Jpn. J. Infect. Dis., 60, 262-266, 2007

Original Article Epidemiology of Adult Staphylococcal Meningitis in Southern Taiwan: a Clinical Comparison of aureus and -Negative Staphylococcal Infection Wen-Neng Chang*, Cheng-Hsien Lu, Chi-Ren Huang, Yao-Chung Chuang, Nai-Wen Tsai, Shu-Fang Chen, Chiung-Chih Chang, Hung-Chen Wang1, Chun-Chih Chien2 and Jiunn-Jong Wu3 Department of Neurology, 1Department of Neurosurgery and 2Department of Clinical Pathology, Chang Gung Memorial Hospital-Kaohsiung Medical Center, Chang Gung University College of Medicine, Kaohsiung; and 3Department of Medical Technology, Medical College, National Cheng Kung University, Tainan, Taiwan (Received April 25, 2007. Accepted May 24, 2007) SUMMARY: The clinical and laboratory characteristics of 38 staphylococcal adult bacterial meningitis (ABM) cases (19 and 19 coagulase-negative staphylococcal [CoNS] infections), collected over a period of 6.5 years (July 1999-December 2005; total ABM cases = 181) were analyzed. The results were compared with those of our previous study (January 1986-June 1999; total ABM cases = 202: monomicrobial infection = 180, mixed infection = 22, staphylococcal infection = 27). The 38 staphylococcal meningitis cases were 21 men and 17 women. Fever and altered consciousness were the leading clinical mani- festations. A preceding postneurosurgical state was noted in 12 of the 19 S. aureus infections and all 19 CoNS infections. The ages of onset and mortality rates of S. aureus infection and CoNS infection were 58.21 ± 13.05 years and 36.8% (7/19), and 44.16 ± 17.57 and 5.3% (1/19), respectively. Eleven of the 19 implicated S. aureus strains and 18 of the 19 implicated CoNS strains were mecA gene-positive and methicillin-resistant; all the strains retained their susceptibility to linezolid. The therapeutic results showed a mortality rate of 21% (8/38). This study revealed an increase of methicillin-resistant, postneurosurgical staphylococcal infection in ABM. Patients with CoNS infection had a younger age at onset and a lower mortality rate.

Gung Memorial Hospital (CGMH)-Kaohsiung over a period INTRODUCTION of 6.5 years (July 1999-December 2005). CGMH-Kaohsiung Staphylococcal species, including Staphylococcus aureus is the largest medical center in southern Taiwan. This facility and coagulase-negative staphylococci (CoNS), are not com- is a 2,482-bed acute-care teaching hospital, which provides mon pathogens of adult bacterial meningitis (ABM) (1-3). both primary and tertiary referral care. During this study Adult staphylococcal meningitis is usually seen in patients period, 181 adult patients were determined to have culture- with bacteremia, parameningeal infections or postneurosurgical proven bacterial meningitis, and the study period was chosen states (4-6). Among the preceding postneurosurgical states, to coincide with the time period (13.5 years: January 1986- insertion of an intracranial device is an important risk factor June 1999; 202 cases: 180 with monomicrobial infection, 27 for the development of staphylococcal meningitis (4,5,7, with monomicrobial staphylococcal infection) of our previ- 8). In recent years, adult meningitis caused by methicillin- ous study (3). resistant staphylococcal strains has increased, and resulted in In this study, the criteria for a definite diagnosis of bacte- a therapeutic challenge (4,5,9,10). In this study, the clinical rial meningitis, including the diagnosis of staphylococcal and laboratory characteristics of 38 adult staphylococcal meningitis (1,3-5) were as follows: (i) a positive CSF culture meningitis cases were analyzed, and a clinical comparison of of bacterial pathogen(s), including staphylococcal strains; (ii) S. aureus and CoNS infection was made. In addition, we also clinical features of meningitis, including fever, consciousness made a brief comparison of the clinical features between the disturbance, seizure, or signs of meningeal irritation; and (iii) staphylococcal ABM and non-staphylococcal meningitis. purulent CSF features including at least one of the following parameters: leukocytosis with a leukocyte count > 0.25×109/ L and predominant polymorphonuclear cells, lactate con- MATERIALS AND METHODS centration >3.5 mmol/L, glucose ratio (CSF glucose/serum We retrospectively reviewed the microbiological records glucose) <0.4 or glucose level <2.5 mmol/L if no simulta- of cerebrospinal fluid (CSF) and blood cultures, laboratory neous blood glucose level was determined. Because CoNS is data and medical records of adult patients (≧17 years old) a common contaminant of cultures, the diagnosis of CoNS with culture-proven bacterial meningitis admitted to Chang meningitis was defined with more strict criteria (1,5) only if repeated CSF cultures demonstrated positive results or if *Corresponding author: Mailing address: Department of Neurol- CoNS was cultured from the tip of an indwelling neurosurgical ogy, Chang Gung Memorial Hospital, #123, Ta Pei Road, Niao device. Patients with evidence of concomitant chronic men- Sung Hsiang, Kaohsiung Hsien, Taiwan. Tel: +886-7-731-7123 ingitis or encephalitis that was not due to bacterial pathogens ext. 2283, Fax: +886-7-733-3816, E-mail: cwenneng@ms19. were excluded from this study. Patients were considered to hinet.net have “mixed bacterial meningitis” if at least two bacterial

262 organisms were isolated from the initial CSF culture (11). infection was fatal were compared between these two patient In this study, staphylococcal ABM was classified as groups by means of a Chi-square test or Fisher’s exact test. being either nosocomial or community-acquired infection. Age was compared between the two patient groups by means “Nosocomial” infection was defined as a positive bacterial of a Student’s t test. CSF data for the WBC counts, glucose, infection not present when the patient was admitted to the total proteins and lactate concentrations for the two patient hospital, clinical evidence of an infection no sooner than 48 h groups were logarithmically transformed to improve normal- after admission, or clinical evidence of meningitis within a ity, and comparisons between the two patient groups were short period of time, usually within 1 month after a discharge made using the Student’s t test. Multiple logistic regression from the hospital where the patient had received an invasive analysis was used to evaluate the relationships among vari- procedure, especially a neurosurgical procedure. Otherwise, ables and the two patient groups adjusted for other potential the patient was considered to have “community-acquired” confounding factors. All statistical analyses were conducted infection. Patients who developed meningitis related to head using an SAS software package, version 9.1 (2002; SAS Sta- trauma with skull fractures or neurosurgical procedures were tistical Institute, Cary, N.C., USA). With the same statistical classified as having “postneurosurgical” meningitis. Other- methods, a clinical comparison of the clinical and laboratory wise, patients who demonstrated no clear distinctive disease features between the groups of staphylococcal meningitis and characteristics and who had not undergone any invasive non-staphylococcal meningitis was also made. procedures were classified as having “spontaneous” menin- gitis. RESULTS In this study, the initial consciousness level for adult patients was classified into two groups: (I) normal conscious- Of the 181 culture-proven ABM cases collected during this ness (Glasgow coma scale [GCS] score = 15); and (II) altered study period, 165 cases had monomicrobial infection, and consciousness (GCS < 15). “Leukocytosis” was defined as the other 16, mixed bacterial infection. Table 1 lists the causa- a peripheral white blood cell (WBC) count > 10 × 109/L, tive pathogens of the 181 cases. A postneurosurgical state as “leukopenia” as a WBC count < 5 × 109/L, and “thrombocy- the preceding event was found in 103 of the 181 ABM cases. topenia” as a platelet count < 140×109/L. Blood culture was Of the 165 monomicrobial ABM cases, 101 cases had Gram- carried out in all patients with clinical evidence of meningi- negative pathogen infection, and the other 64, Gram-positive tis and bacteremia was considered as “positive” only when pathogen infection. Among the 64 implicated Gram-positive multiple blood cultures grew the same bacterial pathogen. pathogens, staphylococcal strains were the most common, Over the study period, intravenous administration of peni- including 19 S. aureus strains and 19 CoNS strains. The sub- cillin G or vancomycin with one of the third generation types of the 19 CoNS strains were as follows: S. epidermidis cephalosporins (ceftriaxone, ceftazidime) or cefepime were (n = 13), S. haemolyticus (n = 5) and S. capitis (n = 1). the initial empiric used in the treatment of adult The clinical and laboratory features of the 38 staphylococ- patients with clinical evidence of bacterial meningitis in our cal ABM cases are listed in Table 2. Of the 19 cases with S. hospital. Further adjustment was guided by the aureus infection, 12 were of the postneurosurgical form and results of pathogenic identification and antibiotic suscepti- the other 7 were of the spontaneous form. Nine S. aureus bility tests, and the adjusted antibiotics included oxacillin, strains of the former 12 cases and 2 of the latter 7 cases were imipenem, and meropenem. Antibiotic susceptibility of the isolated pathogen was tested Table 1. Causative pathogens of the 181 adult bacterial meningitis cases using the Kirby-Bauer diffusion method (BBL, Muller-Hinton Spontaneous Post-NS II agars; Becton Dickinson Microbiology Systems, Cockeysville, Pathogen % meningitis meningitis Md., USA). Polymerase chain reaction (PCR) for mecA gene detection was also carried out for those methicillin-resistant Staphylococcus spp. 7 31 21.0 staphylococcal strains. The mecA gene was amplified with pneumoniae 4 2 3.3 specific primer pairs as described by Kampf et al. (12), and Klebsiella pneumoniae 35 7 23.2 this amplification resulted in a 533-bp DNA fragment for Acinetobacter baumannii 21710.5 the mec gene. Ten microliters of the amplified mixture was Escherichia coli 666.6 analyzed on a 1.5% agarose gel. The gel was stained with Pseudomonas spp. 4 6 5.5 ethidium bromide (Sigma Chemical, St. Louis, Mo., USA) Enterobacter cloacae 073.9 and visualized with UV light. Colony hybridization with a 6 2 4.4 32P-labelled probe was performed as described previously Proteus mirabilis 211.7 (13). The probe was the 533-bp DNA fragment of the mecA spp. 1 4 2.8 gene purified from the above PCR products and confirmed Neisseria meningitidis 201.1 by direct nucleotide sequencing analysis with an automated nucleatum 021.1 Model 377 Sequencer (Applied Biosystems Div., Perkin- monocytogenes 201.1 Elmer Corp., Norwalk, Conn., USA). The cultured CoNS Salmonella spp. 2 0 1.1 strains were also examined for subtype identification. API Citrobacter diversus 010.6 STAPH (bioMérieux Vitek, Inc., Hazelwood, Mo., USA) was Fusobacterium nucleatum 100.6 used for the initial subtype identification and the results were Serratia marcescens 010.6 further confirmed by PCR-sequencing of bacterial 16S r-RNA Group A β-hemolytic streptococcus 0 1 0.6 (14). Non-A, non-B, non-D streptococcus 0 1 0.6 A clinical comparison of the 19 S. aureus infections and Micrococcus 0 1 0.6 the 19 CoNS infections was made. Data including gender, Sphingomonas paucimobilis 100.6 type of acquisition of infection, type of infection, underlying Mixed infection 3 13 14.4 conditions, clinical manifestations, and whether or not the Post-NS, postneurosurgical.

263 Table 2. Clinical comparison of Staphylococcus aureus and coagulase-negative staphylococcal meningitis S. aureus infection CoNS infection P-value n = 19 n = 19 (1) Age at meningitis (y) 58.21 ± 13.05 44.16 ± 17.57 0.008 (2) Gender Female 4 13 0.461 Male 15 6 (3) Clinical feature Fever 15 16 1 Altered consciousness 8 10 0.516 Seizure 5 3 0.693 Septic shock 1 0 1 Hydrocephalus 6 12 0.051 CSF leakage 4 1 0.34 (4) Type of infection Spontaneous meningitis 7 0 0.008 Postneurosurgical meningitis 12 19 (5) Methicillin susceptibility Methicillin-resistance 11 18 0.042 Methicillin-sensitivity 8 1 (6) Underlying diseases Diabetes mellitus 3 4 1 Alcoholism 2 1 1 Intravenous drug abuser 2 0 0.486 End-stage renal disease 1 0 1 Systemic lupus erythematosus 0 1 1 Malignancy 3 0 0.23 (7) Peripheral blood study Bacteremia 4 4 1 Leukocytosis 13 14 0.721 (8) CSF data on admission Glucose (mmol/L) 3.67 ± 2.22 3.36 ± 1.03 0.62 Total protein (g/L) 1.76 ± 1.16 0.89 ± 0.96 0.035 Lactate (mmol/L) 5.87 ± 3.25 4.98 ± 2.1 0.392 White cell count (×106/L) 4,313.88 ± 1,6216 964 ± 1,693.14 0.722 (9) Fatality 7 1 0.042 CoNS, coagulase-negative staphylococcus; CSF, cerebrospinal fluid. methicillin-resistant. All 19 cases with CoNS infection had a regression analysis, age at onset of meningitis and fatality postneurosurgical condition as the preceding event, and 18 of were independently associated as differences between these the implicated CoNS strains were methicillin-resistant. Despite two infection groups. Table 3 lists the clinical and laboratory this, all 29 methicillin-resistant staphylococcal strains still data of the non-staphylococcal ABM patients. The compara- retained their susceptibility to linezolid. All 29 methicillin- tive study of staphylococcal infection and non-staphylococcal resistant staphylococcal strains showed positive mecA gene infection revealed that CSF data, presence of CSF leakage, detection. The postneurosurgical states of the 31 staphylo- liver cirrhosis as the underlying disease, acquisition of infec- coccal ABM patients were as follows: ventriculoperitoneal tion and type of infection were significant factors. Variables shunt (n = 12), extraventricular drainage (n = 10), craniectomy used in the logistic regression analysis included the above- (n = 3), head injury (n = 1), traumatic intracranial hemorrhage mentioned significant factors, but only the presence of CSF s/p craniotomy (n = 1), craniotomy (n = 1), subarachnoid leakage at the time of admission (P = 0.004, OR = 54.592, hemorrhage s/p extraventricular drainage (n = 1), traumatic 95% CI = 1.118-2,675.3) was independently associated with subdural hemorrhage s/p craniectomy (n = 1); and lumbar staphylococcal ABM. intrathecal port-A (n = 1). With antibiotic therapy consisting of at least intravenous vancomycin, 30 of the 38 staphylo- DISCUSSION coccal ABM cases survived, while the other 8 died. Of the latter 8 cases, 7 had S. aureus infection, while the remaining When compared with the results of a previous study on the case had CoNS infection. Of these 8 expired cases, 5 had epidemiology of ABM in southern Taiwan (3), this study postneurosurgical infection and the other 3, spontaneous showed an annual increase of ABM cases in our hospital, infection. increasing from 15.0 (202/13.5) cases/year to 27.8 (181/6.5) The comparative analysis of S. aureus and CoNS infection cases/year, and an increase of incidence of postneurosurgical revealed that age at onset of meningitis, methicillin suscepti- states as the preceding event of ABM patients, increasing from bility, CSF total protein level and fatality were significant 40.1% (81/202) to 56.9% (103/181). These increases were factors. After these significant factors were used in the logistic due to the rapid expansion of both the neurological and

264 Table 3. Comparative study of staphylococcal meningitis and monomicrobial non-staphylococcal meningitis Staphylococcal ABM Non-staphylococcal ABM P-value (n = 38) (n = 127) (1) Age at meningitis (y) 58.2 ± 16.8 54.5 ± 16.5 0.247 (2) Gender Female 10 36 1.000 Male 28 91 (3) Clinical feature Fever 31 118 0.057 Disturbed consciousness 18 81 0.089 Seizure 8 45 0.115 Septic shock 1 12 0.302 Hydrocephalus 18 43 0.179 CSF leakage 5 2 0.008 (4) Acquisition of infection Community-acquired 15 75 0.041 Nosocomial-acquired 23 52 (5) Type of infection Spontaneous meningitis 7 68 <0.001 Postneurosurgical meningitis 31 59 (6) Underlying diseases Diabetes mellitus 7 43 0.074 Alcoholism 3 13 1.000 Intravenous drug abuser 2 2 0.227 End-stage renal disease 1 5 1.000 Systemic lupus erythematosus 1 1 0.409 Liver cirrhosis 0 15 0.023 Malignancy 3 11 1.000 (7) Peripheral blood study Bacteremia 8 36 0.412 Leukocytosis 27 84 0.694 (8) CSF data on admission Glucose (mmol/L) 3.49 ± 0.32 2.51 ± 0.27 0.002 Total protein (g/L) 1.37 ± 0.20 4.42 ± 0.39 <0.001 Lactate (mmol/L) 5.42 ± 0.51 13.54 ± 0.83 <0.001 White cell count (×106/L) 2.74 ± 2.09 13.71 ± 6.88 0.002 (9) Survival Alive 30 85 0.227 Death 8 42 AMB, adult bacterial meningitis; CSF, cerebrospinal fluid. neurosurgical departments in our hospital in recent years In this study, most (81.6%; 31/38) of the cases of staphylo- (15). Despite these increases, Klebsiella pneumoniae was still coccal ABM, and particularly those with CoNS infection, were the most common pathogen of ABM. This is a unique epide- considered postneurosurgical infections. The presence of a miologic finding in Taiwan (3,16). In recent years, an increas- postneurosurgical state is well known as an important preced- ing incidence of ABM, especially in patients with nosocomial ing event for the development of staphylococcal ABM (4-8). infection, has also been reported in other Asian countries (17). This epidemiologic trend of staphylococcal ABM in Gram- The same hospital-based study also showed an increase in positive pathogen infections was demonstrated in one of our the incidence of staphylococcal infection, which rose from previous studies on young subjects with ABM (18), in which 15.0% (27/180) to 23.0% (38/165), and a decrease in the 74% of the patients had a preceding postneurosurgical state, incidence of infection, which fell and staphylococcal infections outnumbered S. pneumoniae from 10.6% (19/180) to 3.6% (6/165), in patients with infections, accounting for 29.6% (16/54) of the monomicrobial monomicrobial ABM. Staphylococcal strains had replaced cases of ABM in the young subjects. In this study, S. epidermidis S. pneumoniae, becoming the most common Gram-positive was the most commonly implicated CoNS strain, accounting pathogen of ABM in our hospital. S. pneumoniae is the most for 68.4% (13/19) of the cases, and S. haemolyticus was the commonly implicated Gram-positive pathogen of ABM in second most common, accounting for 26.3% (5/19) of them. both Western and Eastern countries (1,2,16,17). The exact This relative frequency of implicated CoNS subtypes in ABM cause of its decrease in incidence in this study was not clear, is consistent with that of CoNS-related bacteremia (19,20). but it may have been related to the fact that most of the previ- Among the implicated staphylococcal strains, 76% (29/38) ously reported S. pneumoniae ABM patients contracted the of them were methicillin-resistant. The high incidence of mecA infection spontaneously within a community setting, while in gene-positive, methicillin-resistant staphylococcal strains, the present study, the infections were mostly postneurosurgical particularly among cases of postneurosurgical ABM, is a and nosocomial. serious problem of CNS infection in Taiwan (4,5,9,21), and

265 has also resulted in a therapeutic challenge in the choice of Infect., 10, 709-717. initial empiric antibiotics (10,22,23). Together, these problems 3. Lu, C.H., Chang, W.N. and Chang, H.W. (2000): Adult bacterial menin- gitis in southern Taiwan: epidemiologic trend and prognostic factors. J. indicate that the infection control program in the hospitals Neurol. Sci., 182, 36-44. in Taiwan is inadequate (24), and that there is persistent 4. Chang, W.N., Lu, C.H., Wu, J.J., et al. (2000): Staphylococcus aureus dissemination of staphylococcal strains with classic methi- meningitis in adults: a clinical comparison of infections caused by cillin resistance, especially in patients with nosocomial infec- methicillin-resistant and methicillin-sensitive strains. Infection, 29, 245- 250. tions (25). 5. Huang, C.R., Lu, C.H. and Wu, J.J. (2005): Coagulase-negative In the present study, the comparison of S. aureus infection staphylococcal meningitis in adults: clinical characteristics and thera- and CoNS infection among the ABM cases showed that peutic outcome. Infection, 33, 56-60. patients with CoNS infection had a younger age of onset and 6. Pedersen, M., Benfield, T.L. and Skinhoej, P. (2006): Haematologenous Staphylococcus aureus meningitis. A 10-year nationwide study of 96 a lower mortality rate than those with S. aureus infection. consecutive cases. BMC Infect. Dis., 6, 49. This tendency may have resulted from the fact that the ABM 7. Filka, J., Huttova, M. and Tuharsky, J. (1999): Nosocomial meningitis patients with CoNS infection had a higher incidence of in children after ventriculoperitoneal shunt insertion. Acta Paediatr., 88, postneurosurgical state as the preceding event than did the 576-578. 8. Spanu, T., Romano, L. and D’Inzeo, T. (2005): Recurrent ventriculo- ABM patients with S. aureus infection (4,5,18). Despite there peritoneal shunt infection caused by small-colony variants of Staphylo- being no significant difference among the underlying medi- coccus aureus. Clin. Infect. Dis., 41, e48-52. cal and surgical conditions between the patient groups, as 9. Lu, C.H. and Chang, W.N. (2000): Adults with meningitis caused by shown in Table 1, patients with CoNS infection had a higher oxacillin-resistant Staphylococcus aureus. Clin. Infect. Dis., 31, 723- survival rate than those with S. aureus infection. It is known 727. 10. Sabbatani, S., Manfredi, R. and Frank, G. (2005): Linezolid in the treat- that many factors, including comorbidity and underlying criti- ment of severe central nervous system infections resistant to recommended cal illness, may influence the prognosis of ABM, including antimicrobial compounds. Infez. Med., 13, 112-119. staphylococcal ABM (2,4-6). This study also showed a high 11. Chang, W.N., Lu, C.H. and Huang, C.R. (2000): Mixed infections in mortality rate (21%, 8/38) of overall staphylococcal ABM, adult bacterial meningitis. Infection, 28, 8-12. 12. Kampf, G., Weist, K. and Widsinski, S. (1997): Comparison of screening which may have been related to the high incidence of critical methods to identify methicillin-resistant Staphylococcus aureus. Eur. J. postneurosurgical state of the patients with staphylococcal Clin. Microbiol. Infect. Dis., 16, 301-307. ABM. As to the comparison of staphylococcal ABM and 13. Grunstein, M. and Hogness, D.S. (1975): Colony hybridization: a method non-staphylococcal ABM, patients of the former group had a for the isolation of cloned DNAs that contain a specific gene. Proc. Natl. Acad. Sci. USA, 72, 3961-3965. significantly higher incidence of CSF leakage. This differ- 14. Chang, W.N., Wu, J.J. and Huang, C.R. (2002): Identification of viridans ence may have resulted from the relatively higher incidence streptococcal species causing bacterial meningitis in adults in Taiwan. of postneurosurgical state as the preceding event in the cases Eur. J. Clin. Microbiol. Infect. Dis., 21, 393-396. of staphylococcal ABM. 15. Lu, C.H., Chang, W.N. and Lin, Y.C. (2002): Bacterial brain : microbiological features, epidemiological trends and therapeutic out- In conclusion, there is an increasing incidence of staphylo- comes. Q. J. Med., 95, 501-509. coccal infection in ABM patients. Most of these cases are 16. Tang, L.M., Chen, S.T. and Hsu, W.C. (1999): Acute bacterial meningitis associated with a preceding postneurosurgical state, and in adults: a hospital-based epidemiological study. Q. J. Med., 92, 719- most of the implicated staphylococcal strains are mecA gene- 725. positive and methicillin-resistant. S. epidermidis is the most 17. Khwannimit, B., Chayakul, P. and Geater, A. (2004): Acute bacterial meningitis in adults: a 20 year review. Southeast Asian J. Trop. Med. common subtype of the implicated CoNS infection in ABM. Public Health, 35, 886-892. The comparison of S. aureus infection and CoNS infection 18. Tsai, M.H., Lu, C.H. and Huang, C.R. (2006): Bacterial meningitis in among ABM cases revealed that patients with CoNS infec- young adults in southern Taiwan: clinical characteristics and therapeutic tion had a younger age of onset and a lower mortality rate outcome. Infection, 34, 1-8. 19. Herwaldt, L.A., Boyken, L.D. and Pfaller, M.A. (1990): Biotyping of than those with S. aureus infection. However, the extensibility coagulase-negative staphylococci. 108 isolates from nosocomial blood- of these conclusions may be limited due to the small number stream infection. Diagn. Microbiol. Infect. Dis., 13, 461-466. of patients studied. An additional large-scale study will be 20. Simkovicova, M., Hanzen, J. and Milosovic, P. (2001): Coagulase nega- needed for a better delineation of these differences. tive staphylococci isolated on blood cultures. Epidemiol. Mikrobiol. Imunol., 50, 117-120. Nontheless, this study has demonstrated an epidemiological 21. McDonald, L.C., Lauderdale, T.L. and Shiau, Y.R. (2004): TSAR change in staphylococcal infection in ABM patients, and this Paticipating Hospitals: The status of in Taiwan change may present a challenge in terms of the therapeutic among Gram-positive pathogens: The Taiwan Surveillance of Anti- strategy of ABM, especially when facing patients with microbial Resistance (TSAR) programme. 2000. Int. J. Antimicrob. Agents, 23, 362-370. postneurosurgical meningitis and in choosing initial empiric 22. Tunkel, A.R., Hartman, B.J. and Kaplan, S.L. (2004): Practice guide- antibiotics. lines for the management of bacterial meningitis. Clin. Infect. Dis., 39, 1267-1284. 23. Jarlov, J.O. (1999): Phenotypic characteristics of coagulase-negative REFERENCES staphylococci: typing and antibiotic susceptibility. APMIS, Suppl. 91, 1. Durand, M.L., Calderwood, S.B., Weber, D.J., et al. (1993): Acute 1-42. bacterial meningitis in adults: a review of 493 episodes. N. Engl. J. 24. Mollitt, D.L. (2002): Infection control: avoiding the inevitable. Surg. Med., 328, 21-28. Clin. North Am., 82, 365-378. 2. Meyer, C.N., Samuelsson, I.S., Galle, M., et al. (2004): Adult bacterial 25. Huang, A.H., Yan, J.J. and Wu, J.J. (2000): Rapid dissemination of meningitis: aetiology, penicillin-susceptibility, risk factors, prognostic Staphylococcus aureus with classic oxacillin resistance phenotype at a factors and guidelines for empirical antibiotic treatment. Clin. Microbiol. new university hospital. J. Hosp. Infect., 44, 309-315.

266