Faulty. Flawed. Messy. Untrustworthy. The science behind most experts’ advice on diet and health is a popular target these days (thanks, in part, to the food industry). One frequent attack: The advice is based on weak “observational” studies that can’t prove cause and effect. Wrong and wrong. Those types of studies are not necessarily weak, and most advice is also based on randomized clinical trials, the gold standard of scientific research. Here are five game changers.

1. The DASH Studies

The first DASH (Dietary Approaches to Stop Hypertension) study was published in 1997.1

And it’s still the bedrock of today’s advice from the American Heart Association, American College of Cardiology, American Cancer Society, and others: Eat a diet rich in fruits and vegetables, include low-fat dairy, poultry, fish, beans, whole grains, oils, and nuts, and limit sweets, sugar-sweetened beverages, and red meats.2

How did DASH happen?

“The study started with observations of populations where there’s very little high blood pressure and very little rise in blood pressure with age,” says co-author Frank Sacks, professor of cardiovascular disease prevention at the Harvard T.H. Chan School of Public Health.

In contrast, roughly one out of two U.S. adults now have high blood pressure, or hypertension. It’s a major risk factor for strokes and heart attacks.

“Also, studies on vegetarians suggested that diet plays a large role in blood pressure,” says Sacks.

“So scientists at NHLBI—the National Heart, Lung, and Blood Institute—decided that we needed a definitive clinical trial to see whether the DASH dietary pattern would lower blood pressure.” (For more on clinical trials, see “What Kind of Study?” below.)

The trial randomly assigned 459 adults—most had higher-than-normal blood pressure—to eat one of three diets: a typical American diet, a typical American diet with extra fruits and vegetables (instead of snacks and sweets), or a DASH diet (a typical American diet with extra fruits, vegetables, and low-fat dairy and less saturated fat and cholesterol).

The researchers prepared all the food for the participants. (That’s part of what made the study “well controlled.”)

“After eight weeks, the fruits and vegetables had lowered blood pressure significantly” compared to the typical American diet, says Sacks, “but only about half as much as the DASH diet lowered pressure.”


“Even now, we don’t fully understand how those changes—switching from high-fat to low-fat dairy, from beef and pork to fish and chicken, and from butter to oils—lowered blood pressure more than fruits and vegetables alone,” says Sacks.

The next study: DASH-Sodium.

Since all three diets in the DASH study had the same amount of sodium, “the NHLBI scientists thought we should see what happens if you reduce sodium in both the DASH diet and the typical American diet,” Sacks explains.

So researchers randomly assigned 412 people to eat one of those two diets with high levels of sodium (3,300 milligrams a day), with intermediate levels (2,400 mg a day), or with low levels (1,500 mg a day).3

“The blood pressure reduction from cutting salt was substantial, especially in older people,” says Sacks. And if you add the drop from switching to a DASH diet, the difference was huge.

“The DASH diet alone can lower blood pressure as much as taking a blood-pressure-lowering drug,” says Sacks. “If you also cut salt, it can be superior to drug treatment.”

But the researchers had yet another question.

“We wondered what would happen if you took away some of the carbs in DASH and replaced them with unsaturated fats or protein,” says Sacks. (More than half of the extra protein would come from plant foods like beans and nuts.)

That led to a third DASH study, known as OmniHeart.4

“Replacing some of the carbohydrate with protein or unsaturated fat lowered blood pressure, cholesterol, and triglycerides more than the original DASH diet,” says Sacks.

(See “DASH in a Nutshell" for a hybrid of the higher-protein and higher-unsaturated-fat diets.)

“But bear in mind that it wasn’t just a diet in which everything was slathered with cheese or oil,” notes Sacks.

“It was the same healthy DASH dietary pattern.”

2. The Sat Fat Trials

How do we know that saturated fats raise—and unsaturated fats lower—LDL (“bad”) cholesterol, a major risk factor for heart disease?

Early observational studies like the Seven Countries Study reported higher rates of heart disease in people who ate more sat fat.

“But what really made an impression were two parallel series of trials testing which fats raised or lowered blood cholesterol,” says Martijn Katan, a cardiovascular disease expert and emeritus professor of nutrition at Vrije Universiteit in Amsterdam. 5,6

“By 1965 at the latest, it was beyond a reasonable doubt that if you replace saturated fats with polyunsaturated fats, you get a substantial lowering of total cholesterol,” says Katan.

Then, in the early 1970s, researchers started to look at LDL cholesterol separately. By 2016, a World Health Organization report had looked at 91 trials.7

“There wasn’t one single experiment,” says Katan. “There was a mass of well-organized experiments that all showed the same thing: If you replace saturated fats with polyunsaturated fats, LDL goes down.”

And is there any reason to think that lowering LDL might not protect the heart?

“Absolutely not,” says Katan. “The effect of LDL on heart disease risk is one of the best established facts in the whole of medical science.”

The data testing statins or other drugs that slash LDL is massive.

“The latest summary included 27 trials involving 174,000 patients,” says Katan.8 “That’s a staggering number.”

But there’s even more evidence that lower LDL means less heart disease.

“There’s a whole bunch of genetic variants that raise or lower LDL, and they all raise or lower coronary risk,” says Katan.

“So the evidence is coming from all directions, and there’s really no way to explain it all, except by assuming that lowering LDL lowers the risk of coronary heart disease.”

And it’s not just the cholesterol trials.

Several randomized clinical trials from the 1950s, ’60s, and ’70s looked not just at LDL but also at heart attacks and strokes.9

“If you look at the four highest-quality trials together, they provide direct evidence that replacing a diet high in saturated fat with a diet high in polyunsaturated fat prevents heart attacks and strokes,” says Sacks.

3. The Diabetes Prevention Program (DPP)

One out of three adults now have prediabetes. Nine out of ten of them don’t know it.

“The good news is that if you have pre-diabetes, the CDC-led National Diabetes Prevention Program can help you make lifestyle changes to prevent or delay type 2 diabetes and other serious health problems,” says the Centers for Disease Control and Prevention.

“Through the program, you can lower your risk of developing type 2 diabetes by as much as 58 percent (71 percent if you’re over age 60).”

Those results come from the Diabetes Prevention Program (DPP), a trial that randomly assigned 3,234 people with prediabetes to take a placebo or metformin (a drug that lowers blood sugar), or to a “lifestyle” group.

The lifestyle goals: lose excess weight and exercise for at least 2½ hours a week.

“The DPP participants counted the grams of fat they ate because it was easier for them to calculate than counting calories,” says DPP researcher Judith Wylie-Rosett, who heads the division of health promotion and nutrition research at the Albert Einstein College of Medicine in New York.

After nearly three years, the average lifestyle participant had lost only 12 pounds, and only 58 percent had met the exercise goal.

Yet that was enough to slash the risk of diabetes by 58 percent, far more than metformin, which trimmed risk by 31 percent.10

“And people in the lifestyle group still had a 27 percent lower risk of diabetes than the placebo group 15 years later, even though many had regained the lost weight,” says Wylie-Rosett.11

Researchers are still tracking the participants’ risk of diabetes, along with their risk of cancer and memory loss.

“It’s truly amazing to me that we’ve gone from doing a trial to having a lifestyle program that’s reimbursed by Medicare,” says Wylie-Rosett.

“How often does that happen?”

4. The DRINK Study

Do sugary drinks make people gain weight? The DRINK trial answered that question in 2012.

Researchers randomly assigned 641 mostly normal-weight Dutch children who usually drank sugary beverages to get a daily 8 oz. drink sweetened with either sugar (104 calories) or artificial sweeteners (0 calories) at school each day. Neither the children nor the investigators knew who got which drinks.12

“The question was whether the children who got zero calories would sense the difference and compensate by eating more calories from some other source,” says Katan. “Would those kids come home and say, ‘Mom, I’m hungry. I want a snack,’ or would they just have their usual lunch and dinner?”

The result: After 1½ years, the average child who got the sugary drinks had gained roughly two pounds more than the average child who got the sugar-free drinks. And the chubbier kids gained about three pounds more if they got their usual sugary drinks.13

“Even the thinnest children were not immune to the effect of sugary drinks,” says Katan. “But heavier kids may be less likely to sense when they’re eating fewer calories and when they’re overeating.”

More evidence that some people are more vulnerable than others: an observational study published on the same day as the DRINK trial that tracked roughly 11,000 nurses and health professionals for 12 to 18 years.

Among men and women with a genetic risk of obesity, those who drank at least one sugary drink a day were four times more likely to become obese than those who drank less than one sugary drink a month.14

“Sugary drinks make you fat. They circumvent your innate mechanisms for keeping your body weight stable,” says Katan.

“If you eat beans or whole wheat bread, they’re not fattening because you feel full, so you don’t eat other things.”

In contrast, “if you drink sugary drinks, including fruit juices—there’s no difference between fruit juices and soda—they will simply slip in through the back door and you’ll just keep eating as much as you usually do.

“They’re burglars who will rob you of your leanness because you don’t notice them coming in.”

5. Pounds Lost

When you’re trying to lose weight, what matters more: cutting fat, cutting protein, or cutting carbs? Pounds Lost—which began in 2004—wasn’t the first or the last study to explore that question. But it was one of the largest.15

It's not just soda but also sports drinks, lemonade, iced tea, and juices that can lead to weight gain.

“We enrolled over 800 people so we could compare the effect of diets higher or lower in carbohydrate, fat, or protein,” says Harvard’s Frank Sacks.

Pounds Lost was also one of the longest diet trials.

“We did a two-year study, which was quite unusual at the time, because body weight reaches a low point after about six months in a weight-loss program and then you have weight regain,” says Sacks.

Participants got daily meal plans and went to individual and group sessions where they learned what to eat. Each of the diets cut about 750 calories a day.

The healthy low-fat chicken dinner (left) has whole grains and beans. The healthy low-carb dinner uses those calories for extra fat in the salmon, nuts, and avocado.

“We came up with a very simple but somewhat unexpected result,” says Sacks. “The composition of fat, carbohydrate, and protein in a healthy diet had no effect whatsoever on weight loss or regain.” Nor did it affect hunger.

Most other studies—they typically tested low-carb versus low-fat diets—have had similar results.

The most recent: DIETFITS randomly assigned 609 people to cut as much fat or carbs as possible from their diets, rather than count calories.

After one year, each group had lost roughly the same weight—about 12 pounds.16 And that’s no surprise, since both groups ended up cutting a similar number of calories—about 500 to 600 a day, on average. (See “If the Diet Fits,” May 2018.)

“We expected people with insulin resistance to do better on a low-carb diet, but they didn’t,” says lead investigator Christopher Gardner, professor of medicine at the Stanford University School of Medicine. (People with insulin resistance have an increased risk of diabetes.)

Maybe that’s because both diets were healthy.

“Nobody was supposed to eat added sugars or refined grains, and everybody was supposed to eat vegetables,” says Gardner. “Americans get a quarter of their calories from added sugars and refined grains and a woefully low number of calories from vegetables...if you don’t count potatoes.

“If we could just get people to make those changes,” says Gardner, “we’d be well on our way.”


1N. Engl. J. Med. 336: 1117, 1997.
2J. Am. Coll. Cardiol. 63: 2960, 2014.
3N. Engl. J. Med. 344: 3, 2001.
4JAMA 294: 2455, 2005.
5Metabolism 14:747, 759, 766, 776, 1965.
6Am. J. Clin. Nutr. 17: 281, 1965.
7 who.int/nutrition/publications/nutrientrequirements/sfa_systematic_review/en.
8Lancet 385: 1397, 2015.
9Circulation 136: e1, 2017.
10N. Engl. J. Med. 346: 393, 2002.
11Lancet Diabetes Endocrinol. 3: 866, 2015.
12N. Engl. J. Med. 367: 1397, 2012.
13PLoS One 11: e0159771, 2016.
14N. Engl. J. Med. 367: 1387, 2012.
15N. Engl. J. Med. 360: 859, 2009.
16JAMA 319: 667, 2018.

Photos (top to bottom): freshidea/stock.adobe.com, Jennifer Urban/CSPI (GOOD FATS), Kate Sherwood/CSPI (BAD FATS), kazoka303030/stock.adobe.com, nblxer/stock.adobe.com (healthy low-fat chicken dinner), topotishika/stock.adobe.com (healthy low-carb dinner), (DASH Diet, top to bottom) fotolia.com: ©EWA BROZEK, ©Stefanie Leuker, ©angelo.gi, ©Krysiek z Poczty, ©cultureworx, ©Paylessimages, ©pockygallery11.

Illustrations: Loel Barr.