Epidemiology and clinical features of community-acquired, healthcare-associated and nosocomial bloodstream infections in tertiary-care and community hospitals

Epidemiology and clinical ea u es o communi y-acqui ed, heal hca e-associa ed and nosocomial bloods eam in ec ions in e ia y-ca e and communi y hospi als J. Rod ı ´guez-Ban ˜o 1 ,M.D.Lo ´pez-P ie o 2 , M. M. Po illo 1 , P. Re ama 1 , C. Na e a 3 , E. Nun ˜o 4 , M. He e o 5 , A. del A co 6 , A ´. Mun ˜oz 7 ,F.Te ´llez 8 , M. To es-To osa 9 , A. Ma ı ´n-Aspas 10 , A. A oyo 11 , A. Ruiz 12 , R. Moya 13 , J. E. Co zo 14 , L. Leo ´n 15 and J. A. Pe ´ ez-Lo ´pez 16 , on behal o he SAEI/SAMPAC Bac e aemia G oup 1) Seccio ´n de En e medades In ecciosas, Hospi al Uni e si a io Vi gen Maca ena, Se ille, 2) Se icio de Mic obiologı ´a, Hospi al del SAS, Je ez de la F on e a (Ca ´diz), 3) Seccio ´n de En e medades In ecciosas, Hospi al Uni e si a io Reina So ı ´a, Co ´ doba, 4) Seccio ´n de En e medades In ecciosas, Hospi al Uni e si a io Vi gen de la Vic o ia, Ma ´laga, 5) Se icio de En e medades In ecciosas, Hospi ales Uni e si a ios Vi gen del Rocı ´o, Se ille, 6) Se icio de Medicina In e na, Hospi al Cos a del Sol, Ma bella (Ma ´laga), 7) Se icio de Medicina In e na, Hospi al de la Se anı ´a, Ronda (Ma ´laga), 8) Unidad de En e medades In eccio- sas, Hospi al de La Lı ´nea, La Lı ´nea de la Concepcio ´n (Ca ´diz), 9) Seccio ´n de En e medades In ecciosas, Hospi al Pun a de Eu opa, Algeci as (Ca ´diz), 10) Se icio de Medicina In e na, Hospi al Pue a del Ma , Ca ´diz, 11) Seccio ´n de En e medades In ecciosas, Complejo Hospi ala io de Jae ´n, Jae ´n, 12) Unidad de En e medades In ecciosas, Hospi al del SAS, Je ez de la F on e a (Ca ´diz), 13) Se icio de Medicina In e na, Hospi al de An eque a, Ma ´laga, 14) Seccio ´n de En e medades In ecciosas, Hospi al Uni e si a io de Valme, Se ille, 15) Unidad de En e medades In ecciosas, Hospi al To eca ´ denas, Alme ı ´a and 16) Se icio de Mic obiologı ´a, Hospi al Uni e si a io San Cecilio, G anada, Spain Abs ac Classi ica ion o bloods eam in ec ions (BSIs) as communi y-acqui ed (CA), heal hca e-associa ed (HCA) and hospi al-acqui ed (HA) has been p oposed. The epidemiology and clinical ea u es o BSI acco ding o ha classi ica ion in e ia y-ca e (TH) and communi y (CH) hospi als we e in es iga ed in a p ospec i e coho o 821 BSI episodes om 15 hospi als ( en TH and i e CH hospi als) in Andalucı ´a, Spain. Eigh een pe cen we e CA, 24% we e HCA and 58% we e HA. The incidence o CA and HCA BSI was highe in CH han in TH (CA: 3.9 episodes pe 1000 admissions s. 2.2, p <0.01; HCA: 5.0 s. 2.9, p <0.01), whe eas he incidence o HA BSI was lowe (7.7 s. 8.7, p <0.01). In CA and HCA BSI, he espi a o y ac was mo e equen ly he sou ce in CH han in TH (CA: 30% s. 15%; HCA: 20% s. 9%, p £0.03). In HCA BSI, ch onic enal insu iciency and unnelled ca he e s we e less equen in CH han in TH (11% s. 26% and 7% s. 19%, p £0.03), al hough ch onic ulce s we e mo e equen (22% s. 8%, p 0.008). BSIs as a esul o me hicillin- esis an S aphylo- coccus au eus o Pseudomonas ae uginosa we e e y a e in CA episodes, al hough ex ended-spec um b-lac amase-p oducing Esche ichia coli (ESBLEC) caused a simila p opo ion o all BSIs in CA, HCA and HA episodes. Mul i a ia e analysis e ealed no signi ican di e ence in mo ali y a es in CH and TH. HCA in ec ions should be conside ed as a sepa a e class o BSI in bo h TH and CH, al hough di e ences be ween hospi als mus be conside ed. CA BSIs we e no caused by mul id ug- esis an pa hogens, excep o ESBLEC. Keywo ds: Bloods eam in ec ions, communi y-acqui ed in ec ions, ex ended-spec um b-lac amases, heal hca e-associa ed in ec ions, me hicillin- esis an S aphylococcus au eus, mul icen e s udy, nosocomial in ec ions O iginal Submission: 17 July 2009; Re ised Submission: 9 Oc obe 2009; Accep ed: 9 Oc obe 2009 Edi o : M. Paul Clin Mic obiol In ec Co esponding au ho and ep in eques s: J. Rod ı ´guez Ban ˜o, Seccio ´n de En e medades In ecciosas, Hospi al Uni e si a io Vi gen Maca ena, A da D Fed iani 3, 41009 Se ille, Spain E-mail: [email p o ec ed]m In oduc ion Bloods eam in ec ions (BSI) a e an impo an cause o mo - bidi y and mo ali y in hospi alized pa ien s [1–3]. T adi ionally, BSI ha e been di ided in o communi y and nosocomial episodes [4]. Howe e , subsequen o he implemen a ion o ambula o y al e na i es o inpa ien ca e, signi ican changes in he epidemiology o BSI ha e been no ed and a p oposal was made o u he subdi ide communi y-onse BSI in o heal hca e-associa ed (HCA) episodes o pa ien s wi h signi - ican ecen heal hca e con ac and p ocedu es, and s ic ly communi y-acqui ed (CA) episodes o pa ien s wi hou [5]. Howe e , his new classi ica ion has been e alua ed in only a ew s udies [5–9], in which HCA episodes we e CLM 3089 B Dispa ch: 2.2.10 Jou nal: CLM CE: Shiyamala Jou nal Name Manusc ip No. Au ho Recei ed: No. o pages: 6 PE: Gaya h i ª2010 The Au ho s Jou nal Compila ion ª2010 Eu opean Socie y o Clinical Mic obiology and In ec ious Diseases ORIGINAL ARTICLE 10.1111/j.1469-0691.2010.03089.x 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 shown o be somewha mo e simila o hospi al-acqui ed (HA) han o CA BSI in ae iology and p edisposing condi- ions; hese s udies included e ospec i e analysis o included cases om a limi ed numbe o hospi als, and used di e en c i e ia. I s applicabili y o bo h e ia y-ca e hospi- als and communi y hospi als has no been s udied speci i- cally, no is i known whe he his classi ica ion is help ul in iden i ying pa ien s wi h a low isk o acqui ing some eme - gen communi y-onse in ec ion-causing, an ibio ic- esis an pa hogens, such as ex ended-spec um b-lac amase (ESBL)-p oducing Esche ichia coli [10]. The p esen s udy aimed o in es iga e he popula ion-based incidence, epidemi- ology, ae iology and clinical ea u es o BSI acco ding o he epidemiological ype o in ec ion, and o e alua e he newly- p oposed scheme wi hin a wide sample o hospi als ha included bo h e ia y-ca e and communi y se ings. Ma e ials and Me hods Si e and design A mul icen e p ospec i e coho s udy was conduc ed in 15 public hospi als ( en e ia y-ca e and i e communi y hospi- als) in Andalucı ´a, Spain. The pa icipa ing hospi als p o ide heal h ca e o >90% o hei ca chmen a eas (>4 million) [11]. All episodes o clinically signi ican BSI in adul pa ien s (>14 yea s) be ween 15 Oc obe 2006 and 15 Decembe 2006 we e included. Because he aim o he s udy was o include a signi ican numbe o cases om communi y cen- es, he s udy pe iod was ex ended un il 15 Ma ch 2007 in hese hospi als. Blood cul u es we e pe o med, p ocessed and in e - p e ed in acco dance wi h he ecommenda ions o he Spanish Socie y o In ec ious Diseases and Clinical Mic obiol- ogy [12]. Blood cul u es (a leas wo blood d aws) we e ob ained om pe iphe al eins, excep in a ew cases whe e he cen al enous ca he e was he suspec ed cause, when a leas one blood d aw was ob ained om a pe iphe al ein and ano he om he ca he e . Suscep ibili y esul s we e in e p e ed acco ding o he CLSI ecommenda ions [13]; in his analysis, we ocussed on me hicillin esis ance in S aphylo- coccus au eus, ancomycin- esis ance in en e ococci, ESBL p oduc ion in E. coli and Klebsiella spp., and ca bapenem esis ance in En e obac e iaceae and Pseudomonas ae uginosa. Cases we e de ec ed by daily e iew o blood cul u e esul s. All pa ien s wi h posi i e blood cul u es we e consid- e ed eligible. Di e en episodes occu ing in he same pa ien s we e included only when caused by di e en spe- cies. Fo po en ial con aminan s (coagulase-nega i e s aphylo- cocci, diph he oids), only episodes in which he o ganism had been isola ed om wo o mo e blood d aws we e included [4,12]. Pa ien s we e ollowed un il discha ge o dea h. The s udy was app o ed by he E hics Commi ees o he pa icipa ing cen es, which wai ed he need o ob ain- ing in o med consen . Va iables and de ini ions The da a collec ed included: ype o hospi al; epidemiologic ype o in ec ion (CA, HCA o HA); demog aphics; ype and se e i y o ch onic unde lying disease [14]; sou ce o BSI (see below); ascula o u ina y ca he e and mechanical en- ila ion a onse , endoscopic p ocedu es pe o med wi hin he p eceding week, and majo su ge y in he p eceding 3 mon hs; an imic obial use; ae iology; empi ical he apy and ou come. Episodes we e classi ied as HA i he episode occu ed mo e han 48 h a e admission [4]; all o he episodes we e conside ed communi y-onse , which we e hen classi ied as HCA o CA, acco ding o he c i e ia used by F iedman e al. [5]. In b ie , episodes we e classi ied as HCA when any o he ollowing was p esen : in a enous he apy o special- is nu sing ca e a home, haemodialysis in he 30 days be o e he BSI; hospi aliza ion o >2 days in an acu e ca e hospi al in he 90 days be o e he BSI; o esiden s a us in a nu sing home o long- e m ca e acili y. The BSI sou ce was de e mined om clinical and labo a- o y da a, using CDC c i e ia o seconda y BSI [4]. Empi i- cal he apy was conside ed app op ia e when an ac i e an imic obial agen (acco ding o suscep ibili y da a) was adminis e ed a ecommended doses wi hin he i s 24 h a e he blood cul u es we e pe o med. C ude mo ali y was eco ded a days 14 and 30. S a is ical analysis Popula ion-based incidence a es we e calcula ed acco ding o he assigned e e ence popula ion o each cen e [15] and ex apola ed o 1 yea . The e e ence popula ions a e de ined by speci ic geog aphic limi s; pa ien s needing hospi al admission a e ini ially hospi alized in hei co esponding cen e and a e only ans e ed o e e ence cen es when needed. A compa ison o a es was pe o med using a Pois- son eg ession model. Compa isons be ween CA and HCA, and be ween HCA and HA episodes we e pe o med simi- la ly o p e iously epo ed me hods [5]; a iables included in he de ini ions we e no compa ed. Simila me hods we e used o compa isons be ween e ia y-ca e and communi y hospi als. Ca ego ical and con inuous a iables we e com- pa ed using Fishe ’s exac es and he he Mann–Whi ney U- es , espec i ely. Whene e a signi ican c ude associa ion 2Clinical Mic obiology and In ec ion CMI ª2010 The Au ho s Jou nal Compila ion ª2010 Eu opean Socie y o Clinical Mic obiology and In ec ious Diseases, CMI 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 was ound be ween he ype o hospi al and mo ali y, a mul- i a ia e analysis was pe o med by logis ic eg ession o con ol o con oude s using a s epwise o wa d me hod. Resul s Incidence o BSI and epidemiological ype o in ec ion Du ing he s udy pe iod, 821 BSI episodes in adul s we e included; 660 (80%) occu ed in e ia y-ca e hospi als and 161 (20%) occu ed in communi y hospi als. O e all, 476 (58%) we e classi ied as HA, 195 (24%) as HCA, and 150 (18%) as CA. The dis ibu ion o BSI acco ding o p e- de ined epidemiological ypes o in ec ion was somewha di - e en in he wo ypes o hospi als; compa ed o e ia y hospi als, he e was a highe equency o CA episodes in communi y cen es (25% s. 17%, p 0.01), a lowe equency o HA episodes (47% s. 60%, p 0.002) and a simila e- quency o HCA episodes (28% s. 23%, p 0.1). The incidence o BSIs is shown in Table 1. The o e all popula ion-based incidence was highe in e ia y-ca e cen es ela ed o he highe incidence o nosocomial episodes; he admissions- based incidence was highe in communi y hospi als. Because only i e pa ien s had been admi ed om o he cen es ou - side hei ca chmen a ea, he exclusion o hese pa ien s om he analysis did no change he esul s. O pa ien s wi h HCA episodes, 131 (67%) had been p e- iously admi ed, 28 (14%) we e esiden s in long- e m ca e acili ies, 28 (14%) we e ecei ing haemodialysis and wo (1%) we e ecei ing pe i oneal dialysis, 41 (21%) a ended a day hospi al, ou (2%) ecei ed home in a enous he apy, and i e (3%) we e ecei ing o he ypes o specialized home ca e (se e al pa ien s had mo e han one o hese p edispos- ing ac o s). P edisposing ea u es, sou ces and ou come acco ding o epidemiological ype o in ec ion Table 2 shows pa ien cha ac e is ics, clinical ea u es and he ou come o BSI. Ch onic enal insu iciency, haema ologi- cal cance and neu openia we e mo e equen in pa ien s wi h HCA han in CA BSI. Compa ed o HA episodes, HCA occu ed mo e equen ly in pa ien s wi h ch onic enal insu iciency, li e disease and cu aneous ulce , whe eas p e- ious an ibio ic use and in asi e p ocedu es (wi h he excep- ion o unnelled enous ca he e s) we e mo e equen in pa ien s wi h HA BSI. Wi h ega d o he sou ce o in ec ion, u ina y ac in ec ion was mo e equen in CA episodes han in HCA BSI. Endoca di is was mo e equen in HCA han in HA cases, whe eas ascula ca he e - ela ed in ec- ions we e mo e equen in HA BSI. Inapp op ia e empi ical he apy was mo e equen in HA BSI. Al hough he e was a end owa d highe mo ali y in HCA and HA episodes, hese di e ences we e no s a is ically signi ican . TABLE 1. Popula ion-based and admissions-based incidence a es o bloods eam in ec ions All hospi als Te ia y hospi als Communi y hospi als Episodes pe 100 000 popula ion/yea All episodes 109.2 116.4 a 75.3 CA 19.2 19.2 18.0 HCA 25.2 25.2 27.3 HA 64.8 72.0 a 30.0 Episodes pe 1000 admissions All episodes 14.7 13.9 a 16.8 CA 2.7 2.2 a 3.9 HCA 3.5 2.9 a 5.0 HA 8.4 8.7 a 7.7 CA, communi y-acqui ed; HCA, heal hca e associa ed; HA, hospi al-acqui ed. a Fo compa ison wi h communi y hospi als, p <0.01 TABLE 2. Pa ien cha ac e is ics, clinical ea u es and mo ali y o bloods eam in ec ions by epidemiological ype o in ec ion CA (n= 150) HCA (n= 195) HA (n= 476) Age in yea s, median (IQ ange) 67 (46–77) 69 (56–70) 66 (53–75) Male gende 80 (53) 116 (60) 286 (60) Non a al unde lying disease 106 (71) a 91 (47) 326 (68) Ch onic enal insu iciency 5 (3) a 33 (17) b 42 (9) Ch onic li e disease 17 (11) 26 (13) b 31 (7) Solid cance 21 (14) 34 (17) 112 (24) Haema ological cance 4 (3) a 24 (12) 43 (9) HIV in ec ion 6 (4) 5 (3) 5 (1) Ch onic cu aneous ulce – c 12 (6) d 9 (2) Neu openia 2 (1) a 13 (7) 32 (7) Non unnelled cen al enous ca he e – 4 (2) b 182 (38) Tunnelled enous ca he e – 32 (16) b 35 (7) U ina y ca he e 19 (13) 25 (13) b 238 (50) Mechanical en ila ion – 1 (1) b 93 (20) Endoscopic p ocedu e 1 (1) 1 (1) b 28 (6) Su ge y – 9 (5) b 106 (22) P e ious an imic obials 26 (17) a 62 (32) b 253 (53) Sou ce Unknown 22 (15) 45 (23) 127 (27) U ina y ac 46 (31) a 41 (21) 73 (15) Respi a o y ac 28 (19) 22 (11) 55 (12) In a-abdominal in ec ion 15 (10) 22 (11) 44 (9) Bilia y ac 18 (12) 18 (9) 26 (5) Endoca di is 8 (5) 9 (5) b 6 (1) Skin and skin s uc u es 8 (5) 10 (5) 21 (4) Vascula ca he e – 24 (12) b 116 (24) O he s 5 (3) 4 (2) 8 (2) App op ia e empi ical he apy 115 (77) b 141 (72) b 296 (62) Mo ali y 14-day 22 (15) 36 (18) 97 (20) 30-day 29 (19) 43 (22) 116 (24) Da a a e exp essed as he numbe o cases (%), excep whe e speci ied. p alue o all s a is ical compa isons ‡0.05, excep whe e speci ied. CA, communi y-acqui ed. HCA, heal hca e-associa ed. HA, hospi al-acqui ed. a Fo compa ison wi h heal hca e-associa ed episodes, p <0.01 (Fishe ’s exac es ). b Fo compa ison wi h hospi al-acqui ed episodes, p £0.01 (Fishe ’s exac es ). c Fo compa ison wi h heal hca e-associa ed episodes, p 0.02 (Fishe ’s exac es ). d Fo compa ison wi h hospi al-acqui ed episodes, p 0.04 (Fishe ’s exac es ). 2 CMI Rod ı ´guez-Ban ˜oe al. Bac e aemia in e ia y-ca e and communi y hospi als 3 ª2010 The Au ho s Jou nal Compila ion ª2010 Eu opean Socie y o Clinical Mic obiology and In ec ious Diseases, CMI 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 Ae iology and an imic obial suscep ibili y The mic oo ganisms causing BSI a e shown in Table 3. O e - all, G am-nega i es we e mo e equen in HCA han in CA and HA in ec ions. S. au eus, coagulase-nega i e s aphylo- cocci, En e ococcus spp., En e obac e spp. and P. ae uginosa we e less equen in CA in ec ions, whe eas S aphylococcus pneumoniae and E. coli we e mo e equen ly associa ed wi h his g oup. None o he CA cases we e caused by me hicillin- esis an S. au eus, S eno ophomonas mal ophilia o Acine obac- e baumannii, and only one by P. ae uginosa. By con as , he pe cen ages o isola es p oducing ESBLs among E. coli o Kle- bsiella spp. we e simila among CA, HCA and HA BSI. The e we e no episodes as a esul o ancomycin- esis an en e o- cocci o ca bapenem- esis an En e obac e iceae. Analysis o BSI by hospi al ype Among pa ien s wi h CA episodes (40 om communi y hos- pi als and 110 om e ia y hospi als), he e we e no signi i- can di e ences in demog aphics, unde lying condi ions, p edisposing ac o s, sou ce, mic oo ganisms o mo ali y by ype o hospi al (da a no shown), wi h he ollowing excep ions: he espi a o y ac was a mo e equen sou ce in communi y cen es (30% s. 15%, p 0.01), whe eas endo- ca di is was a he less equen (0 s. 7%, p 0.07); E. coli was less equen in communi y cen es (23% s. 44%, p 0.01), whe eas S. pneumoniae was mo e equen (28% s. 15%, p 0.06). Mo e di e ences we e ound in pa ien s wi h HCA epi- sodes (45 om communi y and 150 om e ia y-ca e hospi- als). Ch onic enal insu iciency was less equen in communi y cen es (11% s. 26%, p 0.03) and ch onic cu a- neous ulce s we e mo e equen (22% s. 8%, p 0.008). Tunnelled enous ca he e s we e mo e equen in e ia y- ca e cen es (19% s. 7%, p 0.04), as was haemodialysis (18% s. 2%, p 0.008). The espi a o y ac was he mos equen sou ce o BSI in communi y hospi als (20% s. 9%, p 0.03). Mo ali y a day 14 was highe in communi y cen es (16% s. 29%, p 0.03). Howe e , he di e ence was no longe s a- is ically signi ican when adjus ed o p esen a ion wi h se e e sepsis o sep ic shock, sou ce o in ec ion and se e - i y o unde lying disease (adjus ed OR 2.5, 95% CI 0.8–6.9, p 0.1). The e was no signi ican di e ence in mo ali y a day 30 (20% s. 13%, p 0.2). Fo pa ien s wi h HA episodes (76 om communi y cen- es and 400 om e ia y hospi als), he ollowing di e - ences we e ound: haema ological cance and neu openia we e mo e equen in e ia y-ca e cen es (11% s. 1%, p 0.01, and 8% s. 1%, p 0.04); enous ca he e s we e also mo e equen (90% s. 73%, p <0.001) bu p e ious endo- scopic p ocedu es we e less equen (5% s. 12%, p 0.01). Skin and so issue in ec ions we e he mos equen sou ce o BSI in e ia y-ca e hospi als (21% s. 0, p 0.04), whe eas u ina y ac and in a-abdominal in ec ions we e less equen (14% s. 24%, p 0.02, and 7% s. 20%, p <0.001). The e we e no signi ican di e ences in ae iology o mo ali y. Discussion The conside a ion o HCA as a speci ic ca ego y o BSI has been ound o be clinically ele an in he ew s udies add essing he issue ha we a e awa e o [5–9] because i was disco e ed o be mo e simila o HA han o CA epi- sodes wi h espec o unde lying condi ions and ae iology. We we e in e es ed in in es iga ing whe he he HCA ca e- go y was jus as ele an in a b oade sample o hospi als ha included e ia y-ca e and communi y hospi als. We used he c i e ia empi ically de eloped by F iedman e al. [5] because hey ha e been shown o be o p ognos ic impo - ance [5,8] and p edic i e o ine ec i e ini ial he apy [16]. TABLE 3. Mic oo ganisms isola ed om blood cul u es, by epidemiological ype o in ec ion CA (n= 150) HCA (n= 195) HA (n= 476) G am-posi i e 67 (45) 74 (38) a 240 (50) S aphylococcus au eus 10 (7) 22 (11) 67 (14) Me hicillin- esis an S. au eus b 0 c 6 (3%) 24 (5%) Coagulase-nega i e s aphylococci 8 (5) 20 (10) a 118 (25) En e ococcus 5 (3) 8 (4) a 47 (10) S aphylococcus pneumoniae 27 (18) d 11 (6) a 3 (<1) G am-nega i e 85 (57) 124 (64) a 237 (50) Esche ichia coli 57 (38) 72 (37) a 96 (20) ESBL-p oducing E. coli e 5 (3) 6 (3) 18 (4) Klebsiella spp.12 (8) 13 (7) 39 (8) ESBL-p oducing Klebsiella spp. 1 (<1) 1 (<1) 3 (<1) En e obac e spp.1 (<1) 8 (4) 17 (4) Pseudomonas ae uginosa g 1 (<1) d 13 (7) 26 (5) Acine obac e baumannii 0 2 (1) 23 (5) S eno ophomonas mal ophilia 0 1 (<1) 5 (1) Anae obes 4 (3) 3 (2) 6 (1) Fungi 0 1 (1) h 15 (3) Polymic obial bac e aemia 12 (8) 10 (5) 38 (8) Da a a e exp essed as he numbe o cases (%), excep whe e speci ied. The p alue o all s a is ical compa isons ‡0.05, excep whe e speci ied. CA, communi y-acqui ed; HCA, heal hca e associa ed; HA, hospi al-acqui ed. a Fo compa ison wi h hospi al-acqui ed episodes, p £0.01 (Fishe ’s exac es ). b Me hicillin- esis an isola es we e 0, 27% and 36% o all S. au eus in CA, HCA and HA bloods eam in ec ion (BSI), espec i ely.The p alue o CA s. HA was 0.02 (Fishe ’s exac es ). c Fo compa ison wi h heal hca e-associa ed episodes, p 0.03 (Fishe ’s exac es ). d Fo compa ison wi h heal hca e-associa ed episodes, p £0.01 (Fishe ’s exac es ). e Ex ended-spec um b-lac amase (ESBL)-p oducing isola es we e 9%, 8% and 19% o all E. coli in CA, HCA, and HA BSI, espec i ely. The p alue o HCA s. HA was 0.05 (Fishe ’s exac es ). ESBL-p oducing isola es we e 8% o all Klebsiella spp. in CA, HCA, and HA BSI. g Ca bapenem- esis an isola es we e 0, 15% and 12% all P. ae uginosa in CA, HCA, and HA BSI, espec i ely. The p alue o HCA s. HA was 0.05 (Fishe ’s exac es ). h Fo compa ison wi h hospi al-acqui ed episodes, p 0.04 (Fishe ’s exac es ). 3 4Clinical Mic obiology and In ec ion CMI ª2010 The Au ho s Jou nal Compila ion ª2010 Eu opean Socie y o Clinical Mic obiology and In ec ious Diseases, CMI 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 Twen y- ou pe cen o BSI in ec ions epo ed in his mul icen e s udy including public e ia y-ca e and commu- ni y hospi als in Andalucı ´a, Spain, we e o he HCA ype. This is a simila igu e o ha epo ed by Valle ´se al. [8] in hei s udy pe o med in wo eaching hospi als in Ba celona (24.5%), and lowe han ha epo ed by F iedman e al. [5] in ou hospi als in No h Ca olina (37%), using simila de i- ni ions. Mos o ou HCA BSI pa ien s had ecen ly been admi ed o an acu e ca e hospi al, and some we e esiden s in a long- e m ca e acili y o ecei ing haemodialysis. The equency o BSI in pa ien s wi h ambula o y in a enous ea men o specialized home ca e was low, whe eas i was much highe in No h Ca olina [5]. This p obably e lec s signi ican di e ences be ween he Spanish and US heal hca e sys ems in luencing he epidemiology o BSI. We ound ha he pe cen age o HCA episodes was simila in communi y and e ia y-ca e cen es, indica ing ha his is a ca ego y o BSI ha should be aken in o accoun in bo h ypes o hospi als. The da a ob ained in he p esen s udy con i m ha HCA and CA BSIs show signi ican di e ences ha physicians a ending pa ien s in he eme gency oom should bea in mind when a ending pa ien s wi h communi y-onse sepsis [5–9,17]. Beyond he ea u es included in he de ini ion o HCA BSI, pa ien s in his g oup mo e equen ly showed a al unde lying condi ions, ch onic enal insu iciency, haema- ological cance , p e ious an imic obial ea men and an unknown sou ce o in ec ion. Mo e impo an ly, some ypical heal hca e-associa ed an ibio ic- esis an o ganisms, such as P. ae uginosa o me hicillin- esis an S. au eus (MRSA), we e no p esen o a e in CA episodes bu should be consid- e ed in pa ien s wi h a suspec ed HCA BSI (CA MRSA is s ill anecdo al in Spain) [18]. Cu iously, o he an ibio ic- esis an o ganisms (such as ESBL p oduce s among E. coli o K. pneu- moniae) we e ound a simila equencies in CA, HCA and HA episodes, e lec ing he ac ha ESBL-p oducing en e o- bac e ia a e now signi ican causes o CA BSI [10] and ha he c i e ia used o de ine an HCA episode a e no use ul o uling ou hese o ganisms in pa ien s wi h communi y- onse BSI. As a as we a e awa e, he cha ac e is ics o pa ien s wi h HCA BSI episodes by ype o hospi al ha e no p e iously been s udied. Signi ican di e ences we e ound be ween HCA episodes occu ing in e ia y-ca e and communi y cen- es, e lec ing he ype o ou pa ien s ca ed o in bo h ypes o cen es. Thus, haemodialysis and unnelled ca he e s we e mo e equen in e ia y-ca e hospi als, whe eas pa ien s wi h HCA BSI in communi y hospi als we e mainly ecen ly admi ed pa ien s wi h ch onic condi ions o pa ien s wi h solid cance ecei ing in a enous chemo- he apy in a day hospi al. The e was a highe c ude mo ali y a e o HCA BSI pa ien s ea ed in communi y cen es compa ed o e ia y-ca e hospi als, al hough he esul s o he mul i a ia e analysis sugges ha he c ude associa ion was con ounded by o he a iables. Al hough benchma k a es exis s o HA BSI and o al BSI [1–3], we ha e p o ided incidence a es o HCA BSI. Because he mos app op ia e denomina o o HCA BSI (i.e. he numbe o pa ien s o pa ien -days a isk) is e y di i- cul o ob ain, we op ed o p o ide bo h popula ion-based and admission-based a es. We conside popula ion-based a es o be mo e app op ia e because no e e y pa ien wi h HCA BSI is admi ed o hospi al. Because hospi als pa icipa - ing in he s udy a end >90% o pa ien s in hei ca chmen a ea needing hospi aliza ion, he popula ion-based a es es i- ma ed in he p esen s udy should be conside ed as mini- mum a es. The p esen s udy has se e al limi a ions. The numbe o cases om communi y hospi als was lowe compa ed o ha om e ia y-ca e cen es, which limi s he compa isons be ween he wo ypes o hospi al. Also, he s udy pe iod was longe in he communi y hospi als and, because commu- ni y-onse BSI may show seasonal di e ences, his may ha e in luenced he esul s. Howe e , he s udy pe iods we e chosen aiming o minimize he impac o he di e en pa ame e s. In addi ion, e en hough blood cul u es we e p ocessed in all cen es using s anda dized p o ocols [12,13], he e may ha e been some di e ences among he hospi als. Finally, he esul s ob ained a e p obably no applicable o a eas wi h di e en a epidemiology o esis an bac e ia (such as endemic CA MRSA) and di e en heal hca e sys ems. In summa y, HCA BSI should be conside ed as a dis inc class o BSI, in bo h e ia y-ca e and communi y hospi als, al hough di e ences in he p edisposing condi ions o he pa ien s and clinical ea u es should be aken in o accoun . Acknowledgemen s The pa icipan s om he SAEI/SAMPAC Bac e aemia G oup we e: F. Rod ı ´guez (Hospi al Uni e si a io Reina So ı ´a, Co ´ - doba), Ma ina de Cue o (Hospi al Uni e si a io Vi gen Maca- ena, Se illa), Ma ı ´a V. Ga cı ´a (Hospi al Uni e si a io Vi gen de la Vic o ia, Ma ´laga), Ve o ´nica Gonza ´lez-Gala ´n (Hospi ales Uni e si a ios Vi gen del Rocı ´o, Se illa), Fe nando Fe na ´n- dez-Sa ´nchez (Hospi al Cos a del Sol, Ma bella, Ma ´laga), Ma ı ´a J. Gu ie ´ ez (Hospi al de la Se anı ´a, Ronda, Ma ´laga), An onio Sa ´nchez-Po o (Hospi al de La Lı ´nea, La Lı ´nea de la Concep- cio ´n, Ca ´diz), Be a Bece il (Hospi al Pun a de Eu opa, Algeci as, Ca ´diz), Ana Ga cı ´a-Tapia (Hospi al Pue a del Ma , CMI Rod ı ´guez-Ban ˜oe al. Bac e aemia in e ia y-ca e and communi y hospi als 5 ª2010 The Au ho s Jou nal Compila ion ª2010 Eu opean Socie y o Clinical Mic obiology and In ec ious Diseases, CMI 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 Ca ´diz), Juan C. Alados (Hospi al de Je ez, Je ez de la F on e a, Ca ´diz), Fede ico Acos a (Hospi al de An eque a, Ma ´laga), Ca men Flo ez (Hospi al Uni e si a io de Valme, Se illa), Pe a Na as (Hospi al To eca ´ denas, Alme ı ´a), Ma ı ´a A. Ma ı ´nez-Pe ´ ez (Hospi al Uni e si a io San Cecilio, G anada), Inmaculada Ca azo (Hospi al de Jae ´n, Jae ´n). Au ho con ibu ions JRB and MDLP we e he coo dina o s o he s udy. JRB, MDLP, MMP and PR a e esponsible o he s udy design, which was ac i ely discussed wi h all au ho s and modi ied acco dingly. MMP, PR, CN, EN, MH, AA, AM, FT, MTT, AMA, AA, AR, RM, JEC, LL and JAPL coo dina ed he me h- odology and da a collec ion in each hospi al and e i ied he da abases. JRB, MMP and PR pe o med he p elimina y anal- ysis and w o e he d a o he manusc ip , which was dis- cussed wi h all au ho s and modi ied acco dingly. T anspa ency Decla a ion This s udy was unded by Conseje ı ´a de Salud, Jun a de And- alucı ´a (0063/2006 and PI0048/2008), Minis e io de Sanidad y Consumo, Ins i u o de Salud Ca los III-FEDER, Spanish Ne - wo k o he Resea ch in In ec ious Diseases (REIPI RD06/ 0008) and FIS (PI070190). All au ho s decla e ha he e a e no con lic s o in e es . Re e ences 1. Weins ein MP, Towns ML, Qua e y SM e al. The clinical signi icance o posi i e blood cul u es in he 1990s: a p ospec i e comp ehensi e e alua ion o he mic obiology, epidemiology, and ou come o bac e - emia and ungemia in adul s. Clin In ec Dis 1997; 24: 584–602. 2. Wisplingho H, Bischo T, Tallen SM e al. Nosocomial bloods eam in ec ions in US hospi als: analysis o 24,179 cases om a p ospec i e na ionwide su eillance s udy. Clin In ec Dis 2004; 39: 309–317. 3. Rod ı ´guez-C eixems M, Alcala ´L, Mun ˜oz P e al. Bloods eam in ec- ions. E olu ion and ends in he mic obiology wo kload, incidence and e iology, 1985–2006. Medicine 2008; 87: 234–249. 4. Ga ne JS, Ja is WR, Emo i TG e al. CDC de ini ions o nosoco- mial in ec ions. Am J In ec Con ol 1988; 16: 128–140. 5. F iedman ND, Kaye KS, S ou JE e al. Heal h ca e-associa ed blood- s eam in ec ions in adul s; a eason o change he accep ed de ini- ion o communi y-acqui ed in ec ions. Ann In e n Med 2002; 137: 791–797. 6. Siegman-Ig a Y, Fou e B, O ni-Wassenlau R e al. Reapp aisal o communi y-acqui ed bac e emia: a p oposal o a new classi ica ion o he spec um o acquisi ion o bac e emia. Clin In ec Dis 2002; 34: 1431–1439. 7. Sho AF, Tabak YP, Killiam AD, Gup a V, Liu LZ, Kolle MH. Heal h- ca e-associa ed bloods eam in ec ion: a dis inc en i y? Insigh s om a la ge U.S. da abase C i Ca e Med 2006; 34: 1588–1595. 8. Valles J, Calbo E, Ano o A e al. Bloods eam in ec ions in adul s: impo ance o heal hca e-associa ed in ec ions. J In ec 2008; 56: 27– 34. 9. Raymond NJ, Blackmo e TK, Humble MW e al. Bloods eam in ec- ions in a seconda y and e ia y ca e hospi al se ing. In e n Med J 2006; 36: 765–772. 10. Rod ı ´guez-Ban ˜o J, Na a o MD, Rome o L e al. Bac e emia due o ex ended-spec um be a-lac amase-p oducing Esche ichia coli in he CTX-M e a: a new clinical challenge. Clin In ec Dis 2006; 43: 1407– 1414. 11. Es adı ´s ica de es ablecimien os sani a ios con e ´gimen de in e nado. Agencia de Calidad del Sis ema Nacional de Salud, Ins i u o de In o macio ´n Sani a ia, Minis e io de Sanidad y Consumo 2005. A ail- able a : h p://www.msc.es/es adEs udios/es adis icas/docs/es Hosp05/ publicacionESCRI2005.pd 1 . 12. Loza Fe na ´ndez de Bobadilla E, Planes Reig A, Rod ı ´guez-C eixems M. Hemocul i os. In: P ocedimien os en Mic obiologı ´a Clı ´nica. Socie- dad Espan ˜ola de En e medades In ecciosas y Mic obiologı ´a Clı ´nica, 2003. A ailable a : h p://www.seimc.o g/documen os/p o ocolos/mi- c obiologia. Accessed 15 June 2009. 13. Clinical and Labo a o y S anda ds Ins i u e 2005. Pe o mance s an- da ds o an imic obial suscep ibili y es ing. 15 h in o ma ional supple- men . App o ed s anda d M100-S15. Wayne, PA: CLSI, 2005. 14. McCabe WR, Jackson GG. G am-nega i e bac e emia I. E iology and ecology. A ch In e n Med 1962; 110: 847–855. 15. Se icio Andaluz de Salud. Memo ia 2006. Se ille, Spain: Conseje ı ´a de Salud, Jun a Andalucı ´a, 2007. A ailable a : h p://www.sas.jun a- andalucia.es/publicaciones/lis ado.asp?ma e =7. 16. McDonald JR, F iedman ND, S ou JE e al. Risk ac o s o ine ec i e he apy in pa ien s wi h bloods eam in ec ion. A ch In e n Med 2005; 165: 308–313. 17. Cisne os-He e os JM, Cobo-Reinoso J, Pujol-Rojo M e al. Guı ´a pa a el diagno ´s ico y a amien o del pacien e con bac e iemia. Guı ´as de la Sociedad Espan ˜ola de En e medades In ecciosas y Mic obiologı ´a Clı ´nica (SEIMC). En e m In ecc Mic obiol Clin 2007; 25: 111–130. 18. Rod ı ´guez-Ban ˜o J, Domı ´nguez MA, Milla ´nAe al. Clinical and molecu- la epidemiology o communi y, heal h ca e-associa ed and nosoco- mial me hicillin- esis an S aphylococus au eus in Spain. Clin Mic obiol In ec 2009; 15: 1111–1118. 6Clinical Mic obiology and In ec ion CMI ª2010 The Au ho s Jou nal Compila ion ª2010 Eu opean Socie y o Clinical Mic obiology and In ec ious Diseases, CMI 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 Au ho Que y Fo m Jou nal: CLM A icle: 3089 Dea Au ho , Du ing he copy-edi ing o you pape , he ollowing que ies a ose. Please espond o hese by ma king up you p oo s wi h he necessa y changes/addi ions. Please w i e you answe s on he que y shee i he e is insu icien space on he page p oo s. Please w i e clea ly and ollow he con en ions shown on he a ached co ec ions shee . I e u ning he p oo by ax do no w i e oo close o he pape ’s edge. Please emembe ha illegible ma k-ups may delay publica ion. Many hanks o you assis ance. Que y e e ence Que y Rema ks Q1 AUTHOR: Please check his websi e add ess and con i m ha i is co ec . (Please no e ha i is he esponsibili y o he au ho (s) o ensu e ha all URLs gi en in his a icle a e co ec and useable.) Q2 AUTHOR: some % alues ha e been co ec ed: HCA, 17, 17 and 2, HA, 68 – please e i y. Q3 AUTHOR: some % alues ha e been co ec ed: CA, 57; HCA, 64 and HA, 5 – please e i y. MARKED PROOF Please co ec and e u n his se Ins uc ion o p in e Lea e unchanged unde ma e o emain h ough single cha ac e , ule o unde line New ma e ollowed by o o o o o o o o o and/o and/o e.g. e.g. unde cha ac e o e cha ac e new cha ac e new cha ac e s h ough all cha ac e s o be dele ed h ough le e o h ough cha ac e s unde ma e o be changed unde ma e o be changed unde ma e o be changed unde ma e o be changed unde ma e o be changed Enci cle ma e o be changed (As abo e) (As abo e) (As abo e) (As abo e) (As abo e) (As abo e) (As abo e) (As abo e) linking cha ac e s h ough cha ac e o whe e equi ed be ween cha ac e s o wo ds a ec ed h ough cha ac e o whe e equi ed o indica ed in he ma gin Dele e Subs i u e cha ac e o subs i u e pa o one o mo e wo d(s) Change o i alics Change o capi als Change o small capi als Change o bold ype Change o bold i alic Change o lowe case Change i alic o up igh ype Change bold o non-bold ype Inse ‘supe io ’ cha ac e Inse ‘in e io ’ cha ac e Inse ull s op Inse comma Inse single quo a ion ma ks Inse double quo a ion ma ks Inse hyphen S a new pa ag aph No new pa ag aph T anspose Close up Inse o subs i u e space be ween cha ac e s o wo ds Reduce space be ween cha ac e s o wo ds Inse in ex he ma e Tex ual ma k Ma ginal ma k Please use he p oo co ec ion ma ks shown below o all al e a ions and co ec ions. I you in da k ink and a e made well wi hin he page ma gins. wish o e u n you p oo by ax you should ensu e ha all amendmen s a e w i en clea ly