Incidence of Diabetes Mellitus and Associated Risk Factors in the Adult Population of the Basque Country, Spain

1
Vol.:(0123456789)
Scien i ic Repo s | (2021) 11:3016 | h ps://doi.o g/10.1038/s41598-021-82548-y
www.na u e.com/scien i ic epo s
Incidence o diabe es melli us
and associa ed isk ac o s
in he adul popula ion
o he Basque coun y, Spain
Inés U u ia1,2, Alicia Ma ín‑Nie o1,3, Rosa Ma ínez1,2, J O iol Casano as‑Ma sal1,
Anibal Aguayo1,2, Juan del Olmo4, Euna e A ana1, Elsa Fe nandez‑Rubio1,3,
Luis Cas año1,2,5*, Sonia Gaz ambide1,2,3,5* & The Diabe es Epidemiology G oup*
The aim o his s udy was o es ima e he incidence o diabe es melli us in he Basque Coun y and
he isk ac o s in ol ed in he disease by eassessing an adul popula ion a e 7 yea s o ollow‑up.
In he p e ious p e alence s udy, 847 people olde han 18 yea s we e andomly selec ed om all
o e he Basque Coun y and we e in i ed o answe a medical ques ionnai e, ollowed by a physical
examina ion and an o al glucose ole ance es . In he eassessmen , he same a iables we e
collec ed and he esul ing coho comp ised 517 indi iduals o whom 43 had diabe es a baseline. The
cumula i e incidence o diabe es was 4.64% in 7 yea s and he aw incidence a e was 6.56 cases/1000
pe son‑yea s (95%CI: 4.11–9.93). Among he inciden cases, 59% we e undiagnosed. The mos
s ongly associa ed ma ke s by uni a ia e analyses we e age > 60 yea s, dyslipidaemia, p ediabe es
and insulin esis ance. We also ound associa ion wi h hype ension, obesi y, amily his o y o
diabe es and low educa ion le el. Mul i a ia e analysis adjus ed o age and sex showed ha a se o
isk ac o s assessed oge he (dyslipidaemia, wais ‑ o‑hip‑ a io and amily his o y o diabe es) had
g ea p edic i e alue (AUC‑ROC = 0.899, 95%CI: 0.846–0.953, p = 0.942), which sugges s he need o
ea ly in e en ion be o e he onse o p ediabe es.
Abb e ia ions
IFG Impai ed as ing glucose
IGT Impai ed glucose ole ance
NFG No mal as ing glucose
NGT No mal glucose ole ance
WC Wais ci cum e ence
WHR Wais - o-hip a io
Diabe es melli us is a ch onic disease wi h an inc easing equency o e he las decade ela ed o he aging popu-
la ion and li es yle changes ha p omo e obesi y1. Based on cu en da a om wo ldwide s udies, he In e na-
ional Diabe es Fede a ion (IDF) es ima es ha he global numbe o people wi h diabe es (aged 18–99yea s) will
inc ease o 693 million by 20451. This se ious pa hology has become a p io i y heal h p oblem in he wen y- i s
cen u y acco ding o he Wo ld Heal h O ganiza ion (WHO)2. I s g owing p e alence is p edic ed o inc ease
he signi ican social and economic impac esul ing om he disease associa ed complica ions2–6. Ne e heless,
p e en ion is possible when ac ing on isk ac o s and making an ea ly diagnosis and adequa e ea men ha
p e en long- e m complica ions7–9.
OPEN
1Bioc uces Bizkaia Heal h Resea ch Ins i u e, C uces Uni e si y Hospi al, Uni e si y o he Basque Coun y
UPV/EHU, Leioa, Bizkaia, Spain. 2CIBERDEM (Spanish Biomedical Resea ch Cen e in Diabe es and Associa ed
Me abolic Diso de s), CIBERER (Spanish Biomedical Resea ch Cen e in Ra e Diseases), Ins i u o de Salud Ca los III,
Mad id, Spain. 3Endoc inology and Nu i ion Depa men , C uces Uni e si y Hospi al, Osakide za, Bilbao, Bizkaia,
Spain. 4Clinical Chemis y Labo a o y, C uces Uni e si y Hospi al, Osakide za, Bilbao, Bizkaia, Spain. 5
These
au ho s con ibu ed equally: Luis Cas año and Sonia Gaz ambide. *A lis o au ho s and hei a ilia ions appea s a
he end o he pape . *email: [email p o ec ed]; [email p o ec ed]
2
Vol:.(1234567890)
Scien i ic Repo s | (2021) 11:3016 | h ps://doi.o g/10.1038/s41598-021-82548-y
www.na u e.com/scien i ic epo s/
S udies on he diabe es incidence p o ide mo e eliable in o ma ion on he isk ac o s in ol ed in he
disease p og ession compa ed o he p e alence ones. The e o e, IDF ad oca es ha accu a e incidence s udies
should be made1. Following his ecommenda ion, we ha e e-e alua ed he cu en heal h s a us o he same
coho han p e iously analysed in he diabe es p e alence s udy ca ied ou du ing 2010–2012 in he Basque
Coun y10, o de e mine he incidence o diabe es in ou popula ion and he isk ac o s ha a e in ol ed in he
p og ession o he disease.
Resul s
The eassessmen o he coho ha pa icipa ed in he p e ious p e alence s udy (baseline s udy) a e se en
yea s o ollow-up, made he de e mina ion o diabe es incidence in he Basque Coun y possible, as well as he
e alua ion o he isk ac o s in ol ed in he p og ession o he disease. The baseline cha ac e is ics o pa ici-
pan s and non-pa icipan s in he p esen e-examina ion a e shown in Table1. Excep o he obesi y (16.8% s.
23.0%, p = 0.025) and diabe es pe cen age (8.4% s. 14.3%, p = 0.010) ha a e lowe in he g oup o pa icipan s,
he equency o he es o he a iables we e simila in bo h g oups.
Among he 517 people who ag eed o pa icipa e in he eassessmen , 43 had diabe es a baseline (2010–2012)
and we e excluded om all he diabe es incidence calcula ions. Thus, he g oup a isk was es ablished wi h
474 people o whom 22 de eloped diabe es o e a pe iod o 7yea s. Among he 22 newly diagnosed diabe es
cases, 13 we e no awa e abou hei disease (unknown diabe es). All hese undiagnosed cases unde wen a 75g
OGTT: 5 cases we e diagnosed wi h diabe es by as ing blood glucose, 6 cases by glucose a 120min and 2 cases
by bo h, as ing and 120min glucose, acco ding o he WHO c i e ia11. These da a yield a cumula i e incidence
o diabe es o 4.64% in 7yea s and a aw incidence a e o 6.56 cases/1000 pe son-yea (95%CI 4.11–9.93). The
adjus ed es ima ion o age and sex s uc u e o he Spanish popula ion was o 5.37 cases/1000 pe son-yea (95%
CI 5.35–5.40).
The incidence a es o diabe es acco ding o he a iables es ed in he s udy, as well as he OR (95%CI) o
diabe es esul ed om he uni a ia e binomial eg ession logis ic analyses a e de ailed in Table2. Al hough he
incidence a e was sligh ly highe in men, he di e ence be ween sexes did no each s a is ical signi icance. The
s onges associa ion wi h he diabe es p og ession was ound om 60yea s o age onwa ds, dyslipidaemia, p e-
diabe es a baseline and insulin esis ance. A s a is ically signi ican inc ease in he incidence a e o diabe es in
indi iduals wi h hype ension, amily his o y o diabe es, low educa ion le el and obesi y was also ound. O he
analysed a iables, such as physical ac i i y o smoking, ailed o show any associa ion wi h diabe es de elopmen .
The esul s o he mul i a ia e binomial logis ic eg ession analyses adjus ed o sex and age a e shown in
Table3. Two di e en models we e p oposed o p edic diabe es and we e based on he inclusion o no o
a iables ela ed o he p ediabe es symp oms a baseline. In model A, wais - o-hip a io and amily his o y o
diabe es we e independen ly associa ed wi h diabe es, oge he wi h low HDL-choles e ol and p ediabe es a
baseline. These wo las a iables we e he s onges diabe es p og ession p edic o s in his i s p oposal. Model
B was buil wi h he same a iables han he i s one excep o hose wi h p ediabe es and/o insulin esis ance.
In his second p oposal, low HDL-choles e ol was he s onges p edic o and oge he wi h high iglyce ides
le el, wais - o-hip a io and amily his o y o diabe es, was associa ed wi h diabe es. The a ea unde he ecei e -
ope a ing-cha ac e is ic cu e o diabe es was 0.927 (95% CI: 0.891–0.963) in model A and 0.899 (95% CI:
0.848–0.953) in model B. Bo h models had adequa e goodness-o - i (p = 0.613 and p = 0.942, espec i ely).
Table 1. Baseline cha ac e is ics o he popula ion unde s udy (2010–2012). Compa ison be ween non-
pa icipan s and pa icipan s in he ollow-up. Age is shown as median (IQR: P25–P75). Res o a iables
is shown as n/ o al (%) whe e o al is he numbe o non-pa icipan s o pa icipan s o each a iable, as
app op ia e. Family his o y o diabe es de ined as 1s o 2nd deg ee ela i es wi h diabe es. De ini ions o he
es o ca dio ascula isk ac o s a e de ailed in me hods. a Pea son’s Chi-squa e es . b Mann-Whi ney U- es .
Non-pa icipan s Pa icipan s p- alue
Sex (men) 152/330 (46.1%) 226/517 (43.7%) 0.549a
Age (yea s) 52.7 (37.6–66.6) 52.3 (40.4–63.5) 0.843b
Obesi y 77/330 (23.3%) 87/517 (16.8%) 0.025a
Abdominal obesi y (WC) 181/330 (54.8%) 265/516 (51.4%) 0.357a
Abdominal obesi y (WHR) 81/330 (24.5%) 111/516 (21.5%) 0.345a
Hype ension 157/330 (47.6%) 223/517 (43.1%) 0.231a
High iglyce ides le el 48/320 (15.0%) 98/510 (19.2%) 0.145a
Low HDL-choles e ol 47/319 (14.7%) 68/507 (13.4%) 0.667a
High LDL-choles e ol 85/316 (26.9%) 154/502 (30.7%) 0.281a
Family his o y o diabe es 126/305 (41.3%) 222/493 (45.0%) 0.339a
Educa ion le el (p ima y o lowe ) 115/330 (34.8%) 166/517 (32.1%) 0.453a
Regula exe cise (a leas once a week) 155/330 (47.0%) 265/517 (51.3%) 0.252a
Smoke (> 1 ciga e e/day) 80/330 (24.2%) 121/517 (23.4%) 0.844a
No moglycaemia 243/322 (75.5%) 411/513 (80.1%) 0.133a
P ediabe es 33/322 (10.2%) 59/513 (11.5%) 0.653a
Diabe es 46/322 (14.3%) 43/513 (8.4%) 0.010a
3
Vol.:(0123456789)
Scien i ic Repo s | (2021) 11:3016 | h ps://doi.o g/10.1038/s41598-021-82548-y
www.na u e.com/scien i ic epo s/
Table 2. Diabe es incidence a es and OR (95% CI) o diabe es acco ding o possible isk ac o s. People a
isk we e hose wi hou diabe es in he p e alence s udy (2010–2012). Family his o y o diabe es de ined as
1s o 2nd deg ee ela i es wi h diabe es. De ini ions o he es o ca dio ascula isk ac o s a e de ailed in
me hods. OR (95%CI) we e calcula ed o each a iable by uni a ia e binomial logis ic eg ession analyses.
People a - isk (n) New DM cases (n) Pe son-yea (n)
Incidence a e/1000
pe son-yea s
(95%CI) OR (95% CI) p- alue
Whole coho 474 22 3356 6.56 (4.11–9.93) _ _
Sex
Women 282 12 1997 6.01 (3.11–10.50) 1
Men 192 10 1359 7.36 (3.53–13.53) 1.25 (0.53–2.89) 0.611
Age (yea s)
18–45 114 1 807 1.24 (0.03–6.90) 1
46–60 150 4 1062 3.77 (1.03–9.64) 2.32 (0.42–23.41) 0.348
61–75 150 12 1062 11.30 (5.84–19.74) 6.83 (1.61–63.36) 0.006
≥ 76 60 5 425 11.77 (3.82–27.47) 7.5 (1.45–74.4) 0.015
BMI (Kg/m2)
< 25 164 4 1161 3.44 (0.94–8.82) 1
25–30 207 7 1466 4.78 (1.92–9.84) 1.33 (0.42–4.78) 0.631
≥ 30 100 10 708 14.12 (6.77–25.98) 4.14 (1.40–14.33) 0.010
Abdominal obesi y (WC)
No 155 4 1097 3.64 (0.99–9.33) 1
Yes 232 16 1643 9.74 (5.57–15.82) 2.57 (0.96–8.44) 0.062
Abdominal obesi y (WHR)
No 270 6 1912 3.14 (1.15–6.83) 1
Yes 117 14 828 16.90 (9.24–28.36) 5.7 (2.28–15.84) < 0.001
Hype ension
No 217 5 1536 3.25 (1.06–7.59) 1
Yes 185 16 1310 12.22 (6.98–19.84) 3.76 (1.49–11.10) 0.005
High iglyce ides le el
No 330 10 2336 4.28 (2.05–7.87) 1
Yes 56 10 396 25.22 (12.09–46.38) 6.89 (2.75–17.32) < 0.001
Low HDL-choles e ol
No 273 3 1933 1.55 (0.32–4.54) 1
Yes 126 18 892 20.18 (11.96–31.89) 13.18 (4.60–50.70) < 0.001
High LDL-choles e ol
No 252 5 1784 2.80 (0.91–6.54) 1
Yes 147 16 1041 15.37 (8.79–24.97) 5.65 (2.22–16.69) < 0.001
Glucose a baseline
NFG and/o NGT 415 10 2938 3.40 (1.63–6.26) 1
Isola ed IGT 37 6 262 22.90 (8.41–49.85) 7.97 (2.67–22.29) < 0.001
Isola ed IFG 19 4 135 29.74 (8.10–76.13) 11.21 (3.04–36.41) < 0.001
IFG and IGT 3 2 21 94.16 (11.4–340.15) 64.37 (7.89–750.91) < 0.001
Insulin esis ance (HOMA-IR index)
No 279 5 1975 2.53 (0.82–5.91) 1
Yes 106 15 750 19.99 (11.19–32.97) 8.45 (3.28–25.24) < 0.001
Family his o y o diabe es
No 192 5 1359 3.68 (1.19–8.58) 1
Yes 191 15 1352 11.09 (6.21–18.30) 2.99 (1.17–8.89) 0.021
Educa ion le el (p ima y o lowe )
No 273 8 1933 4.14 (1.79–8.16) 1
Yes 114 12 807 14.87 (7.68–25.97) 3.81 (1.56–9.72) 0.003
Regula exe cise (a leas once a week)
Yes 217 9 1536 5.86 (2.68–11.12) 1
No 170 11 1204 9.14 (4.56–16.35) 1.58 (0.65–3.92) 0.307
Smoke (> 1ciga e e/day)
Non-smoke 159 8 1126 7.11 (3.07–14.00) 1
Ex-smoke 163 10 1154 8.67 (4.16–15.94) 1.22 (0.48–3.18) 0.676
Smoke 65 2 460 4.35 (0.53–15.70) 0.70 (0.13–2.63) 0.620
4
Vol:.(1234567890)
Scien i ic Repo s | (2021) 11:3016 | h ps://doi.o g/10.1038/s41598-021-82548-y
www.na u e.com/scien i ic epo s/
Apa om he 22 new cases o diabe es, he e we e 30 new cases o p ediabe es, meaning ha o e all, 52
people had some glucose me abolism diso de a e he ollow-up pe iod. In con as , among he 102 cases wi h
any glucose me abolism dis u bance in he p e ious p e alence s udy, 36 showed an imp o emen a e he eas-
sessmen : 31 people had no moglycaemia and i e cases wi h diabe es had imp o ed o p ediabe es.
Discussion
The cu en s udy, based on a ep esen a i e sample o he Basque Coun y, es ima es an incidence a e o dia-
be es o 6.56 cases/1000 pe son-yea . The single p e ious incidence s udy on diabe es in gene al popula ion o
he Basque Coun y was ca ied ou 20yea s ago12, when a coho o 594 people wi hou diabe es a baseline
(> 30yea s) was eassessed a e a ollow-up o 10yea s. The es ima ed aw incidence a e was 7.0 cases/1000
pe son-yea , a li le highe han he cu en es ima ion. Conside ing ha in he p esen analysis he only di -
e ence be ween pa icipan s and non-pa icipan s is he lowe pe cen age o obesi y among pa icipan s (23%
s. 17%, p = 0.026) and, as obesi y is a majo isk ac o in he de elopmen o diabe es, we could hink ha
he cu en incidence may be unde es ima ed and hen, assume ha he incidence o diabe es in ou a ea has
emained s able o 20yea s.
I should be no ed he di icul y o compa e he incidence es ima ions om di e en a eas due o he a y-
ing p ocedu es among s udies. Howe e , he p esen s udy ocused on a egion in no he n Spain has ollowed
he same me hodology han he ecen ly published na ionwide s udy, which ole a es alid compa isons. Ou
es ima ed incidence o diabe es in he Basque Coun y adjus ed o he age and sex s uc u e o he Spanish
popula ion is 5.37 cases/1000 pe son-yea s (95%CI 5.35–5.40), lowe han he ecen ly published da a o Spain13,
11.6 cases/1000 pe son-yea s (95%CI 11.1–12.1). The e a e ew da a abou he incidence o diabe es in gene al
popula ion o di e en egions in Spain. Despi e small di e ences in he me hodology, we could say ha he
incidence a e in ou coho , 6.56 cases/1000 pe son-yea , is lowe han ha epo ed in a neighbo ing egion o
no he n Spain14, 10.8 cases/1000 pe son-yea s, and no iceably lowe han he es ima e o he sou he n Spain15,
19.1 cases/1000 pe son-yea s. This ac con i ms a he e ogeneous dis ibu ion o diabe es in he coun y16 and
may be ela ed o he p e alence o obesi y which is almos hal in he Basque Coun y10 compa ed o sou he n
Spain17, 19% e sus 37%, espec i ely.
Unlike many published s udies in which he incidence o diabe es is es ima ed om ques ionnai es o by
e iewing clinical eco ds18,19, in ou s udy, diabe es was diagnosed in gene al popula ion on basis o an OGTT
o wi h a as ing glucose measu emen . This p ocedu e, which was ca ied ou bo h in he baseline s udy and
in he ollow-up, allows a mo e accu a e es ima ion o he incidence o diabe es, since i also includes cases o
unknown diabe es. In his su ey, mo e han hal o he new diabe es cases, 13/22, we e no awa e abou hei
disease. This is a dis u bing igu e ha has no imp o ed since he p e alence s udy in which he unknown dia-
be es igu e accoun ed o 41% o he diabe es cases10. The IDF wa ns ha he isk o de eloping ca dio ascula
e en s is highe in hese people no ecei ing nei he ea men no p e en i e measu es20 he eby, i is o g ea
impo ance o pe o m egula diabe es sc eening in gene al popula ion o educe he numbe o unknown cases.
As expec ed, se e al ca dio ascula isk ac o s such as he age om 60 onwa ds, dyslipidaemia, hype en-
sion, obesi y and al e ed WHR, a e associa ed wi h diabe es in he uni a ia e analyses. Howe e , male sex which
is widely de ined as a isk ac o o de eloping diabe es, does no each s a is ical signi icance in ou s udy,
p obably because o he limi ed numbe o new cases de ec ed in ou coho . The s onges associa ion wi h he
Table 3. Mul i a ia e binomial logis ic eg ession models o assess he con ibu ion o di e en a iables in
diabe es p edic ion. The s a ing model A included all a iables wi h p < 0.2 in he uni a ia e analysis. In model
B, he same a iables we e included, excep o he p e-diabe es s a us a baseline and he HOMA-IR index.
The wo models we e buil manually. A each s ep o he modelling p ocess, we epea edly emo ed he a iable
wi h he highes p- alue and we es ed he model again un il all emaining a iables we e signi ican (p < 0.05).
Age and sex, ega dless o hei s a is ical signi icance, we e kep in he inal model. Family his o y o diabe es
de ined as 1s o 2nd deg ee ela i es wi h diabe es. De ini ions o he es o ca dio ascula isk ac o s a e
de ailed in me hods.
Model A OR (95% CI) p- alue
Sex (Men) 2.02 (0.65–6.43) 0.224
Age 1.01 (0.97–1.06) 0.660
Abdominal obesi y (WHR) 3.93 (1.18–14.20) 0.025
Low HDL-choles e ol 6.66 (2.10–27.18) < 0.001
Family his o y o diabe es 3.61 (1.23–12.47) 0.019
P e-diabe es 7.04 (2.52–20.59) < 0.001
Model B OR (95% CI) p- alue
Sex (Men) 2.68 (0.91–8.26 0.074
Age 1.03 (0.99–1.07) 0.198
Abdominal obesi y (WHR) 3.67 (1.19–12.27) 0.024
High iglyce ides le el 3.55 (1.29–9.77) 0.014
Low HDL choles e ol 7.08 (2.25–28.82) < 0.001
Family his o y o diabe es 3.01 (1.09–9.48) 0.033
5
Vol.:(0123456789)
Scien i ic Repo s | (2021) 11:3016 | h ps://doi.o g/10.1038/s41598-021-82548-y
www.na u e.com/scien i ic epo s/
p og ession o he disease is ound wi h he p ediabe es s a us a baseline which is g ea e in isola ed cases o
IFG han in isola ed cases o IGT, as desc ibed in o he epo s14,21. This is no an unexpec ed esul conside ing
ha he me abolic abno mali ies unde lying bo h diso de s a e di e en 22. Indi iduals wi h isola ed IFG and
isola ed IGT show simila impai men s in insulin ac ion, bu hose wi h isola ed IFG ha e a mo e p onounced
de ec in ea ly insulin sec e ion and, addi ionally, ha e clea abno mali ies in hepa ic glucose egula ion. Mo e
se e e me abolic abno mali ies a e p esen in indi iduals wi h combined IFG and IGT, being he mos s ongly
associa ed diso de wi h he p og ession o diabe es in ou coho .
All es ed a iables in he uni a ia e analyses we e used o build wo mul i a ia e models. The model includ-
ing a iables ela ed o he symp oms o p ediabe es a baseline has he highes p edic i e capaci y, bu he o he
has a high p edic i e alue oo. This second model suppo s he impo ance o abdominal obesi y, dyslipidaemia
and he gene ic componen in he e iopa hogenesis o he disease and sugges s he need o ea ly in e en ion
be o e he onse o p ediabe es o p e en he p og ession o he pa hology.
Specially wo h men ioning is ha in ou coho , 30% o people wi h some glucose me abolism diso de
a baseline e e ed o no moglycaemia a e 7yea s o ollow-up. This igu e is simila o ha ound in o he
s udies8,23. This pe cen age o e e sion could be due o he less s ic c i e ia used in epidemiology o he diagno-
sis o diabe es, which esul s in alse posi i es. Howe e , i should no be uled ou ha weigh loss may con ibu e
o imp o ing glucose me abolism, highligh ing he po en ial o non-pha macological p e en ion s a egies9,24,25.
I could be conside ed a limi a ion o he s udy ha 33% o people o he o iginal coho we e los o ollow-up.
Howe e , he pa icipa ion o hose con ac ed eached 85% and ew di e ences we e ound be ween pa icipan s
and non-pa icipan s, minimizing any possible selec ion bias. Addi ionally, only 19% o pa icipan s e used he
blood analysis and comple ed exclusi ely he basic ques ionnai e o diagnosis. The main s eng hs o ou s udy
a e, i s , he me hodology ha has allowed alid compa isons wi h ecen ly published da a a na ional le el, and
second, he sampling. Indeed, his is he i s s udy ca ied ou on he gene al popula ion in di e en geog aphi-
cal a eas o he Basque Coun y and he e o e has a mo e accu a e es ima e o he incidence o diabe es in he
Basque popula ion, including also undiagnosed cases.
In summa y, he incidence a e o diabe es in he Basque Coun y emains s able o e ime and lowe han
hose epo ed om o he Spanish egions p obably due o he lowe pe cen age o obesi y. Fu he mo e, ou
esul s show ha he c i e ia o iden i ying he popula ion a high isk o de eloping diabe es should no be
based exclusi ely on he p esence o p ediabe es, bu he p edic i e alue o a se o isk ac o s assessed oge he ,
such as dyslipidaemia, wais - o-hip a io and amily his o y o diabe es suppo enough e idence o app oach a
li es yle in e en ion ha a oids he p og ession o he disease.
Pa ien s and me hods
Popula ion. A c oss-sec ional, clus e sampling design s udy was ca ied ou du ing 2010–2012 o es ima e
he p e alence o diabe es in he Basque Coun y10. The Basque Coun y is a egion o 7234km2 in no he n
Spain wi h a popula ion o 2,178,048 inhabi an s, mos ly Caucasian. Adul s (≥ 18yea s) we e andomly selec ed
om 20 public heal h cen es ep esen a i e o he Basque Coun y. A o al o 847 people comple ed he baseline
s udy. This coho was eassessed in 2018–2019 a e 7.1yea s o ollow-up (IQR 6.5–7.6). All indi iduals who
ook pa in he baseline s udy (2010–2012) we e in i ed o a end ano he clinical examina ion by le e and
phone call o assis o medical cen e. Among he 847 ini ial cases, 238 (loss o ollow-up a e 28%) we e no
a ailable (did no answe he phone, mo ed o ano he esidence ou side he a ea o in e es o died), ano he
25 we e excluded due o s udy p o ocol (hospi alized, se e e disease, su gical in e en ion and p egnancy o
ecen ly deli e ed women) and 67 e used o pa icipa e. The esul ing coho consis ed o 517 indi iduals (pa -
icipa ion a e 85%) o whom 43 had diabe es a baseline, 367 (71%) unde wen a 75g OGTT, 52 (10%) accep ed
a as ing blood analysis and 98 (19%) only comple ed a basic ques ionnai e (Fig.1).
The esea ch was ca ied ou in acco dance wi h he Decla a ion o Helsinki (2008) o he Wo ld Medical
Associa ion. The s udy was app o ed by ou local e hic commi ee, CEIC-E (Comi é E ico de In es igación
Clínica de Euskadi) and in o med consen was ob ained om all he pa icipan s.
P ocedu es. The same me hodology was used o bo h he p e alence and he eassessmen s udy. In sum-
ma y, all pa icipan s we e in i ed o a end a medical examina ion a hei heal h cen e wi h a nu se specially
ained o his p ojec . Pa icipan s we e equi ed o answe a ques ionnai e on pe sonal and amily medical
his o y and li es yle. Physical examina ion included blood p essu e (Hem-703 CP, Om on, Ba celona, Spain),
weigh , heigh , wais and hip ci cum e ences measu emen s and, inally, a 75g OGTT was also pe o med. The
OGTT was no comple ed in people wi h capilla y blood glucose measu emen s > 9.9mmol/L (One Touch
Selec Plus, Li escan, Johnson & Johnson, S.A., Mad id, Spain), o in indi iduals who had p e ious diabe es.
When an OGTT was no possible, a as ing glucose analysis was ca ied ou . People who did no wan o ully
pa icipa e in he s udy we e eques ed o answe a basic ques ionnai e in o de o collec in o ma ion abou
pha macological ea men ( o de e mine he exis ence o clinical diabe es, hype ension o dyslipidaemia in
ea men ), he kind o die hey we e ollowing and sel - epo ed weigh .
All samples we e cen i uged insi u 15min a e each blood ex ac ion and anspo ed daily o biochemi-
cal pa ame e s analyses o he Clinical Biochemis y Labo a o y o he C uces Uni e si y Hospi al. Glucose,
HDL-choles e ol and iglyce ides we e analysed in he ADVIA 2400 Chemis y analyse (Siemens Heal hca e
Diagnos ics Inc, Dee ield, IL, USA), acco ding o he co esponding eagen p o ocols. The LDL-choles e ol was
calcula ed by F iedewald equa ion. Insulin was measu ed by chemiluminescence immunoassay in he LIAISON
analyse (DiaSo in, I aly).

6
Vol:.(1234567890)
Scien i ic Repo s | (2021) 11:3016 | h ps://doi.o g/10.1038/s41598-021-82548-y
www.na u e.com/scien i ic epo s/
De ini ion o ca dio ascula isk ac o s. Diabe es and o he ca ego ies o glucose me abolism diso -
de s we e classi ied acco ding o he WHO c i e ia11. Diagnos ic alues o blood enous as ing glucose: NFG
(No mal as ing glucose) < 6.1mmol/L; IFG (Impai ed as ing glucose) 6.1–7.0mmol/L, Diabe es ≥ 7.0mmol/L.
Diagnos ic alues o blood enous glucose concen a ion a 120min (75g OGTT): NGT (No mal glucose ole -
ance) < 7.8mmol/L; IGT (Impai ed glucose ole ance) 7.8–11.1mmol/L; Diabe es ≥ 11.1mmol/L. People wi h
isola ed IFG o IGT and hose wi h bo h diso de s (IFG + IGT) we e conside ed o ha e p ediabe es. Insulin
esis ance was de ined as homeos asis model assessmen o insulin esis ance (HOMA-IR) highe han h ee26.
The ollowing a iables, equally conside ed in baseline s udy, we e also analysed: obesi y de ined as
BMI ≥ 30kg/m2; abdominal obesi y de ined as wais ci cum e ence (WC ≥ 94cm in men and ≥ 80cm in
women) o wais - o-hip- a io (WHR > 1 in men and > 0.85 in women); Hype ension de ined as sys olic p es-
su e ≥ 140mmHg and/o dias olic p essu e ≥ 90mmHg o pa ien unde an ihype ensi e ea men ; High LDL-
choles e ol de ined as alues ≥ 3.9mmol/L o pa ien unde lipid-lowe ing ea men ; High T iglyce ides le el
de ined as alues ≥ 1.7mmol/L; Low HDL-choles e ol de ined as alues < 1.03mmol/L in men and < 1.29mmol/L
in women27.
S a is ical analysis. Quali a i e a iables we e desc ibed as equencies and pe cen ages, and non-pa a-
me ic quan i a i e a iables as median and in e qua ile ange (IQR: P25–P75). Fo compa isons, Chi-squa e
Figu e1. Pa icipa ion low cha .
7
Vol.:(0123456789)
Scien i ic Repo s | (2021) 11:3016 | h ps://doi.o g/10.1038/s41598-021-82548-y
www.na u e.com/scien i ic epo s/
es and Mann–Whi ney U- es we e applied as app op ia e in each case. Pa icipan s wi h missing da a in he
a iable o in e es we e excluded o s a is ical analysis.
A uni a ia e binomial logis ic eg ession model was pe o med o e alua e he associa ion be ween se e al
a iables and diabe es. Va iables wi h clinical ele ance and p- alues < 0.2 we e hen included in a mul i a ia e
binomial logis ic eg ession model o assess he con ibu ion o di e en a iables in diabe es p edic ion. Two
models we e es ed. The s a ing model A included all a iables wi h p < 0.2 in he uni a ia e analysis. In he
model B, he same a iables we e included, excep o he p e-diabe es s a us a baseline and he HOMA IR
index. The wo models we e buil manually. A each s ep o he modelling p ocess, we epea edly elimina ed he
a iable wi h he highes p- alue and we es ed he model again un il all emaining a iables we e signi ican
(p < 0.05). Age and sex we e kep in he inal model ega dless o hei s a is ical signi icance because hey a e
ecognized p edic o s o he de elopmen o diabe es. Resul s we e exp essed as odds a io (OR) and 95% con i-
den in e al (95% CI). Fi h’s me hod was applied in o de o educe he bias o maximum likelihood es ima es
due o a e e en s.
To compa e he disc imina ion powe o he wo models (abili y o he model o sepa a e indi iduals who
will de elop diabe es om hose who will no ) he a ea unde he ecei e ope a ing cha ac e is ic (AUC–ROC)
cu es and hei 95% con idence in e als (95%CI) we e calcula ed using s anda d echniques. The model wi h
highe AUC–ROC cu e is conside ed o ha e be e disc imina ion. The Hosme –Lemeshow es was assessed
o desc ibe he deg ee o adjus men be ween he p edic ions es ima ed by he model and he obse ed esul s
(goodness-o - i ). The es indica es a good i i he esul is no signi ican , indica ing ha he e a e no s a is i-
cally signi ican di e ences be ween he obse ed and expec ed alues.
The incidence a es o diabe es in he coho and acco ding o he s udied a iables we e calcula ed as numbe s
o e en s/pe son- ime a isk o diabe es (people a isk we e hose wi hou diabe es in he p e ious p e alence
s udy). A cons an incidence o e ime was assumed and was exp essed pe 1000 pe son-yea s. To enable com-
pa isons wi h he na ional s udy, he es ima ion o he popula ion incidence was adjus ed o sex and age by
di ec me hod using he Spanish popula ion as e e ence.
S a is ical analyses we e pe o med using R so wa e e sion 4.0.1 (R Founda ion o S a is ical Compu ing).
Resul s we e conside ed s a is ically signi ican when p < 0.05.
Recei ed: 25 June 2020; Accep ed: 15 Janua y 2021
Re e ences
1. Cho, N. H. e al. IDF diabe es A las: global es ima es o diabe es p e alence o 2017 and p ojec ions o 2045. Diabe es Res. Clin.
P ac . 138, 271–281 (2018).
2. Wo ld Heal h O ganiza ion (WHO) Expe Commi ee. Global Repo on Diabe es (Wo ld Heal h O ganiza ion, Gene a, 2016).
3. Sa wa , N. e al. Diabe es melli us, as ing blood glucose concen a ion, and isk o ascula disease: a collabo a i e me a-analysis
o 102 p ospec i e s udies. Lance 375, 2215–2222 (2010).
4. Dall, T. M. e al. The economic bu den o ele a ed blood glucose le els in 2012: diagnosed and undiagnosed diabe es, ges a ional
diabe es melli us, and p ediabe es. Diabe es Ca e 37, 3172–3179 (2014).
5. Bhupa hi aju, S. N. & Hu, F. B. Epidemiology o obesi y and diabe es and hei ca dio ascula complica ions. Ci c. Res. 118,
1723–1735 (2016).
6. Seng, J. J. B. e al. Di e en ial heal h ca e use, diabe es- ela ed complica ions, and mo ali y among i e unique classes o pa ien s
wi h ype 2 diabe es in Singapo e: a la en class analysis o 71,125 pa ien s. Diabe es Ca e 43, 1048–1056 (2020).
7. Saga a, R., Cos a, B., Cab é, J. J., Solá-Mo ales, O. & Ba io, F. Li es yle in e en ions o diabe es melli us ype 2 p e en ion. Re .
Clín. Española 214, 59–68 (2014).
8. Campbell, M. D. e al. Bene i o li es yle-based T2DM p e en ion is in luenced by p ediabe es pheno ype. Na . Re . Endoc inol.
16, 395–400 (2020).
9. Tuomileh o, J. e al. P e en ion o ype 2 diabe es melli us by changes in li es yle among subjec s wi h impai ed glucose ole ance.
N. Engl. J. Med. 344, 1343–1350 (2001).
10. Aguayo, A. e al. P e alence o diabe es melli us and impai ed glucose me abolism in he adul popula ion o he Basque Coun y
Spain. Diabe . Med. 34, 662–666 (2016).
11. Albe i, K. G. M. M. & Zimme , P. Z. De ini ion, diagnosis and classi ica ion o diabe es melli us and i s complica ions. Pa 1:
diagnosis and classi ica ion o diabe es melli us p o isional epo o a WHO consul a ion. Diabe . Med. 15, 539–553 (1998).
12. Vázquez, J. A., Gaz ambide, S. & So o-Ped e, E. Incidence and isk ac o s o ype 2 diabe es melli us: a 10-yea p ospec i e s udy.
Med. Clin. (Ba c) 115, 534–539 (2000).
13. Rojo-Ma ínez, G. e al. Incidence o diabe es melli us in Spain as esul s o he na ion-wide coho [email p o ec ed] s udy. Sci. Rep. 10,
1–9 (2020).
14. Valdés, S., Bo as, P., Delgado, E., Ál a ez, F. & Diaz-Cadó niga, F. Popula ion-based incidence o ype 2 diabe es in No he n Spain.
The as u ias s udy. Diabe es Ca e 30, 2258–2263 (2007).
15. So igue , F. e al. Incidence o ype 2 diabe es in sou he n Spain (Piza a S udy). Eu . J. Clin. In es . 38, 126–133 (2008).
16. Valdés, S., Rojo-Ma ínez, G. & So igue , F. E olu ion o p e alence o ype 2 diabe es in adul Spanish popula ion. Med. Clin.
(Ba c) 129, 352–355 (2007).
17. Valdés, S. e al. P e alence o obesi y, diabe es and o he ca dio ascula isk ac o s in Andalusia (Sou he n Spain). Compa ison
wi h na ional p e alence da a. The [email p o ec ed] s udy. Re . Española Ca diol. (English Ed) 67, 442–448 (2014).
18. Vázquez, L. A. e al. Incidence o diabe es in he wo king popula ion in Spain: esul s om he ICARIA coho . Diabe es The . 10,
57–69 (2019).
19. A eagoi ia, J. M., La añaga, M. I., Rod iguez, J. L., Fe nandez, I. & Piniés, J. A. Incidence, p e alence and co ona y hea disease
isk le el in known Type 2 diabe es: a sen inel p ac ice ne wo k s udy in he Basque Coun y Spain. Diabe ologia 46, 899–909
(2003).
20. Beagley, J., Gua igua a, L., Weil, C. & Mo ala, A. A. Global es ima es o undiagnosed diabe es in adul s. Diabe es Res. Clin. P ac .
103, 150–160 (2014).
21. Bono a, E. e al. Popula ion-based incidence a es and isk ac o s o ype 2 diabe es in whi e indi iduals: he B uneck s udy.
Diabe es 53, 1782–1789 (2004).
22. Weye , C., Boga dus, C. & P a ley, R. E. Me abolic cha ac e is ics o indi iduals wi h impai ed as ing glucose and/o impai ed
glucose ole ance. Diabe es 48, 2197–2203 (1999).
8
Vol:.(1234567890)
Scien i ic Repo s | (2021) 11:3016 | h ps://doi.o g/10.1038/s41598-021-82548-y
www.na u e.com/scien i ic epo s/
23. Ch is ensen, P. e al. Men and women espond di e en ly o apid weigh loss: me abolic ou comes o a mul i-cen e in e en ion
s udy a e a low-ene gy die in 2500 o e weigh , indi iduals wi h p e-diabe es (PREVIEW). Diabe es Obes Me ab 20, 2840–2851
(2018).
24. Weye , C., Hanson, K., Boga dus, C. & P a ley, R. E. Long- e m changes in insulin ac ion and insulin sec e ion associa ed wi h
gain, loss, egain and main enance o body weigh . Diabe ologia 43, 36–46 (2000).
25. Schellenbe g, E. S., D yden, D. M., Vande mee , B., Ha, C. & Ko ownyk, C. Li es yle in e en ions o pa ien s wi h and a isk o
ype 2 diabe es. Ann. In e n. Med. 159, 543–551 (2013).
26. Ma hews, D. R. e al. Homeos asis model assessmen : insulin esis ance and β-cell unc ion om as ing plasma glucose and
insulin concen a ions in man. Diabe ologia 28, 412–419 (1985).
27. Albe i, K. G. M. M., Zimme , P. & Shaw, J. Me abolic synd ome—a new wo ld-wide de ini ion. A consensus s a emen om he
In e na ional Diabe es Fede a ion. Diabe . Med. 23, 469–480 (2006).
Acknowledgemen s
The au ho s hank all heal h cen e di ec o s, nu ses, adminis a i e s a and o he pe sonnel who kindly helped
de elop his wo k, especially he ield wo ke s Alicia Cobo, Oihana Fullaondo, Teba González-F u os and Viole a
Lago. We a e also g a e ul o Lía Ti ado and C is ina Llina es o echnical suppo and o all he people who
olun a ily ag eed o pa icipa e in he s udy.
Au ho con ibu ions
L.C. and S.G. con ibu ed o he concep ion and design o he s udy and he acquisi ion o unding. A.M.-N.,
J.-O.C.-M. and J.O. we e mainly conce ned wi h he da a collec ion and I.U., A.M.-N., R.M., J.-O.C.-M., A.A.,
E.A., E.F.-R. and he “Diabe es Epidemiology G oup” pe o med he da a analysis and in e p e a ion. I.U. and
A.M.-N. we e esponsible o d a ing he a icle, and all au ho s e ised i c i ically o impo an in ellec ual
con en . All au ho s app o ed he e sion o be published. LC and SG a e he gua an o s o his wo k and, as
such, ha e ull access o all he da a in he s udy and ake esponsibili y o he in eg i y o he da a and he
accu acy o he da a analyses.
Funding
This wo k was pa ially suppo ed by g an s om he Depa men o Heal h o he Basque Coun y Go e nmen
(2015111020); ISCIII (PI14/01104), co- unded by ERDF/ESF, “A way o make Eu ope”/”In es ing in you u u e”;
UPV/EHU (IT1281-19); Mena ini G oup Spain (BCA16/029); Endoc ine-Eu opean Re e ence Ne wo k (Endo-
ERN 739527); and CIBERDEM (Spanish Biomedical Resea ch Cen e in Diabe es and Associa ed Me abolic
Diso de s). The s udy unde s we e no in ol ed in he design o he s udy; he collec ion, analysis, and in e -
p e a ion o da a; w i ing he epo ; and did no impose any es ic ions ega ding he publica ion o he epo .
Compe ing in e es s
The au ho s decla e no compe ing in e es s.
Addi ional in o ma ion
Co espondence and eques s o ma e ials should be add essed o L.C.o S.G.
Rep in s and pe missions in o ma ion is a ailable a www.na u e.com/ ep in s.
Publishe ’s no e Sp inge Na u e emains neu al wi h ega d o ju isdic ional claims in published maps and
ins i u ional a ilia ions.
Open Access This a icle is licensed unde a C ea i e Commons A ibu ion 4.0 In e na ional
License, which pe mi s use, sha ing, adap a ion, dis ibu ion and ep oduc ion in any medium o
o ma , as long as you gi e app op ia e c edi o he o iginal au ho (s) and he sou ce, p o ide a link o he
C ea i e Commons licence, and indica e i changes we e made. The images o o he hi d pa y ma e ial in his
a icle a e included in he a icle’s C ea i e Commons licence, unless indica ed o he wise in a c edi line o he
ma e ial. I ma e ial is no included in he a icle’s C ea i e Commons licence and you in ended use is no
pe mi ed by s a u o y egula ion o exceeds he pe mi ed use, you will need o ob ain pe mission di ec ly om
he copy igh holde . To iew a copy o his licence, isi h p://c ea i eco mmons .o g/licen ses/by/4.0/.
© The Au ho (s) 2021
The Diabe es Epidemiology G oup
Alejand o Ga cía‑Cas año1,2, Sa a Gómez‑Conde1, Sa a La au i4,
Idoia Ma ínez de LaPiscina1,2, Lau a Saso1 & Olaia Velasco1