Keywords
bacteremia - community-acquired - healthcare-associated - hospital-acquired
Introduction
Bacteremia is a major pediatric health care problem. Despite great developments in
therapy and supportive care, bacteremia still represents a major cause of morbidity
and mortality.[1]
[2] In particular, gram-negative bacteremia has increasingly been reported among the
pediatric population.[3] Recently, community-onset bacteremia was reclassified, due to increasing numbers
of clinical outpatient treatments, into healthcare-associated (HCA) and community-acquired
(CA) bacteremia.[4]
[5]
[6]
[7]
[8] Previous studies that made a comparison between hospital-acquired nosocomial (HA)
gram-negative bacteremia and CA bacteremia found that there was an increased rate
of bacteremia due to drug-resistant pathogens.[9]
[10]
[11] Studying the varying microbiological patterns of bacteremia in children, gram-negative
septicemia was categorized in two studies to be up to 71.87 and 73%, respectively,
of the culture-positive cases.[12]
[13] Furthermore, recently, there has been an increase in the number of HCA gram-negative
bacteremia among hospitalized patients with high rate of multidrug-resistant (MRO)
bacteremia.[14]
[15] The objective of this study was to analyze the differences between the three types
of gram-negative bacteremia with regard to the comorbidities, clinical characteristics,
etiology, and outcomes among Saudi children in the King Khalid University Hospital.
Methods
Study Design
This study was conducted at the King Khalid University Hospital, a 1,100-bed hospital
with 135 monthly admissions in pediatric wards. It is a major teaching hospital in
Riyadh, Saudi Arabia, providing both primary and tertiary medical care. The study
was conducted between January 11, 2015, and December 27, 2015. Data were purposely
collected from pediatric patients below 15 years of age with gram-negative blood infections,
using daily computerized laboratory data and medical record information system. Standardized
data forms were used to record demographic details, including underlying diseases,
hospital unit, and exposure to the healthcare system in the previous years.
Statistical Analysis
All data were collected in an Excel sheet and were analyzed using the analysis of
variance (ANOVA) method for finding differences among the three types of bacteremia.
Results were considered significant if p < 0.05. Ethical approval for the study was obtained from the ethics committee of the
hospital.
Definitions
Gram-negative bacteremia is defined as the isolation of gram-negative bacilli in a
blood culture specimen. Clinically, significant bacteremia is defined as at least
one positive blood culture, together with clinical features compatible with systemic
inflammatory response syndrome. Patients were diagnosed with CA, gram-negative bacteremia
if their first positive blood culture results were obtained from blood samples drawn
within 48 hours after hospital admission. Cases of HCA bacteremia were diagnosed if
one or more of the following criteria was fulfilled: outpatient treatment (patient
treated in the clinic or emergency room) hemodialysis or intravenous chemotherapy
during the past 30 days; hospitalization for at least 1 day during the past 90 days;
home intravenous therapy or wound care during the past 30 days; or residence in a
long-term care facility.
In addition, HA infection was defined as an infection that occurred more than 48 hours
after admission to the hospital.
Identification and Antimicrobial Susceptibility Testing
Isolates of gram-negative bacteria were identified by standard microbiologic methods
in the microbiology laboratory using an automated identification system (Vitek System;
bioMerieux, Marcy l'Etoile, France). Susceptibilities to antimicrobial agents were
determined by the use of an automated susceptibility testing system (Vitek System;
bioMerieux). Extended spectrum β-lactamase (ESBL) production was confirmed by double
disk synergy testing, in accordance with the Clinical and Laboratory Standards Institute
(CLSI) standards .[15]
Results
Ninety-two children were diagnosed with gram-negative bacteremia, 64 (71.1%) with
HA bacteremia, 20 (21.1%) with CA bacteremia, and 8 (7.8%) with HCA bacteremia. The
most recorded criteria found in HA bacteremia were outpatient treatment in the past
30 days, intravascular chemotherapy in the past 30 days, and hospitalization for > 1
day in the past 90 days. Most of the patients' age was less than 1year ([Table 1]). There was no gender difference, and for the underlying comorbidities, there were
12 cases of renal disorder, 12 cases of gastrointestinal disorder, 8 cases of oncology
diseases, 6 cases of respiratory disorder, and 13 patients with premature immune deficiency,
and there were 3 patients with endocrine disorder ([Table 2]) Fever was significantly more common in HA type. Urinary tract infection constituted
the major clinical presentation for CA and HCA bacteremia. On the other hand, septic
shock/hypotension was the main clinical presentation in HA bacteremia indicating that
the gram-negative bacteremia is clinically significant. The most common gram-negative
bacteria causing bacteremia was Klebsiella pneumoniae (29.3%), which was isolated in cases with HA bacteremia. This was followed by Escherichia coli (isolated in two types: HA and CA bacteremia, [Table 3]). In CA type bacteremia, 25% of the isolated organisms were E. coli. The antimicrobial susceptibility of the isolated organisms of CA/HCA, and HA bacteremia
showed that majority of the organisms were non-ESBL in 90%, and 10% of the isolates
were ESBL. Three isolates were multi-resistant organism (MRO) among HA bacteremia
([Table 4]). In comparing the difference between ESBL and non-ESBL isolate among the three
types of bacteremia, there was no significant difference in the percentage of ESBL
and non-ESBL in CA and HA bacteremia (p = 0.928). However, there was a significant difference in the total frequency of isolates
between CA and HA bacteremia, regardless whether they were in the ESBL or non-ESBL
groups (p < 0.001, [Table 5]). However, there was a high mortality rate constituting 17 mortality cases (18.7%),
14 cases for HA bacteremia and 3 for CA bacteremia ([Table 6]).
Table 1
Age range of pediatric patients with gram-negative bacteremia
|
Age range (y)
|
Community acquired
|
Health care associated
|
Hospital acquired
|
p-Value
|
|
<1 ( No. 26)
|
4 (15.4%)
|
1 (3.8%)
|
21 (80.8%)
|
0.1770
|
|
1–3 ( No. 33)
|
10 (30.3%)
|
3 (9.1%)
|
20 (60.6%)
|
0.9870
|
|
3–6 ( No. 16)
|
3 (18%)
|
2 (12.5%)
|
11 (68.6%)
|
0.2774
|
|
6–12 ( No. 15)
|
3 (20%)
|
1 (6.7%)
|
11 (73.3%)
|
0.2150
|
|
12–14 ( No. 2)
|
1 (6.6%)
|
1 (6.7%)
|
0 (0%)
|
0.1
|
|
Total ( No. 92)
|
20 (21.1%)
|
8 (7.8%)
|
64 (71.1%)
|
0.7801
|
Table 2
Comorbidities and clinical characteristics of 92 pediatric patients with community-acquired
and health-associated and hospital-acquired gram-negative bacteremia
|
Characteristics
|
Community acquired
(n = 20)
|
Health associated
(n = 8)
|
Hospital acquired
(n = 64)
|
p-Value
|
|
Underlying disease
|
|
Hematological/malignancy[8]
|
4 (25%)
|
4 (25%)
|
4 (50%)
|
0.4795
|
|
Renal disease[12]
|
4 (17.5%)
|
4 (17.5%)
|
8 (67%)
|
0.6162
|
|
Gastroenterology[12]
|
8 (67%)
|
0 (0%)
|
4 (33%)
|
0.6162
|
|
Respiratory[6]
|
2 (33%)
|
0 (0%)
|
4 (67%)
|
0.9352
|
|
Endocrine[3]
|
0 (0%)
|
0 (0%)
|
3 (100%)
|
|
|
Immune deficiency[1]
|
0 (0%)
|
0 (0%)
|
1 (100%)
|
|
|
Prematurity[13]
|
0 (0%)
|
0 (0%)
|
13 (100%)
|
|
|
Clinical presentation
|
|
Fever54
|
18 (22%)
|
6 (11%)
|
36 (67%)
|
0.0375
|
|
Septic shock/hypotension26
|
2 (8%)
|
0 (0%)
|
24 (92%)
|
0.0279
|
|
Urinary tract infection[12]
|
10 (50%)
|
4 (33%)
|
2 (17%)
|
0.3006
|
Table 3
Classification of 92 patients with gram-negative bacteremia among community-acquired,
healthcare-associated, and hospital-acquired bacteremia
|
Organism
|
Community acquired
n = 20 (%)
|
Healthcare associated
n = 8 (%)
|
Hospital acquired
n = 64 (%)
|
Total
n = 92 (%)
|
|
Acinetobacter baumannii
|
1 (5)
|
2 (25)
|
2 (3.1)
|
5 (5.4)
|
|
Acinetobacter lwoffii
|
0 (0)
|
1 (12.5)
|
0 (0)
|
1 (1.1)
|
|
Brucella
|
2 (10)
|
0 (0)
|
0 (0)
|
2 (2.2)
|
|
Burkholderia cepacia
|
0 (0)
|
0 (0)
|
1 (1.5)
|
1 (1.1)
|
|
Escherichia coli
|
5 (25)
|
0 (0)
|
10 (15.6)
|
15 (16.3)
|
|
Enterobacter cloacae
|
1 (5)
|
2 (25)
|
4 (6.2)
|
7 (7.6)
|
|
Haemophilus influenzae
|
3 (15)
|
0 (0)
|
0 (0)
|
3 (3.3)
|
|
Klebsiella pneumoniae
|
0 (0)
|
0 (0)
|
27 (42.1)
|
27 (29.3)
|
|
Pseudomonas aeruginosa
|
0 (0)
|
2 (25)
|
12 (18.7)
|
14 (15.2)
|
|
Pseudomonas flourescens
|
0 (0)
|
0 (0)
|
1 (1.5)
|
1 (1.1)
|
|
Salmonella species
|
6 (30)
|
2 (25)
|
0 (0)
|
8 (8.6)
|
|
Salmonella typhi
|
1 (10)
|
0 (0)
|
0 (0)
|
1 (1.1)
|
|
Serratia marcescens
|
0 (0)
|
0 (0)
|
5 (7.8)
|
5 (5.4)
|
|
Stenotrophomonas maltophilia
|
0 (0)
|
0 (0)
|
2 (3.1)
|
2 (2.2)
|
|
Total
|
20
|
8
|
64
|
92 (100%)
|
Table 4
The antimicrobial susceptibility of the isolated organisms among community-acquired/health-associated,
and hospital-acquired bacteremia (n = 92)
|
Antibiotics
|
Resistance
n = 92 (%)
|
Community acquired and health associated
n = 28 (%)
|
Hospital acquired
n = 64 (%)
|
p-Value
|
|
Pipracillin/tazobactam TAZ
|
5 (5.5)
|
0 (0)
|
5 (7.8)
|
0
|
|
Bactrim/cotrimoxazole SXT
|
18 (19.8)
|
5 (18.5)
|
13 (20.3)
|
0.5657
|
|
Meropenem MEP
|
5 (5.5)
|
2 (7.41)
|
3 (4.7)
|
0.0674
|
|
Imipenem IMP
|
6 (6.6)
|
5 (18.5)
|
1 (1.6)
|
0.2793
|
|
Gentamicin GM
|
14 (15.4)
|
10 (37)
|
4 (6.3)
|
0.6201
|
|
Amikacin AK
|
4 (4.4)
|
2 (7.4)
|
2 (3.1)
|
0.0405
|
|
Augmentin AMC
|
16 (17.6)
|
4 (14.8)
|
12 (18.8)
|
0.5663
|
|
Ampicillin/Ampiclox AMP
|
48 (52.7)
|
6 (22.2)
|
42 (65.6)
|
0.1128
|
|
Cephradine CRD
|
14 (15.4)
|
4 (14.8)
|
10 (15.6)
|
0.4338
|
|
Ciprofloxacin CIP
|
11 (12.1)
|
4 (14.8)
|
7 (10.9)
|
0.4447
|
|
Aztreonam ATM
|
4 (4.4)
|
2 (7.41)
|
2 (3.1)
|
0.0406
|
|
Cefoxitin FOX
|
13 (14.3)
|
3 (11.1)
|
10 (15.6)
|
0.4597
|
|
Cefuroxime/Zinacef CXM
|
12 (13.2)
|
0 (0)
|
12 (18.8)
|
0
|
|
Cephaloridine CPL
|
4 (4.4)
|
0 (0)
|
4 (6.3)
|
0
|
|
Chloramphenicol CHL
|
2 (2.2)
|
0 (0)
|
2 (3.1)
|
0
|
Table 5
Difference between ESBL and non-ESBL isolate among community-acquired, healthcare-associated,
and hospital-acquired bacteremia
|
Organism
|
Community acquired (n = 13) and healthcare associated (n = 7)
|
Hospital acquired (n = 46)
|
Total (n = 66)
|
p-Value
|
|
ESBL
|
5 (25.0%)
|
11 (23.9%)
|
16 (24.2%)
|
0.928
|
|
non-ESBL
|
15 (75.0%)
|
35 (76.1%)
|
50 (75.8%)
|
|
Total
|
20 (30.3%)
|
46 (69.7%)
|
66 (100%)
|
<0.001
|
Abbreviations: ESBL, extended spectrum β-lactamase; non-ESBL, nonextended spectrum
β-lactamase.
Note: There is no significant difference in the percentage of ESBL and non-ESBL in
community-acquired and hospital-acquired bacteremia (p = 0.928).However, there is a significant difference in the total frequency of isolates
between community-acquired and hospital-acquired bacteremia, regardless of ESBL or
non-ESBL (p < 0.001).
Table 6
Difference between the mortality rate in hospital-acquired and community-acquired
bacteremia
|
Hospital acquired (n = 64)
|
Community acquired (n = 20)
|
Total
|
p-Value
|
|
14 (82.4%)
|
3 (17.6%)
|
17
|
0.1228
|
Discussion
Recent data exist on the differentiation of HCA bacteremia from HA bacteremia and
CA bacteremia.[16] In this study, comparing the epidemiology of CA, HCA, and HA bloodstream infections,
it was found that health-associated bacteremia (7.8%) was of low incidence unlike
other studies in which there were higher incidences of HCA bacteremia. This is similar
to a study performed in Spain that found that 18% were CA, 24% were HCA, and 58% were
HA.[17] Hoenigl et al found that of the total 1,143 patients diagnosed with bacteremia,
HCA accounted for 63.7% and CA for 36.3% cases.[18] Similarly, Friedman's group reported that 50.9% of their 159 cases of gram-negative
bacteremia were HCA.[9] Unlike this study, in the study by Hoenigl et al, the study cohort had slightly
more male subjects (56.1 vs. 50.2%, p = 0.044) than females.[18] In this study, analyzing the causative pathogens, we found that the most common
gram-negative bacteria causing bacteremia was K. pneumoniae and was found to be significantly associated with the HA bacteremia. The causative
pathogens were comparable between patients with CA and those with HCA bacteremia.
In a recent prospective cohort study, in 672 patients enrolled with positive peripheral
blood cultures (192 CA, 85 HCA, and 395 HA), E. coli was found to be the most frequently isolated pathogens. Escherichia coli was isolated more frequently in patients with community-onset bacteremia, and Pseudomonas species were isolated more frequently among those with HA bacteremia.[18] In agreement with previous reports, the most frequently seen underlying diseases
in patients with gram-negative bacteremia were malignancies.[19] Compared with patients with HCA bacteremia and CA bacteremia for patients with HA
bacteremia, the clinical presentation of urinary tract infection was higher in HCA
bacteremia and CA bacteremia–4 (50%) and 10 (50%), respectively, versus 2 (1.3%) for
HA bacteremia. Marschall et al reported contradicting results with findings of 38
(28.8%) versus 13 (14.4%) cases of bacteremia for the urinary tract when compared
with HA bacteremia and HCA bacteremia.[20] This finding is in accordance with the previous study that reported that HCA bacteremia
was associated with increased 30- and 90-day mortality rates when compared with CA
bacteremia.[21]
In other studies, the mortality rate of HCA infection seemed to be generally higher
than that of CA infection and was similar to that of hospital infection.[10]
In this study, all the isolates of HCA were multisensitive, contrary to a study from
Korea in which a total of 240 patients were infected with community-onset K. pneumoniae bacteremia, 140 (58.3%) were defined as HCA infection cases, and the remaining 100
patients were classified as CA infections. Patients infected with HCA bacteremia showed
significantly different clinical and microbiological characteristics compared with
those infected with CA bacteremia. HCA K. pneumoniae bacteremia was characterized by more antibiotic-resistant pathogens (ciprofloxacin
resistance, 12.9 [18/140] vs. 4.0% [4/100], p = 0.02) and ESBL production (12.1 [17/140] vs. 4.0% [4/100], p = 0.03) than CA bacteremia.[22]
A recent study from Saudi Arabia reported their experience of HA catheter-related
blood stream infections (CRBSIs). CRBSI resulted in 60 admissions with a median of
182 days of hospital stay and 74 changes of central venous catheters. The rate of
CRBSI was 2.9 per 1,000 catheter days. Staphylococcus species were the most prevalent pathogens (32%), followed by K. pneumoniae (5%).[23]
Conclusion
This study concluded that the epidemiology, causative pathogens, and mortality differed
markedly between community-onset and HA bacteremia, while differences between CA and
HCA bacteremia were by far less pronounced. HA bacteremia was associated with an increase
in the risk of mortality when compared with CA bacteremia.
Our study had some limitations. One is the small sample size, a larger study should
be performed to validate the findings of this study. Future studies should focus on
recognizing CA and HCA bacteremia due to resistant organisms. Nonetheless, to the
best of our knowledge, this is the first study comparing the clinical and bacteriological
characteristics of HCA, HA, and CA gram-negative bacteremia in Saudi pediatric patients.
