Plethysmometer article

1. Journal of Antimicrobial Chemotherapy (2003) 52, 651–655
DOI: 10.1093/jac/dkg417
Advance Access publication 1 September 2003
651
.. . .. . .. . .. . .. . .. .. . .. . .. . .. . .. . .. .. . .. . .. . .. . .. .. . .. . .. . .. . .. . .. .. . .. . .. . .. . .. .. . .. . .. . .. . .. . .. .. . .. . .. . .. . .. .. . .. . .. . .. . .. . .. .. . .. . .. . .. . .. . .. .. . .. . .. . .. . .. .. . .. . .. . .. . .. . .. .. . .. . .. . .. . .. .. . .. . .. . .. . .. . .. .. . .. . .. . .. . .. .. . .. . .. . .. . .. . .. .. . .. . .. . .. . .. . .. .. . ..
© The British Society for Antimicrobial Chemotherapy 2003; all rights reserved.
Factors influencing the anti-inflammatory effect of dexamethasone
therapy in experimental pneumococcal meningitis
I. Lutsar*, I. R. Friedland, H. S. Jafri, L. Wubbel, A. Ahmed, M. Trujillo, C. C. McCoig
and G. H. McCracken Jr
Department of Pediatrics, University of Texas Southwestern Medical Center, Dallas, TX, USA
Received 11 January 2003; returned 24 February 2003; revised 14 July 2003; accepted 15 July 2003
Dexamethasone (DXM) interferes with the production of tumour necrosis factor-α (TNF-α) and interleukin-1
(IL-1) and can thereby diminish the secondary inflammatory response that follows initiation of antibacterial
therapy. A beneficial effect on the outcome of Haemophilus meningitis in children has been proven, but until
recently the effect of DXM therapy in pneumococcal meningitis was uncertain. The aim of the present study
was to evaluate factors that might influence the modulatory effect of DXM on the antibiotic-induced inflam-
matory response in a rabbit model of pneumococcal meningitis. DXM (1 mg/kg) was given intravenously 30
min before or 1 h after administration of a pneumococcal cell wall extract, or the first dose of ampicillin. In
meningitis induced by cell wall extract, DXM therapy prevented the increase in cerebrospinal fluid (CSF)
leucocyte and lactate concentrations, but only if given 30 min before the cell wall extract. In meningitis
caused by live organisms, initiation of ampicillin therapy resulted in an increase in CSF TNF-α and lactate
concentrations only in animals with initial CSF bacterial concentrations ≥5.6 log10 cfu/mL. In those animals,
DXM therapy prevented significant elevations in CSF TNF-α [medianchange–184pg/mL,–114pg/mLversus
+683 pg/mL with DXM (30 min before or 1 h after ampicillin) versus controls (no DXM), respectively, P = 0.02]
and lactate concentrations [median change –10.6 mmol/L, –1.5 mmol/L versus +14.3 mmol/L with DXM (30
min before or 1 h after ampicillin) versus controls (no DXM), respectively, P = 0.01]. These effects were inde-
pendent of the timing of DXM administration. In this model of experimental pneumococcal meningitis, an
antibiotic-induced secondary inflammatory response in the CSF was demonstrated only in animals with
high initial CSF bacterial concentrations (≥5.6 log10 cfu/mL). These effects were modulated by DXM therapy
whether it was given 30 min before or 1 h after the first dose of ampicillin.
Keywords: animal models, CSF, experimental meningitis, inflammatory response, S. pneumoniae
Introduction
Pneumococcal meningitis remains a significant cause of morbidity
and mortality. It is the host’s own inflammatory response that is
responsible for the central nervous system injury characteristic of the
disease.1This inflammatory response can be exacerbated by antibac-
terial therapy that increases rapidly the release of proinflammatory
pneumococcal cell wall products.2–5
Currently, dexamethasone (DXM) is the only anti-inflammatory
agent that has been documented to improve the outcome of bacterial
meningitis in clinical trials, most convincingly in children with Hae-
mophilus influenzae meningitis.6 DXM inhibits production of TNF-α
and IL-1, reverses development of brain oedema and limits the
increase in cerebrospinal fluid (CSF) lactate and leucocyte concen-
trations.1,7,8 Studies have suggested that in H. influenzae meningitis
the timing of DXM in relation to the first antibiotic injection is critical.
The antibiotic-induced inflammatory response is prevented only if
DXM therapy is given simultaneously with ceftriaxone; DXM given
1 h later failed to modulate the antibiotic-induced inflammatory
response.3 A meta-analysis of 10 clinical trials in children suggested
that the timing of DXM therapy might also be critical in pneumo-
coccal meningitis. DXM therapy prevented the development of
severe hearing loss only when it was given before or at the same time
as the first antibiotic injection (OR = 0.09; 95% CI: 0.00–0.71).6
Recently de Gans & van de Beek9 in a prospective, randomized,
placebo-controlled, multicentre trial demonstrated the beneficial
effect of adjunctive DXM therapy on the outcome of pneumococcal
meningitis in adults.
The present study was conducted to evaluate the modulatory
effect of DXM therapy given 30 min before, compared with 1 h after,
.. . .. .. . .. . .. . .. . .. . .. .. . .. . .. . .. . .. .. . .. . .. . .. . .. . .. .. . .. . .. . .. . .. .. . .. . .. . .. . .. . .. .. . .. . .. . .. . .. .. . .. . .. . .. . .. . .. .. . .. . .. . .. . . . . .. .. . .. . .. . .. . .. .. . .. . .. . .. . .. . .. .. . .. . .. . .. . .. .. . .. . .. . .. . .. . .. .. . .. . .. . .. . .. .. . .. . .. . .. . .. . .. .. . .. . .. . .. . .. . .. .. . .. . .. . .. . .. ..
**Correspondence address. Clinical Development, Pfizer Ltd., Ramsgate Road ,CT13 9NJ, UK. Tel: +44-1304-645173;
Fax:+44-1304-655669; E-mail: irja_lutsar@sandwich.pfizer.com
byguestonFebruary19,2012http://jac.oxfordjournals.org/Downloadedfrom

2. I. Lutsar et al.
652
intracisternal injection of pneumococcal cell walls, or after the first
injection of ampicillin, in experimental meningitis caused by live
pneumococci.
Material and methods
Inocula
Pneumococcal cell wall (1 cm3 of this product was obtained from
2.5 × 109 live organisms) was produced and provided by Dr Elaine
Tuomanen.10 Streptococcus pneumoniae (MIC and MBC of ampicillin,
0.01 mg/L) was originally isolated from a patient with bacterial menin-
gitis. To induce meningitis, 0.2 mL pneumococcal cell wall product, or
104–105 cfu/mL oflive organisms, were inoculated intracisternally.
Meningitis model and treatment
A rabbit meningitis model originally described by Dacey & Sande11
was used. In the first set of experiments, DXM (1 mg/kg) was given
intravenously 30 min before or 1 h after the administration of pneumo-
coccal cell walls to six and seven animals, respectively. In the second set
ofexperiments, 16–18 h afterinoculation ofliveorganisms,animalswere
treated with ampicillin alone (75 mg/kg every 12 h) for 24 h, or with the
combinationofampicillinand intravenousDXM(1mg/kg)given30min
before or 1 h after the first ampicillin dose. Each treatment group con-
sisted of 12–13 animals. No treatment was given to control animals.
CSF sample collection and analysis
CSF was collected directly from the cisterna magna under acepromazine
and ketamine anaesthesia before and 2, 4, 6, 12 and 24 h after start of
therapy. For the first four CSF collections, animals were restrained
under anaesthesia in stereotactic frames. Leucocytes were counted in
a Neubauer haemocytometer. Bacterial concentrations in CSF were
measured by plating undiluted and serial dilutions of CSF on sheep
blood agar and incubating in 5% CO2 at 35°C for 24 h. The lower limit
of detection was 10 cfu/mL. The remaining CSF was centrifuged and
the supernatant stored at –70°C. CSF lactate concentrations were meas-
ured by a photocolorimetric assay (Behring Diagnostics Inc, Milton
Keynes, UK). TNF-α concentrations were measured by cytotoxic assay
using L929 tumorigenic murine fibroblasts.12 The standard curve for
theTNF-α assay was linearfrom 40–2500 pg/mL.
Statistical analysis
Normallydistributeddataarepresentedas mean ±S.D. andnon-normally
distributed data as median and range. Student’s t-test was used for
comparison of parametric data and the Kruskall–Wallis analysis of
variance for non-parametric variables.
Results
Meningitis induced by pneumococcal cell walls
Administration of pneumococcal cell walls resulted in the release of
TNF-α, an influx of leucocytes and increased lactate concentrations
in CSF (Figure 1). The elevation in TNF-α concentrations was pre-
vented by DXM therapy when given 30 min before or 1 h after pneu-
mococcal cell walls. The increase in CSF leucocyte and lactate
concentrations,however,wasinhibitedonlywhenDXMtherapywas
given 30 minbefore the cell wall products.
Meningitis induced by live organisms
The addition of DXM to ampicillin therapy resulted in a lower initial
bacterial killing rate compared with ampicillin therapy alone (0.39 ±
0.1 cfu/mL/h versus 0.57 ± 0.12 cfu/mL/h; P = 0.04). The changes in
CSF inflammatory indices were similar in all study groups and were
not influenced bythe co-administrationofDXM(Figure 2).
For further analysis, animals were divided in two groups based on
CSF bacterialconcentrationsat the start oftherapy(≤5.5log10 or≥5.6
log10 cfu/mL) (Figure 3). After the first dose of ampicillin, animals
with high initial bacterial concentrations demonstrated significantly
greater changes in CSF TNF-α [median ∆+683 pg/mL (quartiles
+246 to +758) versus ∆–16.3 pg/mL (quartiles –6 to –29)], white
bloodcells (WBC)[median ∆+2175 cells/mL (quartiles 437to 6987)
versus ∆–325 cells/mL (quartiles –787 to +25)] and lactate [median
∆+14.3 mmol/L (quartiles –7.6 to –14.6) versus ∆–8.5 mmol/L
(quartiles–5.1to–16.4)] concentrationscomparedwithanimalswith
lowerinitial bacterial concentrations.
Figure 1. Concentrations of WBC (cells/mL; left), TNF-α (pg/mL; middle) and lactate (mmol/L; right) in the CSF. Meningitis was induced by the intracisternal
administration ofpneumococcalcellwalls.Animalsweretreated with DXMgiven 30 min before (open squares)or 1h after (solid squares) administration of cellwalls.
No treatment was given to the control animals (solid triangles). *P = 0.02 versus controls or those treated with DXM 1 h after introduction of cell walls.
byguestonFebruary19,2012http://jac.oxfordjournals.org/Downloadedfrom

3. Dexamethasone in pneumococcal meningitis
653
In those with high initial bacterial concentrations (≥5.6 log10 cfu/
mL), DXM therapy prevented elevations in CSF TNF-α [median
∆–184 pg/mL (quartiles –116 to –258) or ∆–114 pg/mL (quartiles –
39 to –175) versus ∆+683 pg/mL (quartiles 246 to +758) with DXM
given 30 min before or 1 h after ampicillin versus without DXM,
respectively; P = 0.02] and lactate concentrations [median ∆–10.6
mmol/L (quartiles 7.6 to 17.4) or ∆–1.5 mmol/L (quartiles –19.7 to –
1.3) versus ∆+14.3 mmol/L (quartiles –7.6 to –14.6) with DXM
given 30 min before or 1 h after ampicillin and without DXM,
respectively; P = 0.01]. These effects were not significantly different
as a result of the timing of DXM administration, although there was a
trend indicating that changes in TNF-αand lactate values were lower
in animals given DXM before ampicillin therapy (Figure 4). The
changes in leucocyte concentrations were not affected by DXM ther-
apy. In animals with low CSF bacterial concentrations (≤5.5 log10
cfu/mL), there were no differences in TNF-α, WBC or lactate con-
centrations between those treated with or without DXM (data not
shown). There was no correlation between changes in TNF-α, leuco-
cyte and lactate concentrations in CSF, and the degree of bacterial
killing after introduction of ampicillin therapy (r = 0.06; r = 0.47 and
r =0.46, respectively; P> 0.05).
Discussion
In this model of pneumococcal meningitis, we demonstrated that the
antibiotic-induced secondary inflammatory response, as evidenced
Figure 2. Concentrations of WBC (cells/mL), TNF-α (pg/mL), lactate (mmol/
L) and S. pneumoniae (log10 cfu/mL) in CSF in animals with pneumococcal
meningitis. Animals were treated with ampicillin alone (solid triangles), with
DXM given 30 min before (open squares) or 1 h after ampicillin therapy (solid
squares). No treatment was given to the control (open triangles) animals, and
they died after 6–10 h. Error bars demonstrate lower and upper quartile.
*P = 0.04 versus those not receiving DXM therapy.
Figure 4. Concentrations of WBC (cells/mL, left), TNF-α (pg/mL, middle), lactate (mmol/L, right) in CSF in animals with initial CSF bacterial concentrations ≥5.6
log10 cfu/mL. DXM therapy was given 30 min before (open squares) or 1 h after the first dose of ampicillin (solid squares). Control animals (solid triangles) were
treated with ampicillin only. Error bars demonstrate lower and upper quartiles. *P < 0.05 versus those treated with ampicillin and DXM.
Figure 3. Changes in CSF concentrations of WBC (cells/mL), TNF-α (pg/mL),
lactate (mmol/L) and S. pneumoniae (log10 cfu/mL) in pneumococcal meningi-
tis after introduction of ampicillin therapy. The CSF bacterial concentrations at
the introduction of ampicillin therapy were ≤5.5 log10 cfu/mL (open squares) or
≥5.6 log10 cfu/mL (solid squares). Error bars demonstrate lower and upper quar-
tile. *P < 0.05
byguestonFebruary19,2012http://jac.oxfordjournals.org/Downloadedfrom

4. I. Lutsar et al.
654
by an elevation of CSF TNF-α and lactate concentrations, occurred
only in animals with high bacterial concentrations before the start
of ampicillin therapy. The increase in CSF lactate and TNF-α con-
centrations were inhibited by adjunctive DXM therapy regardless
of whether it was given 30 min before or 1 h after the first dose of
ampicillin.
Liberation of free endotoxin or cell-wall components by cell-wall
active antibiotics has been demonstrated in vitro and in experimental
meningitis.3,4,5,13 This is associated with enhanced inflammation in
the subarachnoid space, as evidenced by an increase in leucocyte,
TNF-αand lactate concentrations.3,14 Ourstudyshowed that the anti-
biotic-induced inflammatory burst did not occur in all animals and
was seen only in those with greater CSF bacterial concentrations
(≥5.6 log10 cfu/mL). Although this has been intimated previously,
data were not provided.14 In experimental meningitis, the magnitude
oftheinflammatoryresponseinCSFdepends ontheconcentrationof
inoculated cell walls10 and thus it is not surprising that animals with
higher bacterial counts demonstrate a more pronounced host inflam-
matory response after antibiotic therapy.
There is concern that rapid bacterial killing by antibiotics could
result in an enhanced inflammatory response and a worsening of the
clinical outcome in meningitis.15,16 The results of this and previous
studies, however, do not support these speculations. On the contrary,
clinical and experimental studies have demonstrated that rapid effec-
tive bactericidal therapy protects against the development of deaf-
ness and other neurological disabilities.17,18 Moreover, this and
previous studies4 found no correlation between CSF inflammatory
markers and the magnitude of bacterial killing. Uncontrolled bac-
terial growth eventually results in a much greater release of cell wall
components—and consequent enhancement of the inflammatory
response—than that induced by antibacterial therapy.4,13 These find-
ings suggest that early antibacterial therapy and rapid clearance of
bacteria from CSF outweigh the potential adverse effects caused by
the antibiotic-induced inflammatoryburst.
DXM, as an adjunct to antibacterial therapy in bacterial meningi-
tis, has been shown to be beneficial in H. influenzae meningitis in
children.6 In pneumococcal meningitis, however, its modulatory
effect was uncertain because of the relatively small numbers of chil-
dren enrolled in prospective controlled trials.19 In adults with pneu-
mococcal meningitis, an unfavourable outcome was seen in 26% of
patientsreceivingadjunctiveDXMcomparedwith52%amongthose
receiving placebo.9 The results of our study clearly supported previ-
ous findingsbyTuomanenet al.8 and showed that,similarto H. influ-
enzae meningitis, DXM therapy prevents the antibiotic-induced
release ofTNF-αand lactateconcentrationsinCSF inpneumococcal
meningitis.3
Early institution of DXM therapy has been suggested because of
its delayed onset of action.20 In our study, the timing of DXM
appeared to be critical only in meningitis induced by the pneumo-
coccal cell wall. Inmeningitis induced bylive microorganisms, how-
ever, once CSF inflammation was established, administration of
DXM before the dose of ampicillin did not appear to have a signifi-
cantlygreater salutaryeffectthan1hafter ampicillin;DXM adminis-
tration that was delayed further was not assessed. These results
contrast with those obtained in experimental H. influenzae meningi-
tis, where the antibiotic-induced inflammatory response was modu-
lated only if DXM was given before or simultaneously with
ceftriaxonetherapy.3 The effectivenessofearly DXM administration
was also highlighted ina recent meta-analysis of 10 randomized con-
trolled clinical trials.6
At the time of diagnosis, ∼40% of patients with pneumococcal
meningitishave CSFbacterial concentrations<106 cfu/mL21,andour
findings suggest that antibiotic-inducedenhancedinflammationmay
not occur in such patients. Whether these patients benefit from DXM
therapy, and how they can be identified at diagnosis, requires further
clarification. Detection of bacteria in Gram-stained specimens of
CSF maybeuseful inidentifyingpatientswithlarge bacterialloads.22
One possible detrimental effect of DXM therapy, also demon-
strated in this study, is decreased bacterial clearance from CSF.23,24
DXMdecreasesthepermeabilityoftheblood–brainbarrier,resulting
in decreased concentrations of hydrophilic antibiotics in the CSF.25
Also,highconcentrationsofDXM(400 µg/mL)inhibitphagocytosis
by CSF leucocytes.26The clinical relevance of these findings, how-
ever, has not been demonstrated.6,9
In this model of pneumococcal meningitis, CSF bacterial concen-
trations at the start of therapy appeared to be more important than the
timing ofDXMtherapy in influencing the antibiotic-induced inflam-
matory response. It is likely that there is a time beyond which DXM
loses its effectiveness, but this point has not been clearly defined.
Acknowledgements
I. Lutsar was a recipient of a fellowship award from the European
Society for Paediatric Infectious Diseases, supported by Lederle-
PraxisBiologicals.Partofthisstudywaspresentedatthe1997annual
meeting of Infectious Diseases Society of America (IDSA), San
Francisco. The study has followed animal experimentation guide-
lines and was approved by the Institutional Animal Care and
Research Advisory Committee of the University ofTexas.
References
1. Spellerberg, B. & Tuomanen, E. I. (1994). Pathophysiology of
pneumococcal meningitis. Annals of Medicine 26, 411–18.
2. Nau, R., Zysk, G., Schmidt, H. et al. (1997). Trovafloxacin delays
the antibiotic-induced inflammatory response in experimental pneumo-
coccal meningitis. Journal of Antimicrobial Chemotherapy 39, 781–8.
3. Mustafa, M. M., Ramilo, O., Mertsola, J. et al. (1989). Modulation of
inflammation and cachectin activity in relation to treatment of experimen-
tal Haemophilus influenzae type b meningitis. Journal of Infectious
Diseases 160, 818–25.
4. Friedland, I. R., Jafari, H., Ehrett, S. et al. (1993). Comparison of
endotoxin release by different antimicrobial agents and the effect on
inflammation in experimental E. coli meningitis. Journal of Infectious
Diseases 168, 657–62.
5. Fischer, H. & Tomasz, A. (1984). Production and release of pepti-
doglycan and wall teichoic acid polymers in pneumococci treated with
beta lactam antibiotics. Journal of Bacteriology 157, 507–13.
6. McIntyre, P. B., Barkey, C. S., King, S. M. et al. (1997). Dexame-
thasone as adjunctive therapy in bacterial meningitis: a meta-analysis of
randomized clinical trials since 1988. Journal of the American Medical
Association 278, 925–31.
7. Täuber, M. G., Khayam-Bashi, H. & Sande, M. A. (1985). Effects
of ampicillin and corticosteroids on brain water content, cerebrospinal
fluid pressure, and cerebrospinal fluid lactate levels in experimental
pneumococcal meningitis. Journal of Infectious Diseases 151, 528–34.
8. Tuomanen, E., Hengstler, B., Rich, R. et al. (1987). Nonsteroidal
anti-inflammatory agents in the therapy for experimental pneumococcal
meningitis. Journal of Infectious Diseases 155, 985–90.
9. de Gans, J. & van de Beek, D. (2002). Dexamethasone in adults
with bacterial meningitis. New England Journal of Medicine 347, 1549–56.
10. Tuomanen, E., Liu, H., Hengstler, B. et al. (1985). The induction of
meningeal inflammation by components of the pneumococcal cell wall.
Journal of Infectious Diseases 151, 859–68.
byguestonFebruary19,2012http://jac.oxfordjournals.org/Downloadedfrom

5. Dexamethasone in pneumococcal meningitis
655
11. Dacey, R. G. & Sande, M. A. (1974). Effect of probenecid on
cerebrospinal fluid concentration of penicillin and cephalosporin derivates.
Antimicrobial Agents and Chemotherapy 6, 437–41.
12. Flick, D. A. & Gifford, G. E. (1984). Comparison of in vitro cell cyto-
toxic assays for tumor necrosis factor. Journal of Immunological Methods
68, 167–75.
13. Stuertz, K., Schmidt, H., Eiffert, H. et al. (1998). Differential
release of lipoteichoic and teichoic acids from Streptococcus pneumo-
niae as a result of exposure to β-lactam antibiotics, rifamycins, trova-
floxacin, and quinupristin–dalfopristin. Antimicrobial Agents and
Chemotherapy 42, 277–81.
14. Pfister, H.-W., Fontana, A., Täuber, M. G. et al. (1994). Mecha-
nism of brain injury in bacterial meningitis: workshop summary. Clinical
Infectious Diseases 19, 463–79.
15. Tuomanen, E. I., Austrian, R. & Masure, H. R. (1995). Pathogene-
sis of pneumococcal infection. New England Journal of Medicine 332,
1280–4.
16. Täuber, M. G., Shibl, A. M., Hackbarth, C. J. et al. (1987). Anti-
biotic therapy, endotoxin concentration in cerebrospinal fluid, and brain
edema in experimental Escherichia coli meningitis in rabbits. Journal of
Infectious Diseases 156, 456–62.
17. Winter, A. J., Comis, S. D., Osborne, M. P. et al. (1998). Ototoxicity
resulting from intracochlear perfusion of Streptococcus pneumoniae in
the guinea pig is modified by cefotaxime or amoxycillin pretreatment.
Journal of Infection 36, 73–7.
18. Lebel, M. H. & McCracken, G. H., Jr (1989). Delayed cerebro-
spinal fluid sterilization and adverse outcome of bacterial meningitis in
infants and children. Pediatrics 83, 161–7.
19. Arditi, M., Mason, E. O., Bradley, J. S. et al. (1998). Three-year
multicenter surveillance of pneumococcal meningitis in children: clinical
characteristics, and outcome related to penicillin susceptibility and dexa-
methasone use. Pediatrics 102, 1087–97.
20. Nakamura, H., Mizushima, Y., Seto, Y. et al. (1985). Dexametha-
sone fails to produce antipyretic and analgesic actions in experimental
animals. Agents and Actions 16, 542–7.
21. Bingen, E., Lambert-Zechovsky, N., Mariani-Kurkdjian, P. et al.
(1990). Bacterial counts in cerebrospinal fluid of children with meningitis.
European Journal of Clinical Microbiology and Infectious Diseases 9,
278–81.
22. Feldman, W. E., Ginsburg, C. M., McCracken, G. H., Jr et al.
(1982). Relation of concentrations of Haemophilus influenzae type b in
cerebrospinal fluid to late sequelae of patients with meningitis. Journal of
Pediatrics 100, 209–12.
23. Schaad, U. B., Kaplan, S. L. & McCracken, G. H. (1994). Steroid
therapy for bacterial meningitis. Clinical Infectious Diseases 20, 685–90.
24. Cabellos, C., Martinez-Lacasa, J., Tubau, F. et al. (2000). Evalua-
tion of combined ceftriaxone and dexamethasone therapy in experimental
cephalosporin-resistant pneumococcal meningitis. Journal of Antimicrobial
Chemotherapy 45, 315–20.
25. Lutsar, I., McCracken, G. H. & Friedland, I. R. (1998). Antibiotic
pharmacodynamics in cerebrospinal fluid. Clinical Infectious Diseases
27, 1117–29.
26. Weitbrecht, W. U. (1980). Influence of milieu and dexamethasone
on in vitro phagocytosis of cerebrospinal fluid phagocytes. Pathological
Research and Practice 167, 393–9.
byguestonFebruary19,2012http://jac.oxfordjournals.org/Downloadedfrom