Bacterial communities found in placental tissues are associated with severe chorioamnionitis and adverse birth outcomes

RESEARCH ARTICLE Bac e ial communi ies ound in placen al issues a e associa ed wi h se e e cho ioamnioni is and ad e se bi h ou comes Ronan M. Doyle 1,2 *, Ka h yn Ha is 2 , S e e Kamiza 3 , Ulla Ha junmaa 4 , Ulla Asho n 5 , Minyanga Nkhoma 5 , Ka h yn G. Dewey 6 , Kenne h Male a 7 , Pe Asho n 5,8,9 , Nigel Klein 1 1UCL G ea O mond S ee Ins i u e o Child Heal h, Uni e si y College London, London, Uni ed Kingdom, 2Depa men o Mic obiology, Vi ology and In ec ion Con ol, G ea O mond S ee Hospi al NHS Founda ion T us , London, Uni ed Kingdom, 3Depa men o Pa hology, Uni e si y o Malawi College o Medicine, Blan y e, Malawi, 4Cen e o Child Heal h Resea ch, Uni e si y o Tampe e Facul y o Medicine and Li e Sciences, and Tampe e Uni e si y Hospi al, Tampe e, Finland, 5Depa men o In e na ional Heal h, Uni e si y o Tampe e School o Medicine, Tampe e, Finland, 6Depa men o Nu i ion, Uni e si y o Cali o nia Da is, Da is, Cali o nia, Uni ed S a es o Ame ica, 7Depa men o Communi y Heal h, Uni e si y o Malawi College o Medicine, Blan y e, Malawi, 8Depa men o Paedia ics, Uni e si y o Tampe e School o Medicine, Tampe e, Finland, 9Depa men o Paedia ics, Tampe e Uni e si y Hospi al, Tampe e, Finland *[email p o ec ed]c.uk Abs ac P e e m bi h is a majo cause o neona al mo ali y and mo bidi y wo ldwide. Bac e ial in ec- ion and he subsequen in lamma o y esponse a e ecognised as an impo an cause o p e e m bi h. I is hypo hesised ha hese o ganisms ascend he ce ical canal, colonise placen al issues, cause cho ioamnioni is and in se e e cases in ec amnio ic luid and he oe us. Howe e , he p esence o bac e ia wi hin he in au e ine ca i y does no always p e- cede cho ioamnioni is o p e e m bi h. Whe eas p e ious s udies obse ing he ypes o bac e ia p esen ha e been limi ed in size and he speci ici y o a ew p ede e mined o gan- isms, in his s udy we cha ac e ised bac e ia ound in placen al issues om a coho o 1391 women in u al Malawi using 16S ibosomal RNA gene sequencing. We ound ha speci ic bac e ia ound concu en ly on placen al issues associa e wi h cho ioamnioni is and deli - e y o a smalle newbo n. Se e e cho ioamnioni is was associa ed wi h a dis inc di e ence in communi y membe s, a highe bac e ial load and lowe species ichness. Fu he mo e, Snea hia sanguinengens and Pep os ep ococcus anae obius ound in bo h ma ched pa ic- ipan aginal and placen al samples we e associa ed wi h a lowe newbo n leng h- o -age Z-sco e. This is he la ges s udy o da e o examine he placen al mic obiome and i s impac o bi h ou comes. Ou esul s p o ide da a on he ole o he aginal mic obiome as a sou ce o placen al in ec ion as well as he possibili y o he apeu ic in e en ions agains a ge ed o ganisms du ing p egnancy. PLOS ONE | h ps://doi.o g/10.1371/jou nal.pone.0180167 July 12, 2017 1 / 23 a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 OPEN ACCESS Ci a ion: Doyle RM, Ha is K, Kamiza S, Ha junmaa U, Asho n U, Nkhoma M, e al. (2017) Bac e ial communi ies ound in placen al issues a e associa ed wi h se e e cho ioamnioni is and ad e se bi h ou comes. PLoS ONE 12(7): e0180167. h ps://doi.o g/10.1371/jou nal. pone.0180167 Edi o : Je e son Te y, BC Child en’s Hospi al, CANADA Recei ed: Feb ua y 22, 2016 Accep ed: June 12, 2017 Published: July 12, 2017 Copy igh : ©2017 Doyle e al. This is an open access a icle dis ibu ed unde he e ms o he C ea i e Commons A ibu ion License, which pe mi s un es ic ed use, dis ibu ion, and ep oduc ion in any medium, p o ided he o iginal au ho and sou ce a e c edi ed. Da a A ailabili y S a emen : All sequence ead da a a e a ailable om he EMBl-EBI Eu opean Nucleo ide A chi e a s udy accession numbe PRJEB15035. Funding: This publica ion is based on esea ch unded by a g an o he Uni e si y o Cali o nia, Da is om he O ice o Heal h, In ec ious Diseases, and Nu i ion, Bu eau o Global Heal h, U.S. Agency o In e na ional De elopmen (USAID) unde e ms o Coope a i e Ag eemen In oduc ion P e e m bi h is he la ges cause o neona al dea hs in he wo ld [1]. Dea h a es a e pa icu- la ly high in de eloping coun ies. Compa ed wi h o he egions, sub-Saha an A ica has had consis en ly highe incidence o p e e m deli e ies [2] wi h ecen es ima es o be ween 10.0% [3] and 16.3% [4] o all bi hs in Malawi. E en in he de eloped wo ld, incidences o p e e m bi h a e inc easing and i emains he majo cause o pe ina al mo ali y [2]. O he child en who su i e ex eme p ema u i y (less han 26 weeks), 19% will de elop se e e dis- abili y [5]. While he ae iology o spon aneous p e e m labou emains elusi e, a ole o bac e ial in ec ion and colonisa ion o oe al issue and he associa ed ma e nal in lamma o y immune esponse is now ecognized as he p obable igge in some cases. Spon aneous p e e m bi h (SPTB) is dis inguished by a highe equency o bac e ial colonisa ion and by he na u e and di e si y o bac e ial species [6]. The e is inc easing e idence o sugges ha i is he ype o bac e ia p esen in p e e m deli e ies, a he han simply he p esence o bac e ia, ha di e s om e m deli e ies [7–9]. While he e a e hypo heses ha bac e ia can in ec oe al issues h ough haema ogenous sp ead, in he majo i y o cases he common ou e is hough o be ascending in ec ion om he agina in o he ce ical canal [10]. I appea s ha i is hese bac e- ia ha lead o in lamma ion o he cho ioamnio ic issues (cho ioamnioni is), well ecognised as being highly associa ed wi h ea lie deli e y [10,11]. Mos s udies o da e ha e been conduc ed in Eu ope and No h Ame ica. In his s udy we aimed o desc ibe he mic obio a ound in placen al issue and e al memb anes (cho ion and amnion) om a coho o women in u al sou he n Malawi. We examined he mic obial com- muni y s uc u e and explo ed i his di e ed in placen al issue exhibi ing cho ioamnioni is. We also de e mined he ela ionship o he placen al mic obiome wi h bi h weigh , newbo n leng h, du a ion o p egnancy and head ci cum e ence. We also collec ed o al and aginal swabs om he same indi iduals o s udy he po en ial sou ce o o ganisms ound in he mic obiomes o he placen a and e al memb anes. Ma e ials and me hods S udy design and en olmen This c oss-sec ional s udy was pa o a la ge clinical ial assessing whe he p o iding Lipid-based Nu ien Supplemen s (LNS) o mo he s du ing p egnancy and o 6 mon hs pos pa um, and o he in an om 6 o 18 mon hs o age, imp o es child g ow h o a g ea e ex en han p ena al i on and olic acid (IFA) o mul iple mic onu ien (MMN) able s [3]. Pa icipan s we e en olled p ospec i ely in o he main ial be o e 20 weeks ges a ion and ollowed h oughou p egnancy, childbi h and beyond. The main in e en ion ou comes in he ial we e bi h weigh and newbo n leng h. A e deli e y we s udied he associa ion be ween placen al and e al memb ane mic obio a, in lamma ion and bi h ou comes. All women en olled as pa o he main ial who deli e ed we e included in hese analyses and all analyses we e adjus ed o in e en ion g oup. Fu he me hodological de ails a ound s udy design, se ing, logis ics and da a managemen a e p o ided in he ull ial esul s [12– 14]. S udy se ing Rec ui men ook place in 4 cen es in Mangochi Dis ic , Sou he n Malawi: Mangochi dis ic hospi al, Malindi hospi al, Lungwena heal h cen e and Namwe a heal h cen e. Placen al mic obio a associa ed wi h bi h ou comes PLOS ONE | h ps://doi.o g/10.1371/jou nal.pone.0180167 July 12, 2017 2 / 23 No. AID-OAA-A-12-00005, h ough he Food and Nu i ion Technical Assis ance III P ojec (FANTA), managed by FHI 360. Addi ional suppo was p o ided by he Bill & Melinda Ga es Founda ion h ough a g an o he Uni e si y o Cali o nia Da is, he Academy o Finland (G an 252075), and he Compe i i e S a e Resea ch Financing o he Expe Responsibili y A ea o Tampe e Uni e si y Hospi al (G an 9S001). The indings and conclusions con ained wi hin he a icle a e hose o he au ho s and do no necessa ily e lec posi ions o policies o he Bill & Melinda Ga es Founda ion, USAID, he US go e nmen , o he o he unde s. The unde s had no ole in s udy design, da a collec ion and analysis, decision o publish, o p epa a ion o he manusc ip . Compe ing in e es s: The au ho s ha e decla ed ha no compe ing in e es s exis . Collec ion o bi h ou come and baseline da a A en olmen , pa icipan s’ weigh , heigh and haemoglobin concen a ion we e measu ed and obs e ic his o y was eco ded. Du a ion o p egnancy was measu ed using ul asound. All pa icipan s we e es ed o mala ial in ec ion and HIV (unless hey we e al eady known o be HIV posi i e o op ed ou ). A he i s home isi o pa icipan s 1–2 weeks pos -en ol- men , in o ma ion was ga he ed on demog aphic, social and economic backg ound. Bi h weigh was measu ed as soon as possible a e deli e y while newbo n leng h and head ci cum- e ence we e measu ed a he in an ’s i s clinic isi a 1–2 weeks old. Sample collec ion A e deli e y he placen a was ans e ed o a s e ile con aine in he hospi al, heal h cen e o home (whe e e deli e y ook place), whe e i was kep in a co e ed con aine a empe a u es anging be ween 20˚C-40˚C. Tissue sampling o he placen a occu ed immedia ely a e deli - e y, unless deli e y occu ed o e nigh (in which case he placen a was sampled he ollowing mo ning), o a home (in which case he con aine had o be anspo ed o he nea es s udy clinic i s ). Two 5 cm x 1 cm pieces o he e al (cho ionic and amnio ic) memb ane we e aken om he edge o he up u e si e and wo 0.5 cm x 0.5 cm pieces o placen al issue a ull hick- ness we e aken om nea he umbilical co d inse ion. One e al memb ane and one placen a sample we e placed in sepa a e c yo ials. I he sample collec ion ook place in Mangochi dis ic hospi al, he c yo ials we e placed a -80˚C. I sample collec ion ook place a an ou lying heal h cen e o Malindi hospi al, he samples we e s o ed a -20˚C o a maximum o wo days be o e being ans e ed o -80˚C s o age a Mangochi dis ic hospi al. Vaginal mucus samples we e aken a a pos na al isi (app ox. 1 week a e deli e y). Swabs we e inse ed app oxima ely 7 cm deep in o he pa icipan ’s agina, wi hou a isual con ol, and o a ed be o e wi hd awal. A e he sample collec ion, swabs we e s o ed a -80˚C. Den al swabs we e collec ed one week a e deli e y o as soon as possible by ubbing he gingi al ma gin o each oo h wi h he swab, a oiding skin con ac . Swabs we e placed in o a -20˚C eeze and as soon as possible mo ed he swab in o a -80˚C eeze . Timing o each s age o he collec ion p ocedu e was eco ded. His ologic examina ion The emaining placen a and e al memb ane samples no in c yo ials we e placed in 10% neu- al bu e ed o malin ixa i e, p ocessed and embedded in pa a in wax. These we e sec ioned a 3–5 mic on hick and s ained wi h haema oxylin and eosin be o e being ead. All placen a and e al memb ane issue samples we e sco ed o his ologic cho ioamnioni is by SK who was blinded o all clinical in o ma ion and samples we e only iden i iable by s udy numbe . Cho - ioamnioni is was de ined as he p esence o in lamma ion (speci ically, he p esence o neu o- phils) in ei he he cho ion o he amnion ( e al memb anes) issues ha su ound he e us. Cho ioamnioni is can di e g ea ly in scale, so we de ined cho ioamnioni is as 5 neu ophil g anulocy es on a e age pe 10 high powe ields p esen in ei he he cho ionic pla e o he amnio ic memb ane, and se e e cho ioamnioni is was de ined as >25 neu ophil g anulo- cy es, which was adap ed om a p e ious s udy [15]. We included p ema u e up u e o mem- b anes as a cause o p e e m bi h and excluded o he causes such as p e e m labou (no known cause), an epa um haemo hage, hype ensi e diso de s and an incompe en ce ix. DNA ex ac ion DNA ex ac ion was ca ied ou o all sample ypes using he QIAmp DNA mini ki (Qiagen, Ge many) as pe he manu ac u e ’s p o ocol wi h an addi ional cell dis up ion s ep a e lysis Placen al mic obio a associa ed wi h bi h ou comes PLOS ONE | h ps://doi.o g/10.1371/jou nal.pone.0180167 July 12, 2017 3 / 23 wi h P o einase K. In he addi ional s ep, 0.1mm glass beads (Lysing Ma ix B, MP Biomedi- cals) we e added o each sample and he 2ml ubes we e shaken on a cell dis up e (Vo ex Genie 2, Scien i ic Indus ies) o 10 minu es a he highes speed. Fo e e y 10 ex ac ions, a nega i e ex ac ion con ol was included (200μl bu e AE). 16S DNA b oad- ange qPCR All DNA samples pu i ied om placen a and e al memb ane samples we e sc eened o bac e- ia using a quan i a i e PCR (qPCR) SYBR g een luo escen dye assay. The ollowing p ime pai a ge ed he V5-7 egions o he 16S RNA gene, 785F: 50-GGATTAGATACCCBRGT AGTC-30, 1175R: 50-ACGTCRTCCCCDCCTTCCTC-3 0[7]. Each PCR eac ion was ca ied ou wi h he ollowing, 1x Powe SYBR G een mas e mix (Li e echnologies), 0.4μM o o wa d and e e se p ime s, 1μl o empla e DNA and molecula g ade wa e (Bioline) o gi e a inal olume o 25μl. Ampli ica ion ook place in an ABI 7300 Real-Time sys em (Li e echnologies) unde he ollowing condi ions: 95˚C×10 min, 40 cycles o 95˚C×15 sec and 60˚C×1 min. Each PCR un included h ee nega i e PCR con ols (1μl bu e AE om QIAmp DNA mini ki ), and a se ial dilu ion o a known amoun o posi i e con ol om a pu e Esche ichia coli cul u e o quan i ica ion used o calcula e he bac e ial load o each sample. 16S RNA copy numbe o each E.coli dilu ion was calcula ed om he genome molecula weigh and DNA concen- a ion as measu ed by a Qubi 2.0 (Li e echnologies) and co ec ed o he se en 16S RNA copies wi hin he E.coli genome. Samples we e de ined as posi i e o bac e ial DNA i hei C alue was lowe han he lowe limi o de ec ion o his assay. This was de e mined o be 28 ±3 cycles o equi alen o 40 CFU/μl depending on a ia ion be ween uns and was de e - mined using calib a ion cu es gene a ed om se ial dilu ion o a known amoun o posi i e con ol om a pu e Esche ichia coli cul u e. 16S DNA amplicon high- h oughpu sequencing Placen al and e al memb ane samples posi i e o bac e ial DNA by qPCR we e selec ed o sequencing. Lib a y p epa a ion was ca ied ou using dual-indexed o wa d and e e se p im- e s, wi h ba codes aken om a p e ious s udy [16]. The ull lis and combina ions o adap ed p ime s used can be ound in S1 Table. Each lib a y p epa a ion PCR was ca ied ou wi h 1X Molzym PCR Bu e , 200 μM dNTPs (Bioline), 0.4 μM o wa d and e e se p ime , 25 mM Mol aq, 5μl empla e DNA and molecula g ade wa e (Bioline) o gi e a inal eac ion olume o 25μl. The eac ion was ampli ied unde he ollowing condi ions depending on he sample ype. Placen a and e al memb ane samples we e ampli ied unde he ollowing condi ions: 94˚C×3 min, 32 cycles o 94˚C×30 sec, 60˚C×40 sec and 72˚C×90 sec, wi h a inal ex ension cycle o 72˚C×10 min. Vaginal and o al swab samples we e ampli ied unde he ollowing con- di ions: 94˚C×3 min, 30 cycles o 94˚C×30 sec, 60˚C×40 sec and 72˚C×90 sec, wi h a inal ex ension cycle o 72˚C×10 min. The esul ing amplicon was cleaned and pooled using Sequal- P ep no maliza ion pla e ki s (In i ogen) and AMPu e XP beads (Beckman Coul e ) bo h as pe manu ac u e ’s p o ocol. Each pla e was pooled in o an equimola inal lib a y a e quan- i ica ion using a Qubi 2.0 (Li e echnologies). Lib a y was loaded on o a MiSeq (Illumina) as pe manu ac u e ’s p o ocol o 500 cycle V2 ki s wi h he addi ion o cus om sequencing p ime s o ead 1 (TACCGGGACTTAGGATTAGATACCCBRGTAGTC), ead 2 (AACACGTTTTA ACGTCRTCCCCDCCTTCCTC)and index 1 (GAGGAAGGHGGGGAYGACGTTAAAACGTGTT). Bioin o ma ics and s a is ical analysis Pai ed-end sequenced eads om each MiSeq un we e me ged using FLASH [17] and demul- iplexed, pooled and assigned OTUs (Ope a ional Taxonomic Uni s) using QIIME 1.8.0 [18] Placen al mic obio a associa ed wi h bi h ou comes PLOS ONE | h ps://doi.o g/10.1371/jou nal.pone.0180167 July 12, 2017 4 / 23 by clus e ing sequences a 97% simila i y agains a small cus om da abase o ull leng h 16S DNA sequences. Any sequences ha ailed o ma ch a 97% we e assigned agains he ull G eengenes da abase. Mock communi ies o known quan i ies o bac e ial DNA om a mix o species we e sequenced alongside samples and we e used o il e he da ase o e o . Nega i e sequencing and ex ac ion con ols we e also sequenced and con amina ing axonomy we e il- e ed om he da ase . A e il e ing s eps, any samples wi h less han 1000 eads we e emo ed. Alpha di e si y (numbe o unique OTUs) and be a di e si y (in a-indi idual unweigh ed UniF ac dis ances) we e bo h calcula ed as implemen ed in QIIME a e andom subsampling wi hou eplacemen o 1000 eads pe sample. Di e si y o mic obes wi hin an indi idual’s body si e was summa ised by aking he median alue o hei in a-indi idual Uni- F ac dis ances. This alue is calcula ed h ough a pai wise compa ison be ween he bac e ial communi ies o all samples, wi h a high alue ep esen ing a a iable communi y compa ed o o he indi iduals and a lowe alue ep esen ing a s able communi y [19]. The bac e ial load o pa icula phyla, amilies o species was calcula ed by mul iplying he o e all bac e ia load gen- e a ed by qPCR by he p opo ion o hei en i e bac e ial communi y ha was occupied by ha pa icula phylum. Whe e bac e ial loads o di e en indi idual species we e calcula ed, he 16S RNA copy numbe was adjus ed o hose species acco ding o how many 16S RNA genes a e known o be ound in he genome. Bi h weigh as measu ed was used i eco ded wi hin 48 h o deli e y; i no , bi h weigh was back calcula ed om weigh measu ed a 6 o 13 days. I he weigh was i s measu ed wi hin 2 o 5 days a e deli e y, when in an s usually lose weigh , bi h weigh was es ima ed by applying an age-dependen mul iplica i e ac o o he measu ed weigh [20]. Mean p oxy o socioeconomic s a us was calcula ed as p e iously s a ed [21]. We de ined p e e m bi h as deli e y occu ing be o e 37 ull week’s ges a ion and low bi h weigh as <2500g. Leng h- o -age and head ci cum e ence- o -age z sco es (LAZ and HCZ) we e cal- cula ed using he WHO Child G ow h S anda ds [22] and de ined s un ing as <-2.0 LAZ and small head ci cum e ence as <-2.0 HCZ. S a is ical analysis was ca ied ou wi h S a a 13 and R 3.1.0. Fo bi a ia e analyses we used ei he independen - es s o chi-squa ed es s o inde- pendence as app op ia e. Adjus ed linea eg ession models we e used o ela e bac e ial load, alpha and be a di e si y o cho ioamnioni is and bi h ou comes. All eg ession models used o examine associa ions be ween OTUs and bi h ou comes we e adjus ed o he possible e ec s o he in e en ion, ma e nal BMI a en olmen , ma e nal age, p oxy o socioeconomic s a us, numbe o p e ious p egnancies, anaemia a en olmen , si e o en olmen , mode o deli e y and ime be ween deli e y and placen a sampling. Co a ia es en e ed in o adjus ed models we e en e ed in o he model in a one s ep, o ced en y me hod. Fo compa isons be ween abundances o indi idual bac e ial species and bi h ou comes, he q- alue was calcula ed using he Benjamini-Hochbe g co ec ion o con ol o he alse disco e y a e. E hical app o al and consen W i en consen was ob ained om he mo he a en olmen by ei he signa u e o humb- p in i no li e a e. I a humbp in was ob ained an impa ial wi ness also a ended and signed. The ins i u ional e iew boa d app o ed his consen p ocedu e. E hical app o al was ob ained om College o Medicine Resea ch and E hics Commi ee (COMREC), Malawi (P o- ocol numbe : P.08/10/972). The ial was egis e ed a clinical ials.go as NCT01239693. Resul s Sample collec ion A o al o 1391 pa icipan s we e ec ui ed in o he iLiNS-DYAD-M ial and en olmen began in Feb ua y 2011 wi h he las deli e y aking place in Feb ua y 2013. 1097 (78.9%) Placen al mic obio a associa ed wi h bi h ou comes PLOS ONE | h ps://doi.o g/10.1371/jou nal.pone.0180167 July 12, 2017 5 / 23 pa icipan s had a leas one placen al o e al memb ane issue analysed o bac e ia and his- ology. Fig 1 shows a low diag am ha documen s easons o he loss o ollow-up o he 21.1% o pa icipan s who we e no included in his mic obiome s udy. O hese, 49 we e excluded because hey had wins, a placen al sample wasn’ collec ed o no his ological esul was ob ained. Included pa icipan s had a lowe BMI (22.1 s 22.6 kg/m2, p = 0.005), we e olde (25 s 24 yea s, p = 0.025), had comple ed less educa ion (3.9 s 4.7 yea s, p<0.001), had a lowe sco e o socioeconomic s a us (-0.07 s 0.36, p<0.001) and we e less likely o be p i- mipa ous (20.4% s 28.1%, p = 0.006) han hose ha we e excluded (Table 1). Fig 1. S udy pa icipan low diag am. h ps://doi.o g/10.1371/jou nal.pone.0180167.g001 Placen al mic obio a associa ed wi h bi h ou comes PLOS ONE | h ps://doi.o g/10.1371/jou nal.pone.0180167 July 12, 2017 6 / 23 De ec ion o bac e ial DNA Bac e ial DNA was de ec ed in 738 (681%) o e al memb anes and in 476 (468%) o placen al samples. O hose pa icipan s who had de ec able bac e ia he mean (SD) bac e ial load was 5.22 (0.84) Log 10 16S DNA copies/μl in he e al memb anes and 4.80 (0.66) Log 10 16S DNA/ μl copies in he placen a. The median sequencing dep hs o placen al issue and e al mem- b anes we e 11,803 and 21,040 eads pe sample espec i ely (S2 Table). Impac o delay in eezing samples As desc ibed in he me hods, he low-income se ing o he popula ion ec ui ed o his s udy we e la gely u al, subsis ing on a ming and ishing, and as such many did no ha e easy access o heal hca e acili ies. As an icipa ed, his in luenced he ime be ween deli e y, collec- ion o placen al samples and ime o eezing. These delays could in luence bo h he abili y o de ec bac e ial DNA, as well as he bac e ial con en and ela i e abundance. The du a ion o s o age a oom empe a u e was eco ded and we analysed he po en ial o co ounding asso- cia ions. We ound ha while he majo i y o samples we e aken wi hin he i s hou o deli - e y, he ange o imings ex ended un il mo e han 24 hou s (Fig 2A). This delay in sampling in luenced he chances o de ec ing 16S RNA amplicon. Analysis o samples posi i e o bac- e ial DNA we e collec ed a a mean di e ence o 3.2 hou s la e han hose ha had unde ec - able le els (p<0.001, Fig 2B). This enhanced de ec ion o bac e ial DNA may be because o g ow h o bac e ia wi hin/on issues a bi h, con amina ing bac e ia a he ime o deli e y ha hen had ime o g ow, o a he ime o subsequen sample collec ion and p ocessing. We in es iga ed o see i we could iden i y which o hese scena ios was mos likely o be ope a ing in his s udy by compa ing he ypes o bac e ia ound on bo h placen a and e al memb ane issues in ela ion o he delay in p ocessing. We ound no s a is ically signi ican di e ence in ei he in species ichness o di e si y in ela ion o he ime a e deli e y ha samples we e aken (Fig 2C and 2D). We also examined he co ela ion be ween he ime in e al and each indi idual OTU iden i ied in bo h placen a and e al memb anes. The e was no impac o ime o sampling on he OTUs ound in placen al issue. The e was a posi i e co ela ion be ween he ela i e abundance o wo OTUs ound in e al memb anes and he leng h o ime be ween deli e y and p ocessing, a e adjus ing o mul iple compa isons (S3 and S4 Tables). These wo OTUs we e o low abundance and he e o e we e no included in analyses o subsequen associa ions wi h cho ioamnioni is o bi h ou comes. We ound no signi ican associa ions be ween home deli e y and bac e ial di e si y and no associa ions be ween he ime a e deli e y he placen a was sampled and any o he p ima y ou come measu es (S1 Fig). How- e e , as hese da a do no ule ou possible associa ions be ween hese a iables and he Table 1. Baseline cha ac e is ics o included and excluded pa icipan s (n = 1391). Cha ac e is ic Included (n = 1097) Excluded (n = 294) P alue Mean (SD) BMI, kg/m 2 22.1 (2.8) 22.6 (3.0) 0.005 Mean (SD) ma e nal age, yea s 25.1 (6.1) 24.2 (6.3) 0.025 Mean (SD) ma e nal educa ion, comple ed yea s a school 3.9 (3.4) 4.7 (3.8) <0.001 Mean (SD) p oxy o socioeconomic s a us -0.07 (0.9) 0.36 (1.2) <0.001 P opo ion o aenemic women (Hb <110 g/l) 19.9% 24.2% 0.118 P opo ion o p imipa ous women 20.4% 28.1% 0.006 P opo ion o women wi h a low BMI (<18.5 kg/m 2 ) 5.6% 4.7% 0.656 P opo ion o women wi h a posi i e HIV es 13.3% 15.0% 0.469 Numbe (%) o women wi h a posi i e mala ia es (RDT) 23.4% 22.7% 0.874 h ps://doi.o g/10.1371/jou nal.pone.0180167. 001 Placen al mic obio a associa ed wi h bi h ou comes PLOS ONE | h ps://doi.o g/10.1371/jou nal.pone.0180167 July 12, 2017 7 / 23 ou come measu es, we en e ed ime a e he deli e y he placen a was sampled as a co a ia e in all subsequen eg ession analyses. The mic obiome ound in he placen al issues a deli e y is dis inc Fig 3A shows, ac oss he popula ion analysed, he 25 commones o ganisms de ec ed wi hin placen al issues. Bac e ial pa e ns we e simila in e al memb anes and placen al issue. Fe al memb anes had a highe incidence o Lac obacillus ine s,Ga dne ella aginalis and Snea hia sanguinegens whe eas he placen al issues had highe incidences o Acine obac e spp. and En e obac e iaceae spp. We compa ed UniF ac dis ances be ween indi idual’s mic obiomes Fig 2. The e ec o sample collec ion on de ec ion o bac e ial DNA. (A) His og am o he numbe o hou s be ween deli e y and he placen a being sampled. (B) Box-and-whiske plo showing he associa ion be ween he numbe o hou s a e deli e y sampling ook place and whe he he placen al issue had de ec able le els o bac e ial DNA. Co ela ion be ween he numbe o hou s a e deli e y he placen a was sampled (C), and he obse ed numbe o OTUs and median in e -indi idual unweigh ed UniF ac dis ance (D). h ps://doi.o g/10.1371/jou nal.pone.0180167.g002 Placen al mic obio a associa ed wi h bi h ou comes PLOS ONE | h ps://doi.o g/10.1371/jou nal.pone.0180167 July 12, 2017 8 / 23 Fig 3. Iden i ica ion o a co e, dis inc placen al mic obiome. (A) Rank abundance cu e showing he 25 mos common o ganisms eco e ed om pa icipan s’ placen al issue (n = 476) and e al memb anes (n = 738) ha we e posi i e o bac e ial DNA. (B) P incipal coo dina e analysis o unweigh ed UniF ac dis ances compu ed o ma ched pa icipan placen a (n = 445), e al memb ane (n = 719), o al (n = 725) and aginal (n = 747) samples om 1107 pa icipan s. PC1 and PC2 e e o p incipal coo dina e wo and p incipal coo dina e h ee espec i ely. Each indi idual poin e e s o a single pa icipan ’s mic obiome o ha body Placen al mic obio a associa ed wi h bi h ou comes PLOS ONE | h ps://doi.o g/10.1371/jou nal.pone.0180167 July 12, 2017 9 / 23 Fig 7. OTUs iden i ied in placen al issues a e also iden i ied in pa icipan ’s mou h and agina. (A) The mean ±SEM es ima ed p opo ion o OTUs in placen a and e al memb anes sou ced om aginal, o al ca i y o an unknown sou ce. (B) Compa ison be ween he p opo ion o OTUs ound in placen a and e al memb anes sou ced om he agina s a i ied by pa icipan s who deli e ed aginally o by caesa ean sec ion. h ps://doi.o g/10.1371/jou nal.pone.0180167.g007 Placen al mic obio a associa ed wi h bi h ou comes PLOS ONE | h ps://doi.o g/10.1371/jou nal.pone.0180167 July 12, 2017 16 / 23 heal hy deli e ies also concluded ha he placen al mic obiome could no be di e en ia ed om labo a o y eagen con amina ion [26], he e o e u he a emp s a sequencing hese nega i e samples would only esul in an inc ease in he de ec ion o con amina ing bac e ial species. Due o his s udy aking place in a low-income se ing, he e we e many challenges in he collec ion o clinical da a and biological samples. These we e an icipa ed, and whe e possible, he impac o con ounde s was limi ed by ins i u ing a numbe o p ocedu al s eps o op imise he chances o de ec ing bac e ia on placen al issues deposi ed p io o o a deli e y. These included sampling and eezing issues as soon a e deli e y as possible, and e en in emo e se ings his was ela i ely success ul (see Fig 2). Simila ly, placen al samples used o his ologi- cal analysis we e ixed in o malin a he same ime as sampling be o e s o age a +4˚C un il embedded in pa a in and used o his ology. We ensu ed da a collec ion quali y h ough de ailed isi guides, egula s a aining and moni o ing and ins uc ions abou he use o da a collec ion o ms. An h opome ic measu emen s we e aken only by ained pe sonnel (bi h weigh could also be measu ed by s udy nu ses o s udy coo dina o s) whose measu e- men eliabili y was e i ied a he s a o he s udy and a 6-mon h in e als he ea e . All an h opome y equipmen was calib a ed daily and an ex e nal moni o appoin ed by he s udy eam did one si e moni o ing isi du ing da a collec ion. In spi e o he p ocedu es ins i u ed o minimise mic obial con amina ion, he majo i y o placen al issues in his s udy con ained OTUs ha we e no ound in ei he ma ched pa ici- pan aginal o o al samples. I is likely ha hese e lec con amina ing bac e ia o m he en i- onmen and ma e nal aeces. We he e o e ocused on indi iduals ha had a ela i ely small Fig 8. P esence o OTUs in bo h agina and placen al issue associa ed wi h a lowe leng h- o -age z-sco e. (A) Hea map showing p esence o aginal o ganisms in all indi iduals’ aginal and placen al samples. Hie a chical clus e ing was compu ed using a e age linkage o Euclidean dis ances. (B) Mean ±95%CI LAZ sco e o pa icipan s wi h indi idual bac e ium no p esen in hei agina, p esen in hei agina only and p esen in bo h agina and placen al issue (*q<0.05). h ps://doi.o g/10.1371/jou nal.pone.0180167.g008 Placen al mic obio a associa ed wi h bi h ou comes PLOS ONE | h ps://doi.o g/10.1371/jou nal.pone.0180167 July 12, 2017 17 / 23 numbe o phylogene ically di e se OTUs ha we e also ound in ma ched aginal samples and we e associa ed wi h bo h se e e cho ioamnioni is and ad e se bi h ou comes. I seems likely ha hese ea u es ep esen ‘ ue’ in ec ion o he placen al issue. In e es ingly, we ound high abundances o Fusobac e ium nuclea um and U eaplasma spp. ha we e in e sely co ela ed wi h du a ion o p egnancy. This inding was epo ed in ano he s udy using simila echniques [27]. These species ha e also equen ly been ound in he amnio ic luid o women who deli e p e e m [28–30], indica ing a pa hogenic ole in p e- e m bi h. Recen molecula s udies ha e ound Snea hia sanguinegens,P e o ella spp., Pep os- ep ococcus spp. and Ga dne ella aginalis in bo h amnio ic luid and placen al issues o women who ha e deli e ed p ema u ely [6,31] and is consis en wi h ou esul s. These o gan- isms we e no ound in he s udy by P ince e al [27], which may e lec ha he la ge size o ou s udy, (1097 compa ed o 71 pa icipan s), can iden i y e en a low p e alence o po en ially pa hogenic bac e ia. Impo an ly, se e al OTUs associa ed wi h a sho e du a ion o p eg- nancy and smalle newbo n size in ou s udy ha e been p e iously associa ed wi h bac e ial aginosis such as Ga dne ella spp., U eaplasma spp., Snea hia spp., P e o ella spp. and Lach- nospi aceae spp. [32,33]. This includes a ecen s udy ha ound a highe abundance o Ga d- ne ella spp. and U eaplasma spp. in he aginas o women du ing p egnancy who subsequen ly deli e ed p e e m [34]. Fu he e idence ha he app oaches we ha e aken o iden i y po en ially pa hogenic bac- e ial colonisa ion/in ec ion comes om inding compa able le els o bac e ia om he placen- al issue and agina in bo h caesa ean sec ions and aginally deli e ed samples. Ou da a suppo s he o igin o pa hogenic bac e ia de ec ed in placen al issue as o igina ing om he agina. Using samples om he same pa icipan we could show OTUs on placen al issue we e o aginal o o al o igin. When Pep os ep ococcus anae obious and Snea hia sanguinegens we e ound in bo h he agina and placen a, his was associa ed wi h smalle newbo n size. This p o ides u he e idence ha ascending in ec ion, and hence he aginal mic obiome, may play a ole in bi h ou comes. I is no possible o exclude a ole o he o al mic obiome in seeding placen al issue [35], as we did ind he same OTUs in he o al ca i y and placen a in some indi iduals. Fu he mo e, he o al sample collec ed in his s udy was a swab o he gin- gi al ma gins and may no de ec bac e ia ound in deep pe iodon al pocke s o pe iapical is- sues ha could ha e an associa ion wi h ad e se bi h ou comes in his coho [36,37]. Ine i ably wi h a s udy o his ype and in his se ing, he e a e a numbe o po en ial limi- a ions in addi ion o con amina ion al eady discussed. Sampling o he o al and aginal sam- ples occu ed one week a e he deli e y. Bac e ia in he aginal and o al ca i ies may be in luenced by changes occu ing in human mic obiomes immedia ely a e deli e y, especially in he aginal ac [34]. This may impac on any associa ions wi h placen al mic obio a and bi h ou comes. We also ecognise ha using ela i e abundances adjus ed o 16S qPCR load is no a s anda d app oach. We de eloped his app oach because as many placen al samples did no ha e a de ec able esiden mic obiome, compa ing jus ela i e abundances would be skewed by samples ha we e nega i e o o low di e si y. While his app oach canno co ec he in insic biases using hese PCR-based me hodologies, in ou iew i emains he bes a ail- able me hod o analysis o samples wi h a low bac e ial load. The e a e no compa able da a om A ica, bu he o ganisms we de ec ed a e simila o hose iden i ied in amnio ic luid [38,39] o placen al issues [7,40] om s udies ac oss Eu ope and No h Ame ica. This may indica e ha placen al mic obiome we obse ed may be ela- i ely p ese ed in mul iple geog aphical se ings. I may also p o ide a a ionale o ea ing p egnan women wi h an ibio ics. A ial in Malawi p ospec i ely ea ed women wi h an ibi- o ics du ing p egnancy showed a p o ec i e e ec on he incidence o p e e m bi h and low bi h weigh [41]. This could be explained by he clea ing o po en ial pa hogens esponsible Placen al mic obio a associa ed wi h bi h ou comes PLOS ONE | h ps://doi.o g/10.1371/jou nal.pone.0180167 July 12, 2017 18 / 23 o ascending in ec ion ha lead o p e e m bi h o in au e ine g ow h es ic ion. O he mic obiomes, such as ha o he gu , a e known o di e signi ican ly be ween wes e n and A ican popula ions [42] and we ound a mino i y o OTUs likely o o igina e om he GI ac , such as a Blau ia sp., Phascola c obac e ium succina u ens and a Lachnospi aceae sp. ha ha e no p e iously been associa ed wi h ad e se bi h ou comes. I is unknown a he momen whe he hese a e aecal con aminan s o e lec egional di e ences be ween his s udy and p e ious s udies. In Malawi app oxima ely 21% o p egnan women ca y g oup B s ep ococcus (GBS) [43] and has also been ound o in ol ed wi h placen al in ec ion and cho ioamnioni is [44,45]. GBS was no iden i ied in his coho by ou me hods. This may be due o a limi a ion in spe- cies iden i ica ion o sho eads sequenced om he 16S RNA gene, a he han a genuine lack o GBS in his s udy coho . We ound app oxima ely 10% o pa icipan s had uniden i ied S ep ococcus spp. sequences ha did no ha e enough a ia ion be ween hem o con iden ly assign a species name bu may well be GBS. In summa y, we epo he la ges s udy o da e o examine he placen al mic obiome a bi h. While he se ing u al Malawi was challenging, we ha e iden i ied a dis inc mic obial communi y in he placen a and e al memb anes associa ed wi h se e e cho ioamnioni is and ad e se bi h ou comes. Bac e ia associa ed wi h bo h se e e cho ioamnioni is and poo bi h ou comes we e phylogene ically di e se and no he mos abundan axa eco e ed in he pla- cen a o e al memb anes. In e es ingly, he species associa ed wi h p e e m bi h we e dis inc om hose associa ed wi h he h ee measu es o g ow h es ic ion. Fu he s udies a e needed o elucida e mechanisms by which bac e ia es ic e al g ow h wi hou igge ing cho ioam- nioni is o p e e m labou . S a egic con ol o he mic obiome esiden in he agina o o al ca i y wi h p o en e icacy agains o ganisms iden i ied in his s udy could con ol po en ial e iologic agen s ha sp ead o he placen a. While he use o an ibio ics o educe a es o p e- e m deli e y ha e had limi ed success [46,47], a ge ing bac e ia associa ed wi h poo bi h ou comes may p o e mo e success ul. In his espec , he e is a logic o using an an ibio ic such as clindamycin, and indeed his has been shown o educe he isk o p e e m bi h and la e misca iage [48]. Sc eening o he aginal lo a and/o he p esence o ce ical in lamma- ion du ing p egnancy could iden i y a isk g oups who could be a ge ed o ea men wi h selec i e and limi ed an ibio ics o aginal p obio ics o limi ad e se bi h ou comes in he u u e. Suppo ing in o ma ion S1 Fig. Associa ions be ween ime a e deli e y he placen a was sampled, bi h ou comes and whe he a pa icipan deli e ed a home. Box-and-whiske plo s showing he associa ion be ween he ime a e deli e y he placen a was sampled and p e alence o (A) p e e m bi h, (B) low bi h weigh , (C) s un ing and (D) small-head ci cum e ence. As well as he associa- ion be ween he whe he he pa icipan deli e ed a home and he (E) obse ed numbe o OTUs and (F) median in e -indi idual unweigh ed UniF ac dis ance. (TIFF) S2 Fig. Speci ic combina ions o bac e ia ound in placen al issues ha associa e wi h each o he , se e e cho ioamnioni is and ad e se bi h ou comes. Hea map o Spea man’s co e- la ions be ween he 6 mos abundan bac e ial phyla (A) and 20 mos abundan bac e ial ami- lies (B) eco e ed om e al memb anes. Hie a chical clus e ing was compu ed by comple e linkage o Euclidean dis ances. Hea map is anno a ed wi h mean di e ence in bac e ial load be ween pa icipan s wi h and wi hou se e e cho ioamnioni is, p e e m bi h (<37 weeks), low bi h weigh (<2500g) and neona al s un ing (LAZ <-2) and small head ci cum e ence Placen al mic obio a associa ed wi h bi h ou comes PLOS ONE | h ps://doi.o g/10.1371/jou nal.pone.0180167 July 12, 2017 19 / 23 (HCZ <-2) o each bac e ial phyla o amily. As e isks indica e p<005 associa ion be ween highe load o ha bac e ial phyla o amily and p e alence o se e e cho ioamnioni is. Adjus ed model P alues we e calcula ed using linea eg ession adjus ing o he nu i ional in e en ion, ma e nal BMI a en olmen , ma e nal age, p oxy o socioeconomic s a us, num- be o p e ious p egnancies, anaemia, si e o en olmen , mode o deli e y and ime be ween deli e y and placen a sampling. (TIFF) S3 Fig. OTUs iden i ied om ei he a aginal o o al sou ce. Hea map o Spea man’s co e- la ions be ween he ela i e abundance o an OTU ha is p esen in bo h he placen a and ei he he agina o o al ca i y and he es ima ed p opo ion o OTUs om ei he a aginal o o al sou ce as calcula ed by Sou ceT acke (q<0.05). (TIFF) S1 Table. Combina ion o 16S RNA V5-V7 lib a y p epa a ion p ime s o mul iplex 384 samples. (DOCX) S2 Table. Numbe o sequenced eads gene a ed om placen al and e al memb ane sam- ples. (DOCX) S3 Table. Associa ions be ween OTUs isola ed om placen al issue and he ime be ween deli e y and p ocessing o sample. (DOCX) S4 Table. Associa ions be ween OTUs isola ed om e al memb anes and he ime be ween deli e y and p ocessing o sample. (DOCX) S5 Table. OTUs isola ed om e al memb anes signi ican ly associa ed wi h di e ences in du a ion o p egnancy. (DOCX) S6 Table. OTUs isola ed om e al memb anes signi ican ly associa ed wi h di e ences in bi h weigh . (DOCX) S7 Table. OTUs isola ed om e al memb anes signi ican ly associa ed wi h di e ences in leng h- o -age Z-sco e. (DOCX) Acknowledgmen s We would like o hank all he women who ag eed o ake pa in his s udy. We would also like o hank membe s o he iLiNS S ee ing Commi ee: Kenne h H B own, Anna La ey, Jean Bosco Oued aogo, S ephen A Vos i and Mamane Zeilani. Au ho Con ibu ions Concep ualiza ion: Ka h yn Ha is, Ulla Asho n, Ka h yn G. Dewey, Kenne h Male a, Pe Asho n, Nigel Klein. Da a cu a ion: Ronan M. Doyle, Ulla Asho n, Minyanga Nkhoma, Pe Asho n. Placen al mic obio a associa ed wi h bi h ou comes PLOS ONE | h ps://doi.o g/10.1371/jou nal.pone.0180167 July 12, 2017 20 / 23 Fo mal analysis: Ronan M. Doyle. Funding acquisi ion: Ka h yn Ha is, Ulla Asho n, Ka h yn G. Dewey, Kenne h Male a, Nigel Klein. In es iga ion: Ronan M. Doyle, S e e Kamiza, Ulla Ha junmaa. Me hodology: Ronan M. Doyle, Ka h yn Ha is. P ojec adminis a ion: Ka h yn Ha is, Ulla Asho n, Ka h yn G. Dewey, Kenne h Male a, Pe Asho n, Nigel Klein. So wa e: Ronan M. Doyle. Supe ision: Ka h yn Ha is, Pe Asho n, Nigel Klein. Visualiza ion: Ronan M. Doyle. W i ing – o iginal d a : Ronan M. Doyle. W i ing – e iew & edi ing: Ronan M. Doyle, Ka h yn Ha is, S e e Kamiza, Ulla Ha junmaa, Ulla Asho n, Minyanga Nkhoma, Ka h yn G. Dewey, Kenne h Male a, Pe Asho n, Nigel Klein. Re e ences 1. Blencowe H, Cousens S, Chou D, Oes e gaa d M, Say L, Molle A-B, e al. Bo n Too Soon: The global epidemiology o 15 million p e e m bi hs. Rep od Heal h. 2013 No 15; 10(Suppl 1):S2. 2. Blencowe H, Cousens S, Oes e gaa d MZ, Chou D, Molle A-B, Na wal R, e al. Na ional, egional, and wo ldwide es ima es o p e e m bi h a es in he yea 2010 wi h ime ends since 1990 o selec ed coun ies: a sys ema ic analysis and implica ions. Lance . 2012 Jun 9; 379(9832):2162–72. h ps://doi. o g/10.1016/S0140-6736(12)60820-4 PMID: 22682464 3. Asho n P, Alho L, Asho n U, Cheung YB, Dewey KG, Ha junmaa U, e al. The impac o lipid-based nu ien supplemen p o ision o p egnan women on newbo n size in u al Malawi: a andomized con- olled ial. Am J Clin Nu . 2015 Feb 1;ajcn.088617. 4. an den B oek NR, Jean-Bap is e R, Neilson JP. Fac o s associa ed wi h p e e m, ea ly p e e m and la e p e e m bi h in Malawi. PloS One. 2014; 9(3):e90128. h ps://doi.o g/10.1371/jou nal.pone. 0090128 PMID: 24595186 5. Moo e T, Hennessy EM, Myles J, Johnson SJ, D ape ES, Cos eloe KL, e al. Neu ological and de elop- men al ou come in ex emely p e e m child en bo n in England in 1995 and 2006: he EPICu e s udies. The BMJ. 2012 Dec 4; 345:e7961. h ps://doi.o g/10.1136/bmj.e7961 PMID: 23212880 6. Jones HE, Ha is KA, Azizia M, Bank L, Ca pen e B, Ha ley JC, e al. Di e ing p e alence and di e si y o bac e ial species in e al memb anes om e y p e e m and e m labo . PloS One. 2009; 4(12): e8205. h ps://doi.o g/10.1371/jou nal.pone.0008205 PMID: 19997613 7. Doyle RM, Albe DG, Jones HE, Ha is K, Fi zge ald F, Peebles D, e al. Te m and p e e m labou a e associa ed wi h dis inc mic obial communi y s uc u es in placen al memb anes which a e independen o mode o deli e y. Placen a. 2014 Dec; 35(12):1099–101. h ps://doi.o g/10.1016/j.placen a.2014.10. 007 PMID: 25458966 8. Aagaa d K, Ma J, An ony KM, Ganu R, Pe osino J, Ve salo ic J. The Placen a Ha bo s a Unique Mic o- biome. Sci T ansl Med. 2014 May 21; 6(237):237 a65–237 a65. h ps://doi.o g/10.1126/sci anslmed. 3008599 PMID: 24848255 9. S eel JH, Mala os S, Kennea N, Edwa ds AD, Miles L, Duggan P, e al. Bac e ia and in lamma o y cells in e al memb anes do no always cause p e e m labo . Pedia Res. 2005 Ma ; 57(3):404–11. h ps:// doi.o g/10.1203/01.PDR.0000153869.96337.90 PMID: 15659699 10. Goldenbe g RL, Hau h JC, And ews WW. In au e ine in ec ion and p e e m deli e y. N Engl J Med. 2000 May 18; 342(20):1500–7. h ps://doi.o g/10.1056/NEJM200005183422007 PMID: 10816189 11. Sebi e NJ, Goldin RD, Regan L. His ological cho ioamnioni is in ela ion o clinical p esen a ion a 14– 40 weeks o ges a ion. J Obs e Gynaecol J Ins Obs e Gynaecol. 2001 May; 21(3):242–5. 12. Asho n P, Poelman B, Dewey KG, Male a K, Klein N, Roge son S, e al. The Impac o Die a y Supple- men a ion wi h Lipid-Based Nu ien Supplemen s on Ma e nal Heal h and P egnancy Ou comes in Placen al mic obio a associa ed wi h bi h ou comes PLOS ONE | h ps://doi.o g/10.1371/jou nal.pone.0180167 July 12, 2017 21 / 23 Ru al Malawi—FANTA-LNS-RTI-Pa -I-Ma 2017.pd [In e ne ]. 2017 [ci ed 2017 Jun 2]. h ps://www. an ap ojec .o g/si es/de aul / iles/ esou ces/FANTA-LNS-RTI-Pa -I-Ma 2017.pd 13. Asho n P, Poelman B, Dewey KG, Male a K, Klein N, Roge son S, e al. The Associa ions be ween Nu i ion, S ess, In ec ion, and In lamma ion and Ma e nal Heal h and P egnancy Ou comes in Ru al Malawi—FANTA-LNS-RTI-Pa -II-Ma 2017.pd [In e ne ]. 2017 [ci ed 2017 Jun 2]. h ps://www. an ap ojec .o g/si es/de aul / iles/ esou ces/FANTA-LNS-RTI-Pa -II-Ma 2017.pd 14. Asho n P, Poelman B, Dewey KG, Male a K, Klein N, Roge son S, e al. Pa hways Leading o Ad e se Bi h Ou comes in Ru al Malawi—FANTA-LNS-RTI-Pa -III-Ma 2017.pd [In e ne ]. 2017 [ci ed 2017 Jun 2]. h ps://www. an ap ojec .o g/si es/de aul / iles/ esou ces/FANTA-LNS-RTI-Pa -III-Ma 2017.pd 15. Goldenbe g RL, Mudenda V, Read JS, B own ER, Sinkala M, Kamiza S, e al. HPTN 024 s udy: his o- logic cho ioamnioni is, an ibio ics and ad e se in an ou comes in a p edominan ly HIV-1-in ec ed A i- can popula ion. Am J Obs e Gynecol. 2006 Oc ; 195(4):1065–74. h ps://doi.o g/10.1016/j.ajog.2006. 05.046 PMID: 16875654 16. Capo aso JG, Laube CL, Wal e s WA, Be g-Lyons D, Hun ley J, Fie e N, e al. Ul a-high- h oughpu mic obial communi y analysis on he Illumina HiSeq and MiSeq pla o ms. ISME J. 2012 Aug; 6 (8):1621–4. h ps://doi.o g/10.1038/ismej.2012.8 PMID: 22402401 17. MagočT, Salzbe g SL. FLASH: as leng h adjus men o sho eads o imp o e genome assemblies. Bioin o ma Ox Engl. 2011 No 1; 27(21):2957–63. 18. Capo aso JG, Kuczynski J, S ombaugh J, Bi inge K, Bushman FD, Cos ello EK, e al. QIIME allows analysis o high- h oughpu communi y sequencing da a. Na Me hods. 2010 May; 7(5):335–6. h ps:// doi.o g/10.1038/nme h. .303 PMID: 20383131 19. Lozupone C, Knigh R. UniF ac: a new phylogene ic me hod o compa ing mic obial communi ies. Appl En i on Mic obiol. 2005 Dec; 71(12):8228–35. h ps://doi.o g/10.1128/AEM.71.12.8228-8235.2005 PMID: 16332807 20. Cheung YB. S a is ical Analysis o Human G ow h and De elopmen . 1 edi ion. Boca Ra on: Chapman and Hall/CRC; 2013. 378 p. 21. Filme D, P i che LH. Es ima ing Weal h E ec s wi hou Expendi u e Da a-o Tea s: An Applica ion o Educa ional En ollmen s in S a es o India. Demog aphy. 2001; 38(1):115–32. PMID: 11227840 22. Fall CHD, Fishe DJ, Osmond C, Ma ge s BM, Ma e nal Mic onu ien Supplemen a ion S udy G oup. Mul iple mic onu ien supplemen a ion du ing p egnancy in low-income coun ies: a me a-analysis o e ec s on bi h size and leng h o ges a ion. Food Nu Bull. 2009 Dec; 30(4 Suppl):S533–546. h ps:// doi.o g/10.1177/15648265090304S408 PMID: 20120795 23. Knigh s D, Kuczynski J, Cha lson ES, Zane eld J, Moze MC, Collman RG, e al. Bayesian communi y- wide cul u e-independen mic obial sou ce acking. Na Me hods. 2011 Sep; 8(9):761–3. h ps://doi. o g/10.1038/nme h.1650 PMID: 21765408 24. Sal e SJ, Cox MJ, Tu ek EM, Calus ST, Cookson WO, Mo a MF, e al. Reagen and labo a o y con- amina ion can c i ically impac sequence-based mic obiome analyses. BMC Biol. 2014 No 12; 12 (1):87. 25. Ha is KA, Ha ley JC. De elopmen o b oad- ange 16S DNA PCR o use in he ou ine diagnos ic clinical mic obiology se ice. J Med Mic obiol. 2003 Aug; 52(P 8):685–91. h ps://doi.o g/10.1099/jmm. 0.05213-0 PMID: 12867563 26. Laude AP, Roche AM, She ill-Mix S, Bailey A, Laughlin AL, Bi inge K, e al. Compa ison o placen a samples wi h con amina ion con ols does no p o ide e idence o a dis inc placen a mic obio a. Mic obiome. 2016; 4:29. h ps://doi.o g/10.1186/s40168-016-0172-3 PMID: 27338728 27. P ince AL, Ma J, Kannan PS, Al a ez M, Gisslen T, Ha is RA, e al. The placen al memb ane mic o- biome is al e ed among subjec s wi h spon aneous p e e m bi h wi h and wi hou cho ioamnioni is. Am J Obs e Gynecol. 2016 May; 214(5):627.e1–627.e16. 28. Chaim W, Mazo M. In aamnio ic in ec ion wi h usobac e ia. A ch Gynecol Obs e . 1992 Feb 1; 251 (1):1–7. PMID: 1550388 29. Yoon BH, Rome o R, Pa k JS, Chang JW, Kim YA, Kim JC, e al. Mic obial in asion o he amnio ic ca - i y wi h U eaplasma u ealy icum is associa ed wi h a obus hos esponse in e al, amnio ic, and ma e - nal compa men s. Am J Obs e Gynecol. 1998 No ; 179(5):1254–60. PMID: 9822511 30. Han YW, Shen T, Chung P, Buhimschi IA, Buhimschi CS. Uncul i a ed bac e ia as e iologic agen s o in a-amnio ic in lamma ion leading o p e e m bi h. J Clin Mic obiol. 2009 Jan; 47(1):38–47. h ps://doi. o g/10.1128/JCM.01206-08 PMID: 18971361 31. DiGiulio DB, Rome o R, Amogan HP, Kusano ic JP, Bik EM, Go sch F, e al. Mic obial p e alence, di e si y and abundance in amnio ic luid du ing p e e m labo : a molecula and cul u e-based in es iga- ion. PloS One. 2008; 3(8):e3056. h ps://doi.o g/10.1371/jou nal.pone.0003056 PMID: 18725970 Placen al mic obio a associa ed wi h bi h ou comes PLOS ONE | h ps://doi.o g/10.1371/jou nal.pone.0180167 July 12, 2017 22 / 23 32. Shipi syna E, Roos A, Da cu R, Halle ´n A, F edlund H, Jensen JS, e al. Composi ion o he aginal mic obio a in women o ep oduc i e age—sensi i e and speci ic molecula diagnosis o bac e ial agi- nosis is possible? PloS One. 2013 Ap 9; 8(4):e60670. h ps://doi.o g/10.1371/jou nal.pone.0060670 PMID: 23585843 33. Fe weis JM, B ooks JP, Se ano MG, She h NU, Gi e d PH, Edwa ds DJ, e al. Di e ences in aginal mic obiome in A ican Ame ican women e sus women o Eu opean ances y. Mic obiol Read Engl. 2014 Jul 29; 34. DiGiulio DB, Callahan BJ, McMu die PJ, Cos ello EK, Lyell DJ, Robaczewska A, e al. Tempo al and spa ial a ia ion o he human mic obio a du ing p egnancy. P oc Na l Acad Sci. 2015 Aug 17; 35. Gonzales-Ma in C, Sp a DA, Allake R. Ma e nal o al o igin o Fusobac e ium nuclea um in ad e se p egnancy ou comes as de e mined using he 16S-23S DNA in e genic ansc ibed space egion. J Med Mic obiol [In e ne ]. 2012 Sep 20 [ci ed 2012 Sep 25]; h p://www.ncbi.nlm.nih.go /pubmed/ 23002071 36. Ha junmaa U, Ja ¨ ns ed J, Alho L, Dewey KG, Cheung YB, Dei chle M, e al. Associa ion be ween ma e nal den al pe iapical in ec ions and p egnancy ou comes: esul s om a c oss-sec ional s udy in Malawi. T op Med In Heal h TM IH. 2015 Jul 30; 37. Shaw L, Ha junmaa U, Doyle R, Mulewa S, Cha lie D, Male a K, e al. Dis inguishing he Signals o Gin- gi i is and Pe iodon i is in Sup agingi al Plaque: a C oss-Sec ional Coho S udy in Malawi. Appl En i- on Mic obiol. 2016 Sep 16; 82(19):6057–67. h ps://doi.o g/10.1128/AEM.01756-16 PMID: 27520811 38. Wang X, Buhimschi CS, Temoin S, Bhanda i V, Han YW, Buhimschi IA. Compa a i e mic obial analysis o pai ed amnio ic luid and co d blood om p egnancies complica ed by p e e m bi h and ea ly-onse neona al sepsis. PloS One. 2013 Feb 20; 8(2):e56131. h ps://doi.o g/10.1371/jou nal.pone.0056131 PMID: 23437088 39. DiGiulio DB, Rome o R, Kusano ic JP, Go ´mez R, Kim CJ, Seok KS, e al. P e alence and Di e si y o Mic obes in he Amnio ic Fluid, he Fe al In lamma o y Response, and P egnancy Ou come in Women wi h P e e m P e-Labo Rup u e o Memb anes. Am J Rep od Immunol. 2010 Jul 1; 64(1):38–57. h ps://doi.o g/10.1111/j.1600-0897.2010.00830.x PMID: 20331587 40. Quei os da Mo a V, P odhom G, Yan P, Hohl held P, G eub G, Rouleau C. Co ela ion be ween placen- al bac e ial cul u e esul s and his ological cho ioamnioni is: a p ospec i e s udy on 376 placen as. J Clin Pa hol. 2013 Ma ; 66(3):243–8. h ps://doi.o g/10.1136/jclinpa h-2012-201124 PMID: 23268318 41. Lun amo M, Kulmala T, Mbewe B, Cheung YB, Male a K, Asho n P. E ec o epea ed ea men o p egnan women wi h sul adoxine-py ime hamine and azi h omycin on p e e m deli e y in Malawi: a andomized con olled ial. Am J T op Med Hyg. 2010 Dec; 83(6):1212–20. h ps://doi.o g/10.4269/ aj mh.2010.10-0264 PMID: 21118924 42. Ya sunenko T, Rey FE, Mana y MJ, T ehan I, Dominguez-Bello MG, Con e as M, e al. Human gu mic obiome iewed ac oss age and geog aphy. Na u e. 2012 Jun 14; 486(7402):222–7. h ps://doi.o g/ 10.1038/na u e11053 PMID: 22699611 43. G ay KJ, Ka ula ula G, Ma emba M, Kamdolozi M, Membe G, F ench N. G oup B S ep ococcus and HIV in ec ion in p egnan women, Malawi, 2008–2010. Eme g In ec Dis. 2011 Oc ; 17(10):1932–5. h ps://doi.o g/10.3201/eid1710.102008 PMID: 22000375 44. Whidbey C, Ha ell MI, Bu nside K, Ngo L, Bec a AK, Iye LM, e al. A hemoly ic pigmen o G oup B S ep ococcus allows bac e ial pene a ion o human placen a. J Exp Med. 2013 Ma 6; 210(6):1265– 81. h ps://doi.o g/10.1084/jem.20122753 PMID: 23712433 45. Schmiedel D, Kikhney J, Masseck J, Rojas Mencias PD, Schulze J, Pe ich A, e al. Fluo escence in si u hyb idiza ion o iden i ica ion o mic oo ganisms in acu e cho ioamnioni is. Clin Mic obiol In ec . 2014 Sep 1; 20(9):O538–41. h ps://doi.o g/10.1111/1469-0691.12526 PMID: 24382010 46. Klebano MA, Ca ey JC, Hau h JC, Hillie SL, Nugen RP, Thom EA, e al. Failu e o Me onidazole o P e en P e e m Deli e y among P egnan Women wi h Asymp oma ic T ichomonas aginalis In ec- ion. N Engl J Med. 2001 Aug 16; 345(7):487–93. h ps://doi.o g/10.1056/NEJMoa003329 PMID: 11519502 47. Kenyon SL, Taylo DJ, Ta now-Mo di W, ORACLE Collabo a i e G oup. B oad-spec um an ibio ics o spon aneous p e e m labou : he ORACLE II andomised ial. ORACLE Collabo a i e G oup. Lance . 2001 Ma 31; 357(9261):989–94. PMID: 11293641 48. Lamon RF, Nhan-Chang C-L, Sobel JD, Wo kowski K, Conde-Agudelo A, Rome o R. T ea men o abno mal aginal lo a in ea ly p egnancy wi h clindamycin o he p e en ion o spon aneous p e e m bi h: a sys ema ic e iew and me aanalysis. Am J Obs e Gynecol. 2011 Sep; 205(3):177–90. h ps:// doi.o g/10.1016/j.ajog.2011.03.047 PMID: 22071048 Placen al mic obio a associa ed wi h bi h ou comes PLOS ONE | h ps://doi.o g/10.1371/jou nal.pone.0180167 July 12, 2017 23 / 23