Postneurosurgical infections due to multidrug-resistant (MDR) organisms have become a nightmare for 21st-century neurosurgeons. This period has already been called the dawn of the postantibiotic era.²¹ The accidental discovery of penicillin by Sir Alexander Fleming heralded the era of antibiotics and of human dominance over pathogens. Approximately 1 century later, at least in critical care settings, the pathogens seem to be dominating again. Acinetobacter baumannii is one such pathogen that has become the scourge of neurosurgical ICUs worldwide, resulting in several global outbreaks.²⁷ The mortality rate due to MDR A. baumannii meningitis can range from 15% to 72.7%.^17,21,26

The data from SENTRY (Antimicrobial Surveillance Program), MYSTIC (Meropenem Yearly Susceptibility Test Information Collection), and other sources have reported the emergence of MDR A. baumannii from Europe, North America, Argentina, Brazil, China, Taiwan, Hong Kong, Japan, Korea, and even remote islands like Tahiti.²¹ It has been nicknamed “Iraqibacter” due to the increased number of nosocomial infections in military hospitals during Operation Iraqi Freedom.¹ In this evolutionary arms race, there is an urgent need to review, evaluate, and find new ways to tackle this difficult problem. The intrathecal or intraventricular (IT) administration of antibiotics has been tried as a last resort in many cases of severe neuro-infections not responding to traditional intravenous (IV) regimens.¹⁰ This analysis is an attempt to review the role of intrathecal or intraventricular antibiotic therapy and its effect on the mortality rate.

Methods

The aim of the study was to review the benefit of combined IV + IT therapy over exclusive IV therapy. The outcome studied was mortality attributed specifically to A. baumannii meningitis. A detailed protocol about the literature search, inclusion and exclusion criteria, selection of cases, and statistical methodology was developed. The guidelines laid out by the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) were followed. Two investigators (N.M. and V.N.) performed the literature search. To minimize selection bias, 2 reviewers (A.S. and D.P.) reviewed the articles independently. In case of doubt, a mutual consensus was reached after discussion.

Literature Search Strategy

A computerized search of titles and abstracts from January 1960 to May 2017 was performed in PubMed, Cochrane Library database, ScienceDirect, and EMBASE. A combination of key word searches was made to build a search strategy. The non-English articles were also searched. Two non-English articles were reviewed after translation. An independent Web search was made with the key words to include any unpublished literature. The reference lists of final selected articles were examined and reviewed (Fig. 1). Key words that were used to build the search directory were “Acinetobacter baumannii,” “Neurosurgery,” “postoperative,” and “infections.” The example of a PubMed search directory is as follows: (“acinetobacter baumannii”[MeSH Terms]) OR (“acinetobacter”[All Fields] AND “baumannii”[All Fields]) OR “acinetobacter baumannii”[All Fields]) AND (“neurosurgical procedures”[MeSH Terms] OR (“neurosurgical”[All Fields] AND “procedures”[All Fields]) OR “neurosurgical procedures”[All Fields] OR “neurosurgery”[All Fields] OR “neurosurgery”[MeSH Terms]). Additional search terms used were “intrathecal,” “intraventricular,” “colistin,” “polymyxin,” “meningitis,” “craniotomy,” and “multidrug resistant.”

The PRISMA flow diagram of systematic literature search.

Results

Five studies with 126 patients were selected (Table 1). All patients had undergone a neurosurgical procedure or procedures and had tested positive for infection with A. baumannii. The selected studies were matched to evaluate homogeneity. The test for homogeneity was performed, and the I² value was 28% (Fig. 2). The 2 groups were called “IV” for the patients who received intravenous antibiotics only, and “IV + IT” for those who received intravenous and intrathecal antibiotics. The total number of patients in the IV-only group was 73, and the number of patients in the IV + IT group was 53. The mean age in the study cohort was 48.8 years. The male/female ratio was 2.1. The underlying disease was traumatic brain injury in 35 cases, intracerebral bleed in 56 cases, and brain neoplasm in 20 cases. The intravenous dose of colistin ranged from 3.75 to 8.8 MIU per day, and the mean duration of IV therapy was 27 days. The dose of intrathecal colistin ranged from 125,000 to 250,000 IU per day, and the mean duration of intrathecal colistin was 21 days (Table 2).

Funnel plot showing the distribution of studies for assessment of bias. SE = standard error.

The data from all the subgroups in the 5 studies were analyzed by the Cochran-Mantel-Haenszel method and charted in a forest plot (Fig. 3). The overall calculated odds ratio for mortality after pooling all the data was 0.16 (95% CI 0.06–0.40, p < 0.0001) for the IV + IT group. The patients who received IV + IT therapy had an 84% lower risk of dying due to infection compared with those who received only IV therapy. This is a very significant risk reduction for the cohorts that received IV + IT therapy.

Forest plot with mortality as the outcome showing the calculated odds ratios, weights, and confidence interval of each study, and the overall pooled odds ratio. The event here is the number of deaths that have occurred in the respective groups. M-H = Mantel-Haenszel.

Discussion

Acinetobacter baumannii (from the Greek word “akinetos,” meaning nonmotile) was first isolated by the Dutch microbiologist Beijerinck in 1911 from soil, as described in Antunes et al.² From its low profile of an opportunistic organism with low virulence,²⁰ it has shot into prominence as one of the foremost concerns of the critical care community today.

“Multidrug resistance” in literal terms means resistance to < 1 antimicrobial agent. The scientific definition is “resistance to three or more anti-microbial classes.”¹⁵ The antimicrobial resistance of A. baumannii significantly increased worldwide between 2004 and 2009.¹² The evolution of carbapenem-resistant A. baumannii has been the most worrying trend. Imipenem resistance in A. baumannii has reached > 50% from a worldwide collection between 2005 and 2009. The once-discarded antibiotics have found new relevance in this context. The most important among them in the fight with A. baumannii is colistin, also known as polymixin E, which is an antibiotic produced by some strains of the bacterium Paenibacillus polymyxa. Structurally they are cyclic polypeptides.⁶ Colistin is given in the form of colistimethate sodium (CMS), which has fewer adverse reactions than colistin.

The pharmacokinetics of drug entry into the CNS is a complex process. It depends on the molecular size, electric charge, lipid solubility, plasma protein binding ability, the affinity to the active transport system, CSF flow, and meningeal inflammation.¹⁹ There is a wide variation from one patient to another in the CSF concentration of the drug, despite giving the same dosage. James et al. found that the concentration of drug in CSF in patients receiving the maximum dosage of intravenous antibiotics alone was for the most part ≤ 5 μg/ml and > 5 μg/ml for drugs given by the intraventricular route.⁸ The colistin penetration into CSF may be as low as 5%¹⁶ to 25%.⁸ The poor penetration into CSF may be the reason for very high mortality rates in the subgroup not receiving intrathecal therapy. The risk of nephrotoxicity with polymyxin therapy ranges from 10% to 18%.⁶ There is a substantial risk of nephrotoxicity with use of colistin that may require dosage reduction and further jeopardize the outcome of treatment.⁵ This drug has also been linked to neurotoxicity. Polymyxins can cause neuromuscular blockade due to inhibition of acetylcholine release. The majority of postneurosurgical A. baumannii infections are associated with indwelling ventriculostomy tubes or CSF leaks.²⁴ The median time for development of postoperative meningitis is 12 days (range 1–40 days). The risk factors predisposing patients to postneurosurgical A. baumannii meningitis are CSF leaks,¹¹ prolonged external ventricular drainage,¹² prolonged duration of surgery,¹¹ opening of sinus during surgery,¹² wound infection,¹¹ repeat surgery,²² and critical illness.¹²

The 2017 Infectious Diseases Society of America’s guidelines²⁵ state that intraventricular antimicrobial therapy should be considered in patients with health care–associated meningitis who respond poorly to systemic therapy. It is also recommended to clamp the ventricular or intrathecal drain for a period of 15–60 minutes to allow the drug to achieve an equilibrium. The actual intraventricular dosage is determined by calculating the “inhibitory quotient.” After administering the first dose, the trough levels of antibiotic are determined from the sample of CSF taken just before giving the second dose. The inhibitory quotient is calculated by dividing the trough level by the minimum inhibitory concentration for the organism. The inhibitory quotient should exceed 10–20 for consistent CSF sterilization. The dosage also needs adjustment depending on the size of the ventricle and daily output from the drain. For the infected CSF drain or shunt tube, removal is recommended.

The recommended doses of colistin for intraventricular administration are 5 mg or 62,500 IU per day (1 mg of CMS = 12,500 IU) in adults and 2 mg or 25,000 IU per day in children.²⁵ The recommended pediatric dosage of CMS per day is between 50,000 and 75,000 IU/kg/day (4–6 mg/kg of body weight/day) in Europe and 83,000–166,000 IU/kg/day (6.6–13.3 mg/kg/day) in the US.¹³ The British Society for Antimicrobial Chemotherapy Working Party recommended that the initial intraventricular dose should be based on the ventricular volume as determined by imaging.³ To avoid increasing intracranial pressure before drug instillation, an equal volume of CSF must be removed. After instillation of the drug, a saline flush is done to minimize the amount of drug left over in the drain tube. The patient’s neurological status must be monitored both before and after giving the drug via the intraventricular route.

De Bonis et al.⁴ reported a CSF sterilization rate of 33% in the IV-only group and 100% in the IV + IT group (p = 0.009). The mean time for CSF sterilization was 21 days. Tuon et al.²⁶ reported a high mortality rate of 72.7%. Only 2 patients in their study received intrathecal therapy. Two patients who received IV + IT therapy also died. It is interesting to note that only 69% of patients who died received appropriate, sensitive antibiotics. Thus, an inappropriate antibiotic regimen could have led to the ineffectiveness of IV + IT therapy in this study. Transmission of infection from other sources like the lungs and urinary tract is a concern in the ICU setting. Fotakopoulos et al.⁷ found that 5.6% of external ventricular drains were infected with A. baumannii.

The importance of strict adherence to hand hygiene and hospital infection protocols cannot be overemphasized. It must be remembered that starting IV + IT therapy does not guarantee the survival of patients. In the studies reviewed, 8 deaths occurred in IV + IT group. What has caused the deaths in these cases? This could be explained in part due to late initiation of intraventricular therapy. It must be inferred that in usual clinical practice, IV therapy is started first, and when there is no improvement, then as a last resort IV + IT therapy is started. We challenge this notion and hypothesize that this may lead to worse outcomes. The other reason could be the variation in the sensitivity patterns of the organism to the antibiotics. Moon et al.¹⁸ found that carbapenem-resistant strains had a higher mortality than other strains. It is not possible to answer this question based on the current study; randomized controlled trials are needed to address these issues. In the absence of clear guidelines, clinicians are faced with a difficult dilemma in caring for these patients. In the situation of clinical nonimprovement with systemic therapy, it is prudent to get intrathecal or intraventricular access and commence the therapy at the earliest possible time. If sensitivity is demonstrated only for colistin, then we recommend that IV + IT therapy be started at the earliest opportunity.

Conclusions

The data analyzed from this study show that the inclusion of intrathecal or intraventricular antimicrobial therapy with intravenous therapy is associated with an 84% lower chance of mortality in postneurosurgical infections due to A. baumannii. The intrathecal route should be strongly considered when dealing with postneurosurgical A. baumannii infections.

Disclosures

The authors report no conflict of interest concerning the materials or methods used in this study or the findings specified in this paper.

Author Contributions

Conception and design: Nanda, Mohammed, Savardekar. Acquisition of data: Mohammed, Narayan. Analysis and interpretation of data: Mohammed. Drafting the article: Mohammed, Savardekar, Patra, Narayan. Critically revising the article: all authors. Reviewed submitted version of manuscript: all authors. Approved the final version of the manuscript on behalf of all authors: Nanda. Statistical analysis: Mohammed, Narayan. Administrative/technical/material support: Nanda, Savardekar, Patra. Study supervision: Nanda.