Keywords
bloodstream infection - neonatal sepsis - blood culture - antibiotic resistance -
neonatal ICU
Introduction
The global neonatal mortality rate has declined approximately 50% in the past two
decades; nevertheless, in 2020, 2.4 million newborns lost their lives worldwide, amounting
to 6,500 newborn deaths per day.[1] According to the latest global burden of disease data in 2019, 227,000 neonates
died due to sepsis, accounting for 4.5% (4.1–4.9) of deaths under 5 years.[2]
Neonatal sepsis occurs due to an infection involving the bloodstream, and it remains
a leading cause of mortality and morbidity among neonates especially in middle- and
low-income countries.[3] Neonates with sepsis may present with nonspecific signs and symptoms when infected,
but identifying the pathogen via blood culture or other methods is the key in diagnosing
that it is truly sepsis.[3] Importantly, the sensitivity of blood cultures is significantly improved with appropriate
blood volumes.[4] However, antibiotic treatment decisions cannot await the results of blood cultures.
Instead, antimicrobial therapy is usually implemented empirically on clinical suspicion
of sepsis. To prevent the emergence of antimicrobial-resistant organisms, therapy
should be instituted according to the local pathogen profiles and their antimicrobial
susceptibility patterns. Unfortunately, data on this topic are limited in low- and
middle-income countries.[5]
[6]
Türkiye is a middle-income country in the World Health Organization (WHO) European
region and has the highest ratio of child population to the general population, when
compared with countries in the European Union. As with the rest of the world, in Türkiye,
the infant mortality rate has declined gradually over the last decade, from 13.9 in
2009 to 9.2 in 2021.[7] However, approximately 1 in 10 newborn deaths are due to sepsis and other infections,[8] most of which could be prevented with prompt diagnosis and appropriate and timely
treatment.
In this study, we aimed to investigate the burden of bloodstream infections (BSIs),
determine the spectrum of pathogens and their antimicrobial susceptibility patterns
in a neonatal intensive care unit (NICU), and report the preliminary results of a
new university hospital in Istanbul, Türkiye.
Materials and Methods
In this retrospective cross-sectional study, medical records of all neonates suspected
to have sepsis admitted to NICU between January 1, 2018, and December 31, 2020, were
analyzed. Patients' laboratory records were reviewed for demographic data, blood culture,
and antimicrobial susceptibility test results.
Two sets of blood samples (1 mL) were obtained from each neonate for culture. Each
set consisted of one aerobic and one anaerobic bottle; if adequate volume of blood
sample could not be obtained for two sets, the entire blood sample was inoculated
into the aerobic bottle. Blood samples were incubated for 5 days in the automated
hemoculture system BacT/Alert (bioMerieux); isolates were identified by VITEK 2 compact
(bioMerieux). Blood cultures with no growth after 5 days incubation was considered
as negative. When growth was detected, Gram staining was performed, and the broth
was subcultured onto blood, chocolate, and Sabouraud's dextrose agars. Isolates of
microorganisms that are common skin commensals (e.g., coagulase-negative staphylococci
[CoNS], Corynebacterium spp., Propionibacterium acnes), from a single blood culture bottle, were considered as contaminants. True bacteremia
with these organisms required isolation of the same species with similar antibiograms
from at least two blood culture sets.[9]
[10]
Data were collected from culture-proven BSI cases in NICU between January 2018 and
December 2020. Descriptive data analysis was performed, and the numbers and percentages
of isolates and their antimicrobial susceptibility tests were summarized.
As this study was retrospective and laboratory based, additional clinical information
regarding the probable source of BSIs, risk factors for BSIs, and comorbidities were
not available.
Primary outcome measures were the pathogen distribution and rates of methicillin,
vancomycin, gentamicin, carbapenem resistance and extended-spectrum beta-lactamase
(ESBL) production among isolated strains.
In this study, laboratory data were collected retrospectively and used in compliance
with procedures and principles laid down in Personal Data Protection Law (No: 6698,
date: March 24, 2016). Patient identifying information was removed from the data set
by the laboratory personnel in charge of data control before analysis, and then sent
to authors.
Results
During the study period, there were 117 BSI episodes in 106 neonates; 11 neonates
had 2 episodes of BSI caused by different species. Among the 106 neonates, 75 (70.8%)
were male and 31 (29.2%) were female. Mean age was 13.2 (±8.1) days. The most common
pathogen was Staphylococcus epidermidis (n = 86, 73.5%), followed by Klebsiella pneumoniae (n = 11, 9.4%) ([Table 1]). Among 86 S. epidermidis isolates, 76 (88.4%) were resistant to methicillin, whereas only 1 (14.3%) Staphylococcus aureus isolate was methicillin-resistant. Overall, methicillin resistance among staphylococci
was 82.8% (77/93) ([Fig. 1]). Vancomycin resistance was not detected. All K. pneumoniae and E. coli isolates were resistant to ampicillin and were ESBL producing. The overall rate of
ESBL production among Enterobacterales was 82.4% (14/17) ([Fig. 2]).
Table 1
Number of isolated pathogens
|
Group (N, %)
|
Microorganism
|
Number of isolates (%)
|
|
Gram-positive (96, 82.1%)
|
Staphylococcus epidermidis
|
86 (73.5%)
|
|
Staphylococcus aureus
|
7 (5.9%)
|
|
Streptococcus agalactiae
|
2 (1.7%)
|
|
Streptococcus pyogenes
|
1 (0.9%)
|
|
Gram-negative (21, 17.9%)
|
Klebsiella pneumoniae
|
11 (9.4%)
|
|
Enterobacter cloacae
|
3 (2.6%)
|
|
Acinetobacter baumannii
|
2 (1.7%)
|
|
Escherichia coli
|
2 (1.7%)
|
|
Pseudomonas aeruginosa
|
1 (0.9%)
|
|
Serratia marcescens
|
1 (0.9%)
|
|
Stenotrophomonas maltophilia
|
1 (0.9%)
|
Fig. 1 Antimicrobial resistance in gram-positive bacteria isolated from neonatal blood cultures.
Fig. 2 Antimicrobial resistance in gram-negative bacteria isolated from neonatal blood cultures.
Gentamicin resistance was common among methicillin-resistant staphylococci and ESBL-producing
gram-negative bacilli ([Fig. 3]). Among 77 methicillin-resistant staphylococci, 54 strains (70.1%), 1 of which was
methicillin-resistant S. aureus and 53 were methicillin-resistant S. epidermidis, were resistant to gentamicin. Gentamicin resistance was detected in
11 (78.6%) of the 14 ESBL-producing Enterobacterales (1 E. coli and 10 K. pneumoniae) ([Fig. 3]). Two K. pneumoniae (2/11, 18.2%) and one Acinetobacter baumannii (1/2, 50%) were resistant to meropenem ([Fig. 2]).
Fig. 3 Number of gentamicin-resistant isolates among methicillin-resistant staphylococci
and extended-spectrum beta-lactamase-producing Enterobacteriaceae.
Discussion
Our study underlines the importance of local surveillance of pathogen distribution
and antimicrobial resistance patterns. Rates of methicillin resistance (82.8%) among
staphylococci and of ESBL production (82.4%) among Enterobacterales in our setting
are alarming. Empirical treatment of neonatal sepsis recommended by WHO consists of
ampicillin/penicillin (or cloxacillin if staphylococci are suspected) with gentamicin.[11] In accordance with WHO's recommendations, empirical treatment in our setting consists
of ampicillin with gentamicin. Vancomycin or meropenem usage is limited with the neonates
not responding to the initial treatment. Therefore, we have considered gentamicin
resistance rates among methicillin-resistant staphylococci and ESBL-producing Enterobacterales.
Unfortunately, gentamicin resistance rates were high at 70.1% in methicillin-resistant
staphylococci and 78.6% in ESBL-producing Enterobacterales. Vancomycin and colistin
resistance were not detected.
There are several other studies in the literature investigating the species distribution
of pathogens isolated from neonates with suspected sepsis. In a recent meta-analysis,
investigating the global incidence of neonatal sepsis found that in approximately
one-third of culture-proven sepsis, the causative agents were S. aureus and Klebsiella spp.[12] Another meta-analysis focusing on developing countries determined Klebsiella spp. (26.36%), S. aureus (23.22%), CoNS (23.22%), and E. coli (15.30%) as the most common causes of neonatal sepsis.[13] A systematic review and meta-analysis investigating the etiologies and antimicrobial
resistances of invasive bacterial infections in neonates in sub-Saharan Africa reported
that S. aureus, Klebsiella spp., and E. coli accounted for 25, 21, and 10% of bacteremia or sepsis cases, respectively.[14] A recent literature review[15] of retrospective studies regarding culture-proven neonatal sepsis conducted in China
between 2016 and 2018 showed a predominance of gram-positive pathogens (59.2%) and
among gram-positive bacteria, CoNS (40.2%) ranked first, followed by Streptococcus spp. (6.81%), Enterococcus spp. (6.10%), and S. aureus (5.15%). Enterobacterales were the most common gram-negative pathogens; Klebsiella spp. were isolated in 14.5%, E. coli in 12.1%, and Enterobacter cloacae in 1.9%. Pseudomonas aeruginosa was isolated only in 1.4% of cultures. Similarly, in our study, the most common pathogens
were Staphylococcus and Klebsiella species; however, CoNS predominated to a far greater extent than that in other studies.
Antimicrobial susceptibility patterns vary between settings and countries. In our
study, methicillin resistance among S. epidermidis and ESBL production among K. pneumoniae are striking. In an Italian study, methicillin resistance was detected in 30% of
S. aureus and 94.7% of CoNS isolated from neonates with sepsis.[16] A study from Germany reported high rate of oxacillin resistance (88.2%) among S. epidermidis and ampicillin resistance (73.9%) among E. coli isolates.[17] Methicillin resistance among staphylococci was reported to be high in an Ethiopian
study, in which 69% of S. aureus and 100% of CoNS strains were resistant to methicillin; 56.6 and 91% of these strains
were also resistant to gentamicin. Antimicrobial resistance was also common among
gram-negative bacteria, with ampicillin and gentamicin resistance detected in 66.7
and 55.6% of E. coli strains and 91 and 82% of Klebsiella strains, respectively.[18] In the systematic review and meta-analysis of studies conducted in sub-Saharan Africa,
resistance to beta-lactam and aminoglycoside drugs recommended by WHO was found to
be 68 and 27%, respectively.[14] In the study conducted by the Burden of Antibiotic Resistance in Neonates from Developing
Societies (BARNARDS) group assessing the burden of antimicrobial resistance in neonatal
sepsis cases in low- and middle-income countries, 60% of gram-negative bacteria were
found to be resistant to the first-line empirical treatment for neonatal sepsis (both
ampicillin and gentamicin).[19] Not all studies reported high levels of resistance. In a study conducted in Türkiye,
cases of early-onset neonatal sepsis were investigated and 27.6% (8 out of 29) of
E. coli strains were found to be ESBL positive; all of these ESBL producers were susceptible
to gentamicin.[20] A study investigating the epidemiology of culture-proven neonatal sepsis from Saudi
Arabia reported that all isolated gram-negative and gram-positive bacteria were susceptible
to aminoglycoside antibiotics (amikacin and/or gentamicin) tested.[21]
This retrospective, laboratory data-based study had some limitations. Due to the lack
of baseline characteristics and comorbidities of the patients and data regarding the
treatment applied to patients and its outcome, we could not assess risk factors related
with sepsis due to drug-resistant bacteria or treatment failure. Nevertheless, our
study provided significant data regarding the etiology of neonatal sepsis in our setting
and their susceptibility patterns which may be used to employ better medical decisions.
In conclusion, our study represents the preliminary results of the NICU in a new university
hospital with high levels of resistance which emphasizes the importance of constant
surveillance of local pathogen distribution and their antimicrobial resistance profiles
to keep treatment protocols and guidelines up-to-date. Multicentered, prospective
studies investigating the etiology and antimicrobial susceptibility patterns in neonatal
sepsis are required to establish feasible management strategies.
