Ci a ion: Pham, T.-T.; Nimp sch, K.;
Aleksand o a, K.; Jenab, M.;
Reichmann, R.; Wu, K.; Tjønneland,
A.; Ky ø, C.; Schulze, M.B.; Kaaks, R.;
e al. P e-Diagnos ic Ci cula ing
Resis in Concen a ions A e No
Associa ed wi h Colo ec al Cance
Risk in he Eu opean P ospec i e
In es iga ion in o Cance and
Nu i ion S udy. Cance s 2022,14,
5499. h ps://doi.o g/10.3390/
cance s14225499
Academic Edi o : Ma k Molloy
Recei ed: 20 Sep embe 2022
Accep ed: 1 No embe 2022
Published: 9 No embe 2022
Publishe ’s No e: MDPI s ays neu al
wi h ega d o ju isdic ional claims in
published maps and ins i u ional a il-
ia ions.
Copy igh : © 2022 by he au ho s.
Licensee MDPI, Basel, Swi ze land.
This a icle is an open access a icle
dis ibu ed unde he e ms and
condi ions o he C ea i e Commons
A ibu ion (CC BY) license (h ps://
c ea i ecommons.o g/licenses/by/
4.0/).
cance s
A icle
P e-Diagnos ic Ci cula ing Resis in Concen a ions A e No
Associa ed wi h Colo ec al Cance Risk in he Eu opean
P ospec i e In es iga ion in o Cance and Nu i ion S udy
Thu-Thi Pham 1,2 , Ka ha ina Nimp sch 1, K asimi a Aleksand o a 3,4, Mazda Jenab 5, Robin Reichmann 3,
Kana Wu 6, Anne Tjønneland 7,8 , Cecilie Ky ø 7, Ma hias B. Schulze 9,10 , Rudol Kaaks 11, Ve ena Ka zke 11,
Domenico Palli 12 , Fab izio Pasanisi 13 , Ful io Ricce i 14,15 , Rosa io Tumino 16, Vi o io K ogh 17 ,
Jeanine Roodha 18, Jesús Cas illa 19,20 , Ma ia-Jose Sánchez 20,21,22,23 , Sand a Milena Colo ado-Yoha 20,24,25 ,
Jus in Ha bs 26, Ma in Ru egå d 27,28 , Ke en Papie 29 , Elom K. Aglago 30 , Niki Dimou 5,
Ana-Lucia Mayen-Chacon 5, Elisabe e Weide pass 31 and Tobias Pischon 1,2,32,33,*
1Molecula Epidemiology Resea ch G oup, Max-Delb ueck-Cen e o Molecula Medicine in he Helmhol z
Associa ion (MDC), 13125 Be lin, Ge many
2
Cha i é-Uni e si ä smedizin Be lin, Co po a e Membe o F eie Uni e si ä Be lin and Humbold -Uni e si ä
zu Be lin, 10117 Be lin, Ge many
3
Depa men o Epidemiological Me hods and E iological Resea ch, Leibniz Ins i u e o P e en ion Resea ch
and Epidemiology—BIPS, 28359 B emen, Ge many
4Facul y o Human and Heal h Sciences, Uni e si y o B emen, 28359 B emen, Ge many
5Nu i ion and Me abolism B anch, In e na ional Agency o Resea ch on Cance (IARC-WHO),
Wo ld Heal h O ganiza ion, 150 Cou s Albe Thomas, CEDEX 08, 69372 Lyon, F ance
6Depa men o Nu i ion, Ha a d T.H. Chan School o Public Heal h, Bos on, MA 02115, USA
7Danish Cance Socie y Resea ch Cen e , 2100 Copenhagen, Denma k
8Depa men o Public Heal h, Uni e si y o Copenhagen, DK-1353 Copenhagen, Denma k
9Depa men o Molecula Epidemiology, Ge man Ins i u e o Human Nu i ion Po sdam-Rehb uecke,
14558 Nu he al, Ge many
10 Ins i u e o Nu i ional Science, Uni e si y o Po sdam, 14558 Nu he al, Ge many
11
Depa men o Cance Epidemiology, Ge man Cance Resea ch Cen e (DKFZ), 69120 Heidelbe g, Ge many
12
Cance Risk Fac o s and Li e-S yle Epidemiology Uni , Ins i u e o Cance Resea ch, P e en ion and Clinical
Ne wo k (ISPRO), 50139 Flo ence, I aly
13 Dipa imen o di Medicina Clinica E Chi u gia, Fede ico Ii Uni e si y, 80131 Naples, I aly
14 Cen e o Bios a is ics, Epidemiology, and Public Heal h (C-BEPH), Depa men o Clinical and Biological
Sciences, Uni e si y o Tu in, 10043 O bassano, I aly
15 Uni o Epidemiology, Regional Heal h Se ice ASL TO3, 10095 G ugliasco, I aly
16 Hyblean Associa ion o Epidemiological Resea ch, AIRE ONLUS, 97100 Ragusa, I aly
17 Epidemiology and P e en ion Uni , Fondazione IRCCS Is i u o Nazionale dei Tumo i di Milano,
Via Venezian 1, 20133 Milan, I aly
18 Depa men o Medical Oncology, UMC U ech , 3584 CX U ech , The Ne he lands
19 Na a a Public Heal h Ins i u e—IdiSNA, 31003 Pamplona, Spain
20 CIBER Epidemiología y Salud Pública (CIBERESP), 28029 Mad id, Spain
21 Escuela Andaluza de Salud Pública (EASP), 18011 G anada, Spain
22 Ins i u o de In es igación Biosani a ia ibs.GRANADA, 18012 G anada, Spain
23 Depa men o P e en i e Medicine and Public Heal h, Uni e si y o G anada, 18071 G anada, Spain
24 Depa men o Epidemiology, Mu cia Regional Heal h Council, IMIB-A ixaca, 30008 Mu cia, Spain
25 Resea ch G oup on Demog aphy and Heal h, Na ional Facul y o Public Heal h, Uni e si y o An ioquia,
Medellín 050010, Colombia
26 Depa men o Radia ion Sciences, Oncology, Umeå Uni e si y, SE-901 87 Umeå, Sweden
27 Depa men o Su gical and Pe iope a i e Sciences, Su ge y, Umeå Uni e si y, SE-901 87 Umeå, Sweden
28 Wallenbe g Cen e o Molecula Medicine, Umeå Uni e si y, SE-901 87 Umeå, Sweden
29 Cance Epidemiology Uni , Nu ield Depa men o Popula ion Heal h, Uni e si y o Ox o d,
Ox o d OX3 7LF, UK
30 Depa men o Epidemiology and Bios a is ics, Impe ial College London, London W2 1PG, UK
31 In e na ional Agency o Resea ch on Cance , Wo ld Heal h O ganiza ion, 69372 Lyon, F ance
32 Max-Delb ueck-Cen e o Molecula Medicine in he Helmhol z Associa ion (MDC), Biobank Technology
Pla o m, 13125 Be lin, Ge many
33
Be lin Ins i u e o Heal h, Cha i é-Uni e si ä smedizin Be lin, Co e Facili y Biobank, 13125 Be lin, Ge many
*Co espondence: [email p o ec ed]; Tel.: +49-30-9406-4563
Cance s 2022,14, 5499. h ps://doi.o g/10.3390/cance s14225499 h ps://www.mdpi.com/jou nal/cance s
Cance s 2022,14, 5499 2 o 15
Simple Summa y:
Resis in has been p oposed o link o cance de elopmen ia in lamma o y
p ocesses. P io case-con ol s udies sugges highe pos -diagnosis esis in concen a ions in CRC
cases compa ed o con ols. He e, we ound no associa ion be ween p e-diagnos ic ci cula ing esis in
concen a ions and he isk o CRC; howe e , we obse ed a ma ginally signi ican associa ion among
cases (and hei ma ched con ols) diagnosed wi h CRC wi hin he i s wo yea s o ollow-up,
whe eas no such associa ion was obse ed among cases (and hei ma ched con ols) diagnosed
wi h CRC a e wo yea s o ollow-up. We specula e ha esis in is mo e likely a ma ke o exis ing
umo s han a isk ac o o CRC.
Abs ac :
Resis in is a polypep ide implica ed in in lamma o y p ocesses, and as such could be
linked o colo ec al ca cinogenesis. In case-con ol s udies, highe esis in le els ha e been ound
in colo ec al cance (CRC) pa ien s compa ed o heal hy indi iduals. Howe e , e idence o he
associa ion be ween p e-diagnos ic esis in and CRC isk is sca ce. We in es iga ed p e-diagnos ic
esis in concen a ions and CRC isk wi hin he Eu opean P ospec i e In es iga ion in o Cance and
Nu i ion using a nes ed case-con ol s udy among 1293 inciden CRC-diagnosed cases and 1293
incidence densi y-ma ched con ols. Condi ional logis ic eg ession models con olled o ma ching
ac o s (age, sex, s udy cen e , as ing s a us, and women- ela ed ac o s in women) and po en ial
con ounde s (educa ion, die a y and li es yle ac o s, body mass index (BMI), BMI-adjus ed wais
ci cum e ence esiduals) we e used o es ima e ela i e isks (RRs) and 95% con idence in e als (CIs)
o CRC. Highe ci cula ing esis in concen a ions we e no associa ed wi h CRC (RR pe doubling
esis in, 1.11; 95% CI 0.94–1.30; p= 0.22). The e we e also no associa ions wi h CRC subg oups
de ined by umo subsi e o sex. Howe e , esis in was ma ginally associa ed wi h a highe CRC
isk among pa icipan s ollowed-up maximally wo yea s, bu no among hose ollowed-up a e
mo e han wo yea s. We obse ed no subs an ial co ela ion be ween baseline ci cula ing esis in
concen a ions and adiposi y measu es (BMI, wais ci cum e ence), adipokines (adiponec in, lep in),
o me abolic and in lamma o y bioma ke s (C- eac i e p o ein, C-pep ide, high-densi y lipop o ein
choles e ol, eac i e oxygen me aboli es) among con ols. In his la ge-scale p ospec i e coho , he e
was li le e idence o an associa ion be ween baseline ci cula ing esis in concen a ions and CRC
isk in Eu opean men and women.
Keywo ds: p e-diagnos ic esis in; colo ec al cance ; isk; p ospec i e; in lamma ion
1. In oduc ion
Resis in is a polypep ide consis ing o 108 amino acids named a e “ esis ance o
insulin” ha belongs o he “ esis in-like molecules” amily [
1
]. I was ini ially epo ed in
oden s as a p o ein p ima ily sec e ed by adipocy es and plays a ole in obesi y-induced
insulin esis ance. Highe esis in concen a ions ha e been ound in obese as compa ed o
non-obese animal models [
1
]. In humans, epo s on he co ela ion be ween ci cula ing
esis in concen a ions and adipose issue mass ha e been inconsis en [
2
–
6
], and esis in
was ound o be p edominan ly p oduced by mac ophages and monocy es, a he han
adipocy es [
2
,
3
]. The p oduc ion and up egula ion o esis in occu du ing monocy e–
mac ophage di e en ia ion [
3
] and esis in hen p omo es an M1-like (p o-in lamma o y)
pheno ype in mac ophages [
7
]. Inc eases in adipose issue mass in humans may be ac-
companied by in il a ion o mac ophages and monocy es, which elease esis in and may
he eby a ec mul iple cell ypes and issues [
3
,
8
]. Human esis in is ele an o in lam-
ma o y p ocesses [
8
,
9
]. By a ious mechanisms, esis in ecep o binding may lead o an
up egula ion o in lamma o y cy okines, including in e leukin-6 (IL-6) and umo nec osis
ac o
α
(TNF-
α
) [
8
–
10
], o p omo e NF-kB ac i a ion [
7
], whe eas esis in exp ession is also
up egula ed in he p esence o in lamma o y cy okines by a posi i e eedback loop, which
may esul in a icious cycle and pe pe ua e in lamma o y condi ions [
7
,
10
,
11
]. O he han
in lamma ion, esis in could p omo e he exp ession o adhesion molecules and g ow h
Cance s 2022,14, 5499 3 o 15
ac o s ha may p omo e angiogenesis [
8
,
9
]. Thus, esis in eme ges as an impo an ma ke
implica ed in he de elopmen o cance [8].
Colo ec al cance (CRC) is he hi d mos common and second mos a al cance
wo ldwide [
12
]. The global bu den o CRC is expec ed o inc ease by 60% by 2030 [
13
].
While Eu opean, Aus alia/New Zealand, and No he n Ame ican egions ha e bo h he
highes CRC cance incidence and mo ali y a e, he majo i y egions o A ica and Sou h
Cen al Asia ha e he lowes incidence a e [
12
]. Thus, CRC incidence has been sugges ed o
co ela e wi h socioeconomic de elopmen and may be linked o wes e n li es yles [
12
,
13
].
Fu he mo e, incidence o CRC is closely ela ed o mul iple ac o s, including a amily
his o y o colon polyps o in lamma o y bowel diseases, socioeconomic s a us, li es yle and
die a y ac o s, and he gu mic obiome, which sha es he unde lying mechanism ela ed
o in lamma ion, angiogenesis, and insulin esis ance [14,15].
As bo h in lamma ion and angiogenesis a e ela ed o CRC, i was specula ed whe he
esis in may also be ele an o he isk o CRC [
8
,
16
]. In ac , highe esis in le els ha e
been ound in CRC pa ien s compa ed o heal hy indi iduals in case-con ol s udies [
17
].
Howe e , i is unclea whe he highe esis in le els a e a cause o consequence o CRC
de elopmen . P ospec i e s udies wi h p e-diagnos ic esis in concen a ions a e necessa y
o in es iga e whe he highe esis in concen a ions a e associa ed wi h a highe isk o
de eloping CRC. To da e, he e has been only one p ospec i e s udy using da a om
1224 pos menopausal women (427 CRC cases) in he Women’s Heal h Ini ia i e (WHI),
which ound no signi ican associa ions be ween esis in and CRC isk [
5
]. Howe e ,
di e en associa ions be ween in lamma ion and he isk o CRC ha e been obse ed in
men e sus women [
18
], sugges ing ha he associa ion be ween esis in concen a ions
and he isk o CRC may di e by sex. Fu he mo e, ha s udy did no sepa a e he
associa ion by umo subsi e. Gi en ha in lamma ion is mo e s ongly ela ed o colon
cance han o ec al cance [
19
], an examina ion by umo subsi e is impo an o be e
unde s and he impac o esis in as an exposu e.
The e o e, we conduc ed a nes ed case-con ol s udy wi hin he Eu opean P ospec i e
In es iga ion in o Cance and Nu i ion (EPIC) s udy o in es iga e he associa ion be ween
p e-diagnos ic esis in concen a ions and CRC isk and o examine whe he he associa ion
di e ed by cance subsi e o sex.
2. Ma e ials and Me hods
2.1. The Eu opean P ospec i e In es iga ion in o Cance and Nu i ion S udy
2.1.1. S udy Popula ion
The EPIC s udy is a la ge, ongoing coho s udy, de ails o which ha e been ex ensi ely
epo ed elsewhe e [
20
]. B ie ly, he EPIC s udy was ini ia ed in 1990 wi h he collabo a ion
o 23 cen e s in 10 Eu opean coun ies, including Denma k, F ance, Ge many, G eece, I aly,
The Ne he lands, No way, Spain, Sweden, and he Uni ed Kingdom. S udy pa icipan s’
en ollmen in all EPIC cen e s had been comple ed in 2000, wi h 519,978 pa icipan s aged
35 o 70 yea s. In he cu en nes ed case-con ol s udy, da a om G eece we e no included
due o adminis a i e easons.
2.1.2. Assessmen s o An h opome y, Li es yle Fac o s, and Die a y Exposu e
Weigh and heigh we e measu ed o all pa icipan s in all EPIC cen e s by ained
obse e s, excep in F ance, Ox o d, and No way [
21
]. In F ance and Ox o d, weigh and
heigh we e measu ed in a pa o he popula ion, whe eas sel - epo ed weigh and heigh
we e a ailable o all pa icipan s. Fo pa o he Ox o d coho , linea eg ession models
we e used o p edic sex- and age-speci ic alues om indi iduals wi h bo h measu ed
and sel - epo ed body measu es. Fo F ance, only he measu ed alues we e used. In
No way, only sel - epo ed weigh and heigh da a we e a ailable. Wais ci cum e ence
was measu ed in all cen e s excep in No way and Umeå, Sweden.
Li es yle cha ac e is ics we e collec ed ia ques ionnai es a ec ui men including
ques ions on educa ion, obacco smoking s a us, consump ion o alcoholic be e ages,
Cance s 2022,14, 5499 4 o 15
physical ac i i y acco ding o he alida ed Camb idge physical ac i i y index, and o
women, mens ual s a us, use o con acep ion, and ho mone eplacemen he apy [
20
].
The Camb idge physical ac i i y index desc ibes ou le els o physical ac i i y (inac i e,
mode a ely inac i e, mode a ely ac i e, and ac i e) based on ec ea ional ac i i y ime in
occupa ion, cycling, and o he physical ac i i ies [
22
]. Ques ionnai es on li es yle a iables
we e de eloped and used independen ly in Denma k, Sweden, No way, and he Naples
cen e in I aly, while hey we e p e iously s anda dized in o he coun ies. All li es yle
a iable codes om di e en ques ionnai es we e s anda dized o he co e EPIC li es yle
ques ions using a comp ehensi e ecoding scheme a e wa d. Baseline die a y exposu e
(including, among o he s, consump ion o ed mea , p ocessed mea , die a y ibe , ui ,
ege able, dai y, ish, and shell ish in ake, as well as alcohol consump ion) was assessed
on all EPIC pa icipan s by using locally adop ed ins umen s, including ood equency
ques ionnai es, die his o y logs, and a combined me hod, and was used o es ima e long-
e m usual die a y in ake.
2.1.3. Blood Collec ion
Biological samples including plasma, se um, leukocy es, and e y h ocy es we e col-
lec ed a baseline om mos pa icipan s, wi h he excep ion o hose in F ance, he UK,
Bil ho en (The Ne he lands), and No way, whe e only a p opo ion o he pa icipan s
we e in i ed o blood sampling [
20
]. Fo mos EPIC cen e s, hal o he blood samples
we e s o ed locally, and hal we e anspo ed o he cen al eposi o y o he In e na ional
Agency o Resea ch on Cance (IARC) o be s o ed in s aws in he apo phase o liquid
ni ogen a
−
196
◦
C. Blood samples we e all s o ed locally in eeze s a
−
70
◦
C in Sweden
and ni ogen apo a −150 ◦C in Denma k.
2.1.4. Follow-Up o Cance Incidence
Inciden CRC cases we e iden i ied ia egional cance egis ies in Denma k, I aly, The
Ne he lands, No way, Spain, Sweden, and he Uni ed Kingdom and by a combina ion o
me hods in F ance and Ge many. The combina ion o me hods included a di ec ollow-up
p ocedu e h ough s udy pa icipan s o hei nex o kin, and con i ma ion o umo s om
a e iew o heal h insu ance eco ds and pa hology egis ies. Fo cen e s ha applied a
combina ion o me hods, indi iduals’ end o ollow-up was conside ed he las known
con ac , da e o diagnosis, o da e o dea h, whiche e came i s .
2.2. The Nes ed Case-Con ol S udy
2.2.1. S udy Design
Inciden CRC cases we e de ined as EPIC pa icipan s who de eloped CRC a e
ec ui men and be o e he closu e da es. CRC was de ined as a combina ion o umo s o
he colon ( he 10 h Re ision o he In e na ional Classi ica ion o Diseases (ICD-10) codes
C18.0–C18.7), umo s ha we e o e lapping o unspeci ied (C18.8–C18.9), and umo s o
he ec um (C19–C20). The closu e da es o he nes ed case-con ol s udy o he p esen
analysis anged om Decembe 2001 o Decembe 2005.
The con ol selec ion p ocess was based on an incidence densi y sampling app oach.
One con ol was selec ed o each case om among a sample o hose who we e a isk
a he ime o diagnosis o he index case wi h a ailable blood samples and was ma ched
(1:1) by ec ui men cen e , sex, age a ec ui men (
±
2 yea s), da e o blood collec ion
(
±
3 mon hs), ime o day o blood collec ion (
±
4 h), as ing s a us a blood collec ion (<3,
3–6, and >6 h), and menopausal s a us (p emenopausal, pos menopausal, pe imenopausal,
o su gically pos menopausal) o women. P emenopausal women we e ma ched on
phases o mens ual cycles and use o o al con acep i es, and pos menopausal women
we e ma ched on cu en ho monal eplacemen he apy use.
Cance s 2022,14, 5499 5 o 15
2.2.2. Labo a o y Analysis
Se um esis in concen a ions we e measu ed using human esis in ELISA assays
(BioVendo Labo a o y Medicine, Inc.; B no, Czech Republic). Measu emen s we e pe -
o med acco ding o he manu ac u e ’s p o ocols. The mean in e -assay coe icien s o
a ia ion o he labo a o y analysis we e 7.4% and 6.6% o quali y-con ol high concen a-
ions (18.3 ng/mL) and low concen a ions (5.01 ng/mL), espec i ely. The mean in e -assay
coe icien o a ia ion was <10.4% o all pooled se um quali y con ols. Measu emen s
and in e -assay coe icien s o a ia ion in adiponec in and high-molecula weigh (HMW)
adiponec in [
23
], lep in [
24
], soluble lep in ecep o [
24
], eac i e oxygen me aboli es
(ROM) [
25
], high-densi y lipop o ein choles e ol (HDL-C), C-pep ide [
26
], high-sensi i i y
C- eac i e p o ein (hsCRP) [
27
], and glyca ed hemoglobin A1c (Hba1c) [
28
] ha e been
desc ibed p e iously.
2.2.3. Final Da ase , Handling o Missing Da a
The cu en analysis was based on da a om pa icipan s in he EPIC CRC nes ed case-
con ol s udy, o which plasma samples we e a ailable o labo a o y analysis. Resis in
concen a ions we e success ully analyzed in 1383 CRC cases and 1375 con ols. A e
excluding pa icipan s wi hou hei ma ching pai s, a o al o 1293 i s -inciden CRC cases
and hei ma ched con ols we e included in he cu en s udy.
Wais ci cum e ence measu emen s we e missing in 77 ma ched case se s (6.0%),
including 16 and 61 case se s om No way and Umeå-Sweden, espec i ely. Fo 4 CRC
cases, da a on alcohol consump ion, ene gy in ake, and consump ion o ed mea , p ocessed
mea , die a y ibe , ui , ege able, dai y, ish, and shell ish in ake we e missing. Da a
o o he a iables we e missing in cases and con ols as ollows: as ing s a us (20/21),
smoking s a us (14/15), highes educa ion le el (43/34), physical ac i i y index (16/18),
and diabe es (73/90). We checked he dis ibu ion o esis in concen a ions in pa icipan s
wi h missing and no missing da a on hese a iables and obse ed no di e ences. Toge he
wi h he a bi a y missing da a pa e ns gene a ed by he PROC MI p ocedu e in SAS
®
En e p ise Guide
®
8.3 (SAS Ins i u e Inc., Ca y, NC, USA), we assumed ha he da a we e
missing comple ely a andom. Hence, missing da a we e impu ed by sex-speci ic medians
o disc e e a iables and sex-speci ic modes o ca ego ical a iables.
2.3. S a is ical Analysis
Qua ile cu -o poin s o esis in concen a ions we e de i ed om con ols and ap-
plied o he whole s udy popula ion. We hen analyzed he s udy popula ion cha ac e is ics
desc ip i ely by case-con ol s a us and by qua iles o esis in concen a ions. We cal-
cula ed means (s anda d de ia ion (SD)) o medians (and qua iles) depending on he
dis ibu ion o he a iables.
Spea man pa ial co ela ion coe icien s (and co esponding p- alues) con olling
o age and sex we e es ima ed o assess he co ela ions be ween baseline esis in le -
els and adiposi y measu emen s, o he adipokines (adiponec in and lep in), as well as
me abolic and in lamma o y bioma ke s (hsCRP, C-pep ide, HDL-C, and ROM) in con ols.
A co ela ion coe icien lowe han 0.30 was conside ed as li le i any co ela ion [29].
Condi ional logis ic eg ession models we e used o es ima e RRs and 95% CI o CRC
ac oss qua iles o esis in concen a ions. Since cases and con ols we e selec ed using
an incidence densi y sampling p o ocol, he odds a ios es ima e he incidence a e a ios,
which can be in e p e ed as RRs o he associa ion. Po en ial con ounde s we e selec ed
as co a ia es o he models. Highe in akes o ibe and dai y p oduc s a e es ablished
p o ec i e ac o s o CRC, while highe in akes o ed mea , p ocessed mea , and alcohol
a e isk ac o s [
14
,
15
]. Low in akes o ege ables and ish ha e been associa ed wi h a
highe isk o CRC [
14
,
15
]. Since hese a iables may also a ec esis in concen a ions [
30
],
we adjus ed o hese a iables in ou eg ession models. BMI and wais ci cum e ence
a e isk ac o s o CRC [
15
], howe e , he ela ionship be ween obesi y and esis in is
no en i ely clea [
4
,
6
,
31
]. The e o e, h ee models we e used o es ima e he associa ion.

Cance s 2022,14, 5499 6 o 15
Model 1 was only condi ioned on he ma ching a iables. Model 2 was condi ioned on
ma ching a iables and addi ionally adjus ed o smoking s a us (ne e , o me , o cu en
smoke ), educa ion (none, p ima y school, echnical/p o essional school, seconda y school,
o longe educa ion), alcohol abs aine s (de ined as unde 0.3 g/day; yes/no), alcohol
consump ion (g am/day), physical ac i i y index (inac i e, mode a ely inac i e, mode -
a ely ac i e, ac i e), ene gy in ake (kcal/day), and die a y in akes o ed mea (g am/day),
p ocessed mea (g ams/day), die a y ibe (g ams/day), ui in ake (g ams/day), eg-
e able in ake (g ams/day), dai y in ake (g ams/day), and ish and shell ish (g ams/day).
Model 3 was addi ionally adjus ed o BMI (kg/m
2
) and esiduals o BMI-adjus ed wais
ci cum e ence (de i ed om a eg ession model wi h BMI as an independen a iable
and wais ci cum e ence as a dependen a iable o a oid mul icollinea i y). Addi ional
adjus men o heigh , alcohol in ake du ing li e ime, smoking in ensi y and du a ion, o
diabe es a baseline (de ined as sel - epo ed diabe es diagnosis o HbA1c concen a ions
≥
6.5% a baseline) did no al e he isk es ima es app eciably and esul s a e hence no
shown. Resis in concen a ions we e log- ans o med o app oxima e a no mal (Gaussian)
dis ibu ion and included as a con inuous a iable in condi ional logis ic eg ession mod-
els [
32
]. We also analyzed he associa ions be ween esis in concen a ions and he isk o
CRC by modeling es ic ed cubic polynomial splines wi h kno s a he 5 h, 35 h, 65 h, and
95 h pe cen iles o esis in dis ibu ion. We es ed o non-linea i y using a likelihood a io
es o compa e ull mul i a iable-adjus ed condi ional logis ic eg ession models, includ-
ing bo h he linea and cubic spline e m, and educed mul i a iable-adjus ed condi ional
logis ic eg ession models wi h only he linea e m. The es o non-linea i y was no
signi ican in he main analysis (p= 0.35) o he subg oups, excep o he subg oup o men
(p= 0.03). Resul s o non-linea i y es s did no change implici ly a e excluding pa -
icipan s diagnosed wi h CRC wi hin 2 yea s a e ec ui men . Thus, using bo h log-
ans o med and qua ile scales was conside ed su icien o be e cap u ing he ela ion-
ship be ween esis in concen a ions and he isk o CRC [32].
The associa ions be ween esis in concen a ions and he isk o CRC we e u he
assessed acco ding o umo ana omical subsi e (colon, ec um), sex ( emale, male, as well
as wi h es ic ion o pos menopausal women [
5
]), he combina ion o subsi e and sex, by
he leng h o ollow-up (
≤
2 yea s, 2–5 yea s, >5 yea s), BMI (<25,
≥
25), hsCRP (<3 mg/L,
≥
3 mg/L), baseline diabe es (yes/no), as ing s a us (
≤
6 h, >6 h), and C-pep ide (<2 ng/mL,
≥
2 ng/mL). Sensi i i y analyses we e ca ied ou by epea ing all o he analyses wi h he
exclusion o pa icipan s wi h less han 2 yea s o ollow-up, and pa icipan s wi h ex eme
esis in le els (de ined as concen a ions o 1.5 imes he in e qua ile ange below he i s
and abo e he hi d qua ile [
33
]). Analyses we e also es ic ed o pa icipan s wi h no
missing co a ia es’ da a (comple e case analyses).
We inally pooled he ela i e isk es ima e om ou main analysis wi h he ela i e
isk om he published esul om he WHI [
5
] using a andom-e ec s me a-analysis wi h
an in e se a iance me hod. He e ogenei y was assessed using Coch an’s Q-s a is ic es
and inconsis ency index (I2).
Minimal de ec able RRs o bina y exposu e we e es ima ed o he second, hi d, o
ou h qua iles as compa ed o he i s qua ile in ma ched case-con ol using he so wa e
“Powe ” e sion 2.10 (Channing labo a o y, Bos on, MA, USA) [
34
]. Wi h 1293 s udy
pa icipan s o each RR es ima e, 80% powe , and alpha = 0.05, he minimum de ec able
RR is 1.37 unde he assump ion o no co ela ion o exposu e in ma ched pai s, and 1.42
wi h a ecommended co ela ion o exposu e be ween cases and ma ched con ols o 0.2.
All s a is ical es s we e wo-sided, and p- alues less han 0.05 we e conside ed s a is-
ically signi ican . All analyses we e pe o med using SAS
®
En e p ise Guide
®
8.3 (SAS
Ins i u e Inc., Ca y, NC, USA) and R e sion 4.0.5 (R Founda ion o S a is ical Compu ing,
Vienna, Aus ia).
Cance s 2022,14, 5499 7 o 15
3. Resul s
The mean ollow-up ime be ween ec ui men and CRC diagnosis in cases o he
cu en s udy was 4.8
±
2.7 yea s. Cases and con ols had simila dis ibu ions o he ma ch-
ing ac o s (Supplemen a y Table S1). On a e age, cases a baseline we e less physically
ac i e, had a highe BMI, a highe wais ci cum e ence, and a sligh ly lowe die a y ibe
in ake compa ed o con ols (Supplemen a y Table S1). Cases also had highe ci cula ing
le els o HMW adiponec in, ROM, hsCRP, and HbA1c han con ols. Resis in concen-
a ions a baseline we e simila in inciden CRC cases (4.7
±
2.0 ng/mL) and con ols
(4.7 ±2.2 ng/mL).
The main cha ac e is ics o con ols in he popula ion acco ding o he qua iles o
esis in a e p esen ed in Table 1. Pa icipan s wi h highe esis in concen a ions we e mo e
likely o be women, ha e a highe age, a lowe p opo ion o uni e si y-le el educa ion,
lowe alcohol in ake, highe hsCRP, highe C-pep ide, and lowe HDL-C, while less likely o
be in as ing s a us a blood collec ion. In e es ingly, we obse ed highe le els o esis in in
women compa ed o men (4.82
±
2.48 s. 4.53
±
1.84 ng/mL, p= 0.01), and in pa icipan s
wi h less han o equal o 6 h o as ing be o e blood collec ion compa ed o hose wi h
mo e han 6 h o as ing (4.79 ±2.39 s. 4.38 ±1.56 ng/mL, p≤0.001).
Table 1.
Cha ac e is ics o con ols a baseline (n = 1293), by qua iles o esis in concen a ion in he
nes ed case-con ol s udy, Eu opean P ospec i e In es iga ion in o Cance and Nu i ion, 1992–2005.
Qua iles o Resis in Concen a ion
Q1
(N = 324)
Q2
(N = 325)
Q3
(N = 321)
Q4
(N = 323)
Resis in qua ile anges (ng/mL) ≤3.47 3.47< o ≤4.28 4.28< o ≤5.42 5.42< o ≤34.41
Age a blood collec ion, yea s, mean (SD) a57.6 (6.9) 57.5 (7.2) 58.1 (6.7) 59 (7.1)
Women, n (%) a154 (47.5) 183 (56.3) 157 (48.9) 187 (57.9)
Pos menopausal women, baseline, n (%) a123 (38.0) 135 (41.5) 116 (36.1) 141 (43.7)
Fas ing (>6 h), n (%) a98 (30.2) 98 (30.2) 74 (23.1) 80 (24.8)
BMI, kg/m2, mean (SD) 26.2 (3.5) 26.6 (4.1) 26.3 (3.8) 26.2 (3.7)
Wais ci cum e ence, cm, mean (SD) b89.2 (12.5) 88.4 (12.2) 88.8 (11.8) 87.7 (12.6)
Diagnosed diabe es a baseline (sel - epo ed o
HbA1C ≥6.5%), n (%) 19 (5.9) 18 (5.5) 13 (4.0) 14 (4.3)
Cu en smoke , n (%) 84 (25.9) 86 (26.5) 68 (21.2) 85 (26.3)
Uni e si y deg ee o highe educa ion le el, n (%)
68 (21) 52 (16) 51 (15.9) 58 (18)
Physically mode a ely ac i e, o ac i e,
sex-speci ic, n (%) 178 (54.9) 176 (54.2) 168 (52.3) 195 (60.4)
Alcohol abs aine s (<0.3 g/day), n (%) 38 (11.7) 45 (13.8) 43 (13.4) 51 (15.8)
Alcohol consump ion, g/day, median (Q1–Q3) b10.6 (3.2–28.1) 8.5 (1.6–23.5) 7.7 (1.7–19.9) 6.0 (1.0–16.8)
Ene gy in ake, Kcal/day, median (Q1–Q3) b2084 (1608–2478) 2010 (1638–2404) 2052 (1719–2511) 1997 (1590–2464)
Red mea , g/day, median (Q1–Q3) b43.4 (24.4–75.3) 43.0 (25.3–69.9) 49.6 (27.4–76.5) 47.7 (24.2–74.9)
P ocessed mea , g/day, median (Q1–Q3) b24.6 (13.0–42.6) 23.3 (12.5–45.9) 28.4 (16.8–44.3) 23.2 (12.9–42.6)
Die a y ibe , g/day, median (Q1–Q3) b23.6 (18.0–28.6) 22.8 (18.0–27.5) 23.2 (18.0–27.8) 22.1 (17.9–26.7)
F ui in ake, g/day, median (Q1–Q3) b191.5 (106.4–314.3) 195.2 (108.2–330.2) 176.9 (98.2–279.8) 191.5 (109.3–318.4)
Vege able in ake, g/day, median (Q1–Q3) b160.9 (104.3–230.8) 166.8 (103.8–248.1) 148.1 (96.0–221.6) 157.9 (97.4–241.5)
Dai y in ake, g/day, median (Q1–Q3) b278.9 (155.2–435.3) 310.6 (175.0–451.0) 318.2 (171.2–495.9) 305.9 (161.5–484.5)
Fish and shell ish, g/day, median (Q1–Q3) b29.0 (13.4–52.3) 29.7 (16.4–51.8) 29.4 (14.2–49.3) 30.0 (13.6–52.4)
Adiponec in, µg/mL, median (Q1–Q3) b7.3 (5.2–10.1) 7.3 (5.5–9.7) 6.9 (5.0–9.2) 7.2 (5.3–9.6)
HMW adiponec in, µg/mL (Q1–Q3) b3.7 (2.4–5.7) 3.8 (2.4–5.4) 3.6 (2.2–5.1) 3.7 (2.4–5.2)
Lep in, ng/mL (Q1–Q3) b6.3 (3.4–14.1) 9.4 (5.1–19.2) 7.7 (3.5–16.6) 6.9 (3.6–17.0)
Soluble lep in ecep o , ng/mL, mean (SD) b23 (8.9) 21.9 (7.9) 22.7 (8.6) 23.6 (15.8)
ROM, Ca a elli uni s, mean (SD) b381 (71.8) 382.4 (69.4) 379.1 (64.2) 397.4 (71.7)
hsCRP, mg/L, median (Q1–Q3) b1.9 (0.8–3.7) 2.2 (0.9–4.0) 2.1 (0.9–4.2) 2.9 (1.3–5.8)
C-pep ide, ng/mL, median (Q1–Q3) b3.6 (2.5–5.9) 3.7 (2.6–5.6) 4.0 (2.9–6.3) 4.3 (2.9–5.9)
HDL-C, mmol/L, median (Q1–Q3) b1.5 (1.2–1.8) 1.5 (1.2–1.7) 1.4 (1.2–1.7) 1.4 (1.1–1.7)
HbA1c, %, mean (SD) b5.8 (0.6) 5.8 (0.8) 5.7 (0.6) 5.7 (0.5)
a
Ma ching a iable.
b
Among use s only. Abb e ia ions: Q1–Q3, he i s qua ile ( he 25 h pe cen ile) o he
hi d qua ile ( he 75 h pe cen ile); SD, s anda d de ia ion; BMI: body mass index; HMW, high-molecula weigh ;
ROM, eac i e oxygen me aboli es; HDL-C, high-densi y lipop o ein choles e ol (HDL-C); hsCRP, high-sensi i i y
C- eac i e p o ein; HbA1c, glyca ed hemoglobin A1c.
Cance s 2022,14, 5499 8 o 15
Among con ols, a e adjus men o age and sex, esis in concen a ions we e weakly
in e sely co ela ed wi h adiponec in and HDL-C, and weakly posi i ely co ela ed wi h
C-pep ide and hsCRP (co ela ion coe icien ( ) < 0.15). Resis in was no s a is ically
signi ican ly co ela ed wi h BMI, wais ci cum e ence, lep in, soluble lep in ecep o s,
eac i e oxygen me aboli es, o HbA1c concen a ions (Table 2).
Table 2.
Age- and sex-adjus ed Spea man pa ial ank co ela ions be ween esis in and adiposi y
measu emen s, o he adipokines, as well as me abolic and in lamma o y bioma ke s in con ols
(n = 1293), he Eu opean P ospec i e In es iga ion in o Cance and Nu i ion Coho (1992–2005).
Numbe o Pa icipan s Co ela ion Coe icien ( ) p-Value
BMI, kg/m21293 −0.02 0.52
Wais ci cum e ence, cm 1216 −0.03 0.28
Adiponec in, µg/mL 651 −0.08 0.04
HMW adiponec in, µg/mL 650 −0.07 0.08
Lep in, ng/mL 651 −0.002 0.96
Soluble lep in ecep o , ng/mL 651 −0.02 0.56
ROM, Ca a elli uni s 723 0.07 0.05
CRP-hs, mg/L 727 0.12 <0.01
C-pep ide, ng/mL 622 0.09 0.02
HDL-C, mmol/L 726 −0.12 <0.01
HbA1c, % 606 −0.05 0.24
: co ela ion coe icien ; BMI, body mass index; HMW, high-molecula weigh ; ROM, eac i e oxygen me aboli es;
HDL-C, high-densi y lipop o ein choles e ol (HDL-C); hsCRP, high-sensi i i y C- eac i e p o ein; HbA1c, glyca ed
hemoglobin A1c.
Resis in concen a ions we e no s a is ically signi ican ly associa ed wi h he isk o
CRC. Thus, compa ed o qua ile one, he RRs in qua iles we e, qua ile wo, 1.11; 95% CI:
0.88–1.39, qua ile h ee, 1.21; 95% CI: 0.97–1.53, and qua ile ou , 1.15; 95% CI: 0.91–1.46,
p- end = 0.41. In he con inuous scale, RR o doubling esis in concen a ions was 1.11;
95% CI: 0.94–1.30; p= 0.22 (Table 3, model 3). No s a is ically signi ican ela ionship
be ween esis in and CRC was obse ed when he associa ions we e u he explo ed
by umo si e (colon o ec um) (Table 3) and igh -sided and le -sided colon cance
(Supplemen a y Table S2).
Table 3.
Rela i e isk (RR) and 95% con idence in e al (95% CI) es ima ed o he associa ion
be ween esis in concen a ions and he isk o colo ec al cance in he EPIC s udy da a (1992–2005)
in condi ional logis ic eg ession models.
Qua ile Fo m Con inuous Fo m
Q1 Q2 Q3 Q4 p-T end aDoubling Resis in
Concen a ions bp-Value
Resis in qua ile anges
(ng/mL) ≤3.47 3.47< o ≤4.28 4.28< o ≤5.42
5.42< o
≤
34.41
Colo ec al Cance
No. cases/con ols 297/324 317/325 348/321 331/323 1293/1293
Model 1 e 1.06 (0.85–1.32) 1.19 (0.95–1.49) 1.13 (0.90–1.41) 0.46 1.11 (0.95–1.30) 0.19
Model 2 e 1.10 (0.88–1.37) 1.22 (0.97–1.53) 1.15 (0.91–1.46) 0.38 1.12 (0.95–1.31) 0.18
Model 3 e 1.11 (0.88–1.39) 1.21 (0.97–1.53) 1.15 (0.91–1.46) 0.41 1.11 (0.94–1.30) 0.22
Colon Cance
No. cases/con ols 165/174 178/193 203/188 211/202 757/757
Model 1 e 0.97 (0.73–1.30) 1.16 (0.86–1.56) 1.11 (0.83–1.50) 0.62 1.14 (0.94–1.40) 0.19
Model 2 e 1.01 (0.75–1.37) 1.17 (0.86–1.59) 1.15 (0.84–1.56) 0.67 1.15 (0.93–1.42) 0.19
Model 3 e 1.04 (0.76–1.41) 1.20 (0.88–1.65) 1.20 (0.88–1.65) 0.53 1.18 (0.95–1.47) 0.13
Cance s 2022,14, 5499 9 o 15
Table 3. Con .
Qua ile Fo m Con inuous Fo m
Q1 Q2 Q3 Q4 p-T end aDoubling Resis in
Concen a ions bp-Value
Resis in qua ile anges
(ng/mL) ≤3.47 3.47< o ≤4.28 4.28< o ≤5.42
5.42< o
≤
34.41
Rec al Cance
No. cases/con ols 120/134 119/115 127/118 109/108 475/475
Model 1 e 1.15 (0.81–1.64) 1.20 (0.85–1.71) 1.13 (0.78–1.64) 0.76 1.07 (0.82–1.39) 0.62
Model 2 e 1.17 (0.81–1.68) 1.26 (0.88–1.82) 1.20 (0.81–1.76) 0.64 1.09 (0.83–1.43) 0.55
Model 3 e 1.16 (0.80–1.68) 1.25 (0.86–1.80) 1.17 (0.79–1.73) 0.69 1.06 (0.80–1.41) 0.66
Model 1: Condi ioned on ma ching ac o s only: age, sex, s udy cen e , ime o he day a blood collec ion, and
as ing s a us. Women we e u he ma ched by menopausal s a us, phase o he mens ual cycle, and use o o al
con acep i es a blood collec ion, and pos menopausal women we e ma ched by ho mone eplacemen he apy
use. Model 2: Model 1 + smoking s a us, educa ion, alcohol consump ion, alcohol abs aine s, physical ac i i y
index, ene gy in ake, ed mea , p ocessed mea , die a y ibe , ui in ake, ege able in ake, dai y in ake, ish, and
shell ish in ake. Model 3: Model 2 + body mass index (BMI), and esiduals o BMI-adjus ed wais ci cum e ence.
a
p- alues o end de i ed om models wi h he median esis in concen a ion wi hin qua iles as a con inuous
a iable. bModels wi h con inuous log- ans o med esis in concen a ions by log 2.
Sex-s a i ied analyses showed no signi ican associa ion be ween esis in and CRC
(Table 4). The es o in e ac ion by sex was no s a is ically signi ican in all s udy pa ici-
pan s (p= 0.86), colon cance pa ien s and hei pai s (p= 0.72), and ec al cance pa ien s
and hei pai s (p= 0.10). Howe e , among men, doubling esis in concen a ions we e
ela ed o 1.53- old he isk o ec al cance (95% CI: 1.01–2.33).
In s a i ied analyses, he e was no associa ion be ween esis in wi h CRC among
pe sons wi h BMI < 25 kg/m
2
o BMI
≥
25 kg/m
2
(Table S2), wi h o wi hou baseline
diabe es, C-pep ide
≥
2 ng/mL o <2 ng/mL (da a no shown). Among pe sons wi h hsCRP
≥
3 mg/L, pe sons wi h highe compa ed o hose wi h lowe esis in concen a ions had
highe ela i e isks o CRC, bu hese associa ions we e no s a is ically signi ican (RR pe
doubling esis in in pe sons wi h hsCRP
≥
3 mg/L, 1.31; 95% CI: 0.94–1.81). No associa ion
was obse ed o doubling esis in concen a ions among indi iduals wi h hsCRP < 3 mg/L
(Table S2). We obse ed an inc eased isk o CRC associa ed wi h esis in among hose
wi h mo e han 6 h o as ing as compa ed o hose wi h less han 6 h o as ing (RR pe
doubling, 1.45; 95% CI: 1.03–2.03; p= 0.03).
When we es ic ed he main analysis o cases and ma ched con ols who we e di-
agnosed wi hin he i s 2 yea s o ollow-up, we ound ha pa icipan s in he highes
as compa ed o he lowes qua ile o esis in concen a ions had a 1.97- old isk o CRC
(95% CI: 1.06–3.64; p- end < 0.001); RR pe doubling o esis in, 1.44; 95% CI: 0.97–2.12;
Supplemen a y Table S2. In con as , when excluding cases and ma ched con ols ha
we e diagnosed wi hin he i s 2 yea s o ollow-up, he RR in he highes e sus lowes
qua ile was 1.05; 95% CI: 0.81–1.37; p- end = 0.79; RR pe doubling, 1.03; 95% CI: 0.86–1.24;
Supplemen a y Table S2.
The main esul s we e no subs an ially di e en when pe o ming comple e case anal-
yses o when excluding pa icipan s wi h ex eme esis in le els (
Supplemen a y Table S2
).
All indings om subg oup analyses (including analyses by he combina ion o sex and
umo subsi e, by as ing s a us) we e no longe s a is ically signi ican when pa icipan s
who we e diagnosed wi hin he i s wo yea s o ollow-up we e excluded (da a no shown).
In his analysis, he ela i e isk o ec al cance pe doubling esis in concen a ions among
men was 1.13; 95% CI: 0.70–1.82.
When we combined ou esul s wi h hose published om he WHI, he pooled RRs o
he highes e sus he lowes qua ile o esis in concen a ions we e 1.10; 95% CI: 0.93–1.29
o all s udy pa icipan s combined and 1.09; 95% CI: 0.86–1.39 o pos menopausal women.
No signi ican he e ogenei y was ound (I2 = 0.0%, p= 0.77, iden ical in bo h me a-analyses).