Pandemrix® vaccination is not associated with increased risk of islet autoimmunity or type 1 diabetes in the TEDDY study children

ARTICLE Pandem ix® accina ion is no associa ed wi h inc eased isk o isle au oimmuni y o ype 1 diabe es in he TEDDY s udy child en Helena Elding La sson 1 &K is ian F. Lynch 2 &Ma ia Lönn o 3,4 &Michael J. Halle 5 & Åke Le nma k 1 &William A. Hagopian 6 &Jin-Xiong She 7 &Olli Simell 8 & Jo ma Toppa i 8,9 &Ane e-G. Ziegle 10,11 &Beena Akolka 12 &Je ey P. K ische 2 & Ma ian J. Rewe s 13 &Heikki Hyö y 3,14 & o he TEDDY S udy G oup Recei ed: 2 Ap il 2017 /Accep ed: 14 Augus 2017 /Published online: 9 Oc obe 2017 #The Au ho (s) 2017. This a icle is an open access publica ion Abs ac Aims/hypo hesis Du ing he A/H1N1 2009 (A/Cali o nia/04/ 2009) pandemic, mass accina ion wi h a squalene-con aining accine, Pandem ix®, was pe o med in Sweden and Finland. The accina ion was ound o cause na colepsy in child en and young adul s wi h he HLA-DQ 6.2 haplo ype. The aim o his s udy was o in es iga e i exposu e o Pandem ix® simila ly inc eased he isk o isle au oimmuni y o ype 1 diabe es. Me hods In The En i onmen al De e minan s o Diabe es in he Young (TEDDY) s udy, child en a e ollowed p ospec i ely o he de elopmen o isle au oimmuni y and ype 1 diabe es. In Oc obe 2009, when he mass accina ion began, 3401 child en a isk o isle au oimmuni y and ype 1 diabe es we e ollowed in Sweden and Finland. Vaccina ions we e eco ded and au oan ibodies agains insulin, GAD65 and insulinoma-associa ed p o ein 2 we e asce ained qua e ly be o e he age o 4 yea s and semi-annually he ea e . Resul s By 5 Augus 2010, 2413 o he 3401 (71%) child en obse ed as a isk o an isle au oan ibody o ype 1 diabe es on 1 Oc obe 2009 had been accina ed wi h Pandem ix®. By 31 July 2016, 232 child en had a leas one isle au oan ibody be o e 10 yea s o age, 148 had mul iple isle au oan ibodies and 96 had de eloped ype 1 diabe es. The isk o isle au oimmuni y was no inc eased among accina ed child en. The HR (95% CI) o he appea ance o a leas one isle au oan ibody was 0.75 (0.55, 1.03), a leas wo au oan i- bodies was 0.85 (0.57, 1.26) and ype 1 diabe es was 0.67 Membe s o he TEDDY S udy G oup a e lis ed in Acknowledgemen s. Elec onic supplemen a y ma e ial The online e sion o his a icle (h ps://doi.o g/10.1007/s00125-017-4448-3) con ains pee - e iewed bu unedi ed supplemen a y ma e ial, which is a ailable o au ho ised use s. *Helena Elding La sson helena.la [email protected] 1 Depa men o Clinical Sciences Malmö, Lund Uni e si y CRC, Skåne Uni e si y Hospi al SUS, Jan Waldens öms ga a 35; 60:11, 20502 Malmö, Sweden 2 Heal h In o ma ics Ins i u e, Mo sani College o Medicine, Uni e si y o Sou h Flo ida, Tampa, FL, USA 3 Facul y o Medicine and Li e Sciences, Uni e si y o Tampe e, Tampe e, Finland 4 Depa men o De ma ology, Tampe e Uni e si y Hospi al, Tampe e, Finland 5 Depa men o Pedia ics, Uni e si y o Flo ida, Gaines ille, FL, USA 6 Paci ic No hwes Diabe es Resea ch Ins i u e, Sea le, WA, USA 7 Cen e o Bio echnology and Genomic Medicine, Medical College o Geo gia, Augus a Uni e si y, Augus a, GA, USA 8 Depa men o Pedia ics, Tu ku Uni e si y Hospi al, Tu ku, Finland 9 Depa men o Physiology, Ins i u e o Biomedicine, Uni e si y o Tu ku, Tu ku, Finland 10 Ins i u e o Diabe es Resea ch, Helmhol z Zen um München, and Klinikum ech s de Isa , Technische Uni e si ä München, Munich, Ge many 11 Fo sche g uppe Diabe es e.V., Neuhe be g, Ge many 12 Na ional Ins i u e o Diabe es & Diges i e & Kidney Diseases, Be hesda, MD, USA 13 Ba ba a Da is Cen e o Childhood Diabe es, Uni e si y o Colo ado, Au o a, CO, USA 14 Fimlab Labo a o ies, Pi kanmaa Hospi al Dis ic , Tampe e, Finland Diabe ologia (2018) 61:193–202 DOI 10.1007/s00125-017-4448-3 (0.42, 1.07). In Finland, bu no in Sweden, accina ed child en had a lowe isk o isle au oimmuni y (0.47 [0.29, 0.75]), mul iple au oan ibodies (0.50 [0.28, 0.90]) and ype 1 diabe es (0.38 [0.20, 0.72]) compa ed wi h hose who did no ecei e Pandem ix®. The analyses we e adjus ed o con ounding ac o s. Conclusions/in e p e a ion Child enwi haninc eased gene ic isk o ype 1 diabe es who ecei ed he Pandem ix® accine du ing he A/H1N1 2009 pandemic had no inc eased isk o isle au oimmuni y, mul iple isle au oan ibodies o ype 1 diabe es. In Finland, he accine was associa ed wi h a educed isk o isle au oimmuni y and ype 1 diabe es. Keywo ds In luenza accine .Isle au oimmuni y . Pandem ix .Squalene .Swine lu .Type 1 diabe es . Vaccina ion Abb e ia ions GADA Glu amic acid deca boxylase au oan ibodies IA-2A Insulinoma-associa ed p o ein 2 au oan ibodies IAA Insulin au oan ibodies TEDDY The En i onmen al De e minan s o Diabe es in he Young In oduc ion Clinical diagnosis o ype 1 diabe es is p eceded by an au oimmune des uc i e p ocess agains he panc ea ic isle be a cells; a p od omal pe iod ha may las a ew mon hs o se e al yea s. The isk o de eloping au oimmuni y in gene ically suscep ible child en is widely conside ed o be inc eased by pe ina al o ea ly childhood en i onmen al exposu es [1]. While gene ic isk ac o s, such as HLA-DR- DQ [2] and non-HLA gene ic ac o s [3], ha e been implica ed, he en i onmen al condi ioning o igge exposu es ha e no ye been de ined. In The En i onmen al De e minan s o Diabe es in he Young (TEDDY) s udy, we ha e he oppo uni y o analyse igge s in ela ion o se ocon e sion o isle au oimmuni y as a i s p ima y endpoin , as o e 70% o child en wi h mul iple isle au oan ibodies de elop ype 1 diabe es o e a 10-yea pe iod [4]. I has been specula ed ha accina ions ea ly in li e may al e he immune esponse o in ec ions, leading o a dis u bed capaci y o dis inguish be ween sel and non-sel and he eby inc easing he isk o au oimmune eac ions. Howe e , p e ious s udies do no suppo he no ion ha ype 1 diabe es can be igge ed by accina ions [5,6]. Du ing he H1N1 in luenza pandemic in 2009–2010, mass accina ion o bo h child en and adul s ook place in many coun ies. In Sweden and Finland, Pandem ix®, a accine con aining he squalene-based adju an ASO3, was used while o he coun ies used o he ypes o accine. A ew mon hs a e he Pandem ix® accina ion p og amme, he incidence o new na colepsy diagnoses inc eased sha ply in bo h coun ies. Addi ional in es iga ions seeking an unde s anding o he po en ial mechanisms associa ed wi h Pandem ix® and na colepsy sugges ed ha he Pandem ix® accina ion migh ha e esul ed in he loss o o exin-p oducing neu ons, leading o he de elopmen o na colepsy in hese indi iduals [7–9]. The mechanism ha media es his e ec is no ully unde s ood, bu i seems ha bo h he composi e in luenza i us accine and he squalene adju an could con ibu e o he induc ion o o exin-speci ic au oimmuni y [10]. The aim o he p esen s udy was o in es iga e i exposu e o he Pandem ix® accine a ec ed he incidence o isle au oimmuni y and ype 1 diabe es in gene ically suscep ible child en. This was an obse a ional s udy ca ied ou among high- isk child en ollowed p ospec i ely om bi h [11]. Me hods Pa icipan s Pa icipan s we e included in he TEDDY s udy, a p ospec i e coho s udy unded by he Na ional Ins i u es o Heal h wi h he p ima y goal o iden i y en i onmen al causes o ype 1 diabe es [11]. The associa ion be ween he Pandem ix® accina ion and he isk o na colepsy was ini ially ound in Finland and la e in Sweden, ep esen ing wo o he TEDDY coun ies. The o he coun ies included in TEDDY a e he USA (wi h h ee cen es in Colo ado, Geo gia/Flo ida and Washing on) and one addi ional cen e in Eu ope (Ge many). Al hough he same p o ocol applied o all TEDDY cen es [11], he cu en analyses only include child en om Sweden and Finland as Pandem ix® was exclusi ely, and wi h a high co e age, adminis e ed in hose TEDDY coun ies. A all TEDDY si es, child en (n= 440,000) ep esen ing bo h he gene al popula ion and i s -deg ee ela i es o indi iduals wi h ype 1 diabe es we e sc eened a bi h du ing he pe iod 1 Sep embe 2004 o 1 Ma ch 2010 o gene ic ype 1 diabe es isk, de ined by HLA geno ype, as desc ibed p e iously [12,13]. A high- isk popula ion o 8676 child en was ec ui ed o ollow-up om 3 mon hs o age. Du ing he i s 4 yea s, all child en we e examined e e y hi d mon h. The ea e and cu en ly, child en s ill a isk o isle au oimmuni y a e being ollowed biannually un il he age o 15 yea s [11]. Child en wi h one o se e al isle au oan ibodies con inue moni o ing on a 3-mon h schedule a e 4 yea s o age. All pa icipan s and hei legal ep esen a i es ha e gi en in o med consen o pa icipa e in he s udy. The egional e hics commi ees in all pa icipa ing coun ies app o ed he s udy. 194 Diabe ologia (2018) 61:193–202 The mass accina ion pe iod in bo h Sweden and Finland was du ing he win e o 2009–2010 (1 Oc obe o 31 Ma ch). On 1 Oc obe 2009, 727/4358 (17%) child en we e no longe conside ed a isk o ype 1 diabe es as hey had ei he de eloped he disease (n= 30) o d opped ou o he s udy (n= 697) and we e he e o e excluded om he analyses. An addi ional 181 child en we e excluded o lea ing he TEDDY s udy empo a ily and e-joining a e he mass accina ion, 13 we e excluded o being HLA ineligible and 36 child en we e excluded as hey ei he ailed o ha e s udy ou comes mea- su ed (n= 6) o accina ion da a collec ed (n= 30) a e he mass accina ion. O he emaining 3401 child en s ill consid- e ed a isk o de eloping ype 1 diabe es on 1 Oc obe 2009, 3256 had no ye de eloped isle au oimmuni y and 3329 had no de eloped mul iple isle au oan ibodies (Fig. 1). Au oan ibodies and ype 1 diabe es The p ima y ou come in he TEDDY s udy is he de elopmen o pe sis en con i med au oan ibodies o glu amic acid deca boxylase (GADA), insulinoma-associa ed p o ein 2 (IA-2A) o insulin (IAA), measu ed a all s udy isi s and analysed by adiobinding assays [14,15]. All samples we e ini ially analysed a he e e ence labo a o y a he Uni e si y o B is ol, UK. Isle au oimmuni y was de ined as pe sis en (a leas wo consecu i e isi s) p esence o one o mo e o hese au oan i- bodies, con i med in he second e e ence labo a o y a he To al en olled in TEDDY n=8676 Family om Sweden o Finland n=4358 S ill en olled Oc 2009 n=3631 HLA eligible o s udy and ollowed o a leas one ou come n=3401 A isk o isle au oimmuni y on 1 Oc obe 2009 n=3256 A isk o mul iple isle au oan ibodies on 1 Oc obe 2009 n=3329 A isk o ype 1 diabe es on 1 Oc obe 2009 n=3401 Family om US o Ge many (n=4318) De eloped isle au oan ibodies be o e Oc obe 2009 n=73 De eloped mul iple isle au oan ibodies be o e Oc 2009 n=60 Nega i e o isle au oan ibodies and no longe being es ed (n=12) HLA ineligible (n=13) o no ollowed o ou comes (n=6) o o accina ion in o ma ion (n=30) Le s udy and ejoined (n=181) Ou o s udy: Diagnosed wi h ype 1 diabe es (n=30) o d opped ou o s udy (n=697) In TEDDY con inuously n=3450 Fig. 1 Flow cha o he s udy popula ion Diabe ologia (2018) 61:193–202 195 Ba ba a Da is Cen e o Childhood Diabe es (Uni e si y o Colo ado, Den e , CO, USA) and mul iple isle au oan i- bodies was de ined as mo e han one pe sis en con i med au oan ibody. Type 1 diabe es was diagnosed using he ADA c i e ia [16]. Vaccina ion da a A all TEDDY clinic isi s, he pa en s we e asked o epo i he child had been gi en any accina ion since he las isi . Type, dose and da e o accina ion was eco ded by a TEDDY nu se. I possible, he accina ion was e i ied by checking he child’s accina ion ca d. The mass accina ion wi h Pandem ix® began on 1 Oc obe 2009. Among he 3401 child en conside ed a isk o ype 1 diabe es on his da e, 2413 had a eco ded Pandem ix® accina ion by 5 Augus 2010. I mo e han one dose was ecei ed, he da e o i s accina ion was included in he analyses. S a is ical me hods Child en we e ollowed o ou comes om 1 Oc obe 2009. Fo child en bo n be ween 1 Oc obe 2009and 1 Ma ch 2010 he ollow-up was om bi h. Only 26 child en bo n a e he mass accina ion had begun we e accina ed o H1N1. Di e ences in cumula i e incidence o isle au oan ibodies, mul iple isle au oan ibodies and ype 1 diabe es be ween accina ion g oups we e examined in Kaplan–Meie analyses. Fo child en who we e accina ed, su i al ime was om da e o i s accina ion. Fo child en who we e no accina ed o H1N1, su i al ime was om 1 Oc obe 2009. Time-dependen Cox p opo ional haza d models we e used o explo e i accina ion o H1N1 modi- ied he isk o ype 1 diabe es ou comes. In he Cox models, he obse a ions we e le unca ed be o e 1 Oc obe 2009 and igh censo ed a e 31 July 2016 o a e he age o 10 yea s. All models we e adjus ed o coun y o esidence, he p esence o a i s -deg ee ela i e wi h ype 1 diabe es, HLA, sex and age o child on 1 Ma ch 2010. Final models we e also adjus ed o o he gene ic ype 1 diabe es isk ac o s (INS-23Hph1 [ s689], PTPN22 R620W [ s2476601], CTLA4 T17A [ s231775], and SNPs s2292239 in ERBB3, s3184504 in SH2B3, s10517086 and s12708716 in CLEC16A, s4948088 in COBL), ma e nal educa ion a 9 mon hs o age and p obio ic use be o e 90 days o age. Ma e nal educa ion was ca ego ised as p ima y educa ion (n= 764), some college/ ade school (n= 719), college deg ee (n= 1851) o missing educa ion (n= 67). In a compe ing isk analysis, we explo ed he ela ionship be ween Pandem ix® accina ion and he cause-speci ic isk o IAA o GADA as he i s soli a y appea ing isle au oan ibody [17]. In each model, child en who se ocon e ed o a compe ing isle au oan ibody ha was no o in e es we e censo ed a e he day o se ocon e - sion. The inal models we e also examined by age o child (< 2 and ≥2 yea s) as he incidence o IAA and GADA as he i s appea ing isle au oan ibody is known o di e conside ably be o e and a e 2 yea s o age [17]. The associa ion be ween Pandem ix® accine (yes, no) and ype 1 diabe es ou comes we e epo ed as HRs wi h 95% CIs. p alues < 0.05 we e conside ed s a is ically signi ican . To examine o inc eased isk o ype 1 diabe es ou comes wi hin subg oups, in e ac ions we e es ed be ween whe he o no he child had ecei ed he H1N1 accine and HLA, sex, coun y o esidence, age o amily his o y o ype 1 diabe es. The in e ac ions we e conside ed desc ip i e and seconda y and a p alue < 0.05 was conside ed an impo an in e ac ion o in es iga e u he . S a is ical analysis was pe o med using SAS e sion 90.4 (SAS Ins i u e, Ca y, NC, USA). Resul s O he 3401 child en conside ed a isk o ype 1 diabe es as o 1 Oc obe 2009, 2413 (70.9%) we e accina ed wi h Pandem ix®. O he child en accina ed, 98.8% (2385/2413) ecei ed hei i s accina ion be ween 1 Oc obe 2009 and 31 Ma ch 2010, i e had hei i s accina ion in Sep embe 2009 and 22 had hei accina ion be ween 1 Ap il and 5 Augus 2010. O he child en accina ed in Sweden, 72.9% (1010/1385) ecei ed a second Pandem ix® accina ion wi hin a median 2.0 (in e qua ile ange 1.5–3.0) mon hs; howe e , only 0.6% (6/1028) o child en om Finland ecei ed a second accina ion. The accina ion co e age by coun y and age o he child on 1 Ma ch 2010 (all child en bo n) is shown in Fig. 2.Inbo h Finland and Sweden, he co e age was low be o e he age o 6 mon hs (all p< 0.001) and child en we e mo e likely o ecei e he accine i he mo he had a college deg ee when he child was 9 mon hs o age (see elec onic supplemen a y ma e ial Table 1). Fig. 2 H1N1 co e age as o 1 Ma ch 2010 among child en bo n in Finland and Sweden; co e age shown o child en aged < 6 mon hs (solid black ba s), 6–11 mon hs (solid whi e ba s), 12–23 mon hs (solid ligh g ey ba s), 24–36 mon hs (solid da k g ey ba s) and ≥36 mon hs (ligh g ey s iped ba s) 196 Diabe ologia (2018) 61:193–202 A he ime o he cu en analyses (31 July 2016), 232 child en had de eloped pe sis en con i med isle au oan i- bodies by 10 yea s o age (135 in Sweden and 97 in Finland), 148 child en had de eloped mul iple isle au oan ibodies (81 in Sweden and 67 in Finland) and 96 had de eloped ype 1 dia- be es (47 in Sweden and 49 in Finland) (Table 1). The e was no inc eased isk o any isle au oan ibody (HR 0.76 [95% CI 0.57, 1.02]), mul iple isle au oan ibodies (HR 0.92 [95% CI 0.63, 1.35]) o ype 1 diabe es (HR 0.68 [95% CI 0.43, 1.06]) among he accina ed child en. In con as , he isk o any isle au oan ibody (HR 0.50 [95% CI 0.32, 0.78]), mul iple isle au oan ibodies (HR 0.56 [95% CI 0.33, 0.96]) and ype 1 diabe es (HR 0.41 [95% CI 0.23, 0.76]) was dec eased among accina ed child en in Finland, bu his was no seen in Sweden (HR 1.01 [95% CI 0.68, 1.50], HR 1.41 [95% CI 0.82, 2.46] and HR 1.10 [95% CI 0.53, 2.24], espec i ely) (Table 1). HLA geno ype, age a accina ion, sex o amily his o y o ype 1 diabe es did no modi y he associa ion be ween H1N1 accina ion and ou comes (Table 1). The cumula i e incidence o each o he ou comes in he wo coun ies a e accina ion o while conside ed a isk a e 1 Oc obe 2009 is shown in Fig. 3. In Finland, he dec eased isk o isle au oimmuni y and mul iple isle au o- an ibodies o he accina ed child en was seen p ima ily be ween 6 mon hs and 36 mon hs a e he i s accina ion. The cause-speci ic isk o IAA as a i s soli a y au oan ibody, be o e o he isle au oan ibodies appea , p ima ily occu ed in younge child en and ended o be mo e equen in Finland han in Sweden [17]. The e o e, we examined whe he he incidence o IAA o GADA as i s appea ing soli a y au o- an ibody sepa a ely, o any isle au oan ibody, in Finland and Sweden, in bo h younge and olde child en sho ly a e accina ion we e in luenced by he accina ion a e adjus ing o p e iously epo ed isk ac o s. Gene ic isk ac o s p e iously ound o be associa ed wi h isle au oimmuni y, Table 1 P opo ional haza ds models o HIN1 accina ion (yes s no) on isk o isle au oan ibodies, mul iple isle au oan ibodies and ype 1 diabe es be o e age 10 yea s adjus ing o ac o s in able and also s a i ied by he ac o subg oups Fac o and g oup To al (N a ) H1N1 accina ion in ela ion o isk o isle au oan ibodies H1N1 accina ion in ela ion o isk o mul iple isle au oan ibodies H1N1 accina ion in ela ion o isk o ype 1 diabe es E en s (n) HR (95% CI) p alue b E en s (n) HR (95% CI) p alue b E en s (n) HR (95% CI) p alue b All child en 3401 232 0.76 (0.57, 1.02) 148 0.92 (0.63, 1.35) 96 0.68 (0.43, 1.06) Coun y Finland 1438 97 0.50 (0.32, 0.78) 67 0.56 (0.33, 0.96) 49 0.41 (0.23, 0.76) Sweden 1963 135 1.01 (0.68, 1.50) 0.008 81 1.41 (0.82, 2.46) 0.01 47 1.10 (0.53, 2.24) 0.02 Sex Female 1659 103 0.91 (0.58, 1.43) 63 0.80 (0.45, 1.43) 46 0.60 (0.31, 1.16) Male 1742 129 0.67 (0.46, 0.99) 0.46 85 1.03 (0.62, 1.72) 0.43 50 0.77 (0.41, 1.45) 0.35 Family his o y Gene al popula ion 3123 199 0.69 (0.50, 0.94) 122 0.83 (0.55, 1.26) 78 0.64 (0.39, 1.06) Fi s -deg ee ela i e 278 33 1.30 (0.54, 3.13) 0.12 26 1.66 (0.59, 4.68) 0.15 18 0.71 (0.24, 2.07) 0.65 HLA-DR DR-4/4 665 41 0.85 (0.41, 1.78) 27 1.40 (0.52, 3.74) 20 0.55 (0.20, 1.49) DR-4/8 695 47 0.61 (0.32, 1.17) 25 0.62 (0.25, 1.56) 13 0.13 (0.04, 0.49) DR-3/4 1298 111 0.72 (0.47, 1.10) 80 0.76 (0.46, 1.26) 53 0.74 (0.40, 1.38) DR-3/3 649 28 0.77 (0.32, 1.87) 0.68 12 1.73 (0.41, 7.30) 0.32 7 7.74 (0.80, 74.7) 0.14 Age on 1 Ma ch 2010 < 1 yea 752 71 0.76 (0.46, 1.25) 44 0.75 (0.40, 1.39) 26 0.26 (0.10, 0.70) 1 yea 646 50 1.13 (0.55, 2.33) 33 1.52 (0.58, 4.01) 16 2.00 (0.45, 8.97) 2 yea 649 43 0.75 (0.36, 1.58) 25 0.53 (0.20, 1.39) 23 0.70 (0.26, 1.91) 3 yea 587 34 0.76 (0.34, 1.70) 23 1.07 (0.36, 3.17) 17 0.61 (0.21, 1.83) ≥4 yea s 722 34 0.57 (0.26, 1.24) 0.73 23 2.48 (0.57, 10.9) 0.07 14 1.24 (0.27, 5.71) 0.19 a To al numbe o child en on 1 Oc obe 2009 who we e obse ed a isk o ype 1 diabe es; 145 we e no longe obse ed a isk o isle au oimmuni y, and 72 no longe a isk o mul iple isle au oan ibodies b p alue es o mul iplica i e in e ac ion be ween subg oups; all HRs a e adjus ed o ac o s in able Diabe ologia (2018) 61:193–202 197 including INS-23Hph1 ( s689), PTPN22 R620W ( s2476601), CTLA4, T17A ( s231775), SNPs s2292239 in ERBB3, s3184504 in SH2B3, s10517086 and s12708716 in CLEC16A, and s4948088 in COBL [18] we e included in his analysis, as well as p obio ic use be o e 90 days o age [19], ma e nal age a bi h o child and ma e nal educa ion (Table 2). A e adjus men , he HR o isle au oan ibodies (0.47 [95% CI 0.29, 0.75]), o mul iple isle au oan ibodies (0.50 [95% CI 0.28, 0.90]) and o ype 1 diabe es (0.38 [95% CI 0.20, 0.72]) emained signi ican ly dec eased in Finland. Mo eo e , cause-speci ic isk o IAA as he i s appea ing isle au oan ibody was dec eased, pa icula ly among child en accina ed a younge han 2 yea s o age (Table 2). In Sweden, 1.2% o child en obse ed a isk o ype 1 diabe es we e accina ed wi h he seasonal lu accine be o e (0.2%) o only a e (1.0%) he mass accina ion o H1N1 had begun. The equency o ecei ing he seasonal lu accine was simila be ween child en who we e accina ed o H1N1 (1.3%) compa ed wi h child en who we e no (1.0%). In Finland, 58.7% o child en who we e accina ed o H1N1 had a seasonal lu accine ei he be o e (36.1%) o only a e (22.6%) he H1N1 accina ion. This was signi ican ly highe compa ed wi h he child en who we e no accina ed o H1N1 (28.3% ecei ed seasonal lu accine,11.7% be o e [p< 0.001] and 16.6% only a e [p≤0.01]). Mo he s could also ha e ecei ed he seasonal lu accine du ing p egnancy; howe e , he pe cen age was Fig. 3 Kaplan–Meie cu es showing he cumula i e pe cen age o accina ed (solid line) and un accina ed (dashed line) child en de eloping isle au oan ibodies (IA) in (a) Finland wi h an a e age ollow-up ime o 5.3 yea s o accina ed (n=976) and 4.7 yea s o un accina ed (n= 397) child en and (b) Sweden wi h an a e age ollow- up ime o 5.4 yea s o accina ed (n= 1309) and 5.0 yea s o un accina ed (n= 563) child en; mul iple isle au oan ibodies in (c) Finland wi h an a e age ollow-up ime o 5.3 yea s o accina ed (n= 1000) and 4.8 yea s o un accina ed (n=400)child en and (d) Sweden wi h an a e age ollow-up ime o 5.5 yea s o accina ed (n= 1347) and 5.2 yea s o un accina ed (n= 569) child en. Type 1 diabe es (T1D) in (e) Finland wi h an a e age ollow-up ime o 5.5 yea s o accina ed (n= 1027) and 5.1 yea s o un accina ed (n=410)child en and ( ) Sweden wi h an a e age ollow-up ime o 5.7 yea s o accina ed (n= 1384) and 5.4 yea s o un accina ed (n= 578) child en, a e he child was accina ed o a e 1 Oc obe 2009 o he child en who we e no accina ed 198 Diabe ologia (2018) 61:193–202 simila be ween coun ies (Finland, 2.1%; Sweden 1.4%). In e es ingly, o he 1.7% (58/3381) o mo he s who did ge he seasonal lu accine du ing p egnancy, 6.2% (46/743) we e mo he s o younge child en (aged < 1 yea on 1 Ma ch 2010) and he accine was gi en owa ds he end o he mass accina ion. O hese mo he s, only 2/46 had hei child accina ed o H1N1. Seasonal lu accine adminis e ed o mo he s du ing p egnancy o child en du ing ollow-up did no explain he associa ion be ween H1N1 accina ion and ype 1 diabe es ou comes. The seasonal lu accine did no ha e a modi ying e ec on he associa ion be ween H1N1 accina ion and ype 1 diabe es ou comes (da a no shown). Fu he mo e, in Sweden, a second accina ion dose did no in luence any o he ype 1 diabe es ou comes (da a no shown). When analysing hese esul s o e ime, we did no ind any sho - o long- e m ini ia ing o accele a ing e ec s o he Pandem ix® accina ion on isk o isle au oimmuni y, mul iple isle au oan ibodies o ype 1 diabe es. Discussion In his s udy, we in es iga ed whe he isk o isle au oimmuni y and ype 1 diabe es is inc eased in child en who ha e been accina ed wi h he squalene (ASO3)- con aining H1N1 lu accine (Pandem ix®). No ably, Pandem ix® has been associa ed wi h au oimmuni y o o exin-p oducing cells and he de elopmen o na colepsy. As such, we hypo hesised ha his accine may no only p omo e au oimmuni y o o exin-p oducing cells bu also isle au oan igens. The esul s o ou s udy do no suppo his hypo hesis and a gue agains any connec ion be ween he H1N1 mass accina ion campaign and he isk o ype 1 diabe es. In ac , we ound ha he incidence o ype 1 diabe es ou comes in Finland was ac ually lowe in child en accina ed agains H1N1 han in non- accina ed child en. Theo e ically, his inding could indica e ha he accine may ha e educed he isk o ype 1 diabe es in highe isk popula ions by unknown mechanisms. In he p ospec i e coho o child en ollowed in he TEDDY s udy, we ha e he unique capaci y o analyse en i onmen al ac o s associa ed wi h he ini ia ion o isle au oimmuni y and de elopmen o ype 1 diabe es. Type 1 diabe es is p eceded by p og essi e au oimmune des uc ion o be a cells, which can las jus a ew mon hs o up o many yea s. Since isle au oimmuni y and mul iple isle au oan i- bodies a e known o con e o e 70% isk o he de elopmen o ype 1 diabe es wi hin 10 yea s [4], i was impo an o conside hese ma ke s as su oga e endpoin s when analysing a hypo he ical inc eased isk o he disease du ing he i s ew yea s a e he accina ion. By examining o isk o se ocon e sion o wo o mo e isle au oan ibodies as an endpoin along wi h ype 1 diabe es, we ha e enabled an ea lie de ec ion o an inc eased o dec eased isk o clinical ype 1 diabe es. Among a wide ange o da a cu en ly being collec ed in TEDDY, accina ion eco ds a e p ospec i ely documen ed in he s udy. Pandem ix® was used exclusi ely as he H1N1 accine in Sweden and Finland. O he ypes o accine we e Table 2 Associa ion be ween H1N1 accina ion and ype 1 di- abe es ou comes be o e age 10 yea s adjus ing o sex, amily his o y o ype 1 diabe es, p obio ics be o e 90 days, ma e - nal educa ion, ma e nal age, HLA- DR-DQ high- isk geno ypes, INS- 23Hph1 ( s689), PTPN22 R620W ( s2476601), CTLA4 T17A ( s231775), SNPs s2292239 in ERBB3, s3184504 in SH2B3, s10517086 and s12708716 in CLEC16A,and s4948088 in COBL Va iable Mul iple p opo ional haza d model o H1N1 on ype 1 diabe es ou comes Fi s appea ing isle au oan ibodies Mul iple isle au oan ibodies Type 1 diabe es IAA GADA Any HR p alue HR p alue HR p alue HR p alue HR p alue Adjus ed a 0.64 0.10 0.78 0.29 0.75 0.07 0.85 0.41 0.67 0.09 Finland Adjus ed a 0.38 0.01 0.55 0.12 0.47 0.002 0.50 0.02 0.38 0.003 < 2 yea s b 0.30 0.009 1.18 0.79 0.51 0.03 0.45 0.04 0.25 0.002 ≥2yea s b 0.52 0.35 0.51 0.14 0.51 0.06 0.58 0.21 0.62 0.34 Sweden Adjus ed a 0.88 0.70 0.98 0.96 1.03 0.89 1.33 0.34 1.19 0.67 < 2 yea s b 0.79 0.57 1.08 0.87 0.96 0.88 1.31 0.46 1.81 0.33 ≥2yea s b 0.32 0.15 1.12 0.81 0.97 0.94 1.33 0.57 0.82 0.70 a Va iable included in mul i a ia e p opo ional haza ds model; all a iables a ailable on 2966 child en a isk o isle au oan ibodies, o which 214 de eloped isle au oan ibodies; 3037 obse ed a isk o mul iple isle au o- an ibodies, o which 140 de eloped mul iple isle au oan ibodies; and 3103 obse ed a isk o ype 1 diabe es, o which 91 de eloped ype 1 diabe es b Age o child on 1 Ma ch 2010 Diabe ologia (2018) 61:193–202 199 used in he USA as pa o he H1N1 accina ion p og amme. In Ge many, Pandem ix® was used as pa o he accina ion p og amme, bu i had low co e age. Due o he e ogenei y and low numbe s, we did no include he Ge man TEDDY popula ion in his analysis. The e o e, hese analyses we e limi ed o Swedish and Finnish child en. A weakness in ou s udy is ha ou popula ion was selec ed o HLA geno ypes associa ed wi h isk o ype 1 diabe es [12]. One canno exclude he possibili y ha he accine may inc ease he isk o isle au oimmuni y in child en wi h lowe isk geno ypes, who we e sc eened bu no en olled o ollow-up in TEDDY. This may be u he emphasised by he obse a ion ha all child en who de eloped na colepsy a e he Pandem ix® accina ion had HLA geno ypes con aining HLA- DQB1*06:02 [10]. Because he DQ-0602 geno ype is ac ually p o ec i e agains he de elopmen o ype 1 diabe es, child en wi h his geno ype we e excluded om he TEDDY s udy. As HLA plays such a c i ical ole in he p esen a ion o an igens o he immune sys em and how he immune sys em eac s immunologically o in ec ions, indi iduals wi h di e en HLA geno ypes a e known o eac di e en ially o accina ions [20]. A e he epo s o he inc eased incidence o na colepsy, s udies we e es ablished o examine o a possible in luence o he Pandem ix® accina ion on he incidence o o he au oimmune diseases. In one o hese s udies in es iga ing he incidence o ype 1 diabe es, along wi h o he immunological diseases, he au ho s ound a non-signi ican inc eased incidence o ype 1 diabe es [21]. This inding was u he discussed in a numbe o le e s o he jou nal, indica ing ha missing da a and he exclusion o a ious pa icipan s would ha e esul ed in he lack o signi icance [22–24]. In ou cu en s udy, we do no ind any indica ion o inc eased incidence o isle au oimmuni y o ype 1 diabe es a e accina ion wi h Pandem ix®. I is also impo an o emphasise ha ou p ospec i e coho s udy did no su e om he same kind o missing da a issues ha a ec ed he esul s o he p e ious s udy. Simila o ou s udy, in 355 Swedish child en who we e diagnosed wi h ype 1 diabe es du ing he accina ion campaign, younge indi iduals (< 3 yea s) wi h HLA DQ2/2 o 2/X we e low esponde s o Pandem ix®, measu ed as an ibodies agains A/H1N1 haemagglu inin [25]. As he p opo ion o child en < 3 yea s o age wi h he high- isk HLA DQ2/8 dec eased a e he accina ion, i was sugges ed ha he accine a ec ed clinical onse o ype 1 diabe es by delaying onse in child en wi h his geno ype [25]. Howe e , hese s udies we e pe o med sho ly a e he accina ion and included only clinical ype 1 diabe es and no isle au oimmuni y as an endpoin . The e o e, an inc ease in ype 1 diabe es could ha e been missed, since i would no be ob ious un il long a e an inc ease in isle au oimmuni y. Ou cu en s udy is he i s o in es iga e isle au oimmuni y and mul iple isle au oan ibodies in ela ion o he accina ion. Ano he possibili y would be o in es iga e i Pandem ix® a ec ed p og ession om isle au oan i- bodies o clinical ype 1 diabe es. Since au oan ibodies appea ea ly, while p og ession o ype 1 diabe es may ake many yea s, mo e ime is needed o in es iga e p og ession o ype 1 diabe es p ope ly in ou p ospec i e s udy as he peak incidence has no ye been eached. The TEDDY child en will be ollowed o age 15 yea s, and addi ional analyses will be done la e on in he s udy. In e es ingly, when sepa a ing Swedish and Finnish da a, we ound ha Finnish child en had a dec eased isk o isle au oimmuni y (p ima ily IAA as i s au oan ibody), mul iple isle au oan ibodies and ype 1 diabe es a e accina ion. Incidence o ype 1 diabe es ou comes is known o be highe in Finland compa ed wi h Sweden [26]; howe e , i is unclea as o why he e was no inc eased isk among he accina ed popula ion. We no ed ha he majo i y (58.7%) o he H1N1- accina ed Finnish child en and only 27.6% o hose no H1N1- accina ed had ecei ed a p io seasonal lu accina ion, while only 1.2% o Swedish child en had ecei ed p e ious seasonal lu accina ion wi h no di e ence be ween H1N1 accina ion a es. One could specula e ha he immunological memo y induced by p e ious lu accina ions, ega dless o he adju an in hose accines, in Finnish child en led o imp o ed immune esponses o H1N1 i us (a p iming e ec ). A e adjus ing o seasonal lu accina ion, du ing and a e p egnancy, he associa ions wi h ype 1 diabe es ou comes in Finland emained. Ne e heless, i is possible ha in luenza i us in ec ions may inc ease he isk o ype 1 diabe es, as epo ed by p e ious s udies [27,28], while o he s could no con i m his [29]. I is in e es ing o no e ha epea ed doses o Pandem ix® we e mo e equen in Sweden han in Finland. One could specula e ha hese boos e accina ions may ha e led o be e p o ec ion agains H1N1 in ec ions in Sweden. This, in u n, could ha e educed he ci cula ion o he i us and also pa ially p o ec ed non- accina ed child en in Sweden. While p elimina y, hese indings gene a e addi ional ques ions ha will be u he examined in TEDDY, in which analyses o i ome da a will be pe o med in he nea u u e, and wa an s s udy in la ge coho s o e alua e possible p o ec i e associa ions be ween in luenza in ec ions and accina ions and isle au oimmuni y. In conclusion, ou analyses did no ind any inc eased isk o isle au oimmuni y, mul iple isle au oan ibodies o ype 1 diabe es in child en who we e gi en he ASO3-con aining Pandem ix® lu accina ion du ing he H1N1 pandemic in 2009–2010. Addi ional s udies a e needed o u he explo e he po en ial p o ec i e e ec s o in luenza accina ions. 200 Diabe ologia (2018) 61:193–202 Acknowledgemen s The TEDDY S udy G oup (see appendix). Da a a ailabili y The da a ha suppo he indings o his s udy a e a ailable om he Na ional Ins i u e o Diabe es and Diges i e and Kidney Diseases (NIDDK) Cen al Reposi o y upon eques . Funding Funded by U01 DK63829, U01 DK63861, U01 DK63821, U01 DK63865, U01 DK63863, U01 DK63836, U01 DK63790, UC4 DK63829, UC4 DK63861, UC4 DK63821, UC4 DK63865, UC4 DK63863, UC4 DK63836, UC4 DK95300, UC4 DK100238, UC4 DK106955 and Con ac No. HHSN267200700014C om he NIDDK, Na ional Ins i u e o Alle gy and In ec ious Diseases (NIAID), Na ional Ins i u e o Child Heal h and Human De elopmen (NICHD), Na ional Ins i u e o En i onmen al Heal h Sciences (NIEHS), JDRF, and Cen e s o Disease Con ol and P e en ion (CDC). This wo k was suppo ed in pa by he NIH/NCATS Clinical and T ansla ional Science Awa ds o he Uni e si y o Flo ida (UL1 TR000064) and he Uni e si y o Colo ado (UL1 TR001082). Duali y o in e es s The au ho s decla e ha he e is no duali y o in e es associa ed wi h his manusc ip . Con ibu ion s a emen HEL designed he s udy, acqui ed and in e p e ed da a and d a ed he a icle. KL in e p e ed and analysed da a and e ised he a icle. HH designed he s udy, in e p e ed da a and e ised he a icle. ML and MH con ibu ed o concep ion and design and c i ically e ised he a icle. MR, ÅL, WH, J-XS, OS, JT, AZ, BA and JK designed he TEDDY s udy and c i ically e ised he manusc ip . All au ho s app o ed he inal e sion o he a icle. HEL, KL and HH a e gua an o s o his wo k. Appendix: The TEDDY S udy G oup The a ilia ions o s udy g oup membe s a e indica ed using supe sc ip symbols and s a ed a he end o each sec ion. I no symbols a e gi en, he membe is a ilia ed wi h he main a ilia ion gi en a he end o he pa ag aph. Commi ee in ol emen is indica ed ia supe sc ip numbe s. Commi ees 1 Ancilla y S udies, 2 Die , 3 Gene ics, 4 Human Subjec s/ Publici y/Publica ions, 5 Immune Ma ke s, 6 In ec ious Agen s, 7 Labo a o y Implemen a ion, 8 Ma e nal S udies, 9 Psychosocial, 10 Quali y Assu ance, 11 S ee ing, 12 S udy Coo dina o s, 13 Celiac Disease, 14 Clinical Implemen a ion, 15 Quali y Assu ance Subcommi ee on Da a Quali y. Colo ado Clinical Cen e Ma ian Rewe s 1,4–6,10,11 ,Kimbe ly Bau is a 12 , Judi h Bax e 9,10,12,15 , Ru h Bedoy 2 , Daniel Felipe-Mo ales, Kimbe ly D iscoll 9 , B igi e I. F ohne 2,14 , Pa icia Gesualdo 2,6,12,14,15 , Michelle Ho man 12–14 , Rachel Ka ban 12 , Edwin Liu 13 , Jill No is 2,3,12 , Adela Sampe -Imaz, And ea S eck 3,14 , Ka hleen Waugh 6,7,12,15 , Hali W igh 12 . Uni e si y o Colo ado, Anschu z Medical Campus, Ba ba a Da is Cen e o Childhood Diabe es. Finland Clinical Cen e Jo ma Toppa i ¥^1,4,11,14 ,OlliG. Simell ¥^1,4,11,13 , Annika Adamsson ^12 , Su i Ahonen* ±§ , Heikki Hyö y* ±6 , Jo ma Ilonen ¥¶3 , Sanna Jokipuu ^ , Tiina Kallio ^ ,Leena Ka lsson ^ , Miia Kähönen μ¤ ,MikaelKnip* ±5 , Lea Ko anen* ±§ , Mi a Ko easalo* ±§2 , Kalle Ku ppa* ±13 , Tiina La a-aho μ¤ , Ma ia Lönn o * ±6 , Elina Män ymäki ^ , Ka ja Mul asuo μ¤ , Juha Mykkänen ¥3 ,Tiina Niininen ± * 12 ,Sa iNiinis ö ±§ ,MiaNyblom* ± , Pe a Rajala ^ , Jenna Rau anen ±§ , Anne Riikonen* ±§ , Mika Riikonen ^ , Jenni Rouhiainen ^ , Minna Romo ^ , Tuula Simell, Ville Simell ^¥13 , Maija Sjöbe g ¥^12,14 , Aino S enius μ¤12 , Ma ia Leppänen ^ , Sini Vainionpää ^ ,Ee a Va jonen ¥^12 , Rii a Veijola μ¤14 , Su i M. Vi anen* ±§2 , Ma i Vähä- Mäkilä ^ , Ma i Åke lund* ±§ , Ka i Lind o s* 13 . ¥ Uni e si y o Tu ku; ^ Tu ku Uni e si y Hospi al, Hospi al Dis ic o Sou hwes Finland; *Uni e si y o Tampe e; ± Tampe e Uni e si y Hospi al; §Na ional Ins i u e o Heal h and Wel a e, Finland; ¶ Uni e si y o Kuopio; μ Uni e si y o Oulu; ¤ Oulu Uni e si y Hospi al. Geo gia/Flo ida Clinical Cen e Jin-Xiong She 1,3,4,11 , Desmond Scha z* 4,5,7,8 , Diane Hopkins 12 , Leigh S eed 12–15 , Jamie Thomas* 6,12 , Janey Adams* 12 , Ka he ine Sil is 2 , Michael Halle * 14 , Melissa Ga dine , Richa d McIndoe, Ashok Sha ma, Joshua Williams, Gab iela Young, S ephen W. Ande son ^ , Lau a Jacobsen *14 . Cen e o Bio echnology and Genomic Medicine, Augus a Uni e si y; *Uni e si y o Flo ida; ^ Pedia ic Endoc ine Associa es, A lan a. Ge many Clinical Cen e Ane e-G. Ziegle 1,3,4,11 , And eas Beye lein 2 , Ezio Boni acio* 5 ,MichaelHummel 13 , Sand a Hummel 2 , K is ina Fo e ek ¥2 , Nicole Janz, Ma hilde Ke s ing ¥2 , Anne e Knop 7 , Sibylle Kole zko ¶13 , Claudia Peplow 12 , Roswi h Ro h 9 , Ma lon Scholz, Joanna S ock 9,12,14 , Ka ha ina Wa ncke 14 , Lo ena Wendel, Ch is iane Winkle 2,12,15 . Fo sche g uppe Diabe es e.V. and Ins i u e o Diabe es Resea ch, Helmhol z Zen um München; Klinikum ech s de Isa , Technische Uni e si ä München. *Cen e o Regene a i e The apies, TU D esden; ¥ Resea ch Ins i u e o Child Nu i ion, Do mund; ¶ D on Haune Child en’s Hospi al, Depa men o Gas oen e ology, Ludwig Maximillians Uni e si y Munich. Sweden Clinical Cen e Åke Le nma k 1,3–6,8,10,11,15 ,DanielAga dh 13 , Ca in And én A onsson 2,12,13 , Ma ia Ask, Jenny B eme , Ulla-Ma ie Ca lsson, Co ado Cilio 5 , Emelie E icson-Halls öm, Lina F ansson, Thomas Ga d, Joanna Ge a dsson, Rasmus Benne , Monica Hansen, Ge ie Hansson, Susanne Hybe g, F ed ik Johansen, Be glind Jonsdo i , Helena Elding La sson 6,14 , Ma ielle Linds öm, Ma kus Lundg en 14 ,Ma ia Månsson-Ma inez, Ma ia Ma kan, Jessica Melin 12 , Zeliha Mes an, Ka in O osson, Kob a Rahma i, Ani a Ramelius, Falas in Salami, Sa a Sib ho pe, Bi gi a Sjöbe g, Ul ica Swa ling 9,12 , E elyn Tekum Amboh, Ca ina Tö n 3,15 , Anne Wallin, Åsa Wima 12,14 ,So ieÅbe g. Lund Uni e si y. Washing on Clinical Cen e William A. Hagopian 1,3–7,11,13,14 , Michael Killian 6,7,12,13 , Clai e Cowen C ouch 12,14,15 , Jenni e Skidmo e 2 , Josephine Ca son, Ma ia Dalzell, Kayleen Dunson, Rachel He ey, Co bin Johnson, Rachel Lyons, A lene Meye , Denise Mulenga, Alexande Ta , Mo gan Uland, John Willis. Paci ic No hwes Diabe es Resea ch Ins i u e. Pennsyl ania Sa elli e Cen e Do o hy Becke , Ma ga e F anciscus, Ma yEllen Dalmag o-Elias Smi h 2 , Ashi Da a y, Ma y Be h Klein, Ch ys al Ya es. Child en’s Hospi al o Pi sbu gh o UPMC. Da a Coo dina ing Cen e Je ey P. K ische 1,4,5,10,11 , Michael Abbondondolo, Sa ah Aus in-Gonzalez, Ma you i A endano, Sand a Bae hke, Rasheedah B own 12,15 , B an Bu kha d 5,6 ,Ma ha Bu e wo h 2 , Joanna Clasen, Da id Cu hbe son, Ch is ophe Ebe ha d, S e en Fiske 9 , Dena Ga cia, Jenni e Ga meson, Veena Gowda, Ka hleen Heyman, F ancisco Pe ez La as, Hye-Seung Lee 1,2,13,15 , Shu Liu, Xiang Liu 2,3,9,14 , K is ian Lynch 5,6,9,15 , Jamie Malloy, C is ina McCa hy 12,15 , S e en Meulemans, Hemang Pa ikh 3 , Ch is Sha e , Lau a Smi h 9,12 , Susan Smi h 12,15 , Noah Sulman, Roy Tamu a 1,2,13 , Ulla Uusi alo 2,15 , Kend a Vehik 4–6,14,15 , Ponni Vijayakandipan, Kei h Wood, Jimin Yang 2,15 . Pas s a : Lo i Balla d, Da id Hadley, Wendy McLeod. Uni e si y o Sou h Flo ida. Au oan ibody Re e ence Labo a o ies Liping Yu ^5 , Dongmei Miao ^ , Polly Bingley* 5 , Alis ai Williams*, Kyla Chandle *, Saba Rokni*, Clai e Williams*, Rebecca Wya *, Gi y Geo ge*, Sian G ace*. ^ Ba ba a Da is Cen e o Childhood Diabe es, Uni e si y o Colo ado Den e ; *School o Clinical Sciences, Uni e si y o B is ol, UK. Diabe ologia (2018) 61:193–202 201