Intrathecal and intraventricular antibiotics for postoperative Gram-negative meningitis and ventriculitis.

BACKGROUND: Postoperative meningitis is a growing cause of concern, especially with the evolution of multidrug-resistant organism. The authors evaluate the use of intraventricular/intrathecal (IVT/IT) antibiotics for postoperative gram-negative meningitis in patients whom intravenous antibiotics were ineffective.
METHODS: Medical records were retrospectively reviewed and neurosurgery patients with gram-negative postoperative infection meningitis/ventriculitis were enrolled in the study. Their demographics, hospital course, and outcomes were recorded in a pro forma and analyzed using Statistical Package for the Social Sciences, version 19.
RESULTS: The review identified 21 patients with postneurosurgical gram-negative meningitis/ventriculitis who were treated with IVT or IT antibiotics. The most common organism was Acinetobacter species (n = 14; 66%). Amikacin was used in 7 patients, polymyxin B in 9 patients, and colistin in 5 patients. A combination of antibiotics was used in one patient. Cerebrospinal fluid sterility was achieved in all patients with no incidence of relapse. There was a single death, though that was not related to the infectious process as the patient had a massive pulmonary embolism.
CONCLUSION: The findings of this study suggest that IVT and IT antibiotic therapy is a useful option in patients who are nonresponsive to standard intravenous therapy with little or no side effects.

INTRODUCTION

Postoperative meningitis and ventriculitis are one of the grave delayed complications of neurosurgical procedures. Recently, an increase in incidence of postoperative meningitis has been observed.[26] The concern is further aggravated by rising incidence of multidrug-resistant organisms in such infections[1122] that makes their treatment a daunting challenge. The most common causative agents are gram-positive organisms, such as Staphylococcus aureus;[6] however, in case of gram-negative organisms the prognosis is much poor.[6]

Intraventricular (IVT) and intrathecal (IT) antibiotics were reported to be used as back as a century ago.[4] However, still there exist concerns regarding its indications, efficacy, and even safety. The primary advantage for this route is that it provides a direct path for antibiotic delivery bypassing the blood–brain barrier, providing larger effective concentration of antibiotics at the infected site with reduced, if any, systemic side effects.[19] The disadvantages are limited data available for calculation of effective dosage, and the risk for infection due to access to the subarachnoid space and ventricles. The two common routes of access for administration include ventricular access devices such as external ventricular drain (EVD) and lumbar drain (LD), both of which carry minimal procedure-related complications.[2027]

The use of IVT and IT antibiotic therapies has recently regained interest. An earlier infantile randomized control trial identified a three-fold increased relative risk of mortality with the use of IVT/IT and intravenous[1] antibiotics compared to IV antibiotics only. The trial was subjected to significant criticism due to the duration of IV antibiotics and the route of administration of IT antibiotics, which in this case was serial lumbar puncture.[13] Recent studies on using IVT polymyxin and IVT gentamycin in adults have reported much higher cure rates.[723]

Most papers addressing this very important neurosurgical topic share the common limitations of small sample size, heterogeneity of patient population, and treatment plans.[16182325] We, therefore, retrospectively analyzed our own results with the management of such patients to contribute to the existing body of literature. We have analyzed and shared clinical features, hospital course, treatment plans, and eventual outcomes of patients treated for postoperative gram-negative meningitis/ventriculitis with IVT/IT antibiotic therapy.

MATERIALS AND METHODS

The Aga Khan University Hospital is an ISO certified tertiary care hospital with Joint Commission International Accreditation (JCIA) and established Neurosciences program. The hospital maintains a vigorous medical recording system based on electronic as well as manually compiled patient records, logged through International Classification of Diseases (ICD) coding. We reviewed our medical records for all patients admitted to the hospital from January 2008 to December 2012, with a history of neurosurgical procedure performed within 2 weeks of presentation and later presenting with “culture proven” gram-negative ventriculitis/meningitis. Patients who responded to IV antibiotic therapy alone, not requiring IVT or IT antibiotics were excluded. The study was exempted by the Ethics Review Committee as it was a retrospective chart review.

IV antibiotics were started in all patients as soon as a diagnosis was made on cerebrospinal fluid (CSF) analysis. All patients underwent CSF analysis once at least every third day until the CSF culture was negative or the patient was discharged. Patients were started on IVT/IT antibiotics only in cases where IV antibiotics were ineffective despite 5 days of IV therapy, or if the first CSF analysis showed frank pus. IVT/IT antibiotics were administered either through EVD or LD. Dosages were based on available literature on the respective minimum inhibitory concentrations (MICs). IVT/IT antibiotics were continued until three consecutive CSF cultures were negative.

Data were collected on a standardized pro forma for all relevant clinical, laboratory, and radiological variables. Data were analyzed on Statistical Package for the Social Sciences, version 19 (SPSS IBM v 19). Continuous variables with normal distribution were expressed with mean and standard deviation, while continuous data without normal distribution were expressed with median and range. Categorical data were expressed as percentage and proportions. Mann–Whitney U-test was applied to compare nonparametric numerical data. Categorical data were compared using Chi-square test. A P value of 0.05 was regarded as significant.

Biochemical outcomes were based on the concepts of cure, relapse, and failure. “Cure” was defined as three consecutive negative CSF culture results and no relapse after withdrawal of antibiotics; “Relapse” was defined as isolation of the same organism from the CSF or central nervous system (CNS) lesion within 3 weeks of completing therapy; and “Failure” was defined as no biochemical response to treatment. Clinical outcomes were reported as pretreatment and posttreatment Glasgow Coma Scale (GCS) score. We further classified the standard GCS.[14] GCS score of 3 or less was classified as unfavorable and 4 or more as favorable. Outcomes of the patients were determined from medical records and were based on the detailed assessment of the patient made and documented on the last clinical follow-up by the attending physician.

RESULTS

During the study period, 94 cases were reported of postoperative meningitis/ventriculitis at our institution. Of these, 63 were culture proven gram positive and hence excluded from our study. The remaining 31 patients either had isolated gram-negative meningitis/ventriculitis or mixed gram-positive/negative infection, with 21 patients out of these requiring treatments with IVT/IT antibiotics.

Out of the total 21 patients enrolled in the study, 16 were males and 5 were females. Their mean age was 41.7 ± 11 years (range: 26–58 years). All except one of these patients had cranial surgery. Details of individual cases are shown in Table 1. Of the total, 8 patients were admitted with a GCS of >12, while the remaining had a GCS ≤12. In our setting, postoperative gram-negative meningitis/ventriculitis was most commonly found in patients of traumatic brain injury (n = 8), followed by brain tumor (n = 7). The primary diagnosis of the patients can be observed in Graph 1.

Table 1

Clinical parameters of individual patients

Graph 1

Primary diagnosis of the patients who developed Gram Negative Meningitis/Ventriculitis

Acinetobacter species was the most common organism isolated on CSF culture (n = 14) followed by Klebsiella pneumoniae (n = 3), Pseudomonas aeruginosa (n = 1), and Enterobacter cloacae (n = 1). One patient had polymicrobial growth. Three antibiotics were used in the IVT/IT treatment group: Amikacin, polymyxin B, and colistin. The indication was primarily based on the sensitivity of pathogenic organisms cultured on CSF of individual patients; Table 2 shows the resistance pattern of the organism. Amikacin was used in 7 patients, polymyxin B in 9 patients, and colistin in 5 patients. A combination of two IVT/IT antibiotics, polymyxin B and amikacin, was used in one patient. An EVD was used as the route of administration in 13 cases (62%), while a LD was used in 8 (38%) patients. Median duration between starting of IT/IVT after being diagnosed with postneurosurgical meningitis/ventriculitis was 3 days (range 0–7 days). Median duration of IVT/IT therapy was 15 (9–25) days. CSF sterility was achieved in all patients, with a median time to sterility being 7.1 ± 3.8 (range 2–16) days (EVD 7.08 ± 3.37 days vs. LD 7.1 ± 4.6 days, P value 0.176). Only two patients achieved CSF sterility within 48 hours.

Table 2

Antimicrobial resistance of causative organisms from CSF in 21 episodes of postneurosurgical gram-negative meningitis treated with IVT/IT therapy. Patient 8 and 18 had two causative organisms as shown in Table 1. Patient 13 has not been included in the table

Discharge GCS improved in 2, remained the same in 7, and deteriorated in 12 patients. Median hospital stay was 38 days (20–257 days), with the EVD group [46 (29–72)] having a longer stay than LD group [32.5 (20–76) days)]. At 6 months follow-up, favorable outcomes were seen in 14 (66.7%) patients and 7 (33.3%) patients remained in an unfavorable state. One (5%) patient died during treatment and although the patient had Acinetobacter species infection, the cause of death was found to be a massive pulmonary embolus. No adverse effects attributable to IVT/IT therapy were noted in the study.

DISCUSSION

Management of hospital acquired meningitis and/or ventriculitis is a daunting challenge for clinicians.[24] The increasing incidence of multidrug-resistant organism has further worsen the situation.[5] In neurosurgical postoperative patients the complexity is compounded even further as the underlying CNS pathology causes disruption of the blood–brain barrier and in-turn alters both pharmacokinetics and drug penetration.[28] This has been shown in patients with traumatic brain injury who were administered intravenous vancomycin, but the drug consistently failed to achieve concentrations above the MIC.[3] In these circumstances, IVT/IT provides a direct access, which can potentially overcome these issues. High CSF sterility rates along with relatively short time needed to achieve sterility in recent reports has supported the use of IVT/IT therapy in both bacterial and fungal infections.[28]

We herein report one of the largest series of patients with gram-negative CNS infections treated with IVT or IT therapy [Table 3]. We achieved 95% sterility rate in our study with one patient continuing extended IV course after completing IVT therapy. Time to achieve CSF sterility ranged from 2 to 16 days. Additionally, there was no relapse in our study. These findings are consistent when compared from the international data. In a recent study on postoperative meningo-ventriculitis by Remes et al.,[18] the time to achieve CSF sterility ranged from 1 to 12 days and there was just one case of relapse. Tangden et al.[23] reported their data of 13 patients with gram-negative meningo-ventriculitis treated with IT antibiotics and achieved CSF sterility of all patients without relapse. Though our experience was favorable for both IVT and IT approach, however, we do prefer IT over IVT due to the convenience of the procedure and lower comparative risk of complications.

Table 3

Previous studies with postneurosurgical gram-negative meningitis treated with IVT/IT therapy

Earlier studies have debated the use of LD as the route for antibiotic administration as there remained speculations with regards to effective delivery of administered antibiotic into the ventricular cavity and CSF cistern. Kaiser et al. demonstrated that a 5–10 mg injection of antibiotic via lumbar route showed significantly lower ventricular CSF drug level when compared to lumbar CSF, while with ventricular dug injection the levels in the lumbar CSF were comparatively much higher.[7] We, however, did not see a significant difference in the CSF sterility rates when comparing the two routes in our study. Remes et al. drew the same conclusions in their study and even though lumbar and ventricular CSF drug levels were not measured, the CSF sterility rates showed no significant difference either using IT or IVT therapy.[18]

Postneurosurgical meningitis and/or ventriculitis have a critically high mortality rate, ranging from 3 to 33%.[2121517] The fact that it is a relatively uncommon phenomenon has precluded it from undergoing any prospective data collection or randomized controlled trials. In our study, the mortality rate was 5%, which suggests that IT/IVT has valuable potential, especially in patients who can be started early and have a good baseline GCS as has been suggested by previous authors.[25]

In our series, Acinetobacter species was the most commonly encountered gram-negative organism which makes this series differ from other reports, looking specifically at gram-negative meningitis/ventriculitis, where Pseudomonas species is usually more common.[25] We have previously reported the relatively high incidence of Acinetobacter at our institution.[21] It is becoming a growing concern in patients with nosocomial meningitis especially as most cultured organisms are resistant to third and fourth-generation cephalosporins, and some carbapenems.[9] It has been shown to be an independent predictor of unfavorable outcomes in neonates and children[1016] and is the major cause of mortality in adults in one report.[25] The single mortality in our series also had Acinetobacter species infection though the cause of death was a pulmonary embolism.

The major reason for the cautioned use of IVT/IT therapy has been the significant toxicity that was reported by earlier studies.[814] These included seizures in up to 20% of the patients and chemical ventriculitis in as high as 60% of the patients,[814] though those were considered to be dose related. Few papers have also reported side effects such as transient hearing loss and seizures using gentamicin and vancomycin, though most recent studies have shown little or no serious adverse effects with the use of polymyxin B, colistin, and vancomycin.[182528] Meropenem and netilmicin which were used by Remes et al. for the first time also showed no adverse effects.[18] In our study, we did not encounter any patient with any serious adverse effects which supports the safety of IVT/IT therapy if used in the correct dosage.

CONCLUSION

The study describes the clinical outcomes of patients with gram-negative meningitis and/or ventriculitis being treated with IVT/IT antibiotics. In our study, there was good CSF sterility rate with no relapse. This, along with minimal side effects, makes IVT/IT an effective and relatively safe treatment option, especially in patients nonresponsive to standard intravenous therapy. However, prospective randomized trials with high sample size are needed to validate the findings.

Competing Interest

The authors declare that they have no financial or nonfinancial competing interest.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.