Subclinical inflammation associated with prolonged TIMP-1 upregulation and arterial stiffness after gestational diabetes mellitus: a hospital-based cohort study

Vilmi‑Ke älä e al. Ca dio asc Diabe ol (2017) 16:49 DOI 10.1186/s12933‑017‑0530‑x ORIGINAL INVESTIGATION Subclinical in lamma ion associa ed wi hp olonged TIMP‑1 up egula ion anda e ial s i ness a e ges a ional diabe es melli us: a hospi al‑based coho s udy Tiina Vilmi‑Ke älä1,2* , Anneli Lauhio3,4,5, Taina Te aha iala6, Ou i Palomäki2, Jukka Uo ila1,2, Timo So sa6,7 and A i Palomäki1,8 Abs ac Backg ound: Ges a ional diabe es melli us (GDM) has signi ican implica ions o he u u e heal h o he mo he . Some clinical s udies ha e sugges ed subclinical in lamma ion and ascula dys unc ion a e GDM. We aimed o s udy whe he concen a ions o high‑sensi i i y C‑ eac i e p o ein (hsCRP), issue inhibi o o me allop o einase‑1 (TIMP‑1), ma ix me allop o einase‑8 (MMP‑8) and ‑9, as well as alues o a e ial s i ness di e be ween women wi h and wi hou a his o y o GDM a ew yea s a e deli e y. We also in es iga ed possible e ec s o obesi y on he esul s. Me hods: We s udied wo coho s—120 women wi h a his o y o GDM and 120 con ols—on a e age 3.7 yea s a e deli e y. Se um concen a ions o hsCRP we e de e mined by immunonephelome ic and immuno u bidi‑ me ic me hods, MMP‑8 by immuno luo ome ic assay, and MMP‑9 and TIMP‑1 by enzyme‑linked immunoso b‑ en assays. Pulse wa e eloci y (PWV) was de e mined using he oo ‑ o‑ oo eloci y me hod om ca o id and emo al wa e o ms by using a SphygmoCo de ice. A e ial compliance was measu ed non‑in asi ely by an HDI/ PulseWa e™CR‑2000 a e ial onome e . All 240 women we e also included in subg oup analyses o s udy he e ec o obesi y on he esul s. Mul iple linea eg ession analyses we e pe o med wi h adjus men o con ounding ac o s. Resul s: PWV a e p egnancy complica ed by GDM was signi ican ly highe han a e no mal p egnancy, 6.44 ± 0.83 (SD) s. 6.17 ± 0.74 m/s (p = 0.009). P e ious GDM was also one o he signi ican de e minan s o PWV in mul iple linea eg ession analyses. On he o he hand, compliance indices o bo h la ge (p = 0.092) and small (p = 0.681) a e ies did no di e be ween he s udy coho s. Se um TIMP‑1 le els we e signi ican ly inc eased a e p e ious GDM (p = 0.020). Howe e , no di e ences we e ound in he se um le els o MMP‑8, MMP‑9 o hsCRP. In subg oup analyses, he e we e signi ican ly highe concen a ions o hsCRP (p = 0.015) and highe PWV (p < 0.001) among obese women compa ed wi h non‑obese ones. Conclusions: PWV alues we e signi ican ly highe a e GDM compa ed wi h no moglycemic p egnancies and we e associa ed wi h p olonged TIMP‑1 up egula ion. Ca dio ascula isk ac o s we e mo e common in pa icipan s wi h high BMI han in hose wi h p e ious GDM. Keywo ds: A e ial compliance, Ges a ional diabe es melli us, High‑sensi i i y C‑ eac i e p o ein, Ma ix me allop o einase‑8, Ma ix me allop o einase‑9, Pulse wa e eloci y, Subclinical in lamma ion, Tissue inhibi o o ma ix me allop o einase‑1 © The Au ho (s) 2017. This a icle is dis ibu ed unde he e ms o he C ea i e Commons A ibu ion 4.0 In e na ional License (h p://c ea i ecommons.o g/licenses/by/4.0/), which pe mi s un es ic ed use, dis ibu ion, and ep oduc ion in any medium, p o ided 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 license, and indica e i changes we e made. The C ea i e Commons Public Domain Dedica ion wai e (h p://c ea i ecommons.o g/ publicdomain/ze o/1.0/) applies o he da a made a ailable in his a icle, unless o he wise s a ed. Open Access Ca dio ascula Diabe ology *Co espondence: iina. ilmi‑ke ala@khshp. i 2 Depa men o Obs e ics and Gynecology, Tampe e Uni e si y Hospi al, Box 2000, 33521 Tampe e, Finland Full lis o au ho in o ma ion is a ailable a he end o he a icle Page 2 o 10 Vilmi‑Ke älä e al. Ca dio asc Diabe ol (2017) 16:49 Backg ound In de eloped coun ies, he p e alence o ges a ional diabe es melli us (GDM) has inc eased apidly in ecen decades, along wi h inc easing a es o obesi y [1, 2]. In Finland, GDM complica ed 15.9% o p egnancies in 2015 [2]. A diagnosis o GDM has signi ican implica ions o he u u e heal h o he mo he . Fo ins ance, GDM has been shown o be associa ed wi h pos pa um insulin esis ance, hype ension, and dyslipidemia [3–5], placing a ec ed women a isk o me abolic synd ome (Me S), ype 2 diabe es melli us (T2DM) and/o ca dio ascu- la disease (CVD) la e in li e [5–8]. Incidence o CVD e en s, and speci ically hose o co ona y a e y disease, is known o be inc eased in women wi h p e ious GDM, e en in he absence o T2DM [8]. Clinical s udies ha e also e ealed subclinical in lamma ion and ascula dys- unc ion a e GDM [4]. High-sensi i i y C- eac i e p o ein (hsCRP) is a well- known acu e-phase p o ein and a sensi i e bioma ke o sys emic in lamma ion. Ele a ed le els o hsCRP a e a signi ican isk ac o o a he oscle osis [9]. The g oup o ma ix me allop o einases (MMPs) comp ises o e 20 s uc u ally and unc ionally ela ed bu gene ically dis inc membe s [10, 11]. Exp ession and ac i i y a e no mally low, bu inc eased in many pa hophysiologi- cal condi ions. MMPs can modula e immunological esponses, and MMPs can be ei he de ensi e o des uc- i e [11]. Bo h up egula ion and down- egula ion o MMP-8 and -9 ha e been associa ed wi h se e al nonin- ec ious as well in ec ious in lamma o y s a es [12–18]. MMP-8 may also egula e blood p essu e [19]. MMPs and hei inhibi o s, issue inhibi o s o MMPs (TIMPs) ha e been ela ed o a he oscle osis de elopmen and p og ession in humans [20–22]. I has been sugges ed ha imbalanced concen a ions o MMP amily membe s and TIMPs e en ually exe an impo an ole in ca dio- ascula isk [21–25]. In lamma ion may be pa hogenic, by inducing ascu- la dys unc ion [4, 26]. A e ial s i ness has p o en o be an impo an pa ame e o he assessmen o ca - dio ascula isk, and i has ea lie been associa ed wi h endo helial dys unc ion [27, 28]. Ca o id o emo al pulse wa e eloci y (PWV) has eme ged as he gold s anda d o assess a e ial s i ness [29]. When he a e ies a e s i o less dis ensible, PWV inc eases [30, 31]. PWV inc eases p opo ionally o he numbe o ca dio ascula isk ac- o s p esen , such as diabe es o Me S [27, 32, 33]. In epi- demiological s udies, inc eased PWV has been p edic i e o ca dio ascula e en s [29]. Recen ly, he implica ions o GDM as ega ds wom- en’s u u e heal h ha e been widely discussed. As he p e alence o GDM has inc eased o e he yea s, a be e unde s anding o he connec ions be ween p e ious GDM and bo h subclinical in lamma ion and ascula dys unc ion would be o g ea bene i . In addi- ion, ecen ly i has been sugges ed ha MMP-8 is associ- a ed wi h insulin ecep o deg ada ion, and high se um MMP-8 le els wi h an inc eased isk o diabe es melli us ype II [17]. In p e ious s udies se um le els o MMP-8, -9, TIMP-1 and hsCRP ha e been shown o be bioma k- e s e lec ing low-g ade in lamma ion [11, 23, 24, 34, 35]. In addi ion, TIMP-1 has been shown o exe MMP-inde- penden ac ions such as p o-in lamma o y and g ow h- ac o -like p ope ies [36–38]. Wi h his backg ound ou aim was o de ine whe he o no ca dio ascula isk, assessed by se um concen a- ions o hsCRP, MMP-8, MMP-9 and TIMP-1, and alues o a e ial compliance and PWV a e enhanced al eady a ew yea s a e GDM. We also e alua ed he e ec o obe- si y on he esul s. Me hods In his ollow-up s udy o wo coho s, a o al o 120 women wi h a his o y o GDM du ing he index p eg- nancy we e compa ed wi h 120 age-ma ched women wi h no mal glucose me abolism du ing p egnancy. The ime om he index p egnancy o he ollow-up s udy was also ma ched be ween he s udy g oups. All pa - icipan s had deli e ed on a e age 3.7 ( ange 2–6) yea s ea lie a Kan a-Häme Cen al Hospi al, Finland, i.e. a e he publica ion o Finnish Cu en Guidelines o sc eening GDM. Ou na ional guidelines we e published in 2008 and upda ed in 2013 wi hou any change in he diagnos ic c i e ia o GDM [39]. The comple e inclu- sion and exclusion c i e ia, wi h powe analysis, ha e been desc ibed ea lie [40]. B ie ly, GDM was de ined (using he diagnos ic c i e ia o Finnish Cu en Guide- lines) as a pa hological alue in a 2-h 75-g o al glucose ole ance es (OGTT) du ing p egnancy: enous plasma glucose ≥5.3 mmol/L when as ing, ≥10.0 mmol/L a 1h o ≥8.6mmol/L a 2h [39]. Ou na ional diagnos- ic h esholds o GDM a e simila o hose o he In e - na ional Associa ion o Diabe es and P egnancy S udy G oups (IADPSG): plasma glucose ≥5.1mmol/L when as ing, ≥10.0mmol/L a 1h o ≥8.5mmol/L a 2h [41]. Only single on p egnancies we e included. Women we e excluded i hey had ype 1 o ype 2 diabe es be o e he p egnancy, i hey we e p egnan a ime o he s udy, i hey had suspec ed o e i ied malignan o endoc ine disease, i he e was subs ance abuse o ea men , o a known clinical his o y o psychia ic illness. Con ols had o ha e no mal OGTT esul s du ing p egnancy. I he con ols had expe ienced GDM in an ea lie p egnancy, o he weigh o he newbo n was ≥4.5kg, hey we e excluded. The elec onic da abase o he hospi al was used o pick up he cases and con ols. Bo h ec ui men Page 3 o 10 Vilmi‑Ke älä e al. Ca dio asc Diabe ol (2017) 16:49 and examina ions we e accomplished be ween Augus 2011 and July 2014. We in e iewed he pa icipan s as ega ds hei li e- s yle habi s. Li e ime obacco exposu e was es ima ed as pack-yea s, and one pack-yea was de ined as 20 ciga e es smoked e e y day o 1yea [42]. Fu he , we in e iewed he pa icipan s as ega ds hei his o y o auma o in ec ious diseases du ing he p e ious mon h. We measu ed es ing hea a e, b achial blood p es- su e, weigh (kg) and heigh (cm) o he pa icipan s, and calcula ed body mass index (BMI): weigh in kilog ams di ided by heigh in me e s squa ed (kg/m2). The s udy was conduc ed in acco dance wi h he e hi- cal p inciples ou lined in he Decla a ion o Helsinki [43], and he p o ocol was app o ed by he E hics Commi - ee o Kan a-Häme Hospi al Dis ic ( e e ence numbe 521/2010; da e o app o al 21.12.2010). E e y pa icipan was gi en bo h o al and w i en in o ma ion on he s udy be o e she signed an in o med consen documen . Labo a o y me hods Se um samples we e collec ed a e a leas 12h o as - ing and s o ed a −80°C un il analyzed. Se um concen- a ions o hsCRP we e analyzed acco ding o alida ed immunonephelome ic (Uni ed Medix Labo a o ies L d., Espoo, Finland) and immuno u bidime ic (VITA Heal hca e Se ices L d., Vi a Labo a o y, Helsinki, Finland) me hods [44, 45]. Concen a ions o MMP-8 we e de e mined by immuno luo ome ic assay (IFMA) (Medix Biochemica, Espoo, Finland), as p e iously desc ibed [25]. Se um le els o MMP-9 and TIMP-1 we e analyzed by enzyme-linked immunoso ben assay (ELISA) using comme cial ki s (Bio ak ELISA Sys em; Ame sham Biosciences, GE Heal hca e, Buckingham- shi e, UK) and acco ding o he manu ac u e ’s ins uc- ions [18]. Fas ing se um le els o o al choles e ol (TC) and insulin we e analyzed acco ding o alida ed me h- ods as desc ibed in de ail ea lie [40]. De e mina ion o a e ial compliance andpulse wa e eloci y Th ee expe ienced nu ses measu ed he compliance o la ge and small a e ies a e a leas 10min o es in a semi-si ing posi ion. The eco ding was ca ied ou a e an o e nigh as . The pa icipan s we e asked o e ain om ea ing, ha ing ca eina ed d inks, smoking and aking medica ion o 12h, and d inking alcohol o 2days p io o measu emen . Radial a e y pulse wa es we e eco ded non-in asi ely wi h an a e ial onom- e e (HDI/PulseWa e™CR-2000, Hype ension Diagnos- ics, Inc., Eagan, Minneso a, USA) and he p ocedu e in ol es he use o a modi ied Windkessel pulse-con ou me hod [46]. Blood olume ine ia and sys emic ascula esis ance a e used o analyze a e ial compliance. The capaci i e compliance o la ge a e ies (C1), including he ao a, and he endo helial unc ion o small a e ies (C2) we e au oma ically assessed as a mean o he i e mos simila pulse wa es appea ing du ing 30-s o measu e- men . Th ee consecu i e measu emen s we e pe o med o ob ain mean esul s o e e y pa icipan . Ca o id- emo al PWV was measu ed using he oo - o- oo eloci y me hod om ca o id and emo al wa e- o ms by employing a SphygmoCo de ice (A Co Medical, Sydney, Aus alia). T anscu aneous eadings we e ob ained a he igh common ca o id a e y and he igh emo al a e y wi h he subjec s in a supine posi ion wi h di ec -con ac pulse senso s. The ime delay (D o ansi ime) o he wo wa e o ms was egis e ed, and he dis ance (D) be ween ca o id and emo al eco ding si es was ob ained by sub ac ing he ca o id measu e- men si e o s e nal no ch dis ance om he s e nal no ch o he emo al measu emen si e dis ance. PWV was cal- cula ed as ollows: D/D (m/s) [29, 30]. Th ee measu e- men s we e pe o med o ob ain a e age esul s o e e y pa icipan . Only measu emen s ha me he au oma ic quali y con ol cu o we e used in he inal analysis. All he PWV measu emen s we e pe o med by wo expe i- enced nu ses. S a is ical analysis The da a we e analyzed by using IBM® SPSS® S a is ics Ve sion 23 so wa e (copy igh 2015). Va iables we e es ed o no mali y by way o Shapi o–Wilk o Kolmogo- o –Smi no es s, as app op ia e. Da a a e p esen ed as mean±s anda d de ia ion (SD) i no men ioned o he - wise. Di e ences in con inuous a iables be ween GDM pa icipan s and con ols we e s udied by using S uden ’s es in cases o no mali y and he Mann–Whi ney U es in cases o skewed dis ibu ion o measu emen s. All 240 women we e also included in subg oup analy- ses o s udy he e ec o obesi y on he esul s. Fo hese analyses, we di ided he whole s udy g oup in o ou sub- g oups acco ding o obesi y and p e ious GDM. Obe- si y was classi ied as BMI ≥30kg/m2 [47]. The clinical cha ac e is ics o hese ou subg oups we e s udied by way o one-way ANOVA in cases o no mali y and by using he K uskal–Wallis es in cases o non-no mali y. I he o e all p alue was signi ican , indi idual p alues be ween subg oups we e also calcula ed. Pos hoc anal- yses, wi h a conse a i e Bon e oni co ec ion ac o , we e pe o med in o de o co ec o mul iple es ing. The ela ionships be ween di e en ca dio ascula isk ac o s we e es ed by Pea son’s o Spea man’s co ela- ion analysis, as app op ia e. Fu he , we conduc ed uni a ia e linea eg ession analyses o hsCRP, MMP-8, TIMP-1, PWV and a e ial Page 4 o 10 Vilmi‑Ke älä e al. Ca dio asc Diabe ol (2017) 16:49 compliance index alues o ind possible associa ions wi h clinically ele an co a ia es. Then mul i a iable lin- ea analyses we e ca ied ou o examine whe he simple associa ions we e changed a e adjus men o po en ial con ounde s. Finally, s epwise mul iple linea eg ession analyses we e done o ind ou ele an co a ia es o inal models. The selec ed co a ia es in all o hese analyses we e age, BMI, p e ious GDM, ime a e he index p eg- nancy, pack-yea s o smoking, hea a e, sys olic blood p essu e, hsCRP, TC and as ing insulin. F-s a is ics was used o op imize he sequen ial a iable selec ion p oce- du e. A wo- ailed p obabili y alue o <0.05 was consid- e ed signi ican . Resul s The basic clinical cha ac e is ics o he s udy pa icipan s a e summa ized in Table1. The e we e no signi ican di - e ences be ween he wo coho s in sel - epo ed his- o y o espi a o y in ec ion, o he in ec ious disease o auma du ing he mon h be o e ollow-up labo a o y examina ions. Subclinical in lamma ion Se um TIMP-1 le els we e signi ican ly inc eased a e p e ious GDM (Table2). The e was a signi ican posi i e associa ion be ween p e ious GDM and TIMP-1 le els in bo h uni a ia e and mul i a iable linea eg ession analyses (da a no shown). The e we e no di e ences in he concen a ions o MMP-8 and MMP-9 be ween he g oups (Table2). In s epwise mul iple linea eg ession analyses, hsCRP, p e ious GDM and TC we e impo an de e minan s o MMP-8 le els. Likewise, p e ious GDM, oge he wi h BMI and hea a e associa ed wi h TIMP-1 in s epwise mul iple linea eg ession analyses. Ne e he- less, he signi ican de e minan s explained only 13.8% o MMP-8 and 6.7% o TIMP-1 concen a ions (Table3). We ound no di e ence in he concen a ions o hsCRP be ween GDM cases and con ols (Table2), e en when pa icipan s a ec ed wi h in ec ions o aumas we e excluded (da a no shown). In s epwise mul iple linea eg ession analysis (Table3), only BMI was a signi ican de e minan o hsCRP le els, bu he model explained only 9.6% o hsCRP alues. P e ious GDM did no in lu- ence hsCRP concen a ions in ou da a. Pulse wa e eloci y anda e ial compliance PWV alues di e ed signi ican ly be ween he GDM cases and con ols (Table2). In uni a ia e linea eg es- sion analysis, he e we e signi ican associa ions wi h age (p<0.001), as ing insulin (p<0.001), p e ious GDM (p=0.009), TC (p<0.001), hea a e (p<0.001), sys olic blood p essu e (p<0.001) and BMI (p<0.001). In s ep- wise mul iple linea eg ession analysis, signi ican de e - minan s o PWV alues we e sys olic BP, age, insulin le els, p e ious GDM and ime a e he index p egnancy. Co a ia es explained 47.0% o PWV (Table3). In ou wo s udy coho s, he e we e no in e ac ions be ween p e i- ous GDM and TIMP1 on PWV (da a no shown). The e was a nonsigni ican di e ence in C1 alues be ween he s udy g oups. No di e ence was e ealed in C2 alues, ei he . In uni a ia e linea eg ession analysis, Table 1 Basic clinical cha ac e is ics o women wi hGDM andcon ols Da a a e p esen ed as mean±SD i no men ioned o he wise BMI body mass index, BP blood p essu e, F-Gluc as ing glucose, F-Insu as ing insulin, TC o al choles e ol GDM Con ols p alue A e age ime since deli e y, yea s 3.7 ± 1.0 3.7 ± 0.9 0.818 Age, yea s 35.8 ± 4.4 35.9 ± 4.6 0.854 P imipa ous, n (%) 23 (19.2%) 23 (19.2%) 1.000 The apy o GDM du ing p egnancy Insulin, n (%) 24 (20.0%) Me o min, n (%) 1 (0.8%) Die a y he apy, n (%) 95 (79.2%) Pack‑yea s o smoking 3.8 ± 6.0 2.4 ± 4.6 0.012 Du ing he p e ious mon h, his o y o Respi a o y in ec ion, n (%) 45 (37.5%) 44 (36.7%) 0.854 O he in ec ious disease, n (%) 18 (15.0%) 10 (8.3%) 0.053 T auma, n (%) 9 (7.5%) 5 (4.2%) 0.264 BMI, kg/m228.3 ± 5.0 27.5 ± 5.4 0.069 Sys olic BP, mmHg 122.4 ± 12.5 119.0 ± 11.5 0.034 Dias olic BP, mmHg 73.5 ± 9.0 71.8 ± 8.7 0.176 Hea a e, bea s pe minu e 65.9 ± 9.1 63.8 ± 9.6 0.017 TC, mmol/L 4.7 ± 0.9 4.6 ± 0.8 0.329 F‑Gluc, mmol/L 5.6 ± 0.6 5.3 ± 0.3 <0.001 F‑Insu, mU/L 5.2 ± 3.6 4.6 ± 3.6 0.087 Table 2 Resul s o  p ima y analyses o  GDM and con ol g oups Da a a e p esen ed as mean±SD hsCRP high‑sensi i i y C eac i e p o ein, C1 la ge a e y compliance index, C2 small a e y compliance index, PWV pulse wa e eloci y, MMP-8 ma ix me allop o einase‑8, MMP-9 ma ix me allop o einase‑9 GDM Con ols p alue hsCRP, mg/L 2.50 ± 3.69 2.50 ± 4.19 0.582 MMP‑8, ng/mL 27.83 ± 1.48 32.78 ± 1.90 0.082 MMP‑9, ng/mL 384.27 ± 13.15 392.15 ± 12.60 0.667 TIMP‑1, ng/mL 102.80 ± 29.72 94.58 ± 24.51 0.020 MMP‑8/TIMP‑1, mol a io 0.13 ± 0.009 0.17 ± 0.015 0.035 MMP‑9/TIMP‑1, mol a io 1.32 ± 0.078 1.43 ± 0.085 0.152 C1, mL/mmHg × 10 15.14 ± 3.51 15.85 ± 3.36 0.092 C2, mL/mmHg × 100 8.44 ± 3.08 8.60 ± 3.20 0.681 PWV, m/s 6.44 ± 0.83 6.17 ± 0.74 0.009 Page 5 o 10 Vilmi‑Ke älä e al. Ca dio asc Diabe ol (2017) 16:49 he e was no signi ican associa ion be ween C2 and BMI (p=0.726), bu an in e se associa ion be ween C1 and BMI was signi ican (p=0.025). In s epwise mul iple lin- ea eg ession analysis, sys olic BP, hea a e, BMI and ime a e he index p egnancy we e signi ican co a i- a es explaining 52.4% o C1 alues. Signi ican de e mi- nan s o C2 alues we e sys olic BP, hea a e, BMI, age and pack-yea s o smoking. These co a ia es explained 31.7% o C2 alues (Table3). E ec o obesi y insubg oups Al oge he , he e we e 75 women in he obese g oup (BMI ≥ 30 kg/m2); 43 GDM and 32 con ol pa ici- pan s. The non-obese g oup (BMI<30kg/m2; n=165) consis ed o 77 GDM and 88 con ol pa icipan s [55]. In subg oup analyses, pa icipan s in obese subg oups had highe se um concen a ions o hsCRP han hose in non-obese subg oups, as shown in Fig.1. The con- cen a ions o MMP-8 in he ou subg oups we e as ollows: obese GDM cases, 27.76 ± 1.77 ng/mL, obese con ols 37.10 ± 4.16 ng/mL, non-obese GDM cases, 27.88 ± 2.08 ng/mL and non-obese con ols, 31.21 ± 2.10 ng/mL. The concen a ion o MMP-8 was highes among obese con ols, bu he di e - ences be ween he ou subg oups we e no signi ican (p=0.090). We also ound no di e ences in he le els o MMP-9 o TIMP-1 be ween hese ou subg oups (da a no shown). Be ween he subg oups, he e we e no di - e ences in he MMP-8/TIMP-1 o MMP-9/TIMP-1 a io ei he (da a no shown). In he ou subg oups, di e - ences in PWV alues we e signi ican , bu di e ences in bo h C1 and C2 alues we e no (Figs.2, 3). Table 3 Resul s o s epwise mul iple linea eg ession analyses Co a ia es in hese analyses included age, BMI, p e ious GDM, pack‑yea s o smoking, ime a e he index p egnancy, hea a e, sys olic blood p essu e, hsCRP, TC and as ing insulin. Final models include signi ican co a ia es only. S anda dized β p o ides a measu e o he ela i e s eng h o an associa ion, independen o he measu emen uni s. S anda dized β and p alues a e shown only when p<0.05 BMI body mass index, BP blood p essu e, F-Insu as ing insulin, GDM ges a ional diabe es melli us, hsCRP high‑sensi i i y C eac i e p o ein, C1 la ge a e y compliance index, C2 small a e y compliance index, PWV pulse wa e eloci y, MMP-8 ma ix me allop o einase‑8 Pa ame e s Co a ia es included in he model R2 o model Global pS anda dized β p alue hsCRP 0.096 <0.001 BMI 0.259 <0.001 MMP‑8 0.138 <0.001 hsCRP 0.312 <0.001 P e ious GDM −0.137 0.025 TC 0.129 0.036 TIMP‑1 0.067 0.003 P e ious GDM 0.157 0.015 BMI 0.149 0.025 Hea a e −0.132 0.044 C1 0.524 <0.001 Sys olic BP −0.602 <0.001 Hea a e −0.347 <0.001 BMI 0.232 <0.001 Time a e he index p egnancy −0.095 0.041 C2 0.317 <0.001 Sys olic BP −0.345 <0.001 Hea a e −0.312 <0.001 BMI 0.286 <0.001 Age −0.191 0.001 Pack‑yea s o smoking −0.144 0.012 PWV 0.470 <0.001 Sys olic BP 0.534 <0.001 Age 0.230 <0.001 F‑Insu 0.191 <0.001 P e ious GDM 0.105 0.026 Time a e he index p egnancy −0.102 0.040 Page 6 o 10 Vilmi‑Ke älä e al. Ca dio asc Diabe ol (2017) 16:49 Discussion Ou main inding was ha PWV was signi ican ly highe a e GDM han a e no moglycemic p egnancy. This was suppo ed by a nonsigni ican di e ence in he la ge-a e y compliance index, C1, which indica es ha he a e ies o GDM cases we e less dis ensible han hose o he con ols. Secondly, subclinical low-g ade in lamma ion and educed a e ial compliance especially a ec ed women wi h high BMI. In lamma ion has been shown o be a s ong p edic o o women’s ca dio ascula complica ions [48]. We ound ha le els o TIMP-1 we e signi ican ly up egula ed a e p e ious GDM, e lec ing low-g ade in lamma ion Fig. 1 Se um concen a ions o hsCRP in he ou subg oups. Median alues (minimum, maximum) o hsCRP: among obese GDM women 2.1 (0.0, 12.4) mg/mL, obese con ol women 2.1 (0.3, 18.5) mg/mL, non‑obese GDM women 0.9 (0.0, 32.3) mg/mL, and non‑obese con ol women 0.7 (0.0, 25.7) mg/mL. Values o mo e han 10 mg/mL we e measu ed by u bidime ic immunoassay. The o e all p alue is gi en a he bo om. Indi idual p alues o pai wise compa isons a e also p esen ed Fig. 2 PWV in he ou subg oups. Median alues (minimum, maximum) o PWV: among obese GDM women 6.8 (5.6, 9.7) m/s, obese con ol women 6.6 (4.8, 8.5) m/s, non‑obese GDM women 6.3 (4.9, 9.2) m/s, and non‑obese con ol women 6.0 (4.5, 7.9) m/s. The o e all p alue is gi en a he bo om. Indi idual p alues o pai wise compa isons a e also p esen ed Fig. 3 La ge (a) and small (b) a e y compliance index alues in he ou subg oups. a Median alues (minimum, maximum) o he la ge‑a e y compliance index (C1): among obese GDM women 13.3 (9.1, 21.8) mL/mmHg × 10, obese con ol women 14.7 (10.2, 23.5) mL/mmHg × 10, non‑obese GDM women 15.2 (7.2, 25.2) mL/ mmHg × 10, and non‑obese con ol women 15.9 (7.5, 25.7) mL/ mmHg × 10. The o e all p alue is gi en. b Median alues (minimum, maximum) o he small‑a e y compliance index (C2): among obese GDM women 8.8 (2.8, 15.2) mL/mmHg × 100, obese con ol women 8.6 (2.2, 17.7) mL/mmHg × 100, non‑obese GDM women 8.1 (1.8, 17.6) mL/mmHg × 100, and non‑obese con ol women 8.1 (2.4, 16.0) mL/mmHg × 100. The o e all p alue is gi en Page 7 o 10 Vilmi‑Ke älä e al. Ca dio asc Diabe ol (2017) 16:49 among his ela i ely heal hy and young s udy popula- ion. No di e ences we e ound in ci cula ing le els o MMP-8 o MMP-9 be ween he wo s udy coho s. In subg oup analyses, he highes le els o MMP-8 we e in obese con ols, bu his did no each s a is ical signi i- cance ei he . A sea ch o MEDLINE (English language; 1989–Sep embe 2016; sea ch e ms: “MMP-8, MMP-9, TIMP-1” and “GDM”) e ealed no publica ions conce n- ing emale popula ions whe e le els o MMP-8, MMP-9 o TIMP-1 ha e been s udied in connec ion wi h p e i- ous GDM. The e is e idence ha glucose can modula e he exp ession, p oduc ion and ac i i y o MMPs. Fo exam- ple, endo helial cells cul u ed in hype glycemic condi- ions p esen inc eased exp ession and ac i i y o MMP-9 [49]. I is a pi y ha he e we e no samples le o MMP analysis aken om he pa ien s du ing he pe iod when hey su e ed om ges a ional diabe es. We migh pos u- la e, ha du ing he p egnancy GDM inc ease concen- a ions o MMPs and hey in u n up egula e TIMP-1. A e he deli e y, he dec easing concen a ions o glu- cose, MMPs and TIMP-1 ake place consecu i ely. The p olonged up egula ion o TIMP-1 ound in his s udy wi hou up egula ed MMP le els may also be a esul o he ac ha up egula ed TIMP-1 may supp ess MMP-8 and MMP-9 le els. Fu he , hi d explana ion o p o- longed TIMP-1 up egula ion ound in his wo k may be ha p olonged ele a ion o TIMP-1 le els may media e MMP-independen p o-in lamma o y o g ow h- ac o - like signaling unc ions con ibu ing o low-g ade in lam- ma ion [36–38]. Recen s udies ha e epo ed highe CRP and hsCRP le els in women wi h a his o y o GDM han in age- ma ched no mal con ols a e a 1- o 5-yea pos pa - um pe iod [4, 50, 51]. On he con a y, Ajala e al. ound no di e ence in CRP in women a e p e ious GDM compa ed o con ols 4–10 yea s pos pa um [52]. In ou s udy, when hsCRP was de e mined on a e age a 3.7yea s a e deli e y, he e was no di e ence be ween he age-ma ched s udy coho s. Howe e , low-g ade in lamma ion was e iden among obese women, in con- as o non-obese pa icipan s in subg oup analyses. The GDM and non-GDM women o ou s udy did no di e in BMI, which can pa ly explain he simila hsCRP le els be ween he wo s udy coho s. Only a ew s udies ha e been published conce ning a possible ela ionship be ween PWV and p e ious GDM. Lek a e  al. epo ed an enhanced ca dio ascula isk a 5-yea ollow-up as e lec ed in ele a ed PWV a e p e ious GDM diagnosed using he old c i e ia o he Wo ld Heal h O ganiza ion (WHO) (OGTT: 2-h plasma glucose ≥7.8mmol/L). Howe e , hey did no ind such an associa ion in PWV when using IADPSG diagnos ic c i e ia (OGTT: as ing plasma glucose 5.1–6.9mmol/L, 1-h plasma glucose ≥10.0mmol/L o 2-h plasma glucose 8.5–11.0mmol/L) [41, 53]. Using diagnos ic c i e ia o GDM simila o hose o he IADPSG [39], we obse ed a signi ican inc ease in PWV in women wi h p e ious GDM. P e ious GDM was also a signi ican de e minan o PWV in mul iple linea eg ession analysis. Ou esul s a e in acco dance wi h hose o Tam e al., who epo ed highe PWV in women wi h a his o y o GDM ollowed up a a median o 6yea s pos pa um [54]. In con as o hese indings, Hei i e e al. de ec ed no di e ence in PWV a an a e age o 1yea a e p e ious GDM com- pa ed wi h no moglycemic p egnancy [4]. The e we e no signi ican di e ences in C1 o C2 alues be ween he GDM cases and con ols. In a ecen s udy, no di - e ence was ound in ascula unc ion measu ed also by using HDI/PulseWa e™CR-2000 in women wi h a his o y o GDM when compa ed o heal hy con ols 4–10yea s pos pa um, ei he [52]. S eng hs o ou s udy include he ac ha we used s anda dized measu emen s o a e ial s i ness. De e - mina ion o sys emic a e ial s i ness by using HDI/ PulseWa e™CR-2000 equipmen is widely used, and ca o id- emo al PWV is accep ed as he mos eliable measu emen o a e ial s i ness [29]. We measu ed he le els o MMP-8, MMP-9 and TIMP-1 by speci ic immu- noassays p e iously ound o be sui able o diagnosis and moni o ing o sys emic low-g ade in lamma ion associ- a ed wi h ca dio ascula and in ec ious diseases as well as o he in lamma o y s a es [11, 13–18, 23–25]. Fu he , we pe o med a well cha ac e ized hospi al-based s udy o wo coho s o women wi h a simila ollow-up ime and age. Mo eo e , he e was no signi ican di e ence in BMI be ween he s udy g oups, and all pa icipan s had unde gone OGTT sc eening du ing he index p egnancy. Since low- isk pa u ien s do no ou inely unde go OGTTs in Finland [39], his las s eng h may also u n ou o be a weakness, because he mos low- isk women had o be excluded om ou s udy [40]. Al hough he ela i ely sho ime om deli e y o he ollow-up s udy allowed us o obse e ea ly ca dio ascula changes, i may be one o ou s udy limi a ions as well, since majo di e ences be ween he s udy g oups a e p obably be e obse able la e in hei li e. Fo example, wi hin 7yea s pos pa um, p e ious GDM was iden i ied as a isk ac- o o CVD by Gouesla d e al. They s udied da abase o mo e han 1.5 million deli e ies and ound ha he inci- dence o myoca dial in a c ion was 0.04% in women wi h a his o y o GDM and 0.02% wi hou [7]. In ou subg oup analyses, obesi y was associa ed wi h highe le els o hsCRP and highe alues o PWV. We ha e ea lie e ealed he e ec o obesi y being simi- la wi h many o he ma ke s o ca dio-me abolic isks Page 8 o 10 Vilmi‑Ke älä e al. Ca dio asc Diabe ol (2017) 16:49 among he ou subg oups [40, 55]. Ea lie , BMI has been shown o associa e in e sely wi h a e ial compli- ance [56]. As p esen ed in Fig.3, his seemed o be he case also in ou s udy in C1 alues. Su p isingly, in mul- iple eg ession analyses, BMI seemed o be p o ec i e as ega ds a e ial compliance (C1 and C2). BMI was signi i- can ly co ela ed wi h sys olic blood p essu e and hea a e (da a no shown). Hence, adjus ed indings conce n- ing C1 and C2 migh ha e been a ec ed by hese ela- ionships i espec i e o possible biologic associa ions. In ou opinion, his esul may be explained by mul iple in e ac ions o C1 and C2 measu emen s wi h o he con- ounding a iables. This was suppo ed by he indings o uni a ia e analysis and s epwise mul iple linea eg es- sion analysis wi hou sys olic BP and hea a e as co a i- a es, whe e in e se associa ion be ween BMI and C1 was ound and associa ion be ween BMI and C2 was anished (da a no shown). The p e alence o obesi y is inc easing a ound he wo ld [57]. Speci ically, isce al obesi y modi ies glucose and lipid me abolism. I is associa ed wi h inc eased isk o a e ial s i ness and a he oscle osis bo h in no mal- weigh subjec s and pa ien s wi h T2DM [58, 59]. Ou esul s imply ha in p e en ing ca dio ascula isk among women a e deli e y, we need a comp ehensi e a i ude in clinical ca e ins ead o concen a ing on single ac o s. Conclusions When s udied 3.7yea s a e deli e y, PWV alues we e highe in women wi h p e ious GDM, indica ing ha hei a e ies a e less dis ensible han hose in women wi h p e ious no moglycemic p egnancy. Among o he indings, his ela ionship was e en mo e e iden in obese subjec s. We also ound ha se um le els o TIMP-1 we e signi ican ly up egula ed a e p e ious GDM, e lec ing low-g ade in lamma ion among his ela i ely heal hy and young s udy popula ion. Al oge he , ou esul s dem- ons a e ha p e ious GDM may e lec a subclinical in lamma o y s a e and oge he wi h obesi y may con- ibu e o an ea ly s age o he subclinical a he oscle o ic p ocess e en in ela i ely young and heal hy women. Abb e ia ions C1: la ge a e y compliance index; C2: small a e y compliance index; GDM: ges a ional diabe es melli us; hsCRP: high‑sensi i i y C‑ eac i e p o ein; MMP: ma ix me allop o einase; OGTT: o al glucose ole ance es ; PWV: pulse wa e eloci y; TIMP: issue inhibi o o me allop o einase. Au ho s’ con ibu ions TV‑K pa icipa ed in he design o he s udy, conduc ed expe imen s, pe o med da a analyses and d a ed he manusc ip . AL pa icipa ed in he design o he s udy and con ibu ed o d a ing he manusc ip . TT ca ied ou he analyses o MMP‑8, MMP‑9 and TIMP‑1. OP, JU and TS con ibu ed o d a ing he manusc ip . AP designed he s udy, helped o pe o m da a analyses and d a ed he manusc ip . All au ho s ead and app o ed he inal manusc ip . Au ho de ails 1 School o Medicine, Uni e si y o Tampe e, Tampe e, Finland. 2 Depa ‑ men o Obs e ics and Gynecology, Tampe e Uni e si y Hospi al, Box 2000, 33521 Tampe e, Finland. 3 Depa men o In ec ious Diseases, In lamma ion Cen e , Helsinki Uni e si y Hospi al, Helsinki, Finland. 4 Clinicum, Uni e si y o Helsinki, Helsinki, Finland. 5 The Social Insu ance Ins i u ion o Finland, Ben‑ e i Se ices, Helsinki, Finland. 6 Depa men o O al and Maxillo acial Diseases, Helsinki Uni e si y and Uni e si y Hospi al, Helsinki, Finland. 7 Di ision o Pe i‑ odon ology, Depa men o Den al Medicine, Ka olinska Ins i u e , Huddinge, Sweden. 8 Depa men o Eme gency Medicine, Kan a‑Häme Cen al Hospi al, Hämeenlinna, Finland. Acknowledgemen s We app ecia e he p o essional echnical aid o Anna Silén, Ta u S anden, Hanna Kujanen, A i Vi a, Nick Bol on, Ki s i Räsänen and Piia Suu salmi. We since ely acknowledge he wo k o he clinical s a o Linnan Klinikka and Kan a‑Häme Cen al Hospi al. Compe ing in e es s The au ho s decla e ha hey ha e no compe ing in e es s. A ailabili y o da a and ma e ials The da ase s gene a ed and analyzed du ing he cu en s udy a e no publicly a ailable as a esul o he ac ha indi idual p i acy could be comp omised, bu a e a ailable om he co esponding au ho on easonable eques . Consen o publica ion E e y pa icipan was gi en bo h o al and w i en in o ma ion on he s udy be o e she signed an in o med consen documen . E hics app o al and consen o pa icipa e The s udy was conduc ed in acco dance wi h he e hical p inciples ou lined in he Decla a ion o Helsinki, and he p o ocol was app o ed by he E hics Com‑ mi ee o Kan a‑Häme Hospi al Dis ic (Re e ence Numbe 521/2010; da e o app o al 21.12.2010). Funding This s udy was suppo ed by g an s om he Finnish Cul u al Founda ion, Häme Regional Fund and he Minis y o Heal h and Social Wel a e in Finland ia Medical Resea ch Funds o Kan a‑Häme Cen al Hospi al, Tampe e Uni e ‑ si y Hospi al, Helsinki Uni e si y Hospi al EVO and Ka olinska Ins i u e . Publishe ’s No e Sp inge Na u e emains neu al wi h ega d o ju isdic ional claims in pub‑ lished maps and ins i u ional a ilia ions. Recei ed: 16 Janua y 2017 Accep ed: 4 Ap il 2017 Re e ences 1. Dabelea D, Snell‑Be geon JK, Ha s ield CL, Bischo KJ, Hamman RF, McDu ie RS. Kaise Pe manen e o Colo ado GDM Sc eening P og am: inc easing p e alence o ges a ional diabe es melli us (GDM) o e ime and by bi h coho : Kaise Pe manen e o Colo ado GDM Sc eening P og am. Diabe es Ca e. 2005;28(3):579–84. 2. Vuo i E, Gissle M. Pe ina al s a is ics: pa u ien s, deli e ies and newbo ns 2015. S a is ical epo 16/2016. Helsinki: Na ional Ins i u e o Heal h and Wel a e; 2016. 3. Gobl CS, Bozku L, Ya agudi R, P ikoszo ich T, Tu a A, Pacini G, Kop‑ pens eine R, Kau zky‑Wille A. Bioma ke s o endo helial dys unc ion in ela ion o impai ed ca bohyd a e me abolism ollowing p egnancy wi h ges a ional diabe es melli us. Ca dio asc Diabe ol. 2014;13(1):138. 4. Hei i e SM, Solomon CG, Mi chell GF, Skali‑Ounis N, Seely EW. Subclini‑ cal in lamma ion and ascula dys unc ion in women wi h p e ious ges a ional diabe es melli us. J Clin Endoc inol Me ab. 2005;90(7):3983–8. 5. V achnis N, Augoulea A, Iliod omi i Z, Lamb inoudaki I, Si akis S, C ea sas G. P e ious ges a ional diabe es melli us and ma ke s o ca dio ascula isk. In J Endoc inol. 2012;2012:458610. Page 9 o 10 Vilmi‑Ke älä e al. Ca dio asc Diabe ol (2017) 16:49 6. Shah BR, Re naka an R, Boo h GL. Inc eased isk o ca dio ascula disease in young women ollowing ges a ional diabe es melli us. Diabe es Ca e. 2008;31(8):1668–9. 7. Gouesla d K, Co ene J, Ma ie AS, Gi oud M, Co in Y, Pe i JM, Quan in C. Ea ly ca dio ascula e en s in women wi h a his o y o ges a ional diabe es melli us. Ca dio asc Diabe ol. 2016;15:15. 8. Re naka an R, Shah BR. Role o ype 2 diabe es in de e mining e inal, enal, and ca dio ascula ou comes in women wi h p e ious ges a ional diabe es melli us. Diabe es Ca e. 2017;40(1):101–8. 9. Ka adeniz M, Du an M, Akyel A, Ya lioglues M, Ocek AH, Celik IE, Kilic A, Yalcin AA, E gun G, Mu a SN. High sensi i e c p le el is associa ed wi h in e media e and high syn ax sco e in pa ien s wi h acu e co ona y synd ome. In Hea J. 2015;56(4):377–80. 10. Lengle S, Mach F, Mon ecucco F. Role o ma ix me allop o einase‑8 in a he oscle osis. Media In lamm. 2013;2013:659282. 11. So sa T, Tjade hane L, Kon inen YT, Lauhio A, Salo T, Lee HM, Golub LM, B own DL, Män ylä P. Ma ix me allop o einases: con ibu ion o pa ho‑ genesis, diagnosis and ea men o pe iodon al in lamma ion. Ann Med. 2006;38(5):306–21. 12. Lauhio A, Salo T, Ding Y, Kon inen YT, No ds öm D, Tschesche H, Lähde i a J, Golub LM, So sa T. In i o inhibi ion o human neu ophil collagenase (MMP‑8) ac i i y du ing long‑ e m combina ion he apy o doxycycline and non‑s e oidal an i‑in lamma o y d ugs (NSAID) in acu e eac i e a h i is. Clin Exp Immunol. 1994;98(1):21–8. 13. Lauhio A, Kon inen YT, Tschesche H, No ds öm D, Salo T, Lähde i a J, Golub LM, So sa T. Reduc ion o ma ix me allop o einase 8‑neu ophil collagenase le els du ing long‑ e m doxycycline ea men o eac i e a h i is. An imic ob Agen s Chemo he . 1994;38(2):400–2. 14. Lauhio A, Salo T, Tjade hane L, Lähde i a J, Golub LM, So sa T. Te acy‑ clines in ea men o heuma oid a h i is. Lance . 1995;346(8975):645–6. 15. Lauhio A, Saikku P, Salo T, Tschesche H, Lähde i a J, So sa T. Combina ion ea men in Chlamydia‑ igge ed eac i e a h i is: commen on he a icle by Ca e e al. A h i is Rheum. 2011;63(1):305–7 (au ho eply 307–8). 16. Lauhio A, Häs backa J, Pe ilä V, Te aha iala T, Ka lsson S, Va pula T, Va pula M, Ruokonen E, So sa T, Kolho E. Se um MMP‑8, ‑9 and TIMP‑1 in sepsis: high se um le els o MMP‑8 and TIMP‑1 a e associa ed wi h a al ou come in a mul icen e, p ospec i e coho s udy. Hypo he ical impac o e acyclines. Pha macol Res. 2011;64(6):590–4. 17. Lauhio A, Fä kkilä E, Pie iläinen KH, Ås öm P, Winkelmann A, Te a‑ ha iala T, Pi ilä E, Rissanen A, Kap io J, So sa TA, Salo T. Associa ion o MMP‑8 wi h obesi y, smoking and insulin esis ance. Eu J Clin In es ig. 2016;46(9):757–65. 18. Rau elin HI, Oksanen AM, Veijola LI, Sipponen PI, Te aha iala TI, So sa TA, Lauhio A. Enhanced sys emic ma ix me allop o einase esponse in Helicobac e pylo i gas i is. Ann Med. 2009;41(3):208–15. 19. Cena JJ, Lalu MM, Cho WJ, Chow AK, Bagdan ML, Daniel EE, Cas o MM, Schulz R. Inhibi ion o ma ix me allop o einase ac i i y in i o p o ec s agains ascula hypo eac i i y in endo oxemia. Am J Physiol Hea Ci c Physiol. 2010;298(1):H45–51. 20. Siasos G, Tousoulis D, Kiou is S, Oikonomou E, Siasou Z, Limpe i M, Papa assiliou AG, S e anadis C. In lamma o y mechanisms in a he o‑ scle osis: he impac o ma ix me allop o einases. Cu Top Med Chem. 2012;12(10):1132–48. 21. Paim LR, Sch eibe R, Ma os‑Souza JR, Sil a AA, Campos LF, Aze edo ER, Alonso K, de Rossi G, E chebehe e M, Go la JI, Clique A J , Nad uz W J . Oxidized low‑densi y lipop o ein, ma ix‑me allop o einase‑8 and ca o id a he oscle osis in spinal co d inju ed subjec s. A he oscle osis. 2013;231(2):341–5. 22. Goncal es FM, Jacob‑Fe ei a AL, Gomes VA, Casella‑Filho A, Chagas AC, Ma caccini AM, Ge lach RF, Tanus‑San os JE. Inc eased ci cula ing le els o ma ix me allop o einase (MMP)‑8, MMP‑9, and p o‑in lam‑ ma o y ma ke s in pa ien s wi h me abolic synd ome. Clin Chim Ac a. 2009;403(1–2):173–7. 23. Pussinen PJ, Sa na S, Puolakkainen M, Ohlin H, So sa T, Pesonen E. The bal‑ ance o se um ma ix me allop o einase‑8 and i s issue inhibi o in acu e co ona y synd ome and i s ecu ence. In J Ca diol. 2013;167(2):362–8. 24. So sa T, Te aha iala T, Leppilah i J, He nandez M, Gamonal J, Tuomainen AM, Lauhio A, Pussinen PJ, Män ylä P. Collagenase‑2 (MMP‑8) as a poin ‑ o ‑ca e bioma ke in pe iodon i is and ca dio ascula diseases. The apeu‑ ic esponse o non‑an imic obial p ope ies o e acyclines. Pha macol Res. 2011;63(2):108–13. 25. Tuomainen AM, Nyyssönen K, Laukkanen JA, Te aha iala T, Tuomainen TP, Salonen JT, So sa T, Pussinen PJ. Se um ma ix me allop o einase‑8 concen a ions a e associa ed wi h ca dio ascula ou come in men. A e ioscle Th omb Vasc Biol. 2007;27(12):2722–8. 26. Meigs JB, Hu FB, Ri ai N, Manson JE. Bioma ke s o endo helial dys unc‑ ion and isk o ype 2 diabe es melli us. JAMA. 2004;291(16):1978–86. 27. Bou ouy ie P, T opeano AI, Asma R, Gau ie I, Bene os A, Lacolley P, Lau en S. Ao ic s i ness is an independen p edic o o p ima y co ona y e en s in hype ensi e pa ien s: a longi udinal s udy. Hype ension. 2002;39(1):10–5. 28. Nigam A, Mi chell GF, Lambe J, Ta di J. Rela ion be ween condui essel s i ness (assessed by onome y) and endo helial unc ion (assessed by low‑media ed dila a ion) in pa ien s wi h and wi hou co ona y hea disease. Am J Ca diol. 2003;92(4):395–9. 29. Lau en S, Cockc o J, Van Bo el L, Bou ouy ie P, Gianna asio C, Hayoz D, Pannie B, Vlachopoulos C, Wilkinson I, S uijke ‑Boudie H. Eu opean ne wo k o non‑in asi e in es iga ion o la ge a e ies: expe consensus documen on a e ial s i ness: me hodological issues and clinical appli‑ ca ions. Eu Hea J. 2006;27(21):2588–605. 30. Agabi i‑Rosei E, Mancia G, O’Rou ke MF, Roman MJ, Sa a ME, Smulyan H, Wang J, Wilkinson IB, Williams B, Vlachopoulos C. Cen al blood p essu e measu emen s and an ihype ensi e he apy. A consensus documen . Hype ension. 2007;50:154–60. 31. Nelson MR, S epanek J, Ce e e M, Co alciuc M, Hu s RT, Tajik J. Nonin a‑ si e measu emen o cen al ascula p essu es wi h a e ial onom‑ e y: clinical e i al o he pulse p essu e wa e o m? Mayo Clin P oc. 2010;85(5):460–72. 32. Kim YK. Impac o he me abolic synd ome and i s componen s on pulse wa e eloci y. Ko ean J In e n Med. 2006;21(2):109–15. 33. C uickshank K, Ris e L, Ande son SG, W igh JS, Dunn G, Gosling RG. Ao ‑ ic pulse‑wa e eloci y and i s ela ionship o mo ali y in diabe es and glucose in ole ance: an in eg a ed index o ascula unc ion? Ci cula ion. 2002;106(16):2085–90. 34. Ko mi I, Al ak y H, Te aha iala T, Pussinen PJ, Sinisalo J, So sa T. The e ec o p olonged sys emic doxycycline he apy on se um issue deg ading p o einases in co ona y bypass pa ien s: a andomized, double‑masked, placebo‑con olled clinical ial. In lamm Res. 2014;63(5):329–34. 35. Payne JB, Golub LM, S one JA, Lee HM, Reinha d RA, So sa T, Slepian MJ. The e ec o suban imic obial‑dose‑doxycycline pe iodon al he apy on se um bioma ke s o sys emic in lamma ion: a andomized, double‑masked, placebo‑con olled clinical ial. J Am Den Assoc. 2011;142(3):262–73. 36. Hayakawa T, Yamashi a K, Tanzawa K, Uchijima E, Iwa a K. G ow h‑ p omo ing ac i i y o issue inhibi o o me allop o einases‑1 (TIMP‑1) o a wide ange o cells. A possible new g ow h ac o in se um. FEBS Le . 1992;298(1):29–32. 37. Moo e CS, C ocke SJ. An al e na e pe spec i e on he oles o TIMPs and MMPs in pa hology. Am J Pa hol. 2012;180(1):12–6. 38. S e le ‑S e enson WG. Tissue inhibi o s o me allop o einases in cell signaling: me allop o einase‑independen biological ac i i ies. Sci Signal. 2008;1(27): e6. 39. Kaaja R, Alenius H, Kinnunen T, Komulainen J, Pe änen N, Rönnemaa T, Sa amies J, Soukka H, Te amo K, Vuo ela P, Vää äsmäki M. Ges a ional diabe es (online). Cu en ca e guidelines. Wo king g oup se up by he Finnish Medical Socie y Duodecim, he Medical Ad iso y Boa d o he Finnish Diabe es Associa ion and he Finnish Gynecological Associa ion; 2013. h p://www.kaypahoi o. i. Accessed 25 June 2013. 40. Vilmi‑Ke älä T, Palomäki O, Vainio M, Uo ila J, Palomäki A. The isk o me a‑ bolic synd ome a e ges a ional diabe es melli us—a hospi al‑based coho s udy. Diabe ol Me ab Synd . 2015;7:43. 41. Repo o a Wo ld Heal h O ganiza ion Consul a ion. Diagnos ic c i e ia and classi ica ion o hype glycaemia i s de ec ed in p eg‑ nancy: a Wo ld Heal h O ganiza ion guideline. Diabe es Res Clin P ac . 2014;103(3):341–63. 42. Saquib N, S e anick ML, Na a ajan L, Pie ce JP. Mo ali y isk in o me smoke s wi h b eas cance : pack‑yea s s. smoking s a us. In J Cance . 2013;133(10):2493–7. 43. Wo ld Medical Associa ion Inc. Decla a ion o Helsinki. E hical p inciples o medical esea ch in ol ing human subjec s. J Indian Med Assoc. 2009;107(6):403–5. 44. Chenillo O, Henny J, S einme z J, He be h B, Wagne C, Sies G. High sensi i i y C‑ eac i e p o ein: biological a ia ions and e e ence limi s. Clin Chem Lab Med. 2000;38(10):1003–11.