Secondary Logo

Share this article on:

Methicillin-Resistant Staphylococcus aureus Meningitis in Adults: A Multicenter Study of 86 Cases

Pintado, Vicente MD; Pazos, Rosario MD; Jiménez-Mejías, Manuel E. MD; Rodríguez-Guardado, Azucena MD; Gil, Antonio MD; García-Lechuz, Juan M. MD; Cabellos, Carmen MD; Chaves, Fernando MD; Domingo, Pere MD; Ramos, Antonio MD; Pérez-Cecilia, Elisa MD; Domingo, Diego MD

doi: 10.1097/MD.0b013e318243442b
Original Study

Methicillin-resistant Staphylococcus aureus (MRSA) meningitis is an uncommon disease, and little is known about its epidemiology, clinical features, therapy, and outcome. We performed a multicenter retrospective study of MRSA meningitis in adults. Eighty-six adult patients were included and the following data were obtained: underlying diseases, clinical presentation, analytical and microbiologic data, response to therapy, and outcome.

There were 56 men (65%) and the mean age was 51.5 years; 54 of them (63%) had severe comorbidities. There were 78 cases of postoperative meningitis and 8 of spontaneous meningitis. The infection was nosocomial in 93% (80/86) of the cases. Among the 78 patients with postoperative meningitis, the most common predisposing conditions were cerebrospinal fluid (CSF) devices (74%), neurosurgery (45%), CSF leakage (17%), and head trauma (12%). Most patients had fever (89%), altered mental status (68%), headache (40%), and meningeal signs (29%). The most common CSF findings were pleocytosis (90%), elevated protein level (77%), and hypoglycorrhachia (30%). CSF Gram stain and blood cultures were positive in 49% (32/65) and 36% (16/45) of cases, respectively. An associated MRSA infection and polymicrobial meningitis appeared in 33% (28/86) and 23% (20/86) of cases, respectively. Antimicrobial therapy was given to 84 patients. Most of them received vancomycin (92%) either as monotherapy (64%) or in combination with other antibiotics (28%), for a median of 18 days. Overall 30-day mortality was 31% (27/86). Multivariate study identified 2 independent factors associated with mortality: spontaneous meningitis (odds ratio [OR], 21.4; 95% confidence interval [CI], 2.3–195.4; p = 0.007), and coma (OR, 9.7; 95% CI, 2.2–42.3; p = 0.002).

In conclusion, MRSA is a relatively uncommon but serious disease. Although most cases are nosocomial infections appearing in neurosurgical patients, spontaneous meningitis may present as a community-onset infection in patients with severe comorbidities requiring frequent contact with the health care system. Most patients have a favorable response to vancomycin, but the beneficial effect of combined and intraventricular therapy, or alternative drugs, remains unclear. MRSA meningitis is associated with a high mortality, and the presence of spontaneous infection and coma are the most important prognostic factors.

From Infectious Diseases Service (VP, RP), Hospital Ramón y Cajal, Madrid; Infectious Diseases Service (MEJ-M), Hospital Virgen del Rocío, Sevilla; Infectious Diseases Unit (AR-G), Hospital Central de Asturias, Oviedo; Department of Internal Medicine (AG), Hospital La Paz, Madrid; Department of Clinical Microbiology (JMG-L), Hospital Gregorio Marañón, Madrid; Infectious Diseases Service, Hospital Bellvitge (CC), L’Hospitalet de Llobregat, Barcelona; Microbiology Department (FC), Hospital Doce de Octubre, Madrid; Infectious Diseases Unit (PD), Hospital de la Santa Creu i Sant Pau, Barcelona; Infectious Diseases Unit (AR), Hospital Puerta de Hierro-Majadahonda, Madrid; Clinical Microbiology Service (EP-C), Hospital Clínico San Carlos, Madrid; Microbiology Service (DD), Hospital La Princesa, Madrid, Spain.

The authors have no funding or conflicts of interest to disclose.

Reprints: Vicente Pintado García, Infectious Diseases Service, Hospital Ramón y Cajal, Carretera de Colmenar km 9.1, 28034 Madrid, Spain (e-mail:

Back to Top | Article Outline


Staphylococcus aureus remains a common pathogen and a major cause of morbidity and mortality. Resistance to antibiotics such as oxacillin is increasing, and, in recent years, glycopeptide-resistant strains have been reported worldwide. S. aureus meningitis is an uncommon disease, accounting for only 1%–9% of cases of bacterial meningitis.6,11,12,28 Two different modes of pathogenesis had been described in staphylococcal meningitis: postoperative meningitis, associated with neurosurgical procedures, shunt devices, or head trauma, and “hematogenous” or “spontaneous” meningitis, secondary to staphylococcal infection outside the central nervous system.1,4,7,8,10,13,15,21,22,25

Although few studies about staphylococcal meningitis have been reported in the literature, the incidence of methicillin-resistant S. aureus (MRSA) meningitis seems to be increasing in recent years.2,5,17,26 Most cases of MRSA meningitis are hospital-acquired infections that have appeared in neurosurgical patients with cerebrospinal fluid (CSF) devices or recent surgery, and are associated with a high mortality rate (10%–45%). Although the clinical experience has been limited, vancomycin therapy seems to be useful in the management of MRSA meningitis, with a reported response rate of 54%–61%.5,17,26

However, no large series of MRSA meningitis have been reported.2,5,17,26 We report here the results of a multicenter study on the epidemiology, clinical presentation, response to treatment, and outcome of 86 cases of MRSA meningitis in adults. To our knowledge, this represents the largest study of the infection reported thus far.

Back to Top | Article Outline


Study Design

The reports of all cases of MRSA meningitis in adults diagnosed during a 25-year period (1981–2005) were retrospectively collected from 11 tertiary-care hospitals by use of a uniform questionnaire, as part of a collaborative study. All the participating centers were tertiary teaching hospitals with active neurosurgery programs, as well as referral centers for head trauma.

Back to Top | Article Outline


The present study focuses on episodes of meningitis due to MRSA that were diagnosed from January 1981 to December 2005. The following data were obtained for all patients: age, sex, underlying medical or neurosurgical diseases, associated MRSA infections, clinical features of meningitis, analytical and microbiologic data, response to antibiotic therapy, outcome, and 30-day mortality.

Back to Top | Article Outline

Microbiologic Methods

Microbiologic evaluation was performed in the microbiology laboratories of the participating hospitals. CSF was obtained either by lumbar puncture or from CSF devices. All CSF samples were centrifuged and the sediment was Gram stained and cultured on the usual media for aerobic bacteria following standard procedures. Uncentrifuged aliquots were analyzed for leukocyte count and glucose and protein levels. S. aureus isolates were identified according to standard techniques. Blood cultures were done using automated blood culture systems. Antimicrobial susceptibility was determined by microdilution using Clinical Laboratory Standards Institute methods. An S. aureus strain was considered susceptible to methicillin if the minimal inhibitory concentration (MIC) was ≤2 μg/mL and resistant if the MIC was ≥4 μg/mL.

Back to Top | Article Outline


An adult was defined as an individual aged ≥16 years. Meningitis was defined by isolation of MRSA from 1 or more CSF cultures and/or blood cultures, with clinical manifestations of acute meningitis and typical CSF findings such as pleocytosis (>10 cells/μL), decreased glucose level (<0.40 mg/dL), or increased protein concentration (>0.45 mg/dL). Cases were identified by review of diagnoses from hospital discharge records, records from the database of the infectious disease departments, and clinical microbiology laboratories.

The cases were classified as either postoperative meningitis—that is, meningitis secondary to head trauma, neurosurgery, or CSF devices; and spontaneous meningitis—that is, cases without a history of head trauma or neurosurgical procedures. Cases of meningitis were further classified as community-acquired or nosocomial, based on the 1988 guidelines from the Centers for Disease Control and Prevention.9 Since these guidelines do not specify duration of hospitalization, this was arbitrarily chosen for the study: infection was considered nosocomial if the diagnosis was made after a minimum of 2 days of hospitalization and the infection was neither present nor incubating at the time of admission, the patient was hospitalized within the previous month, or the patient underwent placement of a permanent CSF device within the previous year.

Patients were considered to have mixed bacterial infection when 2 or more bacterial organisms were isolated from CSF cultures in the presence of clinical features of meningitis. In mixed infections, usually contaminant bacteria such as coagulase-negative staphylococci or diphtheroids were considered significant pathogens when they were repeatedly isolated from CSF samples, isolated from other significant clinical samples, or when they were observed on CSF Gram stain.

The following were considered severe comorbidities (underlying medical diseases): cardiovascular disease (congestive heart failure, coronary artery disease, or myocardial infarction), liver disease (clinical or histologic diagnosis of cirrhosis), renal disease (chronic renal disease or abnormal blood urea nitrogen and creatinine concentrations), pulmonary disease (severe asthma or chronic obstructive pulmonary disease), cerebrovascular disease (stroke or transient ischemic attack), diabetes (previously diagnosed disease if patient was receiving hypoglycemic therapy), malignancy (any cancer except malignant brain or skin tumors active at the time of presentation), alcoholism, and immunodeficient states (transplantation, splenectomy, immunosuppressive therapy, or human immunodeficiency virus [HIV] infection). The underlying conditions before the meningitis were rated according to the classic McCabe-Jackson criteria.18

Clinical evidence of meningeal inflammation was assessed by the presence of fever (temperature >38°C), headache, neck stiffness and/or other meningeal signs, altered mental status (confusion or lethargy, coma with response to pain, coma unresponsive to all stimuli), focal neurologic findings, or seizures.6 The severity of the clinical condition was assessed according to the American College of Chest Physicians Consensus Committee as follows: sepsis, severe sepsis, and septic shock.3 Associated MRSA infection was defined by the presence of consistent clinical manifestations at an anatomic site during the episode of meningitis and/or if MRSA was recovered from clinically significant specimens obtained at that site (for example, surgical site, urine, tracheal aspirate).

Antimicrobial therapy was prescribed at the discretion of the responsible medical team according to the general recommendations for treatment of community-acquired and nosocomial bacterial meningitis at the time of diagnosis, and the results of antibiotic susceptibility testing for each isolate. Empirical therapy was defined as the antibiotic administered before the definitive microbiologic testing result was obtained, and it was considered to be appropriate when the isolated MRSA strain was susceptible. Definitive therapy was defined as the antibiotics administered after the identification and susceptibility testing of MRSA. Combined therapy was defined as the use of ≥2 antibiotics active against MRSA. Clinical outcome was assessed by 30-day mortality (defined as death within 30 days from the diagnosis of meningitis). Death was considered to be related to meningitis when it was caused by complications of infection occurring during the first 30 days after diagnosis.

Back to Top | Article Outline

Statistical Analysis

All data were entered into a database and analyzed using SPSS 18.0 software package. Quantitative variables were analyzed with the Student t test or the Mann-Whitney test when appropriate. Qualitative variables were analyzed with the chi-squared test with the Yates correction or the Fisher exact test (2-tailed) when necessary. To identify independent predictors for mortality, a backward logistic regression model was applied. Discriminant analysis was made through the area under the receiver-operation characteristic curve (ROC AUC). Calibration was assessed using the Hosmer-Lemeshow test for goodness-of-fit. All p values were 2-sided and values of 0.05 or less were considered statistically significant.

Back to Top | Article Outline


Demographic and Epidemiologic Data

Over the 25 years of the study, 86 cases of MRSA meningitis were diagnosed at the 11 participating hospitals. Approximately half of the cases (42/86) were detected during the last 6 years (2000–2005) of the study period. The main demographic and epidemiologic data are shown in Table 1. At 1 of the institutions (Hospital Ramon y Cajal, Madrid), S. aureus was responsible for 33 of the 300 cases (11%) of adult bacterial meningitis diagnosed during the 1991–2005 period, and 7 of these 33 cases (21%) were MRSA.



There were 56 men (65%), and the mean age was 51.5 years (range, 16–86 yr). A high proportion of patients (63%) had severe comorbidities such as cerebrovascular disease (34%), cardiovascular disease (33%), diabetes (9%), malignancy (8%), or immunodeficiency (7%). The severity of the underlying conditions according to the McCabe-Jackson classification was as follows: nonfatal, 60 (70%), ultimately fatal, 23 (27%), and rapidly fatal, 3 (3%).

The infection was nosocomial in 80 (93%) cases and community-acquired in 6 (7%) cases. There were 78 (91%) cases of postoperative meningitis and 8 (9%) cases of spontaneous meningitis. The main characteristics of patients with postoperative and spontaneous meningitis are presented in Table 2. Compared with patients with postoperative meningitis, patients with spontaneous meningitis had a significantly higher frequency of community-acquired infection (62% vs. 1%; p < 0.001) and severe comorbidities (100% vs. 59%; p = 0.02). Three patients had nosocomial spontaneous meningitis appearing during hospitalization for severe underlying diseases. In patients with nosocomial meningitis, infection appeared after a median duration of hospitalization of 25.5 days (range, 3–186 d).



The most common neurosurgical condition predisposing to postoperative meningitis was the presence of CSF devices (64 cases), followed by recent neurosurgery (39), CSF leakage (15), and head trauma (10). External ventricular drains were the most common type of CSF device (30 cases), followed by ventriculoperitoneal (VP) shunts (26), and other devices (8). Meningitis was an early complication of CSF devices; infection appeared after a median of 18 days (range, 1–131 d) and 33 days (range, 1–1821 d) after the placement of external ventricular drains and VP shunts, respectively.

Among the 6 patients with community-acquired infection there were 5 cases of spontaneous meningitis and 1 case of postoperative meningitis associated with long-term (>1 yr) VP shunt. Five of the 6 patients (83%) had severe comorbidities and 3 (50%) had ultimately or rapidly fatal conditions, according to the McCabe-Jackson classification. Four of these 5 patients with spontaneous meningitis had severe comorbidities such as lymphoma, renal failure, heart failure, and chronic pulmonary disease that required frequent admissions to the hospital, and meningitis could therefore be considered a health care-associated infection.

Twenty-eight patients (33%) had an associated MRSA infection, with surgical site (16%), pneumonia (9%), central venous catheter (6%), and skin and soft tissue (5%) the most common. The frequency of associated MRSA infection was similar in patients with postoperative and spontaneous infection (32% vs. 37%, respectively; p = 0.54). Among the 5 patients with spontaneous community-onset meningitis there was only 1 case of catheter-related MRSA bacteremia in a patient undergoing hemodialysis. In 2 other patients with spontaneous infection, the acquisition was nosocomial and MRSA meningitis was associated with spondylitis and catheter-related infection, respectively. In addition, a high proportion of patients had other common predisposing factors for nosocomial MRSA infection in the previous month, such as central venous catheter (56%), urinary catheter (51%), antibiotic therapy (49%), or intubation (39%); 15% of them had had a previous episode of MRSA infection during hospitalization.

Back to Top | Article Outline

Clinical Data

The main clinical and analytical data of these patients are shown in Table 3. The clinical course was acute (≤7 d) in 89% of patients. The duration of symptoms before diagnosis ranged from 1 to 30 days (mean, 3.9 d). Most patients presented with fever (89%), altered mental status (68%), headache (40%), and meningeal signs (29%). Focal neurologic deficit (9%), seizures (8%), and petechial rash (2%) were less frequently observed. Abdominal pain was present in 14 (17%) patients, 10 of whom had VP shunts. The clinical presentation was similar in patients with postoperative and spontaneous meningitis with the exception of focal neurologic deficit, which was more frequent in patients with spontaneous infection (see Table 2).



The severity of infection was classified as follows: sepsis (37%), severe sepsis (6%), and septic shock (12%). Septic shock was significantly more frequent in spontaneous than in postoperative meningitis (50% vs. 8%; p < 0.001). Two patients developed disseminated intravascular coagulation. Seven patients developed suppurative complications secondary to MRSA infection during the evolution of meningitis, such as brain abscess (n = 2), and spinal epidural empyema (n = 1); 4 patients with VP shunt developed abdominal complications such as peritonitis (n = 3) or abdominal abscess (n = 1).

Back to Top | Article Outline

Analytical and Microbiologic Data

The most common CSF abnormalities were pleocytosis (90%), elevated protein level (77%), and hypoglycorrhachia (30%). The median CSF leukocyte count was 237 cells/μL, and 52% of patients had >250 cells/μL. CSF Gram stain and CSF culture were positive in 49% (32/65) and 99% (85/86) of patients, respectively. CSF findings were similar in patients with postoperative and spontaneous meningitis (see Table 2). Blood cultures were reported as positive in 36% (16/45) of cases. The incidence of bacteremia was higher in spontaneous (5/8) than in postoperative meningitis (11/37), but the difference was not statistically significant (62% vs. 30%; p = 0.07). Diagnosis was confirmed by blood culture in the only patient with negative CSF culture. MRSA strains were usually resistant to ciprofloxacin (92%), erythromycin (82%), clindamycin (71%), and gentamicin (65%); there was a lower proportion of resistant strains to rifampin (23%) and cotrimoxazole (19%). No cases of vancomycin-resistant strains were detected. MRSA was isolated in other clinical samples in 37 (43%) of the patients: CSF devices (16 cases), surgical wound (10 cases), tracheal aspirate (9 cases), and venous catheter (4 cases) samples were the most common. Twenty patients (23%) with postoperative meningitis (80% of them with CSF devices) had mixed meningeal infection due to gram-negative bacilli (12%), coagulase-negative staphylococci (6%), enterococci (3%), streptococci (3%), or other bacteria (3%) (see Table 3).

Back to Top | Article Outline


Empirical antimicrobial therapy was given to all but 2 patients who died prematurely. Most patients (77%) were initially treated with vancomycin. Empirical antimicrobial therapy was considered appropriate in 81% of cases; initial therapy was considered appropriate more frequently in postoperative than in spontaneous infection (84% v. 50%; p = 0.01). When the diagnosis of MRSA meningitis was established, 84 patients received definitive antimicrobial therapy that was considered appropriate in 82 (98%) of them (Table 4). Most patients received vancomycin (92%) either as monotherapy (64%) or in combination with rifampin (15%), aminoglycosides (6%), or other antibiotics (7%). The median duration of antimicrobial therapy was 18 days (range, 2–55 d). Mean daily dose of vancomycin and rifampin was 1930 mg (range, 1000–3000 mg) and 700 mg (range 600–1200 mg), respectively. Serum vancomycin through level was available for 10 patients and the mean value was 11.3 μg/mL (range, 4–28 μg/mL). Five patients received alternative antimicrobial therapy (see Table 4). In addition, 30% of them (26/86) received intraventricular therapy with vancomycin. The dose and schedule of intraventricular vancomycin was variable, but most patients received 10–20 mg every 24–48 hours, for a median of 10 days (range, 3–22 d). Adjuvant therapy with anticonvulsive drugs, dexamethasone, or mannitol was used in 41%, 27%, and 1% of cases, respectively. CSF devices were removed in 78% (50/64) of patients. The median time from diagnosis of meningitis to device removal was 2.5 days (range, 1–41 d).



Back to Top | Article Outline

Outcome and Prognostic Factors

Overall 30-day mortality was 31% (27/86). Meningitis was the main cause of death in all but 7 patients. Death was an early event in the course of the infection, and most patients (17/27) died during the first 2 weeks after the diagnosis of meningitis (median, 5.5 d; range, 1–30 d). Among the 79 patients receiving definitive therapy with vancomycin (as monotherapy or in combination with other antibiotics), the mortality rate was 29% (23/79).

Table 5 lists the main characteristics of the patients who died compared with those who survived. For univariate and multivariate study all patients presenting with coma (responsive to pain or unresponsive to all stimuli) were analyzed together as a group. Mortality correlated significantly with spontaneous meningitis (p < 0.01), community-acquired infection (p < 0.01), altered mental status (p = 0.01) or coma (p < 0.01), focal neurologic deficit (p < 0.01), septic shock (p < 0.01), and retention of CSF devices (p = 0.01). No other major differences were found between patients who survived or died. The mortality rate was similar in patients receiving monotherapy or combined therapy (30% vs. 33%; p = 0.78) and in patients receiving or not receiving intraventricular therapy (23% vs. 35%; p = 0.32). The mortality was significantly higher in patients whose CSF devices were retained (7/14) compared with patients whose devices were removed (9/50) (50% vs. 18%; p = 0.01).



The following variables were included in the final multivariate analysis: age, sex, spontaneous meningitis, coma, and septic shock. Multivariate analysis identified 2 independent factors associated with mortality: spontaneous meningitis (odds ratio [OR], 21.4; 95% confidence interval [CI], 2.3–195.4; p = 0.007), and coma (OR, 9.7; 95% CI, 2.2–42.3; p = 0.002). The Hosmer-Lemeshow test was successful (0.97) and the value of ROC AUC was 0.72 (95% CI, 0.60–0.85).

Back to Top | Article Outline


S. aureus meningitis is an uncommon infection, accounting for approximately 1%–9% of bacterial meningitis cases.6,11,12,28 Most series describe 2 different forms of staphylococcal meningitis: 1) postoperative, associated with neurosurgical procedures, CSF devices, or head trauma; and 2) spontaneous meningitis, secondary to staphylococcal infection outside the central nervous system.1,4,21,22 Previous studies have focused on infections due to methicillin-susceptible S. aureus. However, in recent years, an increasing proportion of cases (5%–48%) have been caused by MRSA.1,13,22 To our knowledge, only 3 series of MRSA meningitis, including only a small number of patients, have been reported to date.2,17,26

The current multicenter study presents what we believe is the largest series of adult patients with MRSA meningitis, but unfortunately, the study design was not appropriate to assess the temporal trends in the incidence of this infection. However, a high proportion (49%) of cases appeared in the final 6 years (2000–2005) of the 25-year study period, suggesting that the incidence of MRSA meningitis is increasing, as previously reported.1,5,26 One of the limitations of the study was that cases were collected until 2005, and we therefore could not assess the impact of the community-acquired MRSA epidemic on the incidence of MRSA meningitis.

A high proportion of our cases of MRSA meningitis (91%) were postoperative infections, in agreement with the results of previous series (91%–100%).2,17,26 Postoperative meningitis is the most common form of S. aureus meningitis, with a reported frequency of 35%–63%.22,28 CSF devices and recent neurosurgery were the most common predisposing factors for MRSA meningitis, as previously reported.1,2,5,17,26 As a consequence of the postoperative condition of the patients, MRSA meningitis is usually a nosocomial infection that appears in patients with prolonged hospitalization and multiple risk factors for MRSA infection such as central venous or urinary catheters, intubation, or antibiotic therapy.

On the other hand, spontaneous meningitis usually appears as a community-acquired infection in patients with severe underlying diseases,1,4,21,22 although it may also appear in the nosocomial setting. We observed only 6 patients with community-acquired infection. Most of them had severe comorbidities that required frequent hospitalization, and meningitis could therefore be considered a health care-associated infection. We must emphasize that MRSA should be considered a potential cause of meningitis in patients with severe underlying diseases requiring frequent contact with the health care system who present with clinical features of community-onset meningeal infection. Severity of illness and the presence of underlying diseases such as hematologic malignancy, chronic renal failure, and cirrhosis have been associated with an increased risk for MRSA infection.27 A 2009 review described the clinical features of community-acquired MRSA meningitis.19

Approximately one-third of the current patients had an associated MRSA infection. This finding has been reported to be more frequent in patients with spontaneous S. aureus meningitis.1,22 However, we have not observed a significant difference in associated MRSA infection between patients with spontaneous and postoperative meningitis. In spontaneous cases, meningitis develops as a complication of bacteremic S. aureus infection arising from a variety of clinical sources such as paraspinal or epidural abscesses, endocarditis, osteomyelitis, and soft tissue infections.10,12,13,28 In postoperative meningitis, it has been suggested that these infections arise at the time of surgery.7,8,22 It is noteworthy that surgical site infection was the most common associated MRSA infection in our series.

The usual presentation of MRSA meningitis is rapid onset of fever, altered mental status, and headache. This typical clinical course has been previously reported, but we observed a lower proportion of patients with meningeal signs (29%), focal neurologic deficit (9%), and seizures (8%), compared with the results of previous series of MRSA meningitis (60%, 19%–34%, and 27%–30%, respectively).2,17,26 MRSA meningitis is a severe infection, as demonstrated by the high proportion of patients (18%) with severe sepsis or septic shock. Although the clinical presentation of postoperative and spontaneous meningitis was similar, focal neurologic deficit and septic shock were significantly more frequent in patients with spontaneous infection. The higher frequency of septic shock in spontaneous meningitis has been reported in previous series of S. aureus meningitis.1,22

CSF findings in MRSA meningitis are usually consistent with bacterial infection demonstrated by pleocytosis, elevated protein levels, and hypoglycorrhachia, and our data are similar to those described in previous reports.2,17,26 Shunt-related infections have been reported to cause a lower inflammatory meningeal response, and this fact may also explain the lower leukocyte CSF count observed in MRSA meningitis.22,25 CSF Gram stain was positive in approximately half the patients. In contrast to other bacterial meningitis, the lower yield of CSF Gram stain in MRSA meningitis (20%) has been previously reported.2 One of the most prominent features of spontaneous S. aureus meningitis is bacteremia, observed in 64%–100% of patients.7,8,10,12,13,15 Most patients with MRSA meningitis had postoperative infections, and therefore a low proportion of our patients (36%) had bacteremia, in agreement with previous series of postoperative S. aureus meningitis (0%–37%)7,8,12,21 or MRSA meningitis (8%–27%).2,17,26 Although MRSA bacteremia was more frequent in spontaneous than in postoperative cases, the current study did not find a significant difference between the 2 groups.

One of the most relevant findings of the current study was the high proportion (23%) of mixed infections that appeared exclusively in patients with postoperative meningitis. Polymicrobial infection accounts for 1%–14% of bacterial meningitis.6,23 Head trauma, neurosurgery, CSF devices, CSF leakage, and contiguous infected foci are the most important predisposing factors.6,23 Polymicrobial infection is an uncommon finding in S. aureus meningitis, but it seems to be a frequent complication (20%–54%) of MRSA infection that has been associated with the antecedent of neurosurgery and CSF devices.2,17,26

Vancomycin is considered the standard therapy of MRSA meningitis. A high daily dose (45–60 mg/kg) should be administered in adults to maintain serum vancomycin through levels of 15–20 μg/mL.16,29,30 Some researchers have suggested that combinations with rifampin or cotrimoxazole1,10,29 may improve the prognosis, especially in shunt-related infections, but the experience with combined therapy is still inconclusive. Vancomycin has poor CSF penetration in the absence of inflamed meninges or when administered with dexamethasone. Vancomycin failures have been described and may be attributed to its poor CSF penetration.1 However, a 2007 report24 demonstrated that appropriate CSF concentrations of vancomycin can be obtained with high-dose vancomycin (administered as continuous infusion of 60 mg/kg per d after a loading dose of 15 mg/kg), even when concomitant steroids are used. Alternative antibiotics to vancomycin are linezolid or cotrimoxazole.16,29

Most of our patients had postoperative infections and were initially treated with vancomycin, a circumstance that can explain the high proportion of cases receiving appropriate empirical therapy. Most patients received definitive therapy with vancomycin for prolonged periods, and a favorable response was obtained in 71% of them. In addition, some of them received rifampin, aminoglycosides, cotrimoxazole, and/or intraventricular therapy with vancomycin. However, the use of appropriate empirical therapy, combined definitive therapy, or intraventricular vancomycin had no significant impact on the survival rate of our patients. We should acknowledge that the number of patients treated with the different regimens was too small to reach reliable conclusions.

The duration of treatment in S. aureus meningitis and MRSA meningitis has not been established. Although some reports support the use of a 3-week course of therapy,1,2,8 guidelines of the Infectious Diseases Society of America published in 2011 recommend the use of a 2-week course.16 In S. aureus meningitis, the sterilization of CSF cultures usually occurs after a mean period of 1 week.1,11 However, the length of therapy should be based on the primary infection when meningitis is associated with another staphylococcal infection (that is, endocarditis, epidural abscess, etc.).

There have been occasional reports of successful therapy of MRSA meningitis with other antibiotics such as teicoplanin, linezolid, and daptomycin, but no regimen has been proven to be superior for therapy of this infection. In the series of Arda et al,2 6 patients with postoperative MRSA meningitis were treated with teicoplanin (400–800 mg/d) and a favorable response was observed in all of them. In the review by Ntziora et al,20 all 3 patients with MRSA meningitis treated with linezolid (600 mg twice daily) were cured. Linezolid has been proven successful in 3 additional cases of community-acquired MRSA meningitis.19 A 2008 report14 described the favorable response to daptomycin (6 mg/kg every 48 h) in a patient with chronic renal failure and MRSA meningitis secondary to catheter-related bacteremia. However, the experience with these alternative drugs is still limited, and these favorable results should be interpreted with caution because they could be a consequence of publication bias.

Adjuvant steroid therapy has not shown any benefit in staphylococcal meningitis.15 We have observed no beneficial effect of dexamethasone therapy on the mortality rate. However, it is important to remember that the dose, schedule, and timing of administration of the drug were not collected in the study protocol, making it difficult to determine the impact of this therapy on the outcome of infection. In postoperative cases, the removal of the infected CSF shunt is considered 1 of the most important parts of therapy.12,13,22,25 CSF devices were removed in 78% of cases, and this intervention was associated with a significantly lower mortality, as previously described.1,2

Based on our experience and following current guidelines for therapy of severe MRSA infections, we recommend a high dose of vancomycin (15–20 mg/kg per dose, every 8–12 h, to maintain serum vancomycin through levels of 15–20 μg/mL) for initial therapy of MRSA meningitis. Although some experts recommend the addition of rifampin or cotrimoxazole, there is no clear evidence to support the use of combined therapy. Linezolid, cotrimoxazole, and teicoplanin should be considered as alternative antibiotics for patients who fail to respond to vancomycin therapy. Intraventricular vancomycin (10–20 mg/d) should be considered in patients with shunt infections that are difficult to eradicate or who cannot undergo surgery to remove CSF devices. Shunt removal is recommended for shunt infection, and the shunt should not be replaced until CSF cultures are repeatedly negative. There is no clinical evidence to support the use of adjuvant dexamethasone therapy in MRSA meningitis.

The mortality rate of MRSA meningitis is very high (10%–45%), and most patients die as a direct consequence of meningeal infection.2,5,17,26 The wide range of associated mortality may be partially accounted for by the differing patient populations and underlying diseases. Previous reports have identified several factors that may influence mortality in S. aureus meningitis, such as advanced age, severe underlying diseases, spontaneous meningitis, community-acquired infection, altered mental status, bacteremia, septic shock, respiratory failure, hyponatremia, disseminated intravascular coagulation, use of inappropriate antibiotic therapy, and infection with specific staphylococcal strains.1,4,7,8,10,12,13,15,21,22,25,28 No specific risk factors for mortality have been described for MRSA meningitis. It is noteworthy that in 2 small comparative studies,5,26 the mortality rate of MRSA meningitis (35%–45%) was higher than that observed in patients with methicillin-susceptible strains (12%–27%), but this difference was not statistically significant. In agreement with previous reports, we also observed a very high mortality rate (31%). In our experience, the outcome of MRSA meningitis was significantly related to the presence of coma and spontaneous infection. The presence of altered mental status or coma is a classical prognostic factor in bacterial meningitis,6 and previous reports have observed that spontaneous S. aureus meningitis is associated with an adverse outcome.7,8,22

The current multicenter cohort study represents the largest series of adult MRSA meningitis that we know of, and it has enabled us to determine the epidemiology, clinical features, and prognostic factors of this uncommon infection. However, the current study has several limitations. Although the series includes a large number of patients, the data were retrospectively collected and some degree of subjective interpretation could not be eliminated. As an example, we used a categorical classification to assess the alteration of mental status instead a more objective system (such as the Glasgow Coma Score) because this scoring system was missing in most of our clinical records. In addition, some of the clinical features attributed to meningitis could be secondary to the underlying neurosurgical condition.

We have identified spontaneous infection and altered mental status as prognostic factors in MRSA meningitis. However, the classification of meningitis as postoperative or spontaneous infection, although it could be useful in the clinical setting, may be considered arbitrary for investigational purposes. In addition, the researchers had no control over the general management of patients, including the selection of empirical and definitive antimicrobial therapy or the use of adjuvant measures such as dexamethasone, intraventricular therapy, or removal of CSF devices. Although no vancomycin-resistant strains were detected, the influence of vancomycin MIC on the response to therapy could not be established. Finally, although vancomycin was used in most patients at standard doses, vancomycin dose as well as serum through levels were lower than those recommended by current guidelines, and our results are too limited to assess the impact of this parameter on the mortality of the patients.

In conclusion, MRSA is a relatively uncommon but serious disease. Although most cases present as a nosocomial infection complicating the postoperative course of neurosurgical patients, spontaneous meningitis can also appear as a community-acquired infection in patients with severe comorbidities requiring frequent contact with the health care system. Clinical features and CSF abnormalities are similar to those observed in acute meningitis caused by other bacteria. Although most patients have a favorable response to vancomycin, the beneficial effect of combined antimicrobial therapy, intraventricular therapy, or alternative antibiotics needs to be evaluated. MRSA meningitis is associated with a high mortality rate. In our experience, spontaneous infection and coma at presentation were significantly related with an adverse outcome. Removal of CSF devices seems to have a favorable impact on survival.

Back to Top | Article Outline


The authors thank Angel Dominguez (Infectious Diseases Unit, Hospital Virgen Macarena, Sevilla, Spain) and Nieves Sopena (Infectious Diseases Unit, Hospital Germans Trias i Pujol, Badalona, Spain) for their contribution to the study, and Santiago Moreno (Infectious Diseases Service, Hospital Ramón y Cajal, Madrid, Spain) for critical review of the manuscript.

Back to Top | Article Outline


1. Aguilar J, Urday-Cornejo V, Donabedian S, Perri M, Tibbetts R, Zervos M. Staphylococcus aureus meningitis. Case series and literature review. Medicine (Baltimore). 2010; 89: 117–125.
2. Arda B, Yamazhan T, Sipahi OR, Islekel S, Buke C, Ulusoy S. Meningitis due to methicillin-resistant Staphylococcus aureus (MRSA): a review of 10 cases. Int J Antimicrobial Agents. 2005; 25: 414–418.
3. Bone RC, Balk RA, Cerra F, Dellinger RP, Fein AM, Knaus WA, Schein RM, Sibbald WJ. Definitions for sepsis and organ failure and guidelines for the use of innovative therapies in sepsis. The ACCP/SCCM Consensus Conference Committee. American College of Chest Physicians/Society of Critical Care Medicine. Chest. 1992; 101: 1644–1655.
4. Brouwer MC, Keizerweerd GD, De Gans J, Spanjaard L, Van de Beek D. Community acquired Staphylococcus aureus meningitis in adults. Scand J Infect Dis. 2009; 41: 375–377.
5. Chang WN, Lu CH, Wu JJ, Chang HW, Tsai YC, Chen FT, Chien CC. Staphylococcus aureus meningitis in adults: a clinical comparison of infections caused by methicillin-resistant and methicillin-sensitive strains. Infection. 2001; 29: 245–250.
6. Durand M, Calderwood S, Weber D, Miller SI, Southwick FS, Caviness VS, Swartz MN. Acute bacterial meningitis in adults. A review of 493 episodes. N Engl J Med. 1993; 328: 21–28.
7. Falco V, Almirante B, Pahissa A, Gasser I, Fernandez F, Martinez JM. Meningitis caused by Staphylococcus aureus. Analysis of 16 cases. Med Clin (Barc). 1990; 94: 208–211.
8. Fong IW, Ranalli P. Staphylococcus aureus meningitis. Q J Med. 1984; 53: 289–299.
9. Garner JS, Jarvis WR, Emori TG, Horan TC, Hughes JM. CDC definitions for nosocomial infections, 1988. Am J Infect Control. 1988; 16: 128–140.
10. Gordon JJ, Harter DH, Phair JP. Meningitis due to Staphylococcus aureus. Am J Med. 1985; 78: 965–970.
11. Hussein A, Shafran S. Acute bacterial meningitis in adults. A 12-year review. Medicine (Baltimore). 2000; 79: 360–368.
12. Jensen AG, Espersen F, Skinhoj P, Rosdahl VT, Frimodt-Moller N. Staphylococcus aureus meningitis. A review of 104 nationwide, consecutive cases. Arch Intern Med. 1993; 153: 1902–1908.
13. Kim JH, Van der Horst C, Mulrow CD, Corey GR. Staphylococcus aureus meningitis: review of 28 cases. Rev Infect Dis. 1989; 11: 698–706.
14. Lee DH, Palermo B, Chowdhury M. Successful treatment of methicillin-resistant Staphylococcus aureus meningitis with daptomycin. Clin Infect Dis. 2008; 47: 588–590.
15. Lerche A, Rasmussen N, Wandall JH, Bohr VA. Staphylococcus aureus meningitis: a review of 28 consecutive community-acquired cases. Scand J Infect Dis. 1995; 27: 569–573.
16. Liu C, Bayer A, Cosgrove SE, Daum RS, Fridkin SK, Gorwitz RJ, Kaplan SL, Karchmer AW, Levine DP, Murray BE, Rybak M, Talan DA, Chambers HF. Clinical practice guidelines by the Infectious Diseases Society of America for the treatment of methicillin-resistant Staphylococcus aureus infections in adults and children. Clin Infect Dis. 2011; 52: e18–e55.
17. Lu C, Chang W. Adults with meningitis caused by oxacillin-resistant Staphylococcus aureus. Clin Infect Dis. 2000; 31: 723–727.
18. McCabe WR, Jackson GG. Gram negative bacteremia: I. Etiology and ecology. Arch Intern Med. 1962; 110: 847–855.
19. Naessens R, Ronsyn M, Druwe P, Denis O, Ieven M, Jeurissen A. Central nervous invasion by community-acquired methicillin Staphylococcus aureus. J Med Microbiol. 2009; 58: 1247–1251.
20. Ntziora F, Falagas ME. Linezolid for the treatment of patients with central nervous system infection. Ann Pharmacother. 2007; 41: 296–308.
21. Pedersen M, Benfield TL, Skinhoej P, Jensen AG. Haematogenous Staphylococcus aureus meningitis. A 10-year nationwide study of 96 consecutive cases. BMC Infect Dis. 2006; 6: 49.
22. Pintado V, Meseguer MA, Fortun J, Cobo J, Navas E, Quereda C, Corral I, Moreno S. Clinical study of 44 cases of Staphylococcus aureus meningitis. Eur J Clin Microbiol Infect Dis. 2002; 21: 864–868.
23. Pintado V, Cabellos C, Moreno S, Meseguer MA, Ayats J, Viladrich PF. Enterococcal meningitis. A clinical study of 39 cases and review of the literature. Medicine (Baltimore). 2003; 82: 346–364.
24. Ricard JD, Wolff M, Lacherade JC, Mourvillier B, Hidri N, Barnaud G, Chevrel G, Bouadma L, Dreyfuss D. Levels of vancomycin in cerebrospinal fluid of adult patients receiving adjunctive corticosteroids to treat pneumococcal meningitis: a prospective multicenter observational study. Clin Infect Dis. 2007; 44: 250–255.
25. Roberts FJ, Smith JA, Wagner KR. Staphylococcus aureus meningitis: 26 years’ experience at Vancouver General Hospital. Can Med Assoc J. 1983; 128: 1418–1420.
26. Rodriguez-Guardado A, Maradona JA, Perez F, Carton Sanchez JA, Blanco A, Rial JC, Asensi Alvarez V. Postsurgery meningitis by Staphylococcus aureus: comparison between methicillin-sensitive and resistant strains. Med Clin (Barc). 2005; 124: 102–103.
27. Safdar N, Maki DG. The commonality of risk factors for nosocomial colonization and infection with antimicrobial-resistant Staphylococcus aureus, enterococcus, gram-negative bacilli, Clostridium difficile, and Candida. Ann Intern Med. 2002; 136: 834–844.
28. Schlesinger LS, Ross SC, Schaberg DR. Staphylococcus aureus meningitis: a broad-based epidemiologic study. Medicine (Baltimore). 1987; 66: 148–156.
29. Tunkel AR, Hartman BJ, Kaplan SL, Kaufman BA, Roos KL, Scheld WM, Whitley RJ. Practice guidelines for the management of bacterial meningitis. Clin Infect Dis. 2004; 39: 1267–1284.
30. Van de Beek D, Drake JM, Tunkel AR. Nosocomial bacterial meningitis. N Engl J Med. 2010; 362: 146–154.

CI = confidence interval; CSF = cerebrospinal fluid; MIC = minimal inhibitory concentration; MRSA = methicillin-resistant Staphylococcus aureus; OR = odds ratio; ROC AUC = area under the receiver-operation characteristic curve; VP = ventriculoperitoneal

© 2012 Lippincott Williams & Wilkins, Inc.