W Li, D Kandath, S Kouz, N Takirkheli, D Mason, SE Nissen, for the ACCELERATE Investigators. Evacetrapib and cardiovascular outcomes in high-risk vascular disease. N Engl J Med 2017; 376:1933–1942. doi: 10.1056/NEJMoa1609581.
15. L Bowman, JC Hopewell, F Chen, K Wallendszus, W Stevens, R Collins, SD Wiviott, CP Cannon, E Braunwald, E Sammons, MJ Landray MJ. HPS3/TIMI55–REVEAL Collaborative Group. Effects of anacetrapib in patients with atherosclerotic vascular disease. N Engl J Med 2017; 377:1217–1227. doi: 10.1056/NEJMoa1706444.