Number Research
← Back to Explore
Health Sciences Nutrition and Dietetics Diet and Metabolic Markers

Two Machines: Why Diet Fixes Cholesterol and Blood Sugar Differently

A retracted-then-vindicated trial and a buried review finding reveal that LDL answers to food composition while glucose answers to body weight.

S.J. Nam 7 min read
Two Machines: Why Diet Fixes Cholesterol and Blood Sugar Differently

In 2003, a research team at the University of Toronto ran a quiet experiment that should have made bigger headlines than it did. They took people with high cholesterol and fed one group a carefully engineered plant-based diet — almonds, oats, barley, eggplant, soy protein, margarine spiked with plant sterols. After a single month, that group's LDL cholesterol had dropped by roughly 29 percent. A comparison group taking a statin dropped about 31 percent. Food, in other words, had come within a rounding error of a pharmaceutical.

That result — David Jenkins's "portfolio diet" — is the single most useful fact in the entire diet-and-numbers debate, and also the most misleading. Useful, because it proves food can move a hard clinical marker as hard as a drug. Misleading, because it tempts us to believe that whatever a "healthy diet" does for cholesterol, it must do the same way for blood sugar. It doesn't. The evidence points to something a generic explainer almost never says out loud: your cholesterol and your blood sugar respond to food through two entirely different machines. One is driven by what you eat. The other, mostly, by how much of you there is.

The Cholesterol You Can Eat Your Way Out Of

The portfolio diet is not a lifestyle. It is a formula — a portfolio of four cholesterol-lowering foods stacked on top of each other, each attacking LDL through a distinct route. Viscous soluble fiber from oats and barley traps bile acids in the gut so the liver has to pull cholesterol out of the blood to make more. Plant sterols compete with cholesterol for absorption. Soy protein and tree nuts displace the saturated fat that would otherwise push LDL up. Assembled together, the effects are additive rather than redundant.

Across the six controlled studies Jenkins later pooled, the diet lowered LDL cholesterol by about 17 percent, apolipoprotein B by 15 percent, total cholesterol by 12 percent, and triglycerides by 16 percent — and in the observational follow-ups, people who ate this way had lower rates of heart disease, type 2 diabetes, and early death. The tighter, four-week trials produced the flashier numbers, up to a 35 percent LDL drop, precisely because they were controlled feeding studies where nothing was left to willpower.

Notice what's doing the work here. It isn't calorie cutting. It isn't weight loss. It's composition — specific molecules displacing other specific molecules. You could eat the portfolio diet and lose no weight at all, and your LDL would still fall, because the mechanism runs through the chemistry of the food itself.

The Mediterranean pattern tells a compatible story. In the Spanish PREDIMED trial, participants at high cardiovascular risk who were assigned a Mediterranean diet enriched with extra-virgin olive oil or mixed nuts saw major cardiovascular events — heart attack, stroke, cardiovascular death — fall by roughly 30 and 28 percent respectively versus a low-fat control. Researchers have pointed out that this 30 percent reduction sits in the same range as the benefit seen in statin trials, achieved at no cost to the health system. And here disagreement in the literature is worth naming rather than hiding: PREDIMED was retracted in 2018 over irregularities in how some participants were randomized. When the investigators reanalyzed and republished, the headline number barely moved — the roughly 30 percent risk reduction survived the scrutiny. That the result held up after a retraction is, if anything, more persuasive than a clean first pass would have been.

The Blood Sugar Plot Twist

Now watch the machine change. For type 2 diabetes, the loudest fight of the last decade has been low-carb versus low-fat — a near-religious war over macronutrient ratios. Cut the carbohydrates, one camp insists, and you stop spiking the glucose your failing insulin can't handle. There is real evidence for it: systematic reviews find that low-carbohydrate diets, defined as under 26 percent of calories from carbohydrate, can drive meaningful reductions in HbA1c and even push some patients into remission, particularly around the six-month mark.

But then comes the finding most coverage skips, buried in the fine print of the reviews. When researchers compared trials that restricted calories against trials that let people eat low-carb food freely, the calorie-restricted approach was not superior at twelve and twenty-four months. What actually predicted the drop in blood sugar, across studies, was the amount of weight people lost — and the association held at six, twelve, and twenty-four months. One review put it plainly: sustained weight loss is the key to remission, and the weight-independent effect of merely cutting carbs remains unproven.

Read that carefully, because it inverts the popular narrative. The low-carb diet often works — but there's a strong case that it works largely as an unusually effective weight-loss tool, not because carbohydrates are uniquely toxic to a diabetic pancreas. People eat fewer calories on low-carb diets because protein and fat are filling and because eliminating a whole food category simplifies choices. The pounds come off, the fat leaves the liver and pancreas, insulin sensitivity returns. The macronutrient ratio is the delivery truck. Weight loss is the cargo.

Why the Advice Feels Contradictory

Put the two machines side by side and the confusion that haunts every kitchen-table diet argument suddenly makes sense. Cholesterol responds to composition: swap saturated fat for viscous fiber, sterols, nuts, and olive oil, and LDL falls whether or not the scale moves. Blood sugar, in the diabetic range, responds mostly to energy balance and body weight: the specific foods matter less than whether the intervention produces durable weight loss.

This is why a person can do everything "right" for one number and stall on the other. A low-carb dieter loading up on butter, cheese, and fatty red meat may watch their blood sugar normalize as they slim down — while their LDL climbs, because they've inadvertently maximized saturated fat, the exact lever the portfolio diet was built to pull the other way. Meanwhile the careful portfolio eater whose LDL has plunged may see barely any glucose benefit if their weight holds steady. Two virtuous diets, two different scoreboards, and neither is a lie.

The uncomfortable implication is that "eat healthy and your numbers will improve" is not one instruction but two, and they can quietly pull against each other. The good news hiding underneath is that the overlap is large. The Mediterranean pattern lowers cardiovascular risk and tends to produce modest, sustainable weight loss; the portfolio diet's observational arm showed less diabetes, not just less heart disease. The diets that win on both fronts are the ones built around plants, fiber, and unsaturated fat, eaten in a quantity that keeps weight in check — food chosen for what it's made of, in an amount that respects what it does to the body.

So the next time someone waves a single study at you and declares the one true diet, ask which number they're optimizing. Cholesterol and blood sugar are not reading from the same script, and the diet that quiets both is less a discovery than a negotiation between two machines that answer to different masters.

References