; Predictors of Hospital Mortality in Adult Patients with Community
Learning Center
Plans & pricing Sign in
Sign Out
Your Federal Quarterly Tax Payments are due April 15th Get Help Now >>

Predictors of Hospital Mortality in Adult Patients with Community


  • pg 1
									                                                                            Community-acquired bloodstream infection       93

      Predictors of Hospital Mortality in Adult Patients with
       Community-Acquired Bloodstream Infection in the
                        Emergency Room
                    Tsung-Ming Chen, Tsung-hua Tsai1, Fang-Zhi Chen2, sZu-erh Chan3

         Background: Bloodstream infections (BSI) continue to be a serious problem in the community, but
    data on the clinical predictors of hospital mortality associated with BSI patients presenting in the
    emergency department are limited in Taiwan.
         Objectives: To characterize the clinical and microbiological factors for hospital mortality from
    community-acquired BSI in emergency room (ER) patients.
         Methods: This 12-month prospective study was conducted at the Erlin branch of Changhua Christian
    Hospital from January 1, 2005 to December 31, 2005. All admitted patients over 18 years old with
    microbiologically confirmed BSI in the ER were enrolled. Data were collected on demographics, comorbid
    medical conditions, clinical profile, microorganisms, source of infection, antimicrobial susceptibility
    testing, empiric antibiotic therapy, and hospital outcome. The prognostic factors were studied using
    univariate logistic regression.
         Results: During the study period, a total of 90 cases of BSI were reported. The mean age of the
    patients was 70.1 years (range, 39 to 97 years), and 51 (56.6%) patients were women. The most common
    comorbidities were diabetes mellitus (42.2%), neurological disorder with poor performance (24.4%),
    cardiovascular disease (21.1%), liver cirrhosis (15.5%), and underlying malignancy (12.2%). The
    most common pathogens were Escherichia coli (49%), Klebsiella species (8.8%), and Staphylococcus
    aureus (7.8%). The most common sources of BSI were the urinary tract (47.8%), and intra-abdominal
    sites (22.2%). Abnormal body temperature was the most frequent clinical finding (85.5%), and 32% of
    the patients developed acute organ dysfunction in the ER. Appropriate empiric antibiotic therapy was
    administrated in 82.2% of the cases. The hospital mortality rate was 15.5%. Univariate analysis revealed
    the predictors of mortality were comorbid liver cirrhosis (OR 6.37, p=0.005), presence of severe sepsis
    or septic shock (OR 44.68, p<0.001), and pneumonia (OR 42, p=0.004) or an intra-abdominal site (OR
    34.36, p=0.001) as the source of infection.
         Conclusions: The predictors identified in this study provide important prognostic information for
    patients with community-acquired BSI on arrival in the ER.

    Key words: bloodstream infections (BSI), emergency room (ER), community

                   Introduction                                  and remain a significant cause of mortality (1-3).
                                                                 Many studies on the epidemiology, microbiological
    Bloodstream infections (BSI) continue to be                  etiology, and prognosis of hospitalized patients
severe, often life-threatening infectious diseases,              with BSI have been published(4-7); however, data on

Received: October 19, 2006 Accepted for publication: January 4, 2007
From the Departments of Internal Medicine, 1Laboratory Medicine, 2Emergency Medicine
  Surgical Medicine, Changhua Christian Hospital Erlin Branch
Address for reprints: Dr. Szu Erh Chan, Department of Surgical Medicine, The Erlin Branch of Changhua Christian Hospital
558 Section 1, Dacheng Road, Erhlin Town 526, Taiwan (R.O.C.)
Tel: (04)8960128 Fax: (04)8966991
E-mail: 68302@cch.org.tw
94    J Emerg Crit Care Med. Vol. 18, No. 3, 2007

clinical predictors of hospital mortality in patients   positive BSI, included age, sex, underlying diseases
with BSI in the emergency department setting are        and predisposing factors (comorbid factors), clinical
limited in Taiwan.                                      and laboratory parameters in the emergency room,
    The aim of the present study was to describe        microbiological results of cultures from other
the impact of certain factors, such as gender,          sites, antibiotic regimens, results of antimicrobial
age, comorbid medical conditions, causative             susceptibility testing, and hospital outcome
organisms, source of BSI, degree of systemic            (survival or death).
response, and appropriateness of antibiotic therapy,        During the study period, nurses obtained
on the outcomes of ER patients with community-          blood from two peripheral sites. Before collecting
acquired BSI. The purpose of this study was (1)         the blood cultures, the skin was disinfected using
To analyze the clinical data of septic patients with    70% isopropyl alcohol and 10% povidone-iodine.
mortality risk on admission to the ER. (2) To assess    Blood samples were inoculated into both aerobic
which microorganisms are causing BSI in our             and anaerobic broth media for processing with
geographic area. (3) To provide information about       the BACTEC 9240 blood culture system (Becton
in vitro antimicrobial susceptibilities to available    Dickinson Diagnostic Instrument System, Sparks,
antibiotic agents. Our study may help physicians        MD, USA). Broth from positive bottles was
identify septic patients with a higher risk of          identified with use of standard techniques (8) .
hospital mortality and choose appropriate empirical     S u s c e p t i b i l i t y t e s t i n g o f i s o l a t e s w a s done
antibiotics.                                            u s i n g t h e s t a n d a r d d i s k d i ff u s i o n m e t h o d ,
                                                        according to the guidelines of the National
         Materials and Methods                          Committee for Clinical Laboratory Standards
                                                        (NCCLS) (9).
Hospital setting and patients
    The Erlin Branch of Changhua Christian              Definitions
Hospital is a 200 bed community teaching hospital            An episode of bloodstream infection was
providing primary care in the southern Changhua         defined by the first set of positive blood cultures in
area. The emergency room (ER) has more than             a series. If two or more days separated one episode
25,000 adult visits per year.                           from a preceding one, it was considered a new
     For the 12-month period from January 1, 2005       episode.
through December 31, 2005, data were collected               A contaminant was defined as an isolate
on all adult patients more than 18 years of age with    from a blood culture in a situation in which the
positive blood cultures from specimens collected        patient’s clinical condition was not suggestive of
within 48 hours on their arrival at the emergency       sepsis. The judgment was made after a review of
room of our hospital. The episodes were identified      the patient’s clinical findings, and based on the
by daily surveillance of microbiology laboratory        numbers of positive blood cultures, the results of
records. Each patient was observed prospectively        cultures of material from other body sites, imaging
from the time the blood culture became positive,        results, surgical findings, and the clinical course.
unless the patient had been discharged.                 Organisms that are commonly recovered from
                                                        the environment or skin flora (mainly coagulase-
Methods                                                 negative Staphylococci, non-hemolytic Streptococci,
     Data collected on all patients with a proven       Corynebacterium species and Bacillus species)
                                                                     Community-acquired bloodstream infection   95

were mostly determined to be contaminants, unless         or requiring help with most daily activities. The
isolated from 2 or more separate blood culture sets       performance status of the patient was initially
and there was strong evidence indicating a high           assessed according to the Eastern Cooperative
probability of a true BSI.                                Oncology Group (ECOG) scales (10): ECOG scale
    The source of BSI was defined as an obvious           3 or 4 was considered a predisposing factor
focus with signs of infection, with the same              for infection. Immunosuppressive therapy was
organism being isolated from the blood and from           considered as a comorbidity only if a patient had
the local site of infection or, when no organism was      received the equivalent of 30 mg of prednisolone
isolated, an infectious focus was found on (1) a          daily or cytotoxic agents daily for at least 2 weeks
physical examination, which was compatible with           before the bloodstream infection.
a diagnosis of cellulites, subcutaneous or cutaneous           The term “sepsis” defined the condition of all
abscess, or deep neck infection or (2) imaging            patients with a positive blood culture with a clinical
which was compatible with a new pulmonary                 systemic inflammatory response syndrome (SIRS),
infiltrate, acute cholecystitis, or intestinal            which included a varying combination of abnormal
perforation. The source of infection was designated       body temperature (>38ºC or < 36ºC), tachypnea
as one of following: pulmonary, intra-abdominal,          (>20/min.), tachycardia (>90/min.), and changes in
urinary, soft tissue/skin, intravascular device-related   blood leukocyte count (WBC>12000/μL or <4000/μL)
(intravascular catheter, implanted port, double lumen     or morphology (immature leukocytes >10%).
catheter, or tunneled cuffed catheter), or bone/joint.    “Severe sepsis or septic shock” was characterized
      A series of comorbid medical conditions             by a combination of sepsis with signs of acute
frequently cited as predisposing to infection, and        organ dysfunction or hypotension (SBP<90)
obtained through a review of medical records,             which persisted after a fluid challenge of 20-30
were determined to be either present or absent in         ml/kg, in accordance with the conclusions of the
each episode; these included diabetes mellitus,           American College of Chest Physicians/Society of
chronic obstructive pulmonary disease, underlying         Critical Care Medicine (ACCP/SCCM) Consensus
malignancy, liver cirrhosis, renal insufficiency,         Conference (11).
end-stage renal disease requiring hemodialysis,                Antibiotic therapy was considered appropriate
cardiovascular disease, a neurological disorder with      when empirical antibiotics which had been
poor performance, or immunosuppressive therapy.           administrated within 24 hours in appropriate doses
      Underlying malignancy was defined as a solid        were shown to be effective against the causative
or hematological malignancy diagnosed or treated          pathogens by in vitro susceptibility testing. We also
in the past 5 years. Renal insufficiency was defined      recorded the outcome of each case, and whether the
as a serum creatinine level greater than 2.5 mg/          patient was discharged alive or died.
dL. Liver cirrhosis included alcoholic and non-
alcoholic liver cirrhosis. Cardiovascular disease was     Statistical analysis
defined as clinical documentation of at least one of           The X2 and Fisher’s extact tests were used to
following diseases: hypertension, coronary artery         compare categorical variables. The primary data analysis
disease, or congestive heart failure with an ejection     compared hospital mortality with survival; categorical
fraction <40% on echocardiography. Neurological           values were expressed as a percentage of the group from
disorders included major cerebral vascular accident       which they were derived. All p values were two tailed; p
(CVA), dementia, and spinal injury with bedridden         values less than 0.05 were considered significant.
96    J Emerg Crit Care Med. Vol. 18, No. 3, 2007

     Univariate logistic regression was performed      Pseudomonas aeruginosa (n=7, 4.3%) were the
to determine the association between several           most common pathogens. Coagulase-negative
predisposing factors and a single variable             Staphylococci (n=35, 59.3%), and Corynebacterium
of outcome (hospital mortality), and their             species (n=8, 13.5%) constituted most of the
corresponding 95% confidence intervals (95%CI)         contaminants. Only 4 (6%) gram-negative
were calculated. The odds ratios were calculated       organisms and 1 (1.7%) anaerobic organism were
according to the lowest risk categories; an odds       identified as contaminants in cases where there was
ratio of death of 1.0 was subjectively assigned        only one positive culture and the clinical situation
to the lowest risk category within each variable       was not compatible with a real infection.
studied. All statistical analyses were done with           Of the 92 true-positive cases of BSI, 2 patients
SPSS for Windows Release 9.0 (SPSS, Chicago,           were excluded from further analysis due to transfer
IL, USA). All p values obtained were considered        to another hospital. The demographic and clinical
significant when less than 0.05.                       characteristics of the remaining 90 patients with
                                                       community-acquired BSI can be seen in Table 2.
                     Results                           Fifty-one patients (56.6%) were women, 39 (43.3%)
                                                       were men. and the median age was 70.1 years
    From Jan. 1, 2005 to Dec. 31, 2005, 1312           (range, 39 to 97 years). Diabetes mellitus was the
blood cultures were performed in the ER. A total       most common comorbid medical condition (n=38,
of 220 (16.8%) microorganisms were isolated.           42.2%), followed by neurological disorder with
There were a total of 220 isolates in 144 positive     poor performance (n=22, 24.4%), cardiovascular
BSI episodes for clinical evaluation. After careful    disease (n=19, 21.1%), liver cirrhosis (n=14,
analysis, 161 isolates in 92 cases denoted true-       15.5%), and underlying malignancy (n=11, 12.2%).
positive BSI; the other 59 isolates in 52 cases were   Abnormal temperature (n=77, 85.5%) was the
contaminants. During this interval, there were         most frequent presenting SIRS criterion, followed
25,000 adult ER visits; thus the incidence of BSI      by tachycardia and tachypnea (both n=63, 70%);
was 3.68 cases/1000 adult ER visits.                   an abnormal white blood cell count was found
    Of the 220 isolates (Table 1), 77 (35%)            in 58 cases (64.4%), and immature band cells
were gram-positive organisms, and 138 (62.7%)          > 10% were documented in 50 (55.5%) cases.
were gram-negative; the remaining 5 (2.3%)             The urinary tract was the most common source of
were anaerobic. Moreover, certain organisms            BSI (n=43, 47.8%), followed by intra-abdominal
such as coagulase-negative Staphylococci,              sites (n=20, 22.2%). Sepsis was observed in 56
Corynebacterium species, and Bacillus species were     episodes (62.2%), severe sepsis or septic shock
predominantly identified as contaminants, while        was identified in 29 cases (32.2%), and 5 (5.5%)
Staphylococcus aureus, most of the gram-negative       patients had no signs of sepsis. Administration of
bacilli, and Bacteroid fragilis rarely presented       empirical antibiotics within the first 24 hours after
as contaminants. In the 161 true-positive blood        admission was appropriate in 74 (82.2%) cases. The
cultures, Escherichia coli (n=74, 46%), Klebsiella     over-all hospital mortality rate was 15.5%.
species (n=21, 13%), Staphylococcus aureus                  There were 14 deaths among the 90 cases
(n=13, 8%): methicillin-resistant Staphylococcus       of community-acquired BSI in the ER. The
aureus (n=4), methicillin-sensitive Staphylococcus     characteristics of those who died are compared with
aureus (n=9), Proteus species (n=11, 6.8%), and        those who survived during their hospital course in
                                                                 Community-acquired bloodstream infection   97

Table 1   Number of microorganisms isolated from blood culture sets
                                                                           No. of isolate
Microorganism (No. of isolate)                                  True pathogen           Contaminant
Gram-positive (77)
          Staphylococcus aureus (14)
          MRSA (4)                                                     4                       0
          MSSA(10)                                                     9                       1
          Coagulase negative staphylococcus (37)                       2                      35
          Streptococcus pneumoniae (2)                                 2                       0
          Viridans streptococcus (1)                                   0                       1
          Group D nonenterococcus (2)                                  2                       0
          Enterococcus faccalis (2)                                    0                       2
          Group A,B,G Streptococcus (5)                                4                       1
          Bacillus sp (5)                                              0                       5
          Corynebacteriump sp (8)                                      0                       8
          Stomatococcus mucilaginosus (1)                              0                       1
Gram-negative (138)
          Escherichia coli (76)                                       74                       2
          Klebsiella species (21)                                     21                       0
          Acinetobacter speciesi (4)                                   3                       1
          Proteus species (11)                                        11                       0
          Pseudomonas aeruginosa (7)                                   7                       0
          Aeromonas veronii (2)                                        2                       0
          Citrobacter freundii (2)                                     2                       0
          Enterobacter cloacae (3)                                     2                       1
          Moraxella sp (1)                                             1                       0
          Morganella morganii (1)                                      1                       0
          Salmonella enteritidis group D (4)                           4                       0
          Vibrio vulnificus (4)                                        4                       0
          Shewanella putrefaciens (2)                                  2                       0
Anaerobic (5)
          Bacteroides fragilis (4)                                     4                       0
          Fusobacterium nucleatum (1)                                  0                       1
Total     220                                                        161                      59
MRSA: Methicillin-resistant Staphylococcus aureus; MSSA: Methicillin-sensitive Staphylococcus aureus

Table 3 and Table 4). Age was not associated with      (pulse rate > 90/min.), tachypnea (respiratory rate
mortality among the patients with BSI, whereas the     >20/min.), abnormal WBC count (>12000/μL or
proportion of male and female patients who died        <4000/μL), and elevated percentages of bands
was similar. The relationship between comorbid         (immature leukocytes >10%) were not significantly
medical conditions and mortality revealed that         associated with a poor outcome. Similarly, there
liver cirrhosis was significantly associated with a    was no statistically significant difference in
poor outcome (OR: 6.37, p=0.005). Abnormal body        mortality rates between Gram-positive and Gram-
temperature (BT >38ºC or <36ºC), tachycardia           negative pathogens. However, the source of
98    J Emerg Crit Care Med. Vol. 18, No. 3, 2007

Table 2   Demographic and clinical characteristics of 90 community-acquired bloodstream infection episodes
Age, years                                                        70.5 14.8
Male                                                               39 (43.3)
Comorbid Medical condition*
   Diabetes mellitus                                               38 (42.2)
   COPD                                                              4 (4.4)
   Underlying malignancy                                            11 (12.2)
   Renal insufficiency                                               8 (8.9)
   ESRD                                                              8 (8.9)
   Cardiovascular disease                                          19 (21.1)
   Liver cirrhosis                                                 14 (15.5)
   Neurologic disorder, poor performance                           22 (24.4)
   Immumosuppressine therapy                                         3 (3.3)
Clinical and laboratory profile
   Temperature > 38℃ or < 36℃                                      77 (85.5)
   Heart rate > 90/min                                             63 (70)
   Respiratory rate > 20/min                                       63 (70)
   WBC > 12000 or < 4000/μL                                        58 (64.4)
   Bandemia > 10%                                                  50 (55.5)
Source of infection
   Respiatory tract                                                  6 (6.6)
   Intra-abdomen                                                   20 (22.2)
   Urinary tract                                                   43 (47.8)
   Soft tissue/skin                                                10 (11.1)
   Catheter related                                                10 (11.1)
   Bone/Joint                                                        1 (1.1)
Type of microorganism
   Gram-posifive                                                    11 (12.2)
   Gram-negative                                                   72 (80)
   Anaerobic                                                         3 (3.3)
   Polymicorbial                                                     4 (4.4)
Severity of systemic response
   No sepsis                                                         5 (5.5)
   Sepsis                                                          56 (62.2)
   Severe sepsis or septic shock                                   29 (32.2)
Appropriate empiric antibiotic therapy                             74 (82.2)
Hospital mortality                                                 14 (15.5)
*: One case may has no or more than one type of comorbid medical condition.
COPD: chronic obstructive pulmonary disease; ESRD: end-stage renal disease requiring hemodialysis.
Continuous variables are given as mean    standard deviation
Categorical variables are given as number (percentage)
                                                                    Community-acquired bloodstream infection   99

Table 3   Characteristics of patients with episodes of community-acquired bloodstream infections
                                       Survival         Mortality             Odds ratio
Variable (number)                       n=76             n=14                 (95% CI)       P-value
Age, yrs                              70.5 12.9        69.1 16.8                               NS
  <60 (20)                            15(19.7)         5 (35.7)        3.8 (0.8, 17.9)         NS
  60-75 (37)                          34 (44.7)        3 (21.4)        1.0
  >75 (33)                            27 (35.5)        6 (42.8)        2.5 (0.6, 11.0)         NS
  Female (51)                         44 (57.9)        7 (50)          1.0
  Male (39)                           32 (42.1)        7 (50)          1.4 (0.4, 4.3)          NS
  Diabetes mellitus (38)              35 (46.0)        3 (21.4)        0.3 (0.1, 1.2)          NS
  COPD (4)                            4 (5.2)          0               0                       NS
  Malignancy (11)                     7 (9.2)          4 (28.6)        3.9 (0.9, 15.9)         NS
  Renal insufficiency (8)             8 (10.5)         0               0                       NS
  ESRD (8)                            8 (10.5)         0               0                       NS
  Cardiovascular disease (19)         18 (23.7)        1 (7.1)         0.2 (0.02, 1.2)         NS
  Liver cirrhosis (14)                8 (10.5)         6 (42.8)        6.4 (1.8, 23.1)        0.005
  Neurological disorder (22)          19 (25.0)        3 (21.4)        0.8 (0.2, 3.2)          NS
  Immunosuppressive (3)               3 (4.0)          0               0                       NS
  Gram-positive (11)                  10 (13.1)        1 (7.1)         1.0
  Gram-negative (72)                  61 (80.2)        11 (78.6)       1.8 (0.2, 15.5)         NS
  Anaerobic (3)                       2 (2.6)          1 (7.1)         5.0(0.2, 117.9)         NS
  Polymicrobial (4)                   3 (3.9)          1 (7.1)         3.3 (0.1, 70.9)         NS
Clinical Parameters*
  Abnormal temperature (77)           65 (85.5)        12 (85.7)       1.0 (0.1, 5.2)          NS
  Tachycardia (63)                    51 (67.1)        12 (85.7)       2.9 (0.6, 14.1)         NS
  Tachypnea (63)                      53 (69.7)        10 (71.4)       1.1 (0.3, 3.8)          NS
  WBC count change (58)               47 (61.8)        11 (78.6)       2.3 (0.6, 8.8)          NS
  Bandemia (50)                       39 (51.3)        11 (78.6)       3.5 (0.9, 13.5)         NS
*: Abnormal temperature (>38℃ or < 36 ℃); Tachycardia (>90/min.); Tachypnea (>20/min.);
WBC count change (WBC>12000/μL or <4000/μL ); bandemia (immature leukocytes>10%)
COPD: chronic obstructive pulmonary disease; ESRD, end-stage renal disease requiring hemodialysis;
NS: Not Significant
Continuous variables are given as mean ± standard deviation
Categorical variables are given as number (percentage)
100    J Emerg Crit Care Med. Vol. 18, No. 3, 2007

Table 4    Comparison of infectious source, severity of systemic response, and appropriate antibiotic therapy
           of bloodstream infection episodes with the outcome
                                                Survival         Mortality          Odds ratio
Variable (number)                                n=76             n=14               (95%CI)             P-value
Infectious source
  Urinary (43)                                  42   (55.3)        1    (7.1)      1.0
  Intraabdom1nal (20)                           11   (14.5)        9   (64.3)     34.4 (3.9, 300)         0.001
  Pulmonary (6)                                  3    (3.9)        3   (21.4)     42.0 (3.3, 536)         0.004
  Soft tissue/Skin (10)                          9   (11.8)        1    (7.1)      4.6 (0.2, 81.8)         NS
  Catheter (10)                                 10   (13.2)        0               0                       NS
  Bone/Joint (1)                                 1    (1.3)        0               0                       NS
Systemic response
  No sepsis (5)                                  5 (6.6)          0                0                       NS
  Sepsis (56)                                   55 (72.4)         1 (7.1)          1.0
  Severe sepsis or septic shock (29)            16 (21.0)        13 (92.8)        44.7 (5.4, 368)        <0.001
Empiric antibiotics
  Appropriate (74)                              63 (82.9)        11 (78.6)         1.0
  Inappropriate (16)                            13 (17.1)         3 (21.4)         1.3 (0.3, 5.4)          NS
NS: Not Significant
Categorical variables are given as number      (percentage)

infection and severity of BSI were found to be                of BSI due to Klebsiella species, Pseudomonas
strongly associated with mortality; mortality was             aeruginosa, or Staphylococcus aureus (see Table 6).
significantly higher in patients who presented with                The susceptibility of the Gram-positive
severe sepsis or septic shock (OR: 44.7, p<0.001),            pathogens is shown in Table 7. A high rate of
and patients with an intra-abdominal site (OR: 34.4,          resistance to erythromycin was seen in these
p=0.001) or pneumonia (OR: 42, p=0.004) as the                microorganisms. Resistance to penicillin was found in
infectious source.                                            all strains of Staphylococcus aureus, but this antibiotic
     The distribution of pathogens in community-              was effective against isolates of Streptococcus
acquired BSI cases is shown in Table 5. The                   species. Trimethoprim-sulfamethoxazole and
most common causative organisms were                          ciprofloxacin maintained high activity against gram-
Escherichia coli in 49% (44 in 90 patients), and              positive pathogens except for half of the methicillin-
Klebsiella species in 8.8% (8 in 90 patients). Two            resistant Staphylococcus aureus (MRSA) isolates.
microorganisms were isolated in all 4 cases of                Glycopeptides (teicoplanin and vancomycin) were
polymicrobial bacteremia. The majority of cases of            the most active agents against all gram-positive
Staphylococcus aureus bacteremia were associated              pathogens in our study. In addition, all isolates of
with intravascular devices and soft tissue infections,        Bacteroid fragilis were resistance to penicillin and
whereas Escherichia coli BSI was most often                   clindamycin, but susceptible to cefmetazole and
associated with urinary tract and intra-abdominal             metronidazole.
infections. The highest fatality rates by pathogen                 Table 8 shows the susceptibility of gram-
were observed among patients with BSI due to                  negative pathogens causing BSI in the study.
Acinetobacter baumannii. Conversely, no one died              Meropenem and piperacillin/tazobactam were
                                                                       Community-acquired bloodstream infection 101

Table 5   The causative organisms and the sources of infections in 90 episodes with bloodstream infections
                                                          No. of episode                        Source*
Pathogen                                                       (n=90)                      (no. of episode)
  MRSA                                                           2                 IV (1), Bone (1)
  MSSA                                                           5                 IV (2), S (3)
  Coagulase-negative staphylococcus                              1                 IV (1)
  Streptococcus pneumoniae                                       1                 Res (1)
  Group A, B,G Streptococcus                                     2                 IV (2)
  Escherichia coli                                              44                 Abd (8), U (36)
  Klebsiella species                                             8                 Abd (2), U (4), S (2)
  Acinetobacter speciesi                                         2                 Res (2)
  Proteus species                                                4                 Abd (1), U (3)
  Pseudomonas aeruginosa                                         4                 Res (1), IV (3)
  Aeromonas veronii                                              1                 Abd (1)
  Enterobacter cloacae                                           2                 U (1), S (1)
  Moraxella sp                                                   1                 Res (1)
  Morganella morganii                                            1                 Res (1)
  Salmonella enteritidis group D                                 2                 Abd (2)
  Vibrio vulnificus                                              2                 S (2)
  Shewanella putrefaciens                                        1                 Abd (1)
  Bacteroides fragilis                                           3                 Abd (2), S (1)
Polymicrobial**                                                  4                 Abd (2), IV (1), S (1)
 * Res: respiratory tract; Abd: intra-abdominal; U: urinary tract; S: soft tissue/skin; IV: intravascular catheter;
   Bone: bone/joint.
** Pathogens; Escherichia coli & Group D nonenterococcus. Escherichia coli & Klebsiella pneumoniae.
   Escherichia coli & Citrobacter freundii. Escherichia coli & Pseudomonas aeruginosa.

active against all gram-negative pathogens.                 respectively. For Klebsiella species, two of the
Microorganisms which frequently induced severe              21 isolates were extended-spectrum β-lactamase
sepsis or septic shock such as Acinetobacter species        (ESBL)-producing, leading to a 9.5% resistance rate
and Aeromonas species, were most resistant to               against most available antibiotic agents. For Proteus
cefazolin, ampicillin and amoxillin clavulanate;            species and Pseudomonas aeruginosa, a wide range
aminoglygosides (amikacin and gentamicin)                   of resistance rates was observed.
maintained high activity against gram-negative
pathogens except for Proteus species (72.7%
of Proteus species isolates were resistant to
gentamicin). For E. coli, less than 10% of isolates
                                                                 I n t h e p r e s e n t s t u d y, w e r e p o r t t h e
were resistant to cephalosporins, but 64.8%, 32.4%,
                                                            characteristics of ER patients who acquired a BSI in
and 14.8% of E. coli isolates were resistant to
                                                            the community; the hospital mortality rate reaches
ampicillin, amoxillin clavulanate, and ciprofloxacin,
                                                            15.5%, which confirms recent data of a 15% to 20%
102    J Emerg Crit Care Med. Vol. 18, No. 3, 2007

Table 6   Correlation of pathogens of bloodstream infection with the outcome
                                             Fatal rate (%)            Survivors             Mortalities
Pathogen                                          n=90                   n=76                  n=14
  MSSA                                             0                       5                     0
  MRSA                                             0                       2                     0
  Others*                                         25%                      3                     1
  Escherichia coli                                 9.3%                  39                      4
  Klebsiella species                               0                       9                     0
  Pseudomonas aeruginosa                           0                       4                     0
  Proteus species                                 25%                      3                     1
  Acinetobacter baumannii                        100%                      0                     2
  Enterobacter cloacae                             0                       2                     0
  Salmonella enteritidis                          50%                      1                     1
  Vibrio vulnificus                               50%                      1                     1
  Others**                                        50%                      2                     2
  Bacteroides fragilis                            33.3%                    2                     1
Polymicrobial***                                  25%                      3                     1
MRSA: Methicillin-resistant Staphylococcus aureus; MSSA: Methicillin-sensitive Staphylococcus aureus
  * Pathogen (no. of episode/no. of mortality); Coagulase negative staphylococcus (1/0).
     Group G beta streptococcus (1/0). Streptococcus agalactiae (1/0). Streptococcus pneumoniae (1/1)
 ** Pathogen (no. of episode/no. of mortality); Aeromonas veronii (1/1). Moraxella sp (1/0).
     Morganella morganii (1/0). Shewaanella putrefaciens (1/1)
*** Pathogens (no. of episode/no. of mortality); Escherichia coli & Group D nonenterococcus (1/1).
     Escherichia coli & Klebsiella pneumoniae (1/0). Escherichia coli & Citrobacter freundii (1/0).
     Escherichia coli & Pseudomonas aeruginosa (1/0)

mortality rate associated with community-acquired         presentation with severe sepsis or septic shock, and
BSI(4,12,13).                                             an intra-abdominal or pulmonary source of BSI.
     Several demographic and clinical conditions          Most of our findings were consistent with results
have been analyzed to estimate significant factors        from previously published data. However, sex, age,
affecting the prognosis of patients with BSI(4,6,12-4).   systemic inflammatory response syndrome (SIRS)
Patients with predisposing conditions such as old         criteria, and inappropriate antimicrobial treatment
age, “rapidly fatal” underlying disease, leukopenia,      had no influence on mortality.
pneumonia, an intra-abdominal source of infection,             The etiologic pathogens responsible for
a high APACHE score, shock, and inadequate                community-acquired BSI have been elucidated in
empiric antibiotic treatment had high mortality           numerous studies. These data revealed that Gram-
rates. In our study population, 4 variables were          negative bacteria, including E. coli, Klebsiella
associated with an increased risk of death after          species, Pseudomonas aeruginosa and Proteus
community-acquired BSI: comorbid liver cirrhosis,         species, were implicated in 35% to 85% of
                                                                 Community-acquired bloodstream infection 103

Table 7    Resistance to antimicrobial agents of the most frequent gram-positive pathogen causing
           community-acquired bloodstream infection
                         S. pneumoniae            MSSA                 MRSA           Steptocococcus spp*
                              (n=2)                (n=9)                (n=4)                (n=6)
TMP-SMX                         0                    0                   50%                  ND
Ciprofloxacin                  ND                    0                   50%                  ND
Erythromycin                  100%                 100%                100%                  100%
Oxacillin                      ND                    0                 100%                   ND
Penicillin                     ND                  100%                100%                    0
Vancomycin                      0                    0                    0                    0
Teicoplanin                     0                    0                    0                    0
n = number of isolate
*Pathogen; Group A, B, G Streptococcus (n=4). Group D nonenterococcus (n=2)
ND: not determinated; TMP-SMX: Trimethoprim-sulfamethoxazole;
MRSA: Methicillin-resistant Staphylococcus aureus; MSSA: Methicillin-sensitive Staphylococcus aureus;

Table 8   Resistance to antimicrobial agents of the most frequent and fatal gram-negative pathogen causing
          community-acquired bloodstream infection
                        E. coli        Klebsiella spp.    Proteus spp.     P. aeruginosa      Others*
                        (n=74)             (n=21)            (n=11)            (n=7)           (n=7)
Cefazolin                8.1%                19%               18%              ND             71.4%
Cefotaxime               5.4%               9.5%               18%              ND               0
Amikacin                 2.7%               9.5%                0              28.5%             0
Ampicillin              64.8%               100%              100%              ND             71.4%
AMX-CLV                 32.4%                19%               18%              ND             71.4%
Cefuroxim                5.4%               9.5%               18%              ND               0
Ciprfloxacin            14.8%                 0                 0              28.5%           28.5%
Gentamicin              12.2%               9.5%             72.7%             28.5%             0
Pip-Tazo                   0                  0                 0               ND               0
Cefepime                   0                9.5%               18%             28.5%             0
Meropenem                  0                  0                 0                0               0
n=number of isolate
ND: not determinated; AMX-CLV, amoxcillin-clavulanate; Pip-Tazo, piperacillin-tazobactam
*Fatal pathogen (No. of isolate /No. of episode/No. of mortality); Aeromonas veronii (2/1/1).
Shewanella putrefaciens (2/1/1). Acinetobacter baumannii (3/2/2)
104    J Emerg Crit Care Med. Vol. 18, No. 3, 2007

community-acquired BSI cases, Gram-positive                has been less than 5% in our hospital in recent years.
cocci, particularly Staphylococcus aureus and                  Compared with other studies, the most specific
Streptococcus pneumoniae, accounted for 15% to 40%,        finding in our study was that comorbid liver
and 10% to 15% of cases were polymicrobial(12-15).         cirrhosis was the most important predictive variable
Our study shows that Gram-negative organisms were          of hospital mortality in patients with community
implicated as major pathogens (84%) in patients            acquired BSI (OR: 6.37, p=0.005). There were
with BSI, based on emergency room data; this               fourteen patients with liver cirrhosis in our study
finding differs from the nearly equal prevalence           group. The mortality rate reached 43% in this group
between Gram-positive and Gram-negative                    and patients died at a relatively young age. BSI due
pathogens found in other studies. The distributions        to Acinetobacter baumannii, Aeromonas veronii, E.
of pathogens in our study also varied greatly from         coli, Salmonella enteritidis, and Vibrio vulnificus
recent reports (12-14). We compared our data with          were associated with mortality in cirrhotic patients.
that in previous reports on the major pathogens            This result is similar to previous reports (18,19) .
responsible for community-acquired BSI since               Eleven of the fourteen cirrhotic patients (78.5%)
the 1980s. E. coli was the leading pathogen in our         had severe sepsis or septic shock on arrival in the
study, and accounted for almost 50% of all cases:          ER; proposed reasons such as impaired systemic
the prevalence of E. coli was lower than that found        clearance of endotoxins, subtle changes in clinical
in the Isphahani et al. study in 1987 (80%), but           symptoms, hepatic encephalopathy or alcoholism-
much more predominant than that reported by                induced delay in awareness of symptoms, and lack
Lark et al. in 2001 (15%). Staphylococcus aureus           of attention from family may explain why BSI
was the third leading pathogen, and caused 8% of           patients with liver cirrhosis come to the ER at a
cases in our study, compatible with the report by          relatively late stage of sepsis(20).
Pederson et al. in 2003, but less than a 30% rate                The Convention of the American College of
of Staphylococcus aureus community-acquired                Chest Physicians/Society of Critical Care Medicine
BSI reported by Lark et al. in 2001. These changes         (ACCP/SCCM) Consensus (10) in 1991, and the
in the microbiological etiology of community-              SCCM/ESICM/ACCP/ATS/SIS International Sepsis
acquired pathogens in different decades and                Definitions Conference in 2001(21), have defined the
regions may reflect the variant geographic factors,        terms “systemic inflammatory response syndrome”
numbers of immunocompromised patients, and                 (SIRS), “sepsis”, “severe sepsis”, and “septic
medical resource usage. Klebsiella species were the        shock”. These definitions have been well accepted
second most common organism in our series, and             by clinicians; however, the prognostic value of the
Pseudomonas aeruginosa and Proteus species tied            categorizations has remained controversial(22,23). In
for fourth. These results are consistent with other        our study, the variables used to define SIRS, such
studies, and reflect the fact that the prevalence of       as temperature, heart rate, respiratory rate, WBC
Gram-negative bacillus species has remained stable         count, and even elevated percentages of bands,
in cases of community-acquired BSI all over the            did not have predictive power for mortality. Many
world. We found a relatively low prevalence of             reports have supported the limitations of SIRS
community-acquired Streptococcus pneumoniae                criteria in predicting microbial infection (24,25). In
BSI in our study, compared with other reports (2%          contrast to SIRS criteria, the importance of acute
vs. 7-20%); the reasons are uncertain, but the yearly      organ dysfunction as a prognostic factor has long
prevalence of Streptococcus pneumoniae in blood cultures   been established in septic patients(26,27); acute organ
                                                                        Community-acquired bloodstream infection 105

dysfunction that developed after the onset of BSI         the survival of patients with BSI is greater with
markedly affected the outcome. In our study, 5            severe patient conditions(14). In addition, Kang et al.
cases of bacteremia did not fulfill 2 or more SIRS        found that inadequate initial antimicrobial therapy
criteria; all of these patients survived. Only 1 of the   had no influence on fatalities in BSI patients
56 patients who never developed organ dysfunction         with a pancreaticobiliary tract, urinary tract, or
in the ER died. Severe sepsis or septic shock             intravenous catheter source of infection, when
occurred in 29 patients during the ER course, and         compared to those whose source of infection was
13 (44.8%) died. After analysis of the correlation        the lung, peritoneum, or an unknown site. Most of
between the degree of systemic response and               the former patients underwent catheter-removal or
mortality, we found that acute organ dysfunction          decompression of the obstructive tract (33). These
in septic patients (qualifying as severe sepsis or        results are in agreement with our finding that in
septic shock) has a strong prognostic significance        community-acquired BSI cases, the severity of the
for hospital mortality (OR: 44.7, p<0.001), which         systemic response and source of infection may be
confirms the findings of previous reports(28,29).         more responsible for hospital mortality.
     We analyzed the correlation between infectious               Because of the high mortality associated
source and mortality, taking urinary tract infection      with community-acquired BSI, clinicians in the
as the reference category (OR: 1). Infection of           ER setting must be able to provide adequate
the lower respiratory tract (OR: 42, p=0.004) and         intervention in patients with suspected sepsis.
intra-abdominal infection (OR: 34.4, p=0.001)             Moreover, common organisms and their resistance
were associated with a marked risk of death. This         patterns within the geographic area need to be
result was compatible with that of previous studies       taken into account when selecting antibiotic
analyzed (4,12-14). On the other hand, no organism        agents for empiric treatment. The susceptibility of
was significantly associated with an increased            pathogens in our study is listed in Tables 7 and 8.
risk of mortality in our study. Taken together, our       Observations based on susceptibility tests showed
results imply that the influence of the individual        a high prevalence (30%) of methicillin-resistant
microorganism is less important than host factors         Staphylococcus aureus, 100% erythromycin-
(underlying disease and systemic response) and            r e s i s t a n t S t re p t o c o c c u s s p e c i e s , i n c r e a s e d
infectious sources in regard to hospital mortality in     ciprofloxacin-resistant Gram-negative bacilli,
cases of community-acquired BSI.                          and ESBL-producing Klebsiella pneumoniae
     In our study, inappropriate empiric antibiotic       and Proteus mirabilis in community-acquired
therapy in the ER was not found to be significantly       BSI in this study. All of the above findings are
associated with increased hospital mortality in           important issues in our community. Many factors
patients with community-acquired BSI. Several             have been shown to contribute to the development
reports have noted that appropriate antimicrobial         of antimicrobial resistance (34) ; public behavior
therapy does have a favorable effect on the               and frequent cross-infection with resistant
outcomes of critically ill patients with a BSI(30-32).    microorganisms in nursing homes are the most
However, the patient groups in their studies were         likely reasons for the relatively high prevalence of
different from ours: their patients all had severe        some resistant pathogens in our region.
sepsis, and the potential risk factors were rarely            In conclusion, we have reported on a
seen in the community. Valles et al. found that           12-month observation study of clinical conditions
the influence of appropriate antibiotic therapy on        and outcomes of adult patients with community-
106     J Emerg Crit Care Med. Vol. 18, No. 3, 2007

a c q u i r e d B S I i n t h e e m e rg e n c y r o o m o f a         Infect Dis 1992;15:866-73.
community hospital. Our study also provides data                  7. Weinstein MP, Reller B, Murphy JR, et al. The
on antimicrobial resistance patterns in our local                    clinical significance of positive blood cultures:
area. Community-acquired BSI still carries a high                      a comprehensive analysis of 500 episodes of
mortality rate; an ominous prognosis is associated                     bacteremia and fungemia in adults. I: laboratory
with comorbid liver cirrhosis, a respiratory tract                     and epidemiologic observations; and II: clinical
or intra-abdominal source of infection, and the                        observations, with special reference to factors
presence of acute organ dysfunction or shock.                        influencing prognosis. Rev Infect Dis 1983;5:35-70.
These results should be taken into account when                   8. B a r o n E J, C i t r o n D M. A l g o r i t h m f o r
making clinical decisions for patients with                            identification of anaerobic bacteria. In: Murray
suspected sepsis in the ER setting.                                    PR, Baron EJ, Pfaller MA, Tenover FC, Yolken
                                                                       RH, eds. Manual of clinical microbiology. 7th
                     References                                        ed. Washington, DC: ASM Press, 1999:652-3.
                                                                  9.   National Committee for Clinical Laboratory
  1. Reimer LG, Wilson ML, Weinstein MP. Update                        S t a n d a r d s. P e r f o r m a n c e s t a n d a r d s f o r
     on detection of bacteremia and fungemia.                          antimicrobial susceptibility testing. 12 th
     Clinical Microbiology Review 1997; 10:444-65.                     informational supplement. NCCLS document
  2. Diekema DJ, Beekmann SE, Chapin KC, et al.                        M100-S12 [ISBN 1-56238-454-6]. Wayne, PA:
     Epidemiology and outcome of nosocomial and                        National Committee for Clinical Laboratory
     community-onset bloodstream infection. J Clin                     Standards. 2002.
     Microbiol 2003;41:3655-60.                                  10.   Oken MM, Creeth RH, Tormey DC, et al.
  3. Bates DW, Pruess KE, Lee TH. How bad are                          Toxicity and response criteria of the Eastern
     bacteremia and sepsis? Outcomes in a cohort                       Cooperative Oncology Group. Am J Clin Oncol
     with suspected bacteremia. Arch Intem Med                         1982;5:649-55.
     1995;155:593-8.                                             11.   American College of Chest Physicians-
  4. M y l o t t e J M , K a h l e r L , M c C a n n C .               Society of Critical Care Medicine Consensus
     Community-acquired bacteremia at a teaching                       Conference. Definitions for sepsis and organ
     versus a nonteaching hospital: impact of                          failure and guidelines for the use of innovative
     acute severity of illness on 30-day mortality.                    therapies in sepsis. Crit Care Med 1992;
     American J Infection Control 2001;29:13-9.                        20:864-75.
  5. Weinstein MP, Towns ML, Quartey SM,                         12.   Pederson G, Schonbeyder HC, Sorensen HT.
     et al. The clinical significance of positive                      Source of infection and other factors associated
     blood cultures in the 1990s: a prospective                        with case fatality in community-acquired
     comprehensive evaluation of the microbiology,                     bacteremia--- a Danish population-based cohort
     epidemiology, and outcome of bacteremia                           study from 1992 to 1997. Clin Microbiol Infect
     and fungemia in adults. Clin Infect Dis 1997;                     2003;9:793-802.
     24:584-602.                                                 13.   Lark RL, Saint S, Chenoweth C, Zemencuk JK,
  6. Uzun O, Akalin HE, Hayran M, et al. Factors                       Lipsky BA, Plorde JJ. Four-year prospective
     influencing prognosis in bacteremia due to                        evaluation of community-acquired bateremia:
     Gram-negative organisms: evaluation of 448                        epidemiology, microbiology and patient
     episodes in a Turkish university hospital. Clin                   outcome. Diagnostic Microbiology and Infetion
                                                                               Community-acquired bloodstream infection 107

      Disease 2001;41:15-22.                                        23. Salvo I, D e Cian W, Musicco M, et al. The
14.   Valles J, Rello J, Ochagavia A, Garnacho J,                       Italian SEPSIS study: preliminary results on
      Alcala MA. Community-acquired bloodstream                         the incidence and evolution of SIRS, sepsis,
      infection in critically ill adult patients: impact                severe sepsis and septic shock. Intensive Care
      of shock and inappropriate antibiotic therapy                     Med 1995;21:S244-9.
      on survival. Chest 2003;123:1615-24.                          24. Bossink AWF, Groeneveld ABJ, Hack CE, et
15.   Ispahani P, Pearson NJ, Greenwood D. An                           al. The clinical most response to microbial
      analysis of community and hospital-acquired                       infection in medical patients with fever. Chest
      bateremia in a large teaching hospital in the                     1999;116:380-90.
      United Kingdom Quarterly Journal of Med                       25. Fontanarosa PH, Kaeberlein FJ, Gerson LW, et
      1987;241:427-40.                                                  al. Difficulty in predicting bacteremia in elderly
16.   Hauy JB, Harthug S, Kalager T, Digranes                           emergency patients. J Emerg Med 1992;21:842-8.
      A, Solberg CO. Bloodstream infections at a                    26. A n n a n e D, A e g e r t e r P, J a r s-G u i n c e s t r e
      Norwegian University hospital, 1974-1979 and                      MC, Guidet B. Current Epidemiology of
      1988-1989: changing etiology, clinical features,                  septic shock. Am J Respir Crit Care Med
      and outcome. Clin Infect Dis 1994;19:246-56.                      2003;168:165-72.
17.   S t e i n b e rg J P, C l a r k C C, H a c k m a n B O.       27. A n g u s D, L i n d e-Z w i r b l e W T, L i d i c k e r
      N o s o c o m i a l a n d c o m m u n i t y-a c q u i r e d       J, C l e r m o n t G, C a r c i l l o J, P i n s k y M R.
      Staphylococcus aureus bacteremias from                            Epidemiology of severe sepsis in the United
      1980 to 1993: impact of intravascular devices                     States: analysis of incidence, outcome, and
      and methicillin resistance. Clin Infect Dis                       associated costs of care. Crit Care Med
      1996;23:255-9.                                                    2001;29:1303-10.
18.   KUO CH, Chang CS, Yang CY, Sheen IS, Liaw                     28. Alberti C, Brun-Bnisson C, Goodman SV, et al.
      YF. Bacteremia in patients with cirrhosis of the                  Influence of systemic inflammatory response
      liver. Liver 1991; 11:334-339                                     syndrome and sepsis on outcome of Critically
19.   F o r e m a n M G, M a n n i n o D M, M o s s M.                  ill infected patients. Am J Respir Crit Care
      Cirrhosis as a risk factor for sepsis and death:                  Med 2003;168:77-84.
      analysis of the national hospital discharge                   29. Bossink AW, Groeneveld J, Hack CE, Thijs
      survey. Chest 2003;124:1016-20.                                   LG. Prediction of mortality in febrile medical
20.   Barnes PF, Arevalo C, Chan LS, Wong SF,                           patients:how useful are systemic inflammatory
      Reynolds TB. A prospective evaluation of                          response syndrome and sepsis criteria? Chest
      bacteremic patients with chronic liver disease.                   1998;113:1533-41.
    Hepatology 1988;8:1099-103.                                     30. Ibrahim EH, Sherman G, Ward S, et al. The
21. Levy MM, Fink MP, Marshall JC, et al. 2001                          influence of inadequate antimicrobial treatment
      SCCM/ESICM/ACCP/ATS/SIS International                             of bloodstream infections on patient outcomes
      sepsis definitions conference. Crit Care Med                      in the ICU setting. Chest 2000;118:146-55.
    2003;31:1250-6.                                                 31. Leone M, Bourgoin A, Cambon S, Dubuc M,
22. Rangel Frausto MS, Pittet D, Costigan M,                            Albanese J, Martin C. Empirical antimicrobial
    et al. The natural history of the systemic                          therapy of septic shock patient: adequacy
    inflammatory response syndrome (SIRS): a                            and impact on the outcome. Crit. Care Med
    prospective study. JAMA 1995;273:117-23.                            2003;31:462-7.
108     J Emerg Crit Care Med. Vol. 18, No. 3, 2007

32. MacArthur RD, Miller M, Albertson T, et al.              impact of inappropriate initial amtimicrobial
      Adequacy of early empiric antibiotic treatment         therapy on ontcome. Antimcriobial Agents and
      and survival in severe sepsis: experience              Chemotherapy 2005;49:760-6.
      from the MONARCS trial. Clin Infect Dis            34. C i z m a n M. T h e u s e a n d r e s i s t a n c e t o
      2004;38:284-8.                                         antibiotics in the community. Int J Antimicrob
33. Kang CI, Kim SH, Park WB, et al. Bloodstream             Agents 2003;21:297-307.
    infections caused by antibiotic-resistant gram-
      negative bacilli: risk factors for mortality and

                                                                    1                2             3

                  2005      1   1              12       31

                       90                                           70.1             51
(42%)                                    (24.4%)                           (21.1%)            (15.5%)              (12.2%)
                  E.coli (47.8%)      Klebsiella         (10%)          Staphylococcus aureus (7.7%)
         (47.8%)                         (22.2%)             32%
                                              82.2%                             15.5%
(OR 6.37      P    0.005)                                                                          (OR 44.68   P    0.001)
                       (OR 42    P   0.004)                         (OR 34.36        P    0.001)

        95   10   19                 96    1   4
                                     1              2          3

       (04)8960128              (04)8966991
E-mail: 68302@cch.org.tw

To top