Pleasecitethisarticleinpressas:M.deSousa-Uva,etal.,Trendsindiabetesincidencefrom1992to2015andprojectionsfor2024:APortuguese ContentslistsavailableatScienceDirect
Primary
Care
Diabetes
journal homepage:http://www.elsevier.com/locate/pcd
Original
research
Trends
in
diabetes
incidence
from
1992
to
2015
and
projections
for
2024:
A
Portuguese
General
Practitioner’s
Network
study
Mafalda
de
Sousa-Uva
a,∗,
L.
Antunes
a,
B.
Nunes
a,b,
A.P.
Rodrigues
a,
J.A.
Simões
c,d,
R.T.
Ribeiro
e,f,
J.M.
Boavida
g,
C.
Matias-Dias
a,baNationalHealthInstituteDoutorRicardoJorge,Lisbon,Portugal
bNationalSchoolofPublicHealth,NewUniversityofLisbon,Lisbon,Portugal cUnidadedeSaúdeFamiliarMarquêsdeMarialva,Cantanhede,Portugal dFacultyofHealthSciences,UniversityofBeiraInterior,Covilhã,Portugal
eAPDPDiabetesPortugal,EducationandResearchCenter(APDP-ERC),Lisbon,Portugal fCEDOC,ChronicDiseasesResearchCenter,NOVAMedicalSchool,Lisbon,Portugal gPortugueseDiabetesProgramme,DirectorateGeneralofHealth,Lisbon,Portugal
a
r
t
i
c
l
e
i
n
f
o
Articlehistory:
Received4February2016 Receivedinrevisedform 26April2016 Accepted13May2016 Availableonlinexxx Keywords: Diabetes Incidence Projections Portugal Epidemiology
a
b
s
t
r
a
c
t
Background:Diabetes isknownasa majorcause ofmorbidityandmortalityworldwide. Portugal isknownastheEuropeancountrywiththehighestprevalenceofthisdisease. WhilediabetesprevalencedataisupdatedannuallyinPortugal,theGeneralPractitioner’s (GP)SentinelNetworkrepresentstheonlydatasourceondiabetesincidence.Thisstudy describesthetrendsinDiabetesincidence,between1992and2015,andestimateprojections forthefutureincidenceratesinPortugaluntil2024.
Methods:Anecologicaltime-seriesstudywasconductedusingdatafromGPSentinel Net-workbetween1992and2015.FamilydoctorsreportedallnewcasesofDiabetesintheir patients’lists.AnnualtrendswereestimatedthroughPoissonregressionmodelsaswellas thefutureincidencerates(until2024),sexandagegroupstratified.Incidencerate projec-tionswereadjustedtothedistributionoftheresidentPortuguesepopulationgivenStatistics Portugalprojections.
Results:TheaverageincreaseinDiabetesincidenceratewasintotal4.29%(CI95%3.80–4.80) peryearunderstudy.Until1998–2000,theannualincidenceratewashigherinwomen,and from1998–2000to2013–2015turnouttobehigherinmen.Theincidencerateprojectedfor 2022–2024was972.77/105inhabitantsintotal,and846.74/105and1114.42/105,respectively, inwomenandmen.
Conclusions:ThisisthefirststudyinPortugaltoestimatediabetesincidencerateprojections. Thedisturbingreportedprojectionsseemrealisticifthingscontinueasinthepast.Actually, effectivepublichealthpolicieswillneedtobeundertakentominimizethisalarmingfuture scenario.
©2016PrimaryCareDiabetesEurope.PublishedbyElsevierLtd.Allrightsreserved.
∗ Correspondingauthor.
E-mailaddress:mafalda.uva@insa.min-saude.pt(M.deSousa-Uva).
http://dx.doi.org/10.1016/j.pcd.2016.05.003
Pleasecitethisarticleinpressas:M.deSousa-Uva,etal.,Trendsindiabetesincidencefrom1992to2015andprojectionsfor2024:APortuguese
1.
Introduction
DiabetesMellitus(DM)isachronicdiseasecharacterizedby high levelsofblood glucose, that developseitherfrom the destruction of insulin-producing cells (type 1 diabetes) or whenbodytissuesbecomeresistanttotheactionofinsulin, eveninastateofcirculating hyperinsulinemia(type 2 dia-betes)[1].Type1DMisanauto-immuneconditionthatusually developsinchildhoodoradolescence,whiletype2DM devel-ops in adulthood and is associated with obesity, physical inactivityandunhealthydiet.Thelatterrepresents90%ofthe globalcasesofdiabetes[2,3].
DMisknownasamajorcauseofmorbidityandmortality worldwide,beingoneofthemainresponsiblefortheglobal burdenofdisease[4].Itistheleadingcauseofblindness,
kid-neyfailureandamputationworldwide[5],andisresponsible forcausingcardiovascular problemsthatleadto50–80%of deathsinpeoplewithdiabetes[6].TheprevalenceofDMin Portugalin2014wasestimatedat13.1%(10.8%inwomenand 15.8%inmen)[7]andPortugalisknownastheEuropeanUnion countrywiththehighestprevalenceofthischronicdisease
[8].
While prevalence data is updated annually, based on the 2010PREVADIAB nationwide study [9], and taking into considerationdemographicchanges,diabetesincidencehas garnered less attention as a health indicator.The General Practitioner’s (GP) Sentinel Network (GP Sentinel Network) represents,inPortugal,theonlydatasourceondiabetes inci-dence,since1992tonowadays.
Thus,thisstudyaimedtogaininsightinto theDM inci-dence trends, between 1992 and 2015, in the Portuguese population under observation ofthe GP Sentinel Network, aswellastomakeprojectionsforthefutureincidencerates basedonthisknowledge.
2.
Methods
GP Sentinel Network is a network based on the volunteer workofGeneralPractice/FamilyMedicinedoctorswhose pro-fessionalactivityisdevelopedinprimaryhealthcarecenters of the Portuguese National Health System. These doctors report several health events, agreed each year. Between 1992and 2015(except2001and 2002),family doctorsfrom GP Sentinel Network reported all new cases of DM (both types 1 and 2) in their patients’ lists. Diabetes case def-inition used was based on Directorate-General of Health (DGS) guidelines for clinical practice [10]. From 1992 to 1995 the GP Sentinel Network covered only Portuguese mainland, but since 1996 and 1997 Madeira and Azores (Portugueseislands)began,respectively,participationinthe network.
Anecologicaltime-seriesstudywasconductedusingdata fromtheGPNetworkbetween1992and2015,withexception oftheyears2001and2002.
EveryyearGP’supdatethecompositionoftheirpatients’ lists,takingintoaccountsomeuserscharacteristicssuchas ageandgender,whichrepresentvariablesalsoconsideredin thisstudy.
The annual DM incidence rates were calculated divid-ing the observed number ofnew cases by the population undereffectiveobservationfromtheGPNetworkduringthe years understudy.Thepopulationunder effective observa-tion (PUEO) isthe annual averageofthe GP’s patients’ list weeklysum:PUEOt=
Mm=1NmItm,whereNm isthe numberofpatientscomposingtheGPmlist;andItmisavariablethat
takesthevalue1ifthedoctormwasactiveinweekt,and0 otherwise.Theywereonlyconsideredthepatients’listsfrom theGP’swhowereactiveoneachweek(whichmeansthose that reported cases, evenif theywere of zero). The popu-lationundereffectiveobservationusedasthedenominator fortheannualincidencerateistheaverageofthe PUEOin the52weeksoftheyear,namely:PUEOyear=
Tt=1PUEOt/52.
Thepopulation undereffectiveobservationvariedbetween 231,292in1991and 37,363in2015.Theincidencerate esti-matesarepresentedper100,000inhabitants,pertriennium, and disaggregated by sex and age group. The triennium 2001–2003haveonlydatafor2003,sincethenewcasesof dia-beteswerenotreportedbytheGPnetworkintheyears2001 and2002.
Poisson regression models were used to estimate the annualtrendsfrom1992to2015.Theobservedincidencerates were then adjustedforthePortuguesepopulation distribu-tion,accordingtoestimatesfromtheStatisticsPortugal[11]. Theincidencerateprojectionsfrom2015to2024,sexandage groupstratified,werealsoestimatedusingPoissonRegression modelsasproposed byHakulinen andDyba[12].Incidence rateprojectionswerethenadjustedtotheresidentPortuguese population distributiongiventheStatisticsPortugal projec-tions [13]. Prediction intervals forthe projected rates were calculatedbasedontheworkofDybaandHakulinen[14]. Pear-songoodnessoffitstatisticandresidualanalysiswereusedto assessthegoodnessofthefittedmodels.
3.
Results
3.1. Totalandsexstratifiedincidence
In the first triennium under study, 1992–1994, the annual incidenceratewas261.7/105inhabitants,andinthelast
tri-ennium,2013–2015,was630.42/105inhabitants(Fig.1).Until
1998–2000,theannualincidenceratewashigherinwomen, andfrom1998–2000to2013–2015turnouttobehigherinmen (Fig.1).
The average increase in DM incidence rate was 4.29% (CI95% 3.80–4.80) per year understudy.In menthe annual trendwasalsoofincrease(5.16%[CI95% 4.47–5.86])aswell asinwomen,althoughslighter forthelatter(4.29%[CI95% 3.80–4.80]).
3.2. Totalandsexstratifiedprojectedincidence
The incidence rate projected for 2022–2024 was 972.77/105
inhabitants in total (CI95% 864.05/105–1081.49/105), and
846.74/105 (CI95% 707.98/105–985.49/105) and 1114.42/105
(CI95% 944.13/105–1284.72/105), respectively, in women and
Pleasecitethisarticleinpressas:M.deSousa-Uva,etal.,Trendsindiabetesincidencefrom1992to2015andprojectionsfor2024:APortuguese
1992-1994 1995-1997 1998-2000 2001-2003 2004-2006 20072009- 20102012- 20132015- 20162018- 20192021- 20222024
-Women CI95% lower 578.64 642.73 707.98
Women 280.89 283.78 318.38 312.73 493.10 514.41 562.34 586.65 696.02 770.67 846.74
Women CI95% upper 813.41 898.61 985.49
Men CI95% lower 747.99 845.85 944.13
Men 240.96 264.58 331.46 415.65 552.74 604.05 741.66 678.87 893.62 1003.851114.42
Men CI95% upper 1039.26 1161.851284.72
Total CI95% lower 696.97 780.11 864.05
Total 261.65 274.53 324.69 362.37 521.80 557.40 647.91 630.42 789.57 880.74 972.77
Total CI95% upper 882.16 981.36 1081.49
0 200 400 600 800 1000 1200 1400 In cid en ce rate p er 10000 0 inhabitants
Diabetes incidence rate between 1992-2015 and projecons unl 2024
Fig.1–DiabetesMellitusannualincidencerateprogressioninthepopulationunderobservationfromGPSentinelNetwork andprojectionsfor2024,sexstratified.
3.3. Agestratifiedincidence
Asexpected,therewereobservedgreaterDMincidencerate estimatesintheoldestagegroups.Generally,inmentheage groupwithagreaterincidenceratewasthe55–64yearsold (Fig. 2)andinwomenitwasthe65–74yearsoldagegroup (Fig.3).
TherewasasignificantannualincreaseinDMincidence ratein theage groups abovethe 55 yearsold. Thegreater increasewasseeninmeninthe75+agegroup(7.05%peryear [IC95%5.52–8.60]).
3.4. Agestratifiedprojectedincidence
The higher incidence rate projected for 2022–2024 was, in women, in the age group 65–74 years old (1730.06/105
inhabitants [CI95% 1430.56–2029.56]) and, in men, in the age group 55–64 years old (2412.22/105 inhabitants [CI95%
1569.47–3254.97])(Figs.2and3).
4.
Discussion
TherewasanincreasingtrendintheDMincidenceinthetime periodunderstudy,from1992to2015.Weproposethatthis increasecanberelatedwiththreemainpossiblereasonsthat couldbeactingjointlyorindividually.Ononehand,itcould beassociatedwithanincreaseintheprevalenceofobesityin Portugal[15],asobesityisknownintheliteratureasoneofthe mainriskfactorsfortype2diabetes[5].Ontheotherhand,it couldalsoberelatedwiththeincreasinglypronounced popu-lationaging,sincetheestimatesoftheDMincidenceinthis studyarehigheratolderagesandthereisevidencethatthe overallDMincreaseswithage[16].Finally,itcouldalsobe asso-ciatedwiththechanginginDiabetesdiagnosiscriteria,in2002, whicheventuallycouldjustifyahighernumberofdiagnosed
casesinrecentyears[17].Followingtheknowndistributionof diabetesnewcasesbyagegroup,weexpectthattype2 dia-betesrepresentthelargestnumberoftheobserveddiabetes casesreportedherebytheGPsentinelnetwork,andthusalso oftheprojectedrates.
Theinversionofdiabetesincidenceestimatesamongmen and women, until 1998–2000 higher in females, and after that time period higher in males, it is as far as we know not described in the national literature, although already describedintheinternationalliterature[18].However,itseems consistent with available prevalence data, where men are moreaffectedbydiabetesthanwomen[7].Suchaneventmay beassociatedwithanincreaseinthediseasediagnosisinmen duetotheincreaseddemandforhealthcareinrecentyearsby thisgroup.Ontheotherhand,maybeduetopositivechanges ineatingbehaviorsmorepronouncedinwomeninthelast years[8,19].Itcouldbeevenbecauseofagreaterfrequencyof obesityinrecentyearsinmen[8].
InwhatconcernsDMincidencerateprojections,basedon themostlikelyscenario,weprojectedthatthenumberofnew casesofindividualswithdiagnoseddiabetesinPortugalwillbe 972.77/105habitantsby2022–2024.Thefittedmodelactually
producedestimateswhichseemedsimilartothoseobserved. Thisisthe firststudy inPortugal thatestimates projec-tionsfordiabetesincidence,sowehavenotermofcomparison nationally.However,ourresultsseemreasonablewhen com-paredtotheUnitedStates(U.S.)populationwhoseincidence projections are from 800/105 habitantsin 2008to 1500/105
habitantsin2050[20].
Theincreaseindiabetesincidenceprojectedherecouldbe largelyattributabletotwomajordemographicfactors,similar tothosepointedbyBoyleandcolleagues[20]:(1)Portuguese populationaging;and(2)increasingsizeofhigherriskgroups. Furthermore,theprojectionswereportsufferfromseveral limitations,suchasnotbeingabletodiscernbetweentype 1andtype2diabetes.Otherrelevantlimitationsincludenot
Pleasecitethisarticleinpressas:M.deSousa-Uva,etal.,Trendsindiabetesincidencefrom1992to2015andprojectionsfor2024:APortuguese 1992-1994 19951997- 19982000- 20012003- 20042006- 20072009- 20102012- 20132015- 20162018- 20192021- 20222024 -0-14 13.58 12.62 27.74 31.54 16.84 5.12 0.00 67.01 27.01 29.17 31.49 15-24 11.08 19.00 16.59 12.96 34.71 28.04 33.99 35.47 42.90 46.77 50.64 25-34 46.58 29.48 52.93 161.19 79.70 99.14 87.33 91.19 145.93 170.48 199.17 35-44 171.17 176.53 268.89 378.70 460.19 376.48 415.85 225.37 529.06 573.68 618.30 45-54 563.21 576.05 642.20 948.40 995.70 1051.11 991.64 1059.75 1288.48 1383.05 1477.62 55-64 677.39 738.87 830.74 981.79 1375.94 1590.24 2222.99 1530.07 2051.66 2231.94 2412.22 65-74 658.44 723.32 958.80 817.24 1347.43 1467.11 1674.48 1684.86 1922.41 2085.37 2248.33 75+ 337.95 436.90 499.00 478.60 950.94 956.22 1223.82 1273.50 1387.01 1521.92 1656.82 0.00 500.00 1000.00 1500.00 2000.00 2500.00 3000.00 Incidence rate per 10000 0 inhabitants
Diabetes incidence rate between 1992-2015 and projecons unl 2024, in men, per age group
Fig.2–MenDiabetesMellitusannualincidencerateprogressioninthepopulationunderobservationfromGPSentinel Networkandprojectionsfor2024,agegroupstratified.
takingintoconsiderationpossibleeffectsoflocalinitiatives concerningdiabetespreventioninPortugalandthe dropin thepopulationundereffectiveobservationfromtheGP Sen-tinelNetworksince1992tonowadays.Boththeselimitations may,eventually,reducetheprojecteddiabetesincidencerates. Thedropinthepopulationundereffectiveobservationcould beassociatedwithnewtasksattributedtoGPsmedical spe-cialtyinthelastfewyearsinPortugal,whichcouldgavethem lesstimetonotifythenewcasesofdiseaseintheirpatient lists. This is infact the reasonpointed out by the largest number of GPs who left the network. On the other hand, simultaneously,theprimarycaresystemhasbeen recently restructured in Portugal, which could also explain some
dissatisfaction with the working conditions. Additionally, manyofthedoctorswhowerepartoftheGPsnetworkwhen itwascreatedarenowpassingthrougharetirementprocess. Thisdropinthepopulationundereffectiveobservationcould eventuallyberesponsibleforareductionintheprojected dia-betesincidencerates,becausetheGPswhoremainedintheGP SentinelNetworkcouldbethoseparticularlyinterestedinthe reporteddiseases.TheotherGPswholeftthenetwork,could nothavereportedaswellinolderyears,sincetheydidnot havesomuchattentionandinterestinthehealthproblems understudyandintheGPnetwork.Ifthis“lessmotivatedGPs” continuedinthesentinelnetwork,itshouldbeobserved even-tuallylowerincidencerates,whichcouldresultinadecrease
1992-1994 1995-1997 1998-2000 2001-2003 2004-2006 2007-2009 2010-2012 2013-2015 2016-2018 2019-2021 2022-2024 0-14 8.10 18.25 9.49 33.67 10.59 10.79 12.50 27.54 18.75 19.98 21.30 15-24 10.73 15.99 13.22 0.00 44.37 0.00 34.41 0.00 25.87 28.16 30.66 25-34 43.28 45.36 56.37 31.36 57.09 40.69 121.29 71.29 79.91 86.54 93.71 35-44 190.96 171.68 183.58 200.92 270.63 269.31 212.49 169.14 269.97 283.29 297.28 45-54 530.06 436.83 508.45 600.40 762.70 730.57 658.43 745.24 845.91 892.54 939.18 55-64 744.76 677.84 802.06 576.37 1163.40 1156.66 1305.21 1190.29 1384.56 1473.73 1562.90 65-74 714.29 865.82 870.77 856.21 1157.17 1232.73 1391.75 1403.58 1527.53 1628.79 1730.06 75+ 439.50 442.16 493.07 510.79 907.59 1008.85 985.06 1142.75 1247.61 1356.23 1464.85 0 200 400 600 800 1000 1200 1400 1600 1800 2000 Inc id e n ce r at e per 100 0 00 i nh abi ta n ts
Diabetes incidence rate between 1992-2015 and projecons unl 2024, in women, per age group
Fig.3–WomenDiabetesMellitusannualincidencerateprogressioninthepopulationunderobservationfromGPSentinel Networkandprojectionsfor2024,agegroupstratified.
Pleasecitethisarticleinpressas:M.deSousa-Uva,etal.,Trendsindiabetesincidencefrom1992to2015andprojectionsfor2024:APortuguese oftheprojectionsherereported.Quitetheopposite,
unfore-seenincreasesinlifeexpectancy ortheimplementation of furtherdiabetesscreeningactionsmayincreasethenumber ofnewlydiagnoseddiabetescases.Also,changesindiagnostic criteriacouldaffectincidencerates,aswellastheincreased ordecreasedaccesstomedicalcare.
Wehaveassumedalinearincreaseindiabetesincidence likeBoyleetal.,2001[21].However,diabetesincidenceseemed torisedifferentlyuntil2003andafterthistimeperiod.Should theriseindiabetesincidenceratesbenonlinear,our projec-tionsmaybeoverestimated.Nonetheless,wearguethatthe generalizedlinearmodelusedisaconsistentandwidelyused methodusefulforafirstapproachtothissubjectinPortugal givingtheGPsSentinelNetworkdataset[12,14].
Finally,itshouldbenoticedthatthedisturbingreported projectionsseemrealisticifthecircumstancescontinueasin thepast,withnomajorPublicHealthchangesoccurring con-cerningdiabetes. Foraninstance,the projectedincreaseof olderpeoplewithdiabetesunderlinesthe needforabetter supportofsocialandhealthcareservicesinthisagegroup.
EffectivePublicHealthpolicieswillneedtobeundertaken tominimizethisalarmingfuturescenario.Actually,thisstudy couldgivesomesupportforPublicHealthgoalsdesigninorder tominimizetheimpactofdiabetesinthePortuguese popula-tioninfutureyears.
Conflict
of
interest
Theauthorsdeclarenoconflictsofinterest.
Acknowledgments
The authors thank the volunteer work of General Prac-tice/FamilyMedicinedoctorsfromtheGP SentinelNetwork whoprovidedinformationaboutincidentcasesofdiabetesin theirpatients’lists.
r
e
f
e
r
e
n
c
e
s
[1] WHO,DiabetesProgramme,WHO,Switzerland,Geneva, 2013,Availableathttp://www.who.int/diabetes/action online/basics/en/index.html.
[2] WHO,GlobalStatusReportonNCDs2010,WHO, Switzerland,Geneva,2011,Availableathttp://www.who. int/nmh/publications/ncdreport2010/en/.
[3] InternationalDiabetesFederation,IDFDiabetesAtlas, Brussels,Belgium,2014.Availableat
http://www.diabetesatlas.org/.
[4] WHO,CausesofDeath2008:DataSourcesandMethods, WHO,Switzerland,Geneva,2011,Availableat
http://www.who.int/healthinfo/globalburdendisease/ cod2008sourcesmethods.pdf.
[5] A.D.Deshpande,M.Harris-Hayes,M.Schootman, Epidemiologyofdiabetesanddiabetes-related complications,Phys.Ther.88(2008)1254–1264.
[6] E.Gakidou,L.Mallinger,J.Abbott-Klafter,R.Guerrero,S. Villalpando,R.L.Ridaura,etal.,Managementofdiabetesand associatecardiovascularriskfactorsinsevencountries;a comparisonofdatafromnationalhealthexamination surveys,Bull.WorldHealthOrg.89(2011)172–183.
[7] L.Gardete-Correia,etal.,DiabetesFactoseNúmeros, ObservatórioNacionaldaDiabetes,Portugal,Lisboa,2015, Availableathttps://www.dgs.pt/estatisticas-de-saude/ estatisticas-de-saude/publicacoes/diabetes-factos-e-numeros-7-edicao.aspx.
[8] InternationalDiabetesFederation,2015DiabetesAtlas, InternationalDiabetesFederation,Belgium,Brussels,2015, Availableathttp://www.diabetesatlas.org/
resources/2015-atlas.html.
[9] L.Gardete-Correia,J.M.Boavida,J.F.Raposo,A.C.Mesquita, C.Fona,R.Carvalho,etal.,Firstdiabetesprevalencestudyin Portugal:PREVADIABstudy,DiabeticMed.27(2010)879–881.
[10] SociedadePortuguesadeDiabetologia,NormasdaDGS, SociedadePortuguesadeDiabetologia,Portugal,Lisboa. Availableathttp://www.spd.pt/index.php/normas-da-dgs. [11] StatisticsPortugal,Dadosestatísticos,StatisticsPortugal,
Lisboa,2013,Availableathttp://www.ine.pt/xportal/ xmain?xpid=INE&xpgid=inebasedados.
[12] T.Hakulinen,T.Dyba,Precisionofincidencepredictions basedonpoissondistributedobservations,Statist.Med.13 (1994)1513–1523.
[13] StatisticsPortugal,Projectionsfrom2008to2060,Statistics Portugal,Lisbon,2009,Availableathttps://www.ine.pt/ xportal/xmain?PUBLICACOESmodo=2&PUBLICACOESpub boui=65946767&xpgid=inepublicacoes&xpid=INE. [14] T.Dyba,T.Hakulinen,Comparisonofdifferentapproaches
toincidencepredictionbasedonsimpleinterpolation techniques,Statist.Med.19(2000)1741–1752.
[15] I.Carmo,O.Santos,J.Camolas,J.Vieira,M.Carreira,L. Medina,L.Reis,J.Myatt,A.Galvão-Teles,Overweightand obesityinPortugal:nationalprevalencein2003–2005,Obes. Rev.9(2008)11–19.
[16] F.Soriguer,A.Goday,A.Bosch-Comas,E.Bordiú,A. Calle-Pascual,R.Carmena,etal.,Prevalenceofdiabetes mellitusandimpairedglucoseregulationinSpain:the Di@bet.esStudy,Diabetologia55(2012)88–93.
[17] DGS,Actualizac¸ãodosCritériosdeClassificac¸ãoe DiagnósticodaDiabetesMellitus,DGS,Portugal,Lisboa, 2002,Availableathttp://www.dgs.pt/upload/membro.id/ ficheiros/i005604.pdf.
[18] S.Wild,G.Roglic,A.Green,R.Sicree,H.King,Global prevalenceofdiabetes.Estimatesfortheyear2000and projectionsfor2030,DiabetesCare27(2004)
1047–1053.
[19] DGS,PortugalAlimentac¸ãoSaudávelemnúmeros,DGS, Portugal,Lisboa,2014,Availableathttps://www.dgs.pt/ estatisticas-de-saude/estatisticas-de-saude/publicacoes/ portugal-alimentacao-saudavel-em-numeros-2014.aspx. [20] J.P.Boyle,T.J.Thompson,E.W.Gregg,L.E.Barker,D.F.
Williamson,Projectionoftheyear2050burdenofdiabetes intheUSadultpopulation:dynamicmodelingofincidence, mortality,andprediabetesprevalence,Popul.HealthMet.8 (2010)1–12.
[21] J.P.Boyle,A.A.Honeycutt,K.M.Narayan,T.J.Hoerger,L.S. Geiss,H.Chen,etal.,Projectionofdiabetesburdenthrough 2050:impactofchangingdemographyanddisease prevalenceintheU.S.,DiabetesCare24(2001) 1936–1940.