Nosocomial meningitis (NM) is a serious neurosurgical complication that severely affects the survival of patients and the success rate of neurosurgery. According to literature reports, the mortality of NM patients is up to 15%–45% that of non-NM patients (Pintado et al., 2019).

Increasing antibacterial resistance in bacteria is a growing public health crisis that makes many medical care-associated infections difficult to treat with current antibiotics (Bielli et al., 2020; Liu et al., 2020). Multidrug-resistant Enterobacteriaceae (MDRE) is defined as Enterobacteriaceae resistant to at least three types of antibiotics (Teklu et al., 2019), including β-lactams, aminoglycosides, macrolides, and quinolones, among others. Because of their resistance to multiple drugs, clinicians are always limited in their choice of antibiotics. NM caused by MDRE usually leads to serious consequences for neurosurgical patients (Thatrimontrichai et al., 2021). Due to the complexity of neurosurgery and the existence of the blood–brain barrier, the mortality rate in patients with multidrug-resistant (MDR) bacterial meningitis is always higher than that in patients infected with wild-type strains (Liang et al., 2019).

This study focuses on the epidemiology and clinical features of the largest cohort of patients with MDRE-NM reported to date, with the aims of evaluating the clinical outcome of NM caused by MDRE and examining the risk factors for mortality in MDRE-NM patients. To the best of our knowledge, this is the first survival analysis on MDRE-NM patients conducted worldwide.

In total, 45,771 neurosurgery patients and 3,570 NM patients were observed in the two neurosurgical centers mentioned above. Among them, 22.8% (815/3570) of the cases were MDR patients; coagulase-negative staphylococci occupied the highest ratio. A total of 86 patients died of MDR NM, and besides MDRE, MDR Acinetobacter baumannii comprised the highest proportion. The distribution of these patients is shown in Figure 2 . Of these patients, 242 cases of Enterobacteriaceae including 41.7% (101/242) of patients with MDRE-NM were recorded. Of them, two patients younger than 18 years, three patients hospitalized less than 7 days or without antimicrobial treatment, two patients who died within 7 days, and four patients with incomplete medical records (11 in total) were excluded from this study. Of the remaining 90 cases, 79 patients were from Beijing Tiantan Hospital and Capital Medical University (55 survival and 24 non-survival) and 11 were from Daqing Oilfield General Hospital (7 survival and 4 non-survival).

The emergence and spread of MDR pathogens, especially Enterobacteriaceae, is an important nosocomial health concern, making the improved molecular characterization and evaluation of the mortality risk factors for these strains vital. In this cohort study, the molecular characteristics and mortality risk factors for MDRE-NM were evaluated and compared with those of other types of MDR bacteria. The results revealed a high rate (101/242, 41.7%) of MDRE incidence among neurosurgical patients with NM, among which K. pneumoniae (46/90, 51.1%) was the most prevalent Enterobacteriaceae causing MDRE-NM. As evaluated by AST, third-generation cephalosporins (ceftriaxone) and carbapenems (meropenem) had high resistance rates of 92.2% and 44.4%, respectively. MDRE illustrated specific resistance genotypes for various antibiotics, among which bla_KPC (27/40, 67.5%) and bla_CTX-M-1 (60/83, 72.3%) had the highest ratios. Multivariate Cox survival analysis showed that EVD and a GCS score ≤8 were independent mortality risk factors for patients with MDRE-NM.

Due to the limited choice of antibiotics, infections caused by MDRE can lead to more serious outcomes. In recent years, the increasing number of MDRE infections has brought great difficulties in clinical treatment. In our study, the proportion of K. pneumoniae was as high as 51.1%, which exceeded the proportion of Enterobacteriaceae reported in the literature (Balkhair et al., 2019), indicating that MDR K. pneumoniae caused more neurosurgical meningitis. K. pneumoniae possesses not only high drug resistance but also high virulence (Xu et al., 2019) and high invasiveness (Juan et al., 2019), and its infection mortality rate is also higher than those of other types of Enterobacteriaceae (Park et al., 2019). E. coli and K. aerogenes also accounted for certain proportions of the infections. These three types of Enterobacteriaceae account for 86.67% of all MDRE and are the most important pathogenic bacteria.

MDR is defined as bacterial resistance to at least three types of antibiotics. In this study, eight kinds of antibiotics, namely, meropenem, ceftriaxone, amikacin, levofloxacin, SMZ-TMP, tetracycline, chloramphenicol, and polymyxin B, were classified to screen MDRE pathogens. Target MDRE pathogens showed high resistance rates to meropenem (44.4%) and ceftriaxone (92.2%); moreover, resistance to tetracycline, levofloxacin, and chloramphenicol was relatively high. Nevertheless, the rates of resistance to amikacin, SMZ-TMP, and polymyxin B were low. All of the MDRE pathogens assessed were especially sensitive to polymyxin B, which is the last line of defense against MDR Gram-negative bacteria. A study reported that, for patients with severe multiresistant bacterial meningitis, polymyxin B can be injected intrathecally for emergency, which may be a better alternative for MDRE-NM treatment (Abad-Restrepo et al., 2018). However, treatment with polymyxin B causes several adverse reactions, such as nephrotoxicity (Zavascki and Nation, 2017), neuromuscular blockage (Myint et al., 2016), and respiratory depression (Hussein et al., 2010), so its application is limited to NM patients.

The presence of antibacterial resistance genes closely influences the choice of therapeutic drugs. For example, ceftazidime/avibactam has better activity against A- and D-type carbapenemase-producing bacteria, but is ineffective against the B type (such as bla_VIM and bla_IMP ) (Zhanel et al., 2013). The distribution of carbapenemase-producing Enterobacteriaceae genes always varies; for example, the most prominent genotype in Europe is D-type bla_OXA-48 (Van Duin and Doi, 2017), the most prominent strain in India is bla_NDM-1 (Rahman et al., 2018), and the most prominent strain in China is bla_KPC (Chi et al., 2019; Xu et al., 2019). In this study, the majority of the MDRE pathogens possessed more than one type of carbapenem resistance gene; nevertheless, bla_KPC was the most frequent carbapenemase-producing mechanism. Previous studies have reported that undergoing loss of membrane porin protein and demonstrating modifications in permeability caused by efflux pump systems can significantly affect the resistance of bacteria to carbapenem (Petrosillo et al., 2013; Wassef et al., 2015). Our results confirmed that Enterobacteriaceae expressing the ESBL-producing gene+Ompk35 comprised a large proportion; however, since the detection of Ompk35 expression is not included in routine clinical laboratory test, ESBL + Ompk35, as a type of carbapenem resistance mechanism, also deserves the attention of neurosurgeons. Studies have confirmed that the predominant genotype of the MDRE varies in different regions and almost determines the phenotype of the MDRE. For example, an African study showed that bla_SHV was the most frequent genotype, followed by bla_TEM , bla_CTX-M , and bla_SHV (Nwafia et al., 2019). In a longitudinal analysis, bla_CTX-M-15 was the dominant ESBL-producing gene in all European countries, except Greece, where bla_SHV was more common (Kazmierczak et al., 2020). Another study from the Netherlands, however, reported that bla_CTX-M-1 was the dominant gene, followed by bla_CTX-M-15 (Van Hoek et al., 2018). Screening for third-generation cephalosporin-related resistance genes in this study revealed that the three genotypes with the highest frequency of expression (bla_TEM , bla_CTX-M-9 , and bla_SHV ) are all ESBL-producing resistance genes. Most likely because we investigated relatively few genes related to the production of the AmpC enzyme, fewer AmpC-related genes were detected. In fact, K. pneumoniae was the most common MDRE pathogen, and bla_KPC was the predominant carbapenem resistance gene detected. In addition, univariate analysis showed that clinical factors such as AMV and ICU admission were significantly different in the two groups, which suggests that the prevalence of MDRE pathogens may be related to the clinical environment. Therefore, to evaluate the mechanism of resistance transformation, clinical molecular epidemiological approaches, such as multilocus sequence typing or pulsed-field gel electrophoresis, should be carried out (Xiong et al., 2021; Kim et al., 2022). For amikacin, tetracycline, and levofloxacin administration, we found that, although the MDRE pathogens showed resistance, we did not screen out the corresponding resistance genes. The reasons may be that our method for selecting resistance genes was insufficient and that testing for resistance genes was limited.

Survival analysis of MDRE-NM patients will help in their early treatment (Tewabe et al., 2018). Moreover, this analysis is a significant procedure to prevent the transfer of resistant bacteria among neurosurgical patients in the same ward. Control of the spread of resistant bacteria in the hospital will definitely help in maintaining patient health and safety. The prognosis of NM patients usually has poor outcomes, and prior studies reported that the mortality of NM patients with neurosurgery ranged between 20% and 78% (Karvouniaris et al., 2018). Several previous studies have identified mortality and infectious risk factors for meningitis, such as CSF leakage (Chen et al., 2020), coma (Olson et al., 2015), surgical intervention (Hanko et al., 2020), impaired consciousness (Mccormick et al., 2013), long surgery duration, AMV, delayed catheter removal even when clinically indicated, and non-sterile technique used during the surgery (O'Grady et al., 2011). In our study, the overall 30-day MDRE-NM mortality was 31.1%. EVD and a GCS score ≤;8 were associated with poor prognosis in patients with MDRE-NM. EVD is a common complication after brain surgery; nevertheless, the incidence of meningitis caused by EVD is high and increases the morbidity and mortality. Surveys such as those conducted by Lu et al. (2019) have shown that, in high-risk patients, the proportion of EVD-related infections may reach 22% (Arabi et al., 2005). EVD prolongs the length of hospital stay (LOS) of NM patients and increases their morbidity and mortality, increasing the hospitalization costs and even causing reoperations (Lyke et al., 2001; Eymann et al., 2008). The GCS provides an objective and reliable way of recording the conscious state of a patient. Generally, coma can be classified using GCS as severe, with GCS ≤;8; moderate, with GCS 9–12; and minor, with GCS >13 (Fu et al., 2020). Having a low GCS was a mortality risk factor and may have been predictive of the poor prognosis of MDRE-NM patients in our study. It is logical that low GCS scores indicate the severity of NM and coma, which easily leads to poor outcomes. Therefore, patients with EVD and with GCS ≤;8 require more extensive attention, such as the replacement of drainage catheters and the application of high-grade antibiotics, to obtain better treatment outcomes.

In the treatment of infectious disease, the application of combined antibiotics is of great significance; nevertheless, whether the combined application of antibiotics has a positive effect on the patient has not yet been determined. Tofas et al. reported that a combination of active drugs in carbapenemase-producing K. pneumoniae infection was associated with a lower mortality rate (Tofas et al., 2016). The results of this study showed that antibiotic combinations were more common in both AET and ADT. Regarding the outcomes of the patients, there was no significant difference in the effects of monotherapy and dual- or triple-combination drug therapy; this finding differs from those of previous reports. However, this result may be caused by the low number of included target patients. In addition, previous studies showed that the administration of polymyxin B is of great significance in the treatment of meningitis caused by MDR (Zhong et al., 2020), but in our hospital, it had fewer clinical applications.

There are several limitations to our study. Firstly, this was a retrospective study that evaluated patient mortality risk factors. This implies that the results of the study may not be representative of other studies. Secondly, the relatively small number of patients limited our epidemiological analysis. Thirdly, the number of drug resistance genes identified was relatively small, and many types of genes were not included. Further multicenter studies for mortality risk factor screening and more extensive evaluation of the resistance genes of the MDRE-NM causative pathogens may be required.

In summary, to the best of our knowledge, this is the first report to conduct a two-center molecular and cohort study evaluating the mortality risk factors for MDRE-NM patients. Our study revealed that MDRE has multiple drug resistance genotypes, and EVD and a GCS score ≤;8 are independent mortality risk factors for patients with MDRE-NM, which deserve the attention of microbiologists and neurosurgical clinicians.

The original contributions presented in the study are included in the article/supplementary material. Further inquiries can be directed to the corresponding authors.

This study was approved by the Ethical Committee of Beijing Tiantan Hospital and Capital Medical University (approval no. KY-2021-079-02).

GHZ, YS, YC: design and drafting of the article; data collection; and data analysis and interpretation. LQ: statistics. HL: MDR genotypes test. LZ and GJZ: conduct the whole study.

This work was supported by the Beijing Hospital Authority Clinical Medicine Development of Special Funding (grant no. ZYLX202108) and Beijing Municipal Administration of Hospitals Incubating Program (grant no. PX2022021).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.