This is another special guest post from our favorite study-dismantler, Denise Minger . Read all of her previous Mark’s Daily Apple articles here , here , here , here and here , and pay her website a visit. Thanks, Denise, for clearing up the confusion once again!
Sweden is a land of many wonders – most of which put the USA to shame. They’ve got fjords, ABBA, and caviar in a tube. And while Americans get arrested for things like DUIs and stealing socks from Walmart , Swedes get arrested for the more admirable feat of smuggling butter .
Such a delicious felony doesn’t reflect a life of crime so much as a life of fat – the edible kind. For nearly a decade, Sweden has been the unofficial headquarters of the Low Carb High Fat (LCHF) movement, churning out an unprecedented number of lipid lovers. In 2011, a whopping 25% of Sweden’s population was trying to eat more fat and curb their carbohydrate intake, with 5% of Swedes identifying as hardcore LCHF adherents. And those numbers only seem to be growing.
But a new study claims to cast doubt on the safety of Sweden’s fatward trend. In a paper published in Nutrition Journal last week ( PDF available here ), researchers linked the Swedish low-carb boom to rising cholesterol levels, increased heart disease risk, and failure to maintain long-term weight loss. A cascade of predictable headlines ensued, ranging from “ Atkins diet found to be bad for the heart ” to the brazen “ Time to retire the low-carb diet fad .” The latter article – a well-circulated piece from The Atlantic – gave low-carb fans a particularly snarky flogging:
Low carbohydrate evangelists will almost certainly attack today’s announcement – and perhaps this post – with biblical fury. They’ll make their usual claim: that this is yet another conspiracy of scientists who just don’t get it, scientists who don’t understand nutrition, scientists who somehow made it through their PhD’s and MD’s without knowing the first thing about how the human body works. But let’s face it – most of us know in our hearts that eschewing a breakfast of whole grains and fruit crowned with a dab of yogurt for a greasy pile of sausage, bacon, and eggs is not the road to health.
Ouch! Time to whip out the granola and Yoplait?
If something seems fishy about this study, it’s not just the lutfisk: as usual, there’s more going on (or in this case, less going on) than the alarmist media would suggest. Here’s the real scoop.
First and foremost, this study has nothing to do with low carbers, or even heart disease: it’s an observational study of Swedes Not Further Specified (SNFS). Researchers pulled 25 years’ worth of data from 140,000 folks in the North Sweden Diet Database, averaged everyone’s food intake, averaged everyone’s total cholesterol, and made some pretty graphs.
The findings? Sweden’s fat consumption (as a percent of total energy) dropped between 1986 and 1991, held steady until about 2004, and then started rising; the inverse happened with carbs:
(In those early years, fat kerplunked due to the Vasterbotten Intervention Programme – a project launched in the mid ‘80s to encourage more physical activity, scoot Swedes towards a low-fat Mediterranean diet, improve food labeling, and hopefully knock heart disease off its tyrannous throne.)
And like an ant frantically evading the smoosh of a human thumb, average body mass index skittered wherever the heck it wanted (usually up):
And average total cholesterol – not among low carbers, but among north Sweden’s general population – followed a roller-coaster-esque trajectory of its own. It plummeted between 1986 and 1992, crept back up until 1994, rolled downhill again until 2000, stabilized between 2002 and 2008, and then began yet another upward jog until the study ended. A visual for your viewing pleasure:
That’s it. No distinction between HDL and LDL, no reported triglycerides, and – most notably – no mention of how diet or cholesterol changes corresponded with actual heart disease rates. In fact, there’s nary a morbidity or mortality statistic to be found in the whole paper.
So how did the doom-and-gloom warnings about low carbing enter the scene?
Simple: the researchers think Sweden’s recent cholesterol hike might be a result of the LCHF boom – although the design of this study, with dieters of all persuasions stirred together in a giant pot of data-soup, makes it impossible to tell if that’s actually the case. Although the media apparently didn’t get the memo, the researchers even stated in their paper that “our study design does not allow a causal evaluation of the relationship between the increased fat intake since 2004 and the increased cholesterol values after 2007.” Likewise, the media thinks Sweden’s ever-increasing BMI, despite being an average of the general population, means that low-carb adherents fail to maintain their weight loss. (Yeah, I’m scratching my head over that logic too.)
Although it’s possible that saturated-fat-loving, low-carbing Swedes singlehandedly raised the nation’s average cholesterol in 2008, this seems iffy for a few reasons. For one, fat intake and blood cholesterol were hardly mirroring each other for the study’s previous 22 years: cholesterol levels continued to drop even when fat intake remained steady (though this could also be confounded by use of statins and ratios of specific fatty acids). Even more suspiciously, there was a four-to-six-year lag between the rise in fat consumption and the 2008 cholesterol jump – implying that Sweden’s fat-feasting took half a decade to affect blood values. (Usually cholesterol reflects diet changes in a matter of weeks.) And although this particular study tells us nada beyond “total cholesterol,” another paper published just a few months ago – drawing from the same pool of data – shows that after 2008, north Sweden’s triglycerides were frolicking uphill with cholesterol:
If that graph convinces anyone that the Low Carb High Fat movement is spiking triglycerides, I’ve got some oceanfront property in Arizona to sell ‘em, too: the evidence is pretty consistent that low carbing makes triglycerides sink . It seems more likely that other factors are lurking behind Sweden’s shifting blood lipid patterns, especially during the past few years.
Lo and behold, the fine print of our study-du-jour suggests the same. According to the researchers, foods associated with a high fat intake in Sweden were not just the predictable oils and meats, but also pizza, French fries, potato chips, corn chips, cheese-flavored puffed products, and popcorn, as well as “fats used for spreading on bread.” Unless Sweden has a very different definition of “low carb” than the rest of the world, it sounds like their fat intake isn’t coming solely from LCHF-approved fare, but from what most of us affectionately refer to as bona fide junk.
Even when it comes to foods directly associated with high cholesterol, Sweden’s low-carb movement still can’t shoulder all the blame. According to a supplementary table kindly included with the paper (available as a Word document here ), boiled potatoes and coffee made bigger statistical contributions to blood cholesterol than saturated fat did! Other foods associated with high cholesterol included not just high-fat foods, but also low-carb no-gos like sweet buns, crisp bread, white bread, and sweet fruit soups.
Although this study doesn’t reveal anything very useful about what food does to us (particularly where fat is concerned), fear not: the diet-disease data for northern Sweden is a many splendored thing, and other studies have mined it with much more fruitful results. Or in some cases, milkful.
In 2010, the American Journal of Clinical Nutritionpublished a paper looking directly at dairy fat and myocardial infarction among some of the same northern Swedes used in last week’s study. The results? Women with high intakes of dairy fat – confirmed both by food-frequency questionnaires and ruminant-milk-fat biomarkers – were less likely to suffer from a first heart attack than their milk-minimizing counterparts. (Cheese and fermented dairy products looked particularly heart-protective.) And just in case you think that one was a fluke, a 2004 study examining the same Swedish population found dairy fat to be negatively associated with cardiovascular risk factors, with zero indication that full-fat milk products contribute to heart attacks. Want more? Yet another study (PDF), again based on those fat-loving northern Swedes, found dairy fat to be beautifully protective against strokes – especially in women.
So much for all that “artery clogging” hoopla!
It’s a mystery why our current study tried to glean anything about low-carb diets from such tight-lipped data – especially since it’d be easy enough to actually untangle those low carbers from the rest of the population and follow them, up close and personal, like the stalkers we all secretly yearn to be. In fact, another north Sweden study published in February attempted that very feat (sans creepiness), and failed to find any increased mortality among the folks it deemed “low carb.” (Unfortunately, that study also used an invariably terrible diet-score design, but that’s a story for another day.)
As for Sweden’s recent cholesterol uppage? Without more detailed data and a look-see at disease rates, it’s virtually meaningless. But even if there’s no legitimate evidence dooming Swedish low carbers to an early grave, it’s possible that the LCHF movement is having some unhappy consequences in the Land of the Midnight Sun. The health climate there is clearly whipping fat-phobia into remission – which, despite the admirable triumph over bad science, could also breed a new species of half-hearted dieters like Low Carb Weekend Warriors, Low Carb As Long as There’s Not a Cookie in Front of Me, and Why Don’t I Just Put Butter on Everything Edible and Buy New Pants When They Get Too Tight. Folks who jump on the “high fat” bandwagon while still living in high-carb land may indeed find themselves gaining weight and frightening their doctors with ominous lipid panels. Whether or not this is happening in Sweden right now remains to be known, but it is a possibility.
Personally, though, I have other theories:
I’m not saying correlation equals causation. I’m just saying maybe we should think about switching to Google Plus.