BSI in Patients with Malignancies CID 2003:36 (1 May) 1103

M A J O R A R T I C L E
Current Trends in the Epidemiology
of Nosocomial Bloodstream Infections
in Patients with Hematological Malignancies
and Solid Neoplasms in Hospitals
in the United States
Hilmar Wisplinghoff,
1
Harald Seifert,
1
Richard P. Wenzel,
2
and Michael B. Edmond
2
1
Institute for Medical Microbiology, Immunology and Hygiene, University of Cologne, Germany; and
2
Department of Internal Medicine,
Virginia Commonwealth University, Richmond, Virginia
A total of 2340 patients with underlying malignancy were identied among 22,631 episodes of nosocomial
bloodstream infections (BSIs) in a prospectively collected database for 49 hospitals in the United States
(Surveillance and Control of Pathogens of Epidemiologic Importance [SCOPE] Project). Data were obtained
for the period of March 1995 through February 2001. Gram-positive organisms accounted for 62% of all BSIs
in 1995 and for 76% in 2000 ( ), and gram-negative organisms accounted for 22% and 14% of all BSIs P ! .001
for these years, respectively. Neutropenia was observed in 30% of patients, so neutropenic and nonneutropenic
patients were compared. In both, the predominant pathogens were coagulase-negative staphylococci (32% of
isolates recovered from neutropenic patients and 30% of isolates recovered from nonneutropenic patients).
The source of BSI was not determined for 57% of patients. The crude mortality rate was 36% for neutropenic
patients and 31% for nonneutropenic patients.
Bloodstream infections (BSIs) are currently the 13th
leading cause of death in the United States, and the
age-adjusted mortality rate has increased by 78%during
the past 2 decades [1]. Approximately 250,000 cases of
nosocomial BSI occur annually in the United States [2].
The proportion of patients with nosocomial BSI who
have an underlying malignancy is 10% (Surveillance
Received 28 October 2002; accepted 9 January 2003; electronically published
14 April 2003.
Presented in part: 38th Annual Meeting of the Infectious Diseases Society of
America, Denver, Colorado, 30 May3 June 1998 (abstract L-2).
Financial support: The Surveillance and Control of Pathogens of Epidemiologic
Importance Project is funded, in part, by Pzer, Merck, and Cubist Pharmaceuticals.
Reprints or correspondence: Dr. Hilmar Wisplinghoff, Institute for Medical
Microbiology, Immunology and Hygiene, University of Cologne, Goldenfelsstr.
1921, 50935 Cologne, Germany (h.wisplinghoff@uni-koeln.de).
Clinical Infectious Diseases 2003; 36:110310
2003 by the Infectious Diseases Society of America. All rights reserved.
1058-4838/2003/3609-0002$15.00
and Control of Pathogens of Epidemiologic Importance
[SCOPE] Project database; unpublished data). Despite
advances in the treatment and supportive care of pa-
tients with malignancies that have led to improvements
in long-term survival, infection remains the most sig-
nicant complication of anticancer therapy. Through-
out the 1960s and 1970s, gram-negative organisms were
most frequently isolated from neutropenic patients with
cancer who had BSI. However, during the past 20 years,
gram-positive organisms have become increasingly
common. Coagulase-negative staphylococci (CoNS),
viridans group streptococci, Staphylococcus aureus,
Escherichia coli, and Pseudomonas aeruginosa, in various
rank orders, have been found to be the most prevalent
organisms in recent studies [39]. Despite advances in
antimicrobial treatment, BSIs prolong hospital stay, in-
crease direct patient care costs, and cause considerable
mortality [3, 10]. In neutropenic patients with fever of
1104 CID 2003:36 (1 May) Wisplinghoff et al.
unknown origin, the attack rate for BSI is 11%38% [1114].
The clinical manifestations of BSI in these patients may vary
from transient bacteremia to fulminant septic shock. In adult
patients with BSI and underlying malignancies, the crude mor-
tality rate is 18%42% [3, 8, 15, 16]. In the face of emerging
multidrug-resistant organisms, antimicrobial prophylaxis and
treatment have become increasingly difcult in these highly
compromised patients. This study was conducted to assess the
epidemiological features of BSI in adult patients with malig-
nancies and the species distribution and antimicrobial suscep-
tibilities of causative pathogens. In addition, patients who had
neutropenia were compared with those who did not have
neutropenia.
MATERIALS AND METHODS
Background. The SCOPE Project, based at Virginia Com-
monwealth University in Richmond, was established to identify
the predominant pathogens and antimicrobial susceptibilities
of nosocomial bloodstream isolates. The 49 participating hos-
pitals throughout the United States represent medical institu-
tions of various sizes (range, 601200 beds) from a broad range
of geographical regions [9, 17]. The data generated by the
SCOPE Project have shown a high correlation with data from
the National Nosocomial Infection Surveillance (NNIS) pro-
gram of the Centers for Disease Control and Prevention (At-
lanta, GA) [17].
Study design. Clinical data and bloodstream isolates were
prospectively collected by local infection-control practitioners.
Adult patients admitted during the period of 1 March 1995
through 28 February 2001 to one of the 49 hospitals partici-
pating in the SCOPE Project were eligible if they had an un-
derlying malignancy and met the criteria for a nosocomial BSI.
A nosocomial BSI was determined to have occurred if 1
culture of a blood sample obtained 48 h after admission to
the hospital yielded a pathogenic organism. If the bloodstream
isolate was a potential skin contaminant (e.g., diphtheroids,
Propionibacterium species, Bacillus species, CoNS, or micro-
cocci), all of the following criteria were required for the di-
agnosis: the presence of an intravascular catheter, the initiation
of antimicrobial therapy, and 1 of the following: temperature
of 138.0C or !36.0C, chills, and systolic blood pressure of
!90 mm Hg [9]. Episodes of BSI that represented relapses were
excluded. However, second episodes after different hospital ad-
missions were included as separate cases. We studied the data
for all eligible episodes of BSI.
The data that were routinely collected included age, sex,
underlying malignancies, predisposing clinical conditions, spe-
cies distribution and antimicrobial susceptibility of causative
pathogens, and outcome. Sources of secondary BSI were iden-
tied by cultures of samples obtained from distant sites that
yielded the same pathogen with an identical resistance pattern.
Patients who had neutropenia (absolute neutrophil count
[ANC], !1000 neutrophils/mL) at the onset of BSI were then
compared with patients who had hematological malignancies
but not neutropenia.
Microbiological methods. Blood cultures were processed
at the participating hospitals. The identication of blood iso-
lates and susceptibility testing were performed by the routine
methods in use at the afliated laboratories.
Statistical analysis. The results are expressed as the mean
or as a proportion of the total number of patients value SD
or isolates. For continuous variables, mean values were com-
pared by 2-sample Students t tests for independent samples.
Differences in proportions were compared by x
2
test or Fishers
exact test, when appropriate. Mean values are reported 1 SD.
All tests of signicance are 2-tailed; a was set at 0.05. The
Mann-Whitney U test was performed to test the equality of
continuous variables. All statistical analyses were performed
with use of SPSS software, version 10.0 for Windows (SPSS).
RESULTS
Study population and patient characteristics. From1 March
1995 through 28 February 2001, a total of 22,631 cases of BSI
were reported by hospitals participating in the SCOPE Project.
Among these cases, 2340 clinically signicant episodes of BSI
were identied in adult patients with malignancies. Patients
had a mean age of years (median, 58 years; range, 55 17
1696 years). A total of 1151 patients (49%) were women.
Neutropenia was observed in 696 patients (30%) at the time
of BSI. Age and sex distribution did not vary signicantly be-
tween neutropenic and nonneutropenic patients (313 patients
[45%] vs. 838 patients [51%], respectively, were women; mean
age, 52 years vs. 56 years, respectively). Almost all patients were
hospitalized as a result of primary neoplastic disease (2192
patients [94%]). All other causes were seen in !2% of patients.
The mean length of hospital stay before the onset of BSI was
15 days (median, 11 days; range, 0382 days) and did not differ
signicantly between neutropenic and nonneutropenic patients,
but there was a statistical trend (mean, 16 vs. 15 days, respec-
tively; ). P p.06
The crude mortality rate was higher for neutropenic patients
(251 patients [36%] vs. 510 patients [31%]; ). The P p.053
most prevalent potential risk factors were presence of central
venous lines in 626 (90%) of neutropenic patients and in 1319
(80%) of nonneutropenic patients (OR, 2.2; 95% CI, 1.662.93;
), followed by the presence of peripheral intravenous P ! .001
lines, in 97 (14%) and 378 (23%) (OR, 1.8; 95% CI, 1.442.37;
); receipt of total parenteral nutrition, in 112 (16%) P ! .001
and 316 (19%) (OR, 0.8; 95% CI, 0.631.03; ); and the P p.08
presence of urinary catheters, in 83 (12%) and 381 (23%) (OR,
BSI in Patients with Malignancies CID 2003:36 (1 May) 1105
Table 1. Predisposing factors for bloodstream infections in neutropenic and nonneutropenic patients
within 48 h before bloodstream infection.
Factor
No. (%) of patients
OR (95% CI) P
a
ANC, neutrophils/mL
Total
(n p 2340)
!1000
(n p 696)
1000
(n p 1644)
Central venous line present 626 (89.9) 1319 (80.2) 1945 (83.1) 2.2 (1.662.93) !.001
Peripheral intravenous line present 97 (13.9) 378 (23.0) 475 (20.3) 1.8 (1.442.37) !.001
Arterial line present 18 (2.6) 60 (3.6) 78 (3.3) 0.7 (0.391.23) .24
Urinary catheter present 83 (11.9) 381 (23.2) 464 (19.8) 2.2 (1.712.90) !.001
Receipt of total parenteral nutrition 112 (16.1) 316 (19.2) 428 (18.3) 0.8 (0.631.03) .08
Dialysis performed 5 (0.7) 43 (2.6) 48 (2.1) 3.7 (1.4010.69) .005
Stay in intensive care unit 48 (6.9) 223 (13.6) 271 (11.6) 2.1 (1.512.97) !.001
Receipt of ventilator support 27 (3.9) 118 (7.2) 145 (6.2) 1.9 (1.233.01) .003
NOTE. ANC, absolute neutrophil count.
a
Patients with an ANC of !1000 neutrophils/mL vs. those with an ANC of 1000 neutrophils/mL.
Table 2. Distribution of gram-positive, gram-negative, anaerobic, and fungal pathogens in blood-
stream infections in 2652 patients with hematological malignancies and solid neoplasms, 19952000.
Pathogen
No. (%) of patients, by year
1995
(n p 390)
1996
(n p 556)
1997
(n p 508)
1998
(n p 451)
1999
(n p 336)
2000
(n p 411)
Gram-positive organisms 241 (61.8) 339 (61.0) 267 (52.6) 251 (55.7) 201 (59.8) 312 (75.9)
Gram-negative organisms 84 (21.5) 154 (27.7) 145 (28.5) 164 (36.4) 100 (29.8) 59 (14.4)
Anaerobic organisms 7 (1.8) 13 (2.3) 48 (9.4) 10 (2.2) 8 (2.4) 6 (1.5)
Fungi 58 (14.9) 50 (9.0) 48 (9.4) 26 (5.8) 27 (8.0) 34 (8.3)
2.2; 95% CI, 1.72.9; ; table 1). Among neutropenic P ! .001
patients, 48 (7%) required intensive care before the onset of
BSI, in contrast to 223 (14%) of nonneutropenic patients (OR,
2.1; 95% CI, 1.512.97; ). Ventilator support before BSI P ! .001
was required for 27 (4%) and 118 (7%) of patients, respectively
(OR, 1.9; 95% CI, 1.233.01; ). P p.003
Microbiological features. During the study period, a total
of 2711 isolates were recovered during 2340 episodes of BSI.
Sixty-one percent of all episodes were caused by gram-positive
aerobic organisms, and 27% were caused by gram-negative aer-
obic organisms. The proportion of gram-positive organisms
was 62% for BSIs in 1995 and 76% for those in 2000 (P !
). The proportions of gram-negative pathogens during the .001
same periods of time were 22% and 15%, respectively. The
greatest changes in the spectrum of pathogens occurred from
1999 to 2000. The proportion of gram-positive and gram-neg-
ative organisms changed from 60% to 76% and from 30% to
14% in those years, respectively. During that time, there were
no signicant differences in hospitals reporting to SCOPE. Also,
proportions did not change signicantly if only data from the
hospitals that reported continuously for 19952000 were an-
alyzed. Anaerobes were rarely isolated (3% of isolates), and the
proportion remained stable throughout the study period. Fungi
accounted for 10% of isolates, with a nadir of 6% in 1998 and
a peak of 15% in 1995 (table 2).
The organisms that were most frequently isolated from neu-
tropenic patients were CoNS (32% of BSIs), S. aureus (12% of
BSIs), E. coli (7% of BSIs), and enterococci (6% of BSIs). In
nonneutropenic patients, the most common pathogens were
CoNS (30% of BSIs), enterococci (14% of BSIs), and S. aureus
(11% of BSIs). Candida species accounted for 9% and 8% of
isolates recovered from neutropenic and nonneutropenic pa-
tients, respectively (table 3). Viridans group streptococci, which
accounted for 3% of all isolates in neutropenic patients, were
signicantly more common in this group (OR, 1.9; 95% CI,
1.877.60; ). Enterococcus faecium was more frequently P ! .001
isolated from patients without neutropenia than it was from
patients with neutropenia (6% vs. 2%; OR, 3.0; 95% CI,
1.755.04; ). Other frequently isolated organisms were P ! .001
distributed almost equally in both groups. Enterococci were
isolated earliest during the hospital stay (mean, day 10), whereas
viridans group streptococci and Candida species were isolated
latest in the hospital stay (mean, days 17 and 18, respectively).
A total of 329 (14%) of all episodes were polymicrobial.
No signicant seasonal or geographical patterns could be
observed for any of the organisms when different time periods
1106 CID 2003:36 (1 May) Wisplinghoff et al.
Table 3. Species distribution of predominant pathogens in bloodstream infections in neutropenic
and nonneutropenic patients.
Pathogen
No. (%) of patients
OR (95% CI) P
a
ANC, neutrophils/mL
Total
(n p 2711)
!1000
(n p 798)
1000
(n p 1913)
Gram-positive organisms
All 487 (61.0) 1152 (60.2) 1639 (60.5) 1.0 (0.871.23) .7
CoNS 252 (31.6) 566 (29.6) 818 (30.2) 1.1 (0.911.32) .3
Staphylococcus aureus 98 (12.3) 213 (11.1) 311 (11.5) 1.1 (0.861.45) .4
Enterococci
All 50 (6.3) 265 (13.9) 315 (11.6) 2.4 (1.743.34) !.001
Enterococcus faecalis 25 (3.1) 100 (5.2) 125 (4.6) 1.7 (1.072.73) .02
Enterococcus faecium 18 (2.3) 122 (6.4) 140 (5.2) 3.0 (1.755.04) !.001
Streptococci
All 73 (9.1) 90 (4.7) 163 (6.0) 2.0 (1.462.84) !.001
VGS 23 (2.9) 15 (0.8) 38 (1.4) 1.9 (1.877.60) !.001
Other 14 (1.8) 18 (0.9) 32 (1.2) 1.9 (0.883.99) .1
Gram-negative organisms
All 199 (24.9) 521 (27.2) 720 (26.6) 0.9 (0.731.08) .2
Escherichia coli 58 (7.3) 148 (7.7) 206 (7.6) 0.9 (0.671.29) .7
Klebsiella species 43 (5.4) 130 (6.8) 173 (6.4) 0.8 (0.541.13) .2
Pseudomonas aeruginosa 29 (3.6) 90 (4.7) 119 (4.4) 0.8 (0.491.19) .3
Enterobacter species 25 (3.1) 55 (2.9) 80 (3.0) 1.1 (0.661.81) .8
Other Enterobacteriaceae 14 (1.8) 42 (2.2) 56 (2.1) 0.8 (0.411.49) .5
Other gram-negative organisms 30 (3.8) 57 (3.0) 87 (3.2) 1.3 (0.792.04) .4
Anaerobes 38 (4.8) 55 (2.9) 93 (3.4) 1.7 (1.082.63) .02
Fungi
All 74 (9.3) 185 (9.7) 259 (9.6) 0.9 (0.711.28) .8
Candida species 69 (8.6) 161 (8.4) 230 (8.5) 1.0 (0.761.39) .9
Other 5 (0.6) 23 (1.2) 28 (1.0) 0.5 (0.171.44) .3
NOTE. ANC, absolute neutrophil count; CoNS, coagulase-negative staphylococci; VGS, viridans group streptococci.
a
Patients with an ANC of !1000 neutrophils/mL vs. those with an ANC of 1000 neutrophils/mL.
and geographical regions (the northwest, northeast, southwest,
and southeast United States) were compared. Primary BSI, in
which no source could be determined, was seen in 60% of
nonneutropenic patients and in 56% of neutropenic patients.
However, samples from distant sites were infrequently obtained
for culture; therefore, some points of origin might not have
been detected. Secondary BSI most often originated from in-
travenous catheters (in 24% of neutropenic patients and 27%
of nonneutropenic patients) and the urinary tract (5% and 7%,
respectively), followed by the lower respiratory tract and the
gastrointestinal tract (4% and 3% each, respectively). No sig-
nicant differences between neutropenic and nonneutropenic
hosts were observed with regard to sources of infection.
Antimicrobial susceptibility. Methicillin resistance was de-
tected in 29% of S. aureus isolates and in 77% of CoNS isolates.
Vancomycin resistance was found in 56% of E. faecium isolates.
Ceftazidime resistance was noted in 7% of P. aeruginosa isolates.
Decreased susceptibility to penicillin (MIC, 0.12 mg/L) was
detected in 27% of viridans group streptococci. Among En-
terobacteriaceae, resistance to ceftriaxone was most common
among Enterobacter isolates (42%). Imipenem resistance was
observed in 6% of P. aeruginosa isolates and in 1% of Klebsiella
isolates. Resistance to gentamicin was encountered in 15% of
P. aeruginosa isolates, 5% of E. coli isolates, and 4% of Enter-
obacter isolates. Resistance rates of pathogens isolated from
neutropenic hosts did not differ signicantly from that for non-
neutropenic hosts. However, for P. aeruginosa, resistance to
ceftazidime was not detected in isolates recovered frompatients
with an ANC of !1000 neutrophils/mL, but it was observed in
9% of isolates recovered from patients with an ANC of 1000
neutrophils/mL ( ). Resistance to vancomycin was higher P p.2
in E. faecium isolates recovered from nonneutropenic patients,
BSI in Patients with Malignancies CID 2003:36 (1 May) 1107
Table 4. Crude mortality rate for neutropenic and nonneutropenic patients with monomicrobial
bloodstream infections, stratied by pathogen.
Pathogen
No. (%) of patients who died/
no. in each group (%)
OR (95% CI) P
a
ANC, neutrophils/mL
Total
(n p 2011)
!1000
(n p 608)
1000
(n p 1403)
MSSA 12/58 (20.7) 27/113 (23.9) 39/171 (22.8) 0.8 (0.361.91) .7
MRSA 6/21 (28.6) 7/53 (13.2) 13/74 (17.6) 2.2 (0.568.35) .2
CoNS 66/197 (33.5) 115/356 (32.4) 185/553 (33.4) 1.3 (0.861.84) .2
Enterococcus faecalis 6/20 (30.0) 22/88 (25.0) 28/108 (25.9) 0.9 (0.282.78) .99
Vancomycin-susceptible
Enterococcus faecium 2/6 (33.3) 7/43 (16.3) 9/49 (18.4) 1.7 (0.2813.06) .6
VRE 6/9 (66.7) 18/62 (29.0) 24/71 (33.8) 4.9 (0.9328.26) .05
VGS 2/14 (14.3) 2/11 (18.2) 4/25 (16.0) 1.3 (0.0721.78) .99
Escherichia coli 16/42 (38.1) 35/106 (33.0) 51/148 (34.5) 1.3 (0.582.84) .6
Enterobacter species 8/21 (38.1) 9/40 (22.5) 17/61 (27.9) 2.1 (0.587.83) .2
Klebsiella species 6/30 (20.0) 25/100 (25.0) 31/130 (23.8) 0.8 (0.242.23) .6
Pseudomonas aeruginosa 11/23 (47.8) 20/64 (31.2) 28/87 (35.6) 1.9 (0.655.91) .2
Candida species 27/56 (48.2) 57/132 (43.2) 84/188 (44.7) 1.2 (0.622.41) .6
NOTE. ANC, absolute neutrophil count; CoNS, coagulase-negative staphylococci; MRSA, methicillin-resistant S.
aureus; MSSA, methicillin-sensitive S. aureus; VGS, viridans group streptococci; VRE, vancomycin-resistant enterococci.
a
Patients with an ANC of !1000 neutrophils/mL vs. those with an ANC of 1000 neutrophils/mL.
compared with isolates recovered from neutropenic patients
(56% vs. 50%; ). P p.8
Outcome. Outcome data were available for all 2011 pa-
tients with monomicrobial BSI. The overall mortality rate was
32%, ranging from 16% in patients with BSI due to viridans
group streptococci to 45% in patients with BSI due to Candida
species. In general, the rates were higher for neutropenic pa-
tients than they were for nonneutropenic patients. The mor-
tality rate for patients with BSI due to P. aeruginosa was 36%
(48% of neutropenic vs. 31% of nonneutropenic patients;
); for E. coli, it was 35% (38% vs. 33%, respectively; P p.2
); for vancomycin-resistant E. faecium, it was 34% (67% P p.6
vs. 29%, respectively; ); and for CoNS, it was 33% (34% P p.05
vs. 32%, respectively; ; table 4). P p.2
In patients with S. aureus BSI, there were no signicant dif-
ferences in outcome between patients with BSI due to methi-
cillin-resistant strains, compared with those with BSI due to
methicillin-susceptible strains (18% vs. 23% died, respectively;
OR, 0.7; 95% CI, 0.341.52; ). However, because of the P p.4
limited number of patients in each group, the difference was
not statistically signicant (OR, 2.3; 95% CI, 0.885.99; P p
). For polymicrobial BSI, the outcome was worse for neu- .1
tropenic patients (38% died, compared with 27% of nonneu-
tropenic patients; OR, 1.8; 95% CI, 1.22.6; ). P p.002
DISCUSSION
Patients with malignancies and neutropenia are at high risk for
the development of nosocomial BSI. The attack rate of BSI in
neutropenic patients is 11%38% [1214] and is particularly
high in patients who have undergone bone marrow transplan-
tation, with a reported rate of 360 BSIs per 1000 neutropenic
episodes [3]. The proportion of BSIs caused by gram-positive
organisms has increased considerably in recent years [9, 1820],
and gures of 70%81% have been reported [5, 21]. Conrm-
ing these studies, we also found that gram-positive organisms
are the most prevalent pathogens in nosocomial BSIs in patients
with malignancies in hospitals throughout the United States,
ranging from 64% of BSIs in 1995 to 76% of BSIs in 2000.
Others have reported that BSIs in patients with hematologic
malignancies or solid tumors are most frequently caused by
CoNS, viridans group streptococci, S. aureus, E. coli, and P.
aeruginosa, in various rank orders [58].
Except for E. faecium, which was isolated more frequently
from nonneutropenic patients, and viridans group streptococci,
which were isolated more frequently from neutropenic patients,
species distribution did not vary signicantly between neutro-
penic and nonneutropenic hosts. It is of note that viridans
group streptococci were isolated from only 3% of positive blood
cultures of neutropenic patients in the present study. This pro-
1108 CID 2003:36 (1 May) Wisplinghoff et al.
portion is lower than that reported previously fromcomparable
patient populations (9%30%) [5, 2226]. The lower rate ob-
served in our study might be in part attributable to the con-
sideration of only nosocomial BSI in this study, as opposed to
other studies, in which cultures of blood samples obtained at
the time of admission were also included. In addition, our
denition of neutropenia as an ANC of !1000 neutrophils/mL
may also add to the low rate, because viridans group strepto-
coccal bacteremia occurs more frequently in patients with he-
matologic malignancies, profound neutropenia, and severe mu-
cositis than it does in patients with solid tumors and less-severe
neutropenia.
CoNS were the most frequently isolated pathogens both in
neutropenic and nonneutropenic hosts (32% and 30%, re-
spectively). Other studies [3, 58] reported proportions of BSIs
due to CoNS of 10%14%. The difference might be partially
explained by the case denition of BSI used by the SCOPE
Project. Even though strictly dened clinical signs are required
for inclusion, patients with only 1 blood culture yielding CoNS
are eligible. Recent studies have shown that the proportion of
contaminants is lower if 2 positive blood cultures (with sam-
ples obtained within hours of each other) are required for
inclusion [2729]. Therefore, our design might lead to an over-
estimation of the prevalence of BSIs due to CoNS. However,
this should be minimized by the strict clinical denitions used
in the SCOPE Project to ascertain the clinical relevance of the
positive blood culture. In addition, data from the SCOPE Pro-
ject have shown a high correlation with the NNIS data [17].
Mrazova-Studena and colleagues [30] reported a signicantly
higher frequency of anaerobic and fungal BSI in neutropenic
patients who had cancer, compared with nonneutropenic pa-
tients who had cancer. We also found that anaerobic bacteria
were more prevalent among neutropenic patients. However,
concurring with the results of the Club dEtudes des Maladies
Infectieuses en Cancer (CEMIC) group [31], we could not de-
tect any signicant differences between neutropenic and non-
neutropenic hosts with regard to the prevalence of fungi as
causative pathogens of BSI. In contrast, P. aeruginosa infection
was not signicantly more common in either group in our
study. However, a comparison of results is difcult because of
a considerable variability in the patient populations investigated
in different studies, even if only neutropenic patients are taken
into account.
BSI in immunocompetent hosts often derives from dissem-
ination of previously localized infections. In contrast, BSI in
neutropenic patients is more likely to derive from endogenous
sources, such as the gastrointestinal tract [26]. With regard to
identied sources of BSI, we did not observe any signicant
differences in neutropenic versus nonneutropenic patients.
However, sources were not determined for more than one-half
of the patients in both groups, and some potential sources of
infection that are important in neutropenic patients, such as
the gastrointestinal tract, could not be assessed.
The resistance to methicillin among isolates of S. aureus was
higher than in this study that in northern Europe [32] but
comparable to resistance rates reported in other series from the
United States [9]. However, compared with the overall popu-
lation (all BSIs recorded by SCOPE), the rates were still rela-
tively low (29% in the present study vs. 40% for all BSIs re-
corded by SCOPE). The prevalence of resistance to vancomycin
in E. faecium in our study was comparable to that noted in
recent studies from the United States [9] but higher than in
recent European studies [31]. Increasing resistance to penicillin
and other b-lactam antibiotics has been documented in viridans
group streptococci isolated from neutropenic patients with can-
cer [26, 33]. Similar to data reported from a recent series from
Germany [26], we found that 73% of isolates were susceptible
to penicillin. Resistance was observed in 13% of patients, a rate
still lower than those reported from Spain [33]. Several recent
studies found that neutropenia was independently associated
with vancomycin-resistant enterococcal infection when com-
paring patients who had vancomycin-resistant enterococcal BSI
with those who had vancomycin-susceptible enterococcal BSI
[34, 35]. In contrast, Lucas et al. [36], when comparing BSIs
due to vancomycin-susceptible and -resistant enterococci, did
not nd neutropenia to be a risk factor for BSI due to van-
comycin-resistant enterococci. In our series, the prevalence of
resistance to vancomycin among isolates of E. faecium was
higher in nonneutropenic patients, but the difference did not
reach statistical signicance.
One limitation of our study is that several factors that might
have inuenced the epidemiology of causative pathogens and
outcome in neutropenic patients with BSI (such as depth and
duration of neutropenia; presence of lymphopenia; type and
stage of underlying malignancy; and incidence, severity, and
duration of mucositis) could not be assessed, making com-
parison with other studies difcult. Also, because this study
only included nosocomial BSIs, comparison with studies that
have included patients who had community-acquiredinfections
may be difcult. Recent studies involving neutropenic patients
in the intensive care unit who have BSI reported crude mortality
rates of 35%83% [13, 37], and the crude mortality rate for
neutropenic patients with BSI is 18%42% [3, 38]. In our series,
the crude mortality rate was 35% in neutropenic patients and
31% in nonneutropenic patients. In contrast, Elting et al. [39]
reported that BSI in neutropenic patients usually has a more
benign course and is successfully treated in 190% of cases.
Increased mortality rates for immunocompromised patients
were found to be associated with BSI due to E. coli or P. aeru-
ginosa, with mortality rates of 22%33% [7], and with BSI due
to viridans group streptococci and enterococci (30% and 33%,
respectively [7]). In our study, outcome was worst for patients
BSI in Patients with Malignancies CID 2003:36 (1 May) 1109
with monomicrobial BSI due to Candida species (crude mor-
tality rate, 45%), followed by P. aeruginosa, E. coli, and van-
comycin-resistant E. faecium. Of note, the mortality rate was
also high for patients with BSI due to CoNS. When outcomes
stratied by organism were compared between neutropenic and
nonneutropenic hosts, no signicant differences were seen.
However, despite the large number of patients included in this
study, the number of cases in some subsets was low. Therefore,
further studies with a larger number of patients might be
warranted.
In conclusion, we report an ongoing trend toward gram-
positive BSI and emerging and increasing resistance to anti-
microbial agents in S. aureus, enterococci, and streptococci in
patients with hematological malignancies and solid neoplasms
in a large sample of hospitals across the United States.
Acknowledgments
We thank the infection-control practitioners and microbi-
ology laboratory personnel at the participating hospitals for
their active participation in this project, and we thank Donna
McClish for assistance with data management.
References
1. National Nosocomial Infections Surveillance. NNIS systemreport, data
summary from January 1992April 2000. Issued June 2000. Am J Infect
Control 2000; 28:42948.
2. Pittet D, Li N, Woolson RF, Wenzel RP. Microbiological factors inu-
encing the outcome of nosocomial bloodstream infections: a 6-year
validated, population-based model. Clin Infect Dis 1997; 24:106878.
3. Collin BA, Leather HL, Wingard JR, Ramphal R. Evolution, incidence,
and susceptibility of bacterial bloodstream isolates from 519 bone mar-
row transplant patients. Clin Infect Dis 2001; 33:94753.
4. Rolston KV, Elting LS, Bodey GP. Bacteremia due to viridans strep-
tococci in neutropenic patients [letter]. Am J Med 1995; 99:450.
5. Gonzalez-Barca E, Fernandez-Sevilla A, Carratala J, Granena A, Gudiol
F. Prospective study of 288 episodes of bacteremia in neutropenic can-
cer patients in a single institution. Eur J Clin Microbiol Infect Dis
1996; 15:2916.
6. Hann I, Viscoli C, Paesmans M, Gaya H, Glauser M. A comparison
of outcome from febrile neutropenic episodes in children compared
with adults: results from four EORTC studies. International Antimi-
crobial Therapy Cooperative Group (IATCG) of the European Organ-
ization for Research and Treatment of Cancer (EORTC). Br J Haematol
1997; 99:5808.
7. Spanik S, Kukuckova E, Pichna P, et al. Analysis of 553 episodes of
monomicrobial bacteraemia in cancer patients: any associationbetween
risk factors and outcome to particular pathogen? Support Care Cancer
1997; 5:3303.
8. Krupova I, Kaiserova E, Foltinova A, et al. Bacteremia and fungemia in
pediatric versus adult cancer patients after chemotherapy: comparison
of etiology, risk factors and outcome. J Chemother 1998; 10:23642.
9. Edmond MB, Wallace SE, McClish DK, Pfaller MA, Jones RN, Wenzel
RP. Nosocomial bloodstream infections in United States hospitals: a
three-year analysis. Clin Infect Dis 1999; 29:23944.
10. Pittet D, Tarara D, Wenzel RP. Nosocomial bloodstream infection in
critically ill patients: excess length of stay, extra costs, and attributable
mortality [see comments]. JAMA 1994; 271:1598601.
11. Klastersky J. Current attitudes for therapy for febrile neutropenia with
consideration to cost-effectiveness. Curr Opin Oncol 1998; 10:2848.
12. Madani TA. Clinical infections and bloodstreamisolates associatedwith
fever in patients undergoing chemotherapy for acute myeloid leucemia.
Infection 2000; 28:36771.
13. Gaytan-Martinez J, Mateos-Garcia E, Sanchez-Cortes E, Gonzalesz-
Llaven J, Casanova-Cardiel LJ, Fuentes-Allen JL. Microbiological nd-
ings in febrile neutropenia. Arch Med Res 2000; 31:38894.
14. Serody JS. Fever in immunocompromised patients. N Engl J Med
2000; 342:2179.
15. Krcmery V Jr, Spanik S, Krupova I, et al. Bacteremia due to multi-
resistant gram-negative bacilli in neutropenic cancer patients: a case
controlled study. J Chemother 1998; 10:3205.
16. Ehni WF, Reller LB, Ellison RTD. Bacteremia in granulocytopenic pa-
tients in a tertiary-care general hospital. Rev Infect Dis 1991; 13:6139.
17. Voelker R. New group tracks hospitals drug-resistant bugs [news].
JAMA 1996; 275:1778.
18. Cometta A, Calandra T, Gaya H, et al. Monotherapy with meropenem
versus combination therapy with ceftazidime plus amikacin as empiric
therapy for fever in granulocytopenic patients with cancer. The Inter-
national Antimicrobial Therapy Cooperative Group of the European
Organization for Research and Treatment of Cancer and the Gruppo
Italiano Malattie Ematologiche Maligne dellAdulto Infection Program.
Antimicrob Agents Chemother 1996; 40:110815.
19. Cometta A, Zinner S, de Bock R, et al. Piperacillin-tazobactam plus
amikacin versus ceftazidime plus amikacin as empiric therapy for fever
in granulocytopenic patients with cancer. The International Antimi-
crobial Therapy Cooperative Group of the European Organization for
Research and Treatment of Cancer. Antimicrob Agents Chemother
1995; 39:44552.
20. Bochud YP, Cometta A, Francioli P. Virulent infections caused by alpha-
hemolytic streptococci in cancer patients and their management. Curr
Opin Infect Dis 1997; 10:42230.
21. Rubio M, Palau L, Vivas JR, et al. Predominance of gram-positive
microorganisms as a cause of septicemia in patients with hematological
malignancies. Infect Control Hosp Epidemiol 1994; 15:1014.
22. Bochud PY, Eggiman P, Calandra T, Van Melle G, Sagha L, Francioli
P. Bacteremia due to viridans streptococcus in neutropenic patients
with cancer: clinical spectrum and risk factors [see comments] Clin
Infect Dis 1994; 18:2531.
23. Carratala J, Alcaide F, Fernandez-Sevilla A, Corbella X, Linares J, Gudiol
F. Bacteremia due to viridans streptococci that are highly resistant to
penicillin: increase among neutropenic patients with cancer. Clin Infect
Dis 1995; 20:116973.
24. Donnelly JP, Dompeling EC, Meis JF, de Pauw BE. Bacteremia due to
oral viridans streptococci in neutropenic patients with cancer: cyto-
statics are a more important risk factor than antibacterial prophylaxis
[letter; comment]. Clin Infect Dis 1995; 20:46970.
25. Reinert RR, Queck A, Kaufhold A, Kresken M, Lutticken R. Antimi-
crobial resistance and type distribution of Streptococcus pneumoniae
isolates causing systemic infections in Germany, 19921994. Clin Infect
Dis 1995; 21:1398401.
26. Wisplinghoff H, Reinert RR, Cornely O, Seifert H. Molecular rela-
tionships and antimicrobial susceptibilities of viridans group strepto-
cocci isolated from blood of neutropenic cancer patients. J Clin Mi-
crobiol 1999; 37:187680.
27. Peacock SJ, Bowier ICJ, Crook DWM. Positive predictive value of blood
cultures growing coagulase-negative staphylococci. Lancet 1995;
346(8968):1912.
28. Weinstein MP, Reller LB, Murphy JR, Lichtenstein KA. The clinical
signicance of positive blood cultures: a comprehensive analysis of 500
episodes of bacteremia and fungemia in adults. I. Laboratory and ep-
idemiologic observations. Rev Infect Dis 1983; 5:3553.
29. Weinstein MP, Towns ML, Quartey SM. The clinical signicance of
positive blood cultures in the 1990s: a prospective comprehensive eval-
uation of the microbiology, epidemiology, and outcome of bacteremia
and fungemia in adults. Clin Infect Dis 1997; 24:584602.
1110 CID 2003:36 (1 May) Wisplinghoff et al.
30. Mrazova-Studena M, Drgona L, Spanik S, et al. Bacteremia in neu-
tropenic versus nonneutropenic cancer patients: etiology and outcome
in 401 episodes. Neoplasma 1997; 44:3148.
31. Coullioud D, Van der Auwera P, Viot M, Lasset C. Prospective mul-
ticentric study of the etiology of 1051 bacteremic episodes in 782 cancer
patients. CEMIC (French-Belgian Study Club of Infectious Diseases in
Cancer). Support Care Cancer 1993; 1:3446.
32. Kresken M, Hafner D. Drug resistance among clinical isolates of fre-
quently encountered bacterial species in central Europe during
19751995. Study Group Bacterial Resistance of the Paul-Ehrlich-
Society for Chemotherapy. Infection 1999; 27(Suppl 2):S28.
33. Carratala J, Roson B, Fernandez-Sevilla A, Alcaide F, Gudiol F. Bacteremic
pneumonia in neutropenic patients with cancer: causes, empirical anti-
biotic therapy, and outcome. Arch Intern Med 1998; 158:86872.
34. Lautenbach E, Bilker WB, Brennan PJ. Enterococcal bacteremia: risk
factors for vancomycin resistance and predictors of mortality. Infect
Control Hosp Epidemiol 1999; 20:31823.
35. Bhorade SM, Christenson J, Pohlman AS, Arnow PM, Hall JB. The
incidence of and clinical variabilities associated with vancomycin-
resistant enterococcal colonization in mechanical ventilated patients.
Chest 1999; 115:108591.
36. Lucas KG, Brown AE, Armstrong D, Chapman D, Heller G. The iden-
tication of febrile, neutropenic children with neoplastic disease at low
risk for bacteremia and complications of sepsis. Cancer 1996; 77:7918.
37. Warren DK, Zack JE, Elward AM, Cox MJ, Fraser VJ. Nosocomial
primary bloodstream infections in intensive care unit patients in a
nonteaching community medical center: a 21-month prospective study.
Clin Infect Dis 2001; 33:132935.
38. Fagon JY, Chastre J, Vuagnat A, Trouillet JL, Novara A, Gibert C.
Nosocomial pneumonia and mortality among patients in intensive care
units. JAMA 1996; 275:8669.
39. Elting LS, Rubenstein EB, Rolston KV, Bodey GP. Outcomes of bac-
teremia in patients with cancer and neutropenia: observations from
two decades of epidemiological and clinical trials. Clin Infect Dis
1997; 25:24759.