Management of Acute Bacterial and Viral Meningitis in Adults

    Authors: Dayna Lynn Gutsin, MD, Chief, Resident, Department of Emergency Medicine, Emory University School of Medicine.

    Katherine L. Heilpern, MD, Assistant Professor, Department of Emergency Medicine, Emory University School of Medicine.

    Editor: Gideon Bosker, MD, Assistant Clinical Professor, Section of Emergency Services, Yale University School of Medicine, Associate Clinical Professor, Oregon HEalth Sciences University, Portland.

    Peer Reviewer: David E. Manthey, MD, Research Director, Brooke Army Medical Center, Department of Emergency Medicine, San Antonio Uniformed Services Health Education Consortium, San Antonio, TX.

Editor's Note-Meningitis, or inflammation involving the meninges and cerebral spinal fluid (CSF), is a life-threatening emergency that may be associated with devastating neurologic sequelae or death if not promptly recognized and treated. Its causative organisms have been changing over the years, and so has the treatment for this disease. Because of its myriad presentations, it is crucial that the physician be able to diagnose acute bacterial meningitis (ABM) and differentiate this condition from other disease processes that may cause or mimic meningitis.

The incidence of bacterial meningitis is approximately 4-10 cases per 100,000 people in the United States, and this infection is responsible for more than 2000 deaths per year.1 However, a 1997 study reported a 55% reduction in all cases of bacterial meningitis between 1986 and 1995. Specifically, there was an 87% reduction in meningitis in children between the ages of 2 months and 5 years, a group that accounted for two-thirds of all cases reported in 1986.2 The introduction of the Haemophilus influenzae Type b (Hib) vaccine has dramatically reduced the incidence of what was once the most prevalent organism responsible for bacterial meningitis. Currently, the age group at greatest risk for ABM includes children between 1 and 24 months of age, followed by a gradual increase in risk of infection from age 22 years until old age.3 Adults older than 60 years old account for 1000-3000 cases of ABM per year in the United States and for more than 50% of all deaths related to meningitis.4

Understanding age groups at risk and characterization of likely pathogens aids the physician in confirming the diagnosis and implementing life-saving treatment. With these clinical challenges in mind, the purpose of this article is to review critical decisions involved in the diagnosis and management of adult meningitis.

Overview and General Principles

The pathogens responsible for bacterial meningitis are thought to originate in the nasopharyngeal mucosa. From this location, pathogens invade the intravascular space, cross the blood brain barrier (BBB), and eventually enter the CSF. Viruses (enterovirus) are more likely to spread via the fecal-oral route. In this scenario, viral replication occurs in the GI tract, which leads to viremia and subsequent invasion of the BBB.5 Generally speaking, host defenses in the subarachnoid space are inadequate and inefficient; therefore, once organisms gain entry into the CSF, bactericidal activity is lacking6 and pathogens may propagate.

The deleterious consequences of ABM are, to a great extent, caused by the inflammatory response generated by the host. Inflammation involves many cascades and cellular responses, including activated polymorphonuclear leukocytes (PMNs) and inflammatory mediators such as interleukin-1 (IL-1), tumor necrosis factor (TNF), and platelet activating factor (PAF).7 The consequences of this inflammatory response include cerebral edema, vasculitis, increased intracranial pressure (ICP), decreased cerebral blood flow, loss of autoregulation, and cortical hypoxia.1 Similar pathophysiologic consequences occur with non-bacterial (aseptic) meningitis; however, recruitment of leukocytes and activation of inflammatory mediators are not as intense as they are in ABM.

Patients at Risk

A number of underlying medical illnesses and epidemiologic risk factors are associated with an increased risk of bacterial meningitis. Extremes of age, anatomic or functional asplenia (e.g., sickle cell disease), alcoholism, cirrhosis, malnutrition, chronic liver or renal disease, immunoglobulin or complement deficiencies, HIV, malignancy, and diabetes mellitus increase the risk for acquiring meningitis. In addition, patients with recent head injury or neurosurgery, cerebroventricular shunt or CSF leak, or who attend day care or live in close living quarters (i.e., military barracks or college dorms) are at increased risk. Finally, local extension or bacteremia leads to an increased risk of meningitis in persons with such concomitant conditions as otitis media, sinusitis, mastoiditis, brain abscess, or pneumonia.1,6,8,9

Patients in whom the diagnosis of ABM is being considered should be isolated in a respiratory isolation room. This is required to prevent potential transmission of Neisseria meningitidis and Haemophilus influenzae, both of which are spread via respiratory droplets.

Clinical Presentation

It should be stressed that only 10-20% of patients eventually diagnosed with ABM present with fulminant symptoms over a 24-hour period; the remaining 80-90% of individuals develop symptoms over 1-7 days. Eighty-five percent of patients with bacterial meningitis eventually present with fever, headache, meningismus or nuchal rigidity, and altered mental status.8 Other common signs and symptoms include photophobia, vomiting, back pain, myalgias, diaphoresis, and malaise. Generalized seizures can occur in up to 40% of patients with ABM.1 Kernig's sign (resistance to extension of the leg while the hip is flexed) and Brudzinski's sign (involuntary flexion of the hip and knee when the patient's neck is abruptly flexed while laying supine) are observed in up to 50% of patients and are suggestive of meningitis. However, their presence or absence does not rule in or rule out meningitis.1,5,8

Papilledema is present in fewer than 1% of patients on initial exam; in fact, its presence may suggest an alternative diagnosis, such as venous sinus thrombosis, brain abscess, subdural empyema, or syphilis.8,9 Cranial nerve palsies, especially those involving CN III, IV, VI, and VII can occur in up to 20% of patients; visual field defects, dysphasia, and hemiparesis are encountered less frequently. Coma, hypertension, and signs of impending herniation are late findings.1 Importantly, about 50% of patients with N. meningitidis may present with a rash that begins as an erythematous macular rash, and then eventually progresses to petechiae and purpura.1,3,4

From a clinical perspective, it is important to note that viral meningitis shares several clinical features of bacterial meningitis including fever, headache, nausea and vomiting, and back pain. These signs and symptoms may be preceded by several days of a nonspecific acute febrile illness characterized by malaise and anorexia.10 The symptoms can develop abruptly, or they can evolve over several weeks. Although seizures occur less often in viral meningitis, certain viruses such as Arbovirus, mumps, and Varicella-Zoster virus are associated with seizures and encephalitis.10 An erythematous macular rash may be present as well. Although many physicians feel that viral meningitis presents with a more indolent, or less "toxic," course, this is not always the case; therefore, CSF evaluation is mandatory in all suspected cases.

It should be emphasized that certain groups of patients present with atypical signs and symptoms that may be confused with aseptic meningitis or other illnesses. In particular, elderly patients may present only with lethargy or obtundation, with or without fever or signs of meningeal irritation. In fact, fever may be absent in up to 40% of elderly patients with ABM; instead, some cases may present with hypothermia.8,9 Confusion occurs in more than half of elderly patients, while headaches occur in 20-80% of all cases. Nuchal rigidity, which is present in about 60-90% of cases, is neither sensitive nor specific in the geriatric patient. One review suggests that approximately 35% of elderly patients have had nuchal rigidity without meningitis. In these cases, nuchal rigidity is due to such other diseases as cervical spondylosis or Parkinson's disease.4 Accordingly, the physician should maintain a high index of suspicion for meningitis in all elderly patients with altered level of consciousness, whether or not they have fever, headache, or meningismus.

Although this article focuses on bacterial meningitis in adults, it should be pointed out that neonates comprise an important group of patients with atypical presentations; these include poor feeding, listlessness, and altered respiratory pattern. Patients with head trauma (in particular, penetrating wounds) are at increased risk for developing meningitis; however, the physician may erroneously believe the symptoms are secondary to the trauma and may forego performing an LP. Neutropenic or immunosuppressed patients, which include patients with HIV, chronic steroid therapy, or metastatic disease, may not be able to mount the characteristic CSF inflammatory response. Therefore, typical signs of meningeal irritation may be absent. In patients with a cerebroventricular shunt, malfunction may result in shunt infection, and, conversely, shunt infection may result in malfunction. These patients often present with headache, nausea, lethargy, with or without fever, and alterations in level of consciousness.9

Fungal meningitis usually presents in an indolent fashion with focal neurologic signs and altered mental status. Tuberculous meningitis, like fungal meningitis, may be characterized by a slowly evolving course and absence of frank meningeal signs. Cranial nerve deficits, ophthalmoplegia, facial droop, seizures, and altered mental status are common findings.4 In patients with AIDS, cryptococcal meningitis may present with a paucity of signs and symptoms (mild headache, constitutional symptoms); therefore, physicians should maintain a high suspicion for fungal or tuberculous meningitis in patients with HIV, even if patients present with minimal symptoms. Partially treated bacterial meningitis may be clinically indistinguishable from aseptic meningitis; consequently, a thorough history of recent antibiotic use is important in all patients who are being evaluated for meningitis.

Finally, there are many infectious and noninfectious causes of meningeal irritation. In patients who have an indeterminate CSF, the physician should consider unusual pathogens, multisystem disease, or drug reactions. (See Table 1.)

Table 1. Differential Diagnosis of Acute Bacterial Meningitis

Infectious Non-Infectious
• Parameningeal Focus • Systemic Disease
       Brain Abscess        Sarcoidosis
       Subdural Empyema        Kawasaki disease
       Epidural Abscess        SLE
       Pansinusitis        Multiple Sclerosis
• Bacterial Infection        Migraine Headache
       Partially treated ABM        Guillain-Barre
       Lyme disease        Behcet’s disease
       Syphilis        Malignancy
       Leptospirosis        Leukemia
       Neurobrucellosis        Lymphoma
• Fungal Infection • Vaccine Reaction
• Mycobacterial Infection        Mumps
• Parasitic Infection        MMR
• Viral Infection        Polio
       Adenovirus        Poison
       Arbovirus        Lead
       CMV        Mercury
       Coronavirus • Trauma
       Enterovirus        Subarachnoid
       EBV        Hemorrhage
       HSV        s/p Neurosurgery
       HIV        Traumatic LP
       Influenza Virus • Drugs
       Parainfluenza virus        Azathioprine
       Rabies virus        Ibuprofen
       LCV        IVIG
       Rhinovirus        Isoniazid
       Rotavirus        OKT3
       VZV        Sulfamethizole Trimethoprimsulfamethoxazole

Patient Evaluation: Timing and Critical Decisions

Computerized Tomography (CT). One of the most important factors contributing to delayed diagnosis and treatment of bacterial meningitis is the decision to perform cranial CT imaging prior to lumbar puncture (LP).11 Historically, a subset of practitioners has been concerned about the remote risk of herniation that may occur during LP. Systematic guidelines help risk stratify these patients. When evaluating patients for meningitis, patients who require CT prior to LP include those with focal neurologic findings, papilledema, focal seizures, or abnormalities on exam that suggest increased intracranial pressure.1,4 Such findings can also suggest alternative diagnoses such as brain abscess, venous sinus thrombosis, mass lesion, subdural empyema, obstructive hydrocephalus, or tuberculous or fungal meningitis. Any expert would agree that it is usually safe to perform an LP on patients who have a reliable history of generalized tonic-clonic seizures, but who present without neurologic deficits. One other group of patients who should have CT evaluation done prior to LP are those in whom atypical presentations are common. For example, patients with immunosuppression (i.e., HIV, multiple myeloma or metastatic cancer, or patients receiving chronic steroids or chemotherapy) are at risk for mass lesions with or without clinical evidence of elevated ICP. Frequently, elderly patients present atypically (without fever or meningismus) and may have comorbid illnesses that cause altered mentation (stroke, hemorrhage). Alcoholics comprise another high-risk group; not only are they immunosuppressed, but they are at increased risk for subdural hematoma.

The most compelling argument for obtaining a CT prior to performing an LP is that a mass lesion may not be clinically evident, and that empiric treatment can be instituted before the scan. Moreover, some authorities suggest that the delay in time to LP does not affect its accuracy or outcome of the results.11 Opponents feel that CT is usually a waste of time and valued resources, and that CSF culture results may be affected by empiric therapy. If bacterial meningitis is a strong consideration, and the decision is made to perform a CT prior to LP, two sets of blood cultures should be obtained and antibiotics should be administered promptly before sending the patient for neuroimaging. Urine cultures may be helpful in the very young and very old.

Timing of Antibiotic Administration

Delaying administration of antibiotics in patients suspected of having bacterial meningitis is a frequent precipitant of malpractice litigation. Typical arguments take the position that a delay in antibiotic therapy may result in increased morbidity and mortality.11,12,13 Although there is still some controversy concerning clinical outcomes and their relationship to the timing of antibiotic administration, most physicians would agree that meeting the following standard of care is advisable. Antibiotics should be administered within 30 minutes of clinical suspicion/presentation of ABM. Often, patients who appear "sick" or have symptoms such as hypotension, abnormal mental status, or a temperature greater than 40°C will receive antibiotics more quickly, as will children in the age group of 2-10 years, and infants with a bulging fontanelle. Interestingly, the presence of headache is associated with delayed use of antibiotics, probably because alternative diagnoses for headache are sought.13 It is essential to initiate antibiotics quickly in high-risk patients with atypical presentations. Blood cultures followed by antibiotic administration within 30 minutes of presentation are mandatory in all patients suspected of having bacterial meningitis and in whom CT precedes LP.

Table 2. Cerebrospinal Fluid Parameters in Meningitis

  Normal Bacterial Viral Fungal TB Parameningeal Focus or Abscess
WBC count
0-5 > 1000 100-1000 100-500 100-500 10-1000
% lymph
0-15 90 < 50
> 50
< 50
> 50
< 50
> 80
< 50
45-65 <40 45-65 30-45 30-45 45-65
glucose ratio
0.6 < 0.4 0.6 < 0.4 < 0.4 0.6
20-45 >150 50-100 100-500 100-500 > 50
Opening pressure
(cm H20)
6-20 ++ NL or + ++ ++ N/A

Although prompt administration of antibiotics is the rule rather than the exception, there are two groups of patients in whom it may be reasonable to withhold antibiotics pending LP results. In this regard, patients who are young, healthy and "nontoxic" in appearance, and in whom the physician is strongly suspicious of aseptic meningitis or an alternate diagnoses (i.e., there is a very low suspicion for bacterial meningitis), comprise one group. The second group consists of patients who are nontoxic, do not require CT, but in whom the diagnosis of ABM is likely and in whom LP results can be obtained within a reasonable time limit, (i.e., 30-60 minutes).

It should be stressed that, despite immediate initiation of antibiotic therapy, there is a minimal alteration in CSF parameters (cellular and metabolic composition) for 12-24 hours; consequently, there is a sufficient diagnostic window within which to perform LP and still make an accurate diagnosis.8 However, it must be stressed that the diagnostic yield of CSF Gram's stain and culture decreases by approximately 20% and 30%, respectively, if antibiotics are initiated prior to LP.22 Although Gram's stain and cultures may be negative shortly after initiation of antibiotics, as a rule, cell counts, percentage of cell morphologies, and protein and glucose levels remain unaltered for about 24 hours after initiation of IV antibiotics.10

Interpretation of Lumbar Puncture

Examination of the CSF is mandatory for diagnosis evaluation of meningitis. (See Table 2.) The typical CSF findings for ABM can be found in Table 3. Fortunately, prediction rules have been generated to help characterize and confirm the diagnosis of bacterial meningitis. These predictors have been tested in many situations and have performed reliably as individual predictors of bacterial meningitis.15 The following prediction rules developed by Spanos and others have demonstrated certainty of 99% or higher for the diagnosis of ABM:1,3,8,9,15

1. CSF glucose < 34 mg/dL

2. CSF:blood glucose ratio < 0.23

3. CSF protein > 220 mg/dL

4. Total CSF leukocyte count > 2 ´106/L

5. Total CSF neutrophil count > 1180 ´ 106/L

In addition, this author has produced an excellent nomogram for estimating the probability of bacterial vs. viral meningitis.3

In many patients, the LP results will fall in an indeterminate category (i.e., an elevated CSF leukocyte count higher than 500/mcL with near normal CSF glucose and protein). In 25% of patients with bacterial meningitis, no organisms are seen on Gram's stain, and in 30-40% of these patients, the CSF parameters will be nondiagnostic.3,15 In other words, 10% of patients with subsequent confirmation of ABM will have an LP that yields a negative Gram's stain-as well as a total CSF WBC lower than 1000/mcL-and only mild abnormalities in the CSF glucose and protein.

If antibiotics are not given prior to the LP, the accuracy of the Gram's stain is between 60-90%. Overall, the Gram's stain performs poorly as a screening test, with a sensitivity of only 40-60%; however, specificity is greater than 90%.1,8 Furthermore, the sensitivity decreases by 20% if the patient has been pretreated with antibiotics.14 CSF cultures are positive in about 70-85% of cases without prior antibiotic administration and decrease to less than 50% with prior therapy.1,8 Given the relative insensitivity of the Gram's stain and culture, obtaining two sets of blood cultures prior to antibiotic administration is imperative and may help compensate for the decrease in positive CSF culture results.

Table 3. Typical CSF Findings in Patients with Bacterial Meningitis

CSF Parameter Typical Findings
Opening pressure > 180 mm H2Oa
White blood cell count 1000-5000/mcL (range,
< 100 to >1 0,000)b
Percentage of neutrophils > 80% c
Protein 100-500 mg/dL
Glucose < 40 mg/dLd
Lactate > 35 mg/dL
Gram’s stain Positive in 60-90%e
Culture Positive in 70-85%f
Bacterial antigen detection Positive in 50-100%
a Values over 600 mm H2O suggest the presence of cerebral edema, intracranial suppurative foci, or communicating hydrocephalus.
b Patients with very low CSF white blood cell counts (0 to 20/uL) tend to have a poor prognosis.
c About 30% of patients with Listeria monocytogenes meningitis have an initial lymphocyte predominance in CSF.
d The CSF-serum glucose ratio is < 0.31 in ~70% of patients
e The likelihood of detecting the organism by Gram’s stain correlates with the concentration of bacteria in the CSF; concentrations of < 103 cfu/mL is associated with positive Gram’s stain ~25% of the time and concentrations > 105 cfu/mL leads to positive microscopy in up to 97% of cases.
f Yield of CSF cultures may decrease in patients who have received prior antimicrobial therapy.

    Adapted from: Tunkel AR, Scheld WM. Acute bacterial meningitis in adults. In: Remington JS, Swartz MN, eds. Current Clinical Topics in Infectious Diseases. Boston: Blackwell Science, 1995:220.

Preadministration of oral antibiotics prior to presentation does not appear to decrease the total CSF white blood cell count, the glucose, the glucose serum ratio, or the percentage of patients with a positive Latex agglutination (LA). There is, however, a decrease in percentage of neutrophils, a decrease in CSF protein, and a decrease in the percentage of positive Gram's stains. Intravenous antibiotics given 1-2 days prior to LP may not significantly alter the CSF cell count, protein, or glucose concentration, but will substantially decreases the positivity of the Gram's stain and culture.16

Table 4. Antibiotic Choice Based on Age and Comorbid Medical Illness

Neonate E. coli, Group B strep, Listeria
Ampicillin and third-generation cephalosporin
1-3 months S. pneumoniae, N. meningitidis,
H. influenzae, S. agalactiae

Listeria, E. coli
Ampicillin and third-generation cephalosporin
3 months to 18 years N. meningitidis, S. pneumoniae,
H. influenzae
Third-generation cephalosporin
18-50 years S. pneumoniae, N. meningitidis Third-generation cephalosporin
Older than 50 years N. meningitidis, S. pneumoniae
Gram negative bacilli, Listeria
Group B strep
Ampicillin and third-generation
Neurosurgery/ head injury S. aureus, S. epidermidis
Diphtheroids, Gram neg. bacilli
Vancomycin and
Immunosuppression Listeria, Gram negative bacilli
S. pneumoniae, N. meningitidis
(consider adding Vancomycin)
Ampicillin and Ceftazidime
CSF shunt S. aureus, Gram negative bacilli Vancomycin and Ceftazidime

Third generation cephalopsporin = ceftriaxone or cefotaxime

Ancillary CSF tests can be employed to help confirm the diagnosis. For example, if the CSF parameters are nondiagnostic, or the patient has been treated with prior oral antibiotics, and, therefore, the Gram's stain and/or culture are likely to be negative, then LA may be helpful. This test requires a small amount of bacterial antigen to produce a positive test.1 It has a variable sensitivity rate, ranging between 50-100%, and high specificity. Latex agglutination tests are available for Hib, Streptococcus pneumoniae, N. meningitidis, Escherichia coli K1, and S. agalactiae (Group B strep).8 Up to 25% of patients pretreated with antibiotics have a positive LA, so it may be helpful in a small subgroup of patients who have received antibiotics prior to presentation.9 Given the cost of antigen testing (approximately $100) and its limited clinical utility, most experts do not recommend bacterial antigen testing of CSF, except in patients who have received prior antibiotics or who have a negative Gram's stain.4 CSF C-reactive protein (CRP) has been found to be significantly elevated in children with ABM compared to those individuals with aseptic meningitis; however, there are no established recommendations for adults.10 In children, a CRP greater than 20 mg/L corresponds to bacterial meningitis with 100% sensitivity and 96% specificity.5 CSF lactate and pH are also under investigation for their clinical utility in predicting ABM.10

CSF Cryptococcal antigen and India ink stain should be considered in all patients who have known HIV disease or HIV risk factors, including those who live in an area with a high incidence of HIV. Cryptococcal antigen is positive in 95% of confirmed cases, and India ink is positive in 50-75% of cases.4

In the nontoxic patient with a nondiagnostic lumbar puncture, the LP should be repeated in 8-12 hours to help differentiate between bacterial and viral (aseptic) meningitis, regardless of whether the physician has administered antibiotics.10 Early in the course of viral meningitis, there may be a predominance of neutrophils in the CSF initially; this will shift toward a lymphocytic cell line after a few hours.3,8,10 This also holds true for bacterial meningitis that may initially present with a lymphocytic predominance in the CSF. A repeat LP can be performed as an outpatient in patients with reliable caretakers and guaranteed follow-up. Outpatient LP is appropriate for patients who have already received antibiotics within the week prior to presentation or in children younger than 1 year of age.10

A traumatic tap will raise both red and white blood cell counts. To interpret the actual number of CSF white blood cells, a quick rule of thumb is as follows :

True WBC CSF = (WBC actual CSF - WBC blood) ´ RBC CSF - RBC blood

In addition, 1000 RBCs in the CSF artificially increase the CSF protein by 1 mg/dL.16

Table 5. Antibiotic Choice Based on Gram’s Stain

Gram’s (+) cocci S. pneumoniae, S. aureus
S. agalactiae (Group B
Vancomycin and third- generation cephalosporin
Gram’s (-) cocci N. meningitidis Penicillin G
Gram’s (-) coccobacilli H. influenzae Third-generation cephalosporin
Gram’s (+) bacilli Listeria monocytogenes Ampicillin, Pen G + Gentamycin
Gram’s - bacilli E.coli, Klebsiella Serratia, Pseudomonas Ceftazidime +/- aminoglycoside

Acute Viral Meningitis (AVM)

The typical presentation of acute viral meningitis (AVM) is characterized by acute onset of fever, headache, nausea and vomiting, with or without meningismus. AVM occurs most commonly in children between 1 and 10 years of age and has a peak incidence in summer and early fall. Patients with AVM are generally easily arousable and have a nonfocal physical exam.17 Approximately one-half of patients with AVM will present with nuchal rigidity. Myalgias and photophobia are other commonly encountered complaints. Seizures are rare in AVM, except in young children with high fever or in those with a preexisting seizure disorder.18 Most cases (85%) are secondary to enteroviral infection, and, consequently, other evidence of enteroviral illness may accompany the clinical symptoms of meningitis, including pharyngitis, conjunctivitis, rash (macular, maculopapular, petechial, or vesicular), pleurodynia, pericarditis, or myocarditis.19 AVM may be preceded by a nonspecific acute febrile illness with malaise and anorexia suggesting viral disease.18 Typically, the headache of AVM improves significantly after the LP is performed.20 The absence of other sites of bacterial infection such as pneumonia or otitis media supports the diagnosis of AVM. Finally, disease in the community or in a close contact supports the diagnosis of AVM.20 The spread of enterovirus is via the fecal-oral route and has an incubation period of 4-6 days. The illness usually lasts between 1-2 weeks.19

Table 6. Recommended Dosages of Antibiotics

Antibiotic Dosage
May need to adjust for renal or hepatic disease
Ampicillin 2g IV q 4 h
Cefotaxime 2g IV q 4-6 h
Ceftazidime 2g IV q 8 h
Ceftriaxone 2g IV q 12 h
Gentamycin Load 1.5mg/kg IV then 1-2mg/kg q 8 h
Nafcillin/ Oxacillin 1.5-2g IV q 4 h
Penicillin G 4 million units IV q 4 h
Rifampin 600 mg po q 12-24 h
Trimethoprimsulfamethoxazole 10 mg/kg IV q 12 h
Vancomycin 1.5-2 g IV q 12 h

The annual incidence of AVM is between 11 and 27 cases per 100,000 population. Over 7000 cases are reported annually in the United States.17 AVM is well known to result in community-wide outbreaks often leading to infection in more than 1000 patients during an epidemic.21

Evaluating AVM

In every patient evaluated for acute meningitis, the physician must form a pretest probability of whether the patient has AVM or ABM. This is based on the clinical appearance and exam, the characteristics of the patient (age and comorbidities), and epidemiological factors (ill contacts, local epidemics, and the time of year). Once the pretest probability for AVM has been established, the physician should proceed to the LP to help confirm the diagnosis and rule out other life-threatening illnesses such as ABM or subarachnoid hemorrhage. The typical CSF findings of AVM are outlined in Table 2.

AVM Lumbar Puncture

As many as two-thirds of all patients ultimately diagnosed with AVM demonstrate an initial neutrophilic predominance in the CSF.22 As a result, many patients receive several days of IV antibiotics and hospitalization, even though 48 hours later, blood and CSF cultures are negative. Several studies have evaluated the utility of a second LP to differentiate patients with AVM vs. ABM. Feigin and Shackelford recognized the concept of leukocyte shifting during a retrospective study as early as 1973.23 They found that 85 of 230 patients with a discharge diagnosis of aseptic meningitis demonstrated an initial LP with neutrophilic predominance (37 of these patients received antibiotics after the initial LP and were removed from the study). The remaining 48 patients comprised the study group. These patients were not toxic, and all had a total CSF WBC count less than 1000/mcL with a neutrophilic predominance, a near normal CSF glucose concentration, and a negative Gram's stain. None of these patients received IV antibiotics. These 48 patients had a second LP performed 6-72 hours after the initial LP (31 of 48 had it repeated within 6-8 hours). In this study, 87% of patients had a clear shift from a polymorphonuclear cell line in the CSF to a mononuclear line between the initial LP and one performed 6-8 hours later (P < 0.001). Ninety-four percent shifted to a mononuclear cell line 12-72 hours after the initial LP. No significant change occurred in CSF total cell count, glucose, and protein. They concluded that, in most cases, if the patient has not received prior antibiotics, a brief period of observation followed by a repeat LP in 6-12 hours helps the physician distinguish between AVM and ABM.

A similar study was conducted by another group prospectively in 1979.22 It included 16 patients with a history and physical exam consistent with AVM, who received supportive care only (i.e., no antibiotics). The LP was repeated 18-48 hours later in the 10 patients who failed to show significant clinical improvement. Additionally, 11 of 16 patients had an LP performed prior to discharge (5-12 days after the initial LP), although all were asymptomatic at discharge. All patients in this study had a total CSF WBC count less than 500/mcL on the initial LP with a mean percent of PMNs of 41.75 ± 29.00. The percent of PMNs in the CSF fell significantly to a mean percent of 8.0 ± 8.8 (P < 0.001) in all 10 patients who had repeat LP performed; the neutrophil percent remained low at 5-12 days. Total leukocyte count, glucose, and protein content in the CSF showed no significant change throughout the acute phase of the illness. In this study of patients with AVM, all patients undergoing serial LP demonstrated a shift in the cell line from neutrophils to mononuclear cells; none of these patients received antibiotics, and all had good outcomes.

Another study in 1989 confirmed that the season and the patient's age are also important predictors because patients at the extreme of age are more likely to have ABM, and young adults are more likely to have AVM. In this study, neutrophils predominated in the initial spinal tap in 40% of cases eventually diagnosed with AVM. Conversely, 15% of patients with a discharge diagnoses of ABM had lymphocytosis in the CSF on initial LP. In 70% of the patients diagnosed with AVM, the fall in proportion of PMNs in the CSF was seen during the first two days of hospitalization, reaffirming the shift from polymorphic to monomorphic cell lines on repeat LP.

In 1991, a study suggested that the shift from polymorphic to monomorphic cells in the CSF occurs approximately 24 hours after the initial symptoms.24 This prospective study was performed during an outbreak of enteroviral meningitis in Israel. Special attention was paid to the time of onset of symptoms. A second LP was performed 6-14 hours after the initial LP on all patients who did not improve clinically and had a CSF analysis suggestive of AVM, but with a neutrophilic predominance. The patients were divided into four groups according to the interval between onset of symptoms and time of the first LP (LP performed within 12 hours of symptom onset; 12-24 hours of symptom onset; 24-36 hours after symptom onset; and more than 36 hours). The percentage of PMNs in the CSF of the patients who had the initial LP in the first 24 hours of onset of the illness fell significantly more than the patients in the latter group (P < 0.01). In all samples obtained more than 24 hours after the onset of symptoms, fewer than 50% of the cells were PMNs. This study affirmed that a second LP is helpful in distinguishing between AVM and ABM, but that the time interval between the onset of symptoms and the LP may influence the trend toward monomorphic cells. Previous studies have recommended repeating the tap in 6-12 hours after the initial tap, whereas this trial suggests deferring the second tap until 24 hours have elapsed after the onset of symptoms (assuming that the patient has presented within the first few hours of his or her illness) in order to provide more conclusive information. A predominance of neutrophils 2-3 days after the initial onset of the symptoms would be unusual in AVM.

Early Bacterial Meningitis vs. AVM

In 1984, Powers reported a CSF lymphocytic predominance in 14 of 103 cases of bacteriologically proven acute bacterial meningitis.15 Classically, ABM presents with CSF leukocyte counts greater than 1000/ mm3. In this study, however, 41 patients with ABM had CSF leukocyte counts less than 1000/mm3, and 13 of the 41 patients had CSF lymphocytosis as well. Five of the 14 patients in this study were younger than 2 months old. Many of the patients had other CSF parameters that suggested ABM including low glucose (7 of 14), elevated protein (6 of 14) or a positive Gram's stain (7 of 14). There was no difference between the groups in terms of bacterial type or pretreatment with antibiotics. There is not adequate data regarding the shift of monomorphs to polymorphs in the CSF of patients infected with early ABM. However, the study by Powers suggests that up to 10% of patients with ABM may present with a total CSF WBC count less than 1000/mcL and a lymphocytic predominance. In most cases, other CSF parameters will be abnormal and suggestive of ABM.

The majority of the evidence seems to support the use of serial lumbar puncture to aid the physician in differentiating between AVM and ABM in patients with a total CSF WBC count less than 500/mcL. The patient who appears toxic or has other CSF parameters consistent with ABM is admitted and given IV antibiotics. The patient who is nontoxic and has clear evidence of AVM by LP (CSF WBC count less than 500/mcL with a lymphocytic predominance, normal CSF glucose, normal or slightly elevated CSF protein, and a negative Gram's stain) may be discharged home with supportive care (for pain, fever, and dehydration). The patients in whom the history and clinical exam suggest a viral etiology, but the LP results are ambiguous (i.e., a CSF WBC count less than 500/mcL, a neutrophil predominance, a normal glucose, and a slightly elevated protein with a negative Gram's stain), warrant a period of observation and, if symptoms persist, a repeat LP 6-12 hours later.

The patient can be discharged if he has guaranteed follow-up and a family member or friend to closely observe him in the interim. This is not appropriate for the extremes of age, patients with immunocompromising comorbidities (cancer, HIV, steroid therapy, diabetes, chronic alcoholism), or in patients whose mental status cannot be followed. These patients require admission and close observation. In most of the studies mentioned in this article, the majority of patients had a total CSF WBC count less than 500/mcL. There is no available data on whether a repeat LP is valuable in patients with a CSF WBC between 500 and 1000 with a neutrophil predominance, normal glucose, normal protein, and a negative Gram's stain.

The repeat lumbar puncture is most valuable in the group of patients suspected to have AVM who have a neutrophil predominant CSF and persistent CNS symptoms. Patients with persistent headache or meningismus should undergo repeat LP to differentiate AVM from early bacterial meningitis.

Antimicrobial and Corticosteroid Therapy

The initial choice of antibiotic depends upon the patient's age, comorbid medical illnesses, and trends in the community for antimicrobial resistance. (See Table 4.) Clearly, the antibiotic should have a high degree of CSF penetration, achieve a high concentration in the CSF, and have intrinsic activity in infected fluid.1

Knowledge of the patient's underlying medical condition will help in antibiotic selection. For example, Listeria monocytogenes is seen in patients older than 60 years of age, in patients with renal or liver failure, in individuals with connective tissue disease, cancer, diabetes, chronic steroid therapy, and in alcoholics.8,9 Listeria monocytogenes can account for up to 25% of cases of ABM in patients older than 60 years.4 S. aureus is a common pathogen in patients who have undergone neurosurgery, in those who have a CNS shunt, and in patients with a head injury and CSF leak. Enteric Gram-negative bacilli (Klebsiella, E. coli, Pseudomonas, and Serratia) are implicated in neurosurgery patients, in elderly and immunosuppressed patients, and in patients with Gram negative sepsis. S. agalactiae (Group B strep) is also found in the elderly and in parturient women, as well as in patients with liver and kidney failure.9 HIV patients are at increased risk for toxoplasmosis, and for cryptococcal tuberculous and CMV meningitis.6

Knowledge of resistance patterns in your community is important when deciding upon empiric therapy. For example, a 1994 CDC study of Atlanta demonstrated that 25% of cases of invasive pneumococcal infection were resistant to penicillin, and 7% were highly resistant and resistant to third generation cephalosporins.8,11 In some areas, H. influenzae shows a 15-30% resistance to ampicillin for type B and non-type B strains.8 In addition to local epidemiology, awareness of the habitat of the patient is helpful; N. meningitidis in more commonly seen in outbreaks in populations living in close confines such as military barracks, dormitories, or day care.9 In most circumstances, the physician will administer empiric therapy based on age and underlying disease. If he or she is able to obtain Gram's stain results quickly, therapy may be stain-directed.

S. pneumoniae once was adequately covered by use of penicillin G or ampicillin, but with emerging and increasing resistance across the United States, a third-generation cephalosporin is a more prudent initial choice. In areas characterized by high resistance to penicillin, vancomycin plus a third-generation cephalosporin ought to be the first line therapy.9,11,16 This can be tailored after resistance patterns are obtained from cultures. H. influenzae is usually adequately covered by a third-generation cephalosporin. The drug of choice for N. meningitidis is penicillin or ampicillin, because little resistance has emerged to this organism. In patients who are at risk for Listeria meningitis, ampicillin must be added to the regimen. S. agalactiae (Group B) is covered by ampicillin as well, and adding an aminoglycoside provides synergy. Pseudomonas and other Gram-negative bacilli should be treated with a broad spectrum third-generation cephalosporin (ceftazidime) plus an aminoglycoside. S. aureus may be covered by nafcillin or oxacillin; vancomycin may be needed if the patient is at risk for methicillin-resistant S. aureus.9,11,16 (See Table 6.)

If it is possible to obtain a Gram's stain within 30 minutes of the patient's presentation, it may be possible to tailor therapy to the organisms seen by microscopy. These results should be interpreted in conjunction with the patient's underlying risk factors (i.e., a 25-year-old with no medical problems with Gram-positive cocci is more likely to have S. pneumoniae than S. aureus meningitis). (See Table 5.)


Dexamethasone reduces the inflammatory response that, to a great degree, is responsible for the morbidity associated with meningitis. In particular, the literature demonstrates that audiologic and neurological sequelae in infants older than two months of age are markedly reduced by early administration of dexamethasone in patients with H. influenzae (Hib) meningitis. There is also some evidence that dexamethasone also may be beneficial in children with pneumococcal meningitis. However, there is no clear evidence of any beneficial effect of dexamethasone in adult meningitis, although the issue has not been adequately studied.7,26

For optimal effect, dexamethasone should be given at a dose of 0.15 mg/kg every six hours IV for 2-4 days to children with suspected Hib or pneumococcal meningitis. The dose should be given just prior to or with the initiation of antibiotics.7,11 Increasing S. pneumoniae resistance to penicillin has created concern with regard to interactions between vancomycin and dexamethasone. Animal studies suggest that dexamethasone may decrease vancomycin penetration into the CSF, but a recent study in children did not bear this out.27 It is reasonable to believe that dexamethasone should be effective in diluting the inflammatory response in adults as well, but due to the lack of data and increased comorbidities in this age group, some experts recommend the use of corticosteroids only if there are organisms present on the Gram's stain (which implies a high concentration of the organism in the CSF) or in patients with declining mental status (coma), and evidence of increased intracranial pressure by exam or laboratory data.9,7


Without question, patients who appear "toxic," or in whom the diagnosis of bacterial meningitis has been confirmed, require admission. Patients diagnosed with viral meningitis can be sent home, but only if they have appropriate follow-up, easy access to health care, and supportive care at home. However, if the patient appears ill and suspicion for ABM was initially high (a high pretest probability), but the CSF suggests aseptic meningitis, the patient should be admitted for observation or short-stay to ensure this is not a case of early ABM. A repeat LP should be performed within 8-12 hours to reassess the CSF. This strategy also is appropriate for the nontoxic patient with a nondiagnostic LP. The point is, if you are unsure of the diagnosis (i.e., you cannot rule out ABM with certainty), you can observe the patient in the office or admit the patient, and perform a second LP 8-12 hours later to confirm the diagnosis. Patients in high-risk groups for fungal or tuberculous meningitis (HIV, malignancy, organ transplant) as well as those with abnormal CSF parameters should be admitted for observation.

Patients in whom the diagnosis of viral meningitis has been confirmed by CSF findings, and who have not received antibiotics, may go home if there is no alteration in mental status, rash, neurologic abnormalities, comorbid illnesses, or immunosuppressive drugs. Cultures should be followed, and, if clinical deterioration occurs, a repeat LP is mandatory. Treatment for viral meningitis consists of supportive care: treating and preventing dehydration, and managing pain and fever.

Chemoprophylaxis and Vaccination

The primary care physician is frequently the initial clinician to manage a patient with meningitis. Consequently, he or she must assume responsibility for preventing further infection in patient contacts. Chemoprophylaxis is currently recommended for high-risk groups exposed to N. meningitidis. These include household contacts (especially young children, and child care or nursery school contacts in the previous seven days), individuals with direct exposure to the index patient's secretions through kissing or sharing toothbrushes or utensils, mouth to mouth resuscitation or unprotected contact during endotracheal intubation in the seven days before onset of the illness, and people who frequently sleep or eat in the same dwelling as the index patient. Low-risk groups (no prophylaxis recommended) include casual contacts or indirect contact, and medical personnel not directly exposed to the index patient's oral secretions. Rifampin 10 mg/kg every 12 hours for four doses (max 600 mg) is the current drug of choice. Ceftriaxone, ciprofloxacin, and sulfasoxazole are alternatives.28 Chemoprophylaxis is also recommended for H. influenzae if there is another child younger than 4 years old in the same household as the index case. In this case, all members of the household should be treated with rifampin 20 mg/kg (max 600 mg) per day for four days, as well as all day care children contacts if there is more than one case of H. influenzae.8

Vaccination for Hib is recommended for all children. Vaccination for N. meningitidis is recommended for asplenic or complement deficient patients, travelers to highly endemic areas, military personnel, and during outbreaks in college dorms and household contacts. Vaccination against S. pneumoniae is recommended for patients older than 65 years, and in those with diabetes, chronic cardiac or lung disease, cancer immunosuppression, or asplenia. Currently, only 10% of patients eligible for vaccination against S. pneumoniae have actually received it.4 Patients with a CSF leak should receive all three vaccines.8 Advising patients who are eligible is the responsibility of all primary care physicians.


Meningitis is a potentially life-threatening condition that demands rapid diagnosis and treatment. Only patients with atypical presentations or focal neurologic signs require CT prior to LP. Antibiotic choice should be empiric, based on age and comorbid illness, unless the Gram's stain is available within 60 minutes. Morbidity and mortality still remain high, as a result of delayed diagnosis and delayed treatment of bacterial meningitis.


    1. Martin JB, Tyler KL, Scheld W M: Bacterial meningitis. In: Tyler KL, Martin JB, eds. Infectious Diseases of the Central Nervous System (Contemporary Neurology Series). Philadelphia: FA Davis; 1993:176-187.

    2. Schuchat A, Robinson K, Wenger JD, et al. Bacterial meningitis in the United States in 1995. N Engl J Med 1997;337:970-976.

    3. Spanos A, Harrell FE, Durack DT. Differential diagnosis of acute meningitis. JAMA 1989;262:2700-2707.

    4. Miller LG, Choi C. Meningitis in older patients: How to diagnose and treat a deadly infection. Geriatrics 1997;52:43-55.

    5. Nelsen S, Sealy DP, Schneider EF. The aseptic meningitis syndrome. Am Fam Phys 1993;48:809-815.

    6. Scheld WM. Bacterial meningitis in the patient at risk: Intrinsic risk factors and host defense mechanisms. Am J Med 1984;193-207.

    7. Townsend GC, Scheld W M. Review: The use of corticosteroids in the management of bacterial meningitis in adults. J Antimicrob Chemother 1996;37:1051-1061.

    8. Segreti J, Harris A. Acute bacterial meningitis. Infect Dis Clin North Am 1996;10:797-809.

    9. Tunkel AR, Scheld WM. Acute bacterial meningitis in adults. In: Remington JS, Swart MN, eds. Current Clinical Topics in Infectious Diseases. Boston: Blackwell Science; 1995:215-239.

    10. Maxson S, Jacobs RF. Viral meningitis. Postgrad Med 1993;93:153-166.

    11. Quagliarello VJ, Scheld WM. Treatment of bacterial meningitis. N Engl J Med 1997;336:708-716.

    12. Radetsky M. Duration of symptoms and outcome in bacterial meningitis: An analysis of causation and the implications of a delay in diagnosis. Pediatr Infect Dis J 1992;11:694-698.

    13. Talan DA, Zibulewsky J. Relationship of clinical presentation to time to antibiotics for the emergency department management of suspected bacterial meningitis. Ann Emerg Med 1993;22:1733-1738.

    14. Greenlee JE. Approach to diagnosis of meningitis. Infect Dis Clin North Am 1990;4:583-598.

    15. McKinney W, Heudebert GR, Harper SA, et al. Validation of a clinical prediction rule for the differential diagnosis of acute meningitis. J Geriatr Intern Med 1994;9:8-12.

    16. Ashwal S. Neurologic evaluation of the patient with acute bacterial meningitis. Neurol Crit Care 1995;13:549-577.

    17. Anderson M. Management of cerebral infection. J Neurol Neurosurg Psychiatry 1993;56:1243-1258.

    18. Maxson S, Jacobs RF. Viral meningitis. Postgrad Med 1993;93:153-166.

    19. Lipton JD, Schafermeyer RW. Central nervous system infections. Emerg Med Clin North Am 1995;13:417-443.

    20. Overall JC. Is it bacterial or viral? Laboratory differentiation. Pediatr Rev 1993;14:255-257.

    21. Rice SK, Heinl RE, Thornton LL, et al. Clinical characteristics, management strategies, and cost implications of a statewide outbreak of enterovirus meningitis. Clin Infect Dis 1995;20:931-937.

    22. Varki AP, Puthuran P. Value of second lumbar puncture in confirming a diagnosis of aseptic meningitis. A prospective study. Arch Neurol 1979;36:581-582.

    23. Feigin RD, Shackelford PG. Value of repeat lumbar puncture in the differential diagnosis of meningitis. N Engl J Med 1973;289:571-573.

    24. Amir J, Harel L, Frydman M, et al. Shift of cerebrospinal polymorphonuclear cell percentage in the early stage of aseptic meningitis. J Pediatr 1991;119:938-941.

    25. Powers WJ. Cerebrospinal fluid lymphocytosis in acute bacterial meningitis. Am J Med 1985;79:216-219.

    26. McIntyre PB, Berkey CS, King SM, et al. Dexamethasone as adjunctive therapy in bacterial meningitis. JAMA 1997;278:925-931.

    27. Klugman KP. Bactericidal activity against cephalosporin-resistant S. pneumoniae in CSF of children with acute bacterial meningitis. Antimicrob Agents Chemother 1995;39:1988-1992.

    28. American Academy of Pediatrics (Section 3). In: Peter G, ed. 1997 Red Book: Report of the Committee on Infectious Diseases. 24th ed.1997:222-3;359-360.