Is intraventricular antibiotic therapy ready for prime time?

Recently, I observed a discussion on intraventricular antibiotic therapy. A patient in the intensive care unit had developed ventriculitis after dislodgement of a ventriculo-external drain with cerebral spinal fluid (CSF) positive for a quite sensitive K. pneumoniae. One of my colleagues was arguing strongly in favor of starting intraventricular antibiotic therapy to improve her chances of good neurological recovery. Personally, I wasn’t sure what the best option was – and spoiler alert: even after extensive research, I am still unsure what objectively is the best evidence-based approach.

The following questions come to mind:

1. What are the risks and benefits associated with intrathecal antibiotic therapy?

2. What are the main uncertainties surrounding intrathecal antibiotic therapy?

1. What are the risks and benefits associated with intrathecal antibiotic therapy?

The term intrathecal antibiotic therapy encompasses both intraventricular and intralumbar administration of antibiotics (1). This therapy has been applied in both community-acquired meningitis as well as post-neurosurgical infections (2). The latter is especially hard to treat as profound meningeal inflammation is often absent and therefore, systemic antibiotics don’t penetrate well. By directly administering antibiotics in the intraventricular space (or more rarely intralumbar space), the aim is to achieve adequate and predictable drug concentrations at the site of infection (4) while minimizing systemic toxicity. Additionally, bacteria might be eliminated more quickly (5), potentially reducing mortality.

Despite the fact that antibiotics have been administered intrathecally since the 1940s (1), the scientific evidence for this practice remains limited and is mostly derived from case series and retrospective cohort series.

• The strongest evidence available is for the combined intravenous and intrathecal administration of colistin. For instance, one meta-analysis including 296 post-neurosurigcal intracranial infections with multi-drug resistant gram-negative bacteria demonstrated a significantly increased survival and microbiological clearance with combination therapy compared to only systemic therapy (6). A retrospective, single-center, cohort study with 114 patients supported these findings (7). As a reflection of this evidence, the latest IDSA guideline recommends intrathecal administration of colistin/polymyxin B for ventriculitis and meningitis caused by multidrug-resistant gram-negative bacteria (8). Moreover, the FDA has approved intrathecal administration of polymixin B (1).

• In case of vancomycin, the evidence is even scarcer. One randomized controlled trial demonstrated that directly administering vancomycin in the ventricles achieved significantly higher CSF concentrations than systemic administration alone (9). A systematic review concluded that intraventricular administration of vancomycin is most likely safe and efficacious (4). Furthermore, some observational data show the sterilizing effect of intraventricular vancomycin. For instance, in a retrospective cohort study with 105 ICU patients, combined intraventricular and systemic vancomycin and/or aminoglycosides led to CSF sterilization in 88.4% of cases and a mortality of 18.1% (11). A similar story can be told about intrathecal gentamicin.

• Notably, the only randomized controlled trial I could find concerning intraventricular therapy dates back to 1980 and involved 52 infants with ventriculomeningitis. This study unfortunately showed higher mortality with intraventricular administration of gentamicin and was terminated early (11). Although the result of this study has been relativized (8), it is the main reason why a Cochrane review concluded that further trials on intraventricular antibiotic therapy in neonates are no longer justified (13).

Regarding the risk, intrathecal administration of antibiotics seems relatively safe (8). The main side effects, which have been reported:

a. Secondary infections

Intrathecal antibiotics require sterile preparation and administration through an external ventricular drain or lumbar puncture. This procedure may lead to secondary infections (14); however, their true prevalence remains unclear.

b. Chemical meningitis

Usually, chemical meningitis is mild and reversible (1). In a systematic review of 229 patients, this was present in 11% of cases (3). However, differentiating chemical meningitis from the primary CNS pathology or bacterial meningitis can be challenging (3). Moreover, toxicity might be underreported as it is difficult to diagnose in sedated and intubated patients (14).

c. Seizures

In the same review, seizures were present in 7% of cases, although the same limitations apply as with chemical meningitis. Furthermore, routine electroencephalography was not performed (3).

2. What are the main uncertainties surrounding intrathecal antibiotic therapy?

“Primum Non Nocere” (“first, do not harm”) remains one of the most important rules in medicine. Before intrathecal antibiotic therapy can become standard practice, several key uncertainties should be addressed.

1. Timing

Theoretically, direct intrathecal administration should result in faster CSF sterilization than systemic antibiotics alone, potentially improving outcomes. One case series with in total 24 patients also indicated that intrathecal therapy should be started within 48 hours (7). The IDSA guidelines of 2017, however, recommend only early start in case of multidrug-resistant gram-negative ventriculomeningitis as intravenous polymyxin B/colistin won’t penetrate well through the blood-brain barrier (15). In all other scenarios, intrathecal antibiotics are seen as a rescue therapy (8). Whether early combined intrathecal and systemic antibiotic therapy confers additional benefit remains as for now unanswered.

2. Dosing of intrathecal antibiotics

Both pharmacokinetics (PK) and pharmacodynamics (PD) are different in the CSF compartment than in the systemic circulation. First of all, distribution volume is mainly determined by the size of the ventricles (1). Additionally, a net flow of CSF from the ventricles to the cisterna magna causes a concentration gradient between the ventricles and the lumbar spine (1). Furthermore, clearance depends on both the daily CSF drainage volume and the properties of the antibiotic used. For instance, a consensus guideline for dosing, to which most often is referred, mentions a frequency of intrathecal therapy based on the daily volume of CSF drained (8). However, in case of vancomycin, one study reported accumulation during treatment, with an associated increase in half-life (4). Moreover, the concurrent administration of systemic antibiotics might also influence their intrathecal concentration (1). Finally, the optimal PK/PD target isn’t determined (1). The IDSA suggests maintaining a CSF trough concentration 10 to 20 times above MIC but this recommendation is based solely on expert opinion (8).

3. Therapeutic drug monitoring (TDM)

TDM could play a key role in addressing the current gaps in knowledge regarding the pharmacokinetics of intrathecal antibiotic therapy. However, this requires local expertise in measuring CSF concentration of an antibiotic (11). Besides, neither the optimal PK/PD target nor the toxic level for any antibiotic are well established (1). TDM is also not routinely performed in studies. For instance, in one multicenter American study, only 17 out of 105 (16.2%) patients received TDM. Moreover, measured antibiotic levels varied significantly between patients, likely due to differences in daily CSF drainage volume, sampling timing, and local practices (11). Furthermore, a systematic review of five studies found no clear association between the measured antibiotic levels and the clinical outcome or toxicity (10).

4. Duration of intrathecal antibiotic therapy

Considerable heterogeneity exists across studies, with reported median duration ranging from 5 days (11) to 14 days (7). In some cases, patients have been treated with intrathecal vancomycin for up to 31 days (4). A reasonable approach might be to continue intrathecal antibiotic therapy for 2 to 3 days after CSF sterilization (14).

My conclusion:

In my view, intrathecal antibiotic therapy should currently be regarded as an experimental therapy: while it appears safe (8), too many important questions remain unanswered. Awaiting randomized trials, centers performing this therapy should systematically report comprehensive data on indication, dosing, timing, duration and adverse effects. Such data could serve as guidance in the design of future randomized controlled trials.

References:

1. Muller AE, van Vliet P, Koch BCP. Clinical Experience with Off-Label Intrathecal Administration of Selected Antibiotics in Adults: An Overview with Pharmacometric Considerations. Antibiotics (Basel). 2023 Aug 5;12(8):1291.

 2. Arheilger L, Barbagallo M, Rancic GS, et al. Intraventricular antibiotics for severe central nervous system infections: a case series. Sci Rep. 2024 Nov 16;14(1):28267.

 3. Brotis AG, Churis I, Karvouniaris M. Local complications of adjunct intrathecal antibiotics for nosocomial meningitis associated with gram-negative pathogens: a meta-analysis. Neurosurg Rev. 2021 Feb;44(1):139-152.

 4. Liu SP, Xiao J, Liu YL, et al. Systematic review of efficacy, safety and pharmacokinetics of intravenous and intraventricular vancomycin for central nervous system infections. Front Pharmacol. 2022 Nov 18;13:1056148.

 5. Shah SS, Ohlsson A, Shah VS. Intraventricular antibiotics for bacterial meningitis in neonates. Cochrane Database Syst Rev. 2012 Jul 11;2012(7):CD004496.

 6. Hu Y, He W, Yao D, et al. Intrathecal or intraventricular antimicrobial therapy for post-neurosurgical intracranial infection due to multidrug-resistant and extensively drug-resistant Gram-negative bacteria: A systematic review and meta-analysis. Int J Antimicrob Agents. 2019 Nov;54(5):556-561.

 7. Hu Y, Li D, Zhang G, et al. Intraventricular or intrathecal polymyxin B for treatment of post-neurosurgical intracranial infection caused by carbapenem-resistant gram-negative bacteria: a 8-year retrospective study. Eur J Clin Microbiol Infect Dis. 2024 May;43(5):875-884.

 8. Tunkel AR, Hasbun R, Bhimraj A, et al. 2017 Infectious Diseases Society of America's Clinical Practice Guidelines for Healthcare-Associated Ventriculitis and Meningitis. Clin Infect Dis. 2017 Mar 15;64(6):e34-e65.

 9. Pfausler B, Spiss H, Beer R, et al. Treatment of staphylococcal ventriculitis associated with external cerebrospinal fluid drains: a prospective randomized trial of intravenous compared with intraventricular vancomycin therapy. J Neurosurg. 2003 May;98(5):1040-4.

 10. LeBras M, Chow I, Mabasa VH, et al. Systematic Review of Efficacy, Pharmacokinetics, and Administration of Intraventricular Aminoglycosides in Adults. Neurocrit Care. 2016 Dec;25(3):492-507.

 11. Lewin JJ 3rd, Cook AM, Gonzales C, et al. Current Practices of Intraventricular Antibiotic Therapy in the Treatment of Meningitis and Ventriculitis: Results from a Multicenter Retrospective Cohort Study. Neurocrit Care. 2019 Jun;30(3):609-616.

 12. McCracken GH Jr, Mize SG, Threlkeld N. Intraventricular gentamicin therapy in gram-negative bacillary meningitis of infancy. Report of the Second Neonatal Meningitis Cooperative Study Group. Lancet. 1980 Apr 12;1(8172):787-91.

 13. Shah SS, Ohlsson A, Shah VS. Intraventricular antibiotics for bacterial meningitis in neonates. Cochrane Database Syst Rev. 2012 Jul 11;2012(7):CD004496.

 14. Karvouniaris M, Brotis A, Tsiakos K, et al. Current Perspectives on the Diagnosis and Management of Healthcare-Associated Ventriculitis and Meningitis. Infect Drug Resist. 2022 Feb 28;15:697-721.

 15. Shofty B, Neuberger A, Naffaa ME, et al. Intrathecal or intraventricular therapy for post-neurosurgical Gram-negative meningitis: matched cohort study. Clin Microbiol Infect. 2016 Jan;22(1):66-70.