Successful treatment of extensively drug-resistant Acinetobacter baumannii ventriculitis with polymyxin B and tigecycline- a case report

Acinetobacter baumannii has emerged as a major nosocomial central nervous system infection, with mortality rates ranging from 15% to 71% for acinetobacter meningitis [1]. Successful treatments with intraventricular/intrathecal polymyxins have been reported; however, the incidence of potential toxicity was not negligible [2]. Here, we report a successful microbiological cure for extensively drug-resistant A. baumannii ventriculitis using intravenous and intraventricular polymyxin B together with intravenous tigecycline.

A 57-year-old male patient was transferred from another hospital to the neurosurgical intensive care unit of the Tangdu hospital. The patient had experienced a falling accident at work and then was diagnosed as having a severe traumatic brain injury. The patient underwent a decompressive craniectomy and an external ventricular drain (EVD) insertion. Eight days after the operation, the patient presented with remittent fever (peak at 39.0 °C) associated with meningeal signs and altered mental status. Empirical antimicrobial therapy was initiated with meropenem and vancomycin. Twelve days after the operation, the patient was transferred to our intensive care unit. A right frontal EVD was inserted because of bilateral hydrocephalus. Cerebrospinal fluid (CSF) analysis revealed a WBC count of 5550 × 10E6/L, with 70% polymorphonuclear leukocytes, a glucose concentration of 1.11 mmol/L, and protein levels of 3662.1 mg/L. CSF sample was cultured on columbia agar with 5% sheep blood at 35 °C in aerobic conditions for 48 h. The bacteria were identified by an automated mass spectrometry microbial identification system (VITEK MS, bioMérieux). On day 5 of the hospitalization, the patient’s CSF culture showed that the A. baumannii was susceptible only to polymyxin (MIC = 1 μg/mL) and tigecycline (MIC ≤1 μg/mL). According to the Clinical and Laboratory Standards Institute criteria, polymyxin breakpoints are susceptible (≤2 μg/ml) and resistant (≥4 μg/ml). For tigecycline, the U.S. Food Drug Administration proposed breakpoints are susceptible (≤2 μg/ml), intermediate (4 μg/ml) and resistant (8 μg/ml). A chest computed tomography scan showed lung infiltrates, which were suggestive of pneumonia. Sputum sample was cultured on Columbia agar with 5% sheep blood and MacConkey at 35 °C in aerobic conditions for 48 h. The same strain of A. baumannii was isolated from the sputum (polymyxin susceptibility was not tested). The antimicrobial therapy was changed to intraventricular (IVT) polymyxin B (50,000 U q24h), intravenous polymyxin B (450,000 U q12h), and tigecycline (50 mg q12h). The patient became afebrile 5 days after the polymyxin B and tigecycline therapy, with negative CSF cultures thereafter. However, in the treatment process, decreased CSF drainage and a contractible right ventricle were observed gradually. Ten days after the right frontal EVD placement, the right EVD was removed, and another EVD was inserted into the left lateral ventricle. Antimicrobial treatment was switched to IVT polymyxin B (25,000 U q12h), intravenous polymyxin B (475,000 U q12h), and tigecycline (50 mg q12h). A contractible left ventricle was also observed after the IVT polymyxin B administration. Next, the antimicrobial regimen was changed by stopping the IVT polymyxin B administration and continuing the intravenous polymyxin B (500,000 U q12h) and tigecycline (50 mg q12h) administration for another 14 days until the patient’s clinical conditions were stable. On day 31 of the hospitalization, the patient was discharged.

Over the years, Acinetobacter baumannii, which is associated with post-neurosurgical meningitis and ventriculitis, has increasingly been regarded as an important nosocomially acquired pathogen [1, 3]. Statistical data showed that 3.6–11.2% of post-neurosurgical meningitis cases are caused by A. baumannii [2, 4]. Antimicrobial-resistant A. baumannii are divided into three categories: multidrug-resistant (MDR), extensively drug-resistant (XDR) and pandrug-resistant (PDR). XDR is defined as non-susceptibility to all penicillins and cephalosporins (including inhibitor combinations), fluroquinolones, aminoglycosides, and carbapenems. [5]. In the present case, XDR A. baumannii that was susceptible only to polymyxin and tigecycline was cultured from the cerebral spinal fluid (CSF) and sputum. As a result of the poor blood-brain barrier penetration, intraventricular (IVT) therapy polymyxin B was used through the external ventricular derivation (EVD).

Emerging evidence has indicated that intrathecal (IT) or IVT colistin administration is a safe and effective treatment for XDR A. baumannii meningitis [6]. Karaiskos et al. [2]. summarized 36 studies and a total of 81 patients who were diagnosed with meningitis secondary to neurosurgical procedures. A total of 89% (72/81) of the cases treated with IVT/ITH colistin were eventually cured, and the median time to achieve sterilization of the CSF was 4 days. De Bonis et al. [3]. compared the outcomes of the XDR A. baumannii ventriculomeningitis patients treated with intravenous (IV) colistin or IV plus IVT colistin. Compared with 33.3% of the cases in the IV alone group, 100% of the cases achieved CSF sterilization (a negative CSF culture result) in the IV + IVT group. The results showed that IVT colistin administration is significantly more effective than IV colistin alone [3]. To the best of our knowledge, polymyxin B and colistin (polymyxin E) were regarded as equivalent because of their similar chemical structures and their activity spectra [7]. However, compared to colistin, there is limited clinical data for the use of polymyxin B in ventriculitis treatment. Piparsania et al [8]. showed successful treatment of multidrug-resistant A. baumannii neonatal meningo-ventriculitis with IVT polymyxin B in combination with IV netilmicin and polymyxin B. IVT polymyxin B (40,000 units per dose) was given alternate day for four weeks. In the present case, sterilization of CSF was detected 5 days after the IVT polymyxin B administration. Hence, we speculate that IVT polymyxin B administration could be as effective as IVT colistin administration.

In the present case, tigecycline was used intravenously. Tigecycline, which belongs to a new class of antibiotics known as the glycylcyclines [9], has demonstrated excellent activity against Acinetobacter strains [10]. Pallotto et al. [11] showed the weak penetration of tigecycline to the CSF in a patient with ventriculo-atrial shunt infection. The average CSF concentration of tigecycline equals to 7.9% of the serum concentration. That is the reason why tigecycline is not currently recommended for Acinetobacter ventriculitis. Recently, Lauretti et al. reported a successful case of the IVT tigecycline use to treat PDR A. baumannii meningitis [12].

Potential toxicity is a concern associated with local administration of polymyxins. Chemical ventriculitis and meningitis which usually cause fever and altered mental state, are the most severe adverse effects reported with IVT/ITH polymyxin treatment. Karaiskos reviewed the literature and found that out of 81 patients with IVT/ITH colistin administration, chemical meningitis and chemical ventriculitis were diagnosed in 3 (3.7%) and 2 (2.4%) cases, respectively [2]. A recent retrospective study showed that out of 9 patients with IVT colistin administration, no cases of chemical meningitis were encountered [3]. However, to the best of our knowledge, limited reports on the adverse effects of IVT/ITH polymyxin B were published.

In the treatment process, two observations were noted. First, the transient ventricular adhesion due to the IVT polymyxin B administration was observed from the CT scan (Fig. 1). Initially, the IVT polymyxin B was administered through the right EVD. However, decreased CSF drainage and a contractible right ventricle were detected after the IVT polymyxin B administration. Six days later, when the left EVD were used for the polymyxin B administration, left ventricle shrinkage was observed. These phenomena indicate that there is a direct correlation between polymyxin B usage and ventricular adhesion. Second, at the 6-month follow-up, the neurological conditions of the patients had not improved, and the patient was still in a comatose status.