A novel multi-approach protocol for the characterization of occupational exposure to organic dust-swine production case study

oxics A icle A No el Mul i-App oach P o ocol o he Cha ac e iza ion o Occupa ional Exposu e o O ganic Dus —Swine P oduc ion Case S udy Ca la Viegas 1,2,*ID , Tiago Fa ia 1,3, Ana Mon ei o 1, Liliana A anha Cae ano 1,4 ID , Elisabe e Ca olino 1, Ani a Quin al Gomes 1,5 and Susana Viegas 1,2 ID 1GIAS, ESTeSL—Escola Supe io de Tecnologia da Saúde de Lisboa, Ins i u o Poli écnico de Lisboa, 1990-096 Lisbon, Po ugal; [email p o ec ed] (T.F.); [email p o ec ed] (A.M.); [email p o ec ed] (L.A.C.); [email p o ec ed] (E.C.); [email p o ec ed] (A.Q.G.); [email p o ec ed] (S.V.) 2Cen o de In es igação em Saúde Pública, Escola Nacional de Saúde Pública, Uni e sidade NOVA de Lisboa, 1600-560 Lisbon, Po ugal 3Cen o de Ciências e Tecnologias Nuclea es, Ins i u o Supe io Técnico, Uni e sidade de Lisboa, E.N. 10 ao km 139,7, 2695-066 Bobadela LRS, Po ugal 4Resea ch Ins i u e o Medicines (iMed.ULisboa), Facul y o Pha macy, Uni e si y o Lisbon, 1649-004 Lisbon, Po ugal 5Facul y o Medicine, Uni e si y o Lisbon Ins i u e o Molecula Medicine, 1649-028 Lisbon, Po ugal *Co espondence: [email p o ec ed] Recei ed: 13 No embe 2017; Accep ed: 25 Decembe 2017; Published: 27 Decembe 2017 Abs ac : Swine p oduc ion has been associa ed wi h heal h isks and wo ke s’ symp oms. In Po ugal, as in o he coun ies, la ge-scale swine p oduc ion in ol es se e al ac i i ies in he swine en i onmen ha equi e di ec in e en ion, inc easing wo ke s’ exposu e o o ganic dus . This s udy desc ibes an upda ed p o ocol o he assessmen o occupa ional exposu e o o ganic dus , o un eil an accu a e scena io ega ding occupa ional and en i onmen al isks o wo ke s’ heal h. The pa icle size dis ibu ion was cha ac e ized ega ding mass concen a ion in i e di e en size anges (PM0.5, PM1, PM2.5, PM5, PM10). Biobu den was assessed, by bo h ac i e and passi e sampling me hods, in ai , on su aces, loo co e ing and eed samples, and analyzed h ough cul u e based-me hods and qPCR. Smalle size ange pa icles exhibi ed he highes coun s, wi h indoo pa icles showing highe pa icle coun s and mass concen a ion han ou doo pa icles. The limi alues sugges ed o o al bac e ia load we e su passed in 35.7% (10 ou o 28) o samples and o ungi in 65.5% (19 ou o 29) o samples. Among Aspe gillus gene a, sec ion Ci cumda i was he mos p e alen (55%) on mal ex ac aga (MEA) and Ve sicolo es he mos iden i ied (50%) on dichlo an glyce ol (DG18). The esul s documen a wide cha ac e iza ion o occupa ional exposu e o o ganic dus on swine a ms, being use ul o policies and s akeholde s o ac o imp o e wo ke s’ sa e y. The me hods o sampling and analysis employed we e he mos sui able conside ing he pu pose o he s udy and should be adop ed as a p o ocol o be ollowed in u u e exposu e assessmen s in his occupa ional en i onmen . Keywo ds: o ganic dus ; occupa ional exposu e; swine; cul u e-based me hods; molecula ools 1. In oduc ion Du ing he pas ew yea s, mos animal husband y p ac ices in Eu ope and he Uni ed S a es ha e been indus ialized, esul ing in li es ock ope a ions aising housands o animals in a single acili y [ 1 , 2 ]. The con ined p oduc ion o swine p omo es was e by-p oduc s and eed concen a ion, Toxics 2018,6, 5; doi:10.3390/ oxics6010005 www.mdpi.com/jou nal/ oxics Toxics 2018,6, 5 2 o 14 wi h he subsequen gene a ion o high le els o o ganic dus , including dus s, gases, mic oo ganisms, mic obial me aboli es and o he po en ial ai bo ne heal h haza ds [1]. The biobu den, comp ising ungi and bac e ia, should be conside ed among o ganic dus as an impo an ai bo ne pollu an in swine p oduc ion acili ies [ 3 ]. O ganic dus in swine p oduc ion may be gene a ed by a ious mic obial g ow h subs a es, such as eeding ma e ials [ 4 , 5 ], moldy hay, eces [ 6 – 9 ], manu e and bedding ma e ial [ 10 ], en ila ion, il a ion and mis ing applica ions [ 11 ]. This di e si y o con amina ion sou ces wo sens he occupa ional p oblem and hinde s he e ec i eness o con ol measu es [ 8 ]. O no e, his occupa ional en i onmen may be associa ed wi h high ae osoliza ion o pa icula e ma e , boos ing exposu e o mic oo ganisms such as ungi [ 8 , 9 ] and o he myco oxins p oduced by hem [8,10]. Swine wo ke s ha e an inc eased p e alence o se e al espi a o y symp oms and diseases, such as ch onic b onchi is, ch onic obs uc i e pulmona y disease and o ganic dus oxic synd ome [ 9 , 12 – 16 ]. In addi ion, wo k p ac ices such as he ypes and me hods o swine eeding, he use o wood sha ings o animal bedding and he use o some speci ic disin ec an s ha e also been ela ed o wo ke s’ symp oms [ 8 , 10 , 17 – 19 ]. When eeding occu s, abundan o ganic dus pa icles om eed become ae osolized o a long ime o deposi ed on he loo [ 10 ]. High dispe sion o mic oo ganisms and hei me aboli es also occu s as hey a e esuspended du ing he eeding ask [20,21]. Al hough he need o manual wo k in swine p oduc ion is dec easing, mos ly due o he con inemen o swine p oduc ion o acili ies wi h au oma ion, in Po ugal, as in o he coun ies, some ac i i ies s ill equi e a me s’ di ec in e en ion, such as pigle ail cu and accina ion, among o he ac i i ies ela ed o swine b eeding [ 8 ]. In hese ac i i ies, inc eased occupa ional exposu e o o ganic dus occu s [ 5 , 8 , 9 ], as p e iously epo ed o ungal con amina ion [ 5 ], myco oxins [ 8 ] and pa icula e ma e [ 9 ] in Po uguese swine a ms. In his s udy, a di e en p o ocol o he assessmen o exposu e o o ganic dus is desc ibed, implemen ed o un eil a mo e accu a e occupa ional exposu e scena io ega ding he o ganic dus isk ac o . In ac , besides he assessmen o pa icles, bac e ia and ungi in ai and su ace samples, molecula ools we e ex ensi ely applied o a ge ha m ul ungal species, and eed samples we e also sc eened. 2. Ma e ials and Me hods 2.1. Swine Fa ms’ Cha ac e is ics and Collec ion o En i onmen al Samples Fi e Po uguese swine a ms we e p ospec ed in he Lisbon dis ic be ween June and July o 2017. Swine a ms we e selec ed acco ding o he ollowing c i e ia: loca ion wi hin Lisbon dis ic , highes numbe o animals (Table 1) and numbe o wo ke s. The i e a ms had been e alua ed, among o he s, in an ea lie s udy om ou g oup [ 5 ], wi h speci ic ecommenda ions on he need o implemen sa e y measu es in he con ex o en i onmen al and occupa ional heal h. Howe e , no modi ica ions in wo king ac i i ies o sa e y p ocedu es we e obse ed. In he swine a m A, ma e ni y was he main ac i i y, despi e ha ing o he a eas wi h on-going ac i i ies. The loo in he swine ma e ni ies was co e ed wi h jou nal pape . Manu e emo al sys ems we e p esen in all a m acili ies, wi h comple e emo al om he building se e al imes a day. The en ila ion sys ems in he s udied a m buildings consis ed o mechanical en ila ion by wall exhaus ans coupled wi h na u al en ila ion h ough he ope a ion o a winch-cu ain. Fa m wo ke s did no use espi a o y p o ec ion de ices in any o he a ms. Twen y o wen y- i e ai samples om i e dis inc a eas (pig ges a ion si e, ma e ni y, s alls, pig a ening a ea and qua an ine con inemen ) (Table 1) and one ou doo ai sample ( o be used as a e e ence) we e aken in each swine a m. Samples om su aces (wall swab), loo co e age om ma e ni ies (jou nal pape ) and eed om pig ges a ion si e and/o pig a ening we e also collec ed a each a m (Table 1). Toxics 2018,6, 5 3 o 14 Table 1. Numbe o samples collec ed and animal quan i y in each a m. Swine Fa ms No. o Ai Samples Impac ion * No. o Ai Samples Impinge No. o Su aces Samples (Walls) No. o Feed Samples No. o Floo Co e Samples Animal Quan i y A 20 5 5 2 1 1768 B 20 # 5 5 2 1 8000 C 20 4 # 5 2 1 3300 D 20 5 5 2 1 6000 E 16 + 4 4 2 1 7000 * A each wo king si e, 4 ai samples we e aken o each media (mal ex ac aga (MEA), dichlo an glyce ol (DG18), yp ic soy aga (TSA), Viole Red bile aga (VRBA)); +: a m wi hou pig a ening; #: one sample los . 2.2. Pa icula e Ma e Assessmen Pa icle measu emen was pe o med wi h di ec - eading equipmen (Handheld Pa icle Coun e om Ligh house Wo ldwide Solu ions (Model 3016/5016)). This measu emen equipmen gi es in o ma ion ega ding pa icle mass concen a ion (mg × m −3 ) (PMC) a i e di e en sizes (PM0.5, PM1, PM2.5, PM5, PM10). Pa icle coun s (PNC) by pa icle diame e size we e also ob ained wi h he same equipmen o six di e en diame e s (0.3 µ m, 0.5 µ m, 1 µ m, 2.5 µ m, 5 µ m and 10 µ m). Measu emen s we e pe o med nea he nose o each wo ke unde ypical en ila ion condi ions and du ing he de elopmen o asks pe o med in each o he i e a eas o in e es . 2.3. Biobu den Sampling and Analysis by Cul u e-Based Me hods Ai samples (50 L) we e collec ed a a 1-m heigh on o ou media: 2% mal ex ac aga (MEA) wi h 0.05 g/L chlo amphenicol media; dichlo an glyce ol (DG18) aga -based media; yp ic soy aga (TSA) wi h 0.2% nys a in; Viole Red bile aga (VRBA), wi h a low a e o 140 L o ai pe minu e. Fo su ace samples, he walls o he conside ed indoo loca ions we e swabbed using a 10 by 10 cm squa e s encil, disin ec ed wi h 70% alcohol solu ion be ween samples, acco ding o he In e na ional S anda d ISO 18593-2004, and swabs we e pla ed on o he selec ed media. Fo loo co e age and eed samples, 4.4 g o each (no o en-d ied p io o p ocessing, hus e aining na u al wa e con en ) we e washed in 40 mL o s e ilized dis illed wa e o 20 min a 200 pm, and 0.15 mL o his suspension we e sp ead on o he ou media. A e incuba ion o MEA and DG18 a 27 ◦ C o 5 o 7 days o ungi and TSA and VRBA a 30 ◦ C and 35 ◦ C o 7 days o mesophilic bac e ia and coli o ms (G am-nega i e bac e ia), espec i ely, biobu den densi ies (colony- o ming uni s, CFU · m −3 , CFU · m −2 , CFU · g −1 ) we e calcula ed. Fungal species we e iden i ied mic oscopically using ease moun o Sco ch ape moun and lac ophenol co on blue moun p ocedu es. Mo phological iden i ica ion was achie ed h ough mac o- and mic oscopic cha ac e is ics [22]. 2.4. Fungal Sampling and Molecula De ec ion by Real-Time PCR Ai samples (300 L) we e collec ed using he impinge Co iolis µ ai sample (Be in Technologies) wi h a low a e o 300 L o ai pe minu e. Samples we e collec ed on o 10-mL s e ile phospha e-bu e ed saline (PBS) wi h 0.05% T i on X-100, and an aliquo was subsequen ly used o DNA ex ac ion using he ZR Fungal/Bac e ial DNA MiniP ep Ki (Zymo Resea ch, I ine, CA, USA) acco ding o he manu ac u e ’s ins uc ions. Aliquo s o su ace samples, loo co e age (jou nal pape ) and eed samples we e also used o DNA ex ac ion ollowing he same s eps as he impinge samples (Table 1). A Viia7 qPCR De ec ion Sys em (The mo Fishe Scien i ic, Wal ham, MA, USA) was used o pe o m eal- ime PCR (RT-PCR) o he molecula de ec ion o Aspe gillus species/s ains (Table 2). Reac ions included 1 × iQ Supe mix (Bio-Rad, Amado a, Po ugal), 0.5 µ M o each p ime and 0.375 µ M o TaqMan p obe in a o al olume o 20 µ L. Ampli ica ion ollowed a h ee-s ep PCR: 40 cycles wi h dena u a ion a 95 ◦ C o 30 s, annealing a 52 ◦ C o 30 s and ex ension a 72 ◦ C o 30 s. A non- empla e Toxics 2018,6, 5 4 o 14 con ol was used in e e y PCR eac ion. As posi i e con ols o he species, DNA samples we e ob ained om e e ence s ains om he Mycology Labo a o y om he Na ional Ins i u e o Heal h Dou o Rica do Jo ge (INSA). Table 2. Sequence o p ime s and TaqMan p obes used o eal- ime PCR. Aspe gillus Sec ions Ta ge ed Sequences Re e ence Fumiga i Fo wa d P ime 50-CGCGTCCGGTCCTCG-30 Re e se P ime 50-TTAGAAAAATAAAGTTGGGTGTCGG-30C uz-Pe ez e al. 2001 [23] P obe 50-TGTCACCTGCTCTGTAGGCCCG-30 Ve sicolo es Fo wa d P ime 50-CGGCGGGGAGCCCT-30 Re e se P ime 50-CCATTGTTGAAAGTTTTGACTGATcTTA-30 P obe 5 0 -AGACTGCATCACTCTCAGGCATGAAGTTCAG-3 0EPA 2017 [24] 2.5. S a is ical Analysis The so wa e SPSS, Ve sion 24.0 o Windows, IBM, Lisbon, Po ugal, 2016 was used o s a is ical analysis. The esul s we e conside ed signi ican a he 5% signi icance le el. The Shapi o–Wilk es was used o es da a no mali y. F equency analysis (n, %) was used o he quali a i e da a. Minimum, maximum, median and in e qua ile ange we e de e mined in he quan i a i e da a, since no mali y was no e i ied. The F iedman es was used o compa e he pa icle concen a ion o di e en dimensions (ei he he coun s o he mass), since he no mali y assump ion was no e i ied. The K uskal–Wallis es was used o compa e he concen a ion o pa icles o di e en sizes (ei he coun s o mass) be ween he i e swine a ms s udied and be ween wo kplaces, since he no mali y assump ion was no e i ied. The K uskal–Wallis es was used o compa e ungi and bac e ia concen a ion, bo h in ai and on su aces, be ween he i e swine a ms s udied and be ween wo kplaces, since he no mali y assump ion was no e i ied. Spea man’s co ela ion coe icien was used o s udy he ela ionship be ween ungi and bac e ia concen a ion (ai and su ace) and pa icle concen a ion (coun s and mass). 3. Resul s 3.1. Pa icula e Ma e S a is ically-signi ican di e ences we e de ec ed o pa icle coun s (PNC) o di e en dimensions and o pa icle mass concen a ions (PMC) o a ious dimensions ( χ2 F (4) = 120.000, p= 0.000) (F iedman’s mul iple compa isons). Smalle pa icles (0.3 µ m) we e he ones wi h signi ican ly highe coun s and lowe mass concen a ions. Smalle pa icle (0.3 µ m) coun s we e signi ican ly di e en among swine a ms ( χ2 K−W (4) = 27.832, p= 0.000), pa icula ly among Swine Fa m D and Swine Fa ms A (p= 0.002) and E (p= 0.000), and among Swine Fa ms C and E (p= 0.005). Rega ding he coun s o 0.5 µ m-sized pa icles, signi ican di e ences we e also ound be ween swine a ms ( χ2 K−W (4) = 25.353, p= 0.000), namely among Swine Fa m D and Swine Fa ms A (p= 0.000) and E (p= 0.000). Th ough mean anks analysis, i was ound ha 0.3 µ m- and 0.5 µ m-sized pa icle coun s we e signi ican ly highe in Swine Fa m E, ollowed by Swine Fa m A. No s a is ically-signi ican di e ences (p> 0.05) we e de ec ed o pa icle coun s o he emaining sizes. Rega ding pa icle mass esul s, he PM0.5 concen a ion was signi ican ly di e en be ween swine a ms ( χ2 K−W (4) = 27.832, p= 0.000), namely among Swine Fa ms D and Swine Fa ms A ( p= 0.002 ) and E (p= 0.000) and among Swine Fa ms C and E (p= 0.005). Highe mass concen a ions o PM0.5 pa icles we e obse ed in Swine Fa ms E and A. The di e ences ound o PM1.0 concen a ions be ween swine a ms ( χ2 K−W (4) = 27.389, p= 0.000) ollowed he same end, i.e., Swine Fa m D di e ed Toxics 2018,6, 5 5 o 14 signi ican ly om Swine Fa ms A (p= 0.001) and E (p= 0.000) and C di e ed om E ( p= 0.008 ). Swine Fa ms E and A also exhibi ed he highes mass concen a ions o PM1.0 pa icles. As o he mass concen a ion o he pa icles PM2.5, s a is ically-signi ican di e ences we e also de ec ed among he swine a ms ( χ2 K−W (4) = 13.159, p= 0.011), namely among Swine Fa ms D and A (p= 0.007), wi h he highes concen a ions obse ed in Swine Fa m A. No s a is ically-signi ican di e ences (p> 0.05) we e de ec ed o mass concen a ion o he emaining pa icles. Highe alues o pa icle coun s and pa icle mass concen a ion we e ound indoo , in compa ison o ou doo ai samples. No s a is ically-signi ican di e ences we e ound among he wo k a eas (ma e ni y, ges a ion, ba e ies, a ening and qua an ine), nei he in coun s no on mass concen a ions o he di e en pa icle sizes (p> 0.05). 3.2. Biobu den: Bac e ial Con amina ion Resul s o o al bac e ial load in indoo ai anged om 1800 o 54,840 CFU · m −3 , wi h Swine Fa m E p esen ing he highes median alue (28,210 CFU · m −3 ) Swine Fa ms A, B, C and D wi h mean alues o 18,688, 13,660, 11,944 and 14,720 CFU · m −3 , espec i ely. In 35.7% (10 ou o 28) o he sampling si es, he o al bac e ial load exceeded he limi alues al eady sugges ed [ 25 ] (10,000 CFU · m −3 o o al bac e ia o eigh hou s o wo k o ag icul u al en i onmen s). G am-nega i e bac e ia load in he ai anged be ween 0 and 220 CFU · m −3 wi h he highes median alue (72 CFU · m −3 ) ound in Swine Fa ms A and B, ollowed by Swine Fa ms C, D and E, wi h mean alues o 4, 24 and 60 CFU · m −3 , espec i ely. None o he sampled si es exceeded he limi alues sugges ed by Goye [ 25 ] (1000 CFU · m −3 o G am-nega i e bac e ia o eigh hou s o wo k o ag icul u al en i onmen s) (Figu e 1). The bac e ial load in ou doo ai was lowe han in indoo ai in each swine a m, excep o he “ a ening” a ea in Fa m C, p esen ing a sligh ly lowe alue indoo s han he o al bac e ial load ou doo s. Toxics2018,6,5 5o 14  (p=0.008).SwineFa msEandAalsoexhibi ed hehighes massconcen a ionso PM1.0pa icles. As o  hemassconcen a iono  hepa iclesPM2.5,s a is ically‐signi ican di e enceswe ealso de ec edamong heswine a ms(߯௄ିௐ ଶ(4)=13.159,p=0.011),namelyamongSwineFa msDandA (p=0.007),wi h hehighes concen a ionsobse edinSwineFa mA.Nos a is ically‐signi ican  di e ences(p>0.05)we ede ec ed o massconcen a iono  he emainingpa icles. Highe  alueso pa iclecoun sandpa iclemassconcen a ionwe e oundindoo ,in compa ison oou doo ai samples.Nos a is ically‐signi ican di e enceswe e oundamong he wo ka eas(ma e ni y,ges a ion,ba e ies, a eningandqua an ine),nei he incoun sno onmass concen a ionso  hedi e en pa iclesizes(p>0.05). 3.2.Biobu den:Bac e ialCon amina ion Resul s o  o albac e ialloadinindoo ai  anged om1800 o54,840CFU∙m −3 ,wi hSwine Fa mEp esen ing hehighes median alue(28,210CFU∙m −3 )SwineFa msA,B,CandDwi hmean alueso 18,688,13,660,11,944and14,720CFU∙m −3 , espec i ely.In35.7%(10ou o 28)o  he samplingsi es, he o albac e ialloadexceeded helimi  aluesal eadysugges ed[25] (10,000CFU∙m −3  o  o albac e ia o eigh hou so wo k o ag icul u alen i onmen s). G am‐nega i ebac e ialoadin heai  angedbe ween0and220CFU∙m −3 wi h hehighes median alue(72CFU∙m −3 ) oundinSwineFa msAandB, ollowedbySwineFa msC,DandE,wi hmean alueso 4,24and60CFU∙m −3 , espec i ely.Noneo  hesampledsi esexceeded helimi  alues sugges edbyGoye [25](1000CFU∙m −3  o G am‐nega i ebac e ia o eigh hou so wo k o  ag icul u alen i onmen s)(Figu e1).Thebac e ialloadinou doo ai waslowe  haninindoo ai  ineachswine a m,excep  o  he“ a ening”a eainFa mC,p esen ingasligh lylowe  alue indoo s han he o albac e ialloadou doo s.  Figu e1.Bac e ialloadob ained o ai andsu acesamples. Rega dingsu acesamples,mesophilicbac e ialpopula ion anged om3×10 4  o 516×10 4 CFU∙m −2 ,wi h hehighes median alue oundinSwineFa mC(192×10 4 CFU∙m −2 ), ollowedbySwineFa msA,B,DandEwi hmean alueso 93.4×10 4 ,182.4×10 4 ,128.4×10 4 and 162×10 4 CFU∙m −2 , espec i ely.G am‐nega i ebac e ia angedbe ween0and104×10 4 CFU∙m −2 , Figu e 1. Bac e ial load ob ained o ai and su ace samples. Rega ding su ace samples, mesophilic bac e ial popula ion anged om 3 × 10 4 o 516 × 10 4 CFU · m −2 , wi h he highes median alue ound in Swine Fa m C (192 × 10 4 CFU · m −2 ), ollowed by Swine Fa ms A, B, D and E wi h mean alues o 93.4 × 10 4 , 182.4 × 10 4 , 128.4 × 10 4 and 162 × 10 4 CFU · m −2 , espec i ely. G am-nega i e bac e ia anged be ween 0 and 104 × 10 4 CFU · m −2 , Toxics 2018,6, 5 6 o 14 wi h only one sampling si e in each swine a m p esen ing coli o m g ow h (G am-nega i e bac e ia), mainly in he “qua an ine” and in he “ma e ni y” a ea in Swine Fa m B. Feed bac e ial con amina ion anged be ween 81 and 1237 CFU · g −1 , and he “ma e ni y” om Swine Fa m E showed coun less mesophilic bac e ia. G am-nega i e bac e ia concen a ions anged om 0 o 363 CFU · g −1 , and he same sampling si e showed coun less colonies, whe eas in he wo o he sampling si es, isola es we e no obse ed (“s alls” in Swine Fa m A and “qua an ine” in Swine Fa m C). Conce ning bedding esul s, coun less mesophilic bac e ia we e p esen in all swine a ms, excep in Fa m D wi h a alue o 1529 CFU · g −1 . Rega ding G am-nega i e bac e ia, swine Fa m E p esen ed coun less coli o ms, and he o he swine a ms anged be ween 489 and 1552 CFU·g−1. These esul s sugges a highe con ibu ion o G am-posi i e han G am-nega i e bac e ia in he bac e io a load. 3.3. Biobu den: Fungal Con amina ion Fungal load in indoo ai anged om 40 o 3120 CFU · m −3 on MEA, wi h Swine Fa m E p esen ing he highes median alue (2500 CFU · m −3 ), ollowed by Swine Fa ms A, B, C and D wi h mean alues o 124, 140, 604 and 104 CFU · m −3 , espec i ely. No ewo hy, 65.5% (19 ou o 29) o he sampling si es showed highe ungal load han he limi s imposed by he Wo ld Heal h O ganiza ion (WHO) (maximum alue o 150 CFU · m −3 ) (2). Two ou o he 29 (6.9%) ai samples collec ed in he i e swine a ms p esen ed highe ungal load when compa ed o he ou doo sampling. Swine Fa m A was he excep ion, wi h highe ungal load indoo s ( wo ou o i e collec ed samples). Simila esul s we e ound on DG18, wi h ungal load anging om 80 o 3400 CFU · m −3 and wi h Swine Fa m E also p esen ing he highes median alue (2680 CFU · m −3 ), ollowed by Swine Fa ms A, B, C and D, wi h mean alues o 160, 400, 604 and 356 CFU · m −3 , espec i ely. DG18 e ealed an inc eased amoun o sampling si es (82.8%; 24 ou o 29) wi h ungal load exceeding he WHO limi s (maximum alue o 150 CFU · m −3 ) (Figu e 2). Nine ou o he 29 (31%) ai samples p esen ed highe indoo ungal load when compa ed o he ou doo sampling. Swine Fa m B was he one wi h he highes numbe o ai samples p esen ing highe load indoo s ( ou ou o i e collec ed samples). Toxics2018,6,5 6o 14  wi honlyonesamplingsi eineachswine a mp esen ingcoli o mg ow h(G am‐nega i e bac e ia),mainlyin he“qua an ine”andin he“ma e ni y”a eainSwineFa mB. Feedbac e ialcon amina ion angedbe ween81and1237CFU∙g −1 ,and he“ma e ni y” om SwineFa mEshowedcoun lessmesophilicbac e ia.G am‐nega i ebac e iaconcen a ions anged om0 o363CFU∙g −1 ,and hesamesamplingsi eshowedcoun lesscolonies,whe easin he wo o he samplingsi es,isola eswe eno obse ed(“s alls”inSwineFa mAand“qua an ine”in SwineFa mC). Conce ningbedding esul s,coun lessmesophilicbac e iawe ep esen inallswine a ms, excep inFa mDwi ha alueo 1529CFU∙g −1 .Rega dingG am‐nega i ebac e ia,swineFa mE p esen edcoun lesscoli o ms,and heo he swine a ms angedbe ween489and1552CFU∙g −1 . These esul ssugges ahighe con ibu iono G am‐posi i e hanG am‐nega i ebac e iain hebac e io aload. 3.3.Biobu den:FungalCon amina ion Fungalloadinindoo ai  anged om40 o3120CFU∙m −3 onMEA,wi hSwineFa mE p esen ing hehighes median alue(2500CFU∙m −3 ), ollowedbySwineFa msA,B,CandDwi h mean alueso 124,140,604and104CFU∙m −3 , espec i ely.No ewo hy,65.5%(19ou o 29)o  he samplingsi esshowedhighe  ungalload han helimi simposedby heWo ldHeal h O ganiza ion(WHO)(maximum alueo 150CFU∙m −3 )(2).Twoou o  he29(6.9%)ai samples collec edin he i eswine a msp esen edhighe  ungalloadwhencompa ed o heou doo  sampling.Swine Fa mAwas heexcep ion,wi hhighe  ungalloadindoo s( woou o  i ecollec edsamples). Simila  esul swe e oundonDG18,wi h ungalload anging om80 o3400CFU∙m −3 and wi hSwineFa mEalsop esen ing hehighes median alue(2680CFU∙m −3 ), ollowedbySwine Fa msA,B,CandD,wi hmean alueso 160,400,604and356CFU∙m −3 , espec i ely.DG18 e ealedaninc easedamoun o samplingsi es(82.8%;24ou o 29)wi h ungalloadexceeding he WHOlimi s(maximum alueo 150CFU∙m −3 )(Figu e2).Nineou o  he29(31%)ai samples p esen edhighe indoo  ungalloadwhencompa ed o heou doo sampling.SwineFa mBwas heonewi h hehighes numbe o ai samplesp esen inghighe loadindoo s( ou ou o  i e collec edsamples).  Figu e2.Fungalloaddis ibu ionin he i eassessedswine a ms.Thedashedline ep esen s he e e encelimi ssugges edby heWo ldHeal hO ganiza ion(WHO). Twen y i edi e en  ungalspecieswe e oundinai samplesonMEAand18onDG18. Cladospo iumsp.was hemos p e alen inindoo ai samplesinbo hmedia(59.4%MEA;66.5% DG18), ollowedbyFusa iumg aminea umspeciescomplex(13.2%)onMEAandUlocladiumsp. (14.6%)onDG18(Table3). Figu e 2. Fungal load dis ibu ion in he i e assessed swine a ms. The dashed line ep esen s he e e ence limi s sugges ed by he Wo ld Heal h O ganiza ion (WHO). Twen y i e di e en ungal species we e ound in ai samples on MEA and 18 on DG18. Cladospo ium sp. was he mos p e alen in indoo ai samples in bo h media (59.4% MEA; 66.5% DG18), ollowed by Fusa ium g aminea um species complex (13.2%) on MEA and Ulocladium sp. (14.6%) on DG18 (Table 3). Toxics 2018,6, 5 7 o 14 Fungal load in subs a e ( eed and loo co e age) and su ace samples was dis ibu ed as ollows: 0 o 4 CFU · g −1 (MEA) and 0 o 39 CFU · g −1 (DG18) in eed; 0 o 1 CFU · g −1 (MEA) and no isola es (DG18) in loo co e ing; 0 o 59 × 10 4 CFU · m −2 (MEA) (maximum alues o 14 × 10 4 CFU · m −2 in Swine Fa m D, ollowed by 6 × 10 4 CFU · m −2 in Swine Fa m E) and 0 o 370,500 (DG18) (maximum alues o 10 ×104CFU·m−2in Swine Fa m D) on wall su aces. Th ee di e en ungal species we e ound in eed samples on bo h media, he Cladospo ium genus being he mos p e alen (71.4% MEA; 82.2% DG18). Fungi isola es om loo co e ing samples we e only obse ed on MEA, wi h p edominan Penicillium genus (50.0%) ollowed by Al e na ia sp. (37.5%). Ten ungal species we e ound in su ace samples on MEA and ou on DG18. Cladospo ium sp. (53.8%) and Scopula iopsis b e icaulis (33.3%) we e he mos de ec ed species on MEA, whe eas Scopula iopsis candida (50.3%) and Aspe gillus sec ion Ci cumda i (19.9%) we e he mos obse ed on DG18. T icho hecium oseum was ound only on su aces (Table 3). Table 3. Fungal dis ibu ion in en i onmen al and subs a e ma ices a e inocula ion on o MEA and DG18 media. MEA DG18 Ai (CFU·m−3) (%; n)Ai (CFU·m−3) (%; n) Cladospo ium sp. 59.4; 12,100 Cladospo ium sp. 66.5; 14,120 Fusa ium g aminea um 13.2; 2700 Ulocladium sp. 14.6; 3100 Al e na ia sp. 5.7; 1160 Ch ysonilia si ophila 4.7; 1000 O he s 21.7; 4420 O he s 14.2; 3020 Su aces (CFU·m−2) (%; n)Su aces (CFU·m−2) (%; n) Cladospo ium sp. 53.8; 210,000 Scopula iopsis candida 50.3; 580,000 Scopula iopsis b e icaulis 33.3; 130,000 Aspe gillus sec ion Ci cumda i 19.9; 230,000 Penicillium sp. 12.8; 50,000 Cladospo ium sp. 13; 150,000 O he s 0.1; 500 O he s 16.7; 193,000 Feed (CFU·g−1) (%; n)Feed (CFU·g−1) (%; n) Cladospo ium sp. 71.4; 10 Cladospo ium sp. 82.2; 37 Penicillium sp. 21.4; 3 Penicillium sp. 8.9; 4 Fusa ium culmo um 7.1; 1 Fusa ium culmo um 8.9; 4 Floo co e ing (CFU·g−1) (%; n)Floo co e ing (CFU·g−1) (%; n) Penicillium sp. 50; 4 -- Al e na ia sp. 37.5; 3 -- Cladospo ium sp. 12.5; 1 -- Aspe gillus gene a we e obse ed on MEA and DG18 (2.26%). Aspe gillus sec ion Ci cumda i was he mos p e alen (55%) on MEA ollowed by Aspe gilli (25%). Di e en Aspe gillus sec ions we e mo e p e alen on DG18, Ve sicolo es being he mos iden i ied (50%) ollowed by Us i (20.8%) (Table 4). Table 4. Aspe gillus sec ions’ dis ibu ion in ai samples. MEA DG18 Ai (CFU·m−3) (%; n)Ai (CFU·m−3) (%; n) Ci cumda i 55; 220 Ve sicolo es 50; 240 Aspe gilli 25; 100 Us i 20.8; 100 Nig i 10; 40 Aspe gilli 12.5; 60 Ve sicolo es 5; 20 Candidi 12.5; 60 Fla i 5; 20 Nidulan es 4.2; 20 No Aspe gillus sec ion Fumiga i no Aspe gillus sec ion Ve sicolo es we e de ec ed by qPCR. Toxics 2018,6, 5 8 o 14 3.4. Co ela ion and Compa ison Resul s No s a is ically-signi ican di e ences we e de ec ed be ween he swine a ms in ei he su ace o ai samples, ega ding o al bac e ial load ( χ2 K−W (4) = 1.936, p= 0.748, χ2 K−W (4) = 3.676, p= 0.452, espec i ely) o G am-nega i e bac e ia (χ2 K−W(4) = 0.081, p= 0.999, χ2 K−W(4) = 7.132, p= 0.129, espec i ely). S a is ically-signi ican di e ences we e ound be ween he ungal load on MEA in he swine a ms in su ace ( χ2 K−W (4) = 13.699, p= 0.008) and in ai ( χ2 K−W (4) = 14.602, p= 0.001) samples. Fungal load was signi ican ly di e en among Swine Fa ms C and D (p= 0.036) in su ace samples and among Swine Fa m D and Swine Fa ms A (p= 0.007) and B (p= 0.025) in ai samples. Swine Fa m D p esen ed he highes mean ank alues o ungal load in bo h su ace and ai samples. S a is ically-signi ican di e ences we e also obse ed o ungal load on DG18 be ween he ai samples collec ed in he swine a ms ( χ2 K−W (4) = 12.621, p= 0.013), namely, be ween Swine Fa ms B and D (p= 0.005), wi h Fa ms D and B exhibi ing he highes and he lowes alues, espec i ely (Table 5). Table 5. Resul s o he K uskal–Wallis es o he compa ison o ungi and bac e ia concen a ion, bo h on su aces and in ai , be ween he i e swine a ms (n= 24). Bac e ia/Fungus Swine Fa ming nRanks Tes S a is ics aK uskal–Wallis Mul iple Compa isons Mean Rank Chi-Squa e d p To al Bac e ia Su ace (CFU·m−2) A 5 12.00 1.936 40.748 B 5 10.00 C 5 14.10 D 4 10.75 E 5 15.30 G am Nega i e Bac e ia-Su ace (CFU·m−2) A 5 12.40 0.081 40.999 B 5 12.60 C 5 12.00 D 4 12.75 E 5 12.80 Fungi (MEA)-Su ace (CFU·m−2) A 5 17.90 13.699 40.008 * C6=D (p= 0.036) B 5 12.50 C 5 6.00 D 4 19.50 E 5 8.00 Fungi (DG18)-Su ace (CFU·m−2) A 5 18.60 8.430 40.077 B 5 13.60 C 5 10.30 D 4 8.50 E 5 10.70 To al bac e ia-Ai (CFU·m−3) A 5 12.10 3.676 40.452 B 5 14.10 C 5 10.40 D 4 17.50 E 5 9.40 G am Nega i e Bac e ia-Ai (CFU·m−3) A 5 10.20 7.132 40.129 B 5 15.00 C 5 15.00 D 4 16.50 E 5 6.60 Fungi (MEA)-Ai (CFU·m−3) A 5 6.40 17.602 40.001 * A6=D (p= 0.007) B 5 8.20 B 6=D (p= 0.025) C 5 9.40 D 4 22.50 E 5 18.00 Fungi (DG18)-Ai (CFU·m−3) A 5 11.60 12.621 40.013 * B6=D (p= 0.005) B 5 6.10 C 5 10.80 D 4 22.50 E 5 13.50 aK uskal–Wallis es ; * s a is ically-signi ican di e ences a he 5% signi icance le el. Toxics 2018,6, 5 9 o 14 Among he sampling si es (ma e ni y, ges a ion, ba e y, a ening and qua an ine), no s a is ically-signi ican di e ences we e de ec ed in ungi no in bac e ia load o ei he su ace o ai samples (p’s > 0.05). The ela ionship be ween ungi, bac e ia (su ace and ai ) and pa icles’ concen a ion (coun ing and mass) was pe o med sepa a ely o each swine a m. In Swine Fa m A, a signi ican nega i e co ela ion, wi h s ong in ensi y, was ound be ween he ungal concen a ion on DG18 in su ace samples and he ungal concen a ion on MEA in he ai ( s = − 0.975, p= 0.005). This esul indica es ha highe ungal concen a ions on DG18 in he su ace a e ela ed o lowe ungal concen a ions on MEA in he ai . In Swine Fa m B, highe o al bac e ia loads in su ace samples we e ound o be signi ican ly co ela ed wi h lowe coun s o 0.5 mic on ( s = − 0.9, p= 0.037), 1.0 mic on ( s = − 0.9, p= 0.037), 2.5 mic on ( s = − 0.9, p= 0.037), 5.0 mic on ( s = − 0.9, p= 0.037) and 10.0 mic on ( s = − 0.9, p= 0.037) pa icles. Addi ionally, highe ungal loads on DG18 in su ace samples we e ound o be posi i ely co ela ed wi h highe ungal loads on MEA in su ace samples ( s = 0.918, p= 0.028 ) and wi h highe o al bac e ia loads in he ai ( s = 0.894, p= 0.041). A s a is ically-signi ican co ela ion was also ound in Swine Fa m B be ween highe PM0.5 mass concen a ion and highe ungal loads on MEA in he ai ( s = 0.9, p= 0.037). In Swine Fa m C, o al bac e ia load in he ai was ound o be nega i ely co ela ed wi h ungal load on DG18 in he ai ( s = − 0.9, p= 0.037), sugges ing ha highe concen a ions o o al bac e ia in he ai a e ela ed o lowe concen a ions o ungi in he ai . No signi ican co ela ions we e ound o Swine Fa ms D and E. 4. Discussion O ganic dus has been he ocus o se e al epidemiological s udies, as exposu e o o ganic dus is desc ibed as causing mucous memb ane i i a ion in he eyes and uppe and lowe ai ways, in lamma ion by alle gic and non-alle gic mechanisms and ai way obs uc ion [ 26 ]. O ganic dus in swine p oduc ion can be gene a ed om se e al sou ces such as eed, skin cells, hai and d ied eces. Acu e exposu es o high le els o dus may p o oke inc eased phlegm p oduc ion and pulmona y in lamma ion 4 − 10 h a e exposu e. In addi ion, ch onic exposu es may esul in b onchi is and as hma [27]. I has al eady been epo ed ha ae osols p oduced in swine p oduc ion acili ies may ha e pa icle size dis ibu ions conside ably smalle han 1.0 µ m in diame e [ 28 , 29 ], comp ising an impo an size ac ion o he espi able ae osols ha should be assessed in swine p oduc ion [ 28 , 29 ]. Fu he mo e, pa icula e ma e can be he ehicle o he biobu den p esen in swine a ms, eaching wo ke s espi a o y sys ems and, consequen ly, enhancing he occupa ional exposu e o swine wo ke s o o ganic dus [ 9 ]. The e o e, mic oo ganisms adhe ing o pa icula e ma e , along wi h coexis ing gases and oxins a e a gene al cause o conce n ega ding co-exposu e o se e al isk ac o s and possible addi i e and syne gis ic heal h e ec s [ 20 , 30 ]. This scena io ein o ces he heo y ha occupa ional exposu e is a ely associa ed wi h a single ac o , since i is commonly a combina ion o se e al isk ac o s [31]. I has been sugges ed ha s a iona y sample s can be adequa ely used in he assessmen o pe sonal exposu e o ai bo ne ungi in con ined ag icul u al en i onmen s [ 7 ] as is he case o animal p oduc ion. Howe e , we should combine no only mo e han one ac i e me hod o ai sampling [32,33] , bu also couple i wi h passi e me hods [ 34 , 35 ], such as su ace swabs, loo co e ing and eed o ob ain a mo e accu a e isk cha ac e iza ion [ 33 ]. As such, his s udy was pe o med using a mul i-app oach p o ocol o he cha ac e iza ion o occupa ional exposu e o o ganic dus by applying he impac ion me hod o ob ain a biobu den iable ac ion om o ganic dus , he impinge me hod o he de ec ion o ha m ul ungal species and he su ace swabbing o complemen ai samples’ esul s ega ding he di e si y o he biobu den [ 31 , 35 ]. In addi ion, he collec ion o loo co e ing and li e (besides su aces swabs) samples allowed de e mining he con amina ion le els om a la ge pe iod o ime (weeks o se e al mon hs), whe eas ai samples can only e lec he load om a sho e pe iod o ime (mos ly minu es) [33,34,36].