Introduction

Clinical infection with Listeria monocytogenes is well recognized to occur in particular groups including the immunocompromised, pregnant women, elderly, and neonates [1]. The incidence of listeriosis is increasing in the USA and Europe [1]. Around 20–25 % of listeriosis cases involve the central nervous system [2]. L. monocytogenes is the second or third most frequent cause of bacterial meningitis in American and European adults [3, 4]. Among Korean adults, L. monocytogenes is the fourth most common etiology of community-acquired bacterial meningitis [5]. Meningitis due to L. monocytogenes has been mostly reported in immunocompromised hosts and elderly patients [6].

There are several predisposing conditions to increase the susceptibility for bacterial meningitis [7]. In particular, the common risk factors of pneumococcal and Listeria meningitis in adults include age older than 65 years, alcoholism, diabetes mellitus, cancer, and organ transplantation [7]. In this study, we determined the risk factors for predicting L. monocytogenes as the etiology of bacterial meningitis in adults compared to that due to S. pneumoniae.

Patients and methods

Study population

We conducted a multicenter, retrospective case–control study at 16 teaching hospitals in Korea. All patients were 18 years or older and were discharged with a diagnosis of bacterial meningitis between 1998 and 2013. The diagnoses were identified using ICD10 codes (G00–G03, A32.1). Bacterial meningitis was defined with clinical manifestations and cerebrospinal fluid (CSF) abnormalities compatible with meningitis and with a confirmed bacterial pathogen by cultures of CSF or blood or bacterial antigen testing. Cases were defined as patients with meningitis caused by L. monocytogenes. Listeria meningitis was diagnosed when L. monocytogenes was identified in the culture of the CSF and/or blood. Controls were cases of pneumococcal meningitis, diagnosed either by culture of the CSF and/or blood or by a bacterial antigen test. Some data of the meningitis cases between 1998 and 2008 were collected from a prior study’s database of bacterial meningitis cases in adults [5]. This study was approved by the Samsung Medical Center Institutional Review Board (2008-06-052-001).

Clinical and microbiological data

Demographic and clinical information including age, gender, underlying medical conditions, neurologic manifestations, CSF profile, brain imaging, and antibiotic treatment were collected from the medical records. Immunosuppressive agents included anti-cancer drugs, corticosteroids, purine pathway inhibitors (azathioprine, mycophenolate mofetil), leflunomide, calcineurin inhibitors (cyclosporine, tacrolimus), mammalian target of rapamycin (mTOR) inhibitors (sirolimus) and tumor necrosis factor-alpha blockers. The McCabe classification was used to categorize the severity of the underlying medical condition as non-fatal, ultimately fatal, or rapidly fatal [8]. Septic shock and severe sepsis were defined by the Society of Critical Care Medicine/American College of Chest Physicians criteria [9]. Based on data from the medical record and according to the definition from the Centers for Disease Control and Prevention (CDC), heavy alcohol drinking was defined as ingestion of five or more glasses of alcohol at one time (or four or more in women) [10].

The empirical antibiotic regimen was considered appropriate in cases if it included ampicillin/other penicillins or meropenem. In controls, concordant antibiotic therapy based on the in vitro susceptibility of the isolated S. pneumoniae strain was considered appropriate. The antimicrobial susceptibility of S. pneumoniae was determined according to the criteria of the Clinical and Laboratory Standards Institute [11].

Outcome variables

Early clinical outcome was evaluated at 3 days of treatment. All-cause mortality and infection-related mortality were assessed at 30 days of treatment. Neurological complications were evaluated at the time of discharge. The length of hospital stay was defined as the time between admission and discharge.

Statistical analysis

All statistical analyses were performed using Stata version 11.2 (StataCorp, College Station, TX, USA). All tests of significance were two tailed. P values <0.05 were considered to be statistically significant. Categorical variables were compared between the groups using Chi square test or Fisher’s exact test. Continuous variables were compared using the Mann–Whitney U test.

Multivariate logistic regression analysis was used to determine the risk factors of Listeria meningitis compared to those of pneumococcal meningitis. Independent variables were selected for backward stepwise regression and included age 50 years or older, chronic kidney disease, chronic liver disease, malignancy, and prior history of receiving immunosuppressive therapy. The covariants showing P ≥ 0.2 were removed from the final model. Time to death curves during the follow-up period were calculated using the Kaplan–Meier method and compared between groups using the log-rank test. Multivariate cox regression analysis was used to determine the risk factors of mortality.

Results

Demographic characteristics and underlying diseases

A total of 36 meningitis cases caused by L. monocytogenes were included in the cases, and the controls included 113 cases of pneumococcal meningitis (Table 1). Among cases of L. monocytogenes meningitis, 9 had positive cultures in both CSF and blood, 20 had positive CSF culture only, and 7 had positive blood culture only. The cases of pneumococcal meningitis comprised 31 with positive cultures in both CSF and blood, 71 with positive cultures in CSF only (39 were also antigen positive), 8 with positive cultures in blood only, and 3 with antigen-positive test only. There were more patients who were ≥50-year-old in the cases than in the controls. Hypertension, chronic kidney disease, chronic liver disease, autoimmune disease, and a prior history of receiving immunosuppressive therapy were also more frequent in the cases than in the controls.

Table 1 Comparison of demographic characteristics and underlying diseases between patients with Listeria meningitis (cases) and pneumococcal meningitis (controls)
Full size table

Predictive risk factors of Listeria meningitis

The most significant risk factor of Listeria meningitis was a prior history of receiving immunosuppressive therapy (odds ratio 8.12) (Table 2). Chronic liver disease was the second most important risk factor associated with the cases (OR 5.03). Chronic kidney disease also had a tendency to be associated with Listeria meningitis (OR 9.86, P = 0.063).

Table 2 Risk factors for Listeria meningitis compared with pneumococcal meningitis
Full size table

Clinical characteristics of Listeria meningitis

At the time of diagnosis, there were no significant differences between the two groups with regard to frequency of neurological abnormalities including seizure, cranial nerve palsy, aphasia, hemiparesis, and other neurological deficits (Table 3). Similarly, there were no differences between the two groups with regard to brain imaging results. The CSF profile of the cases was quite different from that of the controls (Table 3). The median white blood cell (WBC) count, neutrophil percentage, and protein level were much lower in the cases than in the controls. The median CSF/serum glucose ratio was 0.28 in the cases, but only 0.07 in the controls. The frequency with which the microscopic examination of Gram-stained concentrated CSF showed the presence of microorganisms was much lower in the cases than in the controls (12.5 vs. 79.1 %).

Table 3 Comparison of clinical characteristics of patients with Listeria meningitis (cases) and patients with pneumococcal meningitis (controls)
Full size table

Treatment and clinical outcome of Listeria meningitis

In the case, the most frequent initial empirical antibiotic regimen was a combination of vancomycin, third- or fourth-generation cephalosporin, and ampicillin. In contrast, the most frequent regimen in the controls was a combination of vancomycin and third- or fourth-generation cephalosporin (Table 4). The frequency of appropriate empirical antibiotic therapy was significantly lower in the cases than in the controls (63.9 vs. 96.1 %). The time to appropriate antibiotic therapy was also significantly longer in the cases than in the controls.

Table 4 Comparison of antibiotic regimen and clinical outcome of patients with Listeria meningitis (cases) and pneumococcal meningitis (controls)
Full size table

Early outcome assessment in the cases at 3 days showed no mortality; however, 54.8 % of the cases had worsened. In contrast, the fatality rate was 11.6 % in the controls. The overall 30-day mortality rate was higher in the cases than in the controls, although the difference was not significant (Table 4). There were no differences between the two groups with regard to neurological complications at discharge. The length of hospital stay was longer in the cases than in the controls.

Comparison of the Kaplan–Meier curves using a log-rank test demonstrated a significant difference in survival between the two groups (Fig. 1). There was a lower number of patient survival among the Listeria cases than among the controls. However, multivariate Cox regression analysis found that Listeria versus S. pneumoniae was not a significant risk factor of mortality. In contrast, delaying appropriate antibiotic therapy by more than 6 h had a 2.78-fold increased risk of mortality compared to cases in which the patient received appropriate therapy within 6 h (95 % CI 1.13–6.87, P = 0.027). Fatal underlying disease was also associated with increased fatality rate (hazard ratio 2.88, 95 % CI 1.16–7.18, P = 0.023).

Fig. 1
figure 1

Comparison of the Kaplan–Meier curves showing the cumulative probability of survival between Listeria and pneumococcal meningitis cases. The numbers below the plots indicate the number of patients at risk in each group at the given time points

Full size image

Among the 13 cases of Listeria meningitis in which patients received inappropriate initial empirical antibiotic therapy, the inappropriate therapy was continued in four cases even after L. monocytogenes was identified. Three of these four patients died. The median time to appropriate antibiotic therapy in the nine remaining cases was 72 h (range 24–288 h). Two of these nine patients died. Nine of 13 patients developed neurological complications.

Discussion

Adult meningitis patients with a prior history of receiving immunosuppressive therapy had an 8.1-fold increased risk of infection with L. monocytogenes versus with S. pneumoniae. Chronic liver disease also increased the risk of L. monocytogenes by fivefold.

Previous studies have found that most patients (>60 %) with Listeria meningitis are immunocompromised or elderly [6, 12]. There were many reported immunocompromised conditions, including immunosuppressive therapy, malignancy, connective tissue disease, diabetes mellitus, alcoholism, asplenia, liver cirrhosis, end-stage renal disease, and HIV infection [6]. However, those factors in addition to old age have long been recognized as important risks predisposing to pneumococcal meningitis [13]. To the best of our knowledge, there is only one prior report that compares the risk factors between Listeria and pneumococcal meningitis. A prospective observational study in Italian patients older than 50 years at a single center compared 22 cases of Listeria meningitis and 109 cases of pneumococcal meningitis. The group found an association between Listeria and current immunosuppressive therapy (OR 4.0, 95 % CI 1.2–6.9, P = 0.008) [14].

In the present study, other risk factors such as old age, diabetes mellitus, and malignancy were not different in association with the etiology between Listeria and pneumococcal meningitis. In univariate analysis, the proportion of persons aged 50 years or elder was higher in Listeria meningitis than in pneumococcal meningitis. However, this difference was not significant in multivariate analysis. Chronic kidney disease tended to be associated with Listeria meningitis. Therefore, it may be a risk factor predicting Listeria (versus S. pneumoniae) as the etiology of bacterial meningitis in adults.

In our study, the initial empirical antibiotic treatment was inappropriate in 36.1 %. Previous studies found that 30–50 % of patients with Listeria meningitis initially received inappropriate empiric treatment [6, 15]. A delay in treatment increased mortality and the risk of developing neurologic sequelae [15]. Interestingly, 55 % of Listeria cases in this study worsened clinically at 3 days of treatment, although there was no mortality. However, this finding is not unexpected, as 42 % of the patients did not receive appropriate antibiotic therapy for at least 24 h. The fatality rate in our study was similar to that of the previous reports (17–27 %) [13]. Vancomycin had been suggested as an alternative agent for treatment of Listeria meningitis in some recommendations due to in vitro susceptibility of vancomycin against L. monocytogenes [16]. However, the practice guidelines for the management of bacterial meningitis proposed by the Infectious Diseases Society of America (IDSA) in 2004 do not recommend the use of vancomycin for either empirical or definitive therapy of Listeria meningitis [17], and vancomycin has been associated with treatment failures in patients [13]. Therefore, vancomycin use was determined to be inappropriate for Listeria meningitis in our study. There were high rates (55.6 %) of neurological complications in this study. This may have occurred because we included the patients with neurological complications who eventually died. In contrast, Brouwer reported that neurological sequelae occurred in 25 % of patients who survived Listeria meningitis [13].

This study has several limitations. First, this was a retrospective study based on medical records. Therefore, our results may have been susceptible to information bias. For example, time to appropriate antibiotic therapy could not be calculated from all the enrolled patients. Secondly, the sample size of cases was small. Therefore, we were unable to analyze independent risk factors of mortality or neurological complications in patients with Listeria meningitis. Thirdly, the controls in this study were patients with pneumococcal meningitis. Therefore, we were not able to address the common risk factors shared by Listeria and pneumococcal meningitis. Fourthly, our study did not analyze any genotypic information of L. monocytogenes isolates. A recent report showed that the clinical outcomes differed according to the genotype of the L. monocytogenes strains [18]. Therefore, genotypic difference between the L. monocytogenes strains could have affected our analysis. Lastly, our data on the underlying diseases did not include sinusitis or mastoiditis, which are known to be important foci of infection in pneumococcal meningitis [13]. As our study was a retrospective design, consultation of the cases with the otorhinolaryngologists for examination was not done. This might have affected the risk factor analysis.

In conclusion, adult patients with a prior history of receiving immunosuppressive therapy within 1 month and those with chronic liver disease have 8.1-fold and fivefold increased risk of meningitis caused by L. monocytogenes compared to S. pneumoniae, respectively. Empirical antibiotic coverage targeting L. monocytogenes should be considered in the treatment of meningitis in patients with such risk factors.