CLINICAL MICROBIOLOGY REVIEWS, Apr. 1993, p. 118-136 Vol. 6, No.

2
0893-8512/93/020118-19$02.00/0
Copyright © 1993, American Society for Microbiology

Pathogenesis and Pathophysiology of Bacterial Meningitis

ALLAN R. TUNKELl* AND W. MICHAEL SCHELD2
Department of Internal Medicine (Infectious Diseases), Medical College of Pennsylvania, Philadelphia,
Pennsylvania 19129,1 and Departments of Internal Medicine (Infectious Diseases) and Neurosurgety,
University of Virginia Health Sciences Center, Charlottesville, Virginia 229082
INTRODUCTION ...................................................... 118
Epidemiology ...................................................... 118
Animal Models ...................................................... 119
MUCOSAL COLONIZATION AND SYSTEMIC INVASION ...................................................... 119
Fimbriae ...................................................... 119
Polysaccharide Capsule ...................................................... 120
Antibodies ...................................................... 120
Other Bacterial Components ...................................................... 121
BACTEREMIA ...................................................... 121
Polysaccharide Capsule ...................................................... 121
Host Defense Mechanisms ...................................................... 121
MENINGEAL INVASION ...................................................... 121
Site of Invasion ...................................................... 121
Fimbriae ...................................................... 122
Monocytes ...................................................... 122
Other Bacterial Components ...................................................... 122
Secondary Bacteremia ...................................................... 122
ALTERATIONS OF THE BBB ...................................................... 122
LPS ...................................................... 123
Cytokines ...................................................... 123
Localization of BBB Injury ...................................................... 123
In Vitro Model of the BBB ...................................................... 124
BACTERIAL SURVIVAL WITHIN THE SUBARACHNOID SPACE .............................................. 124
CSF Complement ...................................................... 124
CSF Antibody ...................................................... 125
CSF Leukocytes ...................................................... 125
INDUCTION OF SUBARACHNOID SPACE INFLAMMATION .................................................... 125
Cell Wall ..................................................... 126
LPS ..................................................... 126
Inflammatory Mediators ..................................................... 126
INCREASED INTRACRANIA L PRESSURE ..................................................... 127
CEREBRAL VASCULITIS ..................................................... 128
ALTERATIONS IN CEREBRAL BLOOD FLOW ..................................................... 128
ADJUNCTIVE THERAPEUTIC STRATEGIES ..................................................... 129
Experimental Studies ...................................................... 129
Clinical Trials ..................................................... 131
REFERENCES ..................................................... 132

INTRODUCTION during 1986 were lower (e.g., 19% for meningitis due to S.
pneumoniae) (178), suggesting that improvements in early
Epidemiology detection and antibiotic treatment may have occurred in the
1980s. Bacterial meningitis also remains a significant prob-
Bacterial meningitis remains a relatively common and lem in other parts of the world. A recent review of all cases
devastating disease with an overall annual attack rate in the of bacterial meningitis admitted to an isolation-fever hospital
United States of -3.0 cases per 100,000 population. Mortal- in Salvador, Brazil, for the decade 1973 through 1982 re-
ity rates associated with the three most common causative vealed an approximate annual incidence of 45.8 cases per
agents of bacterial meningitis, Haemophilus influenzae, 100,000 population and an overall mortality rate of 33% (24).
Neisseria meningitidis, and Streptococcus pneumoniae, The three common meningeal pathogens (H. influenzae, N.
were 6.0, 10.3, and 26.3%, respectively, in the United StatesS.
1981 (136). Casese fatality rates mengtdsadS.ppneumonae) accounted for 72% offal all
es in a
from 1978 through uoie)conedfr7%
from1
subsequent study
h
ofh
of fe Can fanglt
s s and Los
five states Angeles Cunty
County cases and 70% of the deaths. In addition to this unacceptable
mortality, there is a high rate of neurologic sequelae in
children and adults who survive their episodes of bacterial
* Corresponding author. meningitis (17, 33, 107, 109, 156). This stable but unsatisfac-
118
VOL. 6, 1993
tory situation indicates the ongoing need to study the patho-
genesis and pathophysiology of bacterial meningitis in an
attempt to improve the response to conventional antimicro-
bial therapy (126, 148, 165, 166).
Animal Models
Experimental animal models have been employed exten-
sively during the past 2 decades to increase our understanding
of the pathogenesis and pathophysiology of bacterial menin-
gitis (163). In the most commonly employed infant rat model,
animals developed meningitis following intranasal challenge
with H. influenzae type b (84). This model most closely
simulates the presumed pathogenesis of H. influenzae menin-
gitis in humans, since there was an initial nasopharyngeal
focus followed by bacteremia and an age-dependent suscep-
tibility to meningeal invasion (83). The incidence of bacterial
meningitis, irrespective of rat host age, was directly related to
the intensity of bacteremia. Infant rats also developed men-
ingitis after orogastric challenge with Eschenchia coli (82).
The infant rat model of H. influenzae meningitis has been
used primarily to study the early pathogenic events of bacte-
rial meningitis. These include the determinants of nasopha-
ryngeal colonization and translocation into the bloodstream,
intravascular survival of the organism, bacteremia, and the
mechanisms of central nervous system (CNS) invasion. The
animal's small size is a disadvantage of the infant rat model
because only small samples (7 to 25 RI) of cerebrospinal fluid
(CSF) are obtainable from 5- to 10-day-old rats, precluding
frequent sampling of CSF. Therefore, this model is less
suitable for study of the pathophysiologic consequences of
bacterial meningitis, since frequent sampling of CSF is usu-
ally required for studying these events. Infant primates have
also been used in the study of the pathogenesis of bacterial
meningitis. In one study (130), bacteremia and meningitis
developed in 89 and 94% of animals, respectively, following
the atraumatic intranasal inoculation of H. influenzae type b.
Although this model closely simulates the spectrum of inva-
sive H. influenzae disease seen in humans, it is employed
infrequently because of expense.
Experimental models of meningitis in adult rabbits or rats
rely on the direct intracisternal inoculation of bacteria for
initiation of infection (30, 110). These animals reliably develop
lethal infections with a predictable time course, although the
natural bacteremia-meningitis sequence is bypassed, thereby
creating an artificial pathogenesis. However, these models
have been extremely useful for the study of the pathophysi-
ologic consequences of bacterial meningitis after organisms
have reached the subarachnoid space. CSF samples from
these adult animals may be obtained often.
The following sections briefly review the pathogenesis and
pathophysiology of bacterial meningitis, as depicted in Fig.
1, emphasizing the relationships between specific bacterial
virulence factors and host defense mechanisms that are
responsible for clinical expression of disease (126, 165, 166).
New areas of investigation, as suggested by several of the
steps depicted in Fig. 1, may lead to improvements in the
refractory mortality and unacceptable morbidity in patients
with bacterial meningitis.
MUCOSAL COLONIZATION AND SYSTEMIC
INVASION
Initiation of most cases of bacterial meningitis begins with
host acquisition of a new organism by nasopharyngeal
colonization. Following infection of human nasopharyngeal
PATHOGENESIS OF BACTERIAL MENINGITIS 119
NASOPHARYNGEAL COLONIZATION
LOCAL INVASION
BACTEREMIA
Z ENDOTHEUAL CELL INJURY
INCREASED BBB
PERMEABIUTY MENINGEAL INVASION
- CEREBRAL
SUBARACHNOID SPACE INFLAMMATION M VASCULJTIS
INCREASED CSF OUTFLOW RESISTANCE
HYDROCEPHALUS
VASOGENIC EDEMA INTERSTITIAL EDEMA CYTCDTOXIC EDEMbJA |
INCREASED INTRACRANIAL PRESSURE
CEREBRAL
INFARCTION
I/
DECREASED CEREBRAL BLOOD FLOW
FIG. 1. Scheme depicting the pathogenesis and pathophysiology
of bacterial meningitis.
cells in organ culture in vitro with meningococci or H.
influenzae type b, several events are observed (143): (i)
association with mucus independent of capsular polysaccha-
ride, fimbriae, or immunoglobulin A subclass 1 (IgAl) pro-
tease production; (ii) cytotoxicity characterized by break-
down of epithelial-cell tight junctions, sloughing of ciliated
cells, and ciliostasis; (iii) selective attachment to nonciliated
epithelial cells; (iv) multiplication and formation of micro-
colonies on the epithelial surface; (v) invasion of the epithe-
lium by intracellular or intercellular routes; and (vi) passage
of organisms to the submucosa. These events require viable
organisms and are not completed by commensal species. In
addition, meningococci and H. influenzae type b appear to
invade nasopharyngeal mucosa by different mechanisms:
meningococci utilize parasite-directed endocytosis, and H.
influenzae type b adheres to cells, causing a breakdown in
tight junctions between epithelial cells and leading to inva-
sion by an intercellular mechanism. The bacterial virulence
factors and the host defense mechanisms responsible for
these events are discussed in detail below.
Fimbriae
Many of the major meningeal pathogens possess surface
characteristics that enhance mucosal colonization. Fim-
briae, or pili, are specific organelles found on many bacteria
and often mediate adhesion of bacteria to host cells (10). The
fimbriae of N. meningitidis mediate adherence of the organ-
ism to nasopharyngeal epithelial cells (Fig. 2) (143, 145).
Meningococci, like gonococci, possess fimbriae that differ in
their morphologic, antigenic, and binding properties (51).
Fimbriae appear morphologically as aggregated bundles or
single filaments. The aggregated bundles are found primarily
among disease isolates and exhibit a low degree of adherence
to human buccal epithelial cells, whereas the single filaments
are found predominantly among colonizing isolates with
medium to high adherence characteristics. These fimbriated
strains account for 80% of primary meningococcal isolates
from nasopharyngeal carriers and from the CSF of patients
120 TUNKEL AND SCHELD
as -- mm vw
FIG. 2. Scanning electron micrograph of nasopharyngeal organ
cultures 6 h after infection with encapsulated N. meningitidis 34NP,
showing attachment of meningococci to nonciliated epithelial cells
and formation of a microcolony on the epithelial surface (magnifi-
cation, x8,775). Reproduced from reference 143 with permission of
The University of Chicago Press.
with meningitis (32), although all fimbriae were lost on serial
subculture in the laboratory. Following attachment via a
specific cell surface receptor, meningococci are transported
within a phagocytic vacuole across nonciliated nasopharyn-
geal columnar epithelial cells (74, 144); this series of events
appears to be essential for development of invasive menin-
gococcal disease. Fimbriae have also been implicated in the
adherence of H. influenzae to upper respiratory tract epithe-
lial cells (72, 146), and lack of fimbrial expression impairs the
ability of H. influenzae type b to colonize the nasopharynx
(177). However, fimbriae have not been found on H. influ-
enzae type b isolated from the CSF or blood of patients with
invasive disease (72, 106), suggesting that although fimbriae
play a role in initial adherence within the nasopharynx, their
presence is not necessary for the organism to cause menin-
gitis. Acquisition and colonization of H. influenzae type b
may also be promoted following viral infection by a variety
of respiratory viruses, including influenza A Victoria and
respiratory syncytial virus (140). The precise role of a
preceding upper respiratory viral infection in the enhance-
ment of nasopharyngeal colonization and the subsequent
development of meningitis is controversial, however, and
requires further study.
Polysaccharide Capsule
Bacterial encapsulation may be important for nasopharyn-
geal colonization and systemic invasion of meningeal patho-
gens. Strains of H. influenzae colonize the nasopharynges of
most children by the age of 3 months, although most of these
strains are unencapsulated (140). Among the six encapsu-
CLIN. MICROBIOL. REV.
lated types of H. influenzae (a through f) type b strains
constitute less than 5% of nasopharyngeal isolates, although
more than 95% of meningeal and systemic infections are
caused by type b strains. Experimental studies with infant
rats and laboratory transformants selected by capsular type
have shown that, while all encapsulated strains of H. influ-
enzae have the potential for systemic invasion after intrap-
eritoneal inoculation, type b strains are the most virulent and
are the only capsular types capable of systemic invasion
following intranasal inoculation (85, 120). Indeed, the pres-
ence of serum antibodies to polyribosyl-ribitol phosphate,
the capsular polysaccharide of type b isolates, is protective
against invasive disease (2). Antibodies to type b capsule are
almost uniformly detectable in humans by the age of 4 years,
even in the absence of known exposure to H. influenzae type
b. The presence of antibodies to type b capsule may be
related to the abilities of other encapsulated strains of H.
influenzae to produce some type b capsular material. It has
also been suggested that the source of bacterial capsule may
be secondary to acquisition of new DNA from organisms
colonizing the gastrointestinal tract (81, 138). For example,
E. coli K100 possesses a capsule that is immunologically
related to the type b capsule of H. influenzae and may
stimulate the production of cross-reacting anticapsular anti-
bodies by the host, suggesting that the protection against
infection with H. influenzae type b is due to priming of serum
anticapsular antibodies.
Polysaccharide capsule may also be an important viru-
lence factor for invasive disease caused by S. pneumoniae.
Of the 84 pneumococcal serotypes known, 18 are responsi-
ble for 82% of cases of bacteremic pneumococcal disease (9,
39), and there is a close correlation between bacteremic
subtypes and those implicated in meningitis (20, 44, 50).
However, low degrees of adherence to human pharyngeal
epithelial cells have been observed among pneumococcal
strains isolated from patients with serious infections such as
septicemia and meningitis (4), suggesting that adherence
may be less important in pneumococcal pathogenicity. A
study of S. pneumoniae adherence to nasopharyngeal cells
demonstrated that these organisms were more often found
on desquamated cells than on cells taken from intact epithe-
lium (71). Nasopharyngeal mucus may provide a protected
nidus from which pneumococci can spread (45), although the
various factors responsible for invasiveness among certain
pneumococcal serotypes remain unknown.
Antibodies
Natural antibodies such as IgA, found predominantly in
mucosal secretions, may inhibit the adherence of microor-
ganisms to mucosal surfaces. Formation of these antibodies
is stimulated by colonization of organisms that share cross-
reactive antigens with pathogenic strains (138). However,
the presence of high concentrations of circulating IgA anti-
bodies to N. meningitidis may paradoxically permit the
development or exacerbate the progression of invasive dis-
ease by preferentially binding to the organism, thereby
blocking the beneficial effects of IgG and IgM antibodies (53,
54, 64). In addition, many pathogenic Neisseria, Haemoph-
ilus, and Streptococcus species produce IgAl proteases that
cleave IgA in the hinge region of the immunoglobulin mole-
cule (108); these enzymes may have a pathogenic role by
facilitating adherence of bacterial strains to mucosal surfaces
through local destruction of IgA (86). The exact role of IgA
protease production in the pathogenic sequelae of bacterial
meningitis remains unclear, however.
VOL. 6, 1993
Other Bacterial Components
Other studies have examined various surface components
(e.g., lipopolysaccharide [LPS] and outer membrane pro-
teins) of H. influenzae to determine their roles in the entry
stages of pathogenesis. Antibodies against these components
confer protection against repeated challenge with the organ-
ism (3, 55, 56, 65). More recently, hemocin, the bacteriocin
produced by H. influenzae, has been shown to be strongly
associated with type b encapsulated strains and may play a
role in host nasopharyngeal colonization and/or systemic
invasion by this organism. After intranasal inoculation of
infant rats with an equal mixture of a non-hemocin-produc-
ing strain and its hemocin-producing transformant, organ-
isms capable of hemocin production predominated in iso-
lates from both nasopharyngeal and blood cultures (70).
However, the precise roles of capsular polysaccharide, LPS,
hemocin, outer membrane proteins, and other surface com-
ponents in mucosal translocation and bloodstream invasion
of H. influenzae and other major meningeal pathogens re-
main incompletely defined.
BACTEREMIA
Polysaccharide Capsule
Once the mucosal barrier is crossed, bacteria gain access
to the bloodstream and must then overcome host defense
mechanisms to survive and invade the CNS. Surface encap-
sulation is the most important virulence factor in this regard.
Bacterial capsule, by effectively inhibiting neutrophil phago-
cytosis and resisting classic complement-mediated bacteri-
cidal activity, may enhance bloodstream survival of the
organism and facilitate intravascular replication (166). In-
deed, the most common meningeal pathogens (H. influen-
zae, N. meningitidis, S. pneumoniae, E. coli, and Strepto-
coccus agalactiae) are all encapsulated. Furthermore,
certain specific capsular types among the great many re-
quired are associated disproportionately with the develop-
ment of bacterial meningitis. For example, about 84% of
cases of neonatal meningitis due to E. coli are caused by
strains bearing the Kl antigen, which is antigenically related
to the capsular material of serogroup B meningococci and
type III group B streptococci (119). In the absence of
Kl-specific host antibody, these organisms are profoundly
resistant to phagocytosis (28). Human monoclonal antibod-
ies with specific reactivity for epitopes on the Kl capsule of
E. coli and/or the group B polysaccharide of N. meningitidis
may be useful for prevention and/or treatment of blood-
stream infections caused by these organisms (114); however,
this concept remains conjectural for humans.
Host Defense Mechanisms
Several host defense mechanisms can counteract the
antiphagocytic effects of bacterial capsule. For example, the
capsular polysaccharides of S. pneumoniae activate the
alternative complement pathway, resulting in cleavage of C3
and subsequent attachment of C3b to the bacterial surface,
thereby facilitating opsonization, phagocytosis, and intra-
vascular clearance of the organism (38). Although antipneu-
mococcal cell wall antibody and antipneumococcal capsular
antibody promote the efficient deposition of C3b on the
pneumococcal surface, C3b deposited on the surface of
pneumococcal capsule is a more efficient opsonin in vitro
and in vivo than C3b activated by anti-cell wall antibody (22,
PATHOGENESIS OF BACTERIAL MENINGITIS 121
23). Impairment of the alternative complement pathway, as
in patients with sickle cell disease (103) and in patients who
have undergone splenectomy, predisposes to the develop-
ment of pneumococcal meningitis. The complement cascade
is also activated by H. influenzae type b (113). C3-depleted
rats show a greater incidence and magnitude of bacteremia
after either intravenous or intraperitoneal challenge with H.
influenzae of various serotypes (a, b, c, or d) than do normal
rats (27, 29, 187). Although the incidence of bacteremia due
to type b organisms increases from 63 to 95% in comple-
ment-depleted rats, the incidence and severity of meningitis
are unaffected.
Activation of the complement system is an essential host
defense mechanism in protection against invasive disease by
N. meningitidis. Patients with deficiencies in the terminal
complement components (C5, C6, C7, C8, and perhaps C9),
the so-called membrane attack complex, are particularly
prone to infection with neisserial species, including N.
meningitidis, though usually with a favorable outcome for
the patient when appropriate treatment is instituted (122).
The reasons for the decreased mortality rate in the comple-
ment-deficient patients are not clear. It has been suggested
that the presence of complement-activating products (i.e.,
the membrane attack complex), in concert with other medi-
ators, may contribute to the development of multiorgan
failure and death. A qualitative relationship exists among the
level of circulating meningococcal LPS, a fatal outcome, and
the degree of complement activation (19), indicating that
prognosis is worse for patients with intact complement
systems.
MENINGEAL INVASION
The mechanisms by which bacterial pathogens gain access
to the CNS are largely unknown. One factor may relate to
the concentration of organisms in the blood (84). In the
experimental infant rat model, the intranasal inoculation of
H. influenzae type b initially produced a low-grade bactere-
mia (about 102 CFU/ml) and no organisms were present in
the CSF (141). Culture-positive meningitis was observed
only after an intense bacteremia (>10W CFU/ml) had been
present for at least 6 h (100). Meningitis was also induced in
an age-dependent manner, with a higher incidence in 5-day-
old than in 20-day-old rats (83). In the animals that ultimately
developed meningitis, sustained bacteremia, as opposed to
transient bacteremia, was documented. However, sustained
bacteremia is not the only factor responsible for meningeal
invasion, because many other organisms (e.g., viridans
streptococci) that produce continuous bacteremia during
infective endocarditis rarely produce bacterial meningitis.
Site of Invasion
The exact sites of CNS invasion by meningeal pathogens
are unclear. Early studies with the experimental rat model
suggested that the route of invasion from the bloodstream to
the CSF was through the dural venous sinus system (141).
However, subsequent experiments with the same animal
model suggested that during the ensuing bacteremia a non-
specific, sterile, focal inflammation above the cribriform
plate facilitated invasion of the CNS at that site. Further
studies of infant rats and primates demonstrated that bacte-
ria may enter the CSF via the choroid plexus, which has an
exceptionally high rate of blood flow (-200 ml/g/min), per-
haps because more bacterial organisms are delivered to this
site than to other anatomic locations in the CNS per unit of
122 TUNKEL AND SCHELD
time. Sampling of CSF compartments early during bacterial
meningitis has demonstrated higher bacterial densities in the
lateral ventricles than in the cisterna magna, lumbar sub-
arachnoid space, or supracortical subarachnoid space. Al-
though with time equilibrium is reached in these other
locations, the data suggest initial bacterial entry into CSF in
the lateral ventricles, presumably through the choroid plexi.
Fimbriae
Recent experimental studies have suggested that receptors
for some meningeal pathogens are present on cells in the
choroid plexus and/or cerebral capillaries, which may facil-
itate movement of these pathogens into the subarachnoid
space. In cryostat sections of infant rat brain cortical slices,
strains of E. coli possessing S fimbriae are seen to bind to the
luminal surfaces of the vascular endothelium and the epithe-
lium lining the choroid plexus and brain ventricles (101).
Pretreatment of the brain sections with neuraminidase or the
trisaccharide receptor analog of S fimbriae abolished this
binding. Subsequent experiments demonstrated that 1 h after
intraperitoneal challenge with the S-fimbriated strain of E.
coli, about 50% of organisms in the CSF were S fimbriated
and 50% were nonfimbriated (128), suggesting that phase
variation to the nonfimbriated form may be necessary for
these bacteria to invade the CNS. Nevertheless, the specific
adherence of meningeal pathogens to sites within the CNS is
one hypothesis raised to explain bacterial neurotropism and
currently is an area of intense investigation.
Monocytes
Additional studies to determine the pathogenic mecha-
nisms responsible for meningeal invasion utilized histologic
and scanning microscopic techniques to examine the
neuraxes of pigs inoculated intravenously with a pathogenic
strain of Streptococcus suis type 2 (180). The animals were
sacrificed 17 to 47 h after intravenous challenge, and the only
pathologic lesions detected were associated with the choroid
plexus, manifested as disruption of the plexus brush border,
decrease in the number of Kolmer cells, and exudation of
fibrin and inflammatory cells into the ventricles. Intracellular
bacteria were demonstrated in the parenchyma of the cho-
roid plexus, in the ventricular monocytes, and within periph-
eral blood monocytes. Circulating monocytes, which were
also found to contain phagocytized bacterium-sized parti-
cles, migrated into the CSF via the choroid plexus, suggest-
ing that bacteria may gain access to the CSF in association
with monocytes migrating along normal pathways (the so-
called Trojan horse hypothesis for CNS invasion).
Other Bacterial Components
Other bacterial virulence factors have been studied to
determine their possible roles in meningeal invasion. The
liberation of LPS from N. meningitidis may contribute to the
pathogenicity of this organism in invasive infections (1).
Meningococci vary in their ability to liberate endotoxin, with
increased amounts liberated from patients with invasive
disease. For example, serogroup B meningococcal strains
released slightly more free endotoxin when isolated from
blood or CSF than when isolated from the nasopharynges of
presumably healthy persons. Outer membrane proteins may
also be important. One report has suggested that H. influen-
zae strains with outer membrane protein subtype lc cause
more episodes of meningitis and fewer episodes of epiglot-
CLIN. MICROBIOL. REV.
titis than do strains of subtype 1 (151), perhaps because of
the ability of each subtype to release LPS under appropriate
circumstances. However, the precise role of outer mem-
brane protein subtypes in meningeal invasion is unclear.
Secondary Bacteremia
Following bacterial invasion of the subarachnoid space, a
secondary bacteremia may result from the local CNS sup-
purative process, allowing the meningeal pathogen to con-
tinuously enter and leave the CSF compartment under quite
dynamic circumstances. In an experimental canine model of
pneumococcal meningitis, the early transport of bacteria
from CSF to blood was presumed to be transendothelial
through arachnoid villi containing pores large enough to
accommodate bacteria, which would then enter the superior
sagittal sinus and return to the central venous blood (135).
This transport occurred only following active bacterial mul-
tiplication in CSF and before the height of the febrile
response or CSF pleocytosis. This phenomenon was also
observed with H. influenzae type b, inoculation of which
into the cisterna magna of experimental animals produced an
almost instantaneous bacteremia (140).
ALTERATIONS OF THE BBB
Bacterial meningitis, like many other disease states, in-
creases the permeability of the blood-brain barrier (BBB).
The major sites of the BBB are the arachnoid membrane,
choroid plexus epithelium, and cerebral microvascular en-
dothelium. Previous extensive morphologic studies have
demonstrated intact arachnoid membranes in animals with
bacterial meningitis (176). Therefore, the increased BBB
permeability seen in this disorder must occur at the level of
the choroid plexus epithelium, the cerebral microvascular
endothelium, or both; the cerebral microvascular endothe-
lium has been the site of intensive study in recent years as a
result of techniques for isolation of cerebral microvessels or
endothelial cells or both. The features that distinguish cere-
bral capillaries from other capillaries throughout the body
are (i) adjacent endothelial cells fused together by pentalam-
inar tight junctions (zonulae occludens) that prevent inter-
cellular transport; (ii) rare or absent pinocytotic vesicles;
and (iii) abundant mitochondria (18, 49). Therefore, the
increased BBB permeability that occurs during bacterial
meningitis at the level of the cerebral capillary endothelial
cell may result from separation of intercellular tight junc-
tions, from increased pinocytosis, from both alterations, or
by processes as yet unknown.
An adult rat model of bacterial meningitis was used to
investigate the propensity for bacterial meningitis to induce
functional and morphologic alterations of the BBB (110).
Following the intracisternal inoculation of either E. coli, S.
pneumoniae, or H. influenzae into rats, a uniform host
response to all three encapsulated pathogens was observed
at the level of the cerebral capillary endothelial cell, charac-
terized morphologically by an early and sustained increase in
pinocytotic vesicle formation and a progressive increase in
separation of intercellular tight junctions from 4 to 18 h
postinoculation (Table 1). These morphologic changes cor-
related with the functional penetration of albumin across the
BBB, with the highest values of albumin entry occurring 18
h after intracisternal inoculation, when both morphologic
changes were evident. Following intracisternal inoculation
of an unencapsulated strain of H. influenzae (Rd strain),
there was an increase in pinocytotic vesicle formation, but
VOL. 6, 1993
TABLE 1. Correlation of morphologic and functional
alterations of the BBB in an experimental rat
model of H. influenzae-caused meningitisa
Time after
treatment Inoculum (n)
% Penetration
of "-albumin
CBBB mor-
(h)
~~~~~~~~(meanSE)
± phooy
4 Saline (3) 0.26 ± 0.08
H. influenzae Rd (3) 4.12 + 1.25c t PV
H. influenzae type b (4) 3.80 + 0.75c t PV
18 Saline (4) 1.25 ± 0.27
H. influenzae Rd (4) 4.64 ± 0.80c t PV
H. influenzae type b (4) 8.10 ± 1.20cd T PV + t SJ
a
Reproduced from reference 110 with permission of the American Society
for Clinical Investigation.
b PV, pinocytotic vesicles; SJ, separated junctions; T, increase.
c P < 0.05 compared with control.
d p < 0.05 compared with H.
influenzae type b at 4 h and H. influenzae Rd
at 18 h.
separation of intercellular tight junctions did not occur. This
discrepancy between encapsulated and unencapsulated
strains of H. influenzae likely occurred secondary to the
removal of unencapsulated organisms from the CSF by host
defense mechanisms, whereas deficient opsonic mechanisms
in the CSF (see below) permitted sustained concentrations of
the encapsulated strain. Therefore, encapsulation of H.
influenzae was not essential for BBB injury but facilitated
the progression of such injury by avoidance of host defense
mechanisms.
The effect of the host leukocyte response on altered BBB
permeability was subsequently examined in the experimen-
tal rat model by first rendering the animals leukopenic 4 days
following intraperitoneal injection of cyclophosphamide
(68). Functional increases in BBB permeability, assessed by
penetration of radioactive albumin from blood to CSF, were
observed in both normal and leukopenic rats at 18 h follow-
ing inoculation of either encapsulated or unencapsulated
strains of H. influenzae, but permeability was greater after
challenge with the encapsulated strain. Significant increases
in BBB permeability occurred in the near absence of leuko-
cytes in CSF late in the disease process, although the
presence of leukocytes augmented changes in permeability.
At 18 h following inoculation, alterations of BBB permeabil-
ity correlated with concentrations of bacteria in the CSF.
LPS
Since bacterial capsule was not essential for BBB injury in
the experimental rat model of bacterial meningitis and since
it was recognized that pneumococcal capsule did not induce
inflammation within the subarachnoid space (see below),
BBB permeability was examined following intracisternal
inoculation of purified H. influenzae type b LPS. After
intracisternal inoculation of LPS into rats, the following
results were observed (183): (i) dose-dependent increases in
BBB permeability from 2 pg to 20 ng, with attenuation in
peak response after challenge with 500 ng or 1 ,ug; (ii)
time-dependent increases in BBB permeability, with maxi-
mal alteration at 4 h and complete reversal at 18 h (Fig. 3);
(iii) greater increases in permeability after challenge with
LPS than after challenge with the live parent strain despite
identical LPS concentrations; and (iv) close correlation
between BBB permeability and CSF pleocytosis 4 h after
intracisternal inoculation. Preincubation of LPS with poly-
myxin B (a cationic antibiotic that binds to the lipid A region
PATHOGENESIS OF BACT7ERIAL MENINGITIS 123
108
v-
50000 -
8 Bi I}
at 40000 -
6 X t
E
LL W 30000 -
m co
en cn
C. +1
*4 0
C 20000 -
co
10000 -
*2 co
44
0
1
TIME (H)
FIG. 3. Kinetics of changes in concentration of leukocytes
(WBC) in CSF and in BBB permeability (BBBP) after inoculation
with LPS in an experimental rat model. Reproduced from reference
183 with permission of the American Society for Clinical Investiga-
tion. SE, standard error of the mean.
of LPS) and neutrophil acyloxyacyl hydrolase (which re-
moves nonhydroxylated fatty acids from the lipid A region of
the LPS molecule) inhibited the effect of LPS on BBB
permeability, strongly implicating the lipid A region of LPS
in the observed effects in the CNS. Monoclonal antibodies
directed against the oligosaccharide portion of LPS did not
decrease permeability. No change in BBB permeability was
observed following intracisternal inoculation of LPS into
leukopenic rats. Similar results in rats were obtained follow-
ing intracistemal inoculation of H. influenzae type b outer
membrane vesicles (182), which may represent a relevant,
nonreplicating vehicle for the delivery of the toxic moieties
of LPS to host cells.
Cytokines
Because the increased BBB permeability induced by LPS
in the experimental rat model was not maximal until 4 h
following intracisternal inoculation, it was suggested that a
common host mediator(s) was responsible. Therefore, it was
next determined whether specific inflammatory cytokines,
which mediate many of the deleterious effects of LPS, also
increased permeability (112). Intracisternal inoculation of
human recombinant interleukin-lp (IL-1ip) into rats led to a
peak increase in BBB permeability about 3 h after inocula-
tion (Fig. 4), which is earlier than the peak response ob-
served with LPS (at 4 h). This effect was significantly
attenuated by preincubation of the cytokine with a mono-
clonal antibody to IL-1l3 and was totally abolished in leuko-
penic animals. Preincubation of IL-1lB with polymyxin B did
not alter IL-1,B activity. No permeability changes were
observed following intracisternal inoculation of human re-
combinant tumor necrosis factor alpha (TNF-a) into rats,
although rabbit TNF-ot clearly elicits subarachnoid space
inflammation following intracistemal inoculation of the ho-
mologous species. All available evidence suggests that both
cytokines are important and that they act synergistically,
since inoculation with submaximal doses of IL-lp plus
TNF-a at concentrations that produced no changes individ-
ually enhanced BBB permeability.
Localization of BBB Injury
The precise localization of BBB injury in bacterial menin-
gitis was subsequently examined by in situ tracer perfusion
124 TUNKEL AND SCHELD
5 mechanisms responsible for this increased ability to perme-
0
Co
x IL-1I
0
4 O TNFa
'D
0
* Control
E AMP and cyclic GMP was increased in response to LPS, but
3 production of cyclic AMP only was evident prior to the
immLu
0.
2 cerebral microvascular endothelium during bacterial menin-
m')0
w
cn
+1
C
1 of bovine brain microvascular endothelial cells (102), found
c
E
I.
0
0 2 4 6 8 10
Time Post Inoculation (hr)
FIG. 4. Kinetics of changes in CSF traversal of systemically
administered 1251-BSA in an experimental rat model after intracis-
ternal inoculation with recombinant IL-1i, recombinant TNF-a,
and controls. *, P < 0.05. BBBP, BBB permeability; SE, standard
error of the mean. Reproduced from reference 112 with permission
of the American Society for Clinical Investigation.
and immunolabeling procedures to identify the topography
and microvascular exit pathways of bovine serum albumin
(BSA) (111). Two tracers were used: colloidal gold, which is
useful in ultrastructural definition of albumin exit pathways,
and monomeric BSA, which has a faster rate of transcytosis
than colloidal gold. After intracisternal challenge of rats with
E. coli (O111:B4) LPS, an inducible increase in immunode-
tectable monomeric BSA binding to the luminal membranes
of all microvascular segments in the pia-arachnoid and
superficial brain cortex was observed by transmission elec-
tron microscopy. Uptake of colloidal gold and monomeric
BSA by plasmalemmal vesicles was similar, but there was no
detectable transcytosis to the abluminal side of the endothe-
lium. Exit of both perfused colloidal gold-BSA and immuno-
detectable BSA was through open intercellular junctions of
venules in the pia-arachnoid, specifically and topographi-
cally localizing the BBB injury in bacterial meningitis to the
meningeal venules.
In Vitro Model of the BBB
Further studies were performed to delineate the precise
effect of H. influenzae type b LPS on the cerebral microvas-
culature. Purified preparations of cerebral microvascular
endothelial cells (identified by positive immunofluorescent
staining for factor VIII-related antigen and by the ability to
take up acetylated low-density lipoproteins) from rat cere-
bral cortices were grown on a semipermeable support to
create an in vitro model of the BBB (162, 164). Isolated
cerebral microvascular endothelium has been shown to
retain many of the characteristics of an intact BBB, includ-
ing the ability to form intercellular tight junctions in vitro.
Treatment of intact cerebral microvascular endothelial cell
monolayers with LPS (0.1 ,ug/ml) in serum-free media led to
a statistically significant increase in the percentage of radio-
active albumin able to permeate the monolayer (2.2 to 5.6%;
P < 0.001) in the absence of host inflammatory cells. This
increase in permeability was not associated with cytotoxicity
as assessed by release of lactate dehydrogenase into the
media. To determine the putative intracellular regulatory
CLIN. MICROBIOL. REV.
ate monolayers in the in vitro model, the effects of LPS on
the formation of various second messenger systems in the
cerebral microvascular endothelial cells in response to LPS
stimulation were examined (161). Production of both cyclic
increased permeation of the cell monolayer. This observa-
tion suggests that the increased BBB permeability within the
gitis occurs via a cyclic AMP-dependent process.
In contrast, other investigators, utilizing primary cultures
that H. influenzae type b or purified type b LPS caused
marked cytotoxicity of these cells in culture, an effect that
could be completely blocked by polymyxin B. The presence
of serum was essential for the LPS-induced cytotoxic effect,
and a monoclonal antibody against CD14, a receptor in-
volved in mediating the actions of LPS in monocytes,
completely blocked the cytotoxic effect. Further studies are
necessary, however, to precisely define the cellular mecha-
nisms of altered BBB permeability during bacterial menin-
gitis.
BACTERIAL SURVIVAL WITHIN THE
SUBARACHNOID SPACE
CSF Complement
Once meningeal pathogens penetrate the subarachnoid
space, host defense mechanisms are inadequate to control
the infection (131). Complement components in CSF are
usually absent or present in only minimal concentrations (25,
115, 139, 157). Meningeal inflammation leads to increased,
but low, concentrations of complement in CSF. The impor-
tance of this relative complement deficiency in normal and
infected CSF may be critical, since specific antibody and/or
complement is essential for opsonization of encapsulated
meningeal pathogens, efficient phagocytosis, and removal of
pathogens by the spleen (21). Opsonic and bactericidal
activities are absent or barely detectable in patients with
various forms of meningitis; similar observations have been
made in experimental animal models of bacterial meningitis.
A serum-sensitive E. coli strain was not opsonized in vitro
when the strain was incubated with CSF from rabbits
challenged intracisternally with E. coli Kl, although CSF
from animals with Staphylococcus aureus-caused meningitis
was opsonically active in vitro against E. coli (13). The
absence of opsonization after E. coli challenge may have
been due to the absorption of a specific opsonin by E. coli
early in the course of infection. Alternatively, more serum
protein, and therefore opsonins, may have penetrated into
CSF in staphylococcal rather than E. coli meningitis. Al-
though there are low functional and bactericidal activities in
purulent CSF, the presence of some measurable opsonic
activity may correlate with a favorable outcome. For exam-
ple, outcome was better in patients whose concentrated CSF
samples demonstrated opsonic activity (15 of 27 patients
[186]), suggesting that complement-mediated opsonic activ-
ity can appear in the CSF during bacterial meningitis,
particularly in patients who recover completely.
Several explanations of the low concentrations of comple-
ment in CSF during meningitis have been advanced: (i)
insufficient traversal across the BBB; (ii) variable subarach-
noid space inflammation; (iii) enhanced clearance or removal
from the subarachnoid space; (iv) low production rates in the
VOL. 6, 1993
CNS; and (v) degradation at the site of infection (131). This
last possibility has been investigated in patients and in
animal models of meningitis. Leukocyte proteases have been
shown to degrade functional complement components (e.g.,
C3b) in CSF from patients with meningococcal meningitis,
with the formation of nonopsonic products (e.g., C3d) (179).
In the experimental rabbit model of pneumococcal meningi-
tis, the intracisternal inoculation of phenylmethylsulfonyl
fluoride, a nonspecific protease inhibitor, led to a decline in
pneumococcal concentrations in CSF compared with those
of saline-inoculated controls (131). Therefore, during bacte-
rial meningitis, complement components crossing the BBB
may be degraded by leukocyte proteases, resulting in ineffi-
cient opsonic activity at the site of infection similar to
mechanisms postulated to occur in the pleural space during
the development of an empyema.
CSF Antibody
Immunoglobulin concentrations are also low in normal
CSF, with an average blood/CSF ratio of IgG of about 800:1.
Immunoglobulin concentrations in CSF increase during bac-
terial meningitis but remain low compared with simultaneous
concentrations in serum, and IgG does not appear in the CSF
until late in the course of disease (142, 179). An experimental
rabbit model of pneumococcal meningitis has been utilized
to study the importance of antibodies in host defense against
bacterial meningitis. Intracisternal inoculation of type-spe-
cific antibodies against S. pneumoniae decreased pneumo-
coccal concentrations in CSF, but at a much slower rate than
did treatment with appropriate bactericidal antibiotics (131).
The systemic administration of type-specific monoclonal
antibodies has been examined in the experimental rabbit
model. Intravenous injection of a bactericidal monoclonal
antibody against the polyribosyl-ribitol phosphate of H.
influenzae type b produced high concentrations of antibody
in serum, but there was poor BBB penetration (c5.5%),
even in the presence of meningeal inflammation (47). These
results suggested that systemic administration of type-spe-
cific antibodies alone is suboptimal. The pioneering work of
Flexner (42), who demonstrated that systemic and intrathe-
cal administration of antimeningococcal antiserum raised in
horses was effective in reducing the mortality rate in menin-
gococcal meningitis from approximately 80 to 30%, suggests
that combined sites of administration may be useful. It
remains to be determined, however, whether the intrathecal
administration of antibodies may be useful in the adjunctive
treatment of bacterial meningitis.
CSF Leukocytes
One of the hallmarks of bacterial meningitis is the devel-
opment of a neutrophilic pleocytosis within the CSF, al-
though the precise mechanisms of leukocyte traversal across
the BBB remain to be defined (52, 184). In an experimental
animal model of pneumococcal meningitis, the complement
component C5a has been suggested as one chemotactic
substance in CSF (35, 96, 158), with chemotactic activity
appearing 2 to 4 h before neutrophil influx into the CSF. The
precise role of C5a as a chemotactic substance has recently
been examined in the experimental rabbit model, in which
the intracistemal inoculation of C5a caused a rapid early
influx of leukocytes into CSF 1 h after inoculation (60). This
response was attenuated by coadministration of CSa with
prostaglandin E2 (PGE2) in a dose-dependent manner, sug-
gesting a direct anti-inflammatory action of PGE2 on CSa-
PATHOGENESIS OF BACTERIAL MENINGITIS 125
generated CSF pleocytosis during bacterial meningitis.
However, despite the entry of leukocytes, host defense
mechanisms in CSF remain suboptimal because of the rela-
tive lack of functional opsonic and bactericidal activity.
Therefore, phagocytosis is inefficient at this protected site,
leading to huge concentrations of bacteria in the CSF during
meningitis (36, 37).
There is controversy over the precise role of neutrophils in
host defense within the CNS during bacterial meningitis. In
experimental animal models, low concentrations of leuko-
cytes in CSF have generally been associated with increased
mortality rates (46, 134). Although a study of dogs with
pneumococcal meningitis revealed that leukopenic animals
had a higher survival rate than animals with normal periph-
eral leukocyte counts (62 versus 47 h) (104), the small
number of animals studied precluded statistical analysis. In
an experimental rabbit model of pneumococcal meningitis,
the parameters of bacterial growth rate; final concentrations
of bacteria in CSF; and concentrations of protein, glucose,
and lactate in CSF were no different in animals rendered
leukopenic by the prior intravenous inoculation of nitrogen
mustard than in nonleukopenic animals. However, the re-
sultant bacteremia was about 100-fold greater in leukopenic
animals (34), suggesting that neutrophils either prevented
traversal of pneumococci from CSF to blood or enhanced
neutrophil-mediated bacterial elimination from the blood-
stream at extraneural sites.
The precise pathway by which neutrophils enter the
subarachnoid space remains unknown. Adherence of neu-
trophils to vascular endothelial cells is a likely prerequisite
for traversal into the CSF. Pretreatment of endothelial cells
with cytokines induces formation of specific adhesion mol-
ecules such as endothelial leukocyte adhesion molecule 1
(16). Neutrophil adherence to vascular endothelial cells is
enhanced by pretreatment of the endothelial cells with LPS
(137) or with inflammatory cytokines such as IL-1 and TNF
(15, 137, 165). Similar mechanisms may also enhance neu-
trophil binding to cerebral vascular endothelium (11). A
recent study of an experimental rabbit model has demon-
strated that the intravenous inoculation of a monoclonal
antibody (IB4) against the CD18 family of receptors on
leukocytes blocked the accumulation of leukocytes in the
subarachnoid space despite intracisternal challenge with H.
influenzae type b, N. meningitidis, pneumococcal cell wall,
or LPS (170). The monoclonal antibody also attenuated a
parameter of BBB permeability (i.e., increased protein con-
centrations in the CSF). Furthermore, development of cere-
bral edema and death were prevented in monoclonal anti-
body-treated animals challenged with lethal doses of S.
pneumoniae. Penetration of antibiotics into CSF, bacterial
concentrations in CSF, and bactericidal response to ampi-
cillin therapy were not affected by monoclonal antibody
administration, although the animals exhibited a delay in the
onset of bacteremia and there was an attenuated CSF
inflammatory response after ampicillin-induced bacterial
killing. These results suggest that systemic inoculation of
monoclonal antibodies directed at leukocyte-endothelium
interactions can block leukocyte-mediated damage within
the CNS during bacterial meningitis.
INDUCTION OF SUBARACHNOID SPACE
INFLAMMATION
The ability of meningeal pathogens to induce a marked
subarachnoid space inflammatory response contributes to
many of the pathophysiologic consequences of bacterial
126 TUNKEL AND SCHELD
meningitis (e.g., cerebral edema and increased intracranial
pressure). For example, this inflammatory response may be
a crucial factor in the development of sensorineural deafness
in patients with bacterial meningitis. Utilizing an infant rat
model of pneumococcal meningitis, multiple intraperitoneal
inocula (1 x 108 to 10 x 108 CFU every 24 h for 3 days) of
S. pneumoniae in 5-day-old rats led to bacteremia and
meningitis in 50 and 46% of animals, respectively (121).
Hematoxylin-eosin-stained sections of brain tissue from rats
with positive CSF cultures revealed inflammation in the
meninges and scala tympani but not in the scala media. In
addition, perilymphatic inflammation occurred significantly
(P < 0.001) more than did endolymphatic inflammation,
suggesting that bacteria reached the cochlea via the cochlear
duct, thereby providing a connection between perilymphatic
fluid and CSF. Recent studies in other animal model systems
have focused on the bacterial virulence factors that are
responsible for this inflammatory response, providing many
new, exciting areas of investigation.
Cell Wall
Despite the fact that bacterial capsule is largely responsi-
ble for intravascular and subarachnoid space survival of
meningeal pathogens, capsular polysaccharides are remark-
ably noninflammatory when injected in purified form into the
subarachnoid space. In contrast, cell walls of gram-positive
bacteria are potent inducers of inflammation, and S. pneu-
moniae cell walls activate the alternative complement path-
way (181). Various surface components of S. pneumoniae
have been examined to determine their roles in subarachnoid
space inflammation. A CSF inflammatory response was
invoked in an experimental rabbit model by intracisternal
inoculation of live encapsulated and unencapsulated S.
pneumoniae cells, heat-killed unencapsulated pneumococci,
and pneumococcal cell walls (169); no inflammation was
induced by intracisternal inoculation of either heat-killed
encapsulated strains or isolated capsular polysaccharide.
Pneumococcal cell walls inoculated intracisternally were the
most potent inducers of CSF inflammation among these
substances (168). In addition, independent intracisternal
injection of the major components of the pneumococcal cell
wall, teichoic acid and peptidoglycan, also induced CSF
inflammation. Teichoic acid had the highest specific activity
of the cell wall fractions tested, and activity was markedly
reduced if either component (teichoic acid or peptidoglycan)
was extensively degraded. These findings suggest that pneu-
mococcal cell wall lytic products released during antibiotic-
induced autolysis during treatment of bacterial meningitis
contribute to the host inflammatory response in the sub-
arachnoid space.
LPS
Intracisternal inoculation of purified H. influenzae type b
LPS (in contrast to purified capsular polysaccharide) also
induces subarachnoid space inflammation. In an experimen-
tal rabbit model, the intracisternal inoculation of purified
LPS (2 fg to 200 ng) produced a dose-dependent increase in
concentrations of leukocytes in CSF, whereas no inflamma-
tion was induced by intracisternal inoculation of purified H.
influenzae type b capsular polysaccharide (149). This inflam-
matory response was blocked by polymyxin B and neutro-
phil acyloxyacyl hydrolase, a fact that supports the impor-
tance of the lipid A region of the LPS in the induction of CSF
inflammation. Pretreatment of the LPS with a monoclonal
CLIN. MICROBIOL. REV.
antibody directed against epitopes on the oligosaccharide
portion of the LPS molecule did not reduce the inflammatory
potential of the LPS. Similar results were observed in an
experimental rat model following the intracisternal inocula-
tion of purified LPS, with the maximal amount of inflamma-
tion observed 8 h after inoculation of a 20-ng dose (Fig. 3)
(183). In addition, the intracisternal inoculation of outer
membrane vesicles from H. influenzae type b induced men-
ingeal inflammation in both rabbits and rats in a dose- and
time-dependent manner (92, 182); this response was blocked
by polymyxin B but not by two monoclonal antibodies
directed against surface epitopes of the oligosaccharide side
chains of LPS within outer membrane vesicles. This obser-
vation supports the concept that the delivery of LPS via
outer membrane vesicles induces subarachnoid space in-
flammation.
Inflammatory Mediators
Recently, evidence that supports the hypothesis that
pneumococcal cell wall or H. influenzae type b LPS induces
subarachnoid space inflammation through the local CNS
release of inflammatory mediators such as IL-1, TNF (14,
80, 159), and/or prostaglandins (77, 95) has been accumulat-
ing. Intact pneumococci, pneumococcal cell wall, or lipotei-
choic acid induced IL-1 production by human peripheral
monocytes in vitro (118). IL-1 production by monocytes was
dramatically reduced by chemical alteration of phosphoryl-
choline, the major determinant of teichoic acid, or by
pretreatment with a phosphorylcholine antibody. TNF pro-
duction was not induced in monocytes incubated with pneu-
mococcal cell wall in vitro.
In an experimental rat model, the intracisternal inocula-
tion of H. influenzae type b LPS led to elevated concentra-
tions of IL-1 and TNF in CSF within 30 to 120 min (166).
Elevated concentrations of TNF in CSF have also been
observed in the experimental rabbit model following intra-
cisternal inoculation of LPS (90). TNF activity in CSF was
first detected 45 min after intracisternal challenge with LPS,
with peak activity at 2 h and persistence for about 5 h.
Following intracisternal challenge with live H. influenzae
type b, peak TNF activity in CSF was comparable to that
after LPS challenge, although activity persisted for -14 h.
TNF activity and meningeal inflammation (i.e., CSF pleocy-
tosis) in CSF were significantly reduced by therapy with
dexamethasone or by the intracisternal administration of
goat anti-human TNF-ao antibodies. Interestingly, simulta-
neous analysis of serum samples revealed no detectable
TNF activity, indicating that TNF was principally produced
within the CNS. This local production of TNF-ot has also
been observed in patients with bacterial meningitis (79). In
addition, the finding of increased CSF concentrations of
TNF-a may be specific for bacterial meningitis, as was
suggested in a recent study. TNF-ot concentrations, mea-
sured in mice or humans with either bacterial or viral
meningitis, were elevated in CSF only during bacterial
meningitis (67); however, only a few patients were analyzed.
These results were confirmed in a subsequent prospective
study of patients with infectious meningitis. The presence of
TNF-a in CSF was indicative of a bacterial etiology, al-
though the absence of TNF-ao did not exclude the diagnosis
of bacterial meningitis (93). In addition, elevated concentra-
tions of PGE2, prostacyclin, IL-11, and TNF in CSF were
found in the majority of infants and children with bacterial
meningitis (91). Systemic dexamethasone administration sig-
VOL. 6, 1993
TABLE 2. Concentrations of endotoxin in CSF and bacterial titers before and after antibiotic therapy in eight patientsa
Time TotaTiter
CFU/ml)
(l (log10
CFU/ml)
Before therapy 6.69 + 0.71
After therapy 5.24 + 0.85
P value b <0.005
a Data from reference 5 with permission of The University of Chicago Press.
b
Significance of paired t test (two tailed).
nificantly reduced concentrations of PGE2 and IL-lp in CSF
and improved several indices of meningeal inflammation.
Inoculation of purified rabbit TNF-a or human recombi-
nant IL-1B into the CSF of rabbits also produced significant
CSF inflammation (117). Simultaneous administration of
lower doses of each cytokine resulted in an apparent syner-
gistic inflammatory response that was manifested by a more
rapid and significantly increased influx of leukocytes into
CSF than occurred with administration of each cytokine
alone. In contrast, in an experimental rabbit model of
pneumococcal meningitis, CSF leukocytosis, BBB perme-
ability, and brain edema were induced by intracisternal
inoculation of human recombinant TNF-a, macrophage in-
flammatory proteins 1 and 2, and IL-la but not by intracis-
ternal inoculation with IL-lp (127). Antibodies homologous
to each mediator inhibited leukocytosis and brain edema. In
rabbits treated with a monoclonal antibody to CD18 to
render neutrophil-endothelial cell interactions dysfunctional,
each cytokine lost the ability to cause leukocytosis and brain
edema. Therefore, these cytokines have multiple complex
and interrelated activities in the CNS that contribute to
tissue damage during pneumococcal meningitis.
These findings have implications with regard to outcomes
in patients with bacterial meningitis. Outcome from menin-
gitis due to gram-negative bacilli has been correlated with
persistence of organisms and higher concentrations of endo-
toxin in CSF (as detected by the Limulus lysate assay) (73).
A study of children with H. influenzae meningitis docu-
mented that treatment with ceftriaxone induced a marked
increase in concentrations of free LPS in CSF within 2 to 6
h (Table 2) that correlated with the Herson-Todd severity
score and the number of febrile hospital days (5). These data
suggest that early release of "free" LPS (i.e., LPS not
present in outer membranes of viable organisms) after anti-
microbial therapy enhances the host subarachnoid space
inflammatory response. The degree of elevated concentra-
tions of IL-1lB in CSF also correlated significantly with
outcome from neonatal gram-negative bacillary meningitis in
children (88). In another study of infants and children
predominantly with H. influenzae type b meningitis, patients
with IL-113 concentrations in CSF of .500 pg/ml were more
likely to develop neurologic sequelae (87). Although ele-
vated concentrations of TNF in CSF were observed in 50 to
75% of patients with bacterial meningitis, there was no
correlation between concentrations of TNF in CSF and
outcome. The roles of other inflammatory cytokines in the
induction of subarachnoid space inflammation are less clear.
Elevated concentrations of platelet-activating factor in CSF
have been demonstrated in children with H. influenzae
meningitis (6) and correlated with bacterial density and with
LPS and TNF-a concentrations in CSF. These increased
concentrations of TNF-a and platelet-activating factor were
associated with severity of disease. Increased concentra-
tions of IL-6, occurring after release of TNF-ot and before
PATHOGENESIS OF BACTERIAL MENINGITIS 127
of endotoxin/mi (mean
Loglo ngBound + SE)
~Total Free (%
2.06 + 0.27 2.04 ± 0.28 0.75 + 0.21 (8.5 ± 4.2)
1.53 + 0.29 1.06 ± 0.33 1.29 + 0.23 (63.5 + 11.5)
<0.02 <0.01 <0.01 (<0.005)
neutrophilic infiltration into CSF (174), have also been
observed in patients with bacterial meningitis (58, 123, 175).
These experimental and clinical studies strongly suggest that
release of inflammatory mediators into the CSF during
bacterial meningitis is responsible for induction of a marked
subarachnoid space inflammatory response and may possi-
bly correlate with morbidity and mortality from this disor-
der.
The source of these inflammatory cytokines within the
CSF of patients with bacterial meningitis is unclear. LPS
stimulation of astrocytes and microglia in vitro leads to
release of various cytokines (43), and vascular endothelial
cells in culture produce IL-1 in response to stimulation with
either LPS or TNF (69, 78, 94). One study utilizing purified
preparations of cerebral microvascular endothelial cells from
rats has demonstrated that these cells release IL-6 in vitro in
response to LPS stimulation (160). Further studies, how-
ever, are warranted to address these issues as they pertain to
the patient with bacterial meningitis.
INCREASED INTRACRANIAL PRESSURE
The major element that contributes to an increase in
intracranial pressure during bacterial meningitis is the devel-
opment of cerebral edema, which may be vasogenic, cyto-
toxic, and/or interstitial in origin and may result in life-
threatening cerebral herniation and other complications (26,
40, 41, 57, 98). Vasogenic cerebral edema is principally a
consequence of increased BBB permeability (see above).
Cytotoxic cerebral edema results from swelling of the cellu-
lar elements of the brain, most likely due, in bacterial
meningitis, to release of toxic factors from neutrophils or
bacteria or both. Secretion of antidiuretic hormone also
contributes to the pathogenesis of cytotoxic edema, with
resultant hypotonicity of extracellular fluid and increased
permeability of the brain to water (62). Interstitial edema
reflects obstruction of flow in normal CSF pathways (e.g.,
from the subarachnoid space to blood), as in hydrocephalus.
In an experimental rabbit model of pneumococcal or E.
coli-caused meningitis, the CSF outflow resistance, defined
(and quantified) as factors that inhibit the flow of CSF from
the subarachnoid space to the major dural sinuses, was
markedly elevated (133); these alterations remained for as
long as 2 weeks despite rapid CSF sterilization with penicil-
lin therapy. An increase in the outflow resistance to CSF
movement may cause interstitial brain edema and/or the
resultant hydrocephalus during bacterial meningitis.
Subsequent studies have examined these concepts in more
detail by measuring the water content of the brain (indicative
of cerebral edema if elevated), concentrations of lactate in
CSF, and CSF pressure in animals with pneumococcal
meningitis (154). All three parameters were elevated in
infected animals. Treatment with ampicillin sterilized the
CSF rapidly and normalized the water content of the brain
128 TUNKEL AND SCHELD
and intracranial pressure within 24 h, but the concentration
of lactate in the CSF remained elevated. The bacterial cell
components responsible for the production of brain edema
were studied in an experimental model of E. coli-caused
meningitis (155). Treatment with cefotaxime but not chlor-
amphenicol induced a marked rise in concentrations of
endotoxin in the CSF that were associated with an increase
in the water content of the brain. These effects were neu-
tralized by either polymyxin B or a monoclonal antibody
against lipid A, indicating that increased concentrations of
endotoxin in CSF may be associated with brain edema. The
putative role of the leukocyte in these processes was re-
cently examined in an experimental meningitis model (152).
In sterile meningitis induced by the intracisternal inoculation
of N-formylmethionylleucylphenylalanine (fMLP), a chemo-
tactic peptide, both high (10-' M) and low (10-5 M) doses
induced a CSF pleocytosis, although only high doses pro-
duced an increase in the water content of the brain; intra-
cranial pressure and the concentrations of lactate and pro-
tein in the CSF were unaltered. No increase in the water
content of the brain was observed in neutropenic animals.
When high doses of N-formylmethionylleucylphenylalanine
were injected during the course of pneumococcal meningitis
in rabbits, the results were similar, suggesting that neutro-
phils contributed to cerebral edema if adequately stimulated.
The parameters of increased intracranial pressure and in-
creased concentrations of lactate and protein in the CSF
seemed unrelated to the presence of neutrophils. This area
remains controversial, however, because neutrophils are
required for the increased BBB permeability seen in re-
sponse to intracisternal injection of bacterial products and
inflammatory mediators (112, 183). Additional studies are
needed to more precisely define the role of the neutrophil
within the CNS in the pathophysiology of bacterial menin-
gitis.
Variability among bacterial strains may also be an impor-
tant determinant in the production of brain edema. A recent
study found that intracisternal injection of three different
pneumococcal isolates resulted in pronounced differences in
the pathophysiologic profiles 24 h after challenge (153).
Following intracisternal inoculation of pneumococcal cell
wall fragments, the chemical composition of the fragments,
specifically the degree of teichoication, was found to influ-
ence the induction of brain edema during bacterial meningi-
tis. Spontaneous release of cell wall is reduced for certain
pneumococcal strains and may thus limit their inflammatory
potential. Further work is needed, however, to determine
whether these differences affect the clinical expression of
disease.
CEREBRAL VASCULITIS
Bacterial meningitis exerts profound effects on blood
vessels coursing through the subarachnoid space (116). The
resulting vasculitis leads to narrowing and/or thrombosis of
cerebral blood vessels and the propensity for ischemia
and/or infarction of underlying brain. Arteriography in chil-
dren with bacterial meningitis uniformly demonstrates leak-
age of contrast material or other vascular abnormalities in
the subarachnoid space, although these changes reverted to
normal following successful antibiotic therapy. Severe neu-
rologic complications (e.g., hemiparesis and quadriparesis)
with permanent sequelae may result from involvement of the
large arteries at the base of the brain (59). Vasospasm may
also occur secondary to release of humoral factors elabo-
rated within the CSF or blood vessel wall and may subse-
CLIN. MICROBIOL. REV.
quently lead to vasodilatation or organic stenosis or both
later in the course of disease (185). Phlebitis of the major
cortical draining vessels or dural sinuses or both may result
in thrombosis with secondary brain infarction, focal neuro-
logic deficits, and prominent seizure activity.
ALTERATIONS IN CEREBRAL BLOOD FLOW
In combination with increased intracranial pressure, cere-
bral vasculitis may result in altered cerebral blood flow in
patients with bacterial meningitis. In the infant rhesus mon-
key model of H. influenzae meningitis, cerebral blood flow
(measured by an autoradiographic technique utilizing
[14C]antipyrine) was lower in certain areas of the cortex
(postcentral, temporal, and occipital areas) than in the
hypothalamus and midbrain while the brain stem was hyper-
perfused, suggesting that one of the initial physiologic
changes in H. influenzae meningitis is cerebral cortical
hypoperfusion with resultant relative cerebral anoxia (141).
A recent report has demonstrated that cerebrovascular au-
toregulation is lost in experimental bacterial meningitis
(171). Cerebral blood flow was increased when systemic
blood pressure was raised and decreased when blood pres-
sure was lowered, indicating that flow was pressure passive.
Similar changes were observed in intracranial pressure;
increased blood flow led to increased intracranial pressure.
Furthermore, in an experimental rat model of meningitis
(105), an increase in cerebral blood flow was observed within
the first few hours of intracisternal inoculation of either live
pneumococci or pneumococcal cell wall fragments. These
results suggested that maintenance of adequate intravascular
volume status and minimization of stimuli that increase
systemic blood pressure may be important in the treatment
of bacterial meningitis and of potentially practical clinical
relevance. Measurement of cerebral blood flow (by the
xenon-133 intra-arterial injection method) in an earlier study
of patients with bacterial meningitis revealed a 30 to 40%
reduction in average total blood flow in five patients with
pneumococcal meningitis (mean age, 54 years) but not in five
patients with meningococcal meningitis (mean age, 20 years)
(165). An inverse relationship between cerebral blood flow
velocity and intracranial pressure has been seen in infants
with bacterial meningitis (76). Among eight patients, these
alterations were detected only in the four older infants (ages,
3 to 10 months) and not in the four neonates (ages, 5 to 30
days), in whom no changes in cerebral blood flow velocity
were observed.
A subsequent study that measured total and regional
cerebral blood flow by stable xenon computed tomography
in 20 children seriously ill with bacterial meningitis revealed
a global decrease in cerebral blood flow and even more
regional variability (7). Autoregulation of cerebral blood flow
was preserved in the patients studied, although hyperventi-
lation reduced cerebral blood flow below the ischemic
threshold, raising important concerns about the routine use
of hyperventilation in the management of increased intracra-
nial pressure in patients with bacterial meningitis. Some
authors have suggested that in infants and children with
bacterial meningitis and initially normal computed tomogra-
phy or magnetic resonance imaging scans, the chance that
cerebral blood flow would be reduced to ischemic levels is
unlikely, and in this situation hyperventilation may safely
reduce elevated intracranial pressure for the first 24 to 48 h
before its effect diminishes (8). However, in children with
cerebral edema revealed by neuroimaging studies, cerebral
blood flow is more likely to be normal or reduced, so that
VOL. 6, 1993
TABLE 3. Effect of cyclooxygenase and lipoxygenase inhibitors on CSF leukocytosis in rabbits after intracisternal
injection of 30 p.g of pneumococcal cell walls'
Compound (mg/k
anDoseand
route"
Control
Methylprednisolone 30, i.m.
Diclofenac sodium 5, i.v.
Indomethacin 5, p.o.
Nordihydroguaiaretic acid 5, p.o.
Oxindanac 5, p.o.
a Reproduced from reference 167 with permission of The University of Chicago Press.
bi.m., intramuscular; i.v., intravenous; p.o., oral.
c Time of administration in reference to time in hours of cell wall challenge dose.
d Mean number of leukocytes per microliter of CSF + standard deviation. Value at 0
e p < 0.01 compared with control at 7 h.
hyperventilation may decrease the intracranial pressure at
the expense of a significant reduction of cerebral blood flow,
possibly approaching ischemic thresholds. These patients
may benefit from early use of diuretics (e.g., furosemide) or
osmotically dehydrating agents such as mannitol, providing
that intravascular volume is protected. Corticosteroids may
also be useful in this situation. However, it must be stressed
that controlled clinical trials examining these issues have yet
to be performed. Although the definitive changes in cerebral
blood flow during bacterial meningitis are controversial and
may vary with the stage of disease, these blood flow alter-
ations may lead to regional hypoxia, increased concentra-
tions of lactate in the brain secondary to utilization of
glucose by anaerobic glycolysis, and CSF acidosis, which
may be a precursor to encephalopathy. This phenomenon
has recently been examined in an experimental rabbit model
of pneumococcal meningitis (173) in which animals given a
lower fluid regimen (50 ml/kg per 24 h) of normal saline had
lower mean arterial blood pressure, lower cerebral blood
flow, and higher concentrations of lactate in the CSF com-
pared with animals that received a higher fluid regimen (150
ml/kg per 24 h). These results suggest that intravascular
volume status may be a critical factor in determination of
cerebral blood flow and, therefore, the degree of cerebral
ischemia in meningitis.
ADJUNCTIVE THERAPEUTIC STRATEGIES
Experimental Studies
Because of the information supporting subarachnoid space
inflammation as a major factor contributing to morbidity and
mortality from bacterial meningitis, several studies in exper-
imental animal models have examined whether attenuation
of this inflammatory response might be beneficial. As stated
above, the generation of pneumococcal cell wall components
after treatment with bacteriolytic antibiotics in the experi-
mental rabbit model may contribute to an increased inflam-
matory response within the subarachnoid space (168, 169).
This CSF inflammatory response was reduced by agents
known to exert their effects by inhibition of the cyclooxyge-
nase pathway of arachidonic acid metabolism (Table 3)
(167). Treatment with methylprednisolone or oxindanac in
addition to antimicrobial agents was particularly effective in
decreasing pneumococcal cell wall-induced inflammation,
whereas another inhibitor, diclofenac sodium, was effective
only when administered 5 and 7 h after inoculation of
bacteria into CSF, but did not produce an inhibitory re-
PATHOGENESIS OF BACTERIAL MENINGITIS 129
Times (h)C CSF leukocytosisd after:
~~~~~~~
~~5
h 7h ~ 24 h
677 34 1,545 ±90 870 ± 92
-1 50 ± 20 198 ± 30e 54 ± 20
-1, +2 110 ± 40 433 ± 210e 1,183 ± 150
-1, +2, +5 155 ± 60 300 ± 50 320 ± 50
-1, +2, +5 1,093 - 500 870 ± 240
-1, +2, +5 118 ± 57 59 ± 14e 500 ± 325
h, 28 + 14 cells per p.I.
sponse 24 h after challenge with cell wall. An inhibitor of the
lipoxygenase pathway, nordihydroguaiaretic acid, was inef-
fective in preventing cell wall-induced inflammation. Similar
results were observed with these adjunctive agents after
intracisternal challenge with live pneumococci. There was a
correlation between the concentrations of the arachidonic
acid metabolite PGE2 and of leukocytes in the CSF after
intracisternal challenge with either live pneumococci or
pneumococcal cell walls, and inhibition of the cyclooxyge-
nase pathway reduced CSF inflammation and the concentra-
tion of PGE2 in the CSF. Administration of the nonsteroidal
anti-inflammatory agent indomethacin also led to a decrease
in both the water content of the brain and concentrations of
PGE2 during experimental rabbit pneumococcal meningitis,
although there was no reduction in intracranial pressure
(172). In addition, an anti-inflammatory agent (dexametha-
sone or oxindanac) lessened the massive influx of serum
albumin and other proteins of high and low molecular masses
into the CSF during the early phases of experimental pneu-
mococcal meningitis (61). Ampicillin given alone or in com-
bination with indomethacin was ineffective in preventing this
influx, and the abnormal protein profile in the CSF persisted
for up to 30 days after the initiation of the experimental
infection.
Several corticosteroid agents have been evaluated in ex-
perimental animal models of bacterial meningitis. Early
studies revealed a significant reduction in the mass of
leukocytes within the meninges of rabbits with pneumococ-
cal meningitis following methylprednisolone administration
compared with concentrations in infected controls (97);
chemotactic activity, chemotactic response, and phagocyto-
sis in CSF were, however, not altered by methylpred-
nisolone treatment, although there was an attenuation of
rabbit neutrophil adherence to a nylon fiber column. CSF
outflow resistance (defined as factors that inhibit the flow of
CSF from the subarachnoid space to the major dural sinuses
and quantified following infusion of mock CSF into the
subarachnoid space) was also reduced by methylpred-
nisolone therapy and to a greater extent than in untreated or
penicillin-treated rabbits with pneumococcal meningitis
(133). The reduction in resistance was apparent within 4 h of
the second injection of methylprednisolone (at a dose of 30
mg/kg of body weight intramuscularly 16 and 20 h following
intracistemal inoculation); a rebound increase in CSF out-
flow resistance was not seen after corticosteroid therapy was
discontinued.
The effects of methylprednisolone or dexamethasone on
water content of the brain, CSF pressure, and lactate con-
130 TUNKEL AND SCHELD
TABLE 4. Influence of methylprednisolone and dexamethasone on manifestations of a 24-h infection with pneumococci in rabbitsa
Treatment Bacterial titer Leukocyte count
(log10o CFU/ml) (103/mm3)
No infection ND <0.01
24-h infection
No treatment 7.0 ± 1.2 2.5 ± 6.6
Methylprednisolone 7.6 ± 0.7c 3.9 ± 4.9
Dexamethasone 6.7 ± 0.6 2.1 ± 2.2
a Reproduced from reference 154 with permission of The University of Chicago Press. All values are expressed as means + standard deviations. ND, not
determined.
b p < 0.02 in comparison with other groups.
c P < 0.05 in comparison with dexamethasone-treated group.
d p < 0.02 in comparison with other groups infected for 24 h.
centrations in the CSF of rabbits with pneumococcal men-
ingitis have also been examined (Table 4) (154). Administra-
tion of both agents completely reversed the development of
brain edema at the times studied, but only dexamethasone
reduced the increases in CSF pressure and lactate; methyl-
prednisolone, but not dexamethasone, was associated with
an increase in concentrations of bacteria in CSF. However,
neither agent was superior to therapy with ampicillin alone in
reducing cerebral edema or intracranial pressure. No com-
parison between ampicillin alone and ampicillin plus corti-
costeroids, a comparison that would be relevant to the
potential clinical efficacy of adjunctive corticosteroids in
humans, was made. In another study, treatment with ceftri-
axone was compared with treatment with ceftriaxone plus
dexamethasone in an experimental rabbit model of H. influ-
enzae meningitis (150). Combination therapy consistently
reduced the water content of the brain, CSF pressure, and
the lactate concentration in the CSF to a greater degree than
ceftriaxone alone, although the differences were not statis-
tically significant. By 29 h after inoculation (a well-estab-
lished disease in this model), the values were comparable
whether the animals received antibiotic alone, dexametha-
sone alone, or the combination. The authors suggested that
dexamethasone might be more beneficial if administered
early during antibiotic therapy after (or even before) antibi-
otic-induced lysis and release of microbial products. In a
subsequent analysis utilizing the rabbit model of H. influen-
zae type b meningitis (89), a significant increase in concen-
trations of endotoxin in CSF was documented 2 h following
ceftriaxone administration; this increase was followed by a
rise in TNF concentrations in CSF (Fig. 5). Simultaneous
administration of dexamethasone and ceftriaxone did not
affect the appearance of endotoxin in CSF, but there was a
marked attenuation in concentrations of TNF in CSF mea-
sured 8 h later. Adjunctive dexamethasone therapy also
resulted in a significant decrease in the resultant CSF leuko-
cytosis and a trend towards earlier improvement in concen-
trations of glucose, lactate, and protein in CSF.
Therefore, adjunctive dexamethasone therapy in concert
with antimicrobial agents appeared to improve a number of
parameters of subarachnoid space inflammation in experi-
mental animal models of bacterial meningitis. These param-
eters improved without any apparent decrease in the rate of
bacterial killing within the CSF in vivo, although other
experimental studies have shown that administration of
methylprednisolone decreased the entry of ampicillin and
gentamicin into CSF (132).
The subarachnoid space inflammatory response may con-
tribute to the pathogenesis of hearing loss in bacterial
CLIN. MICROBIOL. REV.
Concn of lactate Change in CSF Water content (g
in CSF (mg/dl) pressure (mm Hg) [dry wt]) of brain
14.0 ± 2.6 +0.8 ± 1.4 396 ± 14
75.3 ± 25.6 +7.5 ± 6.5 410 ± llb
64.3 ± 33.1 +7.8 ± 5.4 395 ± 9
43.8 ± 12.3 +1.8 ± 2.7d 399 ± 12
meningitis, which is due in part to the development of
labyrinthritis following spread of infection to the inner ear.
The infant rat model has been utilized to determine the
influence of corticosteroid administration on the inflamma-
tory reaction in the cochleas of infected animals (63). Infant
rats were inoculated intraperitoneally with H. influenzae
type b, and 24 h later they were treated with ampicillin or
ampicillin plus dexamethasone. At 48 h, concentrations of
leukocytes in CSF were significantly lower in the dexameth-
asone group than in the ampicillin-alone group, although no
histologic differences in the degree of cochlear inflammation
were noted between groups. The extent of cochlear inflam-
mation was minimal only in the ampicillin group, however.
Pentoxifylline, a phosphodiesterase inhibitor that de-
creases endotoxin-induced TNF-a production and attenu-
ates the inflammatory action of IL-1 and TNF on leukocyte
function (147), has also been examined in the experimental
rabbit model of H. influenzae type b meningitis (125).
Administration of pentoxifylline 20 min before intracisternal
challenge with H. influenzae type b LPS significantly re-
duced concentrations of leukocytes, protein, and lactate in
CSF. Peak concentrations of TNF in CSF were reduced by
more than one-third in pentoxifylline-treated animals, al-
though this reduction was not statistically significant and
was unlikely to be solely responsible for the marked modu-
lation of meningeal inflammation. Dexamethasone was su-
perior to pentoxifylline in modulation of these inflammatory
changes in CSF, and no appreciable synergism was observed
when dexamethasone and pentoxifylline were used together.
Recent studies have examined the effects of a monoclonal
antibody (IB4) directed against the CD18 family of receptors
on leukocytes on reduction of subarachnoid space inflam-
mation. In one study, intravenous inoculation of IB4 blocked
the accumulation of leukocytes in the subarachnoid space
despite intracisternal challenge with H. influenzae type b, N.
meningitidis, pneumococcal cell wall, or LPS (170). In
addition, the parameters of BBB permeability, development
of cerebral edema, and death were prevented in the animals
challenged with lethal doses of S. pneumoniae that also
received the monoclonal antibody. In a second study, IB4
and dexamethasone were administered together in a rabbit
model of H. influenzae type b meningitis. The result was a
marked attenuation of all indices of meningeal inflammation
and a reduction in water accumulation in the brain compared
with the results obtained when each agent was given alone
and when infected animals were left untreated (124). Despite
this profound amelioration of meningeal inflammation, clear-
ance rates of bacteria from CSF to blood and from vascular
compartments were unaffected. These results indicate that
VOL. 6, 1993
25
20
E several recent and relevant clinical studies published in the
C
15
1o
UA.
z
E not in those receiving ceftriaxone. The latter point is impor-
E tant, because cefuroxime has recently been shown to be
-)
co
Time (hours)
FIG. 5. (Upper) Comparative changes in concentrations of
TNF-at in CSF in rabbits that received ceftriaxone (CTX) alone at 6
h (*, K; n = 5), CTX plus dexamethasone (DXM) simultaneously at
6 h (0, 0; n = 6), CTX at 6 h and DXM at 7 h (A, A; n 4) or no
=
treatment (controls) (U, El; n = 5). (Lower) Counts of leukocytes
(WBC) in CSF for the same four groups of rabbits as in the upper
panel. Meningitis was induced at time zero. Reproduced from
reference 89 with permission of The University of Chicago Press.
dual therapy with agents directed against the CSF produc-
tion of cytokines and recruitment of leukocytes into the
subarachnoid space may be associated with improved out-
come in bacterial meningitis.
Clinical Trials
Taken together, the experimental studies described above
indicate that adjunctive dexamethasone therapy diminishes
the CSF inflammatory response and the subsequent patho-
physiologic consequences of this inflammation, specifically,
elevated water content in the brain and elevated CSF pres-
sure. On the basis of these observations, several trials were
undertaken to examine the effects of adjunctive corticoster-
oids on outcomes in patients with bacterial meningitis. Early
clinical trials performed in the 1960s failed to show any
benefit for adjunctive corticosteroids (either methylpred-
nisolone or dexamethasone) (12, 31). However, these trials
were performed before much of the recent information on
PATHOGENESIS OF BACTERIAL MENINGITIS 131
pathogenesis and pathophysiology of bacterial meningitis
had been elucidated, and the studies also suffered from
several flaws (e.g., inadequate corticosteroid dose and/or
control group). Therefore, we concentrate on evaluation of
literature.
One study was a double-blind, placebo-controlled trial of
adjunctive dexamethasone therapy in infants and children
with bacterial meningitis (66). The patients received an
antibiotic (cefuroxime or ceftriaxone) with either dexameth-
asone or placebo. The patients who received an antibiotic
plus dexamethasone became afebrile sooner, had more rapid
normalization of CSF parameters (concentrations of glu-
cose, protein, and lactate), and were significantly less likely
to acquire moderate to severe bilateral sensorineural hearing
loss (15.5 versus 3.3%). In addition, concentrations of IL-1,B
but not TNF-cx in CSF were significantly lower 18 to 36 h
later in patients given adjunctive dexamethasone. However,
these findings were significant only in patients with menin-
gitis caused by H. influenzae type b, and the benefits in
terms of morbidity (sensorineural hearing loss) were statis-
tically significant only in patients receiving cefuroxime and
inferior to ceftriaxone in a randomized prospective study of
the therapy of childhood bacterial meningitis (129). In addi-
tion, four patients who received adjunctive dexamethasone
developed gastrointestinal hemorrhage, and two of these
patients required blood transfusions.
A second study, from Egypt, of children and adults with
bacterial meningitis demonstrated a significant reduction in
mortality rate and overall neurologic sequelae in patients
with pneumococcal meningitis who received adjunctive
dexamethasone therapy concomitant with antibiotics (ampi-
cillin plus chloramphenicol) (48). However, no significant
differences between groups in time to afebrility or improve-
ment in CSF parameters were observed; furthermore, the
antibiotics were given intramuscularly, there was no docu-
mentation of possible adverse effects, and an extraordinarily
high percentage of patients presented in a comatose state. In
fact, most patients (370 of 429) had received inadequate
therapy for the 3 to 5 days prior to hospitalization. There
were no differences between the two groups in mortality rate
or rate of hearing impairment for patients with meningococ-
cal or H. influenzae meningitis, but children with H. influ-
enzae meningitis were too young to be tested audiometri-
cally.
In a third recently published trial centered in Costa Rica
(99), infants and children with bacterial meningitis were
randomized in a placebo-controlled, double-blind fashion to
receive cefotaxime with or without adjunctive dexametha-
sone therapy; in this study, the dexamethasone was admin-
istered 15 to 20 min before the first dose of cefotaxime in an
effort to attenuate the CSF inflammatory response associ-
ated with administration of bacteriolytic antibiotics. Twelve
hours after treatment was begun, meningeal inflammation
and concentrations of TNF-a and platelet-activating factor
in CSF had decreased more rapidly in dexamethasone-
treated patients. In addition, by 24 h, the clinical conditions
and mean prognostic scores were significantly better among
patients receiving adjunctive dexamethasone therapy. When
the patients were monitored for a mean of 15 months, those
who had received adjunctive dexamethasone had a signifi-
cantly decreased incidence of one or more neurologic se-
quelae, although reduction of audiologic impairment was
132 TUNKEL AND SCHELD
only a trend. Overall mortality was not reduced in the
dexamethasone group.
Despite the studies described above, controversy regard-
ing the routine use of adjunctive dexamethasone therapy in
all patients with bacterial meningitis remains (75, 165). The
data support the use of adjunctive dexamethasone in infants
and children with bacterial meningitis caused by H. influen-
zae type b. Routine use of dexamethasone therapy for
adults, however, cannot be presently recommended, pend-
ing results of ongoing studies. If dexamethasone is used, the
timing of administration is crucial. Administration before or
with antibiotics is optimal for attenuating the subarachnoid
space inflammatory response. Patients receiving such ther-
apy also need careful monitoring (in addition to the usual
routine for critically ill patients) for the possibility of gas-
trointestinal hemorrhage. Future studies of the pathogenesis
and pathophysiology of bacterial meningitis should lead to
the development of other adjunctive treatment strategies
that may improve the outcome of this disorder.
REFERENCES
1. Andersen, B. M., and 0. Solberg. 1984. Endotoxin liberation
and invasivity of Neisseria meningitidis. Scand. J. Infect. Dis.
16:247-254.
2. Anderson, P., R. B. Johnston, Jr., and D. H. Smith. 1972.
Human serum activities against Haemophilus influenzae type
b. J. Clin. Invest. 51:31-38.
3. Anderson, P. A., A. Flesher, S. Shaw, A. L. Harding, and D. H.
Smith. 1980. Phenotypic and genetic variation in the suscepti-
bility of Haemophilus influenzae type b to antibodies to
somatic antigens. J. Clin. Invest. 65:885-891.
4. Andersson, B., B. Eriksson, E. Falsen, A. Fogh, L. A. Hanson,
0. Nylen, H. Peterson, and C. Svanborg-Eden. 1981. Adhesion
of Streptococcus pneumoniae to human pharyngeal epithelial
cells in vitro: differences in adhesive capacity among strains
isolated from subjects with otitis media, septicemia, or menin-
gitis or from healthy carriers. Infect. Immun. 32:311-317.
5. Arditi, M., L. Ables, and R. Yogev. 1989. Cerebrospinal fluid
endotoxin levels in children with H. influenzae meningitis
before and after administration of intravenous ceftriaxone. J.
Infect. Dis. 160:1005-1011.
6. Arditi, M., K. R. Manogue, M. Caplan, and R. Yogev. 1990.
Cerebrospinal fluid cachectin/tumor necrosis factor-a and
platelet-activating factor concentrations and severity of bacte-
rial meningitis in children. J. Infect. Dis. 162:139-147.
7. Ashwal, S., W. Stringer, L. Tomasi, S. Schneider, J. Thompson,
and R. Perkin. 1990. Cerebral blood flow and carbon dioxide
reactivity in children with bacterial meningitis. J. Pediatr.
117:523-530.
8. Ashwal, S., L. Tomasi, S. Schneider, R. Perkin, and J. Thomp-
son. 1992. Bacterial meningitis in children: pathophysiology
and treatment. Neurology 42:739-748.
9. Austrian, R., and J. Gold. 1964. Pneumococcal bacteremia
with special reference to bacteremic pneumococcal pneumo-
nia. Ann. Intern. Med. 60:759-776.
10. Beachey, E. H. 1981. Bacterial adherence: adhesin-receptor
interactions mediating the attachment of bacteria to mucosal
surfaces. J. Infect. Dis. 143:325-345.
11. Beilke, M. A. 1989. Vascular endothelium in immunology and
infectious disease. Rev. Infect. Dis. 11:273-283.
12. Belsey, M. A., C. W. Hoffpauir, and M. H. D. Smith. 1969.
Dexamethasone in the treatment of acute bacterial meningitis:
the effect of study design on the interpretation of results.
Pediatrics 44:503-513.
13. Bernhardt, L. L., M. S. Simberkoff, and J. J. Rahal, Jr. 1981.
Deficient cerebrospinal fluid opsonization in experimental Esch-
erichia coli meningitis. Infect. Immun. 32:411-413.
14. Beutler, B., N. Krochin, I. W. Milsark, C. Leudke, and A.
Cerami. 1986. Control of cachectin (tumor necrosis factor)
synthesis: mechanisms of endotoxin resistance. Science 232:
977-980.
CLIN. MICROBIOL. REV.
15. Bevilacqua, M. P., J. S. Pober, M. E. Wheeler, R. S. Cotran,
and M. A. Gimbrone, Jr. 1985. Interleukin 1 acts on cultured
human vascular endothelium to increase the adhesion of poly-
morphonuclear leukocytes, monocytes, and related leukocyte
cell lines. J. Clin. Invest. 76:2003-2011.
16. Bevilacqua, M. P., S. Stengelin, M. A. Gimbrone, Jr., and B.
Seed. 1989. Endothelial leukocyte adhesion molecule 1: an
inducible receptor for neutrophils related to complement reg-
ulatory proteins and lectins. Science 243:1160-1165.
17. Bohr, V., 0. Paulson, and N. Rasmussen. 1984. Pneumococcal
meningitis: late neurologic sequelae and features of prognostic
impact. Arch. Neurol. 41:1045-1049.
18. Bradbury, M. W. 1984. The structure and function of the
blood-brain barrier. Fed. Proc. 43:186-190.
19. Brandtzaeg, P., T. E. Mollnes, and P. Kierulf. 1989. Comple-
ment activation and endotoxin levels in systemic meningococ-
cal disease. J. Infect. Dis. 160:58-65.
20. Broome, C. V., R. R. Facklam, J. R. Allen, D. W. Fraser, and
R. Austrian. 1980. Epidemiology of pneumococcal serotypes in
the United States. J. Infect. Dis. 141:119-123.
21. Brown, E. J., S. W. Hosea, and M. M. Frank. 1981. The role of
the spleen in experimental pneumococcal bacteremia. J. Clin.
Invest. 67:975-982.
22. Brown, E. J., S. W. Hosea, C. H. Hammer, C. G. Burch, and
M. M. Frank. 1982. A quantitative analysis of the interaction of
antipneumococcal antibody and complement in experimental
pneumococcal bacteremia. J. Clin. Invest. 69:85-98.
23. Brown, E. J., K. A. Joiner, R. M. Cole, and M. Berger. 1983.
Localization of complement component 3 on Streptococcus
pneumoniae: anti-capsular antibody causes complement depo-
sition on the pneumococcal capsule. Infect. Immun. 39:403-
409.
24. Bryan, J. P., H. R. de Silva, A. Tavares, H. Roche, and W. M.
Scheld. 1990. Etiology and mortality of bacterial meningitis in
Northeastern Brazil. Rev. Infect. Dis. 12:128-135.
25. Buchanen, N., and G. S. McNab. 1972. Cerebrospinal fluid
complement and immunoglobulins in meningitis and encepha-
litis. S. Afr. Med. J. 46:1376-1382.
26. Chan, P. H., and R. A. Fishman. 1979. Brain edema: induction
in cortical slices by polyunsaturated fatty acids. Science 201:
358-360.
27. Corral, C. J., J. A. Winkelstein, and E. R. Moxon. 1982.
Participation of complement in host defense against encapsu-
lated Haemophilus influenzae types a, c, and d. Infect. Immun.
35:759-763.
28. Cross, A. S., P. Gemski, J. C. Sadoff, F. Orskov, and I. Orskov.
1984. The importance of the Kl capsule in invasive infections
caused by Eschenichia coli. J. Infect. Dis. 149:184-193.
29. Crosson, F. A., Jr., J. A. Winkelstein, and E. R. Moxon. 1976.
Participation of complement in the nonimmune host defense
against experimental Haemophilus influenzae bacteremia and
meningitis. Infect. Immun. 14:882-887.
30. Dacey, R. G., Jr., and M. A. Sande. 1974. Effect of probenecid
on cerebrospinal fluid concentrations of penicillin and cepha-
losporin derivatives. Antimicrob. Agents Chemother. 6:437-
444.
31. deLemos, R. A., and R. J. Haggerty. 1969. Corticosteroids as
an adjunctive to treatment in bacterial meningitis. A controlled
clinical trial. Pediatrics 44:30-34.
32. Devoe, I. W., and J. E. Gilchrist. 1975. Pili on meningococci
from primary cultures of nasopharyngeal carriers and cerebro-
spinal fluid of patients with acute disease. J. Exp. Med.
141:297-305.
33. Dodge, P. R., H. Davis, R. Feigin, S. J. Holmes, S. L. Kaplan,
D. P. Jubelirer, B. W. Stechenberg, and S. K. Hirsh. 1984.
Prospective evaluation of hearing impairment as a sequela of
acute bacterial meningitis. N. Engl. J. Med. 311:869-874.
34. Ernst, J. D., J. M. Decazes, and M. A. Sande. 1983. Experi-
mental pneumococcal meningitis: role of leukocytes in patho-
genesis. Infect. Immun. 41:275-279.
35. Ernst, J. D., K. T. Hartiala, I. M. Goldstein, and M. A. Sande.
1984. Complement (C5)-derived chemotactic activity accounts
for accumulation of polymorphonuclear leukocytes in cerebro-
VOL. 6, 1993
spinal fluid of rabbits with pneumococcal meningitis. Infect.
Immun. 46:81-86.
36. Feldman, W. E. 1976. Concentrations of bacteria in cerebro-
spinal fluid of patients with bacterial meningitis. J. Pediatr.
88:549-552.
37. Feldman, W. E. 1977. Relation of concentrations of bacteria
and bacterial antigens in cerebrospinal fluid to prognosis in
patients with bacterial meningitis. N. Engl. J. Med. 296:433-
435.
38. Fine, D. P. 1975. Pneumococcal type-associated variability in
alternate complement pathway activation. Infect. Immun. 12:
772-778.
39. Finland, M. 1973. Excursions into epidemiology: selected
studies during the past four decades at Boston City Hospital. J.
Infect. Dis. 128:76-124.
40. Fishman, R. A. 1975. Brain edema. N. Engl. J. Med. 293:706-
711.
41. Fishman, R. A., K. Sligar, and R. B. Hake. 1977. Effects of
leukocytes on brain metabolism in granulocytic brain edema.
Ann. Neurol. 2:89-94.
42. Flexner, S. 1913. The results of serum treatment in 1300 cases
of epidemic meningitis. J. Exp. Med. 17:553-576.
43. Fontana, A., F. Kristensen, R. Dubs, D. Gemsa, and E. Weber.
1982. Production of prostaglandin E and an interleukin-1 like
factor by cultured astrocytes and C6 glioma cells. J. Immunol.
129:2413-2419.
44. Fraser, D. W., C. C. Geil, and R. A. Feldman. 1974. Bacterial
meningitis in Bernalillo County, New Mexico: a comparison
with three other American populations. Am. J. Epidemiol.
100:29-34.
45. Freter, R. 1980. Prospects for preventing the association of
harmful bacteria with host mucosal surfaces, p. 441-459. In
E. H. Beachey (ed.), Bacterial adherence, vol. 6. Chapman
and Hall, New York.
46. Giampaolo, C., W. M. Scheld, J. Boyd, J. Savory, M. Sande,
and M. Wills. 1981. Leukocyte and bacterial interrelationships
in experimental meningitis. Ann. Neurol. 9:328-333.
47. Gigliotti, F., D. Lee, R. A. Insel, and W. M. Scheld. 1987. IgG
penetration into the cerebrospinal fluid in a rabbit model of
meningitis. J. Infect. Dis. 156:394-398.
48. Girgis, N. I., Z. Farid, I. A. Mikhail, I. Farrag, Y. Sultan, and
M. E. Kilpatrick 1989. Dexamethasone treatment for bacterial
meningitis in children and adults. Pediatr. Infect. Dis. J.
8:848-851.
49. Goldstein, G. W., and A. L. Betz. 1986. The blood-brain
barrier. Sci. Am. 255:74-83.
50. Gray, B. M., G. M. Converse III, and H. C. Dillon, Jr. 1979.
Serotypes of Streptococcus pneumoniae causing invasive dis-
ease. J. Infect. Dis. 140:979-983.
51. Greenblatt, J. J., K. Floyd, M. W. Philipps, and C. E. Frasch.
1988. Morphologic differences in Neisseria meningitidis pili.
Infect. Immun. 56:2356-2362.
52. Greenwood, B. M. 1978. Chemotactic activity of cerebrospinal
fluid in pyogenic meningitis. J. Clin. Pathol. 31:213-216.
53. Griffis, J. M. 1975. Bactericidal activity of IgA of lytic antibody
in human convalescent sera. J. Immunol. 114:1779-1784.
54. Griffis, J. M., and M. A. Bertram. 1977. Immunoepidemiology
of meningococcal disease in military recruits. II. Blocking of
serum bactericidal activity by circulating IgA early in the
course of invasive disease. J. Infect. Dis. 136:733-739.
55. Gulig, P. A., G. H. McCracken, Jr., C. F. Frisch, K. H.
Johnston, and E. J. Hansen. 1982. Antibody response of infants
to cell surface-exposed outer membrane proteins of Hae-
mophilus influenzae type b after systemic Haemophilus dis-
ease. Infect. Immun. 37:82-88.
56. Hansen, E. J., C. F. Frisch, R. L. McDade, Jr., and K. H.
Johnston. 1982. Identification of immunogenic outer membrane
proteins of Haemophilus influenzae type b in the infant rat
model system. Infect. Immun. 32:1084-1092.
57. Horwitz, S. J., B. Boxerbaum, and J. O'Bell. 1980. Cerebral
herniation in bacterial meningitis in childhood. Ann. Neurol.
7:524-528.
58. Houssiau, F. A., K. Bukasa, C. J. M. Sindic, J. Van Damme,
PATHOGENESIS OF BACTERIAL MENINGITIS 133
and J. Van Snick. 1988. Elevated levels of the 26K human
hybridoma growth factor (interleukin 6) in cerebrospinal fluid
of patients with acute infection of the central nervous system.
Clin. Exp. Immunol. 71:320-323.
59. Igarashi, M., R. C. Gilmartin, B. Gerald, F. Wilburn, and J. T.
Jabbour. 1984. Cerebral arteritis and bacterial meningitis.
Arch. Neurol. 41:531-535.
60. Kadurugamuwa, J. L., B. Hengstler, M. A. Bray, and 0. Zak.
1989. Inhibition of complement-factor-5a-induced inflamma-
tory reactions by prostaglandin E2 in experimental meningitis.
J. Infect. Dis. 160:715-719.
61. Kadurugamuwa, J. L., B. Hengstler, and 0. Zak. 1989. Cere-
brospinal fluid protein profile in experimental pneumococcal
meningitis and its alteration by ampicillin and anti-inflamma-
tory agents. J. Infect. Dis. 159:26-34.
62. Kaplan, S. L., and R. D. Feigin. 1978. The syndrome of
inappropriate secretion of antidiuretic hormone in children
with bacterial meningitis. J. Pediatr. 92:758-761.
63. Kaplan, S. L., E. P. Hawkins, M. W. Kline, G. S. Patrick, and
E. 0. Mason, Jr. 1989. Invasion of the inner ear by Haemo-
philus influenzae type b in experimental meningitis. J. Infect.
Dis. 159:923-930.
64. Kayhty, H., H. Jousimies-Somer, H. Peltola, and P. H. Makela.
1981. Antibody response to capsular polysaccharides of groups
A and C Neisseria meningitidis and Haemophilus influenzae
type b during bacteremic disease. J. Infect. Dis. 143:32-41.
65. Lam, J. S., D. M. Granoff, J. R. Gilsdorf, and J. W. Costerton.
1980. Immunogenicity of outer membrane derivatives of Hae-
mophilus influenzae type b. Curr. Microbiol. 3:359-364.
66. Lebel, M. H., B. J. Frei, G. A. Syrogiannopoulos, D. F.
Chrane, M. J. Hoyt, S. M. Stewart, B. D. Kennard, K. D.
Olsen, and G. H. McCracken, Jr. 1988. Dexamethasone ther-
apy for bacterial meningitis. Results of two double-blind,
placebo-controlled trials. N. Engl. J. Med. 319:964-971.
67. Leist, T. P., K. Frei, S. Kam-Hansen, R. M. Zinkernagel, and
A. Fontana. 1988. Tumor necrosis factor ao in cerebrospinal
fluid during bacterial, but not viral, meningitis. Evaluation in
murine model infections and in patients. J. Exp. Med. 167:
1743-1748.
68. Lesse, A. J., E. R. Moxon, A. Zwahlen, and W. M. Scheld.
1988. Role of cerebrospinal fluid pleocytosis and Haemophilus
influenzae type b capsule on blood brain barrier permeability
during experimental meningitis in the rat. J. Clin. Invest.
82:102-109.
69. Libby, P., J. M. Ordovas, K. R. Auger, A. H. Robbins, L. K.
Birinyi, and C. A. Dinarello. 1986. Endotoxin and tumor
necrosis factor induce interleukin-1 gene expression in adult
human vascular endothelial cells. Am. J. Pathol. 124:179-185.
70. LiPuma, J. J., H. Richman, and T. L. Stull. 1990. Haemocin,
the bacteriocin produced by Haemophilus influenzae: species
distribution and role in colonization. Infect. Immun. 58:1600-
1605.
71. Lundberg, C., J. Lonnroth, and C. E. Nord. 1982. Adherence
in the colonization of Streptococcus pneumoniae in the naso-
pharynx of children. Infection 10:63-69.
72. Mason, E. O., S. L. Kaplan, B. L. Wiedermann, E. P. Norrod,
and W. A. Stenback. 1985. Frequency and properties of natu-
rally occurring adherent piliated strains of Haemophilus influ-
enzae type b. Infect. Immun. 49:98-103.
73. McCracken, G. H., Jr., M. M. Mustafa, 0. Ramilo, K. D.
Olsen, and R. C. Risser. 1989. Cerebrospinal fluid interleukin
1-beta and tumor necrosis factor concentrations and outcome
from neonatal gram-negative enteric bacillary meningitis. Pe-
diatr. Infect. Dis. J. 8:155-159.
74. McGee, Z. A., D. S. Stephens, L. H. Hoffman, W. F. Schlech
III, and R. G. Horn. 1983. Mechanisms of mucosal invasion by
pathogenic Neisseria. Rev. Infect. Dis. 5(Suppl. 4):S708-S714.
75. McGowan, J. E., Jr., P. J. Chesney, K. B. Crossley, and F. M.
LaForce. 1992. Guidelines for the use of systemic glucocorti-
coids in the management of selected infections. J. Infect. Dis.
165:1-13.
76. McMenamin, J. B., and J. J. Volpe. 1984. Bacterial meningitis
in infancy: effects on intracranial pressure and cerebral blood
134 TUNKEL AND SCHELD
flow velocity. Neurology 34:500-504.
77. Meyrick, B., R. Hoover, M. R. Jones, L. C. Berry, Jr., and
K. L. Brigham. 1989. In vitro effects of endotoxin on bovine
and sheep lung microvascular endothelial cells. J. Cell. Phys-
iol. 138:165-174.
78. Miossec, P., D. Cavender, and M. Ziff. 1986. Production of
interleukin 1 by human endothelial cells. J. Immunol. 136:
2486-2491.
79. Moller, B., S. C. Mogensen, P. Wendelboe, K. Bendtzen, and
C. M. Petersen. 1991. Bioactive and inactive forms of tumor
necrosis factor-a in spinal fluid from patients with meningitis.
J. Infect. Dis. 163:886-889.
80. Movat, H. Z. 1987. Tumor necrosis factor and interleukin-1:
role in acute inflammation and microvascular injury. J. Lab.
Clin. Med. 110:668-681.
81. Moxon, E. R., and P. Anderson. 1979. Meningitis caused by
Haemophilus influenzae in infant rats: protective immunity
and antibody priming by gastrointestinal colonization with
Escherichia coli. J. Infect. Dis. 140:471-478.
82. Moxon, E. R., M. P. Glode, A. Sutton, and J. B. Robbins. 1977.
The infant rat as a model of bacterial meningitis. J. Infect. Dis.
136(Suppl.):S186-S190.
83. Moxon, E. R., and P. T. Ostrow. 1977. Haemophilus influenzae
bacteremia in infant rats: role of bacteremia in pathogenesis of
age dependent inflammatory response in cerebrospinal fluid. J.
Infect. Dis. 135:303-307.
84. Moxon, E. R., A. L. Smith, D. R. Averill, and D. H. Smith.
1974. Haemophilus influenzae meningitis in infant rats after
intranasal inoculation. J. Infect. Dis. 129:154-162.
85. Moxon, E. R., and K. A. Vaughn. 1981. The type b capsular
polysaccharide as a virulence determinant of Haemophilus
influenzae: studies using clinical isolates and laboratory trans-
formants. J. Infect. Dis. 143:517-524.
86. Mulks, M. H., S. J. Kornfeld, B. Frangione, and A. G. Plaut.
1982. Relationship between the specificity of IgA proteases and
serotypes in Haemophilus influenzae. J. Infect. Dis. 146:266-
274.
87. Mustafa, M. M., M. H. Lebel, 0. Ramilo, K. D. Olsen, J. S.
Reisch, B. Beutler, and G. H. McCracken, Jr. 1989. Correlation
of interleukin-1l3 and cachectin concentrations in cerebrospinal
fluid and outcome from bacterial meningitis. J. Pediatr. 115:
208-213.
88. Mustafa, M. M., J. Mertsola, 0. Ramilo, X. Saez-Llorens, R. C.
Risser, and G. H. McCracken, Jr. 1989. Increased endotoxin
and interleukin-1lB concentrations in cerebrospinal fluid of
infants with coliform meningitis and ventriculitis associated
with intraventricular gentamicin therapy. J. Infect. Dis. 160:
891-895.
89. Mustafa, M. M., 0. Ramilo, J. Mertsola, R. C. Risser, B.
Beutler, E. J. Hansen, and G. H. McCracken, Jr. 1989. Mod-
ulation of inflammation and cachectin activity in relation to
treatment of experimental Haemophilus influenzae type b
meningitis. J. Infect. Dis. 160:818-825.
90. Mustafa, M. M., 0. Ramilo, K. D. Olsen, P. S. Franklin, E. J.
Hansen, B. Beutler, and G. H. McCracken, Jr. 1989. Tumor
necrosis factor in mediating experimental Haemophilus influ-
enzae type b meningitis. J. Clin. Invest. 84:1253-1259.
91. Mustafa, M. M., 0. Ramilo, X. Saez-Llorens, K. D. Olsen,
R. R. Magness, and G. H. McCracken, Jr. 1990. Cerebrospinal
fluid prostaglandins, interleukin 1l, and tumor necrosis factor
in bacterial meningitis. Clinical and laboratory correlations in
placebo-treated and dexamethasone-treated patients. Am. J.
Dis. Child. 144:883-887.
92. Mustafa, M. M., 0. Ramilo, G. A. Syrogiannopoulos, K. D.
Olsen, G. H. McCracken, Jr., and E. J. Hansen. 1989. Induc-
tion of meningeal inflammation by outer membrane vesicles of
Haemophilus influenzae type b. J. Infect. Dis. 159:917-922.
93. Nadal, D., D. Leppert, K. Frei, P. Gallo, H. Lamche, and A.
Fontana. 1989. Tumor necrosis factor-ax in infectious meningi-
tis. Arch. Dis. Child. 64:1274-1279.
94. Nawroth, P. P., I. Bank, D. Handley, J. Cassimeris, L. Chess,
and D. Stern. 1986. Tumor necrosis factor/cachectin interacts
with endothelial cell receptors to induce release of interleukin
CLIN. MICROBIOL. REV.
1. J. Exp. Med. 163:1363-1375.
95. Nichols, F. C., S. W. Garrison, and H. W. Davis. 1988.
Prostaglandin E2 and thromboxane B2 release from human
monocytes treated with bacterial lipopolysaccharide. J. Leu-
kocyte Biol. 44:376-384.
96. Nolan, C. M., R. A. Clark, and H. N. Beaty. 1975. Experimen-
tal pneumococcal meningitis. III. Chemotactic activity in ce-
rebrospinal fluid. Proc. Soc. Exp. Biol. Med. 150:134-136.
97. Nolan, C. M., C. K. McAllister, E. Walters, and H. N. Beaty.
1978. Experimental pneumococcal meningitis. IV. The effect
of methylprednisolone on meningeal inflammation. J. Lab.
Clin. Med. 91:979-988.
98. Nugent, S. K., J. A. Bausher, E. R. Moxon, and M. C. Rogers.
1979. Raised intracranial pressure: its management in Neisse-
ria meningitidis meningoencephalitis. Am. J. Dis. Child. 133:
260-262.
99. Odio, C. M., I. Faingezicht, M. Paris, M. Nassar, A. Baltodano,
J. Rogers, X. Saez-Llorens, K. D. Olsen, and G. H. McCracken,
Jr. 1991. The beneficial effects of early dexamethasone admin-
istration in infants and children with bacterial meningitis. N.
Engl. J. Med. 324:1525-1531.
100. Ostrow, P. T., E. R. Moxon, N. Vernon, and R. Kapko. 1979.
Pathogenesis of bacterial meningitis. Studies on the route of
meningeal invasion following Haemophilus influenzae inocula-
tion in infant rats. Lab. Invest. 40:678-685.
101. Parkkinen, J., T. K. Korhonen, A. Pere, J. Hacker, and S.
Soinila. 1988. Binding sites in the rat brain for Escherichia coli
S fimbriae associated with neonatal meningitis. J. Clin. Invest.
81:860-865.
102. Patrick, D., J. Betts, E. A. Frey, R. Prameya, K. Dorovini-Zis,
and B. B. Finlay. 1992. Haemophilus influenzae lipopolysac-
charide disrupts confluent monolayers of bovine brain endo-
thelial cells via a serum-dependent cytotoxic pathway. J.
Infect. Dis. 165:865-872.
103. Pearson, H. 1977. Sickle cell anemia and severe infections due
to encapsulated bacteria. J. Infect. Dis. 136:S25-S30.
104. Petersdorf, R. G., and C. N. Luttrell. 1962. Studies on the
pathogenesis of meningitis. I. Intrathecal injection. J. Clin.
Invest. 41:311-319.
105. Pfister, H. W., U. Koedel, R. L. Haberl, U. Dinargl, W. Feiden,
G. Ruckdeschel, and K. M. Einhaupl. 1990. Microvascular
changes during the early phase of experimental bacterial men-
ingitis. J. Cerebral Blood Flow Metab. 10:914-922.
106. Pichichero, M. E., M. Loeb, P. Anderson, and D. H. Smith.
1982. Do pili play a role in pathogenicity of Haemophilus
influenzae type b? Lancet ii:960-962.
107. Pike, M. G., P. K. H. Wong, R. Bencivenga, 0. Flodmark,
D. A. Cabral, D. P. Speert, and K. Farrell. 1990. Electrophys-
iologic studies, computed tomography, and neurologic out-
come in acute bacterial meningitis. J. Pediatr. 116:702-706.
108. Plaut, A. G. 1983. The IgAl proteases of pathogenic bacteria.
Annu. Rev. Microbiol. 37:603-622.
109. Pomeroy, S. L., S. J. Holmes, P. R. Dodge, and R. D. Feigin.
1990. Seizures and other neurologic sequelae of bacterial
meningitis in children. N. Engl. J. Med. 323:1651-1657.
110. Quagliarello, V. J., W. J. Long, Jr., and W. M. Scheld. 1986.
Morphologic alterations of the blood-brain barrier with exper-
imental meningitis in the rat. Temporal sequence and role of
encapsulation. J. Clin. Invest. 77:1084-1095.
111. Quagliarello, V. J., A. Ma, H. Stukenbrok, and G. E. Palade.
1991. Ultrastructural localization of albumin transport across
the cerebral microvasculature during experimental meningitis
in the rat. J. Exp. Med. 174:657-672.
112. Quagliarello, V. J., B. Wispelwey, W. J. Long, Jr., and W. M.
Scheld. 1991. Recombinant human interleukin-1 induces men-
ingitis and blood-brain barrier injury in the rat. Characteriza-
tion and comparison with tumor necrosis factor. J. Clin.
Invest. 87:1360-1366.
113. Quinn, P. H., F. J. Crosson, Jr., J. A. Winkelstein, and E. R.
Moxon. 1977. Activation of the alternative complement path-
way by Haemophilus influenzae type b. Infect. Immun. 16:
400-402.
114. Raff, H. V., D. Devereux, W. Schuford, D. Abbott-Brown, and
VOL. 6, 1993
G. Maloney. 1988. Human monoclonal antibody with protec-
tive activity for Escherichia coli Kl and Neisseria meningitidis
group B infections. J. Infect. Dis. 157:118-126.
115. Rahal, J. J., and M. S. Simberkoff. 1982. Host defense and
antimicrobial therapy in adult gram-negative bacillary menin-
gitis. Ann. Intern. Med. 96:468-474.
116. Raimondi, A. J., and C. DiRocco. 1979. The physiopathoge-
netic basis for the angiographic diagnosis of bacterial infection
of the brain and its coverings in children. I. Leptomeningitis.
Child's Brain 5:398-413.
117. Ramilo, O., X. Saez-Llorens, J. Mertsola, H. Jafari, K. D.
Olsen, E. J. Hansen, M. Yoshinaga, S. Ohkawara, H. Nariuchi,
and G. H. McCracken, Jr. 1990. Tumor necrosis factor
a/cachectin and interleukin 11 initiate meningeal inflammation.
J. Exp. Med. 172:497-507.
118. Riesenfeld-Orn, I., S. Wolpe, J. F. Garcia-Bustos, M. K.
Hoffmann, and E. Tuomanen. 1989. Production of interleukin-1
but not tumor necrosis factor by human monocytes stimulated
with pneumococcal cell surface components. Infect. Immun.
57:1890-1893.
119. Robbins, J. B., G. H. McCracken, Jr., E. L. Gotschich, F.
Orskov, I. Orskov, and L. A. Hanson. 1974. Eschenchia coli
Kl capsular polysaccharide associated with neonatal meningi-
tis. N. Engl. J. Med. 290:1216-1220.
120. Roberts, M., T. L. Stull, and A. L. Smith. 1981. Comparative
virulence of Haemophilus influenzae with a type b or type d
capsule. Infect. Immun. 32:518-524.
121. Rodriguez, A. F., S. L. Kaplan, E. P. Hawkins, and E. 0.
Mason, Jr. 1991. Hematogenous pneumococcal meningitis in
the infant rat: description of a model. J. Infect. Dis. 164:1207-
1209.
122. Ross, S. C., and P. Densen. 1984. Complement deficiency states
and infection: epidemiology, pathogenesis and consequences
of neisserial and other infections in an immune deficiency.
Medicine (Baltimore) 63:243-273.
123. Rusconi, F., F. Parizzi, L. Garlaschi, B. M. Assael, M. Sironi,
P. Ghezzi, A. Mantovani,. and the Collaborative Study on
Meningitis. 1991. Interleukin 6 activity in infants and children
with bacterial meningitis. Pediatr. Infect. Dis. J. 10:117-121.
124. Saez-Llorens, X., H. S. Jafari, C. Severien, F. Parras, K. D.
Olsen, E. J. Hansen, I. I. Singer, and G. H. McCracken, Jr.
1991. Enhanced attenuation of meningeal inflammation and
brain edema by concomitant administration of anti-CD18 mono-
clonal antibodies and dexamethasone in experimental Haemo-
philus meningitis. J. Clin. Invest. 88:2003-2011.
125. Saez-Llorens, X., 0. Ramilo, M. M. Mustafa, J. Mertsola, C. de
Alba, E. Hansen, and G. H. McCracken, Jr. 1990. Pentoxifyl-
line modulates meningeal inflammation in experimental bacte-
rial meningitis. Antimicrob. Agents Chemother. 34:837-843.
126. Saez-Llorens, X., 0. Ramilo, M. M. Mustafa, J. Mertsola, and
G. H. McCracken, Jr. 1990. Molecular pathophysiology of
bacterial meningitis: current concepts and therapeutic implica-
tions. J. Pediatr. 116:671-684.
127. Saukkonen, K., S. Sande, C. Cioffe, S. Wolfe, B. Sherry, A.
Cerami, and E. Tuomanen. 1990. The role of cytokines in the
generation of inflammation and tissue damage in experimental
gram-positive meningitis. J. Exp. Med. 171:439-448.
128. Saukkonen, K. M., B. Nowicki, and M. Leinonen. 1988. Role of
type 1 and S fimbriae in the pathogenesis of Escherichia coli
018:K1 bacteremia and meningitis in the infant rat. Infect.
Immun. 56:892-897.
129. Schaad, U. B., S. Suter, A. Gianella-Bonadori, J. Pfenninger, R.
Auckenthaler, 0. Bernath, J. J. Chesaux, and J. Wedgwood.
1990. A comparison of ceftriaxone and cefuroxime for the
treatment of bacterial meningitis in children. N. Engl. J. Med.
322:141-146.
130. Scheifele, D. W., R. S. Daum, V. P. Syriopoulou, D. R. Averill,
and A. L. Smith. 1980. Haemophilus influenzae bacteremia and
meningitis in infant primates. J. Lab. Clin. Med. 95:450-462.
131. Scheld, W. M. 1984. Bacterial meningitis in the patient at risk:
intrinsic risk factors and host defense mechanisms. Am. J.
Med. 76:(Suppl. SA):193-207.
132. Scheld, W. M., and J. P. Brodeur. 1983. Effect of methylpred-
PATHOGENESIS OF BACTERIAL MENINGITIS 135
nisolone on entry of ampicillin and gentamicin into cerebrospi-
nal fluid in experimental pneumococcal and Escherichia coli
meningitis. Antimicrob. Agents Chemother. 23:108-112.
133. Scheld, W. M., R. G. Dacey, H. R. Winn, J. E. Welsh, J. A.
Jane, and M. A. Sande. 1980. Cerebrospinal fluid outflow
resistance in rabbits with experimental meningitis. Alterations
with penicillin and methylprednisolone. J. Clin. Invest. 66:243-
253.
134. Scheld, W. M., C. Giampaolo, J. Boyd, J. Savory, M. R. Wills,
and M. A. Sande. 1982. Cerebrospinal fluid prognostic indices
in experimental pneumococcal meningitis. J. Lab. Clin. Med.
100:218-229.
135. Scheld, W. M., T. S. Park, R. G. Dacey, H. R. Winn, J. A.
Jane, and M. A. Sande. 1979. Clearance of bacteria from
cerebrospinal fluid to blood in experimental meningitis. Infect.
Immun. 24:102-105.
136. Schlech, W. F., III, J. I. Ward, J. D. Band, A. Hightower,
D. W. Fraser, and C. V. Broome. 1985. Bacterial meningitis in
the United States, 1978 through 1981. The national bacterial
meningitis surveillance study. JAMA 253:1749-1754.
137. Schleimer, R. P., and B. K. Rutledge. 1986. Cultured human
vascular endothelial cells acquire adhesiveness for neutrophils
after stimulation with interleukin 1, endotoxin, and tumor-
promoting phorbol diesters. J. Immunol. 136:649-654.
138. Schneerson, R., and J. B. Robbins. 1975. Induction of serum
Haemophilus influenzae type b capsular antibodies in adult
volunteers fed cross-reacting Escherichia coli 075:K100:H5.
N. Engl. J. Med. 292:1093-1096.
139. Simberkoff, M. S., H. N. Moldover, and J. J. Rahal, Jr. 1980.
Absence of detectable bactericidal and opsonic activities in
normal and infected human cerebrospinal fluids. A regional
host defense deficiency. J. Lab. Clin. Med. 95:362-372.
140. Smith, A. L. 1987. Pathogenesis of Haemophilus influenzae
meningitis. Pediatr. Infect. Dis. J. 6:783-786.
141. Smith, A. L., R. S. Daum, D. Scheifele, V. Syriopolou, D. R.
Averill, M. C. Roberts, and T. L. Stull. 1982. Pathogenesis of
Haemophilus influenzae meningitis, p. 89-109. In S. H. Sell
and P. F. Wright (ed.), Haemophilus influenzae: epidemiology,
immunology, and prevention of disease. Elsevier Science
Publishing Co., New York.
142. Smith, H., B. Bannister, and M. J. O'Shea. 1973. Cerebrospi-
nal-fluid immunoglobulins in meningitis. Lancet i:591-593.
143. Stephens, D. S., and M. M. Farley. 1991. Pathogenic events
during infection of the human nasopharynx with Neisseria
meningitidis and Haemophilus influenzae. Rev. Infect. Dis.
13:22-33.
144. Stephens, D. S., L. H. Hoffman, and Z. A. McGee. 1983.
Interaction of Neisseria meningitidis with human nasopharyn-
geal mucosa: attachment and entry into columnar epithelial
cells. J. Infect. Dis. 148:369-376.
145. Stephens, D. S., and Z. A. McGee. 1981. Attachment of
Neisseria meningitidis to human mucosal surfaces: influence of
pili and type of receptor cell. J. Infect. Dis. 143:525-532.
146. Stull, T. L., P. M. Mendelman, J. L. Haas, M. A. Schoenborn,
K. D. Mack, and A. L. Smith. 1984. Characterization of
Haemophilus influenzae type b fimbriae. Infect. Immun. 46:
787-796.
147. Sullivan, G. W., H. T. Carper, W. J. Novick, Jr., and G. L.
Mandell. 1988. Inhibition of the inflammatory action of inter-
leukin-1 and tumor necrosis factor (alpha) on neutrophil func-
tion by pentoxifylline. Infect. Immun. 56:1722-1729.
148. Swartz, M. N. 1984. Bacterial meningitis. More involved than
just the meninges. N. Engl. J. Med. 311:912-914.
149. Syrogiannopoulos, G. A., E. J. Hansen, A. L. Erwin, R. S.
Munford, J. Rutledge, J. S. Reisch, and G. H. McCracken, Jr.
1988. Haemophilus influenzae type b lipooligosaccharide in-
duces meningeal inflammation. J. Infect. Dis. 157:237-244.
150. Syrogiannopoulos, G. A., K. D. Olsen, J. S. Reisch, and G. H.
McCracken, Jr. 1987. Dexamethasone in the treatment of
experimental Haemophilus influenzae type b meningitis. J.
Infect. Dis. 155:213-219.
151. Takala, A. K., L. van Alphen, J. Eskola, J. Palmgren, P. Bol,
and P. H. Makela. 1987. Haemophilus influenzae type b strains
136 TUNKEL AND SCHELD
of outer membrane subtypes 1 and lc cause different types of
invasive disease. Lancet ii:647-650.
152. Tauber, M. G., U. Borschberg, and M. A. Sande. 1988.
Influence of granulocytes on brain edema, intracranial pres-
sure, and cerebrospinal fluid concentrations of lactate and
protein in experimental meningitis. J. Infect. Dis. 157:456-464.
153. Tauber, M. G., M. Burroughs, U. M. Niemoller, H. Kuster, U.
Borschberg, and E. Tuomanen. 1991. Differences of pathophys-
iology in experimental meningitis caused by three strains of
Streptococcus pneumoniae. J. Infect. Dis. 163:806-811.
154. Tauber, M. G., H. Khayam-Bashi, and M. A. Sande. 1985.
Effects of ampicillin and corticosteroids on brain water con-
tent, cerebrospinal fluid pressure, and cerebrospinal fluid lac-
tate levels in experimental pneumococcal meningitis. J. Infect.
Dis. 151:528-534.
155. Tauber, M. G., A. M. Shibl, C. J. Hackbarth, J. W. Larrick,
and M. A. Sande. 1987. Antibiotic therapy, endotoxin concen-
tration in cerebrospinal fluid, and brain edema in experimental
Escherichia coli meningitis in rabbits. J. Infect. Dis. 156:456-
462.
156. Taylor, H. G., E. L. Mills, A. Ciampi, R. du Berger, G. V.
Watters, R. Gold, N. MacDonald, and R. H. Michaels. 1990.
The sequelae of Haemophilus influenzae meningitis in school-
age children. N. Engl. J. Med. 323:1657-1663.
157. Tofte, R. W., P. K. Peterson, Y. Kim, and P. G. Quie. 1979.
Opsonic activity of normal human cerebrospinal fluid for
selected bacterial species. Infect. Immun. 26:1093-1098.
158. Tonnesen, M. G., L. A. Smedley, and P. M. Henson. 1984.
Neutrophil-endothelial cell interaction. Modulation of neutro-
phil adhesiveness induced by complement fragments C5a and
CSa des arg and formyl-methionyl-leucyl-phenylalanine in
vitro. J. Clin. Invest. 74:1581-1592.
159. Tracey, K. J., S. F. Luwry, and A. Cerami. 1988. Cachectin: a
hormone that triggers acute shock and chronic cachexia. J.
Infect. Dis. 157:413-420.
160. Tunkel, A. R., M. T. Kinter, S. W. Rosser, M. A. Aubin, E. J.
Hansen, and W. M. Scheld. 1990. Haemophilus infiuenzae type
b lipopolysaccharide (LPS)-induced synthesis of prostaglandin
E2 (PGE2) and interleukin-6 (IL-6) by cerebral microvascular
endothelial cells (MECs), abstr. 74, p. 99. Program Abstr. 30th
Intersci. Conf. Antimicrob. Agents Chemother. 1990. Ameri-
can Society for Microbiology, Washington, D.C.
161. Tunkel, A. R., S. W. Rosser, E. J. Hansen, and W. M. Scheld.
1990. Haemophilus influenzae type b lipopolysaccharide
(LPS)-induced formation of cAMP and cGMP in cerebral
microvascular endothelial cells (MECs), abstr. 73, p. 99.
Program Abstr. 30th Intersci. Conf. Antimicrob. Agents
Chemother. 1990. American Society for Microbiology, Wash-
ington, D.C.
162. Tunkel, A. R., S. W. Rosser, E. J. Hansen, and W. M. Scheld.
1991. Blood-brain barrier alterations in bacterial meningitis:
development of an in vitro model and observations on the
effects of lipopolysaccharide. In Vitro Cell. Dev. Biol. 27A:
113-120.
163. Tunkel, A. R., and W. M. Scheld. 1989. Applications of therapy
in animal models to bacterial infection in human disease.
Infect. Dis. Clin. North Am. 3:441-459.
164. Tunkel, A. R., and W. M. Scheld. 1989. Alterations in the
blood-brain barrier in bacterial meningitis: in vivo and in vitro
models. Pediatr. Infect. Dis. J. 8:911-913.
165. Tunkel, A. R., B. Wispelwey, and W. M. Scheld. 1990. Bacte-
rial meningitis: recent advances in pathophysiology and treat-
ment. Ann. Intern. Med. 112:610-623.
166. Tunkel, A. R., B. Wispelwey, and W. M. Scheld. 1990. Patho-
genesis and pathophysiology of meningitis. Infect. Dis. Clin.
North Am. 4:555-581.
167. Tuomanen, E., B. Hengstler, R. Rich, M. A. Bray, 0. Zak, and
A. Tomasz. 1987. Nonsteroidal anti-inflammatory agents in the
therapy for experimental pneumococcal meningitis. J. Infect.
Dis. 155:985-990.
168. Tuomanen, E., H. Liu, B. Hengstler, 0. Zak, and A. Tomasz.
1985. The induction of meningeal inflammation by components
of the pneumococcal cell wall. J. Infect. Dis. 151:859-868.
CLIN. MICROBIOL. REV.
169. Tuomanen, E., A. Tomasz, B. Hengstler, and 0. Zak. 1985. The
relative role of bacterial cell wall and capsule in the induction
of inflammation in pneumococcal meningitis. J. Infect. Dis.
151:535-540.
170. Tuomanen, E. I., K. Saukkonen, S. Sande, C. Cioffe, and S. D.
Wright. 1989. Reduction of inflammation, tissue damage, and
mortality in bacterial meningitis in rabbits treated with mono-
clonal antibodies against adhesion-promoting receptors of leu-
kocytes. J. Exp. Med. 170:959-968.
171. Tureen, J. H., R. J. Dworkin, S. L. Kennedy, M. Sachdeva, and
M. A. Sande. 1990. Loss of cerebrovascular autoregulation in
experimental meningitis in rabbits. J. Clin. Invest. 85:577-581.
172. Tureen, J. H., M. G. Tauber, and M. A. Sande. 1991. Effect of
indomethacin on the pathophysiology of experimental menin-
gitis in rabbits. J. Infect. Dis. 163:647-649.
173. Tureen, J. H., M. G. Tauber, and M. A. Sande. 1992. Effect of
hydration status on cerebral blood flow and cerebrospinal fluid
lactic acidosis in rabbits with experimental meningitis. J. Clin.
Invest. 89:947-953.
174. Waage, A., P. Brandtzaeg, A. Halstensen, P. Kierulf, and T.
Espevik. 1989. The complex pattern of cytokines in serum from
patients with meningococcal septic shock. Association be-
tween interleukin 6, interleukin 1, and fatal outcome. J. Exp.
Med. 169:333-338.
175. Waage, A., A. Halstensen, R. Shalaby, P. Brandtzaeg, P.
Kierulf, and T. Espevik. 1989. Local production of tumor
necrosis factor a, interleukin 1, and interleukin 6 in meningo-
coccal meningitis. Relation to the inflammatory response. J.
Exp. Med. 170:1859-1867.
176. Waggener, J. D. 1974. The pathophysiology of bacterial men-
ingitis and cerebral abscesses: an anatomical interpretation.
Adv. Neurol. 6:1-17.
177. Weber, A., K. Harris, S. Lohrke, L. Forney, and A. L. Smith.
1991. Inability to express fimbriae results in impaired ability of
Haemophilus influenzae b to colonize the nasopharynx. Infect.
Immun. 59:4724-4728.
178. Wenger, J. D., A. W. Hightower, R. R. Facklam, S. Gaventa,
C. V. Broome, and the Bacterial Meningitis Study Group. 1990.
Bacterial meningitis in the United States, 1986: report of a
multistate surveillance study. J. Infect. Dis. 162:1316-1323.
179. Whittle, H. C., and B. M. Greenwood. 1977. Cerebrospinal
fluid immunoglobulins and complement in meningococcal men-
ingitis. J. Clin. Pathol. 30:720-722.
180. Williams, A. E., and W. F. Blakemore. 1990. Pathogenesis of
meningitis caused by Streptococcus suis type 2. J. Infect. Dis.
162:474-481.
181. Winkelstein, J. A., and A. Tomasz. 1977. Activation of the
alternative pathway by pneumococcal cell walls. J. Immunol.
118:451-454.
182. Wispelwey, B., E. J. Hansen, and W. M. Scheld. 1989. Haemo-
philus influenzae outer membrane vesicle-induced blood-brain
barrier permeability during experimental meningitis. Infect.
Immun. 57:2559-2562.
183. Wispelwey, B., A. J. Lesse, E. J. Hansen, and W. M. Scheld.
1988. Haemophilus influenzae lipopolysaccharide-induced
blood brain barrier permeability during experimental meningi-
tis in the rat. J. Clin. Invest. 82:1339-1346.
184. Wyler, D. J., S. I. Wasserman, and A. W. Karchmer. 1979.
Substances which modulate leukocyte migration are present in
CSF during meningitis. Ann. Neurol. 5:322-326.
185. Yamashima, T., K. Kashihara, K. Ikeda, T. Kubota, and S.
Yamamoto. 1985. Three phases of cerebral arteriography in
meningitis: vasospasm and vasodilatation followed by organic
stenosis. Neurosurgery 16:546-553.
186. Zwahlen, A., U. E. Nydegger, P. Vaudaux, P. H. Lambert, and
F. A. Waldvogel. 1982. Complement-mediated opsonic activity
in normal and infected human cerebrospinal fluid: early re-
sponse during bacterial meningitis. J. Infect. Dis. 145:635-646.
187. Zwahlen, A., J. A. Winkelstein, and E. R. Moxon. 1983.
Surface determinants of Haemophilus influenzae pathogenic-
ity: comparative virulence of capsular transformants in normal
and complement-depleted rats. J. Infect. Dis. 148:385-394.