REVI EW Open AccessWhy werestarvation diets promoted fordiabetes
in the pre-insulin period?Allan MazurAbstractIn the decade before
the discovery of insulin,the prominent American physicians
Frederick Allen and Elliott Joslinadvocated severe fasting and
undernutrition to prolong the lives of diabetic patients.Detractors
called thisstarvation dieting, and some patients did indeed starve
to death.Allen and Joslin promoted the therapy as adesperate
application of animal experimentation to clinical treatment,and
texts still describe it that way.Thisjustification was
exaggerated.The public record contains only the briefest account of
relevant animal experiments,and clinical experience at the time
provided little indication that severe undernutrition had better
outcomes thanlow carbohydrate diets then in
use.ReviewDuringthefirstthirdofthe20thcentury,
FrederickM.AllenandElliottP.
Joslinwereamongthemostpromi-nentdiabetesspecialistsinAmerica.
Joslin, theelderman, encouragedAllens earlyexperimental studies,
andthetwophysiciansrapidlyconvergedonthebeliefthatseverely
calorie-restricted dietswere the best therapy fordiabetes.
Forabriefperiod, from1915untiltheintro-ductionof insulinin1922,
theypromotedwhathavebeenpejorativelycalledstarvationdiets-dietsbasedonrepeatedfastingandprolongedundernourishmentasthemostadvancedtreatmentfordiabetesmellitus,notasa
cure, butforreliefofsymptoms andmaximumextension of life.There is
nodoubtthatfastingusually reducesglucoselevelsindiabetics,
butprolongedcalorie-restricted dietsintroduced new hazards, most
obviously death by starva-tion,
whichAllenandJoslineuphemisticallycalledinanition. It is accepted
today that calorie restriction isbeneficial
fordiabeticswhoareoverweight. But forthoseofnormalorlowerweight,
itwasalreadyunder-stoodinAllenstimethat
severecalorierestrictioncouldlessenresistancetoinfection,
andforchildrenitmight stunt growth.
Fastingandundernourishmenttherapywasabalancingact,
overseenbyaspecialistwhilethepatientremainedunderclosesupervisioninahospitalorclinic,
andpronetobackslidingbypatientswhoreturnedhome.
Thepermanentlycalorie-restricteddietwasunpleasant,
difficulttomaintain, andenervat-ing,
withtheresultthatmanypatientswithdrewfromtheregimen.
Observerswonderedifsuchprivationwastoohighapriceforamodestextensionofpoorqualitylife.
Joslincommentedretrospectively, Weliterallystarvedthechildandadult
withthefaint
hopethatsomethingnewintreatmentwouldappear...Itwasnofun to starve a
child to let him live [1].Allens most famous patient was Elizabeth
Hughes, thedaughterof CharlesEvansHughes, Governorof NewYork,
Republicancandidateforthepresidencyin1916,andfinallyChiefJusticeof
theU.S. SupremeCourt. Eli-zabethdevelopeddiabetesin1919atage11,
herheightthen4111/2, herweight75pounds. ShewastreatedinitiallybyDr.
Allenwhoputheronaweekoffastingfollowed by a dietof 500 calories
daily with one fast dayper week, bringingher weight
downto55pounds.Freedof glycosuria(sugarintheurine),
herdietwasraisedto1,250calories, exceptonfastdays,
andherweightroseabove60pounds. Atthattime,
mostchilddiabeticsdiedfromcomawithinmonthstoafewyearsof diagnosis
[2]. It is difficult to judge how much
AllensregimenprolongedElizabethslife. Shedislikedthetreatmentaswell
asDr. Allen, butsheadheredtothediet with oversight from her nurse
companion.Thoughanideal patient,
Elizabethdeterioratedser-iouslybythewinterof
1921/22whensheweighed45pounds. Her mother
pleadedwithCanadiandoctorFrederickBanting, arecent discovererof
insulin, toCorrespondence:[email protected] School of
Citizenship and Public Affairs,435 Crouse-Hinds Hall,Syracuse
University,Syracuse,NY 13244,USAMazur Nutrition Journal 2011,
10:23http://www.nutritionj.com/content/10/1/23 2011 Mazur; licensee
BioMed Central Ltd.This is an Open Access article distributed
underthe terms ofthe CreativeCommonsAttributionLicense
(http://creativecommons.org/licenses/by/2.0),
whichpermitsunrestricted use,distribution,
andreproductioninanymedium,provided theoriginal work is properly
cited.includeElizabethasatrial patient. Thiswasamarve-lous success.
Elizabethregainedweight, eventuallygraduatingfromBarnardCollege.
Shemarried, hadthreechildren, andwasactivethroughout
herlifeincivicaffairs, all thewhileoninsulin, until
shediedofpneumoniain1981attheageof73.
Fewofherfriendsknewshewasdiabetic[3].Allenknewthatcarbohydratesarethecomponentofthedietthatmostaffectsbloodglucose.
Manypracti-tionersofthetime, andearlier, prescribedananimaldietor
otherlowcarbohydratevariationtoclearglyco-suria. Criticsof
theAllen-Joslintherapyclaimedasmuchsuccess, withoutnear-total
foodprivation,
usingdietslowoncarbohydratesbutallowingmorefatcal-ories[4,5].
WhythendidAllenandJoslinfollowinghimsoseverelyrestrict
thecaloricintakeof theirpatients, nearly starving them?Allens Early
WorkFrederickMadisonAllen(1879-1964)graduatedfromtheUniversityofCaliforniaatBerkeleyandcompletedhisM.D.
therein1907. MovingtoBaltimore, hecon-tacted oneofhisfraternity
brothers, now atJohnsHop-kins, andspent
afewweeksperformingsurgeryondogs, meeting professors, and looking
for a goodresearchproblembutwithoutsuccess. MovingtoBos-ton, he
spent the years 1909 to 1912 as a teaching
fellowatHarvardMedicalSchool, conductingexperimentsoncatsanddogs,
largelyathisownexpense. Anaustereman, hedescribes
himselfinanunpublishedmemoirintheprivatecollectionof
AlfredHenderson, BethesdaMD, aslivinglikeahermit,
continuallyworkingsevendays a week.Allen produced
diabetesinhisanimals bypartialpan-createctomy,
thedegreeofdiseasedependingonhowmuchpancreaswasremoved.
AccordingtoAlfredHen-derson, a friend and biographer of Allen,
Thereresultedfrom this experimentalperiod a mass
ofmanu-scriptmaterial, allwritteninlonghandwhichappearedcapableof
considerableimprovement.
Nopublishercouldbefoundtowadethroughhiscrudemanuscript[6]. With a
subsidy fromAllens father, Harvard
Univer-sityPresspublishedthemanuscriptasStudiesConcern-ingGlycosuriaandDiabetes,
whichran1,179pages. Inthisbook,
Allengivesnoprescriptionforcaloriedepri-vationasatherapyforhumandiabetes,
nordoeshedescribeanyfelicitouseffectofstarvation
onhisanimalsubjects. Tothecontrary,
Allenoccasionallyreachesconclusionsthatseemcountertohis
laterposition, e.g.,Fat feeding is not to be feared in diabetics
[7].Promoting StarvationIn1913, theyearhisbookwaspublished,
AllenleftHarvardandwasappointedanonresident
assistantphysiciantoworkondiabetesinthenewlyestablishedHospital of
the Rockefeller Institute in NewYork [8]. Inhis memoir written
years later, Allen recalled,I quickly followed up the first clue
from my Harvardwork,
provingthatdiabetesinpartiallydepancrea-tizeddogs, which
wastoosevereto becontrolled ona customary protein-fat diet or on
any diet while theanimalswerefat,
couldbecontrolledandkeptcon-trolledbystarvingthemandthendietingsoastokeepthemthin...WithinafewmonthsIwasabletoask
for human
patients.ThefirstpublicstatementofhistreatmentthatIcanlocatewasreadatasessionof
theAmericanMedicalAssociationinJune 1914, where Allenclaims
newresults: If enough of a dogs pancreas is removed,
glyco-suriaoccursevenonmeat feeding. Insomeof thesedogs,
afewdaysoffastingwillmaketheirurinesugar-free. If feedingof
proteinandfatisthenbeguncau-tiously,
onlyenoughtomaintaintheanimalinitsthincondition,
suchdogsremainfreefromdiabetes. Hislongestexperiment todate was on
adogkeptfreefromglycosuria for six months and still alive. Allen
cautioned,If an attempt is madeto increase the weight of such
ananimal, glycosuria soon appears and must be checked
byrenewedfasting. Suchdogs, thoughvery thin, are vigor-ous and
lively [9].Allenpresents his results anecdotally,
givingnoquan-titativedata, whichwasnotunusualinpresentationsofthat
era, but he does not even report the number of ani-mals used.
Itwouldhavebeenimpossible
toevenintui-tivelyassessthereliabilityofhisresultsortheimplieddifferencesbetweenexperimentalandcontrolanimals,or
if indeed there were control animals. He regards
dogsasadequatemodelsforhumandiabetics,
andhefunc-tionallyequatesthedogssurgicallyreducedpancreaswith human
diabetes [also see
[10]].Allenaddsthathehastreatedalimitednumberofpatientsbyprolongedfastingandcalorierestriction.Theresultsobtainedindicatethatthesamemethodemployedinrenderingthediabeticdogfreeof
glyco-suriaandprolongingitslifeisefficaciousineliminatingglycosuriaandacidosisinthehumanpatient[9].
Fiveyearslater Allen published detailed case histories,
whichshowthathehadbythentreatedonlythreediabeticpatients, none
longer than four months [11].By fall, Allen had treated eight
patients, and the hospitaldirector thought his results sufficient
to justify undertakingthis work on a larger scale [12]. Allen
envisioned diabetesoccurring when a predisposed pancreas is pushed
over theedge by excessive diet. The overstrained pancreas couldno
longer process glucose adequately, causing sugar toappear in the
urine. Conversely, Allen thought anMazur Nutrition Journal 2011,
10:23http://www.nutritionj.com/content/10/1/23Page 2 of
9overstrained pancreas is rested by diminished diet, andthat this
respite may gradually strengthen its functioning.Theattempt to put
onweight, according to the time-honored traditions of diabetic
treatment, is one of the sur-est ways of bringing back all the
symptoms and sendingthe patient downhill [13].A year later,
addressing another medical group in May1915, Allen is more
emphatic. Reporting nonew experi-ments but referring generallyto
surgically-produced dia-betesinanimals,
henotesthatthemeasuresordinarilyused in human diabetics, namely
brief fasting and carbo-hydraterestriction, maybeinsufficient
tokeepdogssugarfree. Inseverecases,
theinitialfastmustsome-timesbemeasuredinweeksratherthanindays.
Thesubsequent diet must be such as to keep the animal at
alowlevelofweight...Ifglycosuriaisprevented,
theani-malsmayremainlivelyandstrongthoughthin...Thetreatmentof
[human]diabetesattheRockefellerHospi-tal has been based upon these
animal experiments [14].Allen had by this time treated 44 patients
at Rockefel-ler, chosen as the most severe cases...and representing
asufficient variety as respects age, social condition, andother
factors. He claims to have cleared up glycosuriawith an initial
fast, sometimes lasting as long as ten days.Broadlyspeaking,
freedomfromglycosuriaseemsattainable in all cases of uncomplicated
human dia-betesbeforethereisdangerofdeathfromstarva-tion...After
the fasting patient has been completelysugar-free for one or two
days, feeding is begun...andthe tolerance of the patient for
carbohydrate, protein,and fat is determined...[T]he diet is
governed...by theamount of each food that can be given in each
indivi-dual case while keeping the urine clear
[14].ItisdifficulttotellpreciselywhatthefateofAllens44patientswasfromtheclosingstatementinhisMay1915talk:
Amongthepatientstreatedthusfar, duringavariablenumberof
monthsinthehospital andathome, spontaneous downward progress has
not yet beenobserved.Mytabulationof thesecases,
whichwerelaterpublished[11], showsthatof 41patientswhosecondition
was known at the end of 1917, 46% were deador in one casenearly
dead.Allen was deeply committed to his new regimen,
ofteninterpretingcomasordeathstolapsesbypatientsfromtheirprescribeddiets.
Heclaimedrepeatedlythathistherapy was basedon his
animalexperimentation, with-outpublically
reportingtheseexperimentsin detail,
noramIawareofanyonereplicatinghisresults. AnarticleinThe
NewYorkTimes of February13, 1916,
saysAllensnewtherapywasbasedonanimal
experimentsperformedattheRockefellerInstitutebyI. KleinerandS.
Meltzer, but was not true [15].Historian Michael Bliss
characterizes Allen as astern,cold, tirelessscientist,
utterlyconvincedofthevalidityof his approach... His therapy for
diabetes seemedimmenselyhard-heartedintheextremecases,
andmetmuchresistancefromdiabetics, theirfamilies, otherphysicians,
andotherworkersattheRockefellerInsti-tute[16]. Toemphasizehispoint,
BlisscitesCase4,one of Allens first patients at Rockefeller, a
blindtwelve-year-oldboy already in bad shape when first seenin1914,
sevenyearsafteronset. Allenexperimentedwith various diets and
periods of fasting and
waspuzzledbyunaccountableglycosuriaunrelatedtotheknown food intake.
Allen remarks in his write-up,It seemed that a blind boy isolated
in a hospital roomandsoweakthathecouldscarcelyleavehisbedwould not
be able to obtain food surreptitiously
whenonlytrustworthypersonswereadmitted. It turnedout that his
supposed helplessness was the very thingthat gave himopportunities
that other personslacked.... Amongtheunusual things
eatenweretooth-paste and bird-seed, the latter being
obtainedfromthecageofacanarywhichhehadaskedfor.Alsohismotherandhisgovernessonvisitinghimsometimes
brought lunch, which was kept in a closet,supposedly without his
knowledge; nonetheless, inthe short intervals when he was
unwatched, he mana-ged to find it and remove such articles as might
notbe missed. These facts were obtained by confessionafter long and
plausible denials.This boy died after four months of treatment
[11].In his memoir, Allen evaluatesmembers of the
Rocke-fellerstaffaseitherfororagainsthim. Dr. EdgarStill-man, who
had direct care of patients while Allendevoted time to hisanimal
experiments, wasincapableanduseless,laterpartof
theoppositionthatwouldtake over controlofthe diabetes ward. Dr.
Reginald Fitz(sonof ReginaldHeberFitz,
along-timeprofessoratHarvardMedical School
whowasoneofJoslinsearlycollaborators[17]),
workedwiththehospitalsdiabeticpatientsinacompetentandagreeablemanner,
butDr. Fitz, havingacareerbeforehim,
wastoocircum-specttobealliedwithmeonthelosingsideandmadehis
friendships on the stronger
side..BlisssuggeststhatAllensdeterminationtoapplyhismethodsruthlesslyledtoadecisionbytheRockefellerInstitutetotakeawayhiscontrolofthediabetesclinic.Allensfriend,
AlfredHenderson, givesamoresympa-thetic portrait but a congruent
reason for departure:Initiallyalonerinhisworkandinhisbeliefs,
daysandnightsbecametooshortfortheexperimentalanimalworkandthemanagementofoftenseverelyMazur
Nutrition Journal 2011,
10:23http://www.nutritionj.com/content/10/1/23Page 3 of
9illpatientswhoneededcloseattentionandfrequentlaboratorytests.
Inaddition, hefounditnecessarypersonally to supervise the hospital
kitchen and diet-aryservice,
whichwerenotinclinedtotakeontheaddedburdenof[Allensdemands,
including]...thethrice-boilingofvegetablestoreducecarbohydrate,oneof
Allensinnovationstobecomeuniversallypracticed...
[The]DirectoroftheHospital...seemeddisinterestedinAllenspleasforadditionaltimeforrepeatedexperimental
confirmationof hisideas.Thingswentfrombadtoworseinhisrelationshipto
authority, and he soon found his activitiesrestricted to the animal
laboratory [6].In 1917, asthe U.S. enteredWorld WarI, the head
oftheInstitute, SimonFlexner, informedAllenthat
hewouldbeacaptaininchargeof adiabeticserviceatLakewood, NewJersey,
effectively terminating hisemployment.
Givenhischeckeredcareertothispointandthedubiousscientificbasisforhistreatment,
onemaywonderwhyhistherapywouldsoonbecomesopopular and Allen so
prominent.Joslins EnlistmentElliott ProctorJoslin(1869-1962)
wasborninsmalltown Oxford, Massachusettsinto a prosperous family
ofreligiousCongregationalists. HereceivedhisM.D. fromHarvardMedical
School. After further training,
heopenedaprivatepracticeinBostonin1898, thefirstdoctorintheU.S.
tospecializeindiabetes, andwasthefounderof
todaysJoslinDiabetesCenters.
Hisbook,TheTreatmentofDiabetesMellitus[18], wentthroughmany
editions, indicative of Joslins prominence.A man of deep faith in
religion and science, Joslin wasfrugal butimmaculateindress,
risingearly, workinglong hours, eating sparely. According to Chris
Feudtner,whostudiedJoslinslongcareer, hehadasprybody,
aserenesenseof discipline, andasternNewEnglandconscience. He
combined agentlemanly manner with ashowmansknackandapreacherszeal,
alwayslookingto spread his message of diabetic care [19].In 1915
the well-establishedJoslin
generouslycreditedtheyoungerAllenforgreat progressthat
theperiod1914-15 had seen in the treatment of diabetes
mellitus:Theadvantageofmaintainingtheurinesugar-freehasbeenuniversallyrecognized,
butall
havecon-cededthatthiswasimpossiblewithoutdangerfromacidosisandinanition.
Fastingandalowdiethavebeenknown, but it
isonlyfairtogiveAllenthecreditoffirsttoseethetherapeuticsignificanceofinanitionuponaseverecaseofdiabetes,
secondtoproveupondiabeticdogsthat
prolongedfastingwouldrenderthemsugar-free, and third tohave
thecourageofhisconvictionsandapplythisprincipletohumandiabetes.
Thanks toDr.
FrederickM.Allenwenolongernursediabetics-wetreatthem![20].Iseenoreasontodoubt
that JoslinwasgenuinelyimpressedbyAllenstherapeuticachievements,
buttherewasaswellapreexistingrelationshipbetweenthemengoingbacktoAllenstimeatHarvard.
JoslinwasanassistantprofessoratHarvardMedicalSchoolwhileAllenwas a
teachingfellow, and they knew one anotherif
onlyfortheirsharedspecializationindiabetes. Dr.AlfredHenderson,
whoknewbothmenintheirlateryears,
toldmethattheycommunicatedaboutpatientswhilebothwereatHarvard.
Also, theywerefraternitybrothers,
membersintheirrespectiveundergraduatecollegesof AlphaDeltaPhi,
thefraternityof JohnD.Rockefeller, Jr.,
andTheodoreandFranklinRoosevelt[21].
IntheprefacetohisHarvardUniversityPressbook,
AllenacknowledgeshisindebtednesstoJoslinforgoodofficesinconnectionwithpublication.Asafel-lowatHarvard,
Allenreceiveda$300grantforhisdia-betesworkfromthemedical
schoolsProctorFund,establishedbybequestof
EllenOsborneProctorforthestudy of chronic diseases. Ellen Proctor
was Elliott Proc-tor Joslins aunt, and he was instrumental in
establishingthefund[17]. Intheprefacetothe1917editionofhistext,
Joslinwrites, IcanneverrepayDr.
Allensmanyandcontinuedkindnessesorsufficientlyacknowledgetheinspirationwhichhisfruitful
andpersistentworkawakens [18].Asaconsequenceof
havingalargepracticeindia-betesbutfeweffectivetools,
Joslinhadby1916lost62patientsundertheageoffifteen,
withcomaalwaysthecauseofdeath[22]. InadoptingAllenstreatment,
firstremovingfat, thenprotein, thencarbohydratefromthediet,
Joslinforestalledhyperglycemiccrisesandketoaci-dosisthat
areknowntodaytoinducediabeticcoma[23]. Quickly noticingthe
improvement, especiallyamonghischildpatients,
Joslinreportedthegratifyingfactthatinthepastyear,
48cases[probablymostlyadults]havebeentreatedbymeinhospitalswithoutadeath
[22]. Two years later he would
comment,Whereasformerlytheprognosisforchildrenlessthantenyearsof
agewasmeasuredinmonths,
todayitisrareforachildtoliveforlessthanoneyear.. Further-more,
Joslinsawapragmaticbenefitforgeneralpracti-tioners: Thegreatest
advantage of the fasting treatmentintroducedbyDr.
Allenliesinitssimplicityandintheremoval
oftheneedforquantitativeurinaryexamina-tions [18].Hisclinical
experiencedeeplyimpressedJoslin: Formyself, I
considertheimpressionwhichI
havethusobtainedoffarmorevaluethananystatisticswhichmyMazur
Nutrition Journal 2011,
10:23http://www.nutritionj.com/content/10/1/23Page 4 of 9records
afford, but itis almostmyduty to presenttheseaswell.He
summarizedhiscasesin tables butwasnotanadeptstatisticianbythe
standardsofthetime.
(Evenhisarithmeticisincorrect.)HisTableOneshows85%ofhisnewpatientsstillalivebutonly80%ofhisfastedpatientsstillalive,
apuzzlingcomparisontoaffirmthevalueoffasting. Nonetheless,
Joslinendsonanexuber-ant note: I amcomingtofeel that
comanolongerrepresentstheculminationofthedisease, butthatitisan
avoidable accident [22].Inallhissubsequentpublications,
JoslinpraisesAllenas the primary innovator in diabetic therapy. The
secondeditionof histextbook, Treatment, openswithJoslinwriting,
Theadvanceinthetreatment of
diabetes,whichbeganwiththeintroductionoffastingbyDr. F.M. Allen,
continues, andstatisticsarenowavailabletoshowit.
So-calledacutelyfataldiabetesisdisappearingandthefirstyearofdiabetesisnolonger,
aswasonlytoo recently the case, the diabetics danger zone.
AlreadyIhavequiteaseriesofpatientswhohaveoutlivedtheirnormal
expectationoflifeattheageof onsetoftheirdiabetes. And later in the
text:Looking back upon the treatment of diabetes
beforeAllensintroductionofprolongedfasting,
itreallyseems...thatourpatientswerenursedratherthantreated. Only
those who have cared for many
patientsbytheoldermethodscanappreciatetheadvancewhich Allen has
given to diabetic therapy [18].InFebruary1916,
TheNewYorkTimesreportedasymposium held atthe New York Academy of
MedicineontheAllenPlan.WhetherornotJoslinhadahandinsettingupthisevent,
hegaveithisimprimatur. Dis-senters alsoparticipated, butthestorys
headline, Radi-calNewMethodofTreatingDiabetes,suggeststhatitwas a
platformfor Allen, whotoldhis colleagues,Manyof
ourpatientsrunuptheeightflightsof
ourstairsatthehospitaloftheinstitute twentytimesa
day.Thentheywalkeight orten milesin theopen air. Theyalsoskip the
rope and toss medicine balls. Weare mak-ingathletesofthem,
thinastheyare, andsurelyitcanat least be said of them they are not
neurasthemic [24].Thenextyear,
Joslinpublisheddatatablesbasedon1,187casesofdiabeteshesawpersonallybetween1893and1916.
Heclaimedtoshowwiththesedatathatsince1915,
whenmoderntherapywasfirstapplied,outcomeshaveimproved.
Inonesuchdisplay, hecom-pares408ultimatelyfatal
caseshesawintheperiod1893-1915with500ultimatelyfatalcaseshesawintheperiod1893-1916(i.e.,
the lattergroup containsthe408fatalitiesof
theformergroupplus92fatalitiessince1915).
Joslinemphasizesthat16.9%of the1893-1915fatalities died within one
year whereas only 14.8% of the1893-1916fatalities diedwithin
oneyear. He
concludes,Achangeinmortalityfrom16.9percentinthefirstyearofthediseaseto14.8percentdoesnotseemlarge,butthesefigures,
asdothoseforthesubsequentyears,understatethereal
improvementintreatment...Thesetables, tomymind,
presentirrefutableevidenceoftheimprovement in the treatment of
diabetes [18].Thisismuddystatisticalreasoning,
evenbythestan-dardsofhistime. First,
thetwogroupsJoslincompares(1893-1915vs. 1893-1916)
containmostlythesamecases. Second,
incombiningallhiscasesfrom1893to1915,
Joslinimplicitlyassumedthattherewasnogra-dualimprovementinthemortalityofhispatientsovertheseyears,
but elsewherehenotesthat adultspre-viouslyunsuspectedof
havingdiabeteswerebythenbeingidentifiedinexaminationsforinsurancepolicies,andhispracticeincreasinglyincludedolderpatientswhogenerallylivedlongerfromonset.
Toaddressbothoftheseconcerns, a better comparisonwouldhave
beenone-yearmortalityforpatientsnewlyseenin1915-16(afterthe
introductionofAllenstherapy)withtheone-year mortality for patients
newly seen in 1914-15.
Third,Joslinacknowledgesthatthechangeinmortalitydoesnotseemlarge.Ifhehadbeenwritingintodayslan-guage,
hemighthavesaidthe2%differenceisnotsta-tisticallysignificant.ItwouldbeunreasonabletoholdJoslinto
standards not then in place, but one can
calcu-latethatatleasta10%differencewouldbe requiredforthe Allen
treatment to show asignificant improvementin mortality at the
now-conventional level of p = .05.Poignantly, Joslin later addsa
discouraging nature [ofthe new treatment], for a new cause of death
appears inmyrecords-namelyinanition, andtothis3casessuc-cumbed.
Inanition, like coma,
mayberightlyconsideredatpresentasevidenceofunsuccessfultreatment[18].Joslin
blamed himself for these deaths.Allen Back in Civilian
LifeBythewarsend, Allenwasastarinthediabetescom-munity, championed
by Joslin, with imitators and
critics.In1919AllenssecondbookappearedasaRockefellerInstitutepublicationwithhisformerassociatesatthehospital,
Drs. StillmanandFitz, ascoauthors, thoughAllen says in his memoir
that he wrote the entire text.Again a civilian,
Allenpurchasedanabandonedestatewith a golf coursein
suburbanMorristown, New Jersey,convertingitintoa
hospitalandlaboratory. HisPhysia-tricInstituteopenedin1920,
itsGreek-derivednameindicatingtreatmentthroughnaturalmeans.
Ratesvar-iedwiththequalityofaccommodations,
whetherwardsorluxuriousrooms.
Diettherapywasindividualized,requiringalmostasmanydieticiansasnurses.
Patientswhowereablecouldrecreateonthegroundsorthenearby town. Allen
held complete control of theMazur Nutrition Journal 2011,
10:23http://www.nutritionj.com/content/10/1/23Page 5 of 9Institute,
itsmedical, administrative, andfundraisingaspects and, to the
extent he could, its patients.The Institute grew rapidly, holding
one or two month-long teachingcourses per year, drawing physicians
fromacrossthenationanddisseminatingcalorie-restrictiontherapy.
SoonyoungElizabethHugheswasbroughtfortreatment,
andherfamousfatherbecameasupporter.After1922,
bothAllenandJoslinwereamongtheearlyadoptersof insulin, but
whereasJoslinspatient-caremodel flourished, Allens operation
contracted, especiallyafterthecrashof1929.
HewasevictedfromInstituteproperty for defaulting on the mortgage
[6].The Rockefeller SeriesAllen, StillmanandFitztreatedabout
100diabeticpatients in Rockefeller Hospital between February
1914and December 1917, some admitted multiple times
foranaverageof69daysperadmission. The1919bookdescribes 76 cases,
most in considerable detail [11]. Likemany clinicians of his day,
Allen did not think statisticallybut rather in terms of individual
case outcomes. Here,abbreviated, are contrasting examples of
juvenile patients,the first having a poor outcome, the second
good:Case 51: A seven-year-old boy was first seen in 1915,
ayearafteronset, hisheightatadmission121cm, hisweight 18.3 kg. In
August 1916 the boy, weighing 17 kg.,was discharged from his second
stay at the Institute on
aprescribed700caloriedietwithaweeklyfastday. Hedied two months
later; details were not obtained.Theessential cause of trouble lay
in the home conditions ofanuneducatedPolishlaboringfamily...
[N]oobviousrecuperation or repair of the assimilative function
wasdisplayed by this child under these circumstances.Case76:
Afour-year-oldboywasfirstseenin1917,threeweeks after onset, his
heightatadmission107cm,hisweight15kg.
Hewasdischargedweighing14.4kgonaprescribed980caloriedietwithaweeklyfastday.Thisboywasstill
well in1922whenhewasswitchedfromdiettherapytoinsulin.
[T]reatmenthasbeenfol-lowedwiththeutmostfidelity...
Thisboyhasremainedconstantlyfaithful todietwhileattendingschool
andcarrying on all activities normal for his
age.Thereisatautological charactertoAllensposthocinterpretations.
Heexplainsacasewithpooroutcomeas due to dietary lapse, often the
patients or his familysownfault.
Heexplainsagoodoutcomeastheresultoffaithfullystickingtodiet.
Howcredulousof AllentobelievethatCase76,
anormallyactivelittleboyamonghisschool friends, remainedconstant
tohisSpartandiet!
Apagelaterhecomplainsofpatientssupposedlyfollowing a diet at home
[who]... seem
mosttrustworthy[but]sometimesbecomeproficientinviolationsofdietwhich
are hard to detect.LookingoverhisRockefellercases,
Allenconcludes,theacutelythreateningsymptomsofdiabeteshavebeencontrolledbythepresenttreatmentinasuc-cessfulandradicalmannerwhichbearscomparisonwith
the most powerful therapeuticmeasures for anyacuteorchronicdisease;
but thediabetesisnotcured,
anddownwardprogressoccursinpracticallyallpotentially severe
casesunless the sameprincipleof limitationof the total
metabolismandbodyweightisadequatelyobservedatall
times...Sponta-neous or inevitable downward progress has
generallybeeneitherabsentornotdemonstrableintypicalcasesofdiabetesofeventheworsttype...[Itistruethat]undulynumerousexamplesofdownwardpro-gresshaveoccurredinthepresent
series. Thesewereattributedtoblundersandmismanagementinthe
application of the principle of treatment [11].Theseriesof
casespresentedforstudyinthe1919bookbeginswithhisfirstdiabeticpatient,
seeninFeb-ruary1914,
andrunsthroughthoseseenbyJune1916,soastoinsureatleast16monthsof
observationineverypublishedcase.ForvariousreasonsthatIhaveacceptedatfacevalue,
heignoresthisscopeconditionby addingeight casesfirstseen afterJune
1916, givingatotal of76cases. Inthemaintext,
Allenreportstheirconditionasknowninlate1917. Footnotesaddedpriorto
publication update four cases to February
1918.Allennevermadeclaimsofamiraclecureandnotesforthrightlythat43%ofhis76casesweredeadbylate1917.
Heacknowledgesfourdeathsfromstarvation(inanition) under
treatment.Theresult of prolongedfastingthat soimpressedElliott
Joslin was the reduction of coma (and consequentdeath),
especiallyinchildren. Allenreportstheoccur-renceoffull orincipient
comain 21 of hiscases. Heusuallytreatedthesebyfasting,
withfourteenrecover-ing. Whileafullycomatosepatient
isunmistakable,Allenacknowledgesthatitisdifficulttojudgewhenapatient
isinincipient comaif theconditionneverdeteriorates to full coma.
Those incidents that
helabeledincipientaccountformostofhisrecoveries.Furthermorehehasnonon-fastedcontrolgroup,
sohecannot evaluate differential rates of coma with and
with-outfasting. Lackingquantitativedata,
wemighttrustJoslinsexperiencedclinicaleyeinseeingareduction incoma
with
fasting.Didseverecalorierestrictionprolonglifemorethanthesmallextensiongainedbysidesteppingaparticularcoma-and-deathevent?Thisquestionisdifficult
toaddress without a proper control group, but Allens
owndataseemtobelietheoptimismwithwhichhepro-moted the
undernourishment regime. The 44 patients ofMazur Nutrition Journal
2011, 10:23http://www.nutritionj.com/content/10/1/23Page 6 of 9whom
hespoke in 1915 are the first44 in the Rockefel-lerseries.
Heknewtheconditionof41ofthesepeopleinOctober1917,
bywhichtime46%weredead. Elevenamong the 44 were under age 20 at
first admission; 78%weredeadbyOctober1917.
(ImadethesecalculationsbycombiningdatainTableOneof Chapter II
andTable One of Chapter VII in [11].)Indeed,
whenAllenspokeaboutthe44inMay1915,he knewthat20%werealready dead.
Nineofthosestillalivehadbeenunderhiscarelessthanfivemonths.Onewondersabouthiscandorinopaquelyconcluding,Amongthepatientstreatedthusfar...,
spontaneousdownward progress has not yet been
observed.Turningtothefullseriesof76cases,
AllentabulatesmortalitybyageofadmissiontoRockefeller. MyTable1,
whichsimplifiesandupdatesAllenstabulationtoFebruary1918,
showsarelationshipalready wellknownbetweenageandmortality.
Althoughcliniciansofthetimedidnotdistinguishtypes1and2diabetes,
theyknewthatthediseasealmostinevitablymeantdeathforyoungchildren,
if notinmonthstheninafewyears.Olderpeople,
whomwenowknowhadpredominatelytype2 diabetes, could often manage
fairly well. As Joslincommented,
Inourenthusiasmfornewmethodsitshouldnot beforgottenthat
eveninthepast
goodresultswereobtainedwiththemajorityof[older]dia-betics,
andthatgradualrestrictionofcarbohydratewasthe means employed
[18].Lacking a proper control group, we cannot say whetheror not
Allens undernourished patients fared much betterthan those on a
more ample diet with restricted carbohy-drates. Allen would
emphasize that a quarter of those inthe first decade of life
remained alive. However, his twoliving patients below age ten (Case
73, age 3, admittedDecember 18, 1916; and Case 76, age 4, admitted
March9, 1917) were ad hoc additions to the series after the
June1916 cutoff date. The 3-year-old girl, under his care
tenmonths, was already weak and emaciated. The 3-year-oldboy,
hisconditiongood,wasonlyeightmonthsfromonsetofthedisease.
Thereisnotmuchlatitudehere for optimism about the
treatment.AllendidnotfirstseeElizabethHughesuntil
1919whenhewaswellsituatedatthePhysiatricInstitute, soshe is
notincludedamongthe10to 19-year-oldsintheRockefellerseries.
Fourof14patientsinheragecate-gorywerestillaliveinFebruary1918.
Wearetoldinafootnote thatone, an11-year-old girl (Case28),
initiallydescribedingoodcondition, hadarelapseandwasreferred
elsewhere for treatment(p. 292). For a 19-year-old female (Case62),
Allenpredicteddeath fromcomaor inanition unless the patientis
radically taken inhandandundernourishedfarmorerigorously.Thethird,
a12-year-oldboy(Case64), isemaciatedbutalive. Thefourth,
a15-year-oldgirl(Case66), oneyearandsevenmonthsintotreatment (two
years from onset), remainswell and is pronounced by her mother the
strongest andmost energetic member of the family. Wedo not knowif
she survived until insulin became
available.ConclusionsAtthebeginningofthe20thcentury,
physicianshadnouseful weaponagainst diabetesexcept
theabilitytoeasily measure sugar in urine. These tests became
widelyused in insurance examinations, uncoveringa
farhigherincidenceofdiabetesintheadultpopulationthanpre-viouslyrecognized.
Oncediagnosed, theusual
medicaladvicewasforobesepatientstoloseweightandforalldiabeticstoreducedietarystarchandsugar.
Thediffer-ence between types 1 and 2 diabetes was not
understooduntil the1930s, but doctorshadlongknownthat
inchildrenthediseasewasusuallyfatal
throughcoma,whileitmightbetoleratedwithlittleacutedifficultyinsome
older people.Forabrief period, from1915until thediscoveryofinsulin
in 1922, prolonged fasting and permanent
calorierestrictionwerechampionedinNorthAmericabyFre-derickAllenandElliottJoslin.
Offeringhopeforchilddiabeticsandtheirparents,
andforseverelyafflictedadults,
theAllenPlanwasquicklyandwidelyadopted.This episode is worth
examiningforthe view it givesusof thefrail
basisuponwhichnewandextremetreat-ments could become common practice
at that time.Allen developed his therapeutic ideas in 1913 at
Rock-efellerHospital, sayingtheywerebasedonhisexperi-ments with
partially pancreatectomized dogs.
HedoubtlesslydidabundantworkonanimalsatHarvardTable 1 Rockefeller
Series: Mortality Rate by Age of AdmissionAge at first admission to
Rockefeller Hospital (years) Number of patients Percent known dead
as of February 19180-9 8 75%10-19 14 71%20-29 14 64%30-39 16
38%40-49 13 31%50+ 11 18%Mazur Nutrition Journal 2011,
10:23http://www.nutritionj.com/content/10/1/23Page 7 of 9prior to
1913, but these results, published in detail,
havenobearingoncalorierestrictiontherapy.
LateranimalexperimentsatRockefellerdorelatetothetreatmentbutarepublished(inanerabeforepeerreview)onlysketchilyandimpossibletocriticallyevaluate.
Noteventhenumberofcomparisonanimalsisreported,
whichmayhavebeenasfewastwo.
IfAllendidestablishindogsareliablemodelforcalorie-restrictiontherapy,
hedidnotatthetimesupportitwithdetailedresultsinthe public
record.From1914to1917Allen,
StillmanandFitztreatedabout100patientsinthediabeticwardofRockefellerHospital.
Allen never said undernourishment wasa curefordiabetesbut
didclaimsuccessineliminatingoravoidingglycosuriaandacidosis,
mitigatingtheworstoutcomesofdiabetes,
andreinvigoratinghispatientsiftheyadheredstrictlytotheirdiets.
Inthe1919book,Allendescribes76of thesecasesindetail,
andwhilerecordingmanydeaths,
nonethelessseesthisrecordasverifyingtheefficacyofundernutrition.
Writingbeforestatisticalanalysiswasestablishedinthemedicallitera-ture,
Allen interpretedhis clinical material on a case-by-casebasis. He
reliablysaw calorierestriction astherea-sonfor successful cases,
and lapses from diet as the rea-son for
failures.Fastingandunder-nourishmentdidfreetheurineofsugarandapparentlyreducedtheincidenceof
coma-and-deathinchildren. But
childdiabeticssoondiedfrominfectionsandothercauses,
nowincludinginani-tion. Itis uncertain whethercalorierestriction
extendedlife beyond the weeks or months gained from avoiding
aparticularcomaevent. MyreanalysisoftheRockefellerseries, showing
very high mortality among children,
pro-videslittlereasontothinkthatitdid. Thisjudgmentisobviously
limited by the absence of a proper comparisongroup.In reading
Allens publications through1919, I got theimpressionthat nearly all
of the diabeticcases at Rocke-fellerHospital
from1914to1917weretreatedbyhismethod.
IwassurprisedtoreadacontraryrecollectionbyAllen,
abstractedfromthediscussionatamedicalassociation meeting in
1921:Contrary tomybeliefand wish, most
ofthediabeticpatientsintheRockefellerInstituteHospital weregiven
high calorie diets, andall the severecases thustreatedendedfatally.
Theonlytwojuvenilepatientswhoarealivetodayarethetwowhoescapedthistreatmentandthesoleknowndifferencebetweenthemandtheothersisthat
theirdietswerekeptrelativelylowinfatsandcalories.
Oneofthesetwohaslost toleranceseriouslythroughsurreptitioustakingof
fat, especiallybutter, andthisdeclineofassimilation seems to have
been halted byeliminatingthiserror... Otherclinical
experienceconvincesmefirmlythattheattempttogiveahighcaloricdietortobuildupthebodyweighttoohighwith
fator anyotherfood is injurious andleads to afatalresult in every
genuinely severe case of
diabetes[5].Itisimpossibleformetosayifthisstatementisfac-tuallycorrect,
butAllenwasadamantinrestatingthiscomparison the following
year:Unfortunatelythegreatmajorityofthese[Rockefel-ler]
casesweresubjectedtorecklessoverfeedingwithfatandtotalcalories.
Thefactthatonlysomeofthese mildercaseshavesurvived, and
allthechil-drenandyoungpersons andall theadults withseverediabetes
havediedunder this
treatment,standsasasufficientwarningagainstthismethodandservessufficientlyasacontrol
of theresultsobtainedunderthemethodof caloriclimitationwhich I have
always advocated [25].If weaccept Allensclaimthat
undernourishmentworked, itisworthaskingifitworkedanybetterthanmore
filling low-carbohydrate diets. As critics Newburghand Marsh put
the issue,Adiabeticdiet, inordertobesatisfactory,
mustbecapableofenablingthepatienttoleadamoderatelyactivelifeforanindefiniteperiod...[T]heseveredia-betic
may be kept sugar free by a sufficient
reductionofhistotalcaloricintake,
butitisfrequentlyneces-sarytoreducethetotalcaloriessomuch...thatsuchpatientssufferfromslowstarvation,
andarequiteincapableofearningalivelihoodindeedmanyofthem may be
said to merely exist [4].NewburghandMarshreporton73cases,
apparentlymostlyadults, admittedtotheUniversityofMichiganHospital,
wherethestatesphysiciansreferredtheirseverestcasesasacourtoflastappeal.Onenteringtheclinic,
patientswerenot fastedbut placedonadietof900-1,000calories,
highinfat, lowincarbohy-drate.
Afterbeingsugarfreeforoneortwoweeks,thediet
wasincreasedfrom1,400to2,500calories,dependingontheneedsof
theperson, mostcaloriescomingfromfat, leastfromcarbohydrate.
Exceptingthreepatientswhodiedwithinadayof
admission,andanotherwhowentintocomaaftereatingabagof oranges
brought byarelative, theurineof allpatientsbecamesugar free,
nonedevelopedsevereacidosis,
andtheinvestigatorsweregreatlyimpressedbytheexcellent conditionof
our patients monthsafterleavingtheclinic[4].Mazur Nutrition Journal
2011, 10:23http://www.nutritionj.com/content/10/1/23Page 8 of
9IamnotsubmittingtheMichiganstudyasdefinitive,butitfortifiedthepositionofanti-starvationphysicianswho
arguedthatcarbohydrateswere themain source
ofglucoseinurineandblood,
soitwasunnecessarytostarvepatientsaslongascarbohydratewaslimited,
andforseriouslyill childrentherewaslittlethatcouldbedone in any
case. Critics also claimed, as had Allen him-selfin1913,
thatfatfeedingisnottobefearedindia-betics
[7].WecannotsaywhetherElizabethHugheswouldhaveliveduntil
theadventofinsulinifAllenhadntnearlystarved her to death. But it
does appear, more generally,that undernourishment therapywas
promotedas atreatmentfordiabeteswithoutclearlysupportingevi-dence
from either animal or human studies.Why, then, did the treatment
and its promoter
achievesuchprominence?JoslinssupportwascriticalforAllentocarrytheday.
Thismayhavehungpartlyonthepriorrelationshipbetweenthemen.
Joslindidhavesomeobservational groundsforhisenthusiasm.
Pro-longedfastingseemedtoalleviatecomaastheproxi-matecauseofdeathinhispatheticchildpatients,
evenif it barely extended their
lives.PossiblytheselfdisciplinerequiredbytheAllenPlanappealedtothepuritanical
outlookof thetwophysi-cians. Allen andJoslin werevery hardworking,
self dis-ciplinedmen, physicallytrimandaversetoexcessorsloth.
Bothbelievedthat diabeticpatientsmust takechargeof
theirownwellbeing. Allenoftensawdete-riorations in healthashis
patientsownfault for lapsingfromtheirdiets.
Joslinproselytizedthatitwasthedia-betics
ownresponsibilitytorigorouslycontrol theirlives. Indeed,
theirideologyaboutproperlifestylemayhavebeenasimportantfortheiradvocacyasevidence-based
effects.Theentireepisodewasbrief. Theprimacyof
dietdiminishingonceinsulinbecameavailable,
leadingtotheprimacyofmedication. Weightlossapart,
thereisstilltodaynoevidence-basedconsensusonproperdia-betic diets
[26,27]).AcknowledgementsAM appreciates the critical suggestions of
Michael Bliss.The study wasfunded by the author.Authors
contributionsAM carried out all research for this article.Competing
interestsThe author declares that they have no competing
interests.Received:7 December 2010Accepted: 11 March
2011Published:11 March 2011References1. Joslin E:The diabetic.
Canadian Med Assn J 1943,48:488-497.2. Joslin E:The Treatment of
Diabetes Mellitus.2 edition.Philadelphia:Lea &Febiger;1917.3.
Cox C:The Fight to Survive New York:Kaplan;2009.4. Newburg L,Marsh
P:The use of a high fat diet in the treatment ofdiabetes
mellitus.Arch InlMed 1921,27:699-705.5. Woodyatt R:Objects and
methods of diet adjustment in diabetes.TransAssn Amer Phys
1921,36:269-292.6. Henderson A:Frederick M. Allen, M.D., and the
Psychiatric Institute atMorristown, N.J. (1920-1938).Acad Med New
Jersey Bull1970,16(4):40-49.7. Allen F:Studies Concerning
Glycosuria and Diabetes Cambridge MA:HarvardUniversity
Press;1913.8. Corner G:A History of the Rockefeller
Institute:1901-1953 New York:Rockefeller Institute Press;1964.9.
Allen F:Studies concerning diabetes.JAMA 1914,63:939-943.10. Allen
F:Investigative and scientific phases of the diabetic question.
JAMA1916,66:1525-1532.11. Allen F,Stillman E,Fitz R:TotalDietary
Regulation in the Treatment ofDiabetes New York:The Rockefeller
Institute for Medical Research;1919.12. Cole R:In Scientific
Reports to the Corporation and the Board of ScientificDirectors of
the Rockefeller Institute for Medical Research.Volume
3-6.SleepyHollow,NY:Rockefeller Foundation Archives;1914,RG 439.13.
Allen F:The treatment of diabetes.Boston Med Surg J
1915,172:241-247.14. Allen F:Prolonged fasting in diabetes.Amer J
Med Sci 1915,150:480-485.15. Metzer S:Scope of paper by Kleiner and
Meltzer.JAMA 1916,65:895.16. Bliss M:The Discovery of Insulin
Chicago:University of Chicago Press;2007.17. Harvey A:The
Interurban Clinical Club (1905-1976) No city
given:InterurbanClinical Club;1978.18. Joslin E:The Treatment of
Diabetes Mellitus Philadelphia:Lea & Febiger;1917.19. Feudtner
C:Bittersweet:Diabetes,Insulin,and the Transformation of
IllnessChapel Hill,NC:University of North Carolina Press;2003.20.
Joslin E:Present-day treatment and prognosis in diabetes.Amer J Med
Sci1915,150:485-495.21. ADP:Catalogue of the Alpha Delta
Phi,1832-1909 New York:ExecutiveCouncil of the Fraternity;1909.22.
Joslin E:The treatment of diabetes mellitus.Canad Med Assn J
1916,6:673-684.23. Kitabchi A,Murphy M,Umpierrez G,Kriesberg
R:Hyperglycemic crises inadult patients with diabetes.Diabetes Care
2006,29:2739-2748.24. Radical new method of treating diabetes. New
York Times Magazine1916.25. Allen F:Observations on the
progressiveness of diabetes.Med ClinicsNorth Amer
1922,6:465-474.26. Franz M:Evidence-based medical nutrition therapy
for diabetes.Nutr ClinPractice 2004,19:137-144.27. Sawyer L,Gale
E:Diet, delusion and diabetes.Diabetologia 2009,52:1-7,Westman
E,Feinman R,Mavropoulos J,Vernon M,Volek J,Wortman J,Yancy
W,Phinney S:Low-carbohydrate nutrition and metabolism.Amer JClin
Nutr 2007,86:276-284.doi:10.1186/1475-2891-10-23Cite this article
as: Mazur:Why werestarvation diets promoted fordiabetes in the
pre-insulin period? Nutrition Journal2011 10:23.Submit your next
manuscript to BioMed Centraland take full advantage of: Convenient
online submission Thorough peer review No space constraints or
color gure charges Immediate publication on acceptance Inclusion in
PubMed, CAS, Scopus and Google Scholar Research which is freely
available for redistributionSubmit your manuscript at
www.biomedcentral.com/submitMazur Nutrition Journal 2011,
10:23http://www.nutritionj.com/content/10/1/23Page 9 of 9
