Sepsis-related mortality in 497 cases with blood culture-positive sepsis in an emergency department

Sepsis- ela ed mo ali y in 497 cases wi h blood cul u e-posi i e sepsis in an eme gency depa men Juha Rannikko a, *, Jaana Sy jänen a , Tapio Seiska i b , Janne Ai oniemi b , Ree a Hu unen a a Depa men o In e nal Medicine, Tampe e Uni e si y Hospi al, Box 2000, FI-33521 Tampe e, Finland b Depa men o Clinical Mic obiology, Fimlab Labo a o ies, Tampe e, Finland A R T I C L E I N F O A icle his o y: Recei ed 14 Decembe 2016 Recei ed in e ised o m 15 Feb ua y 2017 Accep ed 5 Ma ch 2017 Co esponding Edi o : Eskild Pe e sen, Aa hus, Denma k Keywo ds: A ibu able mo ali y Blood cul u e In ec ious disease Rela ed mo ali y Sepsis Su i al qSOFA S U M M A R Y Objec i e: Few s udies ha e sough o es ablish how o en dea h a e sepsis is ela ed o he sepsis and how o en unde lying diseases ha e a majo ole in case a ali y. Me hods: In his e ospec i e coho s udy, da a we e collec ed on 497 cases wi h blood cul u e-posi i e sepsis in an eme gency depa men (ED). Resul s: Sepsis was ca ego ized as se e e in 31% o cases; 7% had sep ic shock. The quick Sepsis- ela ed O gan Failu e Assessmen sco e was posi i e in 136 ou o 473 cases (29%). Nine y-eigh pa ien s died by day 90; in 16 o hese cases (16%) he dea h was sepsis- ela ed in a pa ien wi hou a apidly a al unde lying disease, in 45 cases (46%) he dea h was sepsis- ela ed in a pa ien wi h a apidly a al unde lying disease, and in 37 cases (38%) he dea h was un ela ed o sepsis. Sepsis- ela ed dea h occu ed in 58 ou o 61 cases (95%) by day 28. Conclusions: Unde lying diseases we e ound o ha e a conside able ole in he dea h o pa ien s su e ing om blood cul u e-posi i e sepsis in an ED o a de eloped coun y, as only 16% o he dea hs by day 90 occu ed whe e dea h was sepsis- ela ed and he pa ien had a li e-expec ancy o mo e han 6 mon hs. Imp o ing he ou come o sepsis wi h new ea men s is hus challenging. I is possible ha day 7 + day 28 mo ali y is a mo e app op ia e endpoin han day 90 mo ali y when s udying he ou come o sepsis, as his ime-span includes mos o he pa ien s whose dea h was ela ed o sepsis. © 2017 The Au ho s. Published by Else ie L d on behal o In e na ional Socie y o In ec ious Diseases. This is an open access a icle unde he CC BY-NC-ND license (h p://c ea i ecommons.o g/licenses/by- nc-nd/4.0/). In oduc ion Sepsis is one o he leading causes o dea h wo ldwide. I has been anked as he ele en h mos common cause o dea h in he USA. 1 Ad anced age, immunosupp ession, diabe es, and cance a e majo isk ac o s o sepsis. 2–6 P ognos ic ac o s o he se e i y and ou come o sepsis include ad anced age, ype o in ec ion (e.g., me hicillin- esis an S aphylococcus au eus (MRSA), polymic o- bial), numbe o o gan dys unc ions, and adequacy o an imic obial he apy. 7–9 S udies ha e been pe o med on sepsis pa ien s in eme gency depa men s (EDs) and in ensi e ca e uni s (ICUs) in o de o de e mine isk ac o s o mo ali y 9–13 and he ae iology o illness. 14 Less a en ion has been paid o he ques ions o how o en hese pa ien s ac ually die o sepsis, how o en sepsis is a con ibu o y ac o in he dea h o a pa ien wi h an ad anced unde lying disease, and how o en he dea h is independen o sepsis. In his e ospec i e coho s udy, da a we e collec ed on 497 adul cases o blood cul u e-posi i e sepsis in he ED o Tampe e Uni e si y Hospi al (TAUH). A ca ego iza ion o causes o dea h was de eloped in o de o es ablish how o en dea h was ela ed o sepsis in pa ien s wi hou a apidly a al unde lying disease (g oup 1), was ela ed o sepsis by weakening o a pa ien wi h a apidly a al unde lying disease (g oup 2), and was independen o sepsis bu caused by he unde lying disease (g oup 3). I was also sough o de e mine he bes cu -o among he commonly used days o mo ali y used in sepsis esea ch, i.e. day 7, day 28, o day 90, o mo ali y ela ed o sepsis (dea hs in g oups 1 and 2). Ma e ials and me hods TAUH is a e ia y hospi al wi h a ca chmen popula ion o app oxima ely 524 700 inhabi an s in Pi kanmaa Coun y. The ED o * Co esponding au ho . Tel: +358 3 31166747, Fax: +358 3 31164333. E-mail add ess: [email p o ec ed] (J. Rannikko). h p://dx.doi.o g/10.1016/j.ijid.2017.03.005 1201-9712/© 2017 The Au ho s. Published by Else ie L d on behal o In e na ional Socie y o In ec ious Diseases. This is an open access a icle unde he CC BY-NC-ND license (h p://c ea i ecommons.o g/licenses/by-nc-nd/4.0/). In e na ional Jou nal o In ec ious Diseases 58 (2017) 52–57 Con en s lis s a ailable a ScienceDi ec In e na ional Jou nal o In ec ious Diseases jou nal homepage: www.else ie .com/loca e/ijid he hospi al handles pa ien s equi ing bo h basic and specialized eme gency ca e. In specialized ca e, he majo i y o pa ien s a e in e nal medicine and su gical pa ien s. Blood cul u es a e aken ou inely om pa ien s wi h signs o symp oms o sys emic in ec ion. The popula ion o he p esen s udy comp ised 497 adul pa ien s admi ed o he ED o TAUH and ea ed in specialized ca e, who had blood cul u e-posi i e sepsis du ing he pe iod Ma ch 1, 2012 o Feb ua y 28, 2014. The s udy was app o ed by he E hics Commi ee o Tampe e Uni e si y Hospi al. The need o in o med consen was wai ed, as no addi ional blood sampling was needed and ou ine pa ien ca e was no modified. TAUH has a 24-bed ICU which includes a se en-bed high dependency uni (HDU). In his a icle, ‘ICU’ e e s o bo h o hese. The e a e ou o he HDUs (ca diology, pulmona y, su gical, and in e nal medicine), all aking ca e o sepsis pa ien s; hese a e e e ed o as ‘HDUs’. Blood cul u es we e collec ed in BacT/Ale Ae obic (FA Plus) and Anae obic (FN Plus) blood cul u e bo les and placed in he au oma ed mic obial de ec ion sys em BacT/Ale 3D (bioMé ieux, Ma cy l’E oile, F ance). All pa ien s wi h a posi i e blood cul u e ob ained du ing specialized ca e in he ED we e iden ified in he mic obiology labo a o y se ing TAUH (Fimlab Labo a o ies plc). Pa ien de ails, he name o he o ganism, and he da e o blood cul u e we e collec ed by clinical mic obiologis s (T.S. and J.A.). Cul u es posi i e o coagulase-nega i e S aphylococcus, P opio- nibac e ium, Mic ococcus, Bacillus, and Co ynebac e ium, wi h de ec ion in a single blood cul u e bo le and wi hou clinical ele ance, we e conside ed o be con aminan s and we e excluded. Pa ien s whose ou ine blood samples aken on admission we e no longe a ailable o u he s udies we e also excluded. This could be due, o example, o he ac ha he blood cul u e became posi i e la e han 72 h a e he day o admission. The clinical da a o he pa ien s included in he s udy we e ga he ed e ospec i ely om he pa ien eco ds by he p incipal in es iga o (J.R.). The si e o in ec ion was decided e ospec i ely by he p incipal in es iga o based on clinical judgemen . Sepsis was deemed o be heal hca e-associa ed i he symp oms had s a ed mo e han 48 h a e admission o a heal hca e ins i u ion, o he bac e aemia was ela ed o a su gical ope a ion wi hin he p eceding 30 days o some o he in asi e p ocedu e wi hin he p e ious 10 days. 15 Da a on cause o dea h we e ga he ed om pa ien eco ds (and au opsy eco ds when applicable) by wo clinicians (J.R. and R.H.) independen ly. In cases o disc epancy, a mee ing was held oge he wi h a hi d clinician (J.S.) and a final decision was made. Cause o dea h by day 90 was classified in o h ee di e en ca ego ies: (1) g oup 1 included cases o sepsis- ela ed mo ali y in pa ien s wi hou a apidly a al unde lying disease. The immedia e cause o dea h in his g oup was sepsis, o sepsis was a ac o in a chain o e en s leading o dea h, and he pa ien had a li e- expec ancy o mo e han 6 mon hs. (2) G oup 2 included cases o sepsis- ela ed mo ali y in pa ien s wi h a apidly a al unde lying disease. In his g oup, he pa ien died o sepsis (immedia e cause o dea h, o sepsis was a ac o in a chain o e en s leading o dea h) by weakening o a pa ien wi h a apidly a al (<6 mon hs) unde lying disease. (3) G oup 3 included cases o mo ali y ela ed o unde lying disease. In his g oup, sepsis was no an immedia e cause o dea h o a ac o in a chain o e en s leading o dea h. The ca ego iza ion was based on clinical decision and he judgemen was based on he se e i y o he pa ien ’s unde lying disease, he pa ien ’s p e-pe o mance, se e i y o he sepsis, and eco e y a e he in ec ion. The main unde lying disease associa ed wi h dea h was de e mined e ospec i ely by he p incipal in es iga o . Diagnoses o sepsis, se e e sepsis, and sep ic shock we e made acco ding o consensus defini ions. 16 Fu he , he quick Sepsis- ela ed O gan Failu e Assessmen sco e (qSOFA) was calcula ed pos -hoc acco ding o ecen ly published defini ions. 17 The qSOFA sco e was posi i e i a leas wo o he ollowing h ee c i e ia we e ulfilled in he ED: espi a o y a e 22/min, al e ed men a ion, and sys olic blood p essu e 100 mmHg. The McCabe classifica ion was de e mined as epo ed by McCabe and Jackson. 18 The Pi bac e aemia sco e was calcula ed as p esen ed by Ko ick e al. 19 IBM SPSS e sion 22.0 so wa e (IBM Co p., A monk, NY, USA) was used o he s a is ical analyses. Ca ego ical da a we e analyzed by Chi-squa e es , o Fishe ’s exac es when app op ia e. Odds a ios (OR) wi h 95% confidence in e als (95% CI) a e also p esen ed. Resul s The e we e 800 consecu i e posi i e blood cul u es in adul pa ien s du ing he s udy pe iod. One hund ed and hi y-six o hese we e conside ed o be con aminan s and 167 we e excluded o o he easons. A o al 497 cases o posi i e blood cul u e among 484 pa ien s we e hus included. All 497 cases had sepsis. Du ing he s udy pe iod, 11 pa ien s had sepsis wice on di e en admissions and one pa ien had sepsis h ee imes. O he o al s udy popula ion, 262 (53%) we e male and 235 (47%) we e emale; hey anged in age om 16 o 95 yea s (median 68 yea s). Table 1 p o ides he demog aphic da a, da a on he causa i e o ganisms, and da a on he unde lying diseases s a ified in o six di e en ca ego ies: g oups 1, 2, and 3 (as no ed in he Ma e ials and me hods sec ion), all pa ien s who died, all pa ien s who su i ed, and all cases. Fo 16 ou o 98 pa ien s (16%) who died by day 90, dea h was ela ed o sepsis and he pa ien did no ha e a apidly a al unde lying disease (g oup 1), i.e. 3% o all 497 sepsis cases. Fo 45 pa ien s (46% o all dea hs by day 90), dea h was ela ed o sepsis in ha i weakened he pa ien leading o he dea h, which was in any case expec ed as he pa ien had a apidly a al unde lying disease (g oup 2). Fo 37 pa ien s (38%), dea h by day 90 was un ela ed o sepsis and was caused by an unde lying disease(s) in he pa ien (g oup 3). O he g oup 1 pa ien s, ou (25%) had alcohol abuse as an unde lying disease (Table 1). One was wi hou any unde lying disease. Fou o he pa ien s in g oup 1 we e o e 80 yea s o age (25%). Th ee g oup 1 pa ien s (19%) we e ound lying a home wi h a low le el o consciousness and we e ans e ed o hospi al. In g oup 2 pa ien s, he apidly a al unde lying disease was a solid umou wi h me as asis in eigh cases (18%) and a haema ological malignancy in nine (20%). In g oup 3 pa ien s, he cause o dea h was a solid umou wi h me as asis in 19 pa ien s (51%) and a haema ological malignancy in wo (5%). Thus, malignancies we e associa ed wi h dea h in 46% o pa ien s in g oups 2 and 3 combined. O he common apidly a al unde lying diseases associa ed wi h dea h among he pa ien s in g oups 2 and 3 we e li e disease (11%), hea disease (11%), and neu ological/neu o- su gical disease (10%). The case a ali y a e by day 7, day 28, and day 90 was 9%, 14%, and 20%, espec i ely. Dea h occu ed by day 7 in 94% o g oup 1 pa ien s, in 56% o g oup 2 pa ien s, and in 11% o g oup 3 pa ien s (Figu e 1). All excep wo cases in g oup 2 died be o e day 28. The dea hs in g oup 3 occu ed mos o en be ween day 29 and day 90 (68%). Nine y-fi e pe cen o all sepsis- ela ed dea hs occu ed wi hin 28 days a e sepsis. Table 2 gi es da a on he se e i y o sepsis s a ified in o he same ca ego ies as used in Table 1. The qSOFA was posi i e in 136 ou o 473 cases (29%). G oup 1 had he highes Pi bac e aemia sco es. Fo y-eigh (10%) cases we e ans e ed om he ED o he ICU and 52 (11%) o HDUs. Thus, he majo i y o sepsis pa ien s and he majo i y o qSOFA-posi i e cases we e aken ca e o in gene al wa ds. Ou o 449 cases who we e ea ed ou side he ICU, 104 J. Rannikko e al. / In e na ional Jou nal o In ec ious Diseases 58 (2017) 52–57 53 (23%) had se e e sepsis and nine (2%) had sep ic shock. O hose wi h sep ic shock, h ee died in he ED and he emainde had unde lying diseases ha had p og essed oo a o hem o benefi om ea men in he ICU. By si e o in ec ion, he mos common in ec ion was u ina y ac in ec ion (131 cases, 26%), ollowed by in a-abdominal in ec ion (84 cases, 17%), in ec ion o he skin, so issues, and bones (72 cases, 15%), and lowe espi a o y ac in ec ion (54 cases, 11%). The si e o in ec ion was unknown in 118 cases Table 1 Cha ac e is ics unde lying diseases, and causa i e o ganisms in ela ion o he ou come. Sepsis- ela ed mo ali y in pa ien s wi hou a apidly a al unde lying disease (G oup 1) (n = 16) Sepsis- ela ed mo ali y in pa ien s wi h a apidly a al unde lying disease (G oup 2) (n = 45) Mo ali y ela ed o unde lying disease (G oup 3) (n = 37) All pa ien s who died (n = 98) Pa ien s who su i ed (n = 399) All cases (N = 497) n (%) n (%) n (%) n (%) n (%) n (%) Median age, yea s ( ange) 77 (51–94) 78 (42–93) 70 (44–91) 74 (42–94) 68 (16–95) 68 (16–95) Male sex 8 (50) 24 (53) 25 (68) 57 (58) 205 (51) 262 (53) Unde lying diseases Hea disease 8 (50) 23 (51) 10 (27) 41 (42) 135 (34) 176 (35) Diabe es melli us, any ype 6 (38) 15 (33) 13 (35) 34 (34) 105 (26) 139 (28) Neu ological 5 (31) 14 (31) 9 (24) 28 (29) 84 (21) 112 (23) Solid umou wi h me as asis a 1 (6) 10 (22) 17 (46) 28 (29) 29 (7) 57 (12) Alcohol abuse b 4 (25) 5 (11) 8 (22) 17 (17) 35 (9) 52 (11) Li e disease 3 (19) 8 (18) 6 (16) 17 (17) 32 (8) 49 (10) Haema ological malignancy 0 (0) 11 (24) 4 (11) 15 (15) 30 (8) 45 (9) No unde lying diseases 1 (6) 0 (0) 0 (0) 1(6) 70 (14) 71 (14) Causa i e o ganism c G am-posi i e 5 (31) 18 (40) 13 (35) 36 (37) 189 (47) 225 (45) G am-nega i e 9 (56) 19 (42) 14 (38) 42 (43) 181 (45) 223 (45) Esche ichia coli u osepsis 3 (19) 6 (13) 3 (8) 12 (12) 85 (21) 97 (20) Polymic obial and anae obes 2 (13) 8 (18) 10 (27) 20 (20) 28 (7) 48 (10) a The mos common solid umou wi h me as asis was cance o he gas oin es inal ac (36 cases). b Social o medical p oblems o alcohol abuse in he pas 12 mon hs. c Excluding ungi (n = 1). N=15 94% N=0 0% N=1 6% N=25 56% N=18 40% N=2 4% N=4 11% N=8 22% N=25 68% 0 5 10 15 20 25 30 Days 0-7 Days 8- 28 Days 29 -90 Sepsis- ela ed mo al i y in pa ie n s wi hou a api dly a al un de lyi ng diseas e (G oup 1) Sepsis- ela ed mo al i y in pa ie n s wi h a api dly a al un de lyi ng diseas e (G oup 2) Unde lyin g disease – ela ed mo ali y (G oup 3) Figu e 1. Numbe o cases in g oups 1, 2, and 3 in ela ion o day o case a ali y in a o al o 98 pa ien s who died wi hin 90 days a e sepsis. The pe cen age is pe ca ego y o cause o dea h. 54 J. Rannikko e al. / In e na ional Jou nal o In ec ious Diseases 58 (2017) 52–57 (24%). Sepsis was heal hca e-associa ed in 81 pa ien s (16%). The p e ious use o immunosupp essi e medica ions was common: 94 (19%) pa ien s we e aking co icos e oids (any dose) and 56 (11%) we e on cance chemo he apy. An ibio ics we e s a ed on he day o admission in 93% o pa ien s. The mos common an ibio ic was ce u oxime (57%), ollowed by ce iaxone (18%) and fluo oquinolone (11%). An an ibio ic combina ion was s a ed in 16% o pa ien s. The median ime om admission o he s a o an ibio ics was 174 min ( ange 0–1269 min). O all cases, 426 (86%) had one o mo e unde lying diseases. As shown in Table 1, he e we e significan ly mo e alcohol abuse s in g oup 1 as compa ed o hose who su i ed o mo e han 90 days (4/16 s. 35/399; OR 3.8, 95% CI 1.2–12.4, p = 0.04). The e we e also significan ly mo e pa ien s wi h a solid umou wi h me as asis and pa ien s wi h haema ological malignancies among hose who died by day 90 han among su i o s (28/98 s. 29/399, OR 5.1, 95% CI 2.9–9.1, p < 0.01 and 15/98 s. 30/399, OR 2.2, 95% CI 1.1–4.3, p = 0.02, espec i ely). The causa i e o ganisms a e lis ed in Table 3. G am-posi i e and G am-nega i e o ganisms we e equally common. The e we e ew cases wi h MRSA and ex ended-spec um be a-lac amase-p oduc- ing En e obac e iaceae and no cases wi h ancomycin- esis an en e ococci o ca bapenemase-p oducing En e obac e iaceae. Discussion In his s udy, he case a ali y a e by day 7, day 28, and day 90 was 9%, 14%, and 20%, espec i ely, in 497 cases wi h blood cul u e-posi i e sepsis ea ed in an ED. The day 7 and day 28 case a ali y a es in his s udy a e simila o hose epo ed in a s udy by Lin e al., who also s udied bac e aemic pa ien s in he ED. 9 In wo o he s udies in es iga ing he day 28 case a ali y a e in bac e aemic pa ien s in he ED, Lee e al. epo ed a a e o 9% and Kao e al. epo ed a a e o 19%. 12,20 Only 16% o all dea hs by day 90 in he p esen s udy we e ela ed o sepsis in pa ien s wi h a li e-expec ancy o mo e han 6 mon hs (g oup 1). Fo y-six pe cen o dea hs occu ed in pa ien s wi h a apidly a al unde lying disease al hough dea h was sepsis- ela ed (g oup 2), and 38% o dea hs we e ela ed o he unde lying disease (g oup 3). I is known ha unde lying diseases ha e a ole in he dea hs o sepsis pa ien s, 2–4,9 bu he ca ego iza ion o he p esen s udy gi es a pic u e o he significance o he unde lying diseases in hese dea hs. E en pa ien s in g oup 1 had many ac o s con ibu ing o dea h, o example alcoholism o e y old age. The e was only one pa ien in his g oup wi hou any unde lying disease. I is possible ha he e we e pa ien s in g oup 1 who sough ea men oo la e, as h ee pa ien s in his g oup we e ound lying a home wi h a low le el o consciousness. The e we e also significan ly mo e alcohol abuse s as compa ed o hose who su i ed mo e han 90 days. Al oge he , in his coho consis ing mos ly o communi y- acqui ed cases seen in a uni e si y hospi al ED in a de eloped coun y, he e we e ew non-su i o s who would ha e had a li e- expec ancy o mo e han 6 mon hs and who migh ha e benefi ed om a no el d ug di ec ed agains sepsis o sep ic shock. This may also explain why so many clinical ials in he field o sepsis ha e ailed. 21 Day 90 mo ali y has been used as he endpoin in many sepsis s udies. 22–24 Howe e , he p esen s udy showed ha mos (95%) o he pa ien s whose dea h was sepsis- ela ed died wi hin 28 days a e sepsis, while he majo i y o dea hs occu ing be ween day 28 Table 2 Se e i y o he sepsis in ela ion o he ou come. Sepsis- ela ed mo ali y in pa ien s wi hou a apidly a al unde lying disease (G oup 1) (n = 16) Sepsis- ela ed mo ali y in pa ien s wi h a apidly a al unde lying disease (G oup 2) (n = 45) Mo ali y ela ed o unde lying disease (G oup 3) (n = 37) All pa ien s who died (n = 98) Pa ien s who su i ed (n = 399) All cases (N = 497) n (%) n (%) n (%) n (%) n (%) n (%) Se e e sepsis 16 (100) 30 (67) 13 (35) 59 (60) 93 (23) 152 (31) Sep ic shock 9 (56) 9 (20) 4 (11) 22 (22) 15 (4) 37 (7) qSOFA-posi i e a 13 (81) 26 (59) 16 (46) 55 (58) 81 (21) 136 (29) Admi ed om ED o ICU 11 (69) 7 (16) 3 (8) 21 (21) 27 (7) 48 (10) Pi bac e aemia sco e b 0–1 1 (6) 16 (36) 20 (57) 37 (39) 269 (68) 306 (62) 2–3 3 (19) 17 (39) 10 (29) 30 (32) 93 (24) 123 (25) 4 12 (75) 11 (25) 5 (14) 28 (29) 33 (8) 61 (12) ED, eme gency depa men ; ICU, in ensi e ca e uni ; qSOFA, quick sepsis- ela ed o gan ailu e assessmen . a Da a a ailable o 473 cases (95 who died and 378 who su i ed). b Da a a ailable o 490 cases. Table 3 Causa i e o ganisms. O ganisms n (%) G am-posi i e 225 (45.3) S aphylococcus au eus 74 (14.9) MRSA 3 (0.6) Coagulase-nega i e S aphylococcus 11 (2.2) S ep ococcus pneumoniae 52 (10.5) b -haemoly ic s ep ococci 45 (9.1) Vi idans s ep ococci 21 (4.2) En e ococci 17 (3.4) O he G am-posi i e 5 (1.0) G am-nega i e 223 (44.9) Esche ichia coli 159 (32.0) ESBL-E. coli 9 (1.8) Klebsiella sp 21 (4.2) ESBL-Klebsiella sp 0 (0.0) Pseudomonas ae uginosa 18 (3.6) O he G am-nega i e 25 (5.0) O he s 49 (9.9) Anae obes 15 (3.0) Fungi 1 (0.2) Polymic obial 33 (6.6) All 497 (100) MRSA, me hicillin- esis an S aphylococcus au eus; ESBL, ex ended- spec um be a-lac amase. J. Rannikko e al. / In e na ional Jou nal o In ec ious Diseases 58 (2017) 52–57 55 and day 90 we e independen o sepsis. This makes day 28 a mo e app op ia e ime-poin o in es iga e sepsis- ela ed mo ali y, as egula o y agencies ha e indeed conside ed. 21 None heless, only 34% o dea hs occu ing be ween day 0 and day 7 we e sepsis- ela ed in pa ien s wi h a li e-expec ancy o mo e han 6 mon hs. Thus, e en when day 7 and day 28 a e used as endpoin s o sepsis mo ali y, dea hs independen o sepsis a e no uled ou . This s udy has some limi a ions. By eason o he s udy design, i was no possible o include all blood cul u e-posi i e cases admi ed o he ED du ing he s udy pe iod. The po en ial ep esen a i eness o he ma e ial was assessed by e iewing he blood cul u e findings o hose cases ha we e no included in he s udy. Al oge he he e we e 167 such cases (con aminan s excluded). The pe cen ages o he mos common blood cul u e findings, S. au eus and Esche ichia coli, did no di e be ween he s udy popula ion and he cases ha we e no included. I is hus unlikely ha missed cases would ha e had a majo impac on he esul s. Among cases no included in he ma e ial as compa ed o hose included, he pe cen age o anae obes and o he G am- nega i e o ganisms (see Table 3) was somewha highe (9.0% s. 3.0% and 10.2% s. 5.0%, espec i ely) and he pe cen age o S ep ococcus pneumoniae was lowe (4.2% s. 10.5%). This may be due o he long incuba ion pe iod o anae obes and o he G am- nega i e o ganisms and he sho incuba ion pe iod o S. pneumo- niae. Again, assuming ha case a ali y among missed cases wi h anae obes o o he G am-nega i es would be app oxima ely he same as among hose included in he s udy, he numbe s would be oo low o al e he majo findings o he s udy. Only blood cul u e-posi i e cases we e included, which would exclude o example hose wi h alse-nega i e blood cul u es. This migh also ha e led o a lowe numbe o sep ic pa ien s wi h espi a o y in ec ion. Howe e , i is also a s eng h ha cul u e- nega i e sepsis cases, which a e di ficul o define, we e no included. Hence he ma e ial consis s mo e o ue in ec ions han pa ien s who we e o example only posi i e o sys emic inflamma o y esponse synd ome (SIRS). The ma e ial was ga he ed in a e ia y le el hospi al o a de eloped coun y wi h a low incidence o mul i- esis an bac e ia. These ypes o esul s canno , he e o e, be gene alized o low-income coun ies, as has been poin ed ou by Rello and Leblebicioglu. 25 I is possible ha he ole o he unde lying diseases would also ha e been di e en i he s udy pa ien s had been only om ICUs. The me hod used o ca ego ize he cause o dea h was de eloped by he au ho s. I encompasses he classifica ion used on he dea h ce ifica e (immedia e cause and unde lying cause) and combines i wi h he McCabe classifica ion o apidly a al disease. This has limi a ions, since in some cases i was no easy o decide whe he he pa ien would ha e li ed o mo e han 6 mon hs wi hou he sepsis o no . E en g oup 1 pa ien s we e elde ly and had many unde lying diseases. Howe e , he au ho s belie e ha he ca ego iza ion used is illus a i e o he cou se o e en s among hese pa ien s, and he use o wo o h ee clinicians o p o ide a judgemen gi es i su ficien alidi y. In conclusion, his s udy showed ha pa ien s wi h sepsis who die wi hin 90 days a e sepsis in a de eloped coun y, die mos ly ei he o sepsis by weakening o he pa ien wi h a apidly a al (<6 mon hs) unde lying disease, o o he unde lying disease i sel . This does no mean ha sepsis pa ien s do no need high-quali y ca e. Indeed, he low numbe o cases o sepsis- ela ed mo ali y among pa ien s wi h a li e-expec ancy o mo e han 6 mon hs may be a ibu ed o he good quali y o he ca e hese pa ien s ecei ed. When in es iga ing sepsis- ela ed mo ali y, day 28 should be used as he endpoin o iden i y mos o he sepsis- ela ed mo ali y, while a la ge pa o sepsis-independen mo ali y is uled ou . Funding This wo k was suppo ed by he Compe i i e Resea ch Financing o Tampe e Uni e si y Hospi al (g an 9N075 o J.R. and X50060 o J.A.). The au ho s’ wo k was independen o he unde ( he unding sou ce had no in ol emen ). E hical app o al The s udy was app o ed by he E hics Commi ee o Tampe e Uni e si y Hospi al. Conflic o in e es JR, JS, TS, JA, and RH epo no conflic o in e es . Re e ences 1. K.D. Kochanek, S.L. Mu phy, J. Xu, B. Tejada-Ve a, Dea hs: final da a o 2014, Na l Vi al S a Rep. 2016; 65(June (4)):1–122. 2. Ma in GS, Mannino DM, Moss M. The e ec o age on he de elopmen and ou come o adul sepsis. C i Ca e Med 2006;34:15–21. 3. Williams MD, B aun LA, Coope LM, Johns on J, Weiss RV, Qualy RL, e al. Hospi alized cance pa ien s wi h se e e sepsis: analysis o incidence, mo ali y, and associa ed cos s o ca e. C i Ca e Med 2004;8:R291–8. 4. Wang HE, Shapi o NI, G i fin R, Sa o d MM, Judd S, Howa d G. Ch onic medical condi ions and isk o sepsis. PLoS One 2012;7:e48307. 5. To es VB, Aze edo LC, Sil a UV, Ca uso P, To elly AP, Sil a E, e al. Sepsis- Associa ed Ou comes in C i ically Ill Pa ien s wi h Malignancies. Ann Am Tho ac Soc 2015;12:1185–92. 6. No osad SA, Sapiano MR, G igg C, Lake J, Robyn M, Dumya i G, e al. Vi al Signs: Epidemiology o Sepsis: P e alence o Heal h Ca e Fac o s and Oppo uni ies o P e en ion. MMWR Mo b Mo al Wkly Rep 2016;65:864–9. 7. Kuma A, Robe s D, Wood KE, Ligh B, Pa illo JE, Sha ma S, e al. Du a ion o hypo ension be o e ini ia ion o e ec i e an imic obial he apy is he c i ical de e minan o su i al in human sep ic shock. C i Ca e Med 2006;34:1589– 96. 8. Ka lsson S, Va pula M, Ruokonen E, Pe ila V, Pa iainen I, Ala-Kokko TI, e al. Incidence, ea men , and ou come o se e e sepsis in ICU- ea ed adul s in Finland: he Finnsepsis s udy. In ensi e Ca e Med 2007;33:435–43. 9. Lin JN, Tsai YS, Lai CH, Chen YH, Tsai SS, Lin HL, e al. Risk ac o s o mo ali y o bac e emic pa ien s in he eme gency depa men . Acad Eme g Med 2009;16:749–55. 10. O ega M, Almela M, Ma inez JA, Ma co F, So iano A, Lopez J, e al. Epidemiology and ou come o p ima y communi y-acqui ed bac e emia in adul pa ien s. Eu J Clin Mic obiol In ec Dis 2007;26:453–7. 11. Gille-Johnson P, Hansson KE, Ga dlund B. Clinical and labo a o y a iables iden i ying bac e ial in ec ion and bac e aemia in he eme gency depa men . Scand J In ec Dis 2012;44:745–52. 12. Lee CC, Lee CH, Chuang MC, Hong MY, Hsu HC, Ko WC. Impac o inapp op ia e empi ical an ibio ic he apy on ou come o bac e emic adul s isi ing he ED. Am J Eme g Med 2012;30:1447–56. 13. Blanco J, Mu iel-Bombin A, Sag edo V, Taboada F, Gandia F, Tamayo L, e al. Incidence, o gan dys unc ion and mo ali y in se e e sepsis: a Spanish mul icen e s udy. C i Ca e 2008;12:R158. 14. He ne AC, Ho on JM, Ma chick MR, Jones AE. E iology o illness in pa ien s wi h se e e sepsis admi ed o he hospi al om he eme gency depa men . Clin In ec Dis 2010;50:814–20. 15. Ho an TC, And us M, Dudeck MA. CDC/NHSN su eillance defini ion o heal h ca e-associa ed in ec ion and c i e ia o specific ypes o in ec ions in he acu e ca e se ing. Am J In ec Con ol 2008;36:309–32. 16. Le y MM, Fink MP, Ma shall JC, Ab aham E, Angus D, Cook D, e al. 2001 SCCM/ ESICM/ACCP/ATS/SIS In e na ional Sepsis Defini ions Con e ence. C i Ca e Med 2003;31:1250–6. 17. Singe M, Deu schman CS, Seymou CW, Shanka -Ha i M, Annane D, Baue M, e al. The Thi d In e na ional Consensus Defini ions o Sepsis and Sep ic Shock (Sepsis-3). JAMA 2016;315:801–10. 18. McCabe W, Jackson G. G am-Nega i e Bac e emia; I. E iology and Ecology. A ch In e n Med 1962;110:847–55. 19. Ko ick JA, Ma sh JW, S a zl TE, Yu VL. Pseudomonas ae uginosa bac e emia in pa ien s unde going li e ansplan a ion: an eme ging p oblem. Su ge y 1991;109:62–8. 20. Kao CH, Kuo YC, Chen CC, Chang YT, Chen YS, Wann SR, e al. Isola ed pa hogens and clinical ou comes o adul bac e emia in he eme gency depa men : a e ospec i e s udy in a e ia y Re e al Cen e . J Mic obiol Immunol In ec 2011;44:215–21. 21. Ma shall JC. Why ha e clinical ials in sepsis ailed? T ends Mol Med 2014;20:195–203. 56 J. Rannikko e al. / In e na ional Jou nal o In ec ious Diseases 58 (2017) 52–57 22. Pe ne A, Haase N, Gu o msen AB, Tenhunen J, Klemenzson G, Aneman A, e al. Hyd oxye hyl s a ch 130/0.42 e sus Ringe ’s ace a e in se e e sepsis. N Engl J Med 2012;367:124–34. 23. Yealy DM, Kellum JA, Huang DT, Ba na o AE, Weiss eld LA, Pike F, e al. A andomized ial o p o ocol-based ca e o ea ly sep ic shock. N Engl J Med 2014;370:1683–93. 24. Russell JA, Walley KR, Singe J, Go don AC, Hebe PC, Coope DJ, e al. Vasop essin e sus no epineph ine in usion in pa ien s wi h sep ic shock. N Engl J Med 2008;358:877–87. 25. Rello J, Leblebicioglu H. Sepsis and sep ic shock in low-income and middle- income coun ies: need o a di e en pa adigm. In J In ec Dis 2016;48:120–2. J. Rannikko e al. / In e na ional Jou nal o In ec ious Diseases 58 (2017) 52–57 57