Health care-associated meningitis and ventriculitis in adults: Treatment and prognosis

INTRODUCTION — Health care-associated meningitis and ventriculitis can occur as a complication of neurosurgery, placement of cerebrospinal fluid (CSF) shunt, intrathecal pump, or deep brain stimulator, and less frequently, following dural puncture. Early recognition and treatment are crucial to reduce morbidity and long-term complications.

Meningitis and ventriculitis can occur on a continuum and often do in patients with health care-associated infection, particularly in those who have undergone neurosurgical procedures or have an indwelling ventricular device.

The treatment and prognosis of health care-associated meningitis and ventriculitis are reviewed in this topic. The epidemiology, microbiology, clinical manifestations, and diagnosis are discussed in detail elsewhere. (See "Health care-associated meningitis and ventriculitis in adults: Clinical features and diagnosis".)

Information specific to CSF shunt infections is also presented elsewhere. (See "Infections of cerebrospinal fluid shunts".)

SYSTEMIC ANTIMICROBIAL THERAPY

Empiric therapy — Empiric antimicrobial therapy should be administered when clinical suspicion is high, ideally as soon as cerebrospinal fluid (CSF) and blood specimens have been collected for microbiologic and other testing. (See "Health care-associated meningitis and ventriculitis in adults: Clinical features and diagnosis", section on 'Timing and components of initial evaluation'.)

We suggest one of the following empiric regimens (the dosing listed is appropriate for patients with normal renal and hepatic function) [1]:

●Vancomycin – loading dose of 20 to 35 mg/kg followed by 15 to 20 mg/kg intravenously (IV) every 8 to 12 hours (not to exceed 2 g per dose or a total daily dose of 60 mg/kg).

Plus

●One of the following antipseudomonal beta-lactams:

•Ceftazidime – 2 g IV every eight hours

Or

•Cefepime – 2 g IV every eight hours

Or

•Meropenem – 2 g IV every eight hours

The suggested empiric regimens are intended to cover the expected microbiologic spectrum of health care-associated meningitis and ventriculitis (see "Health care-associated meningitis and ventriculitis in adults: Clinical features and diagnosis", section on 'Microbiology'), with activity against gram-positive skin flora, especially S. aureus and alpha- and beta-hemolytic streptococci, as well as gram-negative aerobic bacilli, especially Klebsiella pneumoniae and Pseudomonas aeruginosa. The regimens also take into account the prevalence of drug resistance among health care-associated pathogens. These suggested agents are also expected to penetrate the CSF well.

The patient’s prior history of infection and local or institutional rates of resistance should also inform selection among the antipseudomonal beta-lactams. Meropenem is the preferred choice for empiric therapy in patients who have a history of infection with an extended-spectrum beta-lactamase (ESBL)-producing gram-negative bacillus or at institutions that have a high prevalence of such pathogens.

Patients with penicillin allergies can often safely take the later-generation cephalosporins listed above or carbapenems (eg, meropenem), in some cases following a test dose procedure (algorithm 1). For patients with severe beta-lactam allergies (eg, anaphylaxis, Stevens-Johnson syndrome/toxic epidermal necrolysis [SJS/TEN], drug reaction with eosinophilia and systemic symptoms [DRESS], acute generalized exanthematous pustulosis [AGEP]) for whom meropenem is contraindicated, aztreonam (2 g IV every six to eight hours) or ciprofloxacin (400 mg IV every 8 to 12 hours) can replace the cephalosporin or carbapenem, although clinical experience with these agents for health care-associated meningitis and ventriculitis is limited.

Tailoring antimicrobial therapy

Organism identified on CSF culture — Once a specific organism is identified and susceptibility testing resulted, the empirically chosen antimicrobial regimen should be adjusted to target that organism [2]. The recommended treatment regimens for specific bacterial pathogens are discussed in detail separately (table 1). (See "Treatment of bacterial meningitis caused by specific pathogens in adults".)

For treatment of pathogens that are extensively drug resistant (eg, susceptible to polymyxins only), intrathecal or intraventricular therapy is a potential adjunct to systemic antibiotic therapy. (See 'Intrathecal and intraventricular therapy' below.)

Antimicrobial treatment of fungal central nervous system (CNS) infections is discussed separately. (See "Candida infections of the central nervous system" and "Treatment and prevention of invasive aspergillosis", section on 'Antifungal therapy' and "Aseptic meningitis in adults", section on 'Fusarium outbreaks'.)

Negative CSF culture — When the CSF culture is negative, ongoing management should be individualized and depends on the clinical suspicion for infection based on the remainder of the laboratory evaluation and the clinical status. Additional evaluation and features that raise or lower suspicion for infection are discussed in detail elsewhere. (See "Health care-associated meningitis and ventriculitis in adults: Clinical features and diagnosis", section on 'When clinical and microbiologic features are discordant'.)

●If the clinical suspicion for bacterial infection remains high despite negative CSF microbiology, the empirically chosen regimen can be continued provided the patient is clinically improving. For those who have suboptimal clinical improvement, the approach is discussed elsewhere. (See 'Approach to poor response' below.)

●If clinical suspicion for bacterial infection is not high (eg, stable CSF pleocytosis without other features of bacterial infection) and CSF cultures are negative, it is reasonable to discontinue antibiotic therapy in some cases. In particular, if the patient did not receive antibiotics prior to CSF sampling, the CSF culture remains negative after two to three days, and no other features are concerning for bacterial infection, antibiotics may be cautiously discontinued [3]. In one observational study of 75 patients with postoperative meningitis and no intracranial devices, stopping antibiotics after 72 hours if CSF cultures were negative (and no other wound or blood cultures were positive in the setting of an elevated CSF leukocyte count) was a safe approach [4]. All episodes of meningitis resolved (regardless of whether antibiotic therapy was discontinued), and complications were rare. However, we do not take this approach if patients have received antibiotic therapy prior to obtaining CSF cultures, as negative CSF cultures could reflect a partially treated infection.

Decisions regarding continuation of antibiotic therapy can be particularly challenging in health care-associated meningitis and ventriculitis, since up to 50 percent of patients have negative CSF cultures [5], and the diagnosis must be made based on clinical judgement and a synthesis of other features. (See "Health care-associated meningitis and ventriculitis in adults: Clinical features and diagnosis", section on 'When clinical and microbiologic features are discordant'.)

Duration — The optimal duration of antimicrobial therapy for health care-associated meningitis and ventriculitis is uncertain and depends on the causative pathogen. In general, the minimum duration is 10 to 14 days. For gram-negative bacillary meningitis or ventriculitis, some experts treat for at least 21 days because of concern that shorter courses are associated with high rates of relapse, although there are no clear data indicating that this is necessary with contemporary antibiotic regimens [6]. Treatment durations for specific pathogens are discussed elsewhere. (See "Treatment of bacterial meningitis caused by specific pathogens in adults", section on 'Therapy for specific pathogens'.)

Ultimately, the length of therapy should also be tailored to the observed response of individual patients. In patients with repeatedly positive CSF cultures while on appropriate antimicrobial therapy, we generally continue treatment until 10 to 14 days following the last positive culture [1]. There are no specific systematic data to support this approach, which is mainly to try to ensure complete eradication of the organism.

MONITORING AND SUBSEQUENT CARE

Follow-up CSF testing in selected cases — We do not routinely repeat cerebrospinal fluid (CSF) testing following initiation of antibiotic therapy for health care-associated meningitis and ventriculitis if the patient is improving and systemic signs of infection are resolving. Indications for repeat CSF analysis and cultures include the following:

●Lack of clinical improvement several days (eg, five to seven days) after initiation of appropriate antibiotic therapy. (See 'Approach to poor response' below.)

●When a negative CSF culture is required for proof of cure prior to discontinuing antibiotics or a neurosurgical procedure. The most common situation is following explantation of an infected device, when documented CSF clearance is needed prior to reimplantation. (See "Infections of cerebrospinal fluid shunts", section on 'Monitoring the response to therapy' and "Infections of cerebrospinal fluid shunts", section on 'Timing of new shunt placement'.)

There are no definitive data informing the ideal timing to resample the CSF to evaluate for microbial clearance; the average duration of appropriate antibiotic therapy needed to sterilize the CSF is two to four days in adults [7]. However, the time to negative CSF cultures may be longer for some pathogens. In some cases of gram-negative bacillary meningitis, CSF cultures may remain positive for as long as 13 days after treatment initiation despite appropriate therapy and apparent clinical response [7,8]. If an isolate grows on repeat CSF culture, susceptibility testing should also be repeated to ensure that the isolate remains susceptible to the chosen antibiotic regimen. Emergent resistance during therapy has been reported in patients with gram-negative bacillary meningitis [9].

Approach to poor response — Individuals who have poor clinical response or have persistently positive CSF cultures despite five to seven days of appropriate antibiotic therapy should be managed in consultation with experts in infectious diseases. Initial approach should include:

●Repeat CSF culture and susceptibility testing and adjustment of antimicrobial therapy, if appropriate – When available, the antibiotic resistance pattern of the causative organism should guide choice of antibiotics. Since emergent resistance during therapy has been reported, the antibiotic regimen may need to be tailored to a more active agent. If susceptibility testing results are not available, institutional antibiotic resistance patterns should be used to inform antibiotic adjustments.

●Repeat neuroimaging to identify complications and need for potential intervention – Loculated infection or abscess can form around surgical sites, indwelling catheters or other implanted hardware. Antibiotic therapy alone often cannot cure such complications, which may require surgical drainage or removal of hardware. Although magnetic resonance imaging (MRI) is more sensitive, metal hardware can distort field of interest, in which case computed tomography (CT) imaging is preferred. With either modality, imaging with and without contrast agent is most sensitive.

●Adjunctive intrathecal or intraventricular antibiotics in selected cases – This is discussed in detail elsewhere. (See 'Intrathecal and intraventricular therapy' below.)

OTHER MANAGEMENT CONSIDERATIONS

Intrathecal and intraventricular therapy

Indications — Adjunctive intrathecal (IT) or intraventricular (IVT) antibiotics are not routinely necessary for health care-associated meningitis and ventriculitis, and we rarely use them for initial therapy. Their main potential role is when traditional intravenous antibiotics are not optimally effective and IT or IVT administration would be anticipated to provide higher drug levels in the CSF. Circumstances for which IT or IVT administration might be useful as adjunctive therapy to intravenous (IV) antibiotics include:

●Infections refractory to appropriate systemic antibiotic therapy (ie, poor clinical response with ongoing pathogen growth in the cerebrospinal fluid [CSF] despite five to seven days of treatment with a regimen expected to be active against the pathogen). This scenario raises concern about inadequate systemic drug penetration into the central nervous system (CNS).

●Infections due to extensively drug-resistant organisms (eg, carbapenem-resistant Acinetobacter baumannii or Pseudomonas aeruginosa resistant to standard antipseudomonal agents). In particular, IT or IVT antibiotics may be useful for pathogens for which polymyxins are the only active systemic antibiotic, since intravenously administered colistin achieves suboptimal CSF levels [10].

●In patients with implanted devices or drains that cannot be removed.

In such cases, decisions to use IT/IVT antibiotics should be individualized. In patients who do not have existing ventricular or intrathecal access, the risks of obtaining access may outweigh potential benefits.

Administration of antibiotics through the IT or IVT route does not require passage through the blood-brain barrier and is thus thought to achieve high drug concentrations at the site of infection while avoiding the toxicity of systemic doses that would be needed to achieve adequate CSF levels. Evidence supporting IT or IVT antibiotics is limited to small observational studies [11-17]. However, these low-quality data suggest a potential benefit of adjunctive IT or IVT therapy.

In a systematic review of 11 case series and other small observational studies that included 348 patients with health care-associated gram-negative meningitis and ventriculitis, adjunctive IT or IVT antibiotics were associated with a higher likelihood of CSF culture clearance and a lower overall mortality compared with systemic antibiotics alone, but the differences were not statistically significant [13]. The association between adjunctive IT or IVT therapy and improved outcomes may have been stronger for carbapenem-resistant infections. In another study of 31 patients with post-neurosurgical gram-negative meningitis that was not included in the review, intraventricular gentamicin was associated with a lower relapse rate (0 of 13 versus 6 of 18 who received systemic antibiotics alone) [15]. In a small trial of 10 patients with staphylococcal ventriculitis associated with an external ventricular drain who were randomly assigned to receive vancomycin through the IV or IVT route, IVT vancomycin resulted in higher CSF levels but microbiologic clearance was achieved by all patients in both groups within three to four days [14].

Overall, major toxicity has not been reported with antibiotics administered through the IT or IVT route. Specific reported adverse effects with different agents are discussed below.

Dosing and administration — Evidence informing the optimal dosing and administration of IT or IVT antibiotics is limited. Management with IT or IVT antibiotics should be undertaken in consultation with specialists in infectious diseases and neurosurgery. It may also be useful to enlist input from a pharmacist who can ensure that the chosen dose and the diluents used for administration are appropriate.

Antimicrobials that can be administered through the IT or IVT route include vancomycin and daptomycin for gram-positive infections and aminoglycosides (eg, gentamicin) and polymyxins (colistin and polymyxin B) for gram-negative infections (table 2). The choice among them depends on the susceptibility profile of the causative organism.

●Route of administration – IT or IVT antibiotics can be administered through a lumbar puncture, lumbar drain, ventriculostomy, or external ventricular drainage system, such as an Ommaya reservoir. For patients with evidence of ventriculitis, we favor IVT over IT administration because it may be difficult to deliver adequate drug concentrations throughout the ventricular system via the IT route [18]. Additionally, if an IVT delivery system is already in place, IVT administration is more practical. However, for patients with meningitis without ventriculitis, IT and IVT administration of most drugs have been associated with similar microbiologic clearance rates [19].

Many individuals with health care-associated meningitis and ventriculitis need ventricular drainage for other reasons (eg, noncommunicating hydrocephalus or the underlying condition). For individuals without existing ventricular or intrathecal access, the decision to place a device for the purpose of administering antibiotics depends on the availability of neurosurgical expertise, whether there are potential alternative options for systemic antibiotic treatment, and the clinical stability of the patient; in some cases, the risks of obtaining access may outweigh potential benefits of IVT antibiotics. If feasible, serial lumbar punctures for antibiotic delivery is an alternative.

When antimicrobial therapy is administered via a ventricular drain, the drain should be clamped for 15 to 60 minutes to allow the agent to equilibrate throughout the CSF [1]. A small number of case reports suggest clamping the tubing for one hour after administration to prevent dilution if there is a large volume of CSF drainage [20-22].

●Dosing – Standard doses have not been established for IT and IVT antibiotics; we agree with doses recommended by the Infectious Diseases Society of America (IDSA) (table 2), which also suggests that doses and intervals be adjusted to achieve antimicrobial concentrations in the CSF that are 10 to 20 times the MIC of the causative pathogen [1]. CSF concentrations obtained with the same IVT dose have been highly variable in pharmacokinetic studies, likely because of individual differences in volume of distribution, ventricular size, or CSF clearance as a result of CSF drainage [1]. Careful calculation of dose and attention to the diluents used to administer IT or IVT therapy are critically important issues when antibiotics are given through these routes [23,24].

The same dose ranges listed for IVT administration below are also used for IT administration:

•Vancomycin – IVT doses of vancomycin range from 5 to 20 mg total daily dose, depending in part on the size of the ventricles [1,25]. Most reports have described daily dosing, although the frequency of administration may also depend on the rate of ventricular drain output (table 2).

Case series suggest that these doses are generally well tolerated. In a systematic review of 17 studies including over 90 patients receiving IVT vancomycin, no clear adverse effects were reported, even when CSF vancomycin levels >100 mg/mL were reported [25].

•Daptomycin – Described doses of IVT daptomycin range from 2 to 5 mg total daily dose, given once daily (table 2) [1]. Higher daily doses administered every two or three days have also been described. Published reports of IVT daptomycin use are rare, but transient fever may be an adverse effect [26].

•Aminoglycosides – IDSA recommended total daily doses for IVT administration are 4 to 8 mg for gentamicin (in adults), 5 to 20 mg for tobramycin, and 5 to 50 mg for amikacin (table 2) [1]. These are largely consistent with dose ranges described in published reports (although doses of gentamicin up to 10 mg have also been reported); aminoglycosides were most commonly given once daily [27].

Case reports of adverse effects have included temporary hearing loss, aseptic meningitis, and CSF eosinophilia [27]. There is no clear relationship between CSF aminoglycoside levels and toxicity [8,16,23,24,27,28].

•Polymyxins – IDSA-recommended total daily doses for IVT administration are 10 mg colistimethate sodium (CMS) for colistin (equivalent to 125,000 international units CMS or 4.1 mg colistin base activity) and 5 mg for polymyxin B (table 2) [1,10]. Of the two, experience with IVT dosing is greater with colistin. Differences in the available drug formulations between different parts of the world may contribute to confusion around dosing and conversion of units [29]. Dosing of colistin administered should be assessed carefully, in consultation with a pharmacist. (See "Polymyxins: An overview", section on 'Dosing and administration'.)

IT/IVT therapy appears to lack the nephrotoxicity of IV colistin. However, in a systematic review of 234 cases of IT or IVT polymyxin therapy for gram-negative meningitis or ventriculitis, toxicity was reported in 16 cases (7 percent), mainly aseptic chemical meningitis or ventriculitis, but also sporadic cases of seizures, peripheral numbness, and cauda equina syndrome [10]. The neurotoxicity appears to be reversible with dose reduction or discontinuation [11].

Other antibiotics that have been administered through the IT or IVT route include tigecycline and quinupristin-dalfopristin [26]. Penicillins and cephalosporins should not be given IT or IVT because of the risk of neurotoxicity [1].

IT/IVT amphotericin is discussed elsewhere. (See "Pharmacology of amphotericin B", section on 'Intrathecal' and "Coccidioidal meningitis", section on 'Patients without a satisfactory response to initial therapy'.)

●Monitoring of CSF drug levels – The optimal approach to checking CSF drug levels to help inform dosing is uncertain. The IDSA notes that checking a level 24 hours after administration of the first dose likely reflects the trough concentration [1]. As above, targeting a trough CSF concentration that is 10 to 20 times the pathogen MIC for that drug is a reasonable approach.

●Duration – The optimal duration of IT/IVT antibiotics is uncertain but is typically 10 to 14 days. We continue them through microbiologic clearance of the CSF, barring difficulties in administration or other complication; if repeat CSF culture remains positive, we continue IT/IVT antibiotics for another 7 to 10 days.

In many reports, CSF culture clearance occurs within a week of IT/IVT administration. In one multicenter retrospective study of 105 patients who received IVT antibiotics, CSF cultures cleared in 88 percent with a median duration of administration of five to six days [16]. In other studies, the mean time to CSF clearance was three days overall and six days for patients with ventriculitis [19,30].

Longer durations may be warranted when polymyxins are used, which is typically in the setting of extensively drug-resistant pathogens. In one systematic review of 234 cases of IT/IVT polymyxin use, the mean duration was 18 days [10].

Device removal — Ideally, any CNS hardware associated with health care-associated meningitis and ventriculitis should be removed [1]. On the rare occasion that a device cannot be safely removed, the patient should be managed in consultation with an infectious diseases specialist. Other antimicrobial strategies, such as adjunctive intraventricular or intrathecal antibiotics or a longer antimicrobial course, may be warranted. (See 'Intrathecal and intraventricular therapy' above.)

Special consideration is warranted for hardware that is necessary to control intracranial hypertension:

●For patients who have an external ventricular drain (EVD) and have suspected but not yet confirmed ventriculitis, we sometimes keep the EVD in place while empirically treating with antibiotics, then remove and replace the EVD if CNS infection is confirmed.

●The approach to CSF shunt removal and timing of replacement are discussed in detail elsewhere. (See "Infections of cerebrospinal fluid shunts", section on 'Device removal' and "Infections of cerebrospinal fluid shunts", section on 'Timing of new shunt placement'.)

The rationale for device removal is to reduce the risk of treatment failure or relapse. In series of patients with intrathecal infusion pumps or deep brain stimulation devices, almost all patients required removal of the infected devices for cure of infection [31,32].

In particular, S. aureus and some coagulase-negative staphylococci, such as S. epidermidis, form biofilm that enhances adherence of the bacteria to prosthetic material, which decreases the likelihood of successful antibiotic treatment if the device remains in place [33].

Reducing intracranial pressure — Patients with health care-associated meningitis and ventriculitis are at risk for elevated intracranial pressure and resulting complications. Evaluation for and management of elevated intracranial pressure are discussed elsewhere. (See "Initial therapy and prognosis of community-acquired bacterial meningitis in adults", section on 'Reduction of intracranial pressure' and "Evaluation and management of elevated intracranial pressure in adults".)

Interventions with no or limited role

Dexamethasone — We suggest against routine use of dexamethasone for management of inflammation or elevated intracranial pressure in patients with health care-associated meningitis or ventriculitis.

Although dexamethasone has been associated with improved neurologic outcomes in selected children and adults with acute bacterial meningitis, trials that have suggested a benefit have generally not included individuals with health care-associated or gram-negative infections, and the pathogens most associated with a benefit from dexamethasone are not common causes of health care-associated infections. These data are discussed in detail elsewhere. (See "Dexamethasone to prevent neurologic complications of bacterial meningitis in adults" and "Bacterial meningitis in children: Dexamethasone and other measures to prevent neurologic complications".)

Clinical data on dexamethasone therapy for health care-associated meningitis and ventriculitis are limited. Systematic studies evaluating adjunctive steroid therapy for gram-negative meningitis or ventriculitis are unlikely to be performed since it is uncommon and affected patients often have comorbid conditions that are either confounding or could be adversely affected by dexamethasone [34]. In addition, animal models of gram-negative meningitis have suggested a potential harm of dexamethasone [35].

Induced hypothermia — It is unknown whether induced hypothermia offers any benefit for patients with severe meningitis or ventriculitis and we recommend against it for such patients. Evidence informing the effect of targeted temperature management for severe meningitis is limited, inconclusive, and discussed in detail elsewhere. (See "Initial therapy and prognosis of community-acquired bacterial meningitis in adults", section on 'Induced hypothermia'.)

PROGNOSIS — Even when appropriate antibiotic therapy is administered, health care-associated meningitis and ventriculitis are associated with considerable morbidity and mortality.

In a retrospective study of 215 patients with health care-associated meningitis or ventriculitis seen at two large tertiary care centers in the United States, the majority had a poor outcome, including death (9 percent), persistent vegetative state (14 percent), and severe disability resulting in partial or total dependence on assistance for daily living (36 percent) [5]. In another retrospective study from Portugal, nearly half of patients with health care-associated meningitis and ventriculitis had a neurologic complication and 37 percent died [36]. A study of patients with post-neurosurgical meningitis in Italy reported a slightly lower mortality rate of 8 percent [37]

Although post-neurosurgical patients or patients with indwelling central nervous system (CNS) devices often have underlying comorbidities that increase risk for adverse outcomes in general, health care-associated meningitis or ventriculitis in such patients is independently associated with poor prognosis [38,39].

SOCIETY GUIDELINE LINKS — Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Bacterial meningitis in adults".)

SUMMARY AND RECOMMENDATIONS

●Empiric antibiotic therapy – For patients with suspected health care-associated meningitis or ventriculitis, we suggest empiric therapy with vancomycin plus an antipseudomonal beta-lactam (ceftazidime, cefepime, or meropenem) (Grade 2C). These regimens cover the expected microbiologic spectrum, take into account the prevalence of drug resistance among health care-associated pathogens, and are expected to penetrate the cerebrospinal fluid (CSF) well. (See 'Empiric therapy' above.)

●Tailoring therapy:

•Organism identified on CSF testing – The antimicrobial regimen should be adjusted to target the identified organism. Recommended regimens are discussed separately (table 1). (See "Treatment of bacterial meningitis caused by specific pathogens in adults".)

•No organism identified on CSF testing – For patients with ongoing clinical suspicion for health care-associated meningitis or ventriculitis despite negative microbiologic CSF testing, the empirically chosen regimen can be continued to complete the course. If clinical suspicion for bacterial infection is not high, CSF cultures are negative after two to three days, and the patient had not received antibiotics prior to obtaining CSF cultures, it is reasonable to discontinue empiric antibiotics and monitor. (See 'Negative CSF culture' above.)

●Follow-up CSF testing – Routine repeat CSF testing is not necessary in patients who are clinically improving. However, we perform repeat CSF analysis and culture in the following situations (see 'Follow-up CSF testing in selected cases' above):

•Lack of clinical improvement after several days (eg, five to seven) of appropriate therapy

•Prior to implantation of a central nervous system (CNS) device to document CSF clearance

●Duration of antimicrobial therapy – The optimal duration of antimicrobial therapy for health care-associated meningitis and ventriculitis is uncertain. We suggest at least 10 to 14 days of systemic antimicrobial therapy rather than shorter durations (Grade 2C). For individuals who have a repeat positive CSF culture despite appropriate antimicrobial therapy, we continue therapy until at least 10 to 14 days following the last positive culture. Although systematic data are lacking, there is concern for relapse with shorter courses. (See 'Duration' above.)

●Intraventricular or intrathecal antimicrobial therapy – We suggest against routine use of intraventricular or intrathecal antibiotics for health care-associated meningitis and ventriculitis (Grade 2C). However, cautious use as an adjunct to systemic therapy is reasonable for refractory infection despite appropriate systemic therapy, infection with extensively drug-resistant pathogens, or infection associated with an implanted CNS device or catheter that cannot be removed. Local instillation of antimicrobial therapy is unnecessary for most patients but has been associated with improved microbiologic or clinical response in some observational series. For individuals without existing ventricular or intrathecal access, the risks of obtaining access may outweigh any potential benefits.

Doses and dosing intervals of intraventricular or intrathecal therapy target CSF concentrations 10 to 20 times the minimum inhibitory concentration (MIC) of the causative organism (table 2). (See 'Intrathecal and intraventricular therapy' above.)

●Device removal – For patients with meningitis or ventriculitis associated with CNS hardware, we recommend device removal to reduce the risk of treatment failure or relapse (Grade 1B). (See 'Device removal' above.)

The approach to CSF shunt removal and timing of replacement are discussed in detail elsewhere. (See "Infections of cerebrospinal fluid shunts", section on 'Timing of new shunt placement'.)

●Prognosis – Even when appropriate antibiotic therapy is administered, health care-associated meningitis and ventriculitis are associated with considerable morbidity and mortality. (See 'Prognosis' above.)