Cerebral haemodynamics in acute bacterial meningitis in adults
From the Departments of 1Neurology and 3Radiology, Chang Gung Memorial Hospital-Kaohsiung Medical Center, Chang Gung University College of Medicine, and 2Department of Biological Science, National Sun Yat-Sen University, Kaohsiung, Taiwan
Address correspondence to Dr Wen-Neng Chang, Department of Neurology, Chang Gung Memorial Hospital, 123 Ta Pei Road, Niao Sung Hsiang, Kaohsiung Hsien, Taiwan. email: chlu99{at}ms44.url.com.tw
Received 18 March 2006 and in revised form 4 August 2006
 |
Summary |
|---|
 Top Summary IntroductionMethodsResultsDiscussionReferences |
|---|
Background: Vascular complications are an important cause of neurological sequelae among adult survivors of acute bacterial meningitis (ABM).
Aim: To examine the haemodynamic changes associated with ABM.
Methods: Serial transcranial colour-coded sonography (TCCS) and magnetic resonance angiography (MRA) were used to examine cerebrovascular changes in adult ABM patients. Outcome at 3 months was categorized using a modified Barthel index.
Results: We recruited 24 patients, 12 men and 12 women, aged 21–68 years. Mean cerebral blood flow velocity (Vmean) increased from day 1 to day 4 in the middle cerebral artery (MCA), anterior cerebral artery (ACA) and posterior cerebral artery (PCA). On day 4, Vmean values in the MCA, ACA and PCA were all significantly higher than reference values in healthy volunteers. At 3 months follow-up, 16 cases had good outcomes, while the other eight had poor outcomes. Under multiple logistic regression analysis, only Glasgow coma score (GCS) at admission was independently associated with the three-month outcome.
Discussion: In these patient, stenosis as demonstrated by TCCS did not wholly coincide with stenosis as demonstrated by MRA, and the presence of intracranial stenosis was not predictive of a poor outcome at 3 months. Further studies are needed to delineate the characteristics and significance of cerebrovascular changes in adult ABM.
|
Introduction |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionReferences |
|---|
Despite the advent of new antimicrobial drugs and modern imaging techniques, mortality and morbidity in acute bacterial meningitis (ABM) both remain high.1,2 Cerebrovascular complications during the acute phase of bacterial meningitis are an important cause of neurological sequelae among ABM survivors.3–5 In studies where transcranial Doppler sonography was used to predict the neurological outcome of ABM,4–7 cerebral mean blood flow velocity (Vmean) was the only parameter used to evaluate the severity of the narrowing of large intracranial vessels in ABM, although several other factors may interfere with the measurement of this parameter.8–10 The associated risks militate against using cerebral angiography to evaluate the frequency of intracranial arterial stenoses in ABM patients. However, new diagnostic techniques, particularly three-dimensional time-of-flight (3-D TOF) magnetic resonance angiograms (MRA) and transcranial colour-coded sonography (TCCS), have revived interest in haemodynamic changes of the cerebrovascular circulation in ABM. MRA and TCCS offer non-invasive diagnostic methods for evaluating occlusive lesions of the intracranial arteries in symptomatic patients, and studying the pathogenesis, natural course, and dynamics of cerebrovascular complications of ABM.11,12 We report our experience with both MRA and TCCS in the evaluation of cerebral haemodynamics in adult ABM patients.
|
Methods |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionReferences |
|---|
From July 2004 to June 2005, 24 adult (aged
16 years) ABM patients at Chang Gung Memorial Hospital (CGMH)-Kaohsiung were enrolled in this study. CGMH-Kaohsiung, a 2482-bed acute-care teaching hospital, is the largest medical centre in southern Taiwan, providing both primary and tertiary referral care. Criteria for a diagnosis of ABM included positive culture from cerebrospinal fluid (CSF) in patients with clinical presentations of bacterial meningitis, plus at least one of the following CSF indicators of purulent inflammation: (i) leukocytosis (leukocytes >250 x 106/l and predominant polymorphonuclear (PMN) cells); (ii) CSF glucose ratio (CSF glucose:blood glucose) <0.4; or (iii) CSF glucose <2.5 mmol/l if no blood glucose measurement was available.13,14 In patients with negative culture results, a diagnosis of ABM was based on compatible clinical features and pleocytosis of at least 100 x 106 PMN/l.13 Patients were enrolled in this study only if full informed written consent was obtained from the patient or their family. Patients with the following conditions were excluded: (i) diffuse atherosclerotic changes on intracranial and extracranial vessels, with or without evidence of old cerebral infarctions; and (ii) central nervous disorders with various neurological deficits.
All patients received complete medical and neurological examinations, brain magnetic resonance imaging (MRI) and MRA, extracranial colour-coded duplex sonography (ECCS) and TCCS examinations at the time of admission. Neurologists and neuroradiologists integrated the clinical manifestations and the findings of neuroimages.
Ultrasound examinations used a pulsed Doppler device, with a 4 MHz probe for extracranial evaluation and a 2 MHz probe for transcranial evaluation (Acuson). TCCS recordings were made within 24 h of hospitalization according to published criteria.15,16 Follow-up TCCS examinations included the same recordings, and were fixed for days 4, 7, 10, 14, 17, and 21 after admission. The maximum systolic and end-diastolic velocities (Vpeak-systolic and Vend-diastolic) and Vmean were measured in each artery, after blood velocities of 10 cardiac cycles showing a constant velocity steady state. To exclude misinterpretation of the flow velocities caused by ventilation in patients who were mechanically ventilated, arterial pCO2 was measured by blood gas analysis. Data regarding arterial blood pressure, arterial blood gas and Glasgow coma score (GCS) were collected prospectively at the time the recording was made.
Our diagnostic criteria for intracranial arterial stenoses included positive corresponding findings in both MRA and TCCS studies. Criteria used for the diagnosis of middle cerebral artery (MCA) stenosis16–19 were as follows: (i) peak flow velocity >157 cm/s in the affected MCA; (ii) side-to-side difference of mean velocity in the MCAs >30 cm/s; (iii) a >80 cm/s increase in the mean flow velocity in one or two adjacent segments of the MCA, which was insonated in 5-mm steps; (iv) Vmean >120 cm/s in the affected MCA; and (v) a ratio of Vmean between the MCA and internal carotid artery (ICA) (MCA/ICA ratio) of >3. Furthermore, a Vmean >100 cm/s in the affected anterior cerebral artery (ACA),19 Vmean > 95 cm/s in the basilar artery (BA),19 and Vmean > 85 cm/s in the affected posterior cerebral artery (PCA),12,19 indicated stenosis in the respective vessels. Corresponding peak systolic velocity cut-offs for 50% and <50% stenoses were 155 and 120–155 cm/s in ACA, 220 and 155–220 cm/s in MCA, 145 and 100–145 cm/s in PCA, and 140 and 100–140 cm/s in BA, respectively.12
In this study, brain MRIs were done at the time of admission for all patients diagnosed with ABM, as well as a follow-up brain MRI study every week during hospitalization. Brain MRI examinations used a 1.5T scanner (Signa, Horizon GE Medical Systems). Pulse sequences of brain MRI studies, including axial and sagittal T1-weighted and T2-weighted images, and gadolinium-DTPA, were used for all patients, on coronal and axial T1-weighted images. Diffusion-weighted imaging (DWI) and apparent diffusion coefficient (ADC) mapping was performed for all patients. Methods used with the MRA technique included 3-D TOF of intracranial vessels and contrast MRA of neck vessels. Maximum intensity projection (MIP) images were reviewed on a PACS (picture archiving and communications system). MIP MRA images were evaluated for stenosis. The vascular distribution of lesions was documented. Stenosis was measured using the calibration markers on each image. Measurements were standardized using the outer margins of the vessel wall. An assessment of stenosis was made by comparing the diameter of the affected segment of the vessel with the diameter of the nearest normal vessel segment. Lesions were defined as proximal if the most proximal segments of the ACA (A1 segment) or MCA (M1 segment) were involved. Distal was defined as involvement of segments beyond the M1 or A1 segment. The same approach was adopted for the most proximal segment of the PCA (P1 segment).
Severities of neurological deficit were assessed by using the GCS at the time of TCCS recording, during the acute phase of bacterial meningitis.20 Cerebral infarction was defined according to WHO criteria,21 and cerebral infarction secondary to meningo-encephalitis was defined as new-onset cerebral infarction demonstrated by brain MRI during hospitalization. Hydrocephalus was defined on radiographic grounds, requiring the demonstration of progressive large ventricles without sulcal enlargement, compared with previous neuroimaging studies. In every patient, lumbar puncture was performed at the time of admission, and the CSF specimens were analysed for pathogen cultures, cell counts, and levels of protein, lactate and glucose.
The TCCS results were compared with the normal values for our laboratory. These normal values were obtained from 24 healthy volunteers aged from 24 to 63 years (13 males, 11 females; mean ± SD age 48.4 ± 16.4). The mean values of Vmean and pulsatility index (PI) were 54.2 ± 13.5 cm/s and 0.78 ± 0.11, respectively for the MCA; 40.1 ± 8.0 cm/s and 0.79 ± 0.12, respectively, for the ACA; 30.7 ± 5.6 cm/s and 0.70 ± 0.13, respectively, for the PCA; and 37.7 ± 10.7 cm/s and 0.73 ± 0.15, respectively, for the BA.
Therapeutic outcomes >3 months after discharge were evaluated by using a modified Barthel Index (MBI).22 For the purposes of analysis, a score <12 was defined as a poor outcome, and 12 as good. Death was included in the poor outcome group. Data for the time course of the flow velocity data of patients were compared with the control group using independent t-test. Comparisons of clinical manifestations and neuroimaging findings between the two patient groups (good and poor outcome) used the 
2 test or Fisher's exact test. Difference in age between the two patient groups were analysed using the student's t test. CSF glucose, total protein, WBC counts, and glucose ratios were logarithmically transformed to improve normality, and comparisons between the two patient groups were made using Student's t test. GCS comparisons used the Wilcoxon rank sum test. Stepwise logistic regression was used to evaluate the relationships between clinical factors and therapeutic outcomes, with adjustments made for other potential confounding factors. All analyses were done with SAS (1990).
|
Results |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionReferences |
|---|
Baseline characteristics
The clinical characteristics of the 24 enrolled cases and the normal controls are listed in Table 1. The time course data for cerebral Vmean and PI of the intracranial arteries and neuroimaging findings are presented in Tables 2 and 3. Mean ± SD systolic blood pressure (BP), diastolic BP, mean BP, body temperature and Glasgow coma score for the patient group during the TCCS study are listed in Table 2, but no comparison with the normal control group was performed.
|
|
|
TCCS findings
Eighteen of the 24 cases could be insonated bilaterally; six could only be insonated unilaterally, because of a lack of an adequate temporal bone window. TCCS recordings were performed in a total of 42 MCAs, 40 ACAs, 39 PCAs, and 24 BAs. Nine of the 42 MCAs showed stenotic velocities according to our TCCS criteria, but none of the other vessels did so. In total, seven patients had stenotic velocities, four with M1 stenosis (<50%) bilaterally and the remaining three with M1 stenosis (<50%) unilaterally. The Vmean of MCA, ACA and PCA increased from day 1 to day 4; on day 4, Vmean in the MCA, ACA and PCA were significantly higher than those of healthy volunteers. The Vmean of the PCA remained significantly higher than that of the healthy volunteers until day 10. The Vmean of the BA was not significantly different from the healthy volunteers from day 1 to day 21. The mean PIs in ACA, MCA, PCA and BA were higher from day 1 to day 21, although not significantly so when compared to healthy volunteers.
MRA and MRI findings
Six patients (25%) had abnormal results in MRA studies; the remaining 18 had normal results. Distribution of the involvement of the vessels among the former six patients was as follows: 3 M1 stenosis (<50%) bilaterally; 1 M1 stenosis (<50%) unilaterally; 2 M1 stenosis (<50%) bilaterally plus A1 stenosis (<50%) unilaterally. During hospitalization, follow-up of these six cases showed stabilization in one, improvement in three, and progression and developing cerebral infarctions in the other two. The locations of cerebral infarctions in these last two were basal ganglia and corona radiata. In total, cerebral arterial infarctions were found in three cases. The locations of cerebral infarction were: basal ganglia and corona radiata in the two above-mentioned cases, and splenium of corpus callosum in the third. Other neuroimaging findings were leptomeningeal enhancement (5 cases) and hydrocephalus (1 case).
MRA and TCCS correlation
There were four stenotic vessels demonstrated by MRA studies only, including M1 stenosis (<50%) in two and A1 stenosis (A1) in the remaining two. Their TCCS demonstrated mild increased flow velocity that did not fulfil the criteria for stenotic velocities. On the other hand, four vessels were interpreted as MCA stenosis by TCCS study only within the first week of the illness (from day 1 to day 7). In total, only six vessels fulfilled the diagnostic criteria of MCA stenoses in both serial TCCS and MRA studies.
Outcome and prognostic factors
At a minimum 3-month follow-up, 16 cases had good outcomes while the other eight had poor outcomes. Of the eight with poor outcomes, six had severe neurological deficits (BI<12) and two died. Septic shock was the cause of both deaths, at 28 and 36 days of hospitalization, respectively. Mean GCS score differed significantly between the good and poor outcome groups from day 1 to day 14 (p < 0.05).
Potential prognostic factors at 3 months follow-up are listed in Table 3. Under univariate analysis, cerebral infarctions and intracranial arterial stenoses (both OR 5, 95% CI 0.378–66.6, p = 0.249) were risk factors for a poor outcome at 3 months (unadjusted risks). Mean age (p = 0.03), GCS score at the time of admission (p = 0.001), and presence of headache (p = 0.03) and septic shock (p = 0.03) were also significantly different between the good outcome and poor outcome groups. Under multivariate logistic regression, including all variables significant at the univariate level, only GCS at the time of admission (p = 0.008, OR = 1.803, 95%CI 1.17–2.78) was independently associated with outcome at 3 months.
|
Discussion |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionReferences |
|---|
The clinical course of ABM can be complicated by a variety of cerebrovascular insults that may result in global or regional reduction in cerebral blood flow, causing permanent neurological deficits and death.1–3 Cerebrovascular insult is an important explanation of the high incidence of neurological sequelae among the survivors of adult ABM. These ABM-related cerebrovascular insults typically occur within the first 2 weeks of the event, and several studies have focused on their aetiology and pathology using diagnostic measures such as cerebral angiography, cranial Doppler ultrasonography and autopsy.4,6,9,23,24 Of the possible cerebrovascular insults in ABM, inflammatory changes of both large and small intracranial vessels seem to be important in determining outcome.4,6,9,23,24 Three phases of arteriopathy, including vasospasm, vasodilatation and stenosis, have been noted in ABM,23 and several mechanisms have been implicated for their development.24–27 Because of the use of different diagnostic tools and criteria, different causative pathogens and the time course of meningitis, the prevalence of intracranial arterial stenoses during ABM is not easy to study. Some of the variables associated with TCCS, such as the presence/absence of a bone window, the angle of insonation and the experience of the examiner, may also influence the reported prevalence rate of cerebrovascular complications among ABM patients.4,6,9,23,24 There are also several factors that may affect cerebral blood flow, including inflammatory hyperaemia, systemic conditions, Vmean, increased intracranial pressure, arterial CO2, body temperature, mean arterial pressure, the use of mechanical ventilation, and whether patients are sedated during the procedures.30–32
There are several possible explanations for the mismatch between MRA and TCCS findings in several of our patients. First, if only TCCS demonstrated stenotic velocities, with normal MRA findings, it could be due to inflammatory hyperaemia, which is frequent in the first week of illness. Second, if only MRA demonstrated stenosis and TCCS did not reflect the criteria of stenotic velocities, it could be due to (i) mild degree of stenosis (<50% stenosis); (ii) systemic conditions such as presence of septic shock, decreasing cerebral blood flow and velocities, and (iii) the disadvantages of TCCS, such as the absence of a bone window and the angle of insonation. Therefore, if only TCCS is used, without using other methods of measuring cerebral blood flow or vessel calibre to confirm an increased value as a stenosis (as in other studies4–7,19), it may under- or overestimate the incidence of stenosis.
In bacterial meningitis, most cerebrovascular events occur in patients with pneumococcal meningitis,9,19 and the fact that rates of cerebrovascular events in our study are relatively low may be partly due to a low rate of pneumococcal meningitis. In Taiwan, Gram-negative bacilli, especially Klebsiella pneumoniae, are the most prevalent pathogens of adult bacterial meningitis, followed by streptococcal species and staphylococcal species.2,28 The proportion of meningitis cases caused by Streptococcus pneumoniae has decreased in recent years in Taiwan.29
In this study, MRI studies revealed that the striatocapsular region was the area of infarction, which might suggest that the arterial occlusion occurred at the M1 segment of MCA, producing blockage of the orifices of the lenticulostriate arteries. Furthermore, a series of MRA studies revealed that the stenoses can be of slow progression or reversible. Thus our TCCS measurements may point to an obstructive process of a proximal segment of basal cerebral vessels as the predominant pathological mechanism, and the observed decline in the velocities in our patients might reflect a reduction in vessel wall inflammation, secondary to the primary infection process. The maximum haemodynamic disturbance was on day 4, and it took up to 3 weeks before flow velocities returned to normal or initial values. Both MRA and TCCS studies can provide non-invasive evaluation of intracranial arterial lesions, but the prevention and management of cerebral infarctions in adult ABM patients remains controversial. Although the use of heparin in has been suggested in pneumococcal meningitis patients,9 the therapeutic risk in this specific group must be carefully evaluated.
In this study, stenosis as demonstrated by TCCS did not wholly coincide with stenosis as demonstrated by MRA. GCS on admission was predictive of a poor outcome at 3 months. Further large-scale studies are needed to delineate the clinical characteristics and therapeutic influence of cerebrovascular insults among the adult ABM population.
|
References |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionReferences |
|---|
1. Sigurardottir B, Bjornsson OM, Jonsdottir KE, Erlendsdottir H, Gudmundsson S. Acute bacterial meningitis in adults. (1997) A 20-year overview. Arch Intern Med 157 425–30.
2. Lu CH, Chang WN, Chang HW. (2000) Adult bacterial meningitis in southern Taiwan: epidemiologic trend and prognostic factors. J Neurol Sci 182 36–44.[CrossRef][Web of Science][Medline]
3. Pfister HW, Borasio GD, Dirnagl U, Bauer M, Einhäupl KM. (1992) Cerebrovascular complications of bacterial meningitis in adults. Neurology 42 1497–504.
4. Müller M, Merkelbach S, Huss GP, Schimrigk K. (1995) Clinical relevance and frequency of transient stenoses of the middle and anterior cerebral arteries in bacterial meningitis. Stroke 26 1399–403.
5. Ries S, Schminke U, Fassbender K, Daffertshofer M, Steinke W, Hennerici M. (1997) Cerebrovascular involvement in the acute phase of bacterial meningitis. J Neurol 244 51–5.[CrossRef][Web of Science][Medline]
6. Gupta R, Mahapatra AK, Bhatia R. (1995) Serial transcranial Doppler study in meningitis. Acta Neurochir (Wien) 137 74–7.[CrossRef][Medline]
7. Okten A, Ahmetoglu A, Dilber E, Dinc H, Kalyoncu M, Ciftcibais K, Yaris N. (2002) Cranial Doppler ultrasonography as a predictor of neurologic sequelae in infants with bacterial meningitis. Invest Radiol 37 86–90.[CrossRef][Web of Science][Medline]
8. Müller M, Merkelbach S, Hermes M, Schimrigk K. (1996) Transcranial Doppler sonography at the early stage of acute central nervous system infections in adults. Ultrasound Med Biol 22 173–8.[CrossRef][Web of Science][Medline]
9. Haring HP, Rötzer HK, Reindl H, Berek K, Kampfl A, Pfausler B, Schmutzhard E. (1993) Time course of cerebral blood flow velocity in central nervous system infections. A transcranial Doppler sonography study. Arch Neurol 50 98–101.
10. Smith SM, Padayachee S, Modaresi KB, Smithies MN, Bihari DJ. (1998) Cerebral blood flow is proportional to cardiac index in patients with septic shock. J Crit Care 13 104–9.[CrossRef][Web of Science][Medline]
11. Aviv RI, Benseler SM, Silverman ED, Tyrrell PN, Deveber G, Tsang LM, Armstrong D. (2006) MR imaging and angiography of primary CNS vasculitis of childhood. Am J Neuroradiol 27 192–9.
12. Baumgartner RW, Mattle HP, Schroth G. (1999) Assessment of 50% and <50% intracranial stenoses by transcranial color-coded duplex sonography. Stroke 30 87–92.
13. Durand ML, Calderwood SB, Weber DJ, Miller SI, Southwick FS, Caviness VS Jr, Swartz MN. (1993) Acute bacterial meningitis in adults: a review of 493 episodes. N Engl J Med 328 21–8.
14. Hussein AS and Shafran SD. (2000) Acute bacterial meningitis in adults. A 12-year review. Medicine (Baltimore) 79 360–8.[Medline]
15. Arnolds BJ and von Reutern GM. (1986) Transcranial Doppler sonography. Examination technique and normal reference values. Ultrasound Med Biol 12 115–23.[CrossRef][Web of Science][Medline]
16. Sorteberg W, Lindegaard KF, Rootwelt K, Dahl A, Nyberg-Hansen R, Russell D, Nornes H. (1989) Effect of acetazolamide on cerebral artery blood velocity and regional cerebral blood flow in normal subjects. Acta Neurochir (Wien) 97 139–45.[CrossRef][Medline]
17. Czosnyka M, Matta BF, Smielewski P, Kirkpatrick PJ, Pickard JD. (1998) Cerebral perfusion pressure in head-injured patients: a noninvasive assessment using transcranial Doppler ultrasonography. J Neurosurg 88 802–8.[Web of Science][Medline]
18. Tsai NW, Chang HW, Chang WN, Huang CR, Lin TK, Chen SD, Lui CC, Wang KW, Cheng BC, Hung PL, Chang CS, Lu CH. (2005) Prognostic factors and therapeutic outcome of isolated symptomatic middle cerebral artery stenosis. Eur J Neurol 12 519–26.[CrossRef][Web of Science][Medline]
19. Müller M, Merkelbach S, Schimrigk K. (1996) Cerebral hemodynamics in the posterior circulation of patients with bacterial meningitis. Acta Neurol Scand 93 443–9.[Web of Science][Medline]
20. Teasdale G and Jennett B. (1974) Assessment of coma and impaired consciousness. A practical scale. Lancet 2 81–4.[CrossRef][Web of Science][Medline]
21. World Health Organization. (1971) Cerebrovascular Disease, Prevention, Treatment and Rehabilitation. WHO Tech Rep Ser No 469 Geneva.
22. Lan SH, Chang WN, Lu CH, Lui CC, Chang HW. (2001) Cerebral infarction in chronic meningitis: A comparison of tuberculous meningitis and cryptococcal meningitis. Q J Med 94 247–53.
23. Yamashima T, Kashihara K, Ikeda K, Kubota T, Yamamoto S. (1985) Three phases of cerebral arteriopathy in meningitis: vasospasm and vasodilatation followed by organic stenosis. Neurosurgery 16 546–53.[Web of Science][Medline]
24. Davis DO, Dilenge D, Schlaepfer W. (1970) Arterial dilatation in purulent meningitis. Case report. J Neurosurg 32 112–15.[Web of Science][Medline]
25. Leeds NE and Goldberg HI. (1971) Angiographic manifestations in cerebral inflammatory disease. Radiology 98 595–604.[Web of Science][Medline]
26. Igarashi M, Gilmartin RC, Gerald B, Wilburn F, Jabbour JT. (1984) Cerebral arteritis and bacterial meningitis. Arch Neurol 41 531–5.
27. Gado M, Axley J, Appleton DB, Prensky AL. (1974) Angiography in the acute and post-treatment phases of Haemophilus influenzae meningitis. Radiology 110 439–44.[Web of Science][Medline]
28. McMenamin JB and Volpe JJ. (1984) Bacterial meningitis in infancy: effects on intracranial pressure and cerebral blood flow velocity. Neurology 34 500–4.
29. Moller K, Larsen FS, Qvist J, Wandall JH, Knudsen GM, Gjorup IE, Skinhoj P. (2000) Dependency of cerebral blood flow on mean arterial pressure in patients with acute bacterial meningitis. Crit Care Med 28 1027–32.[CrossRef][Web of Science][Medline]
30. Moller K, Skinhoj P, Knudsen GM, Larsen FS. (2000) Effect of short-term hyperventilation on cerebral blood flow autoregulation in patients with acute bacterial meningitis. Stroke 31 1116–22.
31. Tang LM, Chen ST, Hsu WC, Lyu RK. (1999) Adult bacterial meningitis in adults: a hospital-based epidemiological study. Q J Med 92 719–25.
32. Lee LH, Chang WN, Huang CR, Chang CS, Chuang YC, Wang KW, Hung PL, Cheng BC, Chang HW, Chang CJ, Lu CH. (2005) Adult Streptococcus pneumoniae meningitis in Southern Taiwan: epidemiologic trends and prognostic factors. J Clin Neurosci 12 32–5.[CrossRef][Web of Science][Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||