Cholesterol is a molecule required by every cell of the body in fairly large amounts. It can be easily synthesised by these cells, or taken up by them from LDL and other ApoB lipoproteins, but cannot be broken down. Cholesterol is not soluble in water, and thus must be carried through the blood on lipoprotein particles. When the cholesterol produced or taken up by the cells of the body becomes surplus to requirements it is extracted by HDL (ApoA1 lipoproteins) and carried back to the liver for disposal as bile salts and acids (most of this cholesterol is reabsorbed and recycled, but there is also a variable amount lost in faeces). Reverse cholesterol transport (RCT) is the term used for this extraction of unneeded cholesterol. Here we describe a simplified version of reverse cholesterol transport, how this has been modified by new research into HDL, and we explain the effect of raising or lowering insulin and insulin sensitivity on RCT.
This video gives a good overview of the systems we’ll be describing. (The brain has its own, largely separate cholesterol system which we’ll ignore for now).
Cholesterol and insulin
We have about 30g of cholesterol in our bodies, and synthesise well over a gram a day. Only 10% of this is synthesised in the liver, and even less if we eat cholesterol or have a reduced requirement. Our requirement goes up when we are growing (cells are expanding and new cells being made) and down when we are fasting or losing weight (when fat cells and glycogen cells are shrinking, and autophagic processes are clearing unwanted cells). Thus it makes sense, and helps to keep cholesterol in balance with requirements, that cholesterol synthesis is stimulated by insulin (the fed state hormone) and inhibited by glucagon (the fasting state hormone). An additional check on cholesterol synthesis in the fasting state is the activation of AMPK by the ketone body B-hydroxybutyrate. No surprises then that cholesterol synthesis is found to be increased in type 2 diabetes.
If scientists want to create the early signs of heart disease in animals, they need to feed them doses of cholesterol much larger than the total capacity of the body to make cholesterol. However, Jerry Stamler, one of the founding fathers of the diet heart hypothesis, found in the early 1960’s that animals treated in this way got better when the cholesterol feeding stopped – unless they were given extra insulin. This vital clue was missed in the later rush to change our diets – Jerry Stamler advised the population to avoid egg yolk and replace fat with refined carbs, yet human diets never supplied the amount of cholesterol he fed his animals – unfortunately, the new, modified human diet would start to increase insulin to the high levels seen in those chickens once the diabesity epidemic got underway.
Reverse Cholesterol Transport
Fortunately our gut and liver cells make a protein called ApoA1, which the liver turns into something called a nascent HDL particle. Unlike VLDL and the other ApoB particles, which are released from the liver as large, fat and cholesterol laden spheres, HDL is produced in an embryonic state, just a few proteins with little if anything in the way of lipids (lipid-poor ApoA1), and only becomes what we call HDL by performing its cholesterol-gathering role out in the body.
If we focus on the cells believed to play the major role in atherosclerosis, macrophages (large immune cells) which can turn into foam cells if they become overloaded with cholesterol, we can see HDL at work. Macrophages clear the blood of infectious agents and damaged particles, and have a particular affinity for oxidised LDL particles (oxLDL). LDL becomes oxidised if it stays in the blood too long (more likely with higher levels or small, dense particles) and is exposed to excessive glucose and fructose levels after meals, or to smoking and other oxidative stressors.[7,8,9] Brown and Goldstein, who won the Nobel Prize for discovering the LDL receptor, estimated that 30-60% of LDL is cleared from circulation by macrophages. (Macrophages exposed to excess insulin increase their uptake of oxLDL by 80%). The oxLDL is then broken down and the cholesterol stored – remember it can’t be broken down. As in other cells, any excess is sent to the surface of the cell, to transporters and other structures that make it available for HDL to pick up, as free cholesterol (cholesterol efflux). If this doesn’t happen for some reason, over a long period, there’s a risk of foam cell formation and atherosclerosis. (Macrophages exposed to excess insulin decrease their efflux of cholesterol to HDL by 25%).
LCAT and esterification
After HDL picks up free cholesterol, this is esterified by an enzyme called lecithin cholesterol acyltransferase (LCAT), making the HDL particle larger. Cholesteryl ester (CE) is cholesterol joined to a fatty acid, usually an unsaturated fatty acid, which is supplied from the phospholipids also picked up from cells by HDL. The more effectively HDL can esterify cholesterol, the sooner it can return to pick up more from the macrophage (or other cell) – this is called HDL efflux capacity – and the phospholipids found in egg yolk have been shown to increase HDL efflux capacity. Phospholipids, found in all whole foods, especially fatty ones like eggs, nuts, seeds, liver, shellfish, and soya beans, are good things to have in your diet; you won’t get them from eating flour, sugar, and oil.
Cholesteryl oleate – a cholesteryl ester
CETP – swapping cholesteryl esters for triglycerides
HDL renews itself in the bloodstream by moving cholesteryl esters onto VLDL and other ApoB particles, in more-or-less equal exchange for triglycerides (TGs), through a banana-shaped protein tunnel called Cholesteryl Ester Transport Protein (CETP) which docks between ApoA1 and ApoB particles. HDL can then shed the TGs picked up to help feed cells along its path (as ApoB particles also do), turning them into free fatty acids and glycerol by the action of lipase enzymes. However, the CETP exchange is another place where things can go wrong. If there are too many TGs on VLDL, and too many TG-rich VLDL particles, and fats are not being burned by the body (yes, we’re talking about insulin resistance again), then the piling of TGs onto HDL via CETP will result in its recall to the liver after limited efflux. Carrying lots of TGs back to the liver that made them is not a good use of HDL’s time. And the cholesterol esters being transferred to former TG-rich VLDL is what makes the “Pattern B” lipoproteins, small dense LDL, which are more likely to oxidise and more easily taken up by macrophages. LDL really, once it’s done its job of delivering fat, cholesterol, antioxidants and proteins to cells that need them, ought to be helping in reverse cholesterol transport by ferrying the extra cholesterol esters it received from HDL back to the liver. Large, cholesterol-dense LDL particles – “Pattern A” – are better at this. Small, dense LDL particles aren’t taken up as avidly by the liver, so tend to stay in circulation and oxidize. Hence the TG/HDL ratio is a critical predictor of cardiovascular risk, whether or not we factor in LDL.
The exchange via CETP action is thought responsible for the inverse relationship between levels of TG and HDL-C. Specifically, the larger the VLDL pool (higher TG level), the greater the CETP-mediated transfer of CE from HDL to VLDL in exchange for TG, resulting in TG-rich small, dense HDL which are catabolized more rapidly, leading to low levels of HDL-C. These small, dense HDL also have reduced antioxidant and anti-inflammatory properties. Thus, the greater the increase in hepatic VLDL-TG synthesis and secretion that characterizes insulin-resistant/hyperinsulinemic individuals, the lower will be the HDL-C concentration.
Insulin resistance in this population (n=103,000) was stratified by tertiles of TG and HDL, with the insulin sensitive tertile having a mean TG/HDL ratio of 1.1 
Fasting, weight loss, and LDL
People who are naturally lean and active and have good insulin sensitivity are at very low risk of cardiovascular disease; they tend to burn fat and have low TG/HDL ratios on any diet. Paradoxically, LDL rises sharply when such people fast for long periods. Despite this, no-one as far as we know has ever suggested that not eating enough causes atherosclerosis. Of course TGs and insulin also fall, while HDL stays the same. But strangely, this rise in LDL does not happen in obese people or people with atherosclerosis.
In healthy lean individuals, LDL and ApoB rise as Insulin-like growth factor falls during a fast.
In obesity or T2DM, or in this case a patient with arteriosclerosis, cholesterol and LDL do not rise during a fast.
Phinney and colleagues found that LDL first fell, then rose significantly, during major weight loss. They calculated that this was due to the delayed removal of around 100g extra cholesterol from the adipose of obese people. LDL became normal when a weight maintenance diet replaced the (low fat, reduced calorie) weight loss diet.
Think about this – all of this extra 100g of cholesterol, 3x the usual whole body content, was eventually removed by reverse cholesterol transport. Some of it ended up on LDL, increasing the LDL count to the level where statins would be indicated according to guidelines. This did not prevent its removal. There is no “LDL gradient” that forces cholesterol back into the body. The LDL level doesn’t tell you whether cholesterol is coming or going – the TG/HDL ratio is the best guide to that.
Hepatic lipase – burning fat.
ApoA1 and HDL production increases the release of hepatic lipase, so in a sense ApoA1 is a fat-burning protein, which helps to explain why eating fat increases ApoA1 output.[18, 19] More lipase means lower TGs all round. So, making more HDL can lower TGs, just as making too many TGs can lower HDL – but only the latter is likely to be harmful.
Of course, a low fat, high carbohydrate diet decreases ApoA1, but this doesn’t mean it’s bad if you’re insulin sensitive and have low TGs (and low LDL) eating such a diet, as many people do; the lower lipid circulation all round probably just means that less ApoA1 will be required for equilibrium. However, the old assumption that the lower fat higher carb diet is the “Prudent” diet hasn’t aged well.
We have previously reported that apoA-I and HDL directly affect HL-mediated triacylglycerol hydrolysis, and showed that the rate of triacylglycerol hydrolysis is regulated by the amount of HDL in plasma.
The antioxidant and antiinflammatory benefits of HDL.
Reverse cholesterol transport is the core business of HDL, but it isn’t the only business; HDL is like a busy doctor with a useful bag of healing tricks trundling up and down your bloodstream. For example, HDL carries an antioxidant protein, PON1. When a fatty hamburger meal rich in lipid peroxides was fed to 71 subjects, those with higher HDL experienced a much smaller rise in oxLDL.
The pre-meal HDL level was associated with the extent of the postprandial rise in oxidized LDL lipids. From baseline to 6 h after the meal, the concentration of ox-LDL increased by 48, 31, 24, and 16 % in the HDL subgroup 1, 2, 3, and 4, respectively, and the increase was higher in subgroup 1 compared to subgroup 3 (p = 0.028) and subgroup 4 (p = 0.0081), respectively. The pre-meal HDL correlated with both the amount and the rate of increase of oxidized LDL lipids. Results of the present study show that HDL is associated with the postprandial appearance of lipid peroxides in LDL. It is therefore likely that the sequestration and transport of atherogenic lipid peroxides is another significant mechanism contributing to cardioprotection by HDL.
Tregs or T regulatory cells are a type of immune cell that switches off inflammatory responses. They are also a type of cell that takes up HDL, and HDL selectively promotes their survival. This is a good thing.
Can LDL help in reverse cholesterol transport?
The answer is yes – if it’s large LDL particles (Pattern A), not so much small dense ones. Triglycerides and VLDL, on the other hand, are no help at all.
There are two pathways by which RCT can occur. In the first, the scavenger receptor class B type 1 (SRB-1) mediates hepatic uptake of CE from HDL particles without uptake of apoA-I or the whole HDL particle . In the second pathway, cholesteryl ester transfer protein (CETP) catalyzes the transfer of CE from HDL to apoB-containing lipoproteins (VLDL and LDL) in exchange for TG from the apoB-containing lipoproteins (Fig. 1) [21, 75]. This exchange results in apoB-containing lipoproteins which are enriched with CEs and depleted of TGs, and HDL particles which are depleted of CEs and enriched with TGs. The TG-rich and CE-poor HDL particles are catabolized faster than large, CE-rich HDL (apoA-I FCR is increased as noted in Fig. 1), a finding resulting in lower levels of HDL-C in the setting of high TG levels . The apoB-containing lipoproteins, now enriched in CE, can also be taken up by the liver receptors as previously described . When TG levels are high, the apoB particles are TG-enriched and hepatic lipase then hydrolyzes the TGs within the TG-rich LDL to release FFAs, a process which remodels the LDL particles into smaller and denser LDL particles which can enter the arterial intima more easily than larger LDL particles, thus making them more atherogenic (Fig. 1). Small, dense LDL particles also bind less avidly to the LDL receptor, thus prolonging their half-life in the circulation and making these particles more susceptible to oxidative modification and to subsequent uptake by the macrophage scavenger receptors .
An unusual experiment (using a radioactive nanoemulsion mimetic of LDL) showed that LDL cholesterol is removed from circulation more rapidly in resistance-trained healthy men than in sedentary healthy men. oxLDL was 50% lower in the resistance-trained men – but total LDL levels were the same, probably as a result of increased beta-oxidation (fat burning).
Why are doctors being confused about HDL and reverse cholesterol transport?
There’s a trend in mainstream medicine to be dismissive of HDL and treat reverse cholesterol transport as unimportant; LDL lowering is the thing. New evidence from genetics, epidemiology, and drug trials is increasingly misinterpreted in this way – probably because drugs that increase HDL have, with some exceptions, been failures. However, drugs that raise HDL, and lower LDL, by inhibiting CETP are not helping either particle do its job; so far, they have neither decreased nor increased the rate of heart attacks in people taking them. Drugs that raise HDL by increasing ApoA1 and nascent HDL output, like the fibrates (e.g. gemfibrozil), do decrease CHD – but only in people with lower HDL and higher TGs! Moderate alcohol use, which also increases ApoA1 output, seems to have a similar effect, though the first randomised controlled trial of this observational hypothesis is only beginning.[23, 24] Even statins help with RCT by decreasing the synthesis of cholesterol in peripheral tissues, thus leaving more room on HDL for efflux cholesterol – again, statins only seem to reduce CHD in the subgroup of people with lower HDL. Clearly reverse cholesterol transport is very important, and efficient reverse cholesterol transport can best explain why so many people with high LDL and high cholesterol do enjoy long lives free from cardiovascular disease. Some ApoA1 genes that especially promote RCT are associated with reduced CVD risk, notably ApoA1 milano – which is actually associated with low HDL, because HDL clearance is so rapid – a paradox which highlights the trickiness of measuring a dynamic process across all tissues only by what appears in the blood. Efficient RCT is associated with lean genes, but it’s largely something you have to work for – eating right, exercising, and looking after yourself in various ways – including giving up smoking, or not starting – which may be why the drug industry has largely given up on it.
We observed that normolipidemic smokers present higher total plasma and HDL phospholipids (PL) (P < .05), 30% lower postheparin hepatic lipase (HL) activity (P < .01), and 40% lower phospholipid transfer protein (PLTP) activity (P < .01), as compared with nonsmokers. The plasma cholesteryl ester transfer protein (CETP) mass was 17% higher in smokers as compared with controls (P < .05), but the endogenous CETP activity corrected for plasma triglycerides (TG) was in fact 57% lower in smokers than in controls (P < .01). Lipid transfer inhibitor protein activity was also similar in both groups. In conclusion, the habit of smoking induces a severe impairment of many steps of the RCT system even in the absence of overt dyslipidemia.
The latest study on very, very high HDL – why isn’t it good?
Last year we wrote about the CANHEART study, which seemed to show adverse health effects of higher HDL. We wrote then that this was probably showing an effect of alcoholism, hereditary CETP defects, and other factors, and not an increase in heart disease. A new study allows us to look at this problem in more detail.
When compared with the groups with the lowest risk, the multifactorially adjusted hazard ratios for all-cause mortality were 1.36 (95% CI: 1.09–1.70) for men with HDL cholesterol of 2.5–2.99 mmol/L (97–115 mg/dL) and 2.06 (1.44–2.95) for men with HDL cholesterol ≥3.0 mmol/L (116 mg/dL). For women, corresponding hazard ratios were 1.10 (0.83–1.46) for HDL cholesterol of 3.0–3.49 mmol/L (116–134 mg/dL) and 1.68 (1.09–2.58) for HDL cholesterol ≥3.5 mmol/L (135 mg/dL).
Those are some very high HDL levels, and not surprisingly fewer than 4% of men and even fewer women had HDL levels so high that they were associated with any extra risk.
Compare that with 40% of both men and women having low HDL levels that were associated with an equally elevated risk!
Further, the risk associated with very high HDL, though it does include cardiovascular deaths, doesn’t seem to include much increased risk of heart attacks and strokes.
This is consistent with alcoholism (a confounder not measurable with accuracy, as we described in the CANHEART analysis) increasing deaths from heart failure, cancer, and other causes, and with no further benefit (but maybe not much harm overall) from CETP variants elevating HDL. Furthermore, interactions between heavy alcohol consumption and genes associated with higher HDL have been noted in some populations.
Note that the HDL level associated with lowest heart disease and stroke incidence in this study is well to the right of the bell curve of population HDL distribution. Most of these people could have done with more HDL.
Madsen et al discuss their results soberly; although they fail to discuss the potential role of alcohol, which would explain the exact pattern of increased mortality seen well, and don’t highlight the 10-fold larger impact associated with low HDL in their study, there is nothing biased about their analysis. The European Heart Journal’s editorial was also worth reading.
However, as reported in medical media, the message changed a bit.
“It appears that we need to remove the focus from HDL as an important health indicator in research, at hospitals and at the general practitioner. These are the smallest lipoproteins in the blood, and perhaps we ought to examine some of the larger ones instead. For example, looking at blood levels of triglyceride and LDL, the ‘bad’ cholesterol, are probably better health indicators,” he notes.
Well yes, looking at everything is good, and TGs and the TG/HDL ratio as well as LDL will give you extra information about the likely reasons for low HDL and whether you need to worry about it. However, Denmark, where the extremely high HDL study was done, is a place where high LDL (the ‘bad’ cholesterol, remember) is associated with lower all-cause mortality in those over 50 free from diabetes or CVD at the start of the study. Over 50 is when most CVD and type 2 diabetes is diagnosed, so LDL might not be all that informative unless you can look at the subclasses of oxLDL, sdLDL, particle number, and so on (of course part of the effect of LDL in Denmark will be due to that country’s higher dairy fat intake, which will also raise HDL and LDL particle size, maybe helping to explain why the association is so favourable in that population).
If we look at the PURE study, higher fat consumption is associated with both higher LDL and higher ApoA1 and HDL, with saturated fat (like all fat types) tending to improve the ApoB/ApoA1 ratio. This is consistent with many other lines of evidence.
Intake of total fat and each type of fat was associated with higher concentrations of total cholesterol and LDL cholesterol, but also with higher HDL cholesterol and apolipoprotein A1 (ApoA1), and lower triglycerides, ratio of total cholesterol to HDL cholesterol, ratio of triglycerides to HDL cholesterol, and ratio of apolipoprotein B (ApoB) to ApoA1 (all ptrend<0·0001).
This is just what fat-burning does, and there’s maybe not a lot of reason to think it’s good or bad per se. What is good about it is, that fat burning lowers insulin. Insulin is what makes your cells hoard cholesterol, and it’s also one of the things that can mess with reverse cholesterol transport. If you’re making or using excess insulin, the switch to a fat burning metabolism allows the insulin to normalise and causes your cells, including the macrophages, to let go of cholesterol – and when they do, the lipoproteins are there ready to carry it away.
Reverse cholesterol transport manages cholesterol flux through all cells and helps us reach a healthy old age.
LDL cholesterol is not a reliable guide to the state of cholesterol flux unless TG/HDL (and HbA1c) are factored in as well. LDL may increase when cholesterol is being removed or in states where it is not being taken up by cells.
Reverse cholesterol transport can remove prodigious amounts of cholesterol from the body during weight loss.
Excessive triglycerides due to insulin resistance can impair reverse cholesterol transport, as can smoking.
Various nutritional factors found in whole food diets have been found to assist in reverse cholesterol transport (including phospholipids, CLA, and polyphenols).
HDL in the high (if not the “extremely high”) range usually correlates with efficient reverse cholesterol transport and has benefits for cardiovascular health, inflammation, antioxidant status etc, but people with HDL outside (higher or lower than) the ideal range can be equally healthy if their overall metabolic health (insulin sensitivity) is good.
The TG/HDL ratio is a good measure of insulin sensitivity, and if excessive can be improved by lowering excessive insulin levels. A low carb diet, intermittent fasting, exercise, or weight loss are all effective ways to correct the TG/HDL ratio.
 Bae HR, Kim DH, Park MH, et al. β-Hydroxybutyrate suppresses inflammasome formation by ameliorating endoplasmic reticulum stress via AMPK activation. Oncotarget. 2016;7(41):66444-66454. doi:10.18632/oncotarget.12119.
 Gylling H, Hallikainen M, Pihlajamäki J, et al. Insulin sensitivity regulates cholesterol metabolism to a greater extent than obesity: lessons from the METSIM Study. Journal of Lipid Research. 2010;51(8):2422-2427. doi:10.1194/jlr.P006619.
 Regulation of hepatic LDL metabolism in the guinea pig by dietary fat and cholesterol.
Lin ECK, Fernandez ML, Tosca MA, McNamara DJ. Journal of Lipid Research. 1994; 35:446-457
 Stamler J, Pick R, Katz LN. Effect of Insulin in the Induction and Regression of Atherosclerosis in the Chick. Circulation Research. 1960; 8:572-576.
 Brown MS, Goldstein JL. Lipoprotein Metabolism in the Macrophage: Implications for Cholesterol Deposition in Atherosclerosis. Annual Review of Biochemistry 1983 52:1, 223-261
 Kopprasch S, Pietzsch J, Kuhlisch E et al. In Vivo Evidence for Increased Oxidation of Circulating LDL in Impaired Glucose Tolerance. Diabetes 2002; 51(10): 3102-3106. https://doi.org/10.2337/diabetes.51.10.3102
 Vos MB, Weber MB, Welsh J et al. Fructose and Oxidized LDL in Pediatric Nonalcoholic Fatty Liver Disease: A Pilot Study. Archives of pediatrics & adolescent medicine. 2009;163(7):674-675. doi:10.1001/archpediatrics.2009.93.
 Medina-Navarro R, Durán-Reyes G, Díaz-Flores M, Kumate Rodríguez J, Hicks JJ. Glucose autoxidation produces acrolein from lipid peroxidation in vitro. Clin Chim Acta. 2003; 337(1-2):183-5.. PMID: 14568199
 Park YM, Sangeeta R Kashyap SR, Major JA, Silverstein RL. Insulin promotes macrophage foam cell formation: potential implications in diabetes-related atherosclerosis. Laboratory Investigation. 2012; 92, 1171-1780.
 Andersen CJ, Blesso CN, Lee J, et al. Egg Consumption Modulates HDL Lipid Composition and Increases the Cholesterol-Accepting Capacity of Serum in Metabolic Syndrome. Lipids. 2013;48(6):10.1007/s11745-013-3780-8. doi:10.1007/s11745-013-3780-8.
 Welty FK. How Do Elevated Triglycerides and Low HDL-Cholesterol Affect Inflammation and Atherothrombosis? Current cardiology reports. 2013;15(9):400. doi:10.1007/s11886-013-0400-4.
 Bertsch RA, Merchant MA. Study of the Use of Lipid Panels as a Marker of Insulin Resistance to Determine Cardiovascular Risk. The Permanente Journal. 2015;19(4):4-10. doi:10.7812/TPP/14-237.
 Sävendahl L, Underwood LE. Fasting increases serum total cholesterol, LDL cholesterol and apolipoprotein B in healthy, nonobese humans. J Nutr. 1999 Nov;129(11):2005-8.
 Ende N. Starvation studies with special reference to cholesterol. Am. J. Clin. Nutr. 1962. 11:270-280.
 Phinney SD, Tang AB, Waggoner CR, Tezanos-Pinto RG, Davis PA. The transient hypercholesterolemia of major weight loss. Am J Clin Nutr. 1991 Jun;53(6):1404-10.
 da Luz PL, Favarato D, Faria-Neto JR Jr, Lemos P, Chagas ACP. High ratio of triglycerides to HDL-cholesterol predicts extensive coronary disease. Clinics. 2008; 64:427-32
 Ramsamy TA, Boucher J, Brown RJ et al. HDL regulates the displacement of hepatic lipase from cell surface proteoglycans and the hydrolysis of VLDL triacylglycerol. The Journal of Lipid Research. 2003; 44: 733-741.
 Chatterjee C, Sparks DL. Hepatic Lipase, High Density Lipoproteins, and Hypertriglyceridemia. The American Journal of Pathology. 2011;178(4):1429-1433. doi:10.1016/j.ajpath.2010.12.050.
 Tiainen S, Ahotupa M, Ylinen P, Vasankari T. High density lipoprotein level is negatively associated with the increase of oxidized low density lipoprotein lipids after a fatty meal. Lipids. 2014; 49(12): 1225-32. doi: 10.1007/s11745-014-3963-y. Epub 2014 Oct 31.
 Rueda CM, Rodríguez-Perea AL, Moreno-Fernandez M et al. High density lipoproteins selectively promote the survival of human regulatory T cells. J Lipid Res. 2017 Aug;58(8):1514-1523. doi: 10.1194/jlr.M072835. Epub 2017 Apr 4.
 da Silvaa JL, Vinagrea CGCM, Morikawa AT et al. Resistance training changes LDL metabolism in normolipidemic subjects: A study with a nanoemulsion mimetic of LDL.
Atherosclerosis. 2011; 219(2): 532-537.
 Mukamal KJ, Clowry CM, Murray MM, et al. Moderate Alcohol Consumption and Chronic Disease: The Case for a Long-Term Trial. Alcohol Clin Exp Res. 2016 Nov;40(11):2283-2291. doi: 10.1111/acer.13231. Epub 2016 Sep 30. Review.
 Moderate Alcohol and Cardiovascular Health Trial (MACH15) . https://clinicaltrials.gov/ct2/show/NCT03169530
 Leite JO, Fernandez ML. Should we take high-density lipoprotein cholesterol levels at face value? Am J Cardiovasc Drugs. 2010; 10(1):1-3. doi: 10.2165/11319590-000000000-00000.
 Zaratin AC, Quintão EC, Sposito AC et al. Smoking prevents the intravascular remodeling of high-density lipoprotein particles: implications for reverse cholesterol transport. Metabolism. 2004 Jul;53(7):858-62.
 Madsen CM, Varbo A, Nordestgaard BG. Extreme high high-density lipoprotein cholesterol is paradoxically associated with high mortality in men and women: two prospective cohort studies. European Heart Journal,. 2017; 38(32): 2478–2486, https://doi.org/10.1093/eurheartj/ehx163
 Devenyi P, Robinson GM, Roncari DA. Alcohol and high-density lipoproteins. Canadian Medical Association Journal. 1980;123(10):981-984.
 Volcik K, Ballantyne CM, Pownall HJ, Sharrett AR, Eric Boerwinkle E. Interaction Effects of High-Density Lipoprotein Metabolism Gene Variation and Alcohol Consumption on Coronary Heart Disease Risk: The Atherosclerosis Risk in Communities Study. Journal of Studies on Alcohol and Drugs, 68(4), 485–492 (2007).
 Barter PJ, Rye K-A. HDL cholesterol concentration or HDL function: which matters? European Heart Journal. 2017; 0: 1–3
 Bathum L, Depont Christensen R, Engers Pedersen L, Lyngsie Pedersen P, Larsen J, Nexøe J. Association of lipoprotein levels with mortality in subjects aged 50 + without previous diabetes or cardiovascular disease: A population-based register study. Scandinavian Journal of Primary Health Care. 2013;31(3):172-180. doi:10.3109/02813432.2013.824157.
 Mente A, Dehghan M, Rangarajan S et al. Association of dietary nutrients with blood lipids and blood pressure in 18 countries: a cross-sectional analysis from the PURE study.
Lancet Diabetes Endocrinol 2017 Published Online August 29, 2017 http://dx.doi.org/10.1016/S2213-8587(17)30283-8
PURE lipids and BP