Vous êtes sur la page 1sur 10

The Soft Science of Dietary Fat

Author(s): Gary Taubes

Source: Science, New Series, Vol. 291, No. 5513 (Mar. 30, 2001), pp. 2536-2541+2543-2545
Published by: American Association for the Advancement of Science
Stable URL: http://www.jstor.org/stable/3082809
Accessed: 15/03/2009 22:09

Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at
http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless
you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you
may use content in the JSTOR archive only for your personal, non-commercial use.

Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at

Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed
page of such transmission.

JSTOR is a not-for-profit organization founded in 1995 to build trusted digital archives for scholarship. We work with the
scholarly community to preserve their work and the materials they rely upon, and to build a common research platform that
promotes the discovery and use of these resources. For more information about JSTOR, please contact support@jstor.org.

American Association for the Advancement of Science is collaborating with JSTOR to digitize, preserve and
extend access to Science.

K y. *; i ffi z

Mainstream nutritional science has demonized dietary fat, yet 50 years and hundreds of millions of
dollars of research have failed to prove that eating a low-fat diet will help you live longer

The Soft Science of

Dietary Fat
Whenthe U.S. SurgeonGeneral'sOfficeset "Clearlythe thoughtsof yesterda
iy werenot Department of Agriculture's (USDA's)
off in 1988to writethe definitivereporton going to serveus verywell." bookleton dietaryguidelines,publishedev-
the dangersof dietaryfat,the scientifictask Duringthe past 30 years,the conceptof ery 5 years,and its ubiquitousFoodGuide
appearedstraightforward. Fouryearsearlier, eatinghealthyin Americahas be:comesyn- Pyramid,which recommendsthat fats and
the NationalInstitutesof Health(NIH)had onymous with avoiding dietary fat. The oils be eaten"sparingly."The low-fatgospel
begunadvisingeveryAmericanold enough creationandmarketingof reduce ed-fatfood spreadsfartherby a kindof societalosmo-
to walk to restrictfat intake,and the presi- products has become big business; over sis, continuouslyreinforcedby physicians,
dent of the American HeartAssociation 15,000 have appeared on suIpermarket nutritionists,journalists,health organiza-
(AHA) had told Timemagazinethatif ev- shelves.Indeed,an entirereseare
chindustry tions, and consumeradvocacygroupssuch
eryone went along, "we will have has arisen to create palatablenonfat fat as the Centerfor Sciencein the PublicInter-
[atherosclerosis]conquered"by the year substitutes, and the food indiustry now est, which refers to fat as this "greasy
2000. The SurgeonGeneral'sOffice itself spendsbillionsof dollarsyearlyselling the killer.""InAmerica,we no longerfearGod
hadjust publishedits 700-page landmark less-fat-is-good-healthmessage The gov- or the communists,but we fear fat,"says
"Reporton NutritionandHealth,"declaring ernmentweighs in as well, with the U.S. DavidKritchevskyof the WistarInstitutein
fat the single most unwholesomecompo- Philadelphia,who in 1958 wrote the first
nentof theAmericandiet. textbookon cholesterol.
All of this was apparentlybased on "They say,'You re:ally As the Surgeon General'sOffice dis-
soundscience. So the task beforethe pro- covered,however,the scienceof dietaryfat
ject officer was merelyto gatherthat sci- need a high levelI of is not nearlyas simpleas it once appeared.
ence togetherin one volume, have it re- The proposition,now 50 yearsold, thatdi-
viewed by a committeeof experts,which proof to change the etary fat is a bane to health is based
hadbeen promptlyestablished,andpublish , chiefly on the fact thatfat, specificallythe
it. The projectdid not go smoothly,howev- recommendatior lS, hard,saturatedfat foundprimarilyin meat
er.Fourprojectofficerscameandwentover and dairyproducts,elevatesblood choles-
the next decade."Itconsumedprojectoffi- which is ironic:' terol levels. This in turn raises the likeli-
cers," says Marion Nestle, who helped because they ne, fver hood that cholesterolwill clog arteries,a
launchthe projectand now runsthe nutri- conditionknownas atherosclerosis,which
tion and food studies departmentat New had a high level of then increasesrisk of coronaryarterydis-
YorkUniversity(NYU). Membersof the ease, heartattack,and untimelydeath.By
oversightcommitteesaw draftsof an early proof to set thern." the 1970s, each individual step of this
chapteror two, criticizedthem vigorously, chain from fat to cholesterolto heartdis-
andthensawlittleelse. -Walter Willett ease had been demonstratedbeyond rea-
Finally,in June 1999, 11 yearsafterthe sonable doubt, but the veracity of the
projectbegan, the SurgeonGeneral'sOf- chainas a wholehas neverbeenproven.In
fice circulateda letter,authoredby the last otherwords,despitedecadesof research,it
of the projectofficers,explainingthatthe is still a debatablepropositionwhetherthe
report would be killed. There was no consumptionof saturatedfats above rec-
other public announcement and no ommendedlevels (step one in the chain)
pressrelease.The letterexplainedthat by anyonewho'snot alreadyat high riskof
the relevantadministrators"did not heartdiseasewill increasethe likelihood
anticipatefully the magnitudeof the of untimelydeath (outcomethree).Nor
additionalexternalexpertiseand staff have hundredsof millions of dollars in
resources that would be needed." In trials managedto generatecompelling
other words, says Nestle, the subject evidence that healthy individuals
matter"wastoo complicated." Bill Har- MH_i ^-. w can extend their lives by V
lan, a member of the oversight lIk^jw ; / : more than a few weeks, if 8
committeeandassociatedirector that, by eating less fat
of the Office of Disease Preven- (see sidebar on
tion at NIH, says "thereportwas p. 2538). To put it z
initiatedwith a preconceivedopinion simply, the data
of the conclusions,"but the sciencebe- _ ouremain ambigu-
as towheth
, ~
hind those opinions was not holding up. ?:;A:&A:,
. .:
....... ous as to whetherS

2536 30 MARCH2001 VOL291 SCIENCE www.sciencemag.org

lzv,, .
i ! .
low-fat diets will benefit healthyAmeri- the pasta and othercarbohydrates that the nary heart disease epidemic seemed to
cans. Worse, the ubiquitous admonish- Food GuidePyramidsuggestsbe eatenco- sweepthe country(see sidebaron p. 2540).
ments to reduce total fat intake have piously.(The studiesalso suggestthattrans "Middle-aged men,seeminglyhealthy,were
encourageda shift to high-carbohydrate fatty acids are unhealthful.These are the droppingdead,"wrote biochemist Ancel
diets, which may be no better-and may fats in margarine,
for instance,andarewhat Keys of the Universityof Minnesota,Twin
even be worse-than high-fatdiets. manyAmericansstartedeatingwhen they Cities,who was amongthe first to suggest
Since the early 1970s, for instance, were told that the saturatedfats in butter that dietary fats might be the cause. By
Americans'averagefat intakehas dropped might kill them.) Harvardepidemiologist 1952, Keys was arguing that Americans
fromover40%of totalcaloriesto 34%;av- WalterWillett,spokesperson fortheNurses' shouldreducetheir fat intaketo less than
erageserumcholesterollevelshavedropped HealthStudy,pointsout thatNIH has spent 30% of total calories,althoughhe simulta-
as well. But no compellingevidence sug- over $100 millionon the threestudiesand neouslyrecognizedthat"directevidenceon
gests that these decreaseshave improved theeffectof the dieton humanarteriosclero-
health.Althoughheartdiseasedeathrates sis is verylittleandlikelyto remainso for
have dropped-and public healthoffi- sometime."Inthe famousandverycon-
cials insist low-fat diets are partly | troversialSevenCountriesStudy,for in-
responsible-the incidenceof heartdis- stance,Keysandhis colleaguesreported
ease does not seem to be declining,as thattheamountof fatconsumedseemed
wouldbe expectedif lowerfatdietsmade to be the salientdifferencebetweenpop-
a difference.Thiswas the conclusion,for ulationssuchas thosein JapanandCrete
instance,of a 10-yearstudyof heartdis- H thathadlittleheartdiseaseandthose,as
ease mortalitypublishedin TheNew En- in Finland,that were plaguedby it. In
gland Journal of Medicine in 1998, which 1961, the FraminghamHeart Study
suggestedthat deathrates are declining linkedcholesterollevelsto heartdisease,
largelybecause doctorsare treatingthe Keysmadethe coverof Timemagazine,
diseasemore successfully.AHA statistics andtheAHA,underhis influence,began
agree:Between1979and1996,thenumber advocatinglow-fatdietsas a palliativefor
of medicalproceduresfor heartdiseasein- men with high cholesterollevels. Keys
creasedfrom 1.2 million to 5.4 million a had also becomeone of the firstAmeri-
year."I don'tconsiderthatthisdiseasecate- cans to consciouslyadopta heart-healthy
goryhasdisappeared oranythingcloseto it," diet:He andhis wife, Timereported, "donot
saysoneAHAstatistician. eat 'carvingmeat'-steaks, chops,roasts-
Meanwhile,obesity in America,which morethanthreetimesa week."
remainedconstant from the early 1960s "InAmerica, we no Nonetheless,by 1969the stateof the sci-
through 1980, has surged upward since ence could still be summarizedby a single
then-from 14%of the populationto over longer fear God or the sentencefroma reportof theDiet-HeartRe-
22%. Diabeteshas increasedapace. Both view Panel of the NationalHeartInstitute
obesityand diabetesincreaseheartdisease communists, but (now the NationalHeart,Lung,and Blood
risk,whichcouldexplainwhy heartdisease Institute, or NHLBI): "It is not known
incidenceis not decreasing.Thatthis obesi- we fear fat." whetherdietarymanipulation has anyeffect
ty epidemicoccurredjust as the government whatsoeveron coronaryheartdisease."The
-David Kritchevsky
beganbombarding Americanswiththe low- chairof the panelwas E. H. "Pete"Ahrens,
fat messagesuggeststhe possibility,howev- whose laboratoryat RockefellerUniversity
er distant,thatlow-fatdietsmighthaveun- yet not one government agencyhas changed in New YorkCity did muchof the seminal
intended consequences-among them, its primaryguidelinesto fit theseparticular researchon fatandcholesterolmetabolism.
weight gain. "Mostof us would have pre- data."Scandalous," saysWillett."Theysay, Whereas proponents of low-fat diets
dictedthat if we can get the populationto 'You really need a high level of proof to were concernedprimarilyaboutthe effects
changeits fat intake,withits densecalories, change the recommendations,'which is of dietaryfat on cholesterollevelsandheart
we wouldsee a reductionin weight,"admits ironic,becausethey neverhad a high level disease,Ahrensand his panel-10 experts
Harlan."Instead, we see theexactopposite." of proofto set them." in clinical medicine, epidemiology, bio-
In the face of this uncertainty,skeptics Indeed,the historyof the nationalcon- statistics,humannutrition,and metabolism
and apostateshave come along repeatedly, victionthatdietaryfat is deadly,andits evo- -were equallyconcernedthat eating less
onlyto see theirworkalmostreligiouslyig- lutionfromhypothesisto dogma,is one in fat could haveprofoundeffectsthroughout
noredas the mainstream medicalcommuni- which politicians,bureaucrats,the media, the body,manyof whichcouldbe harmful.
ty soughtconsensuson the evils of dietary andthe publichaveplayedas largea roleas The brain,for instance,is 70% fat, which
fat. For20 years,for instance,the Harvard the scientistsandthe science.It'sa storyof chieflyservesto insulateneurons.Fatis also
Schoolof PublicHealthhas runtheNurses' whatcanhappenwhenthe demandsof pub- the primarycomponentof cell membranes.
Health Study and its two sequelae-the lic healthpolicy-and the demandsof the Changingthe proportionof saturated to un-
HealthProfessionalsFollow-UpStudyand publicforsimpleadvice-run up againstthe saturatedfats in the diet changes the fat
a the Nurses'HealthStudyII-accumulating confusingambiguityof realscience. composition in these membranes. This
| overa decadeof dataon the diet andhealth could conceivablychange the membrane
| of almost300,000 Americans.The results Fearof fat permeability,which controlsthe transport
- suggestthattotalfat consumedhas no rela- During the firsthalf of the 20th nu-
century, of everythingfromglucose, signalingpro-
| tion to heartdiseaserisk;thatmonounsatu- tritionistswere moreconcernedaboutmal- teins,andhormonesto bacteria,viruses,and
- lowerrisk;and that
ratedfats like olive oil nutritionthanaboutthe sins of dietaryex- tumor-causingagents into and out of the
u saturatedfats are littleworse,if at all, than cess. AfterWorldWarII, however,a coro- cell.Therelativesaturation of fatsin thediet

www.sciencemag.orgSCIENCEVOL291 30 MARCH
2001 2537

What IfAmericansAte LessSaturated Fat? not,"wrote Taylorand his colleagues.

The following year, the U.S. Surgeon General'sOffice funded a
Eat less saturated fat, live longer. For 30 years, this has stood as study at the Universityof California,San Francisco,with the expec-
one cornerstoneof nutritionaladvice given to Americans(see main tation that its results would counterbalancethose of the Harvard
text). But how much longer? Between 1987 and 1992, three inde- analysis.Ledby epidemiologistWarrenBrowner,this study conclud-
pendent research groups used computer models to work out the ed that cutting fat consumption in America would delay 42,000
answer.All three analyses agreed, but their conclusions have been deaths each year, but the net increase in life expectancy would av-
buriedin the literature,rarelyif ever cited. erage out to only 3 to 4 months. The key word was "delay,"for
All three models estimated how much longer people might ex- death, like diet, is a trade-off: Everyonehas to die of something.
pect to live, on average, if only 10% of their calories came from "Deaths are not prevented, they are merely delayed," Browner
saturated fat as recommended.Inthe process their total fat intake later wrote. "The'saved' people mainly die of the same things ev-
would drop to the recommended 30% of calories.All three models eryone else dies of; they do so a little later in life."To be precise,a
assumed that LDLcholesterol-the "bad cholesterol"-levels woman who might otherwise die at 65 could expect to live two
would drop accordinglyand that this diet would have no adverse extra weeks after a lifetime of avoidingsaturatedfat. If she lived to
effects, although that was optimistic at the time and has become be 90, she could expect 10 additionalweeks. The third study, from
considerablymore so since then. All three combined national vital researchers at McGill University in Montreal, came to virtually
statistics data with cholesterol riskfactor data from the Framing- identicalconclusions.
ham Heart Study. Brownerreported his results to the Surgeon General's Office,
The first study came out of HarvardMedical School and was then submitted a paper to The Journal of the American Medical
published in the Annals of InternalMedicine in April 1987. Led by Association U(AMA). Meanwhile,the Surgeon General'sOffice-his
William Taylor,it concluded that individuals with a high risk of source of funding-contacted JAMAand tried to prevent publica-
heart disease-smokers, for instance, with high blood pressure- tion, claimingthat the analysis was deeply flawed.JAMAreviewers
could expect to gain, on average,one extra year by shunningsatu- disagreedand publishedhis article,entitled "WhatIfAmericansAte
rated fat. Healthy nonsmokers, however, might add 3 days to Less Fat?"in June 1991. As for Browner,he was left protecting his
3 months. "Althoughthere are undoubtedly persons who would work from his own funding agents. "Shootingthe messenger,"he
choose to participate in a lifelong regimen of dietary change to wrote to the Surgeon General's Office, "or creating a smoke
achieve results of this magnitude, we suspect that some might screen-does not change those estimates." -G.T.

could also influence cellular aging as well cal establishment and the food industry-
tically significant results.Ahrens and his col-
as the clotting ability of blood cells. leagues were pessimistic about whether suchand by countercultureattacks on excessive
Whetherthe potential benefits of low-fat a massive and expensive trial could ever beconsumption, whether manifested in gas-
diets would exceed the potential risks could done. In 1971, an NIH task force estimated guzzling cars or the classic American cui-
be settled by testing whether low-fat diets sine of bacon and eggs and marbled steaks.
such a trial would cost $1 billion, consider-
actually prolong life, but such a test would ably more than NIH was willing to spend. And while the data on fat and health re-
have to be enormous. The effect of diet Instead, NIH administrators opted for a mained ambiguous and the scientific com-
on cholesterol levels is subtle for most handful of smaller studies, two of which munity polarized, the deadlock was broken
individuals-especially those living in the alone would cost $255 million. Perhaps not by any new science, but by politicians.
real world ratherthan the metabolic wards of more important,these studies would take a It was Senator George McGovern's biparti-
nutrition researchers-and the effect of decade. Neither the public, the press, nor the
san, nonlegislative Select Committee on
cholesterol levels on heart disease is also Nutrition and Human Needs-and, to be
U.S. Congress was willing to wait that long.
subtle. As a result, tens of thousandsof indi- precise, a handful of McGovern's staff
viduals would have to switch to low-fat diets Science by committee members-that almost single-handedly
and their subsequenthealth comparedto that Like the flourishing American affinity for changed nutritional policy in this country
of equal numbers who continued eating fat alternative medicine, an antifat movement and initiated the process of turning the di-
to alleged excess. And all these people evolved independently of science in etary fat hypothesis into dogma.
would have to be followed for years until the 1960s. It was fed by distrust of the McGovern's committee was founded in
enough deaths accumulatedto providestatis- establishment-in this case, both the medi- 1968 with a mandate to eradicate malnutri-
tion in America, and
it instituted a series
: ?' "There comes a point when, if you of landmark federal
food assistance pro-
don't make a decision, the conse- grams. As the malnu-
trition work began to
quences can be great as well. If you peter out in the mid-
1970s, however, the
just allow Americans to keep on committee didn't dis- u
band. Rather,its gen- I
consuming 40% of calories from fat,
eral counsel, Mar-
shall Matz, and staff
there's an outcome to that as well." director, Alan Stone, u

both young lawyers,

-Basil Rifkind decided that the com- lu
mittee would address

2538 30 MARCH2001 VOL291 SCIENCE www.sciencemag.org

"overnutrition,"the dietary excesses of next 7 years of controversy.Among those The guidelinesmight have then died a
Americans.It was a "casualendeavor," says testifying,for instance,was NHLBIdirector quiet death when McGovern'scommittee
Matz."Wereallyweretotallynaive,a bunch Robert Levy, who explained that no one came to an end in late 1977 if two federal
of kids, who just thought,'Hell, we should knew if eating less fat or loweringblood agencies had not felt it imperativeto re-
say somethingon this subjectbeforewe go cholesterollevels would preventheartat- spond. Although they took contradictory
out of business.'"McGovernandhis fellow tacks,whichwas whyNHLBIwas spending points of view, one message-with media
senators-all middle-aged men worried $300 million to studythe question.Levy's assistance-won out.
abouttheir girth and their health-signed positionwas awkward,he recalls,because The first was the USDA, where
on; McGovernand his wife had both gone "thegood senatorscameoutwiththe guide- consumer-activistCarolTuckerForeman
throughdiet-guruNathanPritikin's verylow linesandthencalledus in to get advice."He had recently been appointedan assistant
fat diet and exercise program.McGovern was joined by prominentscientists,includ- secretary.Foremanbelievedit was incum-
quittheprogramearly,butPritikinremained ing Ahrens, who testified that advising bent on USDA to turnMcGovern'srecom-
a majorinfluenceon his thinking. Americansto eat less fat on the strengthof mendationsinto official policy, and, like
McGovern's committee listened to such marginalevidence was equivalentto Mottern,she was not deterredby the exis-
2 days of testimonyon diet and disease in conductinga nutritional experiment withthe tenceof scientificcontroversy."Tellus what
July 1976. Then residentwordsmithNick you know and tell us it's not the final an-
Mottern,a formerlabor reporterfor The swer,"she would tell scientists."I have to
Providence Journal, was assigned the task eat and feed my childrenthreetimesa day,
of researchingand writing the first "Di- and I want you to tell me what your best
etary Goals for the United States."Mot- senseof thedatais rightnow."
tern,who hadno scientificbackgroundand Of course, given the controversy,the
no experiencewritingaboutscience,nutri- "bestsense of the data"would dependon
tion, or health,believedhis DietaryGoals which scientistswere asked.The Foodand
would launch a "revolution in diet and NutritionBoardof theNationalAcademyof
agriculturein this country."He avoidedthe Sciences(NAS),whichdecidesthe Recom-
scientificand medicalcontroversyby rely- mended DietaryAllowances, would have
ing almost exclusivelyon HarvardSchool been a naturalchoice, but NAS president
of PublicHealthnutritionistMarkHegsted Philip Handler,an experton metabolism,
for inputon dietaryfat. Hegstedhad stud- had told Foremanthat Mottern'sDietary
ied fat and cholesterolmetabolismin the Goals were "nonsense." Foreman then
early 1960s, and he believeduncondition- turnedto McGovern'sstaffers for advice
ally in the benefitsof restrictingfat intake, and they recommendedshe hire Hegsted,
althoughhe says he was awarethathis was whichshe did. Hegsted,in turn,reliedon a
an extremeopinion. With Hegsted as his state-of-the-sciencereportpublishedby an
muse,Motternsaw dietaryfat as the nutri- expert very divergentcommitteeof the
tionalequivalentof cigarettes,andthe food AmericanSociety for Clinical Nutrition.
industryas akinto the tobaccoindustryin "They were nowherenear unanimouson
its willingnessto suppressscientific truth anything,"says Hegsted,"butthe majority
in the interestsof profits.To Mottern,those supportedsomething like the McGovern
scientistswho spoke out againstfat were committeereport."
those willing to take on the industry."It American public as subjects. Even the The resultingdocumentbecamethe first
took a certainamountof guts,"he says, "to AmericanMedicalAssociation protested, editionof "Usingthe DietaryGuidelinesfor
speak aboutthis because of the financial suggesting that the diet proposedby the Americans." Althoughit acknowledgedthe
interestsinvolved." guidelinesraisedthe "potentialfor harmful existenceof controversy andsuggestedthata
Mottern'sreportsuggestedthatAmeri- effects."But as these scientiststestified,so singledietaryrecommendation mightnotsuit
cans cut theirtotalfat intaketo 30%of the did representativesfromthe dairy,egg, and an entirediversepopulation,the advice to
caloriesthey consumeandsaturatedfat in- cattle industries,who also vigorouslyop- avoidfatandsaturated fatwas,indeed,virtu-
take to 10%,in accordwith AHA recom- posed the guidelinesfor obvious reasons. allyidenticalto McGover'sDietaryGoals.
mendationsfor men at high risk of heart Thisjuxtapositionservedto taintthe scien- Threemonthslater,the NAS Food and
disease.The reportacknowledgedthe exis- tific criticisms: Any scientists arguing NutritionBoardreleasedits ownguidelines:
tence of controversybut insisted Ameri- againstthe committee'sguidelinesappeared "TowardHealthfulDiets."The board,con-
cans had nothingto lose by following its to be eitherhopelesslybehindtheparadigm, sistingof a dozennutritionexperts,conclud-
advice. "The questionto be asked is not whichwas Hegsted'sview,or industryapol- ed thatthe only reliableadvice for healthy
why should we change our diet but why ogists,whichwasMottern's, if notboth. Americanswas to watchtheirweight;ev-
not?"wrote Hegsted in the introduction. Althoughthe committeepublisheda re- erythingelse, dietaryfat included,would
"Thereare [no risks]thatcan be identified vised editionof the DietaryGoals laterin takecareof itself.The advicewas not taken
and importantbenefits can be expected." the year,the thrustof the recommendations kindly,however,at least not by the media.
This was an optimisticbut still debatable remainedunchanged.It didgive in to indus- The first reports-"ratherincredulously,"
position,and when DietaryGoals was re- trypressureby softeningthe suggestionthat said Handlerat the time-criticized the
| leased in January 1977, "all hell broke Americanseat less meat. Motternsays he NAS adviceforconflictingwiththeUSDA's
loose," recalls Hegsted. "Practicallyno- consideredeven that a "disserviceto the and McGovern'sand thus somehowbeing
bodywas in favorof the McGover recom- public,"refusedto do the revisions,andquit irresponsible.Follow-upreportssuggested
mendations.Damnfew people." the committee.(Motternbecamea vegetari- that the boardmembers,in the words of
McGovernrespondedwith threefollow- an whilewritingthe DietaryGoalsandnow JaneBrody,who coveredthe storyfor The
u up hearings,which aptlyforeshadowed the runsa foodco-opin Peekskill,NewYork.) New YorkTimes,were "all in the pocketof

www.sciencemag.org SCIENCE VOL291 30 MARCH2001 2539


The EpidemicThatWasn't? of dyingfroma heartattackat any particularage remainedunchanged:

Rather,the risingnumberof 50-year-oldsdroppingdead of heart at-
Forhalf a century,nutritionistshave pointed to soaring death rates tackswas primarily due to the risingnumberof 50-year-olds.
as the genesis of their research into dietary fat and heart disease The secondary factor was an increase from 1948 to 1968 in the
and as reason to advise Americansto eat less fat (see main text). probabilitythat a death would be classified on a death certificate
"We had an epidemic of heart disease after WorldWar II,"obesity as arterioscleroticdisease or coronaryheart disease. This increase,
expert Jules Hirschof RockefellerUniversityin New YorkCity said however,was a figment of new diagnostictechnologies-the wider
just 3 months ago in TheNew YorkTimes."Therates were growing use of electrocardiograms,for instance-and the changing termi-
higher and higher,and people became suddenly aware of that, and nology of death certificates. In 1949, the InternationalClassifica-
that diet was a factor." tion of Diseases (ICD)added a new category,"arterioscleroticheart
To proponents of the antifat message, this heart disease epi- disease,"underthe more general rubric"diseasesof the heart."The
demic has always been an indisputablereality.Yet,to the statisti- result, as a 1958 report to the AmericanHeartAssociation noted,
cians at the mortality branch of the National Center for Health was dramatic:"Inone year, 1948 to 1949, the effect of this revision
Statistics (NCHS),the source of all the relevant statistics, the epi- was to raise coronarydisease death rates by about 20% for white
demic was illusory.In their view, heart disease deaths have been males and about 35% for white females."In 1965, the ICDadded a
steadily decliningsince the late 1940s. category for coronaryheart disease, which added yet more deaths
Accordingto HarryRosenberg,director of the NCHSmortality and capped off the apparentepidemic.
branchsince 1977, the key factor in the apparentepidemic,paradoxi- To Rosenbergand others at NCHS,the most likely explanation
cally,was a healthierAmericanpopulation.By the 1950s, premature for the postwar upsurge in coronary heart disease deaths is that
deaths from infectiousdiseases and nutritionaldeficiencieshad been physiciansslowly caught on to the new terminology and changed
all but eliminated,whichleft moreAmericanslivinglong enoughto die the wording on death certificates."Thereis absolutely no evidence
of chronicdiseasessuch as heartdisease.Inotherwords,the actualrisk that there was an epidemic,"says Rosenberg. -G.T.

the industriesbeinghurt."To be precise,the ie WashingtonPost), demonstratedthatthe methodologicalflaws had led to the nega-
boardchairandone of its membersconsult- ard's "objectivity and aptitude are in tive results.They did not, at least publicly,
ed for food industries,and fundingfor the ubt"(The New YorkTimes),or represented considertheirresultsreasonto lessen their
boarditself came from industrydonations. theboard'sguidelinesa "blowagainstthe beliefin the evils of fat.
These industryconnectionswere leakedto Adfaddistswho hold the publicin thrall" The sixth study was the $140 million
thepressfromtheUSDA. ;ience).In any case, the NAS boardhad Lipid ResearchClinics (LRC) Coronary
Hegstednow defendsthe NAS board,al- :n publiclydiscredited.Hegsted'sDietary PrimaryPrevention Trial,led by NHLBIad-
thoughhe didn'tat the time, and calls this idelinesfor Americansbecamethe offi- ministratorBasil Rifkind and biochemist
kind of conflictof interest"a hell of an is- 1 U.S. policy on dietaryfat: Eat less fat. Daniel Steinbergof the Universityof Cali-
sue.""Everybody used to complainthatin- velonger. fornia,SanDiego.TheLRCtrialwas a drug
dustrydidn'tdo anythingon nutrition," he trial,not a diettrial,butthe NHLBIherald-
toldScience,"yetanybodywho got involved eating"consensus" ed its outcomeas the end of the dietaryfat
was blackballed becausetheirpositionswere ice politicians,the press, and the public debate.In January1984,LRCinvestigators
presumablyinfluencedby the industry." (In I decideddietaryfat policy,the science reportedthata medicationcalledcholestyra-
1981, Hegstedreturnedto Harvard,where s left to catch up. In the early 1970s, minereducedcholesterollevelsin menwith
his researchwas fundedby Frito-Lay.) The en NIH optedto forgo a $1 billiontrial abnormally high cholesterol levels and
presshad mixedfeelings,claimingthatthe itmightbe definitiveand insteadfunda modestlyreducedheartdiseaseratesin the
connections"soiled"the academy'sreputa- f-dozenstudiesat one-thirdthe cost, ev- process. (The probability of suffering a
tion "fortenderingcarefulscientificadvice" tonehopedthese smallertrialswouldbe heartattackduringthe seven-plusyearsof
2.1 -
sufficientlypersuasiveto conclude the study was reduced from 8.6% in the
thatlow-fatdietsprolonglives. The placebogroupto 7.0%;the probabilityof
TotalMortality results were published between dying from a heart attack droppedfrom
1980 and 1984. Fourof these trials 2.0% to 1.6%.) The investigatorsthen con-
-comparing heartdiseaseratesand cluded,withoutbenefitof dietarydata,that
o 1.5- diet within Honolulu,PuertoRico, cholestyramine's benefitscouldbe extended
1.3- Chicago, and Framingham- to diet as well. And althoughthe trialtested

1.1 - showed no evidencethat men who only middle-aged men with cholesterollev-
ate less fat livedlongeror hadfewer els higherthanthoseof 95%of the popula-
0.9 - heartattacks.A fifth trial,the Mul- tion, they concluded that those benefits
0.7- tiple Risk FactorInterventionTrial "couldandshouldbe extendedto otherage
0.5 _

I, I
(MRFIT), cost $115 million and groups and women and ... other more mod-
<160 160-199
>240 tried to amplify the subtle influ- est elevationsof cholesterollevels."
Bloodtotal cholesterol (mg/dl)
ences of diet on healthby persuad- Why go so far?Rifkindsays theirlogic
ing subjectsto avoid fat while si- was simple:For 20 years,he and his col-
The big picture. Pooled risk ratios of death from <al multaneouslyquittingsmokingand leagueshadarguedthatloweringcholesterol
causes for men and women aged 35 to 69 who haad taking medication for high blood levels preventedheart attacks.They had
shown no coronaryheartdiseaseat least 5 yearsearliE er pressure. That trial suggested, if spent enormous sums trying to prove it.
(A risk ratioof, say, 1.3 indicates a 30% increase in k.) anything,that eatingless fat might They felt they couldneveractuallydemon-
For men, risk ratios are higher at both high and lo)w shortenlife. In each study,however, stratethat low-fat diets prolongedlives-
cholesterollevels.(See page2543.) the investigators concluded that that would be too expensive,and MRFIT

2540 30 MARCH2001 VOL291 SCIENCE www.sciencemag.org

had failed-but now they had establisheda The test of time tions:Fathas nine caloriesper gramcom-
fundamentallink in the causalchain,from To the outside observer,the challenge in paredto fourcaloriesfor carbohydrates and
lower cholesterollevels to cardiovascular makingsense of any suchlong-running sci- protein, and so fat
cutting fromthe dietsurely
health.Withthat,theycouldtakethe leapof entific controversyis to establishwhether wouldcutpounds."Thisis heldalmostto be
faith from cholesterol-lowering drugs and the skepticsaresimplyon the wrongside of a religious truth,"says Harvard'sWillett.
health to cholesterol-lowering diet and the new paradigm,or whethertheirskepti- Considerable data,however,nowsuggestoth-
health.And afterall theireffort,they were cism is well founded.In otherwords,is the erwise.Theresultsof well-controlled clinical
eager-not to mentionurgedby Congress- science at issue based on sound scientific trialsare consistent:Peopleon low-fatdiets
to renderhelpful advice. "Therecomes a thinkingandunambiguous data,or is it what initiallylose a coupleof kilograms,as they
point when, if you don't make a decision, Sir FrancisBacon,for instance,wouldhave wouldon anydiet,andthentheweighttends
the consequencescanbe greatas well,"says called "wishfulscience,"basedon fancies, to return.After 1 to 2 years,littlehas been
Rifkind. "If you just allow Americansto opinions,andthe exclusionof contraryevi- achieved.Consider,for instance,the 50,000
keep on consuming40% of calories from dence?Baconofferedone viablesuggestion womenenrolledin the ongoing$100 million
fat,there'sanoutcometo thatas well." the two: the test of time.
for differentiating Women'sHealthInitiative(WHI).Half of
WiththeLRCresultsin press,theNHLBI Goodscienceis rootedin reality,so it grows these women have been extensivelycoun-
launchedwhatLevycalled"amassivepublic anddevelopsandthe evidencegets increas- seledto consumeonly 20%of theircalories
healthcampaign." Themediaobliginglywent fromfat.After3 yearson thisnear-draconian
along.Time,for instance,reportedthe LRC regime,say WHI sources,the womenhad
findingsunderthe headline"Sorry,It'sTrue. lost,on average,a kilogrameach.
Cholesterolreally is a killer."The article The link betweendietaryfat and heart
abouta drugtrialbegan:"Nowholemilk.No disease is more complicated,because the
butter.No fatty meats ..." Time followed up hypothesishas divergedinto two distinct
3 months later with a cover story: "And propositions:first,thatloweringcholesterol
Cholesterol andNow theBadNews. .." The preventsheartdisease;second,that eating
coverphotowas a frowningface:a breakfast less fat not only lowerscholesterolandpre-
platewith two friedeggs as the eyes and a vents heartdiseasebutprolongslife. Since
baconstripforthe mouth.Riflindwas quot- 1984,the evidencethatcholesterol-lowering
ed sayingthattheirresults"stronglyindicate drugsare beneficial-proposition number
thatthemoreyoulowercholesterolandfatin one-has indeed blossomed, at least for
yourdiet,the moreyou reduceyourriskof those at high risk of heart attack.These
heartdisease,"a statement thatstilllackeddi- drugsreduceserumcholesterollevels dra-
rectscientificsupport. matically,and they preventheart attacks,
The followingDecember,NIH effective- perhapsby othermeansas well. Theirmar-
ly endedthedebatewitha "ConsensusCon- ket has now reached$4 billiona yearin the
ference."The idea of such a conferenceis United States alone, and every new trial
thatan expertpanel,ideallyunbiased,listens seemsto confirmtheirbenefits.
to 2 daysof testimonyandarrivesat a con- The evidence supporting the second
clusionwith whicheveryoneagrees.In this proposition,that eating less fat makes for
case, Rifkindchairedthe planningcommit- a healthierand longer life, however,has
tee, which chose his LRC co-investigator remainedstubbornlyambiguous.If any-
Steinbergto lead the expertpanel.The 20 inglymorecompelling,whereaswishfulsci- thing, it has only become less compelling
speakersdid includea handfulof skeptics ence flourishesmost underits first authors over time. Indeed,since Ancel Keys start-
-including Ahrens,for instance,andcardi- before"goingdownhill." ed advocating low-fat diets almost
ologist MichaelOliverof ImperialCollege Such is the case, for instance,with the 50 yearsago, the scienceof fat andcholes-
in London-who arguedthatit was unsci- propositionthat dietaryfat causes cancer, terol has evolved from a simple story into
entific to equatethe effectsof a drugwith which was an integralpart of dietaryfat a very complicated one. The catch has
the effectsof a diet.Steinberg'spanelmem- anxietyin the late 1970s.By 1982,the evi- been that few involved in this business
bers,however,as Oliverlatercomplainedin dence supportingthis idea was thoughtto were preparedto deal with a complicated
TheLancet,"wereselectedto includeonly be so undeniablethata landmarkNAS re- story.Researchersinitiallypreferredto be-
expertswho would,predictably, say thatall port on nutritionand cancerequatedthose lieve it was simple-that a single un-
levels of blood cholesterol in the United researcherswho remainedskepticalwith wholesome nutrient, in effect, could be
Statesare too high and shouldbe lowered. "certaininterestedparties [who] formerly isolated from the diverse richness of hu-
And, of course, this is exactly what was arguedthat the associationbetween lung man diets;publichealthadministrators re-
said."Indeed,the conferencereport,written cancerand smokingwas not causational." a to to
quired simple story give Congress
by Steinbergandhis panel,revealedno evi- Fifteenyears and hundredsof millions of and the public; and the press needed a
dence of discord.Therewas "no doubt,"it researchdollarslater,a similarlymassive simple story-at least on any particular
concluded,that low-fat diets "will afford expertreportby theWorldCancerResearch day-to give to editorsand readersin 30
significant protection against coronary FundandtheAmericanInstitutefor Cancer columninches.But as contrariandatacon-
heart disease" to every Americanover 2 Researchcould find neither"convincing" tinued to accumulate,the complications
yearsold. The ConsensusConferenceoffi- nor even "probable" reasonto believe that became increasinglymore difficult to ig-
N cially gave the appearanceof unanimity dietaryfat causedcancer. nore or exclude,and the press beganwaf-
wherenone existed.After all, if therehad The hypothesisthatlow-fatdietsarethe fling or addingcaveats.The scientiststhen
0 been a trueconsensus,as Steinberghimself requisiterouteto weightloss hastakena sim- got the blamefor not stickingto the origi-
told Science, "you wouldn't have had to ilar downwardpath.This was the ultimate nal simple story, which had, regrettably,
s havea consensusconference." fallbackpositionin all low-fatrecommenda- neverexisted.

www.sciencemag.org SCIENCE VOL291 30 MARCH2001 2541


Morefats, fewer answers National CholesterolEducationProgram,

The originalsimplestoryin the 1950s was also advisinglow-fatdietsfor everyoneover
that high cholesterollevels increaseheart 2 yearsold, epidemiologist DavidJacobsof
disease risk. The seminal Framingham the Universityof Minnesota,TwinCities,
Heart Study,for instance,which revealed visitedJapan.Therehe learnedthatJapanese
the association between cholesterol and physicianswere advisingpatientsto raise
heartdisease,originallymeasuredonly to- theircholesterollevels,becauselow choles-
tal serumcholesterol.But cholesterolshut- terol levels were linked to hemorrhagic
tles throughthe blood in an arrayof pack- stroke.At the time,Japanesemen weredy-
ages. Low-density lipoprotein particles ing from stroke almost as frequently as
(LDL, the "bad"cholesterol) deliver fat Americanmen were succumbingto heart
and cholesterol from the liver to tissues disease.Back in Minnesota,Jacobslooked
that need it, including the arterialcells, for this low-cholesterol-stroke relationship
where it can lead to atherosclerotic in the MRFITdataand foundit there,too.
plaques.High-densitylipoproteins(HDLs, And the relationshiptranscendedstroke:
the "good"cholesterol)returncholesterol Menwithverylow cholesterollevelsseemed
to the liver.The higherthe HDL,the lower proneto prematuredeath;below 160 mil-
the heart disease risk. Then there are ligramsper deciliter(mg/dl),the lowerthe
triglycerides, which contain fatty acids, cholesterollevel,theshorterthelife.
and very low density lipoproteins The antifat movement Jacobsreportedhis resultsto NHLBI,
(VLDLs),whichtransporttriglycerides. which in 1990 hosted a conferenceto dis-
All of these particleshave some effect was founded on the cuss the issue, bringing together re-
on heartdiseaserisk,while the fats, carbo- searchers from 19 studies around the
hydrates,andproteinin the diet havevary- Puritan notion that world.The datawere consistent:Whenin-
ing effects on all these particles. The vestigatorstrackedall deaths, instead of
1950s story was that saturated fats in- "something bad had to just heart disease deaths, the cholesterol
crease total cholesterol, polyunsaturated curveswere U-shapedfor men and flat for
fats decreaseit, and monounsaturated fats have an evil cause, and women. In otherwords,men with choles-
are neutral.By the late 1970s-when re- terol levels above240 mg/dl tendedto die
searchersacceptedthe benefits of HDL- you got a heart attack prematurelyfromheartdisease.But below
theyrealizedthatmonounsaturated fats are 160 mg/dl, the men tendedto die prema-
not neutral. Rather, they raise HDL, at because you did some- turelyfrom cancer,respiratoryand diges-
least comparedto carbohydrates, and low- tive diseases,andtrauma.As for women,if
er LDL.This makestheman ideal nutrient thing wrong, which anything,the highertheir cholesterol,the
as far as cholesterol goes. Furthermore, was eating too much longerthey lived (see graphon p. 2540).
saturatedfats cannotbe quite so evil be- These mortalitydatacan be interpreted
cause, while they elevate LDL, which is of a bad thing, rather in two ways. One, preferredby low-fatad-
bad,they also elevateHDL,whichis good. vocates,is thatthey cannotbe meaningful.
And some saturatedfats-stearic acid, in than not having enough Rifkind,for instance,told Science thatthe
particular, the fat in chocolate-are at excess deathsat low cholesterollevels must
worstneutral.Stearicacid raisesHDL lev- of a good thing." be due to preexistingconditions.In other
els but does little or nothingto LDL.And words,chronicillness leads to low choles-
thentherearetransfattyacids,whichraise -John Powles terollevels,notvice versa.He pointedto the
LDL,just like saturatedfat, but also lower 1990 conference reportas the definitive
HDL. Today,none of this is controversial, documenton theissueandas supportforhis
althoughit has yet to be reflectedin any much as 70%-of the fat content of a argument,althoughthe report states un-
FoodGuidePyramid. porterhousewill improvecholesterollevels equivocallythat this interpretationis not
To understandwhere this complexity comparedto what they would be if bread, supported by thedata.
can lead in a simple example,considera or
potatoes, pasta were consumed instead. The other interpretationis that what a
steak-to be precise,a porterhouse,select The remaining 30% will raise LDL but low-fatdiet does to serumcholesterollev-
cut, with a half-centimeter layerof fat, the will also raise HDL. All of this suggests els, and what that in turndoes to arteries,
nutritionalconstituentsof which can be thateatinga porterhousesteakratherthan may be only one componentof the diet's
found in the NutrientDatabasefor Stan- carbohydrates mightactuallyimproveheart effect on health. In other words, while
dardReferenceat the USDA Website. Af- disease risk, although no nutritionalau- low-fatdiets mighthelp preventheartdis-
ter broiling,this porterhousereducesto a thoritywho hasn't writtena high-fat diet ease, they mightalso raisesusceptibilityto
servingof almostequalpartsfat and pro- bookwill saythispublicly. otherconditions.This is whatalwayswor-
tein. Fifty-onepercentof the fat is mono- As for the scientificstudies,in the years ried Ahrens.It's also one reasonwhy the
- unsaturated, of whichvirtuallyall (90%)is since the 1984 consensusconference,the American College of Physicians, for in-
g oleic acid, the same healthy fat that's in one thingtheyhavenot doneis pile up evi- stance, now suggests that cholesterol re-
2 olive oil. Saturated fat constitutes45% of dence in supportof the low-fat-for-allap- ductionis certainlyworthwhilefor thoseat
S the total fat, but a thirdof that is stearic proachto the publicgood. If anything,they high, short-termrisk of dying of coronary
, which is, at the very least, harmless. have added weight to Ahrens'sfears that heart disease but of "much smaller or ...
Theremaining4% of the fat is polyunsatu- theremaybe a downsideto populationwide uncertain"benefitfor everyoneelse.
| rated,whichalso improvescholesterollev- low-fatrecommendations. In 1986, for in- This interpretation-thatthe connection
o els. In sum,well overhalf-and
perhapsas 1
stance,just year after NIH launched the between diet and health far transcends

www.sciencemag.org SCIENCE VOL291 30 MARCH2001 2543

NEWS Focus
cholesterol-is also supportedby the single could be cut fromthe diet and the calories mixed carbohydrates? Do they add green
most dramaticdiet-hearttrialeverconduct- with it, but that'snot the case. Despite all leafy vegetables, or do they add pasta?
ed: the Lyon Diet Heart Study, led by expectationsto the contrary,peopletendto And so it goes. "Thesky'sthe limit,"says
Michel de Lorgerilof the FrenchNational consumethe same numberof caloriesde- nutritionist Alice Lichtenstein of Tufts
Instituteof Healthand Medical Research spitewhateverdiet they try.If they eat less Universityin Boston."Thereare a million
(INSERM)andpublishedin Circulationin total fat, for instance,they will eat more perturbations."
February1999. The investigatorsrandom- carbohydratesand probablyless protein, Thesetrade-offsalso confoundthe kind
ized 605 heart attack survivors, all on because most proteincomes in foods like of epidemiologicalstudiesthat demonized
cholesterol-lowering drugs, into two saturatedfat from the 1950s onward.In
groups.Theycounseledone to eat an AHA particular,individuals who eat copious
"prudentdiet,"very similarto thatrecom- "Most of us would amountsof meat and dairyproducts,and
mendedfor all Americans.They counseled plentyof saturatedfats in the process,tend
the otherto eat a Mediterranean-type diet, have predicted that if not to eat copious amountsof vegetables
with more bread,cereals,legumes,beans, andfruits.The sameholdsfor entirepopu-
vegetables,fruits, and fish and less meat.
we can get the popula- lations. The eastern Finns, for instance,
Totalfat andtypes of fat differedmarkedly whose lofty heartdisease ratesconvinced
in the two diets,butthe HDL,LDL,andto- tion to change its fat AncelKeysanda generationof researchers
tal cholesterollevels in the two groupsre- of the evils of fat, live within500 kilome-
mained virtually identical. Nonetheless, intake, with its dense tersof theArcticCircleandrarelysee fresh
over 4 years of follow-up, the Mediter- calories, we would see produceor a green vegetable.The Scots,
ranean-diet group had only 14 cardiac infamous for eating perhaps the least
deathsandnonfatalheartattackscompared a reduction in weight. wholesomediet in the developedworld,are
to 44 for the "Western-type"diet group. in a similarfix. Basil Rifkindrecallsbeing
The likely explanation,wrote de Lorgeril Instead, we see the laughedat once on this pointwhenhe lec-
and his colleagues, is that the "protective tured to Scottish physicians on healthy
effects[of the Mediterranean diet]werenot exact opposite." diets:"Onesaid,'Youtalkaboutincreasing
related to serum concentrationsof total, fruitsandvegetableconsumption, butin the
LDLor HDLcholesterol." -William Harlan areaI workin there'snot a single grocery
Manyresearchersfind the Lyondata store.'" In bothcases,researchers joke that
so perplexingthatthey'releft question- c the onlygreenleafyvegetablethesepopula-
ing the methodology of the trial. tions consumeregularlyis tobacco.As for
Nonetheless, says NIH's Harlan,the the purportedbenefitsof the widelyhailed
data "arevery provocative.They do Mediterranean diet, is it the fish, the olive
bringup the issue of whetherif we oil, or the freshvegetables?Afterall, says
look only at cholesterollevels we j Harvard epidemiologist Dimitrios
aren't going to miss something Trichopoulos,a native of Greece, the
very important." De Lorgerilbe- olive oil is used eitherto cook vegeta-
lieves the diet's protectiveeffect f bles or as dressingover salads. "The
comes primarilyfrom omega-3 quantityof vegetablesconsumedis al-
fatty acids, found in seed oils, most a pound [half a kilogram]a day,"
meat, cereals, green leafy vegeta- he says, "andyou cannot eat it without
bles, and fish, and from antioxidant olive oil. And we eat a lot of legumes,and
compounds,includingvitamins,traceel- we cannoteat legumeswithoutolive oil."
ements,andflavonoids.He toldSciencethat Indeed, recent data on heart disease
mostresearchers andjournalistsin the field trendsin Europesuggest that a likely ex-
areprisonersof the "cholesterolparadigm." planationfor the differencesbetweencoun-
Althoughdietaryfat and serumcholesterol tries and over time is the availability of
"are obviously connected,"he says, "the freshproduceyear-round ratherthandiffer-
connection is not a robust one" when it Ln rpretathatalso ences in fat intake.While the press often
comesto heartdisease. have con- plays up the Frenchparadox-the French
siderable have little heartdisease despiteseemingly
Dietary trade-offs amountsof fat. high saturatedfat consumption-the real
One inescapablereality is that death is a This plus- paradoxis throughoutSouthernEurope,
trade-off,and so is diet. "Youhaveto eat minus problem whereheartdiseasedeathrateshavesteadi-
something,"says epidemiologist Hugh suggests a different ly droppedwhile animal fat consumption
Tunstall Pedoe of the University of for virtuallyev-
interpretation has steadilyrisen,says Universityof Cam-
Dundee, U.K., spokespersonfor the 21- ery diet studyever done, including,for in- bridge epidemiologistJohn Powles, who
nationMonitoringCardiovascular Disease stance,the kind of metabolic-wardstudies studies nationaldisease trends.The same
Project by the World HealthOrganiza- that originallydemonstratedthe ability of trendappearsin Japan."Wehavethis idea ,
tion. "Ifyou eat moreof one thing,you eat saturatedfats to raise cholesterol. If re- that it's the Arcadianpast, the life in the 2
a lot less of somethingelse. So for every searchersreducethe amountof saturated village, the utopiathatwe've lost,"Powles |
theorysayingthis disease is causedby an fat in the test diet, they have to make up says; "thatthe really protectiveMediter- V
excess in x, you can producean alternative the calories elsewhere. Do they add ranean diet is what people ate in the |
theory saying it's a deficiency in y." It polyunsaturatedfats, for instance,or add 1950s."But that notionisn't supportedby 3
would be simple if, say, saturated fats carbohydrates?A single carbohydrateor the data:As these Mediterraneannations u

2544 30 MARCH2001 VOL291 SCIENCE www.sciencemag.org

became more affluent, says Powles, they Reaven.How this balancesout is the un- David Ludwig, director of the obesity
beganto eat proportionally moremeat and known. "It's a bitch of a question,"says clinic at Children'sHospitalBoston. "The
with it moreanimalfat.Theirheartdisease MarcHellerstein,a nutritionalbiochemist body appearsto run out of fuel." A few
rates,however,continuedto improvecom- at the Universityof California,Berkeley, hoursaftereating,hungerreturns.
pared to populations that consumed as "maybethe great public health nutrition If the theory is correct, calories from
much animal fat but had less access to questionof ourera." the kind of processed carbohydratesthat
fresh vegetables throughoutthe year. To The otherworrisomeaspectof the car- have become the staple of the American
Powles,the antifatmovementwas founded bohydratetrade-offis the possibilitythat, diet are not the same as caloriesfrom fat,
on the Puritannotionthat "somethingbad for some individuals, at least, it might protein,or complexcarbohydrates when it
had to have an evil cause, and you got a actually be easier to gain comes to controllingweight. "They may
heart attack because you did something weighton low-fat/high- cause a hormonalchange that stimulates
wrong, which was eating too much of a carbohydrateregi- hungerand leadsto overeating,"says Lud-
badthing,ratherthannot havingenoughof mens than on wig, "especially in environmentswhere
a good thing." food is abundant...."
The other salient trade-off in In 1979, 2 yearsafterMcGovern'scom-
the plus-minusproblemof human AI: mitteereleasedits DietaryGoals,Ahrens
diets is carbohydrates.When the , wroteto TheLancetdescribingwhathe
federal governmentbegan pushing \? had learnedover30 yearsof studying
low-fat diets, the scientists and |: fat and cholesterolmetabolism:"It
administrators, and virtuallyeveryone |^ : is absolutelycertainthat no one
else involved,hoped that Americans can reliablypredictwhethera
would replace fat calories with fruits change in dietary regimens
and vegetables and legumes, but it ' will haveany effect what-
didn'thappen.If nothingelse, economics -,p ~/ soever on the incidence of
workedagainstit. The food industryhas :o"y ~new events of [coronary
little incentiveto advertisenonproprietary heartdisease],norin whom."To-
items:broccoli,for instance.Instead,says day,many nutritionresearchers,acknowl-
NYU's Nestle, the great bulk of the $30- edgingthe complexityof the situation,find
billion-plusspentyearlyon food advertis- themselvessidingwithAhrens.Krauss,for
ing goes to selling carbohydratesin the instance, who chairs the AHA Dietary
guise of fast food,sodas,snacks,andcandy GuidelinesCommittee,now calls it "scien-
bars.And carbohydratesare all too often "When you don't have tifically naive"to expect that a single di-
whatAmericanseat. etary regime can be beneficial for every-
Carbohydrates are what Harvard'sWil- any real good answers body:"The'goodness'or 'badness'of any-
lett calls the flip side of the calorietrade- thingas complexas dietaryfat andits sub-
off problem.Becauseit is exceedinglydif- in this business, you typeswill ultimatelydependon the context
ficult to add pure proteinto a diet in any of the individual."
have to accept a few Giventheprovensuccessandlow cost of
quantity,a low-fat diet is, by definition,a
high-carbohydrate diet-just as a low-fat not so good ones as the cholesterol-lowering drugs,mostphysicians
cookie or low-fatyogurtare,by definition, now prescribedrugtreatmentfor patientsat
high in carbohydrates.Numerousstudies next best thing." highriskof heartdisease.The drugsreduce
now suggest that high-carbohydrate diets LDL cholesterollevelsby as muchas 30%.
can raise triglyceridelevels, create small, -Ron Krauss Diet rarelydropsLDL by morethan 10%,
dense LDL particles,and reduceHDL-a which is effectivelytrivialfor healthyindi-
combination, along with a condition viduals,althoughit maybe worththe effort
known as "insulinresistance,"that Stan- higher fat diets. One of the many factors for thoseat highriskof heartdiseasewhose
fordendocrinologistGeraldReavenhas la- that influence hungeris the glycemic in- cholesterollevelsrespondwellto it.
beled "syndrome X." Thirty percent of dex, which measures how fast carbohy- The logic underlyingpopulationwide
adult males and 10% to 15% of post- dratesarebrokendowninto simple sugars recommendations such as the latestUSDA
menopausal women have this particular and moved into the bloodstream.Foods DietaryGuidelinesis that limitingsaturat-
syndromeX profile, which is associated with the highest glycemic index are sim- ed fat intake-even if it does littleor noth-
with a several-foldincrease in heart dis- ple sugars and processed grain products ing to extendthe lives of healthyindividu-
ease risk, says Reaven,even amongthose like pasta and white rice, which cause a als and even if not all saturatedfats are
patientswhose LDL levels appearother- rapid rise in blood sugar after a meal. equallybad-might still delaytens of thou-
wise normal. Reaven and Ron Krauss, Fruits,vegetables,legumes, and even un- sands of deathseach year throughoutthe
who studies fats and lipids at Lawrence processed starches-pasta al dente, for entirecountry.Limitingtotal fat consump-
BerkeleyNational Laboratoryin Califor- instance-cause a much slower rise in tion is consideredreasonableadvice be-
nia, have shown that when men eat high- blood sugar. Researchershave hypothe- cause it's simple and easy to understand,
1 carbohydratediets their cholesterol pro- sized that eating high-glycemic index andit maylimitcalorieintake.Whetherit's
| files may shift from normalto syndrome foods increases hunger later because in- scientificallyjustifiablemay simplynot be
I X. In otherwords,the morecarbohydrates sulin overreactsto the spike in blood sug- relevant."Whenyou don't have any real
y replacesaturatedfats, the more likely the ar. "Thehigh insulin levels cause the nu- good answers in this business," says
I end resultwill be syndromeX and an in- trientsfrom the meal to get absorbedand Krauss,"you have to accept a few not so
- creasedheartdiseaserisk."Theproblemis very avidly stored away, and once they good ones as the nextbestthing."
u so clearrightnow it's almosta joke,"says are, the body can't access them," says -GARYTAUBES

www.sciencemag.org SCIENCE VOL291 30 MARCH2001 2545