Neonate Bloodstream Infections (BSI) in China: Emerging Threat of Multidrug-resistant (MDR) pathogens


 Background: Bloodstream Infections (BSIs) continue to be associated with significant morbidity and mortality worldwide, which multidrug-resistant (MDR) pathogens present an ever-growing burden in the hospital and community settings. Data on antimicrobial resistance and prevalence of BSIs pathogens among hospitalized neonates in China are few and not clearly defined. Methods: we retrospectively collected clinical and bacteria data about patients with BSI from 2013 to 2017 in Chongqing Health Center for Women and Children. The patients were divided into the early- and late-onset BSI groups according to if BSI occurred within or beyond 72 hours after admission.Results: 210 BSIs occurred from January 1st, 2014 to December 31th, 2019. Of all included 210 patients, there were 99(47.14%), 105 (50.00%), and 6 (2.86%) infected with gram-positive bacteria, gram-negative bacteria, and fungi, respectively. For early-onset BSI, the predominated pathogen was E. coli, accounting for 59.6% with 29.72% producing of extended-spectrum beta-lactamases (ESBLs). CNS (40.3%), K. pneumoniae (23.8%), and E. coli (20.9%) constitute the main causes of late-onset BSIs. Notably, 75.0% of K. pneumoniae produced ESBLs and 81.3% were MDR, which were higher than early-onset BSIs. Late onset BSIs and antibiotic exposure were significantly associated with MDR infection. Conclusion: Gram-negative bacteria gradually became the main pathogenic bacteria, among which E. coli and K. pneumoniae accounted for the largest proportion. The phenomenon of multi-drug resistance of bacteria is serious, and the first-line drug can't meet the practical needs. Late onset sepsis and antibiotic exposure were significantly associated with MDR infection.


Background
Premature birth, severe infection and suffocation are the leading causes of neonatal death and the most frequently observed infection in neonatal wards is bloodstream infections (BSIs) 1 . In 2017, an estimated 48.9 million incident cases of BSIs were recorded worldwide and 11.0 million BSIs-related deaths were reported, representing 19.7% of all global deaths 2 . It is important that progress in neonatal survival has not kept pace with global reductions in child mortality and neonatal deaths now account for 47% of deaths in children under the age of ve years 3 .
Appropriate empirical treatment of BSIs is crucially important in reducing mortality. To make matters worse, the emergence of antimicrobial resistance (AMR) pathogens makes it more di cult to treat bacteremia.
AMR pathogens exhibit a high degree of resistance to rst line drugs recommended by the World Health Organization-namely, ampicillin, gentamicin, and third generation cephalosporins such as cefotaxime, and even multidrug-resistant (MDR) 4 . A 2020 study of neonatal BSIs in Pune, India reported a high prevalence of AMR with high mortality 3 . Effective interventions are urgently needed to reduce burden of BSIs and death due to AMR pathogens in hospital neonates 5 . However, there are scant data on the epidemiology of neonatal sepsis and associated prevalence of AMR in LMIC, especially in China. It is critical to better understand the regional epidemiology of neonatal sepsis and AMR among hospitalized neonates in order to guide future prevention efforts and policies

Study Design and Patients Collection
We conducted a retrospectively cohort study from January 2014 to December 2019 in Chongqing Health Center for Women and Children. Blood cultures were performed on suspected BSIs patients with at least 2 blood samples from different sites at the same time and the data of neonates with positive blood culture results were collected.

Bacterial Strains
Non-repetitive nosocomial bacterial strains were collected between 2014 and 2019 in the Chongqing Health Center for Women and Children. All the isolates were identi ed at the species level by the VITEK MS (bioMerieux, Hazelwood, MO, United States) system, and routine antimicrobial susceptibility testing was performed by the VITEK2 compact (bioMérieux, Inc., Durham, NC) system.

De nitions
Cases of BSI were de ned as a blood culture positive for a known neonatal pathogen. Patients who were positive for well-described contaminants and without any infection symptoms were not considered as BSI. Further, patients were subclassi ed into 2 groups according to the period: (1) Early onset BSIs was de ned as BSIs occurring day of life (DOL) 0 through DOL 3 day. (2) Late onset BSIs was de ned as BSIs DOL after 3d. AMR and MDR were acceding to the criteria of the Clinical and Laboratory Standards Institute 2019 (CLSI 2019). For gram-positive bacteria, we de ned methicillin-resistant Staphylococcus aureus (MRSA), vancomycin-resistant enterococci (VRE) as MDR pathogens. For gram-negative bacteria, MDR were de ned as bacteria were resistant to at least three antimicrobial classes. Extended-spectrum-βlactamase (ESBLs) producing organisms were tested by the phenotypic tests of ESBL production.

Statistical Analysis
All analyses were performed using SPSS v.25.0 software (SPSS Inc., Chicago, IL, USA). Univariate analyses were performed separately for each of the variables. The odds ratio (OR) and 95% con dence interval (CI) were calculated to evaluate the strength of any association. For all calculations, statistical signi cance was de ned at P < 0.05 for 2-tailed tests.

Results
210 BSIs occurred from January 1st, 2014 to December 31th, 2019. Clinical data were collected from 179 neonates. As shown in Table 1, 59.8% of the patients were male, the mean birthweight of enrolled neonates was 2517.8 g (SD = 934.8), and mean gestation was 35.68 weeks (SD = 4.2), Nearly half of the patients were premature infants and low weight infants, accounting for 46.36% and 48.04% respectively. More than half of enrolled neonates (51.4%) were born by caesarean section. In addition, 14.5% of neonates were positive for embryonic membrane culture. Although 13.4% mothers received antibiotics within 7 days before delivery, less than half (5.6%) received antibiotics longer than 48 hours. Among all in ammatory indicators, the sensitivity of WBC was the worst, with only 13.41% newborns greater than 20*10^9/L. While the sensitivity of PCT was much higher, about three-quarters of patients (77.7%) had increased after the occurrence of BSIs. and 6 (2.9%) infected with gram-positive bacteria, gram-negative bacteria, and fungi, respectively. Interestingly, before 2018, gram-positive bacteria (50.9%) were the main pathogen, but the isolation rate decreased year by year. After 2018, gram-negative bacteria became dominated pathogen, even, nearly three-quarters of the pathogen were gram-negative bacteria (72.7%) in 2019. Fungus (2.9%) is the rarest pathogen of neonatal BSIs in our hospital, and there has been no fungus bacteremia for three consecutive years (Fig. 1). The distribution of bacteria was different in newborns with different weights (Fig. 2). Escherichia coli (E. coli) is the main cause of extremely low birth weight and macrosomia infants, accounting for 60.0% and 50.0% respectively. Klebsiella pneumoniae (K. pneumoniae) accounted for 33.3% of pathogenic isolates, followed by E. coli at 29.2% in very low birth weight infants. For low birth weight infants BSI from 2014 to2019, the most common causes were E. coli (38.5%) and K. pneumoniae (28.8%). However, in normal weight infants, pathogens were predominantly other staphylococcus (45.8%) such as coagulase-negative Staphylococcus (CNS) and E. coli (28.1%).

Resistance Patterns for Early-and Late-Onset Pathogen
Pathogen showed different resistant to ampicillin, gentamicin, and third generation cephalosporins such as cefotaxime which were rst line drugs recommended by the World Health Organization (Fig. 3). Gramnegative bacterial (90.6%) and Staphylococcus spp. (89.28%) had the highest resistance rates to ampicillin, however, Streptococcus spp (13.3%). remained sensitive to ampicillin. About a third of grampositive bacteria and a quarter of gram-negative bacteria are resistant to aminoglycosides each year. In addition, 50% of bacteria are resistant to cephalosporins.

Risk Factors For Mdr Infection
We assessed the risk factors of BSIs with MDR and non-MDR. As shown in the Table 3, Late onset sepsis and antibiotic exposure were signi cantly associated with MDR infection (P < 0.05).

Discussion
In the present study, we found that the microbiota and distribution of pathogens causing BSIs have changed signi cantly in the last 2 years. Gram-negative bacteria emerged as the predominant causative organisms. In addition, about half of BSI patients were infected with MDROs.
This study deserves attention for some major reasons. First, we found about two-thirds of sepsis occurred after 72 h, and the average age was about 9 days. This pattern is in sharp contrast with that reported by Elsevier, who said most episodes of sepsis occurred at a quite early age, with nearly a quarter of culture-positive sepsis episodes occurring within 24 h of birth and two-thirds within 72h 6 . However, China 7,8 and Korea 9 have shown a predominance of late-onset sepsis. At the same time, we also found differences in the distribution of pathogens in different countries. A high prevalence of Gram-positive organisms was found before 2018 in our hospital, which contrasts with previous ndings elsewhere where Gram-negative bacteria were the predominant organisms isolated 4,10−12 . Actually, our study found CNS played a signi cant role in BSIs in neonates with normal weight infants. However, CNS in neonatal bloodstream infections are not clear, as isolation from a single blood culture may represent true bacteremia or contamination 13,14 . Interestingly, Gram-negative bacteria especially Enterobacter have become the major pathogens in our hospital in the last two years 7,9,15−17 . In our study, expect normal weight newborns, the most prevalent pathogens were E. coli and K. pneumoniae. Elsevier had found that Acinetobacter spp. (22%) emerged as the most common isolates in their cohort, which was rare in our hospital.
Second, we found the pathogens associated with BSI and their individual resistance patterns may differ between early-and late-onset BSIs. Remarkably, only about one-third of E. coli produced ESBLs among early-onset BSIs, which was far less than K. pneumoniae did among late-onset BSIs. This result was quite contrast with a study, of 368 398 pediatric clinical isolates characterized in the United States from 1999 to 2011, that only 1.97% were identi ed as third-generation cephalosporin-resistant (G3CR), and 0.47% as ESBL producers and the majority of G3CR and ESBL isolates were E. coli 18  Third, we found late-onset sepsis and antibiotic exposure were signi cantly associated with MDR infection. Doctors often use ampicillin or cephalosporin empirically to treat bacteremia, but these antibiotics could induce high expression of Ampc and ESBLs, leading to the resistance to β-lactam antibiotics 21 . Piperacillin-tazobactam (TZP), a weak inducer of AmpC beta-lactamases and inhibitor of ESBLs may be a valuable treatment option for BSIs 22 .
This study had several limitations. First, the present study was performed at a single center. Second, the sample of this study was relatively small, which could be a source of bias. Third, we may have overestimated the incidence of CNS because its diagnosis was not based on two simultaneous blood cultures.

Conclusion
Our study showed that BSI was a serious clinical challenge in the China. Gram-negative bacteria gradually became the main pathogenic bacteria, among which E. coli and K. pneumoniae accounted for the largest proportion. The phenomenon of multi-drug resistance of bacteria is serious, and the rst-line drug can't meet the practical needs. Late onset sepsis and antibiotic exposure were signi cantly associated with MDR infection. Other antibiotics such as TZP should be the rst choice in empirical antimicrobial therapy.

Declarations
Ethics approval and consent to participate For this study, samples were collected at the microbiology laboratory of our hospital, with no contact with the patient. This study was retrospective and there was no patient identi cation performed during data collection. Therefore, the ethics committee determined that informed consent was not required.

Consent for publication
Not applicable Availability of data and materials All data generated or analysed during this study are included in this published article [and its supplementary information les].
HZ and QH designed the study, HZ, QH, CL and YS performed the experiments.
LG and YS analyzed data. HZ, QH and CL wrote this manuscript. All authors read and approved the nal manuscript.