Successful Treatment of Multidrug-Resistant Acinetobacter baumannii Central Nervous System Infections with Colistin

Acinetobacter baumannii may cause severe central nervous system (CNS) infections, such as meningitis and ventriculitis, especially in patients undergoing neurosurgical procedures or head trauma. Mortality ranges from 20% to 27% in different case series (3). Carbapenems have been considered the treatment of choice for severe infections. However, increasing numbers of carbapenem-resistant Acinetobacter isolates have been reported worldwide, dramatically reducing the existing therapeutic options (1).

Over the past 30 years, colistin use has been limited due to concerns regarding its toxicity along with the development of newer antibiotics with better safety profiles (5, 7). However, the increasing incidence of multidrug-resistant A. baumannii in addition to a lack of new antimicrobial agents has reawakened interest in the utilization of colistin due to its good activity against this organism (7).

Although of particular interest, reports on the management of CNS infections due to multidrug-resistant A. baumannii with colistin are relatively scarce. In order to build on this experience, we reviewed the literature and summarized the up-to-date data. Reports for this review were found through a search of PubMed and of references cited in relevant articles. To the best of our knowledge, eight reports (one prospective nonrandomized study [11], one retrospective study [4], and six case reports [2, 6, 8-10, 13]) overall have been published. These include a total of 14 patients who suffered from CNS infection (meningitis or ventriculitis) due to multidrug-resistant A. baumannii and were treated with colistin intravenously and/or either intrathecally or intraventricularly (Table 1).

TABLE 1.

Reported cases of multidrug-resistant A. baumannii CNS infections treated with colistin

Characteristic	Data from reference no.:
	11	8	9	13	2	4	6	6	10
No. of patients	5	1	1	1	1	2	1	1	1
Age (yr)/sex	NRa	55/F	14/Male	41/F	49/F	NR	16/Male	34/F	28/Male
Underlying condition	NR	MEN	CT	SH	MEN	NR	HEM	SH	CT
CNS infection	M, V	M	M	M	M	NR	V	V	M
Ventricular tubeb	NR	Yes (20)	Yes (42)	Yes (8)	NR	NR	Yes (16)	Yes (7)	NR
MIC (μg/ml)	NR	1	0.3	0.5	NR	NR	≤4	≤4	NR
Colistin methanesulfonate dosage, in mg/kg of body wt/day (route)	2.5-5 (i.v.)	5 (i.v.)	5 (i.v.)	5-10 (ith)e	3.2 (ith)e	4 (i.v.) + 20 (ith)e	5-10 mg/12 h (iventr)	5-10 mg/12 h (iventr)	80 mg/8 h (i.v.) + 1.6-3.2 mg/d (iventr)
Duration of colistin treatment, in days	NR	15	30	22	17	NR	19	17	63
Antibiotic(s) coadministeredc	No	No	No	No	SAM + VAN (i.v.)	NR	SAM (3 days) + TOB (19 days) (i.v.)	TOB (17 days) (i.v.)	AMK + TEC (63 days) (i.v. + iventr)
CSF sterilizationd (days)	NR	4	5	1	6	NR	2	6	NR
Levels of colistin	No	Yes	Yes	Titers	No	No	Titers	Titers	No
Infection outcome	4 of 5 Cured	Cured	Cured	Cured	Cured	Cured	Cured	Cured	Cured

^aNR, not reported; MEN, meningioma; CT, cranial trauma; SH, subarachnoid hemorrhage; HEM, hemangioblastoma; M, meningitis; V, ventriculitis; F, female; i.v., intravenous; ith, intrathecal; iventr, intraventricular.

^bIn parentheses is the number of days from catheter insertion to diagnosis of infection.

^cSAM, ampicillin-sulbactam; VAN, vancomycin; TOB, tobramycin; AMK, amikacin; TEC, teicoplanin. The numbers in parentheses are the duration of treatment.

^dTime after initiation of colistin treatment.

^eMilligrams per day.

In none of these cases was treatment discontinued due to adverse effects; however, some authors emphasized the need of dose modification, especially in patients with renal insufficiency (11). Although nephrotoxicity, neurotoxicity, and neuromuscular blockage have been described in association with systemic use of colistin (5), in the reports reviewed here intrathecal or intraventricular administration exhibited a safe profile even after prolonged use. Generally, intrathecal or intraventricular administration of colistin alone or in combination with systematic administration was well tolerated and could be an effective salvage therapy in patients with CNS infections due to A. baumannii strains resistant to conventional antibiotics. It is of note that sterilization of cerebrospinal fluid (CSF) was achieved in all reported cases within a median of 4.5 days (range 1 to 6 days). Moreover, cure was achieved in all but one case reported (13 out of 14 cases; cure rate, 93%).

By using a broth dilution assay of polymyxins, Jimenez-Mejias et al. determined that about 25% of colistin penetrated the CSF and reached bactericidal concentrations for the entire dosing period. Colistin produced a peak of 5 μg/ml in serum and a peak of 1.25 μg/ml in CSF 1 h after intravenous administration (8, 9). To date it is well known that colistin methanesulfonate, once administered, is hydrolyzed to a mixture of products (colistin sulfate or base, which is more microbiologically active, and its partially sulfomethylated derivatives). Therefore, microbiological assays lack specificity, since the concentrations quantified are apparent values for the combined antibacterial activity of the mixture of hydrolytic products of colistin methanesulfonate (12).

From this analysis it becomes evident that, indeed, experience in multidrug-resistant Acinetobacter CNS infections is scattered. Besides, some of the existing reports provide inadequate information to come to safe conclusions. Consequently, more extensive pharmacokinetic and pharmacodynamic studies as well as randomized controlled trials are needed to evaluate the clinical use of colistin and the desirable concentrations in CSF after intravenous administration.