Broken Science Journalism: Eating Less Protein Could Help You Live Longer
On July 31, 2026, Scientific American ran a story by Adam Kovac under this headline:
Eating less protein could help you live longer
The subheading promised that cutting protein "could add years to your life."
It was brought to my attention by my mom living in India, who forwarded it to me from a local WhatsApp group.
The science behind the story is real. A lab at the University of Wisconsin-Madison, led by Dudley Lamming, has spent years feeding mice less protein. Those mice tend to stay leaner, handle sugar better, and often live longer. The story was built on two new papers from the lab: a review of more than 350 studies, and a new mouse experiment. Within days, Yahoo, TIME, ScienceDaily and many health sites carried their own versions.
The problem is what happened between the research and the headline. The publisher of the review, Cell Press, sent out a press release the same day. It carried this warning:
"...some people, like pregnant women and some older adults, have higher protein needs..."
Scientific American kept the pregnant women. It dropped the older adults.
While the public and policy makers are rightfully concerned about conspiracies spreading through WhatsApp and social media, the biggest health threat may actually be reputable science journalists publishing sensationalized headlines based on animal studies without proper context or nuance. Rather than outright misinformation, it is this type of half-truth with a scientific veil from a well-respected publication that has a large societal impact.
Here is what the story stretched, what it left out, who would be worse off for believing it, and how it was built so that no single sentence is to blame.
1. The evidence comes from mice
The headline talks to you. The evidence comes from mice.
In the new experiment, mice ate less valine, one of the building blocks that protein is made of. Male mice lived about 23 percent longer. Female mice were healthier but did not live any longer (Calubag et al. 2026). No study has shown that eating less protein makes people live longer. The human studies mentioned in the press release measured weight, body fat and blood sugar. None measured lifespan.
The story quotes Lamming saying mice are good models because they "basically eat human food." Mice and people do get a similar share of their calories from protein. That still does not make a mouse's lifespan a guide to yours:
- Mice live fast. A lab mouse lives about two and a half years, and its body burns through energy far faster than ours.
- Mice die of different things. Lab mice of this breed die mostly of cancers and blood disorders (Turturro et al. 2002). A diet that slows cancer in them can add a lot of time. The valine mice did get less cancer. People die of a much wider mix of causes.
- Diet tricks shrink as animals get closer to us. Eating fewer calories reliably lengthens the lives of rats and mice. In monkeys, two long studies disagreed on whether it helps at all (Mattison et al. 2017).
"Older than 120"
The story says cutting protein has stretched animal lifespans by "as high as 53 percent." It then asks readers to imagine an 80-year-old "living to be older than age 120." It never says which study, which animal or which sex the 53 percent came from. The 120 is simply 80 multiplied by 1.53.
To see how far that is from anything real, I built a simple model of how people die. It uses a pattern first described almost 200 years ago: once you are an adult, your chance of dying in any given year roughly doubles every eight years. Then I asked what it would take to push the average lifespan from 80 to 122.
The answer: your risk of dying would have to be cut to about one-fortieth, at every age, from your twenties onward. No diet, drug or medical advance has come anywhere close.
For comparison, I also fed in the strongest links between protein and death found in a large study of real people (Section 3). They move average lifespan by about two years, one way or the other.
2. What "5 percent protein" means on your plate
The story gives its numbers as percentages. People eat about 17 percent of their calories as protein. The ordinary lab mice ate about 20 percent. The longest-lived mice ate 5 to 8 percent. Those percentages are hard to picture, so here they are as food.
A gram of protein carries 4 calories. For someone eating 2,000 calories a day:
- 5 percent protein is 25 grams a day.
- 8 percent protein is 40 grams a day.
That is the total for the whole day, counting everything: the rice, the roti, the vegetables, the milk in your tea. Here is what 25 to 40 grams looks like (values from the USDA FoodData Central database):
| Food | Portion | Protein |
|---|---|---|
| Chicken breast, raw | 170 g | about 38 g |
| Canned tuna, drained | 1 can | about 29 g |
| Eggs | 3 large | about 19 g |
| Greek yogurt, plain | 200 g | about 18 to 20 g |
| Dal, cooked | 1 bowl | about 6 to 9 g |
| White rice, cooked | 1 cup | about 4 g |
Now compare that with what people need. US health authorities say a 70 kg (154 lb) adult needs about 46 grams a day on average, and recommend 56 grams so that almost everyone is covered (IOM 2005). India recommends 54 grams for a 65 kg man and 46 grams for a 55 kg woman (ICMR-NIN 2020). Turned into a human day, the mouse diets fall below even the average need.
How protein is spread over the day matters too, and more so with age. One analysis pooled muscle experiments in men. A young man's muscles stopped responding at about 17 grams of protein in a meal. A man around 71 needed about 28 grams to get the same effect (Moore et al. 2015). Spread the mouse diets over three meals and each meal gets 8 to 13 grams.
3. What human studies say
Age changes the answer
The best-known study of protein and death in people followed 6,381 Americans aged 50 and older for 18 years (Levine et al. 2014). The story does not mention it. Its results depended on age:
- Ages 50 to 65: people in the study who ate the most protein (20 percent of calories or more) had a 74 percent higher death rate than people who ate the least (under 10 percent). Their cancer death rate was about four times higher.
- Ages 66 and over: the pattern flipped. The high-protein group had a 28 percent lower death rate, and less than half the cancer death rate.
To be fair, this partly supports the story. For people in their fifties and early sixties, the study leans toward eating less protein. After 65 it leans the other way. The researchers themselves suggested less protein in middle age and moderate to high protein in old age.
Does that hold for everyone aged 50 to 65?
No. The 74 percent is an average across one group of Americans, not a rule for each person in their fifties. Several things in the study itself narrow what it can tell you.
- It shows a link, not a cause. People who ate a lot of protein may have differed in other ways. The researchers adjusted for smoking, waist size, existing disease and more, but no study can adjust for everything.
- The risk seemed to come from animal protein, but that is hard to pin down. When the researchers took into account how much protein came from meat, eggs and dairy, the extra risk for 50 to 65 year olds shrank to 18 percent and was no longer statistically clear. Taking plant protein into account changed nothing. The authors read this as a sign that animal protein drove the risk, and said it does not show that plant protein protects. Separating the two is tricky, because in these American diets most protein was animal protein: 11 of the 16 percent of calories. The study never looked at people who ate mostly plant protein, so it says little about a diet built on dal, beans and grains.
- It depended on the person. In a subgroup of 2,253 people, the researchers measured IGF-1, a growth hormone that rises with protein intake. The extra cancer risk from high protein was larger in people with more IGF-1. The same diet did not carry the same risk for everyone.
- The comparison group was small. Only 437 people ate low protein, split across both age groups. Each age comparison rests on a few hundred people.
- Diet was measured once. Each person described what they ate on a single day, and was then followed for up to 18 years. Most said it was a normal day, but diets change over 18 years.
- 65 is not a cliff. The researchers found that the apparent benefit of low protein faded with age. The split at 65 was a way to analyze the data, not a birthday on which your body changes.
- Some groups are missing from the picture. The results were not broken down for people who exercise, people with little muscle, or people who already eat too little. And the authors noted that the low-protein group may include people who were malnourished, frail or already falling ill, since sick people often eat less. That can make low protein look worse than it is, especially in older people.
Other large studies also point to who matters more than how much. A study of 131,342 American nurses and health professionals, followed for up to 32 years, found that total animal protein was not linked to overall deaths once lifestyle was taken into account (it did go with slightly more heart deaths), while more plant protein went with fewer deaths (Song et al. 2016). Even those links showed up only among people who smoked, drank heavily, were overweight or did not exercise. Among people with none of those habits, protein made no measurable difference. In that study, replacing protein from processed red meat with plant protein went with the biggest drop in deaths, a hint that the food around the protein may matter as much as the protein itself.
The widest view comes from a review that pooled 31 such studies, covering about 715,000 people (Naghshi et al. 2020). People who ate more protein overall were slightly less likely to die during follow-up, about 6 percent, and those who ate more plant protein about 8 percent less likely. Animal protein was not linked to deaths from heart disease or cancer. That runs opposite to the headline. It also comes with the usual caution: all of these studies watch what people choose to eat rather than test diets, and people who eat more beans and fewer burgers tend to live differently in other ways too.
So what does this mean for a 55-year-old? At most, an inactive middle-aged person eating a lot of animal protein might do better eating less of it, or swapping some meat, especially processed meat, for plant protein. That is a long way from "eat less protein, live longer." It gives no reason to eat below what the body needs, and it says nothing to someone who is active or already eats little. Not one of these limits appears in the Scientific American story, because the study itself does not appear.
What experts tell older adults
The story mentions frailty, but only in mice. It never mentions that people lose muscle as they age, or what doctors who specialize in older patients recommend.
Two expert groups on aging recommend that healthy people over 65 eat at least 1.0 to 1.2 grams of protein for every kilogram they weigh (Bauer et al. 2013; Deutz et al. 2014). For a 70 kg older adult that is 70 to 84 grams a day, about double the high end of the mouse diets. People who are ill need more.
This is the warning Cell Press put in its press release. Scientific American did not pass it on.
A guideline set against a mouse review
The story also pits the lab's work against the new US dietary guidelines, released in January 2026. They raised the protein target to 1.2 to 1.6 grams per kilogram a day, up from 0.8. The story reports that doctors raised conflicts of interest among some of the guideline authors, which is fair. Then it calls the higher target "at odds with" the benefits of eating less protein.
That treats a review of mostly animal studies and advice written for people as the same kind of evidence. They are not. The new target is debatable. It is higher than what most expert bodies had advised for inactive younger adults. But there is human research behind eating more protein for older adults, people who exercise and people trying to lose weight (details in the Notes). A careful story would have said where human evidence supports more protein, where it hints at less (middle age), and where it is thin. Readers got a controversial official on one side and mice on the other.
4. Who is at risk if they follow it
I found no evidence that anyone changed their diet because of this story, and I am not claiming harm. The question is narrower: which readers would be worse off if they took the headline at face value?
Start with who the research fits. The story was aimed at "protein maxxing," the American craze for protein water, protein coffee and protein chips. The people it could help are inactive adults already eating far more protein than they need, mostly in rich countries. For them, "you may not need all that extra protein" is a fair message.
But a headline cannot choose its readers. For almost everyone outside that group, the advice points the wrong way. Three groups stand out.
Whole populations that already eat close to the minimum
This is the largest group, and the one the story never considers. India is the case I know, and Section 5 goes through the numbers. In short: the average Indian eats only a little more protein than the body needs, much of it comes from rice and wheat, which the body uses less well, and the poorest households eat far less than the average. By one estimate, about a third of inactive rural adults may already fall short.
In a population like that, "eat less protein" is not advice for a few people who overdo it. It pushes the typical reader toward too little. The same is likely true wherever meals are built mostly on grains or starchy roots, and wherever money limits what goes on the plate. And a family cooks one meal for everyone, so a change in the kitchen reaches the grandparents too.
Older adults, everywhere
Older muscle responds less to protein, so it takes more to keep it. Lose enough muscle and walking, balance and getting up from a chair all get harder. Falls follow, and falls in old age are dangerous. In one English study of people over 60 who broke a hip, a third had died within a year (Roche et al. 2005).
A 72-year-old who reads "eating less protein could help you live longer" and cuts back is moving away from what doctors for older people advise. The story gives no hint of that. An older adult in India faces both problems at once: an aging body that needs more, and a diet that starts with less.
Anyone whose needs are raised for a while
The story does say that needs rise with pregnancy, injury and heavy exercise. It does not mention people trying to lose weight, who can lose muscle along with fat if protein drops, or people recovering from a serious illness, who often eat less than they need. A reader in any of these situations who takes the headline as general advice is cutting back at exactly the wrong time.
5. The view from India
This story did not stay in America. It reached a WhatsApp group in India, which is how it reached me. An India-based news site republished the press release the next day (Shafaqna India, 2026). So what does "eat less protein" mean for an Indian reader?
How much protein Indians eat
India's latest national survey, covering more than 2.6 lakh households in 2023-24, found (MoSPI 2025):
- About 62 grams of protein per person per day in villages and 63 grams in cities.
- That is about 11 percent of calories from protein.
- Rice, wheat and other grains supply about 46 percent of that protein in villages and 39 percent in cities. That share is falling as more eggs, fish, meat and milk reach the plate.
Averages hide a lot. The poorest 5 percent of rural Indians eat about 1,690 calories a day. The richest 5 percent eat about 2,940. The survey does not report protein by income, but the poorest households very likely eat well below the average.
These figures come from what households buy or grow, divided among family members. They do not track what each person actually ate. An older survey that asked people directly found rural adults eating about 49 grams a day in 2006 (Swaminathan, Vaz and Kurpad 2012).
So the average Indian is not already eating the mouse diet. At about 11 percent of calories, Indians eat more protein than the mice did (5 to 8 percent) and less than Americans (15 to 17 percent). But 11 percent sits right at the edge of the "low protein" group in the US study above (under 10 percent). In that study, after 65, it was the group with the worst survival. The same limits apply: that group was small and American, so treat this as a warning sign, not a verdict.
You will often see the claim that "73 percent of Indians are protein deficient." It comes from a 2017 market-research survey of about 1,800 people in cities (Business Standard, 2017). It is not a medical measurement and should not be quoted as one.
Not all protein is equal
Two things decide how much of that protein the body can actually use.
The first is absorption. Plant foods carry fibre and natural compounds that get in the way of digestion, so less of their protein makes it from the gut into the body. Two studies in healthy Indian adults measured how much of the essential protein building blocks the body absorbed from different foods (Kashyap et al. 2018; Kashyap et al. 2019):
| Food | Share absorbed |
|---|---|
| Meat | about 92% |
| Whole egg | about 89% |
| Chickpeas | about 75% |
| Yellow peas | about 72% |
| Mung beans, husk removed (split moong dal) | about 71% |
| Mung beans, whole | about 63% |
Taking the husk off, as Indian kitchens have long done, makes a real difference.
The second is the mix. Protein is made of building blocks, and the body needs them in a fixed mix, a bit like a recipe that needs every ingredient. Rice and wheat are short of one building block, called lysine. When one ingredient runs short, the body can only use part of the rest. Dal, milk and eggs are rich in lysine, which is one reason dal and rice work so well together, even though dal itself is not absorbed as well as eggs or meat.
Researchers who adjusted India's diet surveys for both found that the protein people could actually use was lower than the grams suggested. By their estimate, about a third of inactive adults in villages may not have been getting enough, even though the grams looked adequate on average.
India's own guidelines reflect this. They recommend 0.83 grams of protein per kilogram of body weight, rising to about 1 gram per kilogram for people whose protein comes mostly from grains such as rice and wheat (ICMR-NIN 2020).
Indian bodies raise the stakes
At the same height and weight, Indians tend to carry more fat and less muscle than white Europeans. Two doctors, one Indian and one British, showed this with a photo of themselves. They had the same body mass index, about 22, but 21 percent and 9 percent body fat (Yajnik and Yudkin 2004). Studies of Indian men living in the US found the same pattern, along with more insulin resistance, an early step toward diabetes (Banerji et al. 1999; Chandalia et al. 1999). That is why Indian guidelines use lower cut-offs for being overweight (Misra et al. 2011).
Muscle is where the body puts much of the sugar from a meal. For an older Indian with little muscle and a higher risk of diabetes, advice that costs muscle comes at a real price.
What an Indian reader should take from this
The research behind the headline is about inactive people eating more protein than they need, mostly in rich countries. For most Indian adults, especially older people and vegetarians whose meals are mostly rice, roti and other grains, India's own guidance is the better guide. For a 60 kg adult, that means about 50 to 60 grams of good-quality protein a day. International advice for people over 65 is a little higher, about 60 to 72 grams at that weight. Dal, milk, curd, paneer and eggs all help. So does exercise, which helps the body put that protein to use.
6. Written for the click
Read the headline again: eat less protein, live longer, add years to your life. That is a promise, and promises get clicked. The careful version, that mouse diets may say something about people who eat too much protein and sit too much, would never have made it to a WhatsApp group in India.
News websites earn money from visitors and track every click. Many test several headlines on the same story and keep whichever one gets clicked most. One study of these tests at a single viral-news site looked at about 105,000 headline versions shown 370 million times (Robertson et al. 2023). I cannot see inside the Scientific American newsroom, and I do not know whether this headline was tested. But a headline built this way does one job well, and informing the reader is not that job.
What makes this case stand out is where the exaggeration started. It usually starts in the press release. One study found that when a press release overstated what animal research meant for people, 86 percent of the news stories that followed did the same. When the release did not, only 10 percent did (Sumner et al. 2014). Here the press release was the careful one. The magazine removed the care on its own.
It did not have to. The same researchers found that warnings in press releases did not reduce news coverage, and exaggeration did not increase it (Sumner et al. 2016). Accuracy would not have cost the story its place in the news. It would only have cost it the most clickable headline.
Built to be defensible
Now imagine someone complains. The magazine has an answer for every line.
- The headline says protein "could" help you live longer. It never says it will.
- The mice are in the story. They just do not appear until the fourth paragraph, after the headline and subhead have done their work.
- The exercise caveat is in the story too, in the last paragraph.
- Much of the argument comes from the lead researcher, in quotes. The magazine is only reporting what an expert said.
Taken one at a time, each of these holds up. Hardly anything in the story is flatly false, and nowhere does it tell anyone to eat less protein. That is the point. This is plausible deniability: a story built so that the misleading part lives in the overall impression, where no single sentence has to answer for it.
It protects the publisher. It does nothing for the reader. Most people read the headline and move on. A WhatsApp forward carries a headline and a link preview, not the fourth paragraph and certainly not the last one. The impression that travels is "eat less protein, live longer," and it is the impression that changes what people put on their plates. A fact-checker grading sentence by sentence would pass this story. A reader should not.
This is also why stories like this are harder to fight than outright misinformation. A false claim can be corrected. A half-truth with every caveat technically present leaves nothing to correct. The magazine can always say the full picture was there for anyone who read to the end.
The rest of the story follows the same pattern.
Few voices. The story quotes two researchers. Lamming, whose lab did the mouse work, carries most of it. The other is Steven Heymsfield of Louisiana State University, who studies body composition and argues that protein beyond what the body needs is simply burned or stored. Nobody quoted treats older patients. A doctor who does would likely have raised the age question within minutes, and complicated the headline.
Selective about conflicts. The story raises conflicts of interest for the guideline authors. It does not mention that Lamming advises Aeovian Pharmaceuticals, a company developing drugs aimed at mTOR, a growth switch inside cells that cutting protein also turns down. Both papers disclose this (Knopf and Lamming 2026; Calubag et al. 2026). Ties like this are common in science and say nothing about whether the work is right. The point is fairness: the same standard for both sides.
The caveat at the bottom. The most useful lines come last. Lamming says that for people who exercise, protein is "probably beneficial and certainly not harmful," and that inactive people eating too much protein are probably hurting themselves. That is a sensible claim, and a different one from the headline. If protein helps active people and the risk lies with inactive people eating too much, the real story is about exercise and excess. The reader who skims gets the headline. Only the reader who finishes gets the real story, and many never get there.
7. Five questions to ask of any diet headline
- Was it tested in people? Results in mice or cells are a reason to study something further. They are not advice for your plate.
- How much is that in food? Turn percentages into grams, and grams into meals. Compare with what your country recommends.
- For what age? Protein can work differently at 50 and at 75. If a story about living longer never mentions age, ask why.
- Who was it written for? Advice for people who eat too much can be wrong for people who eat too little. Check where you start.
- What did the original say? If the press release carried a warning and the story does not, ask who gained from dropping it.
What good coverage would have looked like
The mouse research deserved coverage. A story that served its readers would have given the amounts in grams, said plainly that the lifespan results come from mice (and in the valine study, only male mice), reported that human results depend on age, quoted someone who treats older patients, kept the press release's line about older adults, and put the exercise point at the top.
That story would have been less dramatic. It would also have told a 72-year-old reader in Pune, or in Pittsburgh, something true about what to put on the plate.
Notes for readers who want the details
The lifespan model (survival chart). A simple mortality model: Gompertz's law of mortality (Gompertz 1825) plus a small age-independent background risk of 0.0005 a year. The yearly death risk doubles every 8 years, the human doubling time reported by Finch et al. 1990, and the model is tuned to a life expectancy of 80. Reaching 122 requires multiplying the death risk by about 0.026 (a 39-fold cut) at every age from 20. The Levine et al. 2014 hazard ratios, applied to the ages they were measured in, give 82.5 years (0.72 from age 66) and 78.2 years (1.74 at ages 50 to 65).
Levine et al. 2014 in full (death-rate chart). NHANES III, one 24-hour diet recall, 18-year follow-up. Hazard ratios versus low protein, Table 1, Model 1, with 95% confidence intervals. Ages 50 to 65: all-cause 1.34 (0.81 to 2.22) for moderate and 1.74 (1.02 to 2.97) for high; cancer 3.06 (1.49 to 6.25) and 4.33 (1.96 to 9.56). Ages 66 and over: all-cause 0.79 (0.62 to 0.99) and 0.72 (0.55 to 0.94); cancer 0.67 (0.43 to 1.06) and 0.40 (0.23 to 0.71). The death-rate chart shows the all-cause results as (hazard ratio minus 1) × 100. After adjusting for animal protein (Model 4), the high-protein all-cause hazard ratios became 1.18 (0.60 to 2.31) at ages 50 to 65 and 0.72 (0.50 to 1.02) at 66 and over, neither statistically significant. Adjusting for plant protein left the results unchanged (Table S5); the moderate-protein cancer hazard ratio at 50 to 65 fell only from 3.06 to 2.71 (1.24 to 5.91) after adjusting for animal protein. The plant-versus-animal conclusion rests on these statistical adjustments, not on a separate group of plant-protein eaters. The low-protein reference group was 437 people across both age groups. In the IGF-1 subsample (n = 2,253), each 10 ng/ml of IGF-1 raised the high-versus-low protein cancer hazard at ages 50 to 65 by a further 9 percent (HR 1.09, 1.01 to 1.17). High protein also went with more diabetes deaths, with very wide intervals.
Protein needs. The US figures (IOM 2005) are the Estimated Average Requirement (0.66 g/kg, enough for half of healthy adults) and the Recommended Dietary Allowance (0.8 g/kg). Studies using a tracer method put the average adult requirement at 0.93 g/kg and a safe intake at 1.2 g/kg (Elango et al. 2010); the method has critics (Millward 2025). Phillips and colleagues argued for 1.2 to 1.6 g/kg for adults (Phillips et al. 2016). Sports nutrition bodies recommend 1.2 to 2.0 g/kg for athletes (Thomas et al. 2016). PROT-AGE recommends 1.0 to 1.2 g/kg for healthy people over 65, 1.2 g/kg or more for those who exercise, and 1.2 to 1.5 g/kg for those with acute or chronic illness; ESPEN's numbers are nearly identical. The US guideline's protein review leaned on weight-loss trials.
Plant and animal protein in other studies. Song et al. 2016: animal protein HR 1.02 (0.98 to 1.05) per 10 percent of energy for all-cause mortality and 1.08 (1.01 to 1.16) for cardiovascular mortality; plant protein HR 0.90 (0.86 to 0.95) per 3 percent of energy. Replacing 3 percent of energy from processed red meat protein with plant protein: HR 0.66 (0.59 to 0.75). Naghshi et al. 2020, highest versus lowest intake across 31 cohorts: total protein 0.94 (0.89 to 0.99) and plant protein 0.92 (0.87 to 0.97) for all-cause mortality; plant protein 0.88 (0.80 to 0.96) for cardiovascular mortality.
Per-meal amounts. Moore et al. 2015 found that muscle building plateaued at about 0.24 g/kg per meal in men around 22 and 0.40 g/kg in men around 71, which is 17 and 28 grams for a 70 kg man.
Indian survey figures. HCES 2023-24: 61.8 g per person per day rural and 63.4 g urban; 66.6 g rural per consumer unit (an adult-male equivalent), which is the India survey bar in the protein chart. The 11 percent is 61.8 g × 4 calories ÷ 2,212 calories. The 2006 NNMB survey used 24-hour recall; rural adults ate about 1 g/kg at typical body weights of 47 to 54 kg, and about 0.8 g/kg once protein quality was accounted for (Swaminathan et al. 2012).
Mouse details. The valine study used C57BL/6J mice. The cancer and blood-disorder finding comes from a colony of more than 4,500 C57BL/6 mice (Turturro et al. 2002). Per gram of body weight, a mouse burns energy about 7.5 times faster than a person. This follows from Kleiber's rule that energy use scales with body mass to the power of three-quarters (Kleiber 1947): (80,000 g ÷ 25 g)^0.25 ≈ 7.5.
Body composition. Non-obese Indian men in New York averaged 33 percent body fat at a BMI of 24.5, and two-thirds were insulin resistant (Banerji et al. 1999). Indian men in Dallas cleared sugar from the blood more slowly than white men with the same body fat (Chandalia et al. 1999).

Discussion (0)