I must confess: consistently eating low-carb, high-fat foods breaks two or three of the cardinal taboos of what many believe are part of a “heart healthy diet”: that is, I consume lots of saturated fats and cholesterol-rich foods from animal sources–preferably local, grass-fed or pasture-raised animal sources.
However, after improving and maintaining lots of measurable health markers for over 3 years now, sometimes I ponder part of Queen’s “Bohemian Rhapsody” (“Is this the real life?/Is this just fantasy?”) and reflectively ask myself “why am I not dead or suffering from cardiovascular disease (CVD)?”
Expectations if Cholesterol Caused CVD
If high cholesterol levels were the major cause of CVD as we are normally led to believe, then we should expect data to consistently and positively reflect the following three patterns on large scales:
(1) Those with high cholesterol levels should correlate with a higher incidence of all-cause mortality, specifically from CVD;
(2) Those with lower cholesterol levels should should correlate with a lower incidence of all-cause mortality, specifically from CVD; and
However, a number of such large scale studies have correlated lower cholesterol levels with increased mortality and higher cholesterol levels with increased longevity, especially among the elderly–e.g. Isehara Study, Honolulu Heart Program study, Norwegian HUNT 2 Study.
That is, those with the highest cholesterol levels tend to live longer than those with the lowest cholesterol levels.
The Framingham Heart Study is perhaps the longest (and most cited) cohort study (1948 – present?) to look at the associations of various health markers and health over time. In 1992, 44 years after the project began, the study director, Dr. William Castelli, MD, wrote the following in an editorial to the Archives of Internal Medicine:
…Most of what we know about the effects of diet factors, particularly the saturation of fat and cholesterol, on serum lipid parameters derives from metabolic ward—type studies.2,3 Alas, such findings, within a cohort studied over time have been disappointing, indeed the findings have been contradictory. For example, in Framingham, Mass, the more saturated fat one ate, the more cholesterol one ate, the more calories one ate, the lower the person’s serum cholesterol….
There are other studies that could be listed here, but suffice it to say that such inconsistent findings would have to be reconciled or explained away to order to maintain the notion that higher levels of cholesterol cause or highly correlate with increased mortality over time, specifically from CVD.
(3) Interventions that effectively reduce high (specifically LDL) cholesterol levels–e.g. statin drugs–should correlate to lower incidence of all-cause morality, specifically from CVD, as compared to a control or untreated group.
In another post, I’ll discuss some thoughts on trying to lower cholesterol levels by eating low-cholesterol foods (hint: our body doesn’t seem to quite work this simply!).
But for now, one effective way to lower cholesterol levels is to do so artificially through statin drugs. And so, for the sake of argument, let’s assume that propositions (1) and (2) hold. Then it would seem to follow that if we could lower cholesterol levels (specifically LDL the “bad” type of cholesterol) through statins (which they seem to do pretty effectively), then there should be a lower incidence of all-cause mortality, specifically from CVD, in comparison to similar non-treated groups.
Here, however, the evidence has been underwhelming for the most part. Below is a summary list of eight important randomized control trial (RCT) results of major statin drug intervention trials as they related to stroke preventions, from the Journal of Cerebral Blood Flow and Metabolism
Most of the time, major media outlets will focus on the relative risk reduction (RRR) part of the study, which provides more material for sensational headlines by statistically making mountains out of molehills.
From a patient and potential statin consumer perspective, it’s far more important to know the absolute risk reduction (ARR) difference between the treatment group versus the non-treatment group and also the numbers needed to treat (NNT). The ARR difference is important because I want to know how many bad outcomes (like death or CVD-related events, like stroke) are prevented with statin intervention compared to not taking the statin. In these eight statin drug trials, this difference ranged from 0 to 1.6. In this particular context, this means that there wasn’t too much significant difference between the groups that took statins and the groups that didn’t take statins as it related to stroke prevention over time.
The NNT is also important because I want to know how many people taking statins over time are needed to prevent just one stroke. It’s a measure of how efficacious a drug is at doing what it’s intended to do over a certain period of time, and also a way to evaluate different statin efficacy rates against one another. The lower the number the better. As a patient, it lets me know my odds of benefiting from a particular type of statin intervention. For example, if the NNT is 143, I have a 1 in 143 chance over some time of benefiting from treatment assuming I resemble the control and intervention groups–not great poker odds by any stretch. As a reference point, taking 800 mg of ibuprofen has an NNT of about 1.6 within 4-6 hours in terms of achieving some measure of pain relief.
When looking at RCTs, which are the gold standard in terms of assessing the effectiveness of a single type of intervention (and no other confounding variables), knowing the ARR and NNT gives me a realistic picture of the potential upsides when deciding whether to take a statin or not.
In terms of cost-benefit analysis, I can then properly weigh my decision as patient and potential consumer against potential and known side effects (keeping in mind that no drug is a free lunch in terms of possible consequences), including fatigue, cognitive decline, muscle problems, increased diabetes risk, memory loss, and CoQ10 deficiency among others.
A more recent and controversial study is known as the JUPITER trial (2008). JUPITER was another RCT but pre-selected for 17,000+ people without any history of heart disease who had normal to low LDL cholesterol levels but had high C-Reactive Protein (CRP) markers. The study lasted short of two years. A few peculiar things to note: (A) AstraZeneca, the manufacturer of Crestor, which was the drug used in the intervention group, funded the study, and (B) the lead author, who developed the device to measure CRP, also profits from its use.
The ARR difference between intervention and placebo groups for fatal stroke and/or fatal heart attack ranged from 0.2% to 0.6% and the NNT was 25 over 5 years (or 95 over two years, depending on which time frame one wants to use).
1. LDL cholesterol is not measured directly; instead on most lipid panels, it is estimated indirectly using the Friedewald equation (LDL = TC – HDL – [TG/5]) and seems to hold true if TGs are above 100 but below 400. Total Cholesterol (TC), Triglycerides (TG), and HDL cholesterol are measured directly. See Dr. Michael Eades’s, MD, post about this here.
It might cost a little extra, but I would totally ask for an LDL particle size (“Good” LDL particles are large and fluffy compared to “Bad” LDL particles that are small and dense) test if the standard lipid panel comes back with anything above 130 mg/dL AND the doctor wants to automatically prescribe a statin drug as a result. The reason why is that very-low-carb, high-fat diets tend to raise HDL levels and dramatically drop TGs, both of which are measured directly, unlike LDL.
With an Excel spreadsheet, it’s easy to plug in the Friedewald equation to see how changing certain variables–most notably the TG–affects the calculated LDL number.
2. I’ve always wondered how cholesterol guidelines are set and revised. I mean who gets to decide that 100 mg/dL of LDL cholesterol is “optimal”? And on what basis are such guidelines based upon for an entire population? Is it done by competent but disinterested people?
Well, in 2004, the National Cholesterol and Education Program (NCEP) published an update based on 5 trials (3 of which are listed above–i.e. HPS, ASCOT, and PROSPER) to lower the recommended cholesterol guidelines for adults.
8 out of the 9 doctors who sat on this decision board had significant financial ties to various cholesterol drug manufacturers:
Dr. Grundy has received honoraria from Merck, Pfizer, Sankyo, Bayer, Merck/Schering-Plough, Kos, Abbott, Bristol-Myers Squibb, and AstraZeneca; he has received research grants from Merck, Abbott, and Glaxo Smith Kline.
Dr. Cleeman has no financial relationships to disclose.
Dr. Bairey Merz has received lecture honoraria from Pfizer, Merck, and Kos; she has served as a consultant for Pfizer, Bayer, and EHC (Merck); she has received unrestricted institutional grants for Continuing Medical Education from Pfizer, Procter & Gamble, Novartis, Wyeth, AstraZeneca, and Bristol-Myers Squibb Medical Imaging; she has received a research grant from Merck; she has stock in Boston Scientific, IVAX, Eli Lilly, Medtronic, Johnson & Johnson, SCIPIE Insurance, ATS Medical, and Biosite.
Dr. Brewer has received honoraria from AstraZeneca, Pfizer, Lipid Sciences, Merck, Merck/Schering-Plough, Fournier, Tularik, Esperion, and Novartis; he has served as a consultant for AstraZeneca, Pfizer, Lipid Sciences, Merck, Merck/Schering-Plough, Fournier, Tularik, Sankyo, and Novartis.
Dr. Clark has received honoraria for educational presentations from Abbott, AstraZeneca, Bristol-Myers Squibb, Merck, and Pfizer; he has received grant/research support from Abbott, AstraZeneca, Bristol-Myers Squibb, Merck, and Pfizer.
Dr. Hunninghake has received honoraria for consulting and speakers bureau from AstraZeneca, Merck, Merck/Schering-Plough, and Pfizer, and for consulting from Kos; he has received research grants from AstraZeneca, Bristol-Myers Squibb, Kos, Merck, Merck/Schering-Plough, Novartis, and Pfizer.
Dr. Pasternak has served as a speaker for Pfizer, Merck, Merck/Schering-Plough, Takeda, Kos, BMS-Sanofi, and Novartis; he has served as a consultant for Merck, Merck/Schering-Plough, Sanofi, Pfizer Health Solutions, Johnson & Johnson-Merck, and AstraZeneca.
Dr. Smith has received institutional research support from Merck; he has stock in Medtronic and Johnson & Johnson.
Dr. Stone has received honoraria for educational lectures from Abbott, AstraZeneca, Bristol-Myers Squibb, Kos, Merck, Merck/Schering-Plough, Novartis, Pfizer, Reliant, and Sankyo; he has served as a consultant for Abbott, Merck, Merck/Schering-Plough, Pfizer, and Reliant.
None of this is to say that statins are totally ineffective at avoiding CVD, but the benefits based on the available evidence I’ve seen make it seem like a high-risk, low-reward proposition when the upsides (via absolute risk reduction and numbers needed to treat) and side effects are carefully considered together. And given today’s statin-mania, with pushes to treat healthy people with such drugs to include even children–which would sort of be like vaccination pushes but for CVD prevention–it’s good to be as sober with the data as possible until better evidence comes along. After all, I have more of a stake in my own overall health than anyone else, or at least that’s how it should be.