Sure, I’m losing weight…but is this healthy?
Posted September 15, 2011on:
Someone recently suggested to me that, while my diet is clearly producing weight loss, it is not a healthy diet. That got me thinking. What does “healthy” mean? How is it measured? And sure, I’ve read the books and blog articles demonstrating that a diet high in saturated fat and low in carbohydrates is healthy, but I wanted to see for myself.
First, a bit of background on myself: I spent over 15 years in medical research before I had kids and left my career. My field of study was cardiac electrophysiology, arrhythmia mechanisms, and cardiac pharmacology. I’ve published in journals such as Circulation Research and have served on peer-review panels for submissions to other cardiac health-related journals. I currently own a small consulting business in which I am often called upon to perform searches and reviews of evidence in the medical and scientific literature on various topics. So, I’m not only very comfortable with such search engines as PubMed, but I have access to original articles through my professional affiliations, and I have the training and experience to read and understand them.
I started with the first question: what do we mean by “healthy”? In this day and age, most of us Westerners are concerned about cardiovascular health (i.e. coronary artery disease, or cardiovascular disease – CVD) and diabetes is also gaining epidemic status. There is an important relationship between the two (diabetes and heart health) for which the evidence is rapidly growing. Metabolic syndrome, which is essentially a combination of obesity, diabetes, and heart disease risk factors, is a newly defined indicator of risk that emphasizes the growing understanding of the role of blood sugar regulation (i.e. the insulin system) in cardiovascular health.
How do we measure “good health”? Traditionally people have focussed on overall cholesterol levels, with emphasis on HDL concentration and LDL concentration. However, it is now generally accepted by the medical community that these are not nearly as strong or powerful predictors of risk compared to lipid profiles that go into more detail, specifically with reference to the size of the particles and their relationship with triglyceride levels. It turns out that small, dense LDL particles (often called LDL Pattern B) are much better indicators of risk than total LDL concentration. For example, a report from the Framingham Offspring Study reported that total number of LDL particles (which, for a given amount of cholesterol, would be higher if particle sizes are smaller) was a markedly significant predictor of CVD, whereas overall concentration of LDL was not . Regardless of where a scientist stands on the saturated fat debate, there are certain assumptions to which they all agree: HDL is good, small/dense LDL are especially bad while large/fluffy LDL are neutral, and triglycerides are bad.
Where studies start to conflict in their conclusions is when they look at the effects of diet on these indicators, as well as others such as insulin-sensitivity (a measure of risk for diabetes). This is a difficult field to study because there are so many variables. I found many studies of so-called “high fat” diets that did not distinguish between saturated fats vs. mono- or polyunsaturated fats. Those that did often included a carbohydrate count that would be considered very high by proponents of Primal/Paleo diets (i.e. carbs were kept at the “normal” level for a Western Diet). “High carb” diet regimes often didn’t distinguish between complex vs. simple carbs. Then there’s the fact that a diet low in fat is often lower in calories, so that must be controlled for as well. I found plenty of scientists who are questioning the traditional advice of a diet low in fat – especially saturated animal fats – and high in grains and starches but also plenty who are sticking firmly to that story. I decided instead to focus on what we all agree on – the indicators of cardiovascular and diabetic risk: the lipid profile.
I don’t have the time or space to list every article I read. My findings can be summarized by a recent review paper (essentially a summary of studies to date) that summarized what we know about the effects of diet on “atherogenic dyslipidemia” (which basically means: low HDL, high LDL, and high triglycerides). To summarize: high carbohydrate diets altered lipid profiles towards atherogenic dyslipidemia.
Switching from a low-carb/high-fat diet to a high-carb/low-fat diet caused LDL sizes to shift from big and fluffy (Pattern A = “good”) to small and dense (Pattern B = “bad”). Yup, the low-fat high carb diet made the lipid profile worse. Those who were Pattern B on the high-fat diet showed no change on the low-fat diet. Furthermore, those with Pattern A on the high-fat diet who remained Pattern A after switching to the low-fat diet were later switched to another, even lower-fat diet and then they switched to Pattern B.
And here’s another interesting fact: those with Pattern A or B on the high-fat diet experienced a decrease in overall LDL concentration when they went on the low-fat diet. Lowering LDL concentration is considered a “good thing”. And this decrease was much greater for Pattern B folks when they switched to low-fat compared to Pattern A folks. However, when you look at the size distribution it turns out that Pattern A folks saw their small/dense LDL go up while their large LDL went down a lot. Thus, overall, their LDL went down when on the low-fat diet, but it shifted to an unhealthy prevalence of small/dense LDL over large/fluffy LDL. The Pattern B folks saw an even greater reduction of overall LDL when they switched to a low-fat diet, but here the decrease was due mostly to decreases in large and intermediate sized LDL, with no change in the “bad” small/dense LDL. So while switching to a low-fat diet reduced LDL concentration, it did so while shifting the balance towards the small/dense (“bad”) LDLs.
Similar results have been found in a number of studies covered in this review. And it has been determined that it is the carbohydrates, not the fat, that is the issue. Turns out that low-carb diets are consistently associated with increases in the amount of good, fluffy LDL whereas high-carb diets convert people from that healthy Pattern A to an unhealthy Pattern B, even if the overall LDL concentration has decreased.
I could go on, but most of what I read agreed with this review. Basically, you get a better increase in HDL and a better decrease in triglycerides by eating a low-carb diet versus a high-carb diet. Overall serum cholesterol tends to go up or stay the same and so does LDL with the high fat diet, but only because of the increase in light/fluffy LDL (which have no association with CVD), whereas decreases in serum cholesterol and LDL associated with a low-fat/high-carb diet are the result of preferential decreases in the amount of large/fluffy LDL in favour of small/dense LDL. Overall cholesterol appears to be a rather useless measure of cardiovascular health, and despite decades of research nobody has ever been able to convincingly link total cholesterol to risk of CVD, whereas other measures (such as the triad of HDL, triglycerides, and LDL) have been convincingly demonstrated to be good indicators of risk.
So, I propose that in order to determine whether a diet is truly healthy, one should get a lipid profile that measures HDL, LDL (with distinctions between Pattern A and Pattern B), and triglycerides. Unfortunately it has been a very long time since I had a lipid profile, since I have never been considered at risk for CVD, diabetes, or metabolic syndrome, and when I did they did not routinely test for LDL particle size (even today it must be requested separately).
However, one can make predictions on what the lipid profile should be for someone who has been on a low-carb, high fat diet for some time. Those who follow conventional wisdom might predict that overall cholesterol would be elevated, serum triglycerides would be elevated, HDL would be low and LDL would be high. This would be the profile of someone at high risk of CVD, metabolic syndrome, and perhaps diabetes too. But the latest evidence in the literature does not support that.
Instead, I predict that the profile would look like this: high HDL and low triglycerides, with a possible increase in serum cholesterol and LDL due to a preferential increase in large/fluffy LDL and a significant decrease in small/dense LDL. Here’s an excerpt from an abstract from a recent study in the journal Metabolism that explains my hypothesis:
To test whether a short-term dietary intervention affects LDL particle size, we conducted a randomized, double-blind, crossover study using an intensive dietary modification in 12 nonobese healthy men with normal plasma lipid profile. Participants were subjected to 2 isocaloric 3-day diets: high-fat diet (37% energy from fat and 50% from carbohydrates) and low-fat diet (25% energy from fat and 62% from carbohydrates). Plasma lipid levels and LDL particle size were assessed on fasting blood samples after 3 days of feeding on each diet. The LDL particles were characterized by polyacrylamide gradient gel electrophoresis. Compared with the low-fat diet, plasma cholesterol, LDL cholesterol, and high-density lipoprotein cholesterol were significantly increased (4.45 vs 4.78 mmol/L, P = .04; 2.48 vs 2.90 mmol/L, P = .005; and 1.29 vs 1.41 mmol/L, P = .005, respectively) following the 3-day high-fat diet. Plasma triglycerides and fasting apolipoprotein B-48 levels were significantly decreased after the high-fat diet compared with the low-fat diet (1.48 vs 1.01 mmol/L, P = .0003 and 9.6 vs 5.5 mg/L, P = .008, respectively). The high-fat diet was also associated with a significant increase in LDL particle size (255.0 vs 255.9 Å;P = .01) and a significant decrease in the proportion of small LDL particle (<255.0 Å) (50.7% vs 44.6%, P = .01). As compared with a low-fat diet, the cholesterol-raising effect of a high-fat diet is associated with the formation of large LDL particles after only 3 days of feeding.
If it turns out my HDL is high, my triglycerides are low, and my LDL is Pattern A, then I think I will have demonstrated that my diet does not only produce effective (and nearly effortless) weight loss without restricting caloric intake (i.e. creating an energy deficit and essentially ‘starving’ my cells) but also produces a lipid profile associated with a decreased risk of CVD and metabolic syndrome.
My husband did get a lipid profile when he began our diet. It has been six months now for him and he’s going to go for another one (when he started he had metabolic syndrome – obesity, insulin sensitivity, high LDL and triglycerides and low HDL). I’ll publish his results when he gets them (he has already lost over 40 lbs). I will also go for a lipid profile although I don’t have a baseline from which to make comparisons. Still I will be able to demonstrate whether, after 5 months on my diet, my profile fits that of someone at risk for CVD or not.
 Cromwell WC, Otvos JD, Keyes MJ, Pencina MJ, Sullivan L, Vasan RS, Wilson PW, D’Agostino RB (2007) LDL particle number and risk of future cardiovascular disease in the Framingham Offspring Study—implications for LDL manage- ment. J Clin Lipidol 1:583–592
 Musunuru, K (2010) Atherogenic Dyslipidemia: Cardiovascular Risk and Dietary Intervention. Lipids 45:907–914
 Guay V, Lamarche B, Charest A, Tremblay AJ, Couture P. (2011) Effect of short-term low- and high-fat diets on low-density lipoprotein particle size in normolipidemic subjects. Metabolism [epub ahead of print]