Quantity and diversity of environmental microbial exposure and development of asthma : a birth cohort study

ORIGINAL ARTICLE EPIDEMIOLOGY AND GENETICS Quan i y and di e si y o en i onmen al mic obial exposu e and de elopmen o as hma: a bi h coho s udy A. M. Ka onen 1 , A. Hy € a inen 1 , H. Rin ala 1 , M. Ko ppi 2 ,M.T € aubel 1 , G. Doekes 3 , U. Geh ing 3 , H. Renz 4 , P. I. P e e le 4 , J. Genunei 5 , L. Keski-Nisula 1,6 , S. Remes 7 , J. Lampi 1,8 , E. on Mu ius 9 & J. Pekkanen 1,8 1 Depa men o En i onmen al Heal h, Na ional Ins i u e o Heal h and Wel a e, Kuopio; 2 Pedia ic Resea ch Cen e , Uni e si y o Tampe e and Uni e si y Hospi al, Tampe e, Finland; 3 Ins i u e o Risk Assessmen Sciences, U ech Uni e si y, U ech , he Ne he lands; 4 Ins i u e o Labo a o y Medicine and Pa hobiochemis y, Molecula Diagnos ics, Philipps-Uni e si y Ma bu g, Ma bu g; 5 Ins i u e o Epidemiology and Medical Biome y, Ulm Uni e si y, Ulm, Ge many; 6 Depa men s o Obs e ics and Gynecology, Kuopio Uni e si y Hospi al; 7 Depa men o Pedia ics, Kuopio Uni e si y Hospi al; 8 Uni o Public Heal h and Clinical Nu i ion, Uni e si y o Eas e n Finland, Kuopio, Finland; 9 Child en’s Hospi al, Uni e si y o Munich, Munich, Ge many To ci e his a icle: Ka onen AM, Hy € a inen A, Rin ala H, Ko ppi M, T€ aubel M, Doekes G, Geh ing U, Renz H, P e e le PI, Genunei J, Keski-Nisula L, Remes S, Lampi J, on Mu ius E, Pekkanen J. Quan i y and di e si y o en i onmen al mic obial exposu e and de elopmen o as hma: a bi h coho s udy. Alle gy 2014; 69: 1092–1101. Keywo ds as hma; a opy; child en; mic obial exposu e; espi a o y ac disease. Co espondence Anne M. Ka onen, Depa men o En i onmen al Heal h, Na ional Ins i u e o Heal h and Wel a e, P.O. Box 95, FIN-70701 Kuopio, Finland. Tel.: +358 (0) 29 524 6325 Fax: +358 (0) 29 524 6498 E-mail: [email p o ec ed] Accep ed o publica ion 27 Ap il 2014 DOI:10.1111/all.12439 Edi ed by: Douglas Robinson Abs ac Backg ound: Ea ly-li e exposu e o en i onmen al mic obial agen s may be asso- cia ed wi h he de elopmen o alle gies. The aim o he s udy was o iden i y be - e ways o cha ac e ize mic obial exposu e as a p edic o o espi a o y symp oms and alle gies. Me hods: A bi h coho o 410 child en was ollowed up un il 6 yea s o age. Bac e ial endo oxin, 3-hyd oxy a y acids, N-ace yl-mu amic acid, ungal ex a- cellula polysaccha ides (EPS) om Penicillium and Aspe gillus spp., b-D-glucan, e gos e ol, and bac e ial o ungal quan i a i e polyme ase chain eac ions (qPCRs) we e analyzed om dus samples collec ed a 2 mon hs o age. As hma, wheezing, cough, and a opic de ma i is we e assessed using epea ed ques ion- nai es. Speci ic IgEs we e de e mined a he age o 1 and 6 yea s. Resul s: Only ew associa ions we e ound be ween single mic obial ma ke s and he s udied ou comes. In con as , a sco e o he o al quan i y o mic obial expo- su e, ha is, sum o indica o s o ungi (e gos e ol), G am-posi i e (mu amic acid) bac e ia, and G am-nega i e (endo oxin) bac e ia, was signi ican ly (in e ed-U shape) associa ed wi h as hma incidence (P<0.001): he highes isk was ound a medium le els (adjus ed odds a io (aOR) 2.24, 95% con idence in e al (95% CI) 0.87–5.75 o 3 d quin ile) and he lowes isk a he highes le el (aOR 0.34, 95% CI 0.09–1.36 o 5 h quin ile). The mic obial di e si y sco e, ha is, sum o de ec ed qPCRs, was in e sely associa ed wi h isk o wheezing and was signi ican ly (in e ed-U shape) associa ed wi h sensi iza ion o inhalan alle gens. Conclusion: Sco e o quan i y o mic obial exposu e p edic ed as hma be e han single mic obial ma ke s independen ly o mic obial di e si y and amoun o dus . Be e indica o s o o al quan i y and di e si y o mic obial exposu e a e needed in s udies on he de elopmen o as hma. Mic obial exposu e may con ibu e o he de elopmen o alle gic diseases, bu he ole o di e en agen s and mecha- nisms a e s ill unknown (1). Mos epidemiological s udies wi h mic obial exposu e assessmen among child en ha e been es ic ed o s udies on endo oxin (2) o ha e used only a limi ed se o mic obial ma ke s (3–7). These mic obial ma ke s e lec ypically only a na ow pa o he o al mic obial exposu e. Recen ly, some s udies ha e applied combined indica o s o mic obial exposu e (8, 9), like di e - si y o mic obes, which appea s o ha e a p o ec i e e ec on Alle gy 69 (2014) 1092–1101 ©2014 The Au ho s. Alle gy Published by John Wiley & Sons L d.1092 This is an open access a icle unde he e ms o he C ea i e Commons A ibu ion-NonComme cial License, which pe mi s use, dis ibu ion and ep oduc ion in any medium, p o ided he o iginal wo k is p ope ly ci ed and is no used o comme cial pu poses. he de elopmen o as hma (8). Thus, he e is a clea need o use combined indica o s and o he new app oaches o desc ibe exposu e o en i onmen al mic obes. The objec i e o his s udy was he e o e o s udy p ospec- i ely he associa ions be ween se e al ma ke s o en i onmen- al mic obial exposu e—including hei combina ions—and he de elopmen o doc o -diagnosed as hma, espi a o y symp oms, a opic de ma i is, and sensi iza ion up o he age o 6 yea s. Me hods The s udy popula ion consis ed o child en bo n in Eas e n and Middle Finland: he i s hal o he s udy popula ion (N=214) belongs o a Eu opean bi h coho (PASTURE) (10) among a me s and non a me s, while he second hal o he coho is an ex ension o unselec ed child en (N=228) (11). P egnan women we e ec ui ed, whose child en we e bo n be ween Sep embe 2002 and May 2005. The s udy p o- ocol was app o ed by a local e hics commi ee in Finland (11). A w i en in o med consen was ob ained om he pa en s. Follow-up The i s ques ionnai e was adminis e ed du ing he hi d imes e o p egnancy (11). The ollow-up ques ionnai es (a he age o 2, 12, 18, and 24 mon hs, and he ea e annually), excep he ques ionnai e a 2 mon hs, enqui ed abou any wheezing, o he espi a o y symp oms apa om cold (wheezing, cough, noc u nal cough), and con ounde s o he ime pe iod a e he p eceding ques ionnai e. ‘As hma e e ’ was de ined as i s pa en - epo ed doc o - diagnosed as hma and/o second diagnoses o as hma ic (o obs uc i e) b onchi is. ‘Cu en as hma’ was de ined as hose as hma e e cases who also epo ed he use o as hma medica ion and/o wheezing in pas 12 mon hs in he 6-yea ollow-up. ‘A opic de ma i is e e ’ was de ined as i s pa en - epo ed doc o -diagnosed a opic de ma i is du - ing he ollow-up. House dus samples The p o ocols o dus collec ion a e explained in de ail in he Supplemen a y ma e ial. B ie ly, in he Finnish PAS- TURE s udy, ieldwo ke s ook wo samples om li ing oom loo s o he p esen epo . The i s samples we e p ocessed wi hou sie ing a U ech Uni e si y, whe e hey we e ex ac ed and analyzed o endo oxin (as a ma ke o G am-nega i e bac e ia) and EPS as desc ibed p e iously (12), and he s o ed esidue a e ex ac ion was la e ana- lyzed o b(1,3)-glucans (as a ma ke o ungal biomass) by hea ex ac ion and inhibi ion enzyme immunoassay (EIA) (13). The second samples we e p ocessed and analyzed a Na ional Ins i u e o Heal h and Wel a e (THL), whe e hey we e sie ed o emo e la ge pa icles, d ied, spli and hen s o ed, and inally analyzed o mu amic acid (as a ma ke o G am-posi i e bac e ia), e gos e ol (as a ma ke o ungal biomass), and lipopolysaccha ide (LPS 10:0–16:0 ) (as a ma ke o G am-nega i e bac e ia) by gas ch oma og aphy andem mass spec ome y (GC-MS–MS) (14) and by quan i a i e polyme ase chain eac ion (qPCR) o mic obial DNA (15–17). In he ex ended coho , esul s om only one dus sample om he li ing oom loo collec ed by he amilies a e used in he p esen epo . The sample was sie ed and d ied a THL, hen spli in aliquo s o a ious analyses: One aliquo was sen o U ech whe e i was ex ac ed and analyzed o endo oxin, EPS, and b(1,3)-glucans (13), and o he aliquo s we e analyzed o he chemical ma ke s (mu amic acid, e gos e ol, LPS 10:0–16:0 ) (14) and mic obial DNA (15–17). Quan i a i e polyme ase chain eac ion (qPCR) (15–17) was used o de e mine mic obes ha ha e been sugges ed o be common in mois u e-damaged buildings: wo gene a o G am-posi i e bac e ia, Mycobac e ium spp. and S ep omy- ces spp., and six ungal species, gene a, o g oups: he assay g oup o Aspe gillus umiga us/Neosa o ya ische i (AspNeo g oup); Cladospo ium spp.; he combined assay g oup o Penicillium spp., Aspe gillus spp., and Paecilomyces a io ii (PenAsp g oup); S achybo ys cha a um; T ichode ma i ide/ a o i ide/koningii (T ichode ma i ide g oup); and Wallemia sebi. Exposu e o all ma ke s was exp essed as loads, ha is, he amoun pe m 2 sampling su ace. As he labo a o y p oce- du es de e mine concen a ions (agen s pe mg dus ), loads a e calcula ed by mul iplying he concen a ions wi h he mea- su ed dus weigh pe m 2 . Howe e , as endo oxin, EPS, and glucans in he PASTURE samples we e de e mined in nonsi- e ed dus , while all o he analyses we e pe o med wi h sie ed dus , we calcula ed loads o he o me pa ame e s by mul i- plying wi h he load o sie ed dus in he pa allel sample aken o he o he analyses. Weigh s o hese pa allel sie ed and nonsie ed dus samples in he Finnish PASTURE s udy showed a ai co ela ion (Spea man =0.7), bu wi h a sys- ema ic di e ence: as expec ed, sie ed dus samples con ained on a e age 60% (95% CI 57, 64) less dus . Mean concen a- ions in sie ed and nonsie ed dus om non a m homes we e, howe e , e y simila , which sugges s ha endo oxin, EPS, o glucans a e no p e e en ially bound o he smalle (sie ed) o la ge ( emo ed by sie ing) pa icle ac ions. Abb e ia ions EPS, ungal ex acellula polysaccha ide an igens o Penicillium and Aspe gillus species; GC-MS–MS, gas ch oma og aphy andem mass spec ome y; LPS, lipopolysaccha ide; LPS 10:0-:16:0:0 , amoun (mol) o LPS in each sample. The numbe o mol o 3-hyd oxy a y acids o C 10:0 o C 16:0 ca bon chain leng h di ided by 4; qPCR, quan i a i e polyme ase chain eac ion; PASTURE, P o ec ion agains Alle gy STUdy in Ru al En i onmen s; OR, odds a io; aOR, adjus ed odds a io; sIgE, speci ic immunoglobulin E; kU/L, kilo uni s pe li e ; GEE, gene al es ima ing equa ions; EIA, inhibi ion enzyme immunoassay. Alle gy 69 (2014) 1092–1101 ©2014 The Au ho s. Alle gy Published by John Wiley & Sons L d. 1093 Ka onen e al. En i onmen al mic obial exposu e and as hma Blood sampling Venous blood samples collec ed a one (N=374) and a 6 yea s o age (N=304) we e analyzed o speci ic immuno- globulin E (sIgE) o 13 inhalan (De ma ophagoides p e onys- sinus, D. a inae, alde , bi ch, Eu opean hazel, g ass pollen mix u e, ye, mugwo , plan ain, ca , ho se and dog dande , and Al e na ia al e na a) and 6 ood alle gens (hen’s egg, cow’s milk, peanu , hazelnu , ca o , and whea ) (Mediwiss Analy ic, Moe s, Ge many) (18). The cu o le els o sIgE concen a ions we e 0.35 kU/l a one and 0.70 kU/l a 6 yea s o age. The highe cu o was used a he age o 6 yea s due o high p e alence o sensi iza ion (Table 1). S a is ical analyses Gene alized es ima ing equa ions (GEE) wi h an exchange- able co ela ion s uc u e o accoun o co ela ion be ween epea ed measu es wi hin subjec s we e used o de e mine associa ions be ween mic obial exposu e and epea ed mea- su es o pa en - epo ed wheezing and cough a di e en ages. Su i al analyses (disc e e- ime haza d models) we e used in analyzing as hma e e , cu en as hma, and a opic de ma i is e e . Logis ic eg ession was used o analyze sensi- iza ion o inhalan alle gens. The esul s a e p esen ed as adjus ed odds a ios (aORs) and hei 95% con idence in e - als (95% CI). In he analysis o single mic obial ma ke s, he amoun s o dus and mic obial ma ke s we e gene ally di ided in o h ee ca ego ies using e iles as cu o s. Fo Wallemia sebi, he lowes ca ego y consis ed o le els below he limi o de ec- ion (37.4%) and he es o he alues we e e enly di ided in o wo g oups (medium and high). S achybo ys cha a um and AspNeo g oup we e dicho omized a he de ec ion le el due o high numbe s o nonde ec ed alues. Mic obial di e - si y sco e was de ined as he sum o all he de ec ed qPCR ma ke s ( ange 0–8). Due o a low numbe o obse a ions wi h low and high di e si y, he a iable was ca ego ized in o ou classes: 0–4, 5, 6, and 7–8. To al mic obial quan i y sco e was calcula ed as he sum o h ee ma ke s o he h ee di e en mic obial g oups, ha is, o ungi, G am-nega i e bac e ia, and G am-posi i e bac- e ia (one ma ke pe g oup). Single ma ke s we e i s di ided in o i e g oups using quin iles as cu o s wi h sco es 0, 1, 2, 3, and 4 om lowes o highes and hen summed. Table 1 Incidence o doc o -diagnosed as hma and a opic de ma i is and he poin p e alence o espi a o y symp oms du ing he i s 6 yea s and p e alence o sensi iza ion o inhalan alle gens a he age o 1 and 6 yea s A 1 yea A 1.5 yea s A 2 yea s A 3 yea s A 4 yea s A 5 yea s A 6 yea s Nn%Nn%Nn%Nn%Nn%Nn%Nn % As hma e e * 389 6 1.5 353 9 2.6 332 14 4.2 291 15 5.2 271 9 3.3 257 6 2.3 247 3 1.2 Cu en as hma† 353 3 0.9 331 7 2.1 315 8 2.5 290 7 2.4 281 3 1.1 274 4 1.5 270 3 1.1 Wheezing Apa om cold 391 41 10.5 358 24 6.7 358 25 7.0 338 23 6.8 358 29 8.1 346 22 6.4 357 25 7.0 Any 391 91 23.3 358 58 16.2 358 67 18.7 339 65 19.2 358 68 19.0 348 46 13.2 357 51 14.3 Cough Apa om cold 391 98 25.1 358 66 18.4 358 85 23.7 338 80 23.7 357 91 25.5 348 87 25.0 357 94 26.3 Noc u nal 386 18 4.7 358 47 13.1 358 76 21.2 338 63 18.6 357 78 21.9 348 59 17.0 357 79 22.1 A opic de ma i is e e ‡ 390 61 15.6 300 24 8.0 265 12 4.5 230 14 6.1 211 9 4.3 199 7 3.5 188 2 1.1 Sensi iza ion o inhalan alle gens§ 374 78 20.9 304 115 37.8 N=in espi o y symp oms: o al numbe o child en wi h in o ma ion on gi en symp om a each ollow-up; in sensi iza ion: o al numbe o child en wi h IgE esul s a a gi en ime poin ; in as hma e e : o al numbe o child en a each ollow-up who has in o ma ion on as hma e e a each ollow-up and has no censo ed om he ollow-up due o onse o as hma e e ; in cu en as hma: o al numbe o child en a each ollow-up who has in o ma ion on as hma e e and as hma medica ion/any wheezing a each ollow-up and has no censo ed om he ollow-up due o onse o cu en as hma; in a opic de ma i is e e : o al numbe o child en a each ollow-up who has in o ma ion on a opic de ma i is a each ollow-up and has no censo ed om he ollow-up due o onse o a opic de ma i is. n= he numbe o child en wi h he p esence o wheezing/cough symp om o wi h posi i e IgE esul , o he numbe o newly diagnosed as hma/cu en as hma/a opic de ma i- is cases a each ollow-up. *Doc o -diagnosed as hma e e is de ined as doc o -diagnosed as hma a leas once and/o doc o -diagnosed as hma ic b onchi is mo e han once du ing he 6-yea ollow-up. †Cu en as hma is de ined as as hma e e wi h as hma medica ion and/o wheezing symp om pas 12 mon hs a he age o 6 yea s. ‡A opic de ma i is e e is de ined as doco o -diagnosed as opic de ma i is du ing he ollow-up. §A opic sensi iza ion is de ined as any inhalan sIgE ≥0.35 kU/l a he age o 1 yea and ≥0.70 kU/l a he age o 6 yea s. Alle gy 69 (2014) 1092–1101 ©2014 The Au ho s. Alle gy Published by John Wiley & Sons L d.1094 En i onmen al mic obial exposu e and as hma Ka onen e al. All possible combina ions wi h measu ed mic obial ma ke s we e used o c ea e ou di e en o al quan i y sco es o mic obial exposu e: (1) e gos e ol, endo oxin, and mu amic acid; (2) b-D-glucan, endo oxin, and mu amic acid; (3) e gos- e ol, LPS 10:0-16:0 , and mu amic acid; and (4) b-D-glucan, LPS 10:0-16:0 , and mu amic acid. The c ea ed a iables, wi h anges om 0 o 12, we e di ided in o i e g oups using quin iles as cu o s o s a is ical analyses. Linea and qua- d a ic ends in he associa ions be ween mic obial di e si y o o al quan i y sco es and heal h ou comes we e es ed using polynomial con as s (19). All models we e adjus ed o s udy coho , a ming, and well-known isk ac o s o as hma (ma e nal his o y o alle - gic diseases, gende , numbe o olde siblings, smoking du - ing p egnancy). The models o o al mic obial quan i y sco e (e gos e ol, endo oxin, and mu amic acid) and as hma e e , a opic de ma i is e e , and sensi iza ion o inhalan alle gen we e ca e ully es ed o 20 addi ional con ounding ac o s, which ha e been desc ibed ea lie (12). I a con ounde chan- ged he es ima es o he o al mic obial quan i y sco e by mo e han 10%, all he models o he ou come we e adjus ed o he ac o . All he models o espi a o y symp oms and cu en as hma we e adjus ed o he same con ounding ac- o s as in he as hma e e models. Gi en an exposu e p e a- lence o 25%, powe calcula ions (b=80%, a=0.05) show ha wi h hese da a, we can de ec an odds a io o 2.52 o as hma e e wi h a popula ion p e alence o 14%. The da a we e analyzed using SAS 9.2 o Windows (SAS lnc., Ca y, NC, USA). Resul s The p esen analyses included all hose 410 child en wi h da a on a leas one mic obial ma ke a ailable in any o he quan i y sco es. O hose, 341 (83%) had he esul s o all mic obial de e mina ions. Du ing he 6-yea ollow-up, 62 o 389 child en wi h da a on as hma de eloped as hma, and o hose child en, 35 s ill had cu en as hma a he age o 6 yea s (Table 1). Dis ibu ions o all mic obial ma ke s a e shown in Table 2. Mos co ela ions be ween he amoun o house dus and single mic obial ma ke s as well as in e co ela ions be ween wo mic obial ma ke s oge he we e be ween 0.30 and 0.85 (Table S1). To al mic obial quan i y sco es we e co ela ed Table 2 Le els o mic obial ma ke s N<LOD* (%) Pe cen iles 25 h Median 75 h 95 h Amoun o dus (mg/m 2 )†410 0 173 311 587 1332 G am-posi i e bac e ia o ma ke o G am-posi i e bac e ia Mycobac e ium (cells/m 2 ) 399 21 (5.3) 186 857 641 427 1 753 136 7 041 045 S ep omyces (cells/m 2 ) 399 7 (1.8) 3923 13 089 45 939 462 289 Mu amic acid (ng/m 2 ) 386 7 (1.8) 3122 6616 13 713 34 852 Ma ke o G am-nega i e bac e ia LPS 10:0–16:0 (nmol/m 2 ) 398 2 (0.5) 3 8 15 46 Endo oxin (EU/m 2 ) 372 0 2668 6253 13 485 66 569 Fungi o ma ke o ungi PenAsp g oup (cells/m 2 ) 399 3 (0.8) 13 580 59 405 233 616 1 282 236 T ichode ma i ide g oup (cells/m 2 ) 400 62 (15.5) 340 2028 9906 153 816 Wallemia sebi (cells/m 2 ) 400 152 (38.0) 0 883 11 541 152 041 Cladospo ium (cells/m 2 ) 399 50 (12.5) 3652 25 683 104 690 982 138 AspNeo g oup (cells/m 2 ) 400 320 (80.0) 0 0 0 14 254 S achybo ys cha a um (cells/m 2 ) 400 390 (97.5) 0 0 0 0 EPS (EPSU/m 2 ) 374 15 (2.1) 7079 15 312 38 025 126 107 b-D-glucan (lg/m 2 ) 376 0 511 970 1818 4066 E gos e ol (ng/m 2 ) 401 0 1473 3015 7040 21 978 Quan i y sco es E gos e ol, endo oxin, and mu amic acid 347 3 6 9 12 b-D-glucan, endo oxin, and mu amic acid 352 3 6 9 12 E gos e ol, LPS 10:0–16:0 , and mu amic acid 380 3 6 9 12 b-D-glucan, LPS 10:0–16:0 , and mu amic acid 354 3 7 9 12 Di e si y sco e‡399 5 6 6 7 *LOD = he le el o de ec ion, pe cen age o samples unde de ec ion le el om analyzed samples (%). †The amoun o dus e e s o dus samples used o analyses o qPCRs, e gos e ol, mu amic acid, and LPS (sie ed). ‡The numbe o de ec ed 6 ungal o 2 bac e ial qPCRs. Alle gy 69 (2014) 1092–1101 ©2014 The Au ho s. Alle gy Published by John Wiley & Sons L d. 1095 Ka onen e al. En i onmen al mic obial exposu e and as hma wi h each o he (a ound =0.90) and also wi h amoun o dus (a ound =0.85), bu co ela ions be ween o al mic o- bial quan i y sco es and qPCR esul s we e lowe (a ound =0.30), excep o Mycobac e ium and he PenAsp g oup (a ound =0.50). Di e si y sco e had low co ela ions wi h amoun o dus , all mic obial ma ke s, and o al mic obial quan i y sco es (a ound =0.20), bu highe wi h Wallemia sebi ( =0.66). A B Figu e 1 Adjus ed associa ions be ween amoun o dus and bac- e ial (A) and ungal (B) exposu e a he age o 2 mon hs and as hma and espi a o y symp oms du ing he i s 6 yea s o li e and sensi iza ion o inhalan alle gen a he age o 6 yea s. Models a e adjus ed o s udy coho , a ming s a us, gende , ma e nal his- o y o alle gic diseases (hay e e , a opic de ma i is, and/o as hma), smoking du ing p egnancy, he numbe o olde siblings, and li ing a ea o he home. Models o sensi iza ion o inhalan alle gens (cu poin ≥0.70 kU/l) a e addi ionally adjus ed o loo ype o dus sampling. Alle gy 69 (2014) 1092–1101 ©2014 The Au ho s. Alle gy Published by John Wiley & Sons L d.1096 En i onmen al mic obial exposu e and as hma Ka onen e al. In gene al, he associa ions o single mic obial ma ke s wi h as hma e e and espi a o y symp oms we e mos ly nonsigni - ican , while he shape o he associa ion a ied (Fig. 1). To al mic obial quan i y sco e had a signi ican in e ed-U-shaped associa ion wi h as hma e e : he highes isk o as hma e e was ound wi h he medium le el o o al mic obial quan i y sco e and he lowes isk wi h he highes quan i y sco e (Fig. 2, Table 3). The associa ion was independen o he di e si y sco e. The associa ions we e qui e simila o cu en as hma (Table 3) and when using o he mic obial ma ke s han e gos e ol, endo oxin, and mu amic acid o c ea e he o al mic obial quan i y sco e (Table S2) as well as when using EPS as a ma ke o ungi (da a no shown). To al mic obial quan i y sco es ended o be in e sely associa ed wi h espi a o y symp oms (Table S2). The in e ed-U-shaped associa ion be ween o al mic obial quan i y sco e and as hma e e ended o be s onge a e adjus ing o he amoun o dus (da a no shown). Adjus men o dus also s eng hened he associa ions be ween as hma, and S ep o- myces, endo oxin, PenAsp g oup, b-D-glucan and e gos e ol, bu no he o he single mic obial ma ke s (da a no shown). The associa ions be ween o al mic obial quan i y sco e and any wheezing we e simila among child en who we e sensi ized agains a leas one o he es ed inhalan o ood alle gens a he age o 6 yea s and who we e no sensi ized (P- alue =0.66 o in e ac ion e m). In con as , in he s a i ied analyses, he in e ed-U-shaped associa ions be ween he quan i y sco es and as hma e e we e always s onge among nona opic han among a opic child en, bu he P- alues o in e ac ion anged be ween 0.08 and 0.28 (da a no shown). The e we e no in e - ac ions wi h bi h coho s (P- alues ≥0.27 o in e ac ion e ms). Associa ions be ween o al mic obial quan i y sco e and any wheezing we e also simila in a ms and in u al/sub- u ban a eas (P- alue =0.58 o in e ac ion e m). Due o he low numbe s o child en wi h as hma e e among a me s, in e ac ion es s be ween o al mic obial quan i y sco e and a ming on as hma e e could no be pe o med. Di e si y sco e was associa ed wi h a dec eased isk o any wheezing, and a simila endency was seen wi h as hma e e and cu en as hma (Table 3). The associa ion wi h as hma e e was, howe e , ai ly sensi i e o di e en adjus men s, especially adjus men o a ming. The associa ion wi h any wheezing was simila among child en om a ms and om u al/subu ban a eas as well as child en who we e sensi ized agains a leas one o he es ed inhalan o ood alle gens a he age o 6 yea s and who we e no sensi ized (P- alues ≥0.4 o in e ac ion e ms) (da a no shown). Mos o he associa ions be ween single mic obial ma ke s and a opic de ma i is and sensi iza ion o inhalan alle gens a he age o 1 (Table S3) and 6 yea s (Fig. 1) we e nonsig- ni ican . To al mic obial quan i y sco e was no associa ed wi h a opic de ma i is o sensi iza ion o inhalan alle gen a he age o ei he 1 yea (da a no shown) o 6 yea s (Table 3). Di e si y sco e was signi ican ly (in e ed-U shape) associa ed wi h sensi iza ion o inhalan alle gen a he age o 6 yea s (Table 3), while simila endency, hough nonsigni i- can , was seen wi h sensi iza ion o inhalan alle gen a 1 yea (Table S4). No associa ion was ound be ween di e - si y sco e and a opic de ma i is (Table S4). Discussion The p esen s udy sugges s ha an index o o al quan i y o en i onmen al mic obial exposu e may p edic as hma be e han single mic obial ma ke s, independen ly o mic obial di e si y sco e and amoun o dus . The di e si y sco e dec eased he isk o wheezing and was signi ican ly (in e ed-U shape) associa ed wi h sensi iza ion o inhalan alle gens. Only a ew associa ions we e obse ed be ween sin- gle mic obial ma ke s and he ou comes s udied. This is he i s s udy using an index o o al mic obial quan i y sco e when s udying he isk o as hma. P e iously, one c oss-sec ional s udy ound a endency o in e se associa- ion be ween mic obial quan i y sco e (LPS 10:0–16:0 , e gos- e ol, and mu amic acid) and alle gic sensi iza ion (9). Typically, epidemiological s udies ha e used only one mic o- bial ma ke as a su oga e o o al mic obial exposu e. Howe e , i is clea ha such assessmen s p o ide an incom- ple e pic u e o he o al quan i y and quali y o human mic obial exposu e. Due o high co ela ions, a majo di i- cul y o ea lie s udies has been o sepa a e he e ec s o sin- gle mic obial ma ke s om each o he and om he e ec o he amoun o dus (5, 20). In he p esen s udy, o al mic o- bial quan i y sco e p edic ed he isk o as hma indepen- den ly o mic obial di e si y and amoun o dus . The eason behind he obse ed associa ion be ween o al mic obial quan i y sco e and as hma is unclea . A po en ial explana ion is ha di e en componen s o mic obial g oups o gene a sha e common an igen o molecule s uc u es on hei cell wall su ace, which a e de ec ed by pa e n ecogni- ion ecep o s in inna e immuni y cells and ha e a p o ound in luences on he immune sys em (21). I o al exposu e o Figu e 2 Risk o as hma in ela ion o o al mic obial quan i y sco e in quin iles (sum o e gos e ol, endo oxin, and mu amic acid) (aOR). Model is adjus ed o s udy coho , a ming s a us, gende , ma e - nal his o y o alle gic diseases (hay e e , a opic de ma i is, and/o as hma), smoking du ing p egnancy, he numbe o olde siblings, li ing a ea o he home, and di e si y sco e. Alle gy 69 (2014) 1092–1101 ©2014 The Au ho s. Alle gy Published by John Wiley & Sons L d. 1097 Ka onen e al. En i onmen al mic obial exposu e and as hma Table 3 Adjus ed associa ions be ween mic obial quan i y o di e si y sco es and as hma o espi a o y symp oms up o he age o 6 yea s and cu en as hma and sensi iza ion o inhalan alle gen a 6 yea s o age As hma e e Cu en as hma Any wheezing¥Noc u nal cough¥ Sensi iza ion o inhalan alle gens a 6 yea s¤ N#nN‡aOR (95% CI) N#nN‡aOR (95% CI) aOR (95% CI) aOR (95% CI) NnaOR (95% CI) Quan i y sco e£ Re e ence (1s quin ile) 63 8 354 1 56 5 352 1 1 1 53 19 1 2nd quin ile 73 12 390 1.73 (0.67–4.45) 63 6 380 1.32 (0.38–4.61) 0.65 (0.35–1.22) 0.68 (0.39–1.21) 57 26 2.19 (0.92–5.23) 3 d quin ile 62 13 300 2.24 (0.87–5.75)^ 57 7 297 2.01 (0.59–6.79) 0.89 (0.49–1.62) 0.86 (0.48–1.55) 44 17 1.42 (0.55–3.64) 4 h quin ile 56 11 311 1.78 (0.67–4.69) 53 7 318 1.86 (0.55–6.32) 0.98 (0.55–1.75) 0.70 (0.40–1.24) 47 17 1.32 (0.51–3.40) 5 h quin ile 75 3 438 0.34 (0.09–1.36) 67 2 409 0.39 (0.07–2.11) 0.73 (0.43–1.24) 0.68 (0.38–1.20) 57 20 1.50 (0.61–3.70) P- alue <0.001¶0.03¶0.95¶0.74¶0.46¶ Di e si y sco e§ Re e ence (0–4) 45 9 232 1 37 8 215 1 1 1 33 11 1 5 108 13 566 0.69 (0.27–1.78) 95 7 557 0.30 (0.10–0.89)* 0.68 (0.38–1.20) 0.90 (0.50–1.62) 81 36 2.10 (0.84–5.27) 6 133 22 736 1.03 (0.43–2.49) 122 10 731 0.35 (0.12–0.94)* 0.54 (0.30–0.97)* 1.23 (0.71–2.15) 108 46 2.16 (0.86–5.45) 7–8 43 3 259 0.64 (0.15–2.76) 42 2 253 0.40 (0.07–2.29) 0.14 (0.06–0.32)** 1.21 (0.58–2.50) 36 6 0.42 (0.12–1.50) P- alue 0.67†0.31†<0.001†0.45†0.001¶ £Sum o quin iles o loads o e gos e ol, endo oxin, and mu amic acid. §Numbe o de ec ed qPCRs. Models a e adjus ed o ime, s udy coho , a ming s a us, gende , ma e nal his o y o alle gic diseases (hay e e , a opic de ma i is, and/o as hma), smoking du ing p egnancy, he num- be o olde siblings, li ing a ea o he home, and ei he di e si y sco e o quan i y sco e. N# To al numbe o child en in he beginning o ollow-up, n=numbe o cases in each ca ego y. N‡Subjec s con ibu ed up o se en epea ed obse a ions o his su i al analysis using disc e e- ime haza d models (DTH). ¥Subjec s con ibu ed up o se en epea ed obse a ions o his analysis using gene alized es ima ing equa ions. ¤Models a e addi ionally adjus ed o loo ype o dus sampling. P- alue ^<0.1, *<0.05, **<0.01, †P- alue o linea end es , ¶P- alue o quad a ic end es . Alle gy 69 (2014) 1092–1101 ©2014 The Au ho s. Alle gy Published by John Wiley & Sons L d.1098 En i onmen al mic obial exposu e and as hma Ka onen e al. such s uc u es is he main causal ac o explaining he asso- cia ion be ween mic obial exposu e and lowe isk o alle gic diseases, hen o al mic obial quan i y sco e is likely o be e e lec his exposu e han single mic obial ma ke s. Va ying co ela ions o single mic obial ma ke s wi h o al mic obial exposu e would hen explain he o en weak and con lic ing indings om ea lie s udies. In addi ion, some s udies ha e shown syne gis ic in e ac ions be ween di e en exposing agen s (22), and an index o o al mic obial quan i y sco e may cap u e such syne gis ic e ec s. Da a om deep sequencing o mic obial exposu e (23) will gi e oppo uni ies o explo e u he he abo e po en ial mechanisms. To al mic obial quan i y sco e had an in e ed-U-shaped associa ion wi h as hma. P e ious s udies ha e sugges ed ha he immune sys em may eac i s wi h enhanced esponse and subsequen ly wi h ole ance (24). Associa ions o simila shape ha e also been epo ed ea lie be ween alle gen le els and b-D-glucan, and isk o as hma, wheezing, o a opic sen- si iza ion (25–27). The shape o he associa ion in he p esen s udy may also ela e o some unknown cha ac e is ics o he p esen s udy, gi en ha some single mic obial ma ke s also showed an associa ion o a simila shape. Fu he s udies in la ge coho s and in di e en se ings a e needed o be e de ine he shape o he associa ion be ween o al mic obial quan i y sco e and isk o as hma. We ound a simila in e se associa ion be ween measu ed mic obial di e si y sco e and wheezing, as has been epo ed wi h as hma in wo ecen c oss-sec ional s udies (8). The asso- cia ion wi h as hma was nonsigni ican , hough in he same di ec ion. The e we e di e ences in he mic obiological me h- ods used be ween he p esen s udy and p e ious s udies. We used qPCR, a me hod ha quan i a i ely desc ibes he p es- ence o speci ic iable and non iable o ganisms o p io chosen mic obes (28). The p e ious s udies ei he cul i a ed iable mic obes (GABRIELA s udy) o used single-s and con o ma- ion polymo phism analyses, which is a quali a i e a he han quan i a i e me hod (PARSIFAL s udy) (8). We calcula ed a di e si y sco e based on 6 ungal and 2 bac e ial qPCR assays ep esen ing ei he mic obial species, gene a, o g oups, which hence only p o ides a sugges i e and ough es ima e o he o e all mic obial di e si y. None heless, he esul s a e in line wi h p e ious s udies (8) a ge ing a b oade spec um o mic obes. Taken oge he , he s udies sugges ha mic obial di e si y sco e may p edic as hma and may be a use ul ma ke o desc ibe he quali y o mic obial exposu e. The obse ed associa ions be ween s udied heal h ou - comes and single mic obial ma ke s should be in e p e ed wi h cau ion, as no adjus men s o mul iple es ing we e done. The mos in e es ing indings we e maybe he p o ec- i e associa ions in he case o G am-posi i e bac e ia (Myco- bac e ium spp., S ep omyces spp., and mu amic acid), which ha e been epo ed ea lie o mu amic acid (6, 29). The main s eng hs o he p esen s udy a e he p ospec i e bi h coho design wi h high pa icipa ion a e ( a ied be ween 80% and 95% in each ollow-up) and a good a ie y o measu ed ma ke s o exposu e o en i onmen al mic obes. The small size o ou s udy coho is he main limi a ion o he s udy, and he esul s need o be con i med in u he s ud- ies. The mic obial ma ke s and combina ions o hose used o build o al mic obial quan i y sco es in he p esen s udy do no measu e o al mic obial exposu e wi hou e o , o example using mu amic acid oge he wi h endo oxin o e es i- ma es he amoun o G am-nega i e bac e ia, because mu amic acid is ound also in G am-nega i e bac e ia, al hough o a lesse ex en (30). These weaknesses do no , howe e , appea o be a majo p oblem, as he di e en quan i y sco es used in he p esen s udy showed e y simila associa ions. In conclusion, ou indica o o o al quan i y o mic obial exposu e p edic ed as hma be e han single mic obial ma k- e s. De ec ed associa ions we e independen o he measu ed mic obial di e si y and amoun o dus . Be e indica o s o o al quan i y and di e si y o mic obial exposu e a e needed in s udies on he de elopmen o as hma. Acknowledgmen s We would like o hank he amilies o hei pa icipa ion in he s udy and s udy nu ses Raija Jun unen, Riikka Juola, and Seija An ikainen o ield wo k; Hanna Lepp€ anen, Ka ja Saa nio, Heli Ma ikainen, Susanne an den B ink, Timmo y Wigboldus, Mi ian Boe e, Sieg ied de Wind, and Jack Spi ho en o mic obiological analyses; Pekka Tii anen o suppo in s a is ical analyses and da a managemen ; Hanna-Ma i Takkinen o illus a ions, Asko Veps€ al€ ainen and Timo Kauppila o da a managemen ; and Ma k Phillips o e ising he language o his manusc ip . Funding This s udy was suppo ed by esea ch g an s om he Eu opean Union QLK4-CT-2001-00250; G adua e School in En i onmen al Heal h (SYTYKE); EVO- unding; Fa me s’ Social Insu ance Ins i u ion (Mela); he Academy o Finland (g an 139021); he Juho Vainio Founda ion; he Eu opean commission as pa o HITEA (Heal h E ec s o Indoo Pollu- an s: In eg a ing mic obial, oxicological and epidemiological app oaches), G an ag eemen no. 211488 unde he Se en h F amewo k P og amme, Topic ENV.2007.1.2.1.1. ‘Indoo ai pollu ion in Eu ope: An eme ging en i onmen al heal h issue’; and by he Na ional Ins i u e o Heal h and Wel a e, Finland. Au ho con ibu ions All au ho s app o ed he submi ed e sion. Ka onen in ol ed in s a is ical da a analyses, da a collec ing, manu- sc ip p epa a ion, and d a ing. Hy € a inen and Rin ala in ol ed in dus sampling assessmen , mic obial analyses, and manusc ip p epa a ion. Ko ppi and Remes in ol ed in au ho i y on pedia ic issues and manusc ip p epa a ion. T aubel, Keski-Nisula, and Lampi in ol ed in manusc ip p epa a ion. Doekes and Geh ing in ol ed in mic obial analyses and manusc ip p epa a ion. Renz and P e e le in ol ed in IgE measu emen s and manusc ip p epa a ion. Genunei in ol ed in da a managemen and manusc ip p ep- a a ion. on Mu ius and Pekkanen in ol ed in s udy concep and design, and manusc ip p epa a ion. Alle gy 69 (2014) 1092–1101 ©2014 The Au ho s. Alle gy Published by John Wiley & Sons L d. 1099 Ka onen e al. En i onmen al mic obial exposu e and as hma Con lic s o in e es The au ho s decla e ha hey ha e no con lic s o in e es . Suppo ing In o ma ion Addi ional Suppo ing In o ma ion may be ound in he online e sion o his a icle: Table S1. Spea man co ela ions o mic obial ma ke s pe sampled a ea (/m 2 ). Table S2. Adjus ed associa ions be ween quan i y o di e - si y sco es and as hma o espi a o y symp oms. Table S3. Adjus ed associa ions be ween mic obial expo- su e and a opic de ma i is up o he age o 6 yea s and sensi- iza ion o inhalan alle gen a he age o 1 yea . Table S4. Associa ions be ween mic obial di e si y and a opic de ma i is e e un il he age o 6 yea s and sensi iza- ion o inhalan alle gen a he age o 1 and 6 yea s. Re e ences 1. on Mu ius E. As hma and alle gies in u al a eas o Eu ope. P oc Am Tho ac Soc 2007;4:212–216. 2. Mendy A, Gasana J, Viei a ER, Fo no E, Pa el J, Kadam P e al. Endo oxin exposu e and childhood wheeze and as hma: a me a- analysis o obse a ional s udies. J As hma 2011;48:685–693. 3. Sch am-Bijke k D, Doekes G, Douwes J, Boe e M, Riedle J, Ublagge E e al. Bac e- ial and ungal agen s in house dus and wheeze in child en: he PARSIFAL s udy. Clin Exp Alle gy 2005;35:1272–1278. 4. Douwes J, Siebe s R, Wou e s IM, Doekes G, Fi zha is P, C ane J. Endo oxin, (1 -> 3)-be a-D-glucans and ungal ex a-cellula polysaccha ides in New Zealand homes: a pilo s udy. Ann Ag ic En i on Med 2006;13:361–365. 5. Geh ing U, Hein ich J, Hoek G, Gio annan- gelo M, No dling E, Bellande T e al. Bac- e ia and mould componen s in house dus and child en’s alle gic sensi isa ion. Eu Respi J 2007;29:1144–1153. 6. So dillo JE, Ho man EB, Celedon JC, Li onjua AA, Mil on DK, Gold DR. Mul iple mic obial exposu es in he home may p o ec agains as hma o alle gy in childhood. Clin Exp Alle gy 2010;40: 902–910. 7. Hy € a inen A, Sebas ian A, Pekkanen J, La sson L, Ko ppi M, Pu us T e al. Cha ac e izing mic obial exposu e wi h e gos e ol, 3-hyd oxy a y acids, and ia- ble mic obes in house dus : de e minan s and associa ion wi h childhood as hma. A ch En i on Occup Heal h 2006;61:149– 157. 8. Ege MJ, Maye M, No mand AC, Genunei J, Cookson WO, B aun-Fah lande C e al. Exposu e o en i onmen al mic oo ganisms and childhood as hma. N Engl J Med 2011;364:701–709. 9. on He zen L, Hy a inen A, Laa ikainen T, Makela MJ, Ne alainen A, Va iainen E e al. Risk o a opy associa ed wi h mic o- bial componen s in house dus . Ann Alle gy As hma Immunol 2010;104:269–270. 10. on Mu ius E, Schmid S. The PASTURE p ojec : EU suppo o he imp o emen o knowledge abou isk ac o s and p e en i e ac o s o a opy in Eu ope. Alle gy 2006;61:407–413. 11. Ka onen AM, Hy a inen A, Roponen M, Ho mann M, Ko ppi M, Remes S e al. Con i med mois u e damage a home, espi- a o y symp oms and a opy in ea ly li e: a bi h-coho s udy. Pedia ics 2009;124: e329–e338. 12. Ka onen AM, Hy a inen A, Geh ing U, Ko ppi M, Doekes G, Riedle J e al. Expo- su e o mic obial agen s in house dus and wheezing, a opic de ma i is and a opic sensi- iza ion in ea ly childhood: a bi h coho s udy in u al a eas. Clin Exp Alle gy 2012;42:1246–1256. 13. Sch am D, Doekes G, Boe e M, Douwes J, Riedle J, Ublagge E e al. Bac e ial and ungal componen s in house dus o a m child en, Rudol S eine School child en and e e ence child en - he PARSIFAL S udy. Alle gy 2005;60:611–618. 14. Lappalainen MH, Hy € a inen A, Hi onen MR, Rin ala H, Roi ainen J, Renz H e al. High indoo mic obial le els a e associa ed wi h educed Th1 cy okine sec e ion capaci y in in ancy. In A ch Alle gy Immunol 2012;159:194–203. 15. Rin ala H, Ne alainen A. Quan i a i e measu emen o s ep omyce es using eal- ime PCR. J En i on Moni 2006;8:745–749. 16. To inen E, To kko P, Rin ala AN. Real- ime PCR de ec ion o en i onmen al myco- bac e ia in house dus . J Mic obiol Me hods 2010;82:78–84. 17. Haugland RA, Va ma M, Wyme LJ, Vespe SJ. Quan i a i e PCR analysis o selec ed Aspe gillus, Penicillium and Paecilomyces species. Sys Appl Mic obiol 2004;27:198–210. 18. He zum I, Blume N, Ke s en W, Renz H. Diagnos ic and analy ical pe o mance o a sc eening panel o alle gy. Clin Chem Lab Med 2005;43:963–966. 19. Da is MJ. Con as coding in mul iple eg ession analysis: s eng hs, weak- nesses, and u ili y o popula coding s uc u es. Jou nal o Da a Science 2010;8:61–73. 20. Douwes J, an S ien R, Doekes G, Smi J, Ke kho M, Ge i sen J e al. Does ea ly indoo mic obial exposu e educe he isk o as hma? The P e en ion and Incidence o As hma and Mi e Alle gy bi h coho s udy. J Alle gy Clin Immunol 2006;117:1067–1073. 21. Palm NW, Medzhi o R. Pa e n ecogni ion ecep o s and con ol o adap i e immuni y. Immunol Re 2009;227:221–233. 22. Hu unen K, Pelkonen J, Nielsen KF, Nuu- inen U, Jussila J, Hi onen MR. Syne gis ic in e ac ion in simul aneous exposu e o S ep omyces cali o nicus and S achybo ys cha a um. En i on Heal h Pe spec 2004;112:659–665. 23. Ma gulies M, Egholm M, Al man WE, A iya S, Bade JS, Bemben LA e al. Genome sequencing in mic o ab ica ed high-densi y picoli e eac o s. Na u e 2005;437:376–380. 24. Liu AH. Endo oxin exposu e in alle gy and as hma: econciling a pa adox. J Alle gy Clin Immunol 2002;109:379–392. 25. To ey ER, Almq is C, Li Q, C isa ulli D, Ma ks GB. Nonlinea ela ionship o mi e alle gen exposu e o mi e sensi iza ion and as hma in a bi h coho . J Alle gy Clin Immunol 2008;122:114–118. 26. Pla s-Mills T, Vaughan J, Squillace S, Wood olk J, Spo ik R. Sensi isa ion, as hma, and a modi ied Th2 esponse in child en exposed o ca alle gen: a popula ion-based c oss-sec ional s udy. Lance 2001;357:752– 756. 27. Iossi o a YY, Reponen T, Be ns ein DI, Le in L, Kal a H, Campo P e al. House dus (1-3)-be a-D-glucan and wheezing in in an s. Alle gy 2007;62:504–513. 28. Hi onen MR, Ruo salainen M, Sa olainen K, Ne alainen A. E ec o iabili y o ac i- nomyce e spo es on hei abili y o s imula e p oduc ion o ni ic oxide and eac i e oxygen species in RAW264.7 mac ophages. Toxicology 1997;124:105–114. Alle gy 69 (2014) 1092–1101 ©2014 The Au ho s. Alle gy Published by John Wiley & Sons L d.1100 En i onmen al mic obial exposu e and as hma Ka onen e al.