Download Nosocomial Gram-Negative Bacteremia

Original
Report
Nosocomial
Gram-Negative
Bacteremia
Geoffrey D.Taylor, MD;” Maureen Buchanan-Chell, Bs~N;~Terri Kirkland, BSCN;~
Margaret McKenzie, BScN;+ and Rhoda Wiens, BScNt
ABSTRACT
Objective: To review experience with hospital-acquired
bacteremia due to aerobic gram-negative
bacilli at the University
of Alberta Hospital, episodes of nosocomial bacteremia occurring between August 1986 and June 1995 were analyzed.
Methods: Since 1986, prospective surveillance has been conducted for the occurrence
of hospital-acquired
bloodstream
infections. Cases were detected by monitoring positive blood
culture results reported by the institution’s clinical microbiology
laboratory. Following chart review, isolates were categorized
as representing
a hospital-acquired
infection, a communityacquired infection, or contamination,
based on Centers for Disease Control and Prevention criteria. Cases were followed for
1 week to determine outcome.
Results: There were 584 episodes of gram-negative bacteremia,
28.9% of all hospital-acquired
bloodstream
infections. The
gram-negative
infection
rate per 1000 admissions
was
unchanged from 1986 to 1995, but fell as a proportion of all
infections from 37.9% in 1987 to 19.8% in 1995. Escherichia
coli, Klebsiella spp, Pseudomonas
spp, and Enterobacter spp
were the most commonly identified species, accounting
for
more than 85% of episodes; only the incidence of Enterobacter changed, increasing from 2 per 10,000 admissions in 1987
to 5.2 per 10,000 in 1995. The 1 -week mortality rate gradually fell, from over 18% in 1987 to less than 5% in 1996 and
averaged
14% for monomicrobial
episodes. Pseudomonas
infections were associated with significantly higher mortality
(25.3%, relative risk 2.2) and Enterobacter infections with a significantly lower mortality (5.4%, relative risk 0.4).
Conclusion:
Over the past 9 years, the experience with the
hospital-acquired
bacteremia
due to aerobic gram-negative
bacilli at the University of Alberta Hospital revealed no significant change in the rate, causative organisms, organism-specific rates, or sources. However, as the incidence of infections
due to other organisms has risen, the proportion of nosocomial
* Department
of Medicine,
Unit, University
of Alberta
Received:
November
Address
correspondence
sity ofAlberta
Hospital,
202
University
Hospital,
ofAlberta;
Edmonton,
20,1996;Accepted:January
and +Infection
Alberta.
Control
21,1997.
to Dr. Geoffrey Taylor, 2E4.11 VM.C.,
Univer8440 112 Street, Edmonton,AB,
CanadaTGG
2B7.
bloodstream
infections attributable
isms has gradually fallen.
Key Words:
Int J Infect
bacteremia,
Dis 1997;
to gram-negative
gram-negative,
organ-
nosocomial
1:202-205.
Nosocomial infections occur in 5 to 10% of hospitalized
patients and have been identified as a significant public
health problem. I>2Bacteremic nosocomial infections
account for less than 5%, but result in greater attributable mortality, prolongation of length of hospitalization,
and higher cost of care than other infections.1-3 Bacteremia due to gram-negative organisms has previously
been recognized as an uncommon but often fatal complication of hospital care.4-8
Since 1986, blood culture reports have been prospectively monitored at University of Alberta Hospital to determine the incidence of nosocomial bloodstream infection.
Studies of specific organ system infections resulting in
bloodstream infection have been reported.9J0 The present study examines the incidence, sources, trends, and
outcome of nosocomial bacteremia due to aerobic gramnegative bacilli.
METHODS
The University ofAlberta Hospital is a tertiary acute care
facility and principal teaching hospital of the University
of Alberta Faculty of Medicine. All forms of acute medical
care were provided during the survey period, including
adult and pediatric medicine and surgery, obstetrics-gynecology, oncology, and solid organ transplantation.The hospital has seven intensive care units (ICUs) for critically ill
patients: cardiac, adult medical surgical, cardiac surgery,
burns, neurosurgery, pediatric, and neonatal. In 1986, at
the beginning of the survey, there were approximately
1100 inpatient bedsThrough the study period there was
a gradual reduction in bed numbers reaching 560 in 1995,
an approximate 50% overall reduction; however, annual
patient admissions have been unchanged, averaging
30,500 per year; average length of stay has decreased and
patient severity of illness and complexity has increased.
Nosocomial
Other than in ICUs, where phlebotomy
was performed by registered nurses, blood cultures were obtained
by physicians. Blood cultures were typically performed in
sets of two and incubated aerobically and anaerobically
for 7 days (unless prolonged incubation was requested),
using the Bactec 9240 system (Becton and Dickenson
Diagnostic Instrument Systems, Sparks, MD).
Blood cultures reported as growing microorganisms
were reviewed daily by nurse practitioners in the hospital’s infection control program. Patient charts were
reviewed.The isolate was classified as a contaminant or
resulting from a community-acquired or nosocomial infection based on Centers for Disease Control and Prevention
(CDC) criteria.‘l The source of infection was determined
by chart review. Service or ward where the patient was
located when signs or symptoms of infection first developed was recorded.The patient was followed for 1 week
from first positive blood culture to determine short-term
outcome. Cases were reviewed by a physician (GT) prior
to inclusion in the data-base.
The data-base was reviewed for cases of gram-negative bacteremia occurring between August 1,1986, the initiation of the surveillance program, and June 30,1995.A
case was included in the study if an aerobic gram-negative bacillus had been identified in the blood culture,
either alone or as part of a polymicrobial infection.
RESULTS
There were 584 episodes of nosocomial gram-negative
bacteremia in the 107 months of study (494 monomicrobial and 90 polymicrobial), representing 28.9% of all
nosocomial bloodstream infections. Whereas the incidence of nosocomial bloodstream
infection from all
causes gradually increased through the study period, the
rate for gram-negative infections was unchanged, averaging 2. I6 per 1000 admissions, so that the proportion of
infections due to gram-negative bacteria fell from 37.9%
/
Table
30,
infection
1995.
in Univer-
Nosocomial
1.
/ TuyZor et al
Bacteremia
Gram-Negative
Organisms
Bacteremia*:
Number
n = 610 (%)
E. coli
Klebsiella spp
Pseudomonas
spp
Enterobacter
spp
Serratia spp
Proteus
mirabilis
Acinetobacter
spp
Citrobacter
spp
Stenotrophomonas
Other species+
208
130
94
92
31
14
10
IO
8
13
*Monomicrobial
Aeromonas
malfophilia
and polymicrobial
(Z), Burkholderia
cases: +Morganel/a
cepacia
(l), Alcaiigenes
203
Identified
Organism
(34.3)
(21.5)
(15.5)
(14.5)
(5.1)
(2.3)
(1.6)
(1.6)
(1.3)
(2.2)
(5), Providencia (Z),
(I), Pasteurella (1).
in 1987, the first complete year of the survey, to 19.8%
in the first 6 months of 1995 (Figure 1).
There were 610 aerobic gram-negative bacilli identified in the 584 episodes (Table 1). Escberichia coli
(34.3%), Klebsiella species (21.5%), Pseudomonas spp
(15.5%), and Enterobacter spp (14.5%) were most common, accounting for more than 85% of all episodes. No
change occurred in annual incidence of any bacterial
species, except for an increase in infections due to Enterobacter spp, (from 2.0/10,000 admissions in 1987 to
5.2/10,000 in 1995), which was not statistically significant (P > 0.05 by &i-square for trend).
Table 2 demonstrates the source of infection for
episodes of gram-negative bacteremia; urinary tract was
most common (27.4%), followed by primary infection
(i.e., those for which no underlying source could be identified) (24.8%), gastrointestinal source (17.2%), and lower
respiratory tract (15.7%).The most common source of E.
coli was the urinary tract (43%), followed by the gastrointestinal tract (21%); the respiratory tract was a rare
source of E. coli (8%). Klebsiella and Enterobacter infections were most often attributed to a primary source
(33.8% and 38.6%, respectively), followed by a urinary
tract source (16.9% and 19.3%). Pseudomonas infections
were most commonly attributed to a lower respiratory
tract infection (3 1.9%), followed by primary infection
(21.3%), and were rarely due to a gastrointestinal source
Table
Figure
1.
Incidence
of nosocomial
bloodstream
sity of Alberta
Hospital,
August
1, 1986 to June
Gram-Negative
2.
Nosocomial
Gram-Negative
Bacteremia
Sources
Source
Number
n = 584 (%)
Urinary tract
Primary*
Gastrointestinal
tract
Lower respiratory
tract
Surgical site
Skin (including
burn)
Other
160
145
101
92
39
34
13
(27.4)
(24.8)
(17.2)
(15.7)
(6.7)
(5.9)
(2.2)
*Defined as a case of bloodstream infection that, following review, cannot
attributed to an infection al another body site.
be
204
International
Journal
Figure
2.
Rate of mortality
ture verifying
gram-negative
trend).
of Infectious
Diseases
/ Volume 1, Number 4, April 1997
within 1 week of first positive blood culbacteremia
(P = 0.028 by x2 for linear
(7%). No changes in relative frequency of overall sources
of infection or organism-speciftc sources occurred during
the 9 study years.
Of the 494 monomicrobial episodes, 68 (13.8%) were
associated with death within 1 week of first positive
blood culture. The l-week mortality rate gradually fell
through the study period from 18.3% in 1987 to 4.8% in
the first 6 months of 1995 (p = 0.028 by chi-square for
linear trend) (Figure 2). Mortality for monomicrobial
Pseudomonas cases was 25.3% (28/86) (P < 0.01; relative
risk, [RR] 1.81; 95% confidence interval [CI], 1.15-2.85)
and for Enterobacter cases was 5.4% (5/80) (P = 0.05; RR,
0.41; 95% CI, 0.07-0.99).
DISCUSSION
Despite changes in health care delivery at the University
of Alberta Hospital, including major reductions in bed
numbers, shortened length of stay, and increased patient
acuity resulting in increased susceptibility to infection, the
epidemiology of nosocomial bloodstream infections due
to gram-negative bacteria has remained remarkably constant within the institution. Over the past 9 years there
has been no significant change in the rate, causative organisms, organism-specific rates, or sources of gram-negative
sepsisAs there has been no significant change in annual
blood culture numbers or laboratory methodology there
should be no diagnostic artefact influencing the trend.
One possible explanation for the lack of change in incidence of gram-negative bacteremia is that prevention
techniques have kept pace with increasing patient susceptibility. However, no particular prevention intervention can be identified that would have had widespread
impact on the incidence of gram-negative infections. Four
organisms, namely E. coli, Klebsiella, Pseudomonas, and
Enterobacter accounted for the majority of episodes; all
other gram-negative bacilli together accounted for less
than 15%.
As the incidence of infections due to other organisms, particularly gram-positive cocci, has risen, the proportion of nosocomial bloodstream infections attributable
to gram-negatives has gradually fallen, so that by the first
half of 1995 less than 20% were associated with gram-negative bacteria, compared to 37.9% in 1987, the first full
year of the study. The rise in incidence of gram-positive
infections has been documented elsewhere and has been
attributed to increased use of intravascular catheters of
various types.4,5,‘2
The originating sources of infection for episodes of
nosocomial gram-negative sepsis in the present study generally were not unexpected.The prominence of the urinary tract as a source represents a potential opportunity
to reduce the incidence of nosocomial gram-negative sepsis. Since a large majority of bacteremic nosocomial urinary tract infections are a consequence
of urinary
catheter use, and since studies from other institutions
have shown a high rate of inappropriate use of catheters,
reducing inappropriate
catheter use could potentially
reduce the rate of nosocomial gram-negative sepsis.13x14
Some of the 24.8% of cases that were considered “primary” may have been attributable to vascular catheters,
but it seems likely many were not. It is surprising that in
so many cases of gram-negative sepsis neither the authors
nor the clinicians providing care for the patient could
determine the source of infection.
A l-week follow-up period was chosen in part for
surveillance convenience and in part to estimate the
immediate mortality Impact of septicemia.The overall lweek mortality rate of 13.9% for monomicrobial
cases
cannot be considered an attributable mortality, since
many of these patients had other significant medical problems, but can provide an estimate of the short-term
impact of gram-negative septicemia. The much higher
mortality associated with Pseudomonas sepsis was not
unexpected, as this organism previously has been associated with a high mortality risk6-8;however, the trend to
a lower mortality risk with Enterobacter sepsis was unexpected and should be confirmed by further studies.
The downward trend in l-week mortality was steady,
from 18.3% in 1987 to 4.8% in the first half of 1995, but
whether this reflects a similar reduction in hospital mortality or attributable mortality is uncertain and would
need to be determined by further studies.As there was
no change in source or organism causing gram-negative
sepsis, some other factor or factors must account for this
downward trend. Supportive therapy and specific antimicrobial therapy has gradually been modified over the past
decade, but no single intervention has been shown to
dramatically improve the outcome of patients experiencing gram-negative sepsis. 15,16
This would be consistent
with a steady decline in mortality rather than a dramatic
Nosocomial
reduction
vention.
at one point
in time
owing
to a single
inter-
REFERENCES
1. Haley RW, Culver DH,White JW Morgan WM, Emori TG.The
nationwide nosocomial infection rate.A need for a new vital
statistic.Am J Epidemiol 1985; 121:159-167.
2. Public health focus. Surveillance, prevention, and control of
nosocomial infections. MMWR 1992; 41:783.
3. Pittet D,Tarara D,Wenzel RF! Nosocomial bloodstream infection in critically ill patients. Excess length of stay, extra costs,
and attributable mortality JAMA 1994; 271:1598-1601.
4. McGowan JE Jr. Changing etiology of nosocomial bacteremia
and fungemia and other hospital-acquired
infections. Rev
Infect Dis 1985; 7(Suppl 3):S357-S370.
5. Schaberg DR, Culver DH, Gaynes RI? Major trends in the
microbial etiology of nosocomial infection. Am J Med 1991;
91(Suppl 3B):72S-75s.
6. Bryan CS, Reynolds KL, Brenner ER. Analysis of 1186
episodes of gram-negative bacteremia in non-university
hospitals: the effects of antimicrobial
therapy Rev Infect Dis
1983; 5:629-638.
7. Myerowitz RL, Medeiros AA, O’Brien TE Recent experience
with bacillemia due to gram-negative organisms. J Infect Dis
1971; 124:239-246
Gram-Negative
Bacteremia
/ Taylor et aE
205
8. Miller PJ,Wenzel RI? Etiologic organisms as independent predictors of death and morbidity associated with bloodstream
infections. J Infect Dis 1987; 156:471-477.
9. Taylor GD, Buchanan-Chell
M, Kirkland T, McKenzie M,
Wiens R. Bacteremic nosocomial pneumonia: a 7-year experience in one institution. Chest 1995; 108:786-788.
10. Taylor GD, Buchanan-Chell M, Kirkland T, McKenzie M,Wiens
R. Nosocomial urosepsis: an analysis of 218 cases. Int J Infect
Dis 1997;1:92-94.
11. Garner JS, Jarvis WR, Emori TG, Horan TC, Hughes JM. CDC
definitions for nosocomial infections, 1988. Am J Infect Control 1988; 16:128-140.
12. Banerjee SN, Emori TG, Culver DH, et al. Secular trends in
nosocomial primary bloodstream
infections in the United
States, 1980-1989.Am
J Med 1991; 91(Suppl 3B):86S-89s.
13. Krieger JN, Kaiser DL, Wenzel RI? Urinary tract etiology of
bloodstream infection in hospitalized patients. J Infect Dis
1983; 148:57-62.
14. Hartstein AI, Garber SB, Ward TT, Jones SR, Morthland VII.
Nosocomial urinary tract infection: a prospective evaluation
of 108 catheterized patients. Infect Control 1981; 2:380-386.
15. Calandra T, Cometta A. Antibiotic therapy for gram-negative
bacteremia. Infect Dis Clin North Am 1991; 4:817-834.
16. Young LS. Sepsis syndrome. In: Mandell GL, Bennett JE, Dolin
R, eds. Mandell, Douglas, and Bennett’s principles and practice of infectious diseases. 4th Ed.Vol 1. NewYork: Churchill
Livingstone, 1995:690-705.