Article
Vol. 11. No. 12
1115
Eur. J. Clin. Microbiol. Infect. Dis., December 1992, p. 1115-1128 0934-9723/92/12 01115-14 $3.00/0
An International Multicenter Study of Blood Culture Practices
J.A. W a s h i n g t o n 1 . a n d t h e I n t e r n a t i o n a l C o l l a b o r a t i v e B l o o d C u l t u r e S t u d y G r o u p 2
An international study was organized to review blood culture practices in 67 medical centers, most of which were teaching hospitals with a total of over 58,000 active hospital beds. The number of blood cultures per admission was generally greater than 0.5 in the USA and less than 0.5 in other countries. Criteria varied for defining a septic episode, as well as for ascribing clinical importance to isolates of coagulase-negative staphylococci; however, septicemia rates tended to be lower in centers in which clinical evaluation was included among these criteria. Staphylococci were ranked first or second among etiologic agents of septicemia in the USA, whereas Escherichia coli was most frequently ranked first among European and Asian centers. All USA centers recommended collection of two blood cultures per septic episode and all but one recommended a maximum number of blood cultures per septic episode, whereas similar recommendations were less common in Europe and Asia. Collection of more than 10 ml per blood culture was more common in the USA than in Europe or Asia. A variety of broth-based systems were used, often in combination with lysis-centrifugation for special (fungal, mycobacterial) or, on occasion, routine purposes.
The detection of bacteremia and fungemia remains one o f the most important functions of clinical microbiology laboratories. Although many of the characteristics associated with bacteremia and candidemia have been elucidated in recent years and guidelines have been published for the use of blood cultures to document septicemia (1, 2), there remain many issues and controversies over blood culture procedure (3, 4) and, therefore, potential substantive differences among medical institutions in blood culture practices. Only a small number of surveys of blood culture practices have been published, and these have been largely specific for a nation or region
(5-7). The objectives of this retrospective study were to survey c o n t e m p o r a r y blood culture practices at 1Department of Clinical Pathology, Section of Microbiology, Cleveland CIinic Foundation, 9500 Euclid Avenue, Cleveland, Ohio 44195, USA. 2R. Auckenthaler, I--I6pitalCantonal, Geneva, Switzerland; E.B. Berezin, Bichat-Claude Bernard Hospital, Paris, France; K. Yin-Ching Chuang, National Cheng Kung University Hospital, Taipei, Taiwan; E.H. Sng, Singapore General Hospital, Singapore; H. Kanno, Chiba University Hospital, Chiba, Japan; G.A. Meloni, University of Padua, Padua, Italy; A.R. Noriega, Hospital 12 de Octubre, Madrid, Spain; P.M. Shah, J.W. Goethe University, Frankfurt, Germany; D.C. Shanson, Charing Cross and Westminster Medical School, london, UK.
medical centers, most of which were teaching hospitals in the USA, Western E u r o p e and Asia, and to elucidate any differences in blood culture practices in different parts of the world. T h e study encompassed 67 medical centers comprising at least 58,000 active hospital beds.
Materials and Methods Collaborative study group members were asked to solicit participation by other medical centers in their respective countries. Each participant, including collaborative study group members, completed a questionnaire (available upon request to J.A. Washington) consisting of several sections. The first section provided a basic demographic description of each medical center; the second focused on specimen collection; the third on blood culture systems and procedures; the fourth on results and interpretation of positive blood cultures; and the fifth on the role of the participant in patient care, selection of antimicrobial agents for the hospital formulary, coordination of antimicrobial agents tested with those in the formulary, and evaluation of antimicrobial susceptibility test results in patient care. Medical centers participating in this survey, including those representing the international collaborative study group, were in the USA by region: Harlford Hospital, Hartford, CT; Massachusetts General Hospital, Boston, MA; Memorial Sloan Kettering Cancer Center, New York, NY; Robert Wood Johnson University Hospital, New Brunswick, NJ; Johns Hopkins Hospital, Baltimore,
teaching teaching teaching teaching
teaching
teaching teaching teaching teaching
teaching teaching
teaching
Northeast 1 2 3 4
Mid Atlantic 1
Midwest 1 2 3 4
West 1 2
Northwest 1
300
400 385
880 880 975 560
870
730 490 380 908
Estimated active beds
a Per 1,000 discharges. b NS: Not specified. c Combined inpatient and outpatient totals.
Hospital description
Region and hospital no.
Table 1: Demographics of US hospitals.
22,100
28,500 24,500
30,50O 16,394 32,860 23,000
37,000
27,120 18,870 21,000 35,100
Admissions
13.3 a
29.3 NS
11 13 47.4 100
59
22.6 a 55 17 NS
Per 1000 admissions
NS
5.1 NS
1.3 1.2 5.8 NS
NS
NS NS NS NS
Per 1000 patient days
Bacteremia rate b
5,058
9,600
18,000
36,000
33,000 10,200 8,750
Inpatients
19,000 c
32,000 e 21,000 c
25,000 e
1,132
2,400
600
8,000
3,500 1,200 4,250
Outpatients
Per year
Blood culture rate
0.9
1.1 1.0
0.2
-0.7 0.4
-
-
1.0
~ 0.9 1.7 0.5 0.8
Per admission
g
o o" o
K
~..~.
ox
VO1.11,1992
MD; Cleveland Clinic Foundation, Cleveland, OH; Indiana University School of Medicine, Indianapolis, IN; Washington University School of Medicine, St. Louis, MO; University of Iowa Hospitals, Iowa City, IA; Stanford University Medical Centel, Palo Alto, CA; University of California Los Angeles, Los Angeles, CA; and St, Vincent's Hospital, Portland, OR. In Europe participants were as follows:France: Bichat-Claude Bernard Hospital, Paris; 13 university hospitals, three community hospitals, one pediatric university-affiliated hospital and one cancer university-affiliated institute (multicenter survey); Germany: J.W. Goethe University, Frankfurt; Institute of Hygiene, Frankfurt; University of Erlangen-Nuremberg, Erlangen; Italy: University of Padua, Padua; Vicenza Hospital, Vicenza; Careggi Hospital, Careggi; Verona Hospital, Verona; Treviso Hospital, Treviso; Brescia Hospital, Brescia; Parma Hospital, Parma; Spain: Hospital 12 de Octubre, Madrid; Hospital San Carlos, Madrid; Hospital La Paz, Madrid; Switzerland: Universities of Geneva, Berne, Lausanne, and Zurich; UK: Westminster Hospital, London; University Hospital, Nottingham; Addenbrooks Teaching Hospital, Cambridge; University CoUege and Middlesex Hospital, London; Whittington Hospital, London; Scarborough Hospital, Yorkshire; Luton and Dunstable Hospital, Luton; Altngelvin Hospital, Londonderry. In Asia participants were as follows: Japan: Chiba University Hospital, Chiba; Singapore: Singapore General Hospital, Tan Tock Seng Hospital, Toa Payoh Hospital, Kandang Kerbau Hospital, Alexandra Hospital, Changi Hospital; Taiwan: National Cheng Kung University Hospital, Taipei; National Taiwan University Hospital, Taipei; Tri Service General Hospital, Taipei; Veterans General Hospital, Kaohsiung. Results In all but one instance, 1990 data were reported. The distribution of participants by country was as follows (Tables 1-3): France 19, Germany 3, Italy 7, Japan 1, Singapore 6, Spain 3, Switzerland 4, Taiwan 4, UK 8, U S A 12. Of the 67 medical centers, 62 were described as teaching hospitals, 43 of which were estimated to have at least 700 active beds. The number of hospital admissions per year varied widely from 5,800 to nearly 100,000. The rates of bacteremia and fungemia per 1,000 patient admissions were available from many medical centers and, when specified, ranged from 2 to 100 (approximately one-third reported rates of > 20/1,000 admissions). Varying criteria were used among the centers in defining septicemia on the basis of microbiologic data and in assessing the significance of isolation of coagulase-negative staphylococci. Clinical evaluation was specified as a criterion for defining a septic episode on the basis of a positive blood culture by 3 of 12 USA participants, 15 of 44 European participants, and one Asian participant (Table 4). Other than participants who did not specify a criterion for defining a septic episode, the remain-
1117
ing participants defined such episodes either by the first positive blood culture or by a new positive blood culture occurring at least two days after an initially positive blood culture, by a combination of these two criteria, by the species of microorganism isolated, or by the number of positive blood cultures with the same species (Table 4). When more than one organism was identified within a set of blood cultures, each organism was separately tabulated. The incidence of polymicrobial bacteremia was, therefore, not ascertained. Participants' evaluations of blood cultures yielding coagulase-negative staphylococci were quite diverse (Table 5). As may be seen in Table 5, clinical evaluation of the isolation of coagulase-negative staphylococci was specified by only 15 participants, either alone or in combination with isolation characteristics of the organism (i.e. isolation from more than one blood culture set, both bottles within a set, within 48 hours, species identification, or quantitation). Contamination rates of blood cultures listed by participants are given in Table 6. Contamination rates of 2,000 a
< 700a 700-1,000 a > 1,000a
1,350
Estimated active beds
Table 2: Demographics of European hospitals.
37,000 NS 30,000 NS
39,570 46,563 32,655
80,000 32,800 65,000 53,000 38,000 100,000 50,976
5,800 25,585 55,000
42,870
Admissions
,
I7 NS 14 NS
24 23 16
21 26 4.6 16.2 NS 12 NS
17 NS NS
2-25
25
Per 1000 admissions/discharges
Bacterernia
NS NS NS NS
NS NS NS
NS NS NS NS 2.5 NS NS
NS
NS
0.2-5
NS
Per 1000 patient days
rate b
13,1300 16,000c 12,00(F NS
11,365 11,465 5,185
8,000 6,106 6,500 5,500 5,000 12,000 21,232
4,2oo
5,000
NS
37,203
Inpatient
4,000~
Peryear
NS
4,500
132 0 0
150 236 20 300 2 300 462
NS NS
NS
NS
Outpatient
Blood culture rate
0.4 NS ~ 0.4 NS
0.1 0.25 0.16
0.1 0.2 0.1 0.I 0.13 0.12 0.4
0.02
-0.6
0.5-3
1.15
Per admission
t.ca
o
o o"
C3
O0
Vol. 11, 1992
1119
quent etiologic agents as Escherichia coli and Klebsiella spp. Enterococci were relatively more frequently seen among the top five etiologic agents in the U S A than in Europe, whereas Streptococcus pneumoniae was more frequently encountered in Europe than in the USA. Pseudomonas aeruginosa was ranked within the three most frequent etiologic agents in only three instances and was predominantly ranked sixth to tenth. Bacteroides were not ranked among the top ten etiologic agents in the USA or Asian medical centers but were ranked as high as fourth, although predominantly sixth to tenth, in frequency among European centers. Candida spp. were, with one exception in Europe, ranked sixth to tenth.
~,...
'.g e~ V
ZZm~
©
Mycobacterium avium-intracellulare was ranked among the ten most frequent etiologic agents by two centers; however, since participants were not specifically requested to rank microorganisms other than aerobic, facultatively anaerobic and anaerobic bacteria, it is uncertain how often Mycobacterium avium-intracellulare was encountered elsewhere. Cryptococcus neoformans was ranked fourth in frequency by one center and less than sixth in frequency by another in Europe.
ZZMZZZZZ
A
Q
~NNNNN
Vol. 11. No. 12
1115
Eur. J. Clin. Microbiol. Infect. Dis., December 1992, p. 1115-1128 0934-9723/92/12 01115-14 $3.00/0
An International Multicenter Study of Blood Culture Practices
J.A. W a s h i n g t o n 1 . a n d t h e I n t e r n a t i o n a l C o l l a b o r a t i v e B l o o d C u l t u r e S t u d y G r o u p 2
An international study was organized to review blood culture practices in 67 medical centers, most of which were teaching hospitals with a total of over 58,000 active hospital beds. The number of blood cultures per admission was generally greater than 0.5 in the USA and less than 0.5 in other countries. Criteria varied for defining a septic episode, as well as for ascribing clinical importance to isolates of coagulase-negative staphylococci; however, septicemia rates tended to be lower in centers in which clinical evaluation was included among these criteria. Staphylococci were ranked first or second among etiologic agents of septicemia in the USA, whereas Escherichia coli was most frequently ranked first among European and Asian centers. All USA centers recommended collection of two blood cultures per septic episode and all but one recommended a maximum number of blood cultures per septic episode, whereas similar recommendations were less common in Europe and Asia. Collection of more than 10 ml per blood culture was more common in the USA than in Europe or Asia. A variety of broth-based systems were used, often in combination with lysis-centrifugation for special (fungal, mycobacterial) or, on occasion, routine purposes.
The detection of bacteremia and fungemia remains one o f the most important functions of clinical microbiology laboratories. Although many of the characteristics associated with bacteremia and candidemia have been elucidated in recent years and guidelines have been published for the use of blood cultures to document septicemia (1, 2), there remain many issues and controversies over blood culture procedure (3, 4) and, therefore, potential substantive differences among medical institutions in blood culture practices. Only a small number of surveys of blood culture practices have been published, and these have been largely specific for a nation or region
(5-7). The objectives of this retrospective study were to survey c o n t e m p o r a r y blood culture practices at 1Department of Clinical Pathology, Section of Microbiology, Cleveland CIinic Foundation, 9500 Euclid Avenue, Cleveland, Ohio 44195, USA. 2R. Auckenthaler, I--I6pitalCantonal, Geneva, Switzerland; E.B. Berezin, Bichat-Claude Bernard Hospital, Paris, France; K. Yin-Ching Chuang, National Cheng Kung University Hospital, Taipei, Taiwan; E.H. Sng, Singapore General Hospital, Singapore; H. Kanno, Chiba University Hospital, Chiba, Japan; G.A. Meloni, University of Padua, Padua, Italy; A.R. Noriega, Hospital 12 de Octubre, Madrid, Spain; P.M. Shah, J.W. Goethe University, Frankfurt, Germany; D.C. Shanson, Charing Cross and Westminster Medical School, london, UK.
medical centers, most of which were teaching hospitals in the USA, Western E u r o p e and Asia, and to elucidate any differences in blood culture practices in different parts of the world. T h e study encompassed 67 medical centers comprising at least 58,000 active hospital beds.
Materials and Methods Collaborative study group members were asked to solicit participation by other medical centers in their respective countries. Each participant, including collaborative study group members, completed a questionnaire (available upon request to J.A. Washington) consisting of several sections. The first section provided a basic demographic description of each medical center; the second focused on specimen collection; the third on blood culture systems and procedures; the fourth on results and interpretation of positive blood cultures; and the fifth on the role of the participant in patient care, selection of antimicrobial agents for the hospital formulary, coordination of antimicrobial agents tested with those in the formulary, and evaluation of antimicrobial susceptibility test results in patient care. Medical centers participating in this survey, including those representing the international collaborative study group, were in the USA by region: Harlford Hospital, Hartford, CT; Massachusetts General Hospital, Boston, MA; Memorial Sloan Kettering Cancer Center, New York, NY; Robert Wood Johnson University Hospital, New Brunswick, NJ; Johns Hopkins Hospital, Baltimore,
teaching teaching teaching teaching
teaching
teaching teaching teaching teaching
teaching teaching
teaching
Northeast 1 2 3 4
Mid Atlantic 1
Midwest 1 2 3 4
West 1 2
Northwest 1
300
400 385
880 880 975 560
870
730 490 380 908
Estimated active beds
a Per 1,000 discharges. b NS: Not specified. c Combined inpatient and outpatient totals.
Hospital description
Region and hospital no.
Table 1: Demographics of US hospitals.
22,100
28,500 24,500
30,50O 16,394 32,860 23,000
37,000
27,120 18,870 21,000 35,100
Admissions
13.3 a
29.3 NS
11 13 47.4 100
59
22.6 a 55 17 NS
Per 1000 admissions
NS
5.1 NS
1.3 1.2 5.8 NS
NS
NS NS NS NS
Per 1000 patient days
Bacteremia rate b
5,058
9,600
18,000
36,000
33,000 10,200 8,750
Inpatients
19,000 c
32,000 e 21,000 c
25,000 e
1,132
2,400
600
8,000
3,500 1,200 4,250
Outpatients
Per year
Blood culture rate
0.9
1.1 1.0
0.2
-0.7 0.4
-
-
1.0
~ 0.9 1.7 0.5 0.8
Per admission
g
o o" o
K
~..~.
ox
VO1.11,1992
MD; Cleveland Clinic Foundation, Cleveland, OH; Indiana University School of Medicine, Indianapolis, IN; Washington University School of Medicine, St. Louis, MO; University of Iowa Hospitals, Iowa City, IA; Stanford University Medical Centel, Palo Alto, CA; University of California Los Angeles, Los Angeles, CA; and St, Vincent's Hospital, Portland, OR. In Europe participants were as follows:France: Bichat-Claude Bernard Hospital, Paris; 13 university hospitals, three community hospitals, one pediatric university-affiliated hospital and one cancer university-affiliated institute (multicenter survey); Germany: J.W. Goethe University, Frankfurt; Institute of Hygiene, Frankfurt; University of Erlangen-Nuremberg, Erlangen; Italy: University of Padua, Padua; Vicenza Hospital, Vicenza; Careggi Hospital, Careggi; Verona Hospital, Verona; Treviso Hospital, Treviso; Brescia Hospital, Brescia; Parma Hospital, Parma; Spain: Hospital 12 de Octubre, Madrid; Hospital San Carlos, Madrid; Hospital La Paz, Madrid; Switzerland: Universities of Geneva, Berne, Lausanne, and Zurich; UK: Westminster Hospital, London; University Hospital, Nottingham; Addenbrooks Teaching Hospital, Cambridge; University CoUege and Middlesex Hospital, London; Whittington Hospital, London; Scarborough Hospital, Yorkshire; Luton and Dunstable Hospital, Luton; Altngelvin Hospital, Londonderry. In Asia participants were as follows: Japan: Chiba University Hospital, Chiba; Singapore: Singapore General Hospital, Tan Tock Seng Hospital, Toa Payoh Hospital, Kandang Kerbau Hospital, Alexandra Hospital, Changi Hospital; Taiwan: National Cheng Kung University Hospital, Taipei; National Taiwan University Hospital, Taipei; Tri Service General Hospital, Taipei; Veterans General Hospital, Kaohsiung. Results In all but one instance, 1990 data were reported. The distribution of participants by country was as follows (Tables 1-3): France 19, Germany 3, Italy 7, Japan 1, Singapore 6, Spain 3, Switzerland 4, Taiwan 4, UK 8, U S A 12. Of the 67 medical centers, 62 were described as teaching hospitals, 43 of which were estimated to have at least 700 active beds. The number of hospital admissions per year varied widely from 5,800 to nearly 100,000. The rates of bacteremia and fungemia per 1,000 patient admissions were available from many medical centers and, when specified, ranged from 2 to 100 (approximately one-third reported rates of > 20/1,000 admissions). Varying criteria were used among the centers in defining septicemia on the basis of microbiologic data and in assessing the significance of isolation of coagulase-negative staphylococci. Clinical evaluation was specified as a criterion for defining a septic episode on the basis of a positive blood culture by 3 of 12 USA participants, 15 of 44 European participants, and one Asian participant (Table 4). Other than participants who did not specify a criterion for defining a septic episode, the remain-
1117
ing participants defined such episodes either by the first positive blood culture or by a new positive blood culture occurring at least two days after an initially positive blood culture, by a combination of these two criteria, by the species of microorganism isolated, or by the number of positive blood cultures with the same species (Table 4). When more than one organism was identified within a set of blood cultures, each organism was separately tabulated. The incidence of polymicrobial bacteremia was, therefore, not ascertained. Participants' evaluations of blood cultures yielding coagulase-negative staphylococci were quite diverse (Table 5). As may be seen in Table 5, clinical evaluation of the isolation of coagulase-negative staphylococci was specified by only 15 participants, either alone or in combination with isolation characteristics of the organism (i.e. isolation from more than one blood culture set, both bottles within a set, within 48 hours, species identification, or quantitation). Contamination rates of blood cultures listed by participants are given in Table 6. Contamination rates of 2,000 a
< 700a 700-1,000 a > 1,000a
1,350
Estimated active beds
Table 2: Demographics of European hospitals.
37,000 NS 30,000 NS
39,570 46,563 32,655
80,000 32,800 65,000 53,000 38,000 100,000 50,976
5,800 25,585 55,000
42,870
Admissions
,
I7 NS 14 NS
24 23 16
21 26 4.6 16.2 NS 12 NS
17 NS NS
2-25
25
Per 1000 admissions/discharges
Bacterernia
NS NS NS NS
NS NS NS
NS NS NS NS 2.5 NS NS
NS
NS
0.2-5
NS
Per 1000 patient days
rate b
13,1300 16,000c 12,00(F NS
11,365 11,465 5,185
8,000 6,106 6,500 5,500 5,000 12,000 21,232
4,2oo
5,000
NS
37,203
Inpatient
4,000~
Peryear
NS
4,500
132 0 0
150 236 20 300 2 300 462
NS NS
NS
NS
Outpatient
Blood culture rate
0.4 NS ~ 0.4 NS
0.1 0.25 0.16
0.1 0.2 0.1 0.I 0.13 0.12 0.4
0.02
-0.6
0.5-3
1.15
Per admission
t.ca
o
o o"
C3
O0
Vol. 11, 1992
1119
quent etiologic agents as Escherichia coli and Klebsiella spp. Enterococci were relatively more frequently seen among the top five etiologic agents in the U S A than in Europe, whereas Streptococcus pneumoniae was more frequently encountered in Europe than in the USA. Pseudomonas aeruginosa was ranked within the three most frequent etiologic agents in only three instances and was predominantly ranked sixth to tenth. Bacteroides were not ranked among the top ten etiologic agents in the USA or Asian medical centers but were ranked as high as fourth, although predominantly sixth to tenth, in frequency among European centers. Candida spp. were, with one exception in Europe, ranked sixth to tenth.
~,...
'.g e~ V
ZZm~
©
Mycobacterium avium-intracellulare was ranked among the ten most frequent etiologic agents by two centers; however, since participants were not specifically requested to rank microorganisms other than aerobic, facultatively anaerobic and anaerobic bacteria, it is uncertain how often Mycobacterium avium-intracellulare was encountered elsewhere. Cryptococcus neoformans was ranked fourth in frequency by one center and less than sixth in frequency by another in Europe.
ZZMZZZZZ
A
Q
~NNNNN
