What it is:an infection of the meninges and the CSF that bathes the brain and spinal cord, most dangerously from bacteria, but also viral, fungal, or TB in origin. This chapter is really about bacterial meningitis, because it's the form that kills or deafens someone within hours if you're slow.
The core problem:the same blood-brain and blood-CSF barriers that protect the brain from everyday toxins also keep most antibiotics out. You're fighting an infection in a compartment your drugs weren't built to reach, and every hour of delay before the first antibiotic dose costs neurons.
What you do about it:the moment bacterial meningitis is suspected, give empiric antibiotics chosen by age and likely pathogen (plus dexamethasone right before or with that first dose) within an hour, don't wait on the LP or imaging. Narrow once culture and susceptibility data come back.
Think of this chapter as two problems stacked on top of each other: a timing problem(you must treat before you have proof) and a penetration problem(your drug has to cross a barrier designed to keep it out, and only does so well once the meninges are already inflamed). Almost everything else in the chapter, the "give the first dose within an hour" rule, the drug choices favoring cephalosporins over aminoglycosides, the idea that inflammation itself opens the door for penetration, all traces back to those two problems.
Before culture results exist, the CSF profile itself is often the first real clue to what you're dealing with. Learn this table cold, it's the single highest-yield table in the chapter.
| Parameter | Normal | Bacterial | Viral | Fungal | Tuberculous |
|---|---|---|---|---|---|
| WBC (cells/mm³) | <5 (<30 in newborns) | 1000–5000 | 50–1000 | 20–500 | 25–500 |
| Predominant cell | Monocytes | Neutrophils | Lymphocytes | Lymphocytes | Lymphocytes |
| Protein (mg/dL) | <50 (<500 mg/L) | Elevated | Mild elevation | Elevated | Elevated |
| Glucose (mg/dL) | 45–80 | Low | Normal | Low | Low |
| CSF/blood glucose ratio | 50%–60% | Decreased | Normal | Decreased | Decreased |
Polymorphonuclear pleocytosis + protein >50 mg/dL (500 mg/L) + CSF glucose <50% of the simultaneous serum glucosetogether point hard at bacterial meningitis. Notice bacterial is the only category with a genuinely highWBC count (1000–5000) and the only one dominated by neutrophils instead of lymphocytes, that neutrophil predominance is your fastest visual differentiator on a differential.
Tuberculous meningitis can show a lymphocytic-to-neutrophilic shift afterstarting antituberculous treatment, a so-called "therapeutic paradox." If a TB meningitis patient's differential looks more bacterial a few weeks into therapy, that's not necessarily treatment failure, it can be the expected paradox.
Bacterial meningitis develops through a predictable sequence: (1)nasopharyngeal mucosal colonization and invasion, (2)bacteremia and hematogenous spread (though direct spread from sinusitis or otitis media explains the high rate of pneumococcal meningitis in those patients too), (3)bacterial replication in the subarachnoid space, and (4)a progressive inflammatory cascade that raises intracranial pressure (ICP) and causes cerebral edema, which is what actually damages neurons.
A common thread among the classic CNS pathogens (H. influenzae, E. coli, N. meningitidis) is a thick polysaccharide capsulethat resists neutrophil phagocytosis and complement opsonization, that capsule is a big part of why these organisms make it past the host's first line of defense at all. Passive or active cigarette smoke exposure and cochlear implants that include a positioner both raise the risk of bacterial meningitis.
This is the part your CNS pharmacology lecture drilled into: the brain sits behind two distinct barriers, not one.
Meningeal inflammation itself increases antibiotic penetrationinto CSF, because the same cytokine-driven barrier breakdown that's hurting the patient is also the thing letting your drug in. That's exactly why Table 36-3's drug list splits into "penetrates regardless of inflammation" versus "only penetrates when the meninges are inflamed." It also explains why finishing a full course matters: as inflammation resolves with treatment, penetration drops, so stopping early risks under-dosing a partially treated infection.
Bacterial cell lysis releases wall components (lipopolysaccharide/lipid A endotoxin, lipoteichoic acid, teichoic acid, peptidoglycan, depending on gram-positive versus gram-negative). These trigger capillary endothelial cells and CNS macrophages to release IL-1, TNF, and other cytokines. Proteolytic products and toxic oxygen radicals break down the blood-brain barrier, platelet-activating factor kicks off coagulation, and arachidonic acid metabolites drive vasodilation. The end result: cerebral edema, elevated ICP, CSF pleocytosis, decreased cerebral blood flow, ischemia, and if unchecked, death.
Antibiotic-induced bacterial lysis releases even moreof these cell wall components acutely, which is part of why adjunctive dexamethasone (given before or with the first antibiotic dose, never after) exists: it dampens the host's own inflammatory overreaction to that lysis, not the bacteria themselves.
Classic acute bacterial meningitis presents with fever, nuchal rigidity, altered mental status, chills, vomiting, photophobia, and severe headache. Up to 95% of patients have at least twoof: fever, nuchal rigidity, headache, and altered mental status, that "two out of four" pattern is a useful screening heuristic when the full classic triad isn't present.
| Finding | Notes |
|---|---|
| Kernig's sign | Resistance/pain on passive knee extension with the hip flexed, a marker of meningeal irritation |
| Brudzinski's sign | Passive neck flexion causes involuntary hip and knee flexion |
| Nuchal rigidity | Involuntary neck stiffness resisting passive flexion, the classic bedside finding |
Both signs are poorly sensitive and frequently absent in children. A negative Kernig's or Brudzinski's sign does not clear a patient of meningitis, especially a pediatric one. In infants, look instead for a bulging fontanelle, apnea, purpuric rash, and convulsions, the exam picture in young children is genuinely different from adults.
Purpuric and petechial skin lesions typically indicate N. meningitidis, sometimes with an accompanying picture of disseminated intravascular coagulation (DIC). That said, the same rash can appear with H. influenzae meningitis, and rashes are rare with pneumococcal meningitis. Use the rash to raise suspicion for meningococcus, not to exclude other organisms.
LP(lumbar puncture) is how you get the sample. ICP(intracranial pressure) is what you're worried about raising the whole time. A lumbar puncture is the essential step for establishing the diagnosis, identifying the organism, and getting susceptibility data, but it never delays the first dose of antibiotic (see Treatment below).
Goals:eradicate the infection, resolve signs and symptoms, and cut morbidity and mortality, plus prevent disease in the first place through timely vaccination and chemoprophylaxis of contacts.
Time from suspected diagnosis to the first dose of antibiotic should not exceed 1 hour.The first dose is never withheldfor a delayed LP or pending neuroimaging. If you're waiting on a head CT before your LP (say, to rule out a mass lesion first), draw blood cultures and give empiric antibiotics while that workup happens, don't let logistics eat into the golden hour.
You will not know the organism when you write this order. Age is your best proxy for the likely pathogen, and it's how the empiric regimen is chosen. Every recommendation below is rated A-III(good evidence supporting use, based on expert opinion/clinical experience rather than randomized trial data specific to this indication).
| Age | Most likely organisms | Empiric regimen |
|---|---|---|
| <1 month | S. agalactiae, gram-negative enterics (E. coli, Klebsiella, Enterobacter), L. monocytogenes | Ampicillin + cefotaxime orampicillin + an aminoglycoside |
| 1–23 months | S. pneumoniae, N. meningitidis, H. influenzae, S. agalactiae | Vancomycin + third-generation cephalosporin (cefotaxime or ceftriaxone) |
| 2–50 years | N. meningitidis, S. pneumoniae | Vancomycin + third-generation cephalosporin (cefotaxime or ceftriaxone) |
| >50 years | S. pneumoniae, N. meningitidis, gram-negative enterics, L. monocytogenes | Vancomycin + ampicillin + third-generation cephalosporin (cefotaxime or ceftriaxone) |
Ampicillin is the only regimen ingredient covering Listeria monocytogenes, and Listeria risk is concentrated at the two ends of life, neonates and adults over 50. That's why it's in the <1 month and >50 year regimens but drops out for the 1 month to 50 year range, where meningococcus and pneumococcus dominate and Listeria risk is low.
Vancomycin's role is to cover penicillin/cephalosporin-resistant S. pneumoniaeuntil susceptibility data return, not to independently treat pneumococcus. Its continued use should be guided by local resistance ratesand stopped or continued based on the actual cefotaxime/ceftriaxone MIC once it's back.
An antibiotic can be perfectly active against the organism in a culture plate and still fail clinically if it can't reach therapeutic concentrations in CSF. This table (compared against target-pathogen MIC, using recommended CNS dosing) is why the drug choices above look the way they do.
Therapeutic levels with or without meningeal inflammation(the most reliable group): acyclovir, chloramphenicol, ciprofloxacin, fluconazole, flucytosine, foscarnet, fosfomycin, ganciclovir, isoniazid, levofloxacin, linezolid, metronidazole, moxifloxacin, pyrazinamide, rifampin, sulfonamides, trimethoprim, voriconazole.
Therapeutic levels only when meninges are inflamed(this is most beta-lactams, so timing and ongoing inflammation genuinely matter for these): ampicillin ± sulbactam, aztreonam, cefepime, cefotaxime, ceftazidime, ceftriaxone, cefuroxime, colistin, daptomycin, imipenem, meropenem, nafcillin, ofloxacin, penicillin G, piperacillin/tazobactam, pyrimethamine, quinupristin/dalfopristin, ticarcillin ± clavulanic acid, vancomycin.
Nontherapeutic regardless of inflammation(avoid for meningitis no matter how good the in-vitro susceptibility looks): aminoglycosides, amphotericin B, first- and second-generation cephalosporins (cefuroxime is the one second-gen exception, it's in the middle group above), beta-lactamase inhibitors alone (clavulanic acid, sulbactam, tazobactam), doxycycline (except for documented B. burgdorferi), itraconazole (except reaching therapeutic levels for Cryptococcus neoformans).
Tazobactam does not cross the blood-brain barriereven though piperacillin does, and piperacillin/tazobactam may not achieve therapeutic levels against higher-MIC organisms like P. aeruginosa. And when combinations like pip/tazo are used for CNS infection, you're really relying on the piperacillin component, tazobactam is along for the ride systemically but contributes nothing in the CSF.
Aminoglycosides sit in the nontherapeuticgroup regardless of inflammation, that's why the neonatal regimen pairs ampicillin with an aminoglycoside for systemicListeria coverage rather than relying on the aminoglycoside for CSF penetration, and why gentamicin is always an add-onto a penicillin for Listeria or Group B strep, never monotherapy. Daptomycin technically makes the "therapeutic with inflammation" list here, but in practice it's not a first-line CNS choice; the safer teaching point is that first-generation cephalosporins and aminoglycosides are dead ends for meningitis no matter how active they look on paper.
Once culture and susceptibility data are in hand, switch off the age-based empiric regimen and match the drug to the confirmed organism.
| Organism | First choice | Alternative | Duration |
|---|---|---|---|
| Streptococcus pneumoniae | |||
| Penicillin susceptible (MIC ≤0.06 mcg/mL) | Penicillin G or ampicillin | Cefotaxime, ceftriaxone, cefepime, or meropenem | 10–14 days |
| Penicillin resistant (MIC >0.06 mcg/mL) | Vancomycin + cefotaxime or ceftriaxone | Moxifloxacin | |
| Ceftriaxone resistant (MIC >0.5 mcg/mL) | Vancomycin + cefotaxime or ceftriaxone | Moxifloxacin | |
| Staphylococcus aureus | |||
| Methicillin susceptible | Nafcillin or oxacillin | Vancomycin or meropenem | 14–21 days |
| Methicillin resistant | Vancomycin | Trimethoprim-sulfamethoxazole or linezolid | |
| Other gram-positives | |||
| Group B Streptococcus | Penicillin G or ampicillin ± gentamicin | Ceftriaxone or cefotaxime | 14–21 days |
| S. epidermidis | Vancomycin | Linezolid | 14–21 days |
| L. monocytogenes | Penicillin G or ampicillin ± gentamicin | Trimethoprim-sulfamethoxazole or meropenem | ≥21 days |
European guidelines suggest considering adding rifampinto vancomycin for resistant pneumococcus, and adding gentamicin for the first 7 daysof Listeria treatment. Serum drug level monitoring is recommended whenever an aminoglycoside is on board.
| Organism | First choice | Alternative | Duration |
|---|---|---|---|
| Neisseria meningitidis | |||
| Penicillin susceptible | Penicillin G or ampicillin | Cefotaxime or ceftriaxone | 7–10 days |
| Penicillin resistant | Cefotaxime or ceftriaxone | Meropenem or moxifloxacin | |
| Haemophilus influenzae | |||
| β-lactamase negative | Ampicillin | Cefotaxime, ceftriaxone, cefepime, or moxifloxacin | 7–10 days |
| β-lactamase positive | Cefotaxime or ceftriaxone | Cefepime or moxifloxacin | |
| Enterobacteriaceae (E. coli, Klebsiella) | Cefotaxime or ceftriaxone | Cefepime, moxifloxacin, meropenem, or aztreonam | 21 days |
| Pseudomonas aeruginosa | Cefepime or ceftazidime ± tobramycin | Ciprofloxacin, meropenem, piperacillin + tobramycin, colistin sulfomethate, or aztreonam | 21 days |
Colistinshould be reserved for multidrug-resistant Pseudomonas or Acinetobacter once every other option has been exhausted. Direct CNS administration (intrathecal/intraventricular) may be considered for gram-negative infections that fail conventional systemic treatment.
Meningococcus and H. influenzae:7–10 days. Pneumococcus:10–14 days. Staph aureus and Group B strep:14–21 days. Enterobacteriaceae and Pseudomonas:21 days. Listeria:≥21 days, the longest of all. The pattern roughly tracks how hard the organism is to fully clear from the CNS compartment, and duration should still be individualized to clinical response.
Dexamethasone doesn't treat the infection, it immunomodulates the inflammatory responsethat's doing the actual neurologic damage (see Pathophysiology above).
| Population | Recommendation |
|---|---|
| Infants/children ≥6 weeks, H. influenzae meningitis | IV 0.15 mg/kg every 6 hours for 2–4 days |
| Infants/children, pneumococcal meningitis | May be considered after weighing potential benefits against risks |
| Adults, suspected or proven pneumococcal meningitis | 0.15 mg/kg (up to 10 mg) every 6 hours for 2–4 days |
The first dexamethasone dose must be given 10–20 minutes prior to, or concomitant with, the first dose of antibiotic, never after. The entire rationale depends on blunting the inflammatory burst from bacterial lysis before it happens; giving steroids after the antibiotics have already been running defeats the purpose.
In infants and children with Hib meningitis, adjunctive dexamethasone reduces the risk of hearing loss, a complication that otherwise shows up disproportionately with this organism.
A leading cause of bacterial meningitis in children and young adults worldwide, spread person-to-person via respiratory droplets and pharyngeal secretions. Petechiae are often the clue that points to this organism, and patients can have an obvious or subclinical picture of disseminated intravascular coagulation. Unilateral or bilateral deafnesscan develop early or late in the disease course, an underappreciated complication.
Empiric treatment:third-generation cephalosporins (cefotaxime, ceftriaxone). Penicillin G or ampicillin if the isolate is penicillin-susceptible. Duration is typically 7 dayswith a good clinical response.
Close contacts need antimicrobial chemoprophylaxis as soon as possible, ideally within 24 hoursof identifying the index case. Ciprofloxacin and rifampinare the two most commonly used agents. Waiting days for this defeats the purpose, secondary cases cluster early.
Coma, hearing impairment, and seizuresare common neurologic complications with this organism. Penicillin should never be used as empiric therapyif pneumococcal meningitis is suspected, resistance is too common to gamble on it up front.
Ceftriaxone and cefotaxime serve as the alternatives to penicillin for penicillin-nonsusceptible strains. For cephalosporin-resistantpneumococci, add vancomycin (and rifampin to a lesser extent). Vancomycin plus ceftriaxoneis a reasonable empiric combination while susceptibility data are pending. IV linezolid, daptomycin, and ceftaroline have also emerged as options for multidrug-resistant gram-positive CNS infections.
Widespread infant and childhood vaccination has sharply cut Hib meningitis incidence in the 1 month to 5 year age range, and driven down bacterial meningitis rates overall.
Third-generation cephalosporins (cefotaxime, ceftriaxone)are drugs of choice for empiric therapy since they cover both β-lactamase-producing and non-producing strains. Cefepime and fluoroquinolones are suitable alternatives regardless of β-lactamase status. Duration is 7 days in adults, 7–10 days in children. Dexamethasone reduces hearing-loss risk when given before or with the first antibiotic dose (see above).
Rifampinreduces secondary invasive Hib disease in close contacts by eliminating nasopharyngeal and oropharyngeal carriage. Dosing: 20 mg/kg/dose (max 600 mg) orally once daily for 4 days.Who specifically needs it is guided by American Academy of Pediatrics recommendations.
Implicated in roughly 10% of meningitis cases in patients over 65, and carries a case-fatality rate around 18%in the US, meaningfully worse than most other bacterial causes. This is exactly why ampicillin reappears in the empiric regimen for both neonates and adults over 50.
Treatment:penicillin G or ampicillin, with an aminoglycoside added for proven infection in both children and adults. Treat for a minimum of 21 days, the longest course of any organism in this chapter. Trimethoprim-sulfamethoxazole and meropenem are effective alternatives since both achieve adequate CSF penetration.
| Parameter | When | Watching for |
|---|---|---|
| Clinical response(mental status, fever curve, meningeal signs) | Continuously through the first 48–72 hours and beyond | Improvement supporting the empiric regimen, or lack of it prompting a re-think |
| Culture and susceptibility results | As soon as available | Opportunity to switch from empiric to pathogen-directed, narrower therapy |
| Repeat LP / CSF profile | If clinical response is poor, or organism is resistant/unusual | Persistent pleocytosis, failure of glucose/protein to normalize |
| Aminoglycoside or vancomycin levels | Whenever either is part of the regimen | Therapeutic, non-toxic serum concentrations, since neither reliably penetrates CSF and both carry nephro/ototoxicity risk |
| Hearing assessment | After treatment, especially in children with Hib or pneumococcal meningitis | Sensorineural hearing loss, a common sequela |
| Neurologic exam | Throughout admission and at follow-up | Seizures, focal deficits, signs of raised ICP |