INTRODUCTION — Health care-associated meningitis and ventriculitis can occur as a complication of neurosurgery, placement of cerebrospinal fluid (CSF) shunt or external ventricular drain (EVD), 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 epidemiology, microbiology, clinical manifestations, and diagnosis of health care-associated meningitis and ventriculitis are reviewed in this topic. Treatment and prognosis of health care-associated meningitis and ventriculitis are discussed elsewhere. (See "Health care-associated meningitis and ventriculitis in adults: Treatment and prognosis".)
Information specific to CSF shunt infections is also presented elsewhere. (See "Infections of cerebrospinal fluid shunts".)
INCIDENCE AND RISK FACTORS — Overall, health care-associated meningitis and ventriculitis are uncommon infections and occur primarily as a complication of a neurosurgical procedure or central nervous system (CNS) device. As an example, in a study from Turkey, health care-associated meningitis occurred in 0.3 percent of all admissions and accounted for only 0.5 percent of all health care-associated infections over an eight-year period [1].
Post-neurosurgical infection — Meningitis and ventriculitis complicate fewer than 5 percent of craniotomies. In one large series of 1587 patients who underwent cranial operations, only 14 (0.8 percent) were complicated by CNS infection [2]. Factors associated with an increased risk of postoperative infection include surgery for tumor resection (especially resection of recurrent gliomas), surgical procedures traversing areas of high bacterial colonization (such as the paranasal sinuses), cerebrospinal fluid (CSF) leak, and concomitant infection outside the CNS [3,4].
Craniotomy can also be complicated by bone-flap infection, epidural or brain abscess, or subdural empyema. (See "Pathogenesis, clinical manifestations, and diagnosis of brain abscess" and "Intracranial epidural abscess".)
Endoscopic endonasal surgery accesses the skull base through the nose or sinuses, thereby obviating a craniotomy or facial incision; it is associated with a low rate of postoperative meningitis (1.8 percent in one series at a center that used preoperative prophylaxis) [5].
Ventricular catheter infection
●CSF shunts – The epidemiology of CSF shunt infection is discussed in detail elsewhere. (See "Infections of cerebrospinal fluid shunts", section on 'Incidence and risk factors'.)
●External ventricular drain (EVD) – The reported rates of infection associated with EVDs, also known as ventriculostomy catheters, are variable. In a systematic review of 23 studies that included 5733 EVD placements, the rate of a positive CSF culture ranged from 2.3 to 23 percent, with an average pooled rate of 8.8 percent [6]. Among a subset of studies using a more stringent definition of infection (eg, positive CSF cultures in the absence of both clinical changes and CSF pleocytosis were not counted as infection), the rate was 6.6 percent. However, in another systematic review of 42 studies that included 3035 patients with an EVD, the average rate of drain-associated ventriculitis (which also included culture negative cases) was 23 percent [7]. These differences in reported rates may be related to definition of infection, underlying patient populations, and differences in catheter management across studies.
Features associated with an increased risk of EVD include intraventricular or subarachnoid hemorrhage, history of cranial fracture with CSF leak, history of craniotomy, irrigation of the EVD, concomitant systemic infections, and duration of catheterization [6]. Details on the time to onset of infection following EVD placement are found elsewhere. (See 'Time to onset' below.)
Lumbar puncture and catheter-related infection — The incidence of health care-associated meningitis following lumbar puncture (LP) is extremely low. In one study of almost 2 million instances of spinal or epidural anesthesia, the rate of meningitis was approximately 1 in 50,000 [8]. Women in labor and immunocompromised patients who have an increased risk of bacteremia may be at increased risk of health care-associated meningitis following dural puncture [9].
Few of the reported cases of meningitis following dural puncture have been fatal, suggesting that meningitis in this setting may be less severe than community-acquired meningitis, may be caused by less virulent organisms, or may be recognized and treated earlier [10].
The risk of infection with indwelling lumbar catheters is higher than with LP alone but is also relatively low, especially if catheter care minimizes instrumentation of the closed drainage system and limits duration. In one study of 233 patients with external lumbar catheter for normal pressure hydrocephalus, only two (0.8 percent) developed meningitis [11].
Infection of other CNS devices
●Intracranial pressure monitor – An intracranial pressure (ICP) monitor may be placed within the subdural space, parenchyma, or ventricle to provide a quantitative measure of ICP and cerebral perfusion pressure following open- or closed-head injury or neurosurgery. Central nervous system (CNS) infection occurs as a complication of ICP monitor placement in 2.9 to 10.3 percent of patients [12,13].
Risk factors for infection include placement following open-head injury or intraparenchymal hemorrhage, placement in the intraventricular space, CSF leak, concomitant infection outside the CNS, and placement of the monitor for more than five days [12,13]. Unlike the increased rate of CNS infection seen with placement of a CSF shunt by less-experienced surgeons, CNS infection rates following ICP monitor placement are similar when a neurosurgeon, general surgeon, or midlevel practitioner has performed the procedure [14].
●Deep-brain stimulators – Deep-brain stimulators are used for treatment of a variety of movement disorders. Infections complicate 0 to 15 percent of deep-brain stimulator placement surgery [15-18]. Risk factors for infection are uncertain. Age of the patient, reason for placement, experience of the surgeon, staged placement of leads, and duration of surgery have not been consistently associated with increased risk of developing infection.
●Ommaya reservoirs – Ommaya reservoirs permit direct instillation of medications into the subarachnoid space, resulting in higher and more consistent concentrations for the treatment of leptomeningeal carcinomatosis, pain, and chronic or recurrent CNS infection [19]. Rates of bacterial infection following reservoir placement vary widely, from 7 to 50 percent, and are higher in patients who undergo access of the reservoir more than 20 times and in patients with lymphoproliferative disorders [20].
PATHOGENESIS — The most common mechanisms for health care-associated meningitis and ventriculitis include [21]:
●Contamination at the time of surgery from surrounding skin or sinuses
●Extension of a skin infection at the surgical site into the central nervous system (CNS)
●Transmission of infection with an implanted device
●Hematogenous spread of infection from another site
Although no study has directly demonstrated that a lumbar puncture (LP) needle can introduce skin flora into the cerebrospinal fluid (CSF), several pieces of evidence support this mechanism. Clusters of meningitis cases occurring after LP or placement of an epidural catheter performed by the same operator suggest that organisms may be introduced through disruption of sterile technique [22]. Bacterial cultures of the skin surrounding puncture sites have detected streptococci and other bacteria, supporting the hypothesis that patients who develop meningitis following dural puncture acquire the organism from skin flora introduced with the spinal needle [23]. In addition, culture of spinal needles immediately after dural puncture revealed bacterial contamination in 17.9 percent of the needles, mostly by coagulase-negative staphylococci (87.5 percent) [24].
Droplet transmission (eg, from the upper airways of the operator) as a potential but likely rare cause of health care-associated meningitis is supported by a study that used polymerase chain reaction (PCR) assays to match isolates of Streptococcus salivarius from a patient who developed meningitis following LP with isolates from a throat swab of the neurologist performing the LP [25].
MICROBIOLOGY — The typical causative organisms of health care-associated meningitis and ventriculitis are distinct from those associated with community-acquired meningitis [26]. In patients with health care-associated meningitis and ventriculitis, gram-positive and gram-negative bacteria are both common pathogens [27]. Specifically, the most frequently associated organisms include normal skin flora, such as coagulase-negative Staphylococci spp (CoNS) and Cutibacterium (formerly Propionibacterium) acnes, as well as Staphylococcus aureus, streptococci, and gram-negative bacilli [28,29]. Gram-negative bacilli that cause health care-associated meningitis and ventriculitis include Enterobacterales (eg, Escherichia coli, Klebsiella pneumoniae, Enterobacter spp, Serratia marcescens), Pseudomonas aeruginosa, and Acinetobacter spp [27].
In one study that included 178 cases of health care-associated meningitis and ventriculitis with a positive CSF culture, the proportions of cases due to gram-positive and gram-negative organisms were similar (53 and 47 percent, respectively) [27]. However, the frequency of specific causative organisms varies by study and may depend on the underlying predisposing features:
●External ventricular drains – In a systematic review of studies that included 545 external ventricular drain (EVD) infections in which a causative pathogen was identified, gram-positive organisms predominated, with CoNS in 32 percent and S. aureus in 13 percent [7]. Gram-negative organisms were identified in 32 percent of cases; the most common were Acinetobacter, Enterobacter, Klebsiella, and Pseudomonas spp (9, 5, 5, and 4 percent, respectively).
●Cerebrospinal fluid shunts – The microbiology associated with cerebrospinal fluid (CSF) shunt infections is discussed elsewhere. (See "Infections of cerebrospinal fluid shunts", section on 'Microbiology and pathogenesis'.)
●Dural puncture – In contrast, in a review of 179 cases of meningitis that occurred following dural puncture, the most common cause was alpha-hemolytic streptococci, identified in 49 percent of cases [30]. Meningitis caused by alpha-hemolytic streptococci most often follows a procedure requiring prolonged dural penetration, such as spinal anesthesia or myelography, but is uncommon after a short procedure, such as lumbar puncture (LP) [31].
●Sinus surgery or CSF leaks – Bacteria that colonize the nasopharynx, such as Streptococcus pneumoniae, Haemophilus influenzae, and Streptococcus pyogenes, can cause meningitis following sinus surgery or in the setting of CSF leaks following skull base fracture but are otherwise uncommon in health care-associated infections [32,33].
However, a microbiologic spectrum more typical for health care-associated infection has been reported following endoscopic endonasal surgery, which accesses the brain through a nasal approach. In one series of 37 such cases, one-third were polymicrobial, and S. aureus and P. aeruginosa were the most commonly isolated pathogens [5].
Among gram-negative health care-associated meningitis and ventriculitis, antimicrobial resistance can be a major challenge in some locations. In one series of 40 cases from Ireland, 25 percent of the isolates were resistant to ceftriaxone, and this resistance had increased over the course of the study [34]. In some regions, carbapenem resistance is also highly prevalent. In a review of 115 post-neurosurgical gram-negative meningitis cases at two tertiary care centers in Israel, which included 55 Enterobacterales, 33 Acinetobacter, and 16 Pseudomonas infections, the rate of imipenem resistance was 35 percent [35].
Fungal meningitis, particularly due to Candida spp, is a rare cause of device-related infections; there is often an antecedent history of bacterial infection prior to onset of the fungal meningitis [36] (see "Candida infections of the central nervous system"). Additionally, well-documented sporadic outbreaks of fungal meningitis associated with spinal instrumentation have been reported. As an example, a multistate outbreak of Exserohilum spp meningitis that was associated with contaminated methylprednisolone spinal injections and involved over 700 cases occurred in the United States from 2012 to 2013 [37]. In May 2023, an outbreak of Fusarium solani meningitis (based on elevated CSF beta-glucan levels) was reported among individuals in the United States who had undergone procedures using epidural anesthesia at certain clinics in Matamoros, Mexico [38]. (See "Aseptic meningitis in adults", section on 'Fusarium outbreaks'.)
CLINICAL MANIFESTATIONS
Presenting features — The classic features of health care-associated meningitis and ventriculitis are similar to those of community-acquired meningitis and include fever, headache, neck pain or stiffness, and mental status change. However, only a minority of cases present with all features. Other presenting features include wound drainage (purulent or nonpurulent) at the site of the procedure or device, focal neurologic deficits, seizures, and for patients with meningitis related to lumbar instrumentation, low back pain.
As an example, in a series of 215 patients with health care-associated meningitis and ventriculitis, the following features were reported [39]:
●Fever (temperature >38.0°C) – 41 percent
●Headache – 49 percent
●Changes in mental status – 41 percent
●Nausea/vomiting – 40 percent
●Focal neurologic deficit – 33 percent
●Neck stiffness – 19 percent
●Seizures – 10 percent
●Photophobia – 7 percent
These features are not specific to central nervous system (CNS) infection in postoperative patients or patients with CNS devices. As an example, in a systematic review of 162 external ventricular drain (EVD)-associated ventriculitis cases, fever, headache, and neck stiffness were reported in 72, 62, and 27 percent, respectively; however, fever and meningeal signs were also reported in 29 and 20 percent of 275 individuals who had a drain in place but no ventriculitis [7]. Decline in consciousness occurred in similar proportion of patients with (38 percent) and without (41 percent) ventriculitis.
Infections of cerebrospinal fluid (CSF) shunts can present subtly or with symptoms at the distal end of the shunt; the clinical manifestations of CSF shunt infections are detailed elsewhere. (See "Infections of cerebrospinal fluid shunts", section on 'Clinical manifestations'.)
The CSF findings of health care-associated meningitis and ventriculitis are also discussed elsewhere. (See 'Tests to perform and their interpretation' below.)
Health care-associated meningitis and ventriculitis can be complicated by or coexist with other pyogenic CNS infections, such as brain abscesses and epidural or subdural abscesses. These are discussed in detail elsewhere. (See "Pathogenesis, clinical manifestations, and diagnosis of brain abscess" and "Intracranial epidural abscess" and "Spinal epidural abscess".)
Time to onset — The time to onset of meningitis or ventriculitis following a neurosurgical procedure depends on the procedure and infecting organism. Infections with indolent organisms (eg, coagulase-negative staphylococci or C. acnes) may present over a longer time period than those caused by more virulent organisms.
●Craniotomy – Following craniotomy, most CNS infections occur within the first two weeks. In one prospective study of 6243 patients who underwent craniotomy, 95 developed meningitis at a mean of 14 days postoperatively; one-third of cases occurred in the first week, one-third in the second week, and one-third between the third and ninth weeks [40]. Infection more than three months after surgery is uncommon [41].
●Ventricular drains/catheters – For EVDs, the risk of infection increases the longer the drain is in place. In one review of 17 studies evaluating the duration for catheterization and risk for infection, most studies found that infections were most likely to occur after 5 to 10 days of catheterization [6]. Meningitis can also occur in the days following removal of an EVD [42].
Timing of infection following internal CSF shunt placement is discussed elsewhere. (See "Infections of cerebrospinal fluid shunts", section on 'Incidence and risk factors'.)
●Dural puncture – Symptoms of health care-associated meningitis complicating dural puncture typically develop within the first day after the procedure [43]. Symptoms of epidural abscess after placement of an epidural catheter typically occur within five days of placement [44].
EVALUATION — The diagnosis of meningitis, regardless of whether it is community- or health care-associated, predominantly relies upon examination of the cerebrospinal fluid (CSF). For health care-associated meningitis and ventriculitis, CSF culture is the most important diagnostic test, although the diagnosis depends on a constellation of findings.
Clinical suspicion — Among patients who have undergone a recent neurosurgical procedure or have an indwelling central nervous system (CNS) device, the possibility of meningitis or ventriculitis should be suspected in the setting of new fever or leukocytosis with new headache, meningeal signs, focal neurologic defects, seizures, or worsening mental status. It should also be suspected in patients who have erythema, tenderness, or drainage at the site of a craniotomy incision or CNS device placement (eg, external ventricular drain [EVD], intracranial pressure monitor). (See 'Presenting features' above.)
A low threshold for suspicion is warranted when considering the possibility of health care-associated meningitis and ventriculitis in a neurosurgical patient. Recognition can be challenging for various reasons. Many patients do not present with the full constellation of typical findings. Obtaining an accurate history and review of symptoms may be limited in patients with impaired consciousness. Changes in underlying neurologic deficits may be difficult to assess. Symptoms of systemic infection (eg, fever, leukocytosis, confusion) might be attributed to common postoperative infections, which can co-occur with meningitis.
However, health care-associated meningitis and ventriculitis rarely occur in the absence of head trauma, neurosurgical or invasive neurological procedure (including lumbar puncture [LP]), or immunosuppression. Thus, performing LP to exclude health care-associated meningitis and ventriculitis in an immunocompetent hospitalized patient without a history of trauma or intervention has a low yield and rarely changes management [45].
Timing and components of initial evaluation — When meningitis or ventriculitis is suspected following neurosurgery or in patients with indwelling CNS devices, we perform the following workup (each step is discussed in detail elsewhere):
●Submit CSF fluid for testing, including microbiology (see 'Obtaining a specimen' below and 'Tests to perform and their interpretation' below)
●Check blood cultures (see 'Blood cultures' below)
●Assess for other more common causes of health care-associated infection, such as urinary tract infection or pneumonia (see "Fever in the surgical patient", section on 'Nosocomial infection')
●Image of the surgical site, with and without a contrast agent (see 'Neuroimaging' below)
When suspicion for CNS infection is high based on the clinical assessment and initial results of this evaluation, (eg, in a patient with CSF pleocytosis or fever that cannot be explained by another etiology), we recommend initiation of empiric antibiotic therapy for meningitis/ventriculitis.
We obtain all CSF and blood cultures prior to initiating antibiotics because cultures can become negative within hours of antibiotic administration. In a retrospective review of 326 cases of health care-associated meningitis and ventriculitis, CSF specimens obtained after antibiotics were given were less likely to be positive on Gram stain (13 versus 26 percent) and culture (49 versus 66 percent) compared with specimens obtained prior to antibiotics [27]. Among those who received antibiotics prior to CSF sampling, a shorter interval between antibiotic administration and specimen collection was associated with a higher likelihood of a positive CSF culture.
CSF analysis
Obtaining a specimen — Cerebrospinal fluid (CSF) can be obtained via LP, cisternal puncture (preferably under fluoroscopic guidance), or from a shunt or Ommaya reservoir. In patients with an EVD, initial CSF samples are typically obtained from that drain. Nevertheless, even if an EVD is present, CSF obtained via LP can provide additional useful information, especially in patients with obstruction of CSF flow who may have infection initially confined to the extraventricular compartment.
Accessing an Ommaya reservoir to obtain CSF is usually a last resort, as this requires puncturing the skin, which carries a risk of introducing infection.
Tests to perform and their interpretation — Initial CSF testing for patients with suspected health care-associated meningitis and ventriculitis includes the following:
●Gram stain and culture, including anaerobic culture – Cultures should be held for at least 10 days, as some organisms (eg, C. acnes) grow slowly. A positive CSF Gram stain and culture are considered the gold standard for diagnosis of health care-associated meningitis and ventriculitis and are the most important assays to establish the diagnosis and etiology of meningitis [46]. (See 'Diagnostic criteria' below.)
Identification of an organism on CSF Gram stain is sufficient for diagnosing infectious meningitis in a patient with clinical suspicion for infection. Culture is also critical to identify a specific pathogen and determine its antimicrobial susceptibilities. However, negative results do not exclude infection, especially in patients who have received antibiotics. Up to 70 percent of patients with positive CSF culture have a negative Gram stain [47]. As examples, in a study of 91 patients with health care-associated meningitis and ventriculitis, the CSF Gram stain was positive in 65 (71 percent) of patients with culture-proven infection [48]. However, in another study of 216 patients with health care-associated infection, culture was positive in 48 percent but Gram stain in only 20 percent [39].
Furthermore, a positive culture may rarely be a false positive. Specifically, in a patient with an indwelling CNS device without clinical concern for infection, a single positive culture of CSF taken from the device (eg, an external ventricular drain [EVD]) may represent contamination of the culture or drainage tubing rather than a true infection [26]. (See 'When clinical and microbiologic features are discordant' below.)
●Cell count, protein, glucose – Typical CSF findings in health care-associated meningitis and ventriculitis include an elevated white blood cell (WBC) count with a polymorphonuclear leukocyte (PMN) predominance, elevated protein, low glucose. However, these features do not always reliably distinguish bacterial meningitis from nonbacterial processes [26,49,50].
•Cell count – Although all forms of meningitis are typically associated with elevated WBC count, and elevated WBC is correlated with a positive CSF culture, there is no specific cutoff at which an elevated WBC can confirm health care-associated meningitis and ventriculitis nor at which a low WBC count can exclude it [51]. In a study of 70 patients with meningitis following a neurosurgical procedure, only those who had bacterial meningitis (rather than aseptic meningitis) had a CSF WBC count of >7500/microL and glucose level of <10 mg/dL [52], but this threshold has not been well validated. In studies of patients with gram-negative bacillary meningitis, CSF WBC counts ranged from 0 to 80,600 cells/microL [53-55]; PMNs accounted for more than 50 percent of WBCs in 90 percent of cases.
Although data are limited on the significance of trends in WBC count, increasing CSF pleocytosis should raise concern for CNS infection.
Clinicians should also be aware that the CSF WBC count often differs between compartments; a CSF WBC count obtained from a shunt reservoir or ventricular drain may not reflect the WBC in CSF obtained through an LP, especially if obstructive hydrocephalus is present [28]. (See 'Obtaining a specimen' above.)
•Protein – CSF protein level is usually but not invariably high and is not a reliable predictor of health care-associated meningitis and ventriculitis. In a series of 215 patients with health care-associated meningitis and ventriculitis, the protein level ranged from 14 to 1774 mg/dL (mean 131); 64 percent had a level ≥100 mg/dL [39]. In another review, mean protein levels ranged from 171 to 1123 mg/dL [55]. In a study of 130 patients with an EVD, daily protein and glucose levels were not predictive of developing meningitis CSF [51].
•Glucose – The CSF glucose concentration is usually low in health care-associated meningitis and ventriculitis. In a series of 215 patients with health care-associated meningitis and ventriculitis, the glucose level ranged from 1 to 121 mg/dL (mean 49); 36 percent had a level <40 mg/dL [39]. The glucose level in another case series ranged from 32 to 72 mg/dL [49]. CSF-to-serum glucose ratios of less than 50 percent are infrequent after uncomplicated neurosurgical procedures [46].
A very low CSF glucose may be indicative of infection. In a study of 70 patients with meningitis following a neurosurgical procedure, only those who had bacterial meningitis (rather than aseptic meningitis) had a glucose level of <10 mg/dL [52], but this threshold has not been well validated.
●Extra specimen to hold for additional molecular testing if initial microbiologic testing is uninformative – We often request that our laboratory freeze an aliquot of CSF to save for future testing, should the initial cultures and assays be unrevealing. In particular, broad-range polymerase chain reaction (PCR) assay is an important diagnostic method when antimicrobial therapy has already started or when cultures remain negative. Metagenomic next-generation sequencing (mNGS) of the CSF may also improve pathogen detection, but it is not widely available, and sensitivity and specificity are not fully determined [56].
•Broad-range PCR – This can be performed on CSF for the diagnosis of meningitis and identification of a bacterial, mycobacterial, or fungal pathogen. Specifically, the 16S ribosomal ribonucleic acid (16S rRNA) PCR test can identify a broad range of bacterial pathogens with a single assay. One study reported a sensitivity of 86 percent and a specificity of 97 percent compared with culture [57]. In another study of 86 CSF samples from 28 patients with health care-associated meningitis and ventriculitis, PCR detected bacteria in 49 percent of culture-negative patients, and 81 percent of these samples were from patients who had received antibiotics [58].
•Metagenomic sequencing – mNGS of CSF is an emerging technology that is only available through a limited number of institutions [59]. It has the potential to identify a broad range of pathogens through sequencing of all nucleic acid in a single assay [56]. Although this assay is considered highly specific, the assay parameters have not been fully defined, and the negative predictive value is uncertain. Therefore, it should only be used in consultation with experts in infectious diseases. In a prospective study in hospitalized children and adults, NGS identified 58 CNS infections in 57 patients; 13 (22 percent) of these infections had not been identified through traditional laboratory testing; this study included four patients with health care-associated meningitis and ventriculitis who had negative CSF cultures [60].
●Additional initial testing for specific circumstances
•For patients who undergo CNS device removal in the context of suspected infection, we also submit the explanted device for Gram stain and culture.
•If fungal infection is suspected (eg, immunocompromised patients, or in an outbreak setting), CSF should also be submitted for fungal wet prep and culture, beta-D-glucan, and galactomannan. Molecular testing can also identify fungal pathogens if other methods are uninformative. (See "Candida infections of the central nervous system", section on 'CSF analysis' and "Diagnosis of invasive aspergillosis", section on 'Diagnostic modalities' and "Aseptic meningitis in adults", section on 'Fusarium outbreaks'.)
●Tests we do not routinely perform – We do not routinely submit CSF for lactate and procalcitonin. If such testing is available, it can provide additional supportive evidence of an infection, but there are conflicting data, and positive or negative results must be interpreted with caution.
•Lactate – Measurement of CSF lactate concentration is a rapid and inexpensive ancillary test but is not widely available. Elevations in lactate levels result from a combination of bacterial growth, anaerobic metabolism, and release of lactate by neurons and glial cells affected by meningitis-induced brain edema [61,62]. Because lactate levels are not affected by the presence of red blood cells (RBCs) in the CSF, lactate testing can sometimes be helpful when the CSF is bloody or when routine tests give inconclusive results. In one study of CSF lactate levels in the diagnosis of bacterial meningitis following neurosurgery, lactate levels had a higher sensitivity and specificity compared with the CSF-to-blood glucose ratios (88 versus 77 percent and 98 versus 87 percent, respectively) [49]. Cases of culture-proven bacterial meningitis in this study had mean CSF lactate levels of 7.8 mmol/L, while nonbacterial cases had mean lactate levels of 2.3 mmol/L. In patients with bacterial meningitis following craniotomy, elevation of CSF lactate level ≥4 mmol/L was a sensitive and specific predictor for bacterial meningitis [49].
Despite these data, testing for CSF lactate levels is not routinely ordered or performed in clinical practice because of limited availability and because many clinicians perceive that this test does not offer substantially more information than standard CSF analysis. There have also been inconsistencies in the reported diagnostic power of the test [49].
•Procalcitonin – Measurement of the procalcitonin concentration in CSF can provide supportive evidence of bacterial CNS infection. A small study that used CSF levels of procalcitonin (cutoff value of 0.075 ng/mL) and CSF lactate (cutoff level of 3.45 mmol/L) reported a high diagnostic accuracy for distinguishing between post-neurosurgical bacterial infection and aseptic meningitis (sensitivity of 96 percent and negative predictive value of 97.6 percent) [63].
Various models and parameters to improve diagnostic certainty of health care-associated meningitis and ventriculitis (eg, CSF lactate, CSF procalcitonin, cell index) have been studied, but they have not been adequately validated in the clinical setting [26,64,65]. Some clinicians calculate the cell index, a ratio of the WBC to RBC in the CSF divided by the ratio of WBC to RBC in the peripheral blood [64]. In a case-control study in 111 patients with intracranial hemorrhage, the cell index had good discrimination capacity in identifying patients with culture-confirmed cases [64]. The mean cell index for those with culture-positive meningitis was 4.299 versus 0.007 in the controls without infection.
Blood cultures — Although infrequently positive, blood cultures are an important ancillary test and can help identify a pathogen in patients with meningitis. In a series of 215 patients with health care-associated meningitis and ventriculitis, only 3 percent had bacteremia [39]. Despite this low yield, we typically send blood cultures in all patients with fever and possible health care-associated meningitis and ventriculitis.
The rate of bacteremia may be higher in subsets of patients. As an example, positive blood cultures have been reported in up to 50 percent of adult patients with gram-negative bacillary meningitis, in whom bacteremia has been associated with worse prognosis [46,66]. In patients with ventriculoatrial shunts, bacteremia almost always accompanies shunt infection [26]. (See "Infections of cerebrospinal fluid shunts", section on 'Blood cultures'.)
Other blood tests — Measurement of the procalcitonin concentration in serum can provide supportive evidence of bacterial CNS infection. Although a serum procalcitonin cutoff level of 0.5 ng/mL has been associated with high specificity, the sensitivity was only 68 percent [67]. Although we do not use procalcitonin levels, usage varies by region and institution.
Serum beta-D-glucan and galactomannan are sent when fungal infection is suspected. (See "Candida infections of the central nervous system", section on 'Diagnosis' and "Diagnosis of invasive aspergillosis", section on 'Diagnostic modalities' and "Aseptic meningitis in adults", section on 'Fusarium outbreaks'.)
Neuroimaging — We routinely perform neuroimaging as part of the evaluation for health care-associated meningitis and ventriculitis. In patients who have undergone neurosurgery, imaging that includes the cranial surgical site should be done to assess for potential infectious fluid collections at or adjacent to the surgical site. For patients with ventricular drains, layering of purulent material within the ventricle supports the diagnosis of infection. Neuroimaging may also be useful for determining if there is an adjacent soft tissue or sinus infection and for identifying potential complications of infection, such as hydrocephalus, vasculitis, or cerebral thrombosis [68,69]. For patients with internalized shunts, imaging of the distal site can detect abnormalities that support infection; this is discussed in detail elsewhere. (See "Infections of cerebrospinal fluid shunts", section on 'Imaging'.)
Between magnetic resonance imaging (MRI) and computed tomography (CT), MRI provides higher resolution and is generally our preferred modality; however, CT is an appropriate alternative if it can be obtained more expediently, if metal at the site of interest would distort the MR image or the patient has a contraindication to MRI. Whichever imaging modality is selected, images should be obtained with and without contrast agent.
In one study of 17 patients with ventriculitis who underwent either CT or MRI, ventricular debris was present in 94 percent, ependymal enhancement was present in 64 percent, and meningeal enhancement was present in 76 percent [70]. In patients with ventriculitis, CT scans may reveal ependymal enhancement, but MRI is more sensitive than CT for detecting ventriculitis; with diffusion-weighted imaging demonstrating a bright signal when purulent debris is present.
Neuroimaging cannot reliably differentiate between aseptic or infectious meningitis as inflammation and contrast enhancement occurs in up to 80 percent of post-craniotomy patients who do not have CNS infection [71].
DIAGNOSIS
Diagnostic criteria — Making a diagnosis of health care-associated meningitis and ventriculitis requires synthesis of clinical features, cerebrospinal fluid (CSF) microbiologic testing, and results of other CSF tests. The diagnosis can be made relatively straightforwardly in a patient with a recent central nervous system (CNS) intervention or indwelling CNS device who has new fever, new neurologic deficits or symptoms, and identification of an organism on CSF culture or molecular testing, particularly if there is also CSF pleocytosis or hypoglycorrhachia. (See 'Tests to perform and their interpretation' above.)
However, the diagnosis can be more challenging when clinical and microbiologic features are discordant, as discussed below. (See 'When clinical and microbiologic features are discordant' below.)
The National Healthcare Safety Network of the Centers for Disease Control and Prevention has published criteria for the diagnosis of health care-associated meningitis and ventriculitis for health care facilities to use to track specific health care-associated infections (table 1) [72]. Our clinical approach to making a diagnosis is generally consistent with these criteria, although the specific criteria used have not been validated for clinical use.
When clinical and microbiologic features are discordant — Many patients do not present with all the clinical and microbiologic features of health care-associated meningitis and ventriculitis (table 1), and in such cases the diagnosis can be more challenging. Although culture of CSF or CNS devices is the most important diagnostic test in establishing the presence of infection, no single clinical or laboratory finding can confirm or exclude the diagnosis. Critical clinical judgment is often necessary when the clinical and microbiologic features are discordant:
●Negative CSF culture in setting of clinical suspicion – If initial CSF cultures are negative yet CNS infection is still suspected clinically, we repeat CSF microbiologic testing with routine, anaerobic, and fungal culture; PCR testing; fungal markers; and, if available, metagenomic next-generation sequencing. Additional assays, such as lactate and procalcitonin, may be used depending on availability at the institution and regional practice. (See 'Tests to perform and their interpretation' above.)
If testing remains unrevealing, the diagnosis of health care-associated meningitis and ventriculitis remains presumptive. Without a pathogen identified, symptoms and CSF abnormalities could represent aseptic meningitis due to CNS instrumentation or side effect of a medication. However, CSF Gram stain and culture can be negative despite bacterial (or fungal) infection, especially if antibiotics were administered prior to CSF sampling, and molecular tests do not have perfect sensitivity. Features that support the diagnosis of infection over aseptic meningitis in the setting of negative microbiologic testing include:
•Fever without other obvious cause
•Increasing CSF leukocyte level on repeat sampling
•Purulent drainage or other signs of infection at the surgical or drain site
•New neurologic defects, such as seizures, focal defects, or loss of consciousness
•Neuroimaging suggestive of purulent fluid or abscess formation
If the initial clinical suspicion for CNS infection was not especially high, none of the features noted above are present, and antibiotics had not been administered prior to CSF sampling, it is reasonable to attribute CSF pleocytosis to aseptic rather than bacterial meningitis, discontinue empiric antibiotics, and cautiously monitor. (See "Health care-associated meningitis and ventriculitis in adults: Treatment and prognosis", section on 'Negative CSF culture'.)
Additional details on postoperative aseptic meningitis are found elsewhere. (See 'Differential diagnosis' below.)
●Positive CSF microbiology in setting of low clinical suspicion – We typically avoid sending CSF specimens for culture or other microbiologic testing in the absence of clinical suspicion for infection (ie, we avoid surveillance cultures in patients with ventricular drains). A positive CSF culture in the absence of clinical features (such as fever, leukocytosis, change in neurologic status) may reflect contamination or colonization. However, indolent organisms such as coagulase-negative staphylococci and C. acnes may cause infection without substantial symptoms or signs, so the lack of these cannot rule out infection. In such patients, we repeat CSF testing (cell count and microbiology). Repeated positive culture with the same organism, particularly if the CSF white blood cell (WBC) count increases on repeat testing, is suggestive of infection. (See 'Evaluation' above.)
Overall, we generally have a low threshold to diagnose and treat health care-associated meningitis and ventriculitis because of the potential for severe morbidity with untreated infection. (See "Health care-associated meningitis and ventriculitis in adults: Treatment and prognosis", section on 'Prognosis'.)
DIFFERENTIAL DIAGNOSIS — The main differential diagnosis of health care-associated meningitis and ventriculitis is aseptic meningitis, which can also complicate neurosurgical procedures or dural puncture and can present with similar symptoms and signs, such as fever, neurologic symptoms or deficits, and abnormal cerebrospinal fluid (CSF) cell count, protein, and glucose. By definition, aseptic meningitis requires negative CSF Gram stain and culture, as well as recovery of the patient without administration of antibiotics [73]. Thus, aseptic meningitis cannot be confirmed until several days after initiating the evaluation.
Until culture results are available, it is difficult to distinguish between aseptic and infectious meningitis in the postoperative neurosurgical patient. Following a neurosurgical procedure, CSF protein, glucose, white blood cell (WBC) count, and Gram stain do not reliably distinguish bacterial meningitis from nonbacterial processes [26,49,50]. Quick resolution of an elevated CSF WBC count without administration of antibiotics is more consistent with aseptic meningitis. As an example, in a study of patients who underwent spinal anesthesia but did not develop meningitis, CSF pleocytosis was present in 65 percent at 24 hours, 30 percent at 48 hours, and 18 percent at 72 hours; the highest cell count in this study was 1950 cells/microL [74]. A more recent study noted that none of the CSF characteristics, other than a positive culture, could clearly distinguish between bacterial and aseptic meningitis [75].
Certain clinical features may suggest against aseptic meningitis. In a study of 70 patients who developed meningitis following craniotomy or paranasal or spinal surgery, those with aseptic meningitis did not have purulent wound drainage, wound erythema or tenderness, coma, new focal neurologic deficits, or seizures [52]. Fever (temperature ≥39.4°C) occurred but was uncommon.
The etiology of aseptic meningitis following neurosurgery or dural puncture is not completely understood. One hypothesis is that red blood cells or other materials released during surgery induce meningeal inflammation and pleocytosis. Experimental injection of saline, air, casein, or blood into the subarachnoid space produced symptoms of meningitis and a neutrophilic pleocytosis, the latter persisting up to three to four weeks [76].
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
●Types of infection – Health care-associated meningitis and ventriculitis are uncommon infections that occur primarily as a complication of central nervous system (CNS) devices, such as external ventricular drains (EVDs), cerebrospinal fluid (CSF) shunts, and lumbar drains. Post-craniotomy infections are less common. The risk of infection following lumbar puncture (LP) without an indwelling device is extremely low. (See 'Incidence and risk factors' above.)
Issues related to CSF shunt infections, specifically, are discussed in detail elsewhere. (See "Infections of cerebrospinal fluid shunts".)
●Microbiology – The most frequently associated organisms include normal skin flora, such as coagulase-negative Staphylococci spp and Cutibacterium (formerly Propionibacterium) acnes, as well as Staphylococcus aureus, streptococci, and gram-negative bacilli (eg, Enterobacterales, Pseudomonas aeruginosa, and Acinetobacter spp). Candida spp are a rare cause of device-related infections. (See 'Microbiology' above.)
●Time to onset – For EVDs, infections mostly occur after they have been in place for at least 5 to 10 days; most post-craniotomy infections occur within the first two weeks postoperatively. (See 'Time to onset' above.)
●Clinical suspicion – Following neurosurgery or in the setting of an indwelling CNS device, the possibility of meningitis or ventriculitis should be suspected in the setting of new fever or leukocytosis with new headache, meningeal signs, focal neurologic defects, seizures, or worsening mental status. It should also be suspected in patients who have erythema, tenderness, or drainage at the site of a craniotomy incision or CNS device. (See 'Presenting features' above.)
A low threshold for suspicion is warranted, as most patients do not present with the full constellation of findings and underlying neurologic deficits or illness may mask clinical changes. (See 'Clinical suspicion' above.)
●CSF evaluation – Initial CSF testing includes Gram stain and aerobic and anaerobic culture (held for 10 to 14 days), cell count, protein, and glucose. In immunocompromised patients or in other situations when fungal infection is suspected (eg, outbreak situation), fungal wet prep and culture, beta-D-glucan, and galactomannan should also be performed. We also request that the laboratory save an aliquot of CSF for additional testing, such as 16S ribosomal DNA polymerase chain reaction (PCR) testing, in case cultures are negative. We do not routinely test CSF for lactate and procalcitonin, although if available, can provide additional supportive evidence of an infection. (See 'CSF analysis' above.)
●Other testing – Blood culture and neuroimaging are also routine components of evaluation. Computed tomography (CT) and magnetic resonance imaging (MRI) can both identify ventriculitis or surgical site complications, but MRI is more sensitive and provides higher-resolution imaging. Images should be obtained with and without contrast agent. (See 'Blood cultures' above and 'Neuroimaging' above.)
●Diagnosis – The diagnosis is relatively straightforward in a patient with a recent CNS intervention or indwelling CNS device who has new fever, new neurologic deficits or symptoms, and growth of an organism on CSF culture (or pathogen identification on molecular testing), particularly if there is also CSF pleocytosis or hypoglycorrhachia (table 1). (See 'Diagnostic criteria' above.)
When clinical and microbiologic features are discordant, making the diagnosis often requires repeating CSF analysis and synthesis of other findings that could support CNS infection (lack of other explanation, increasing CSF white blood cell [WBC] count, signs of infection at surgical or drain site, new neurologic deficits, purulent fluid or abscess on imaging). Overall, we have a low threshold to presumptively diagnose and treat health care-associated meningitis and ventriculitis because of the potential for severe morbidity. (See 'When clinical and microbiologic features are discordant' above.)
آیا می خواهید مدیلیب را به صفحه اصلی خود اضافه کنید؟