Share this article on:

Treatment With Intrathecal Colistin of a Resistant Extended-Spectrum β-Lactamase-Producing Klebsiella pneumoniae in the Cerebrospinal Fluid

Enfanto, Christopher M. PharmD, BCPS*; Shin, Sonya MD, MPH

Infectious Diseases in Clinical Practice: July 2007 - Volume 15 - Issue 4 - p 274-277
doi: 10.1097/IPC.0b013e31802df0d7
Case Reports

A 24-year-old man hospitalized 2 weeks prior for emergent craniotomy and craniectomy after a high-speed motor vehicle accident was readmitted to the hospital with high fevers of unknown origin, hydrocephalus, and mental status changes. The patient did not improve on empirical antibiotic therapy or subsequent antibiotic therapy combinations based on cultures and sensitivities. A cerebrospinal fluid culture revealed a multi-drug-resistant extended-spectrum β-lactamase-producing Klebsiella pneumoniae. The patient was started on a 7-day course of intrathecal colistin 10 mg daily. After the 7-day course of intrathecal colistin was completed, repeated cerebrospinal fluid surveillance cultures were obtained and were documented to be sterile. The patient recovered from the infection and was transferred to a step-down unit.

*Department of Pharmacy, and †Division of Infectious Diseases, Brigham and Women's Hospital, Boston, MA.

Address correspondence and reprint requests to Christopher M. Enfanto, PharmD, BCPS, Department of Pharmacy, Brigham and Women's Hospital, Tower L2, 75 Francis St, Boston, MA 02115. E-mail:

Central nervous system (CNS) infections can result in serious neurological disability and death.1 Aggressive treatment with appropriate antibiotics is imperative to eradicate the offending pathogen. Like in all infections, delivering the antibiotic medication to the site of infection is the main goal of therapy. Effective treatment of CNS infections requires understanding factors that influence both drug penetration to the CNS and optimization of the pharmacodynamic properties of the antibiotic agent chosen.

Antibiotics can be roughly divided into 2 major groups based on their activity of bacterial killing in the body. For example, aminoglycosides and fluoroquinolones are concentration-dependent antibiotics. Their efficacy depends on high peak to minimum inhibitory concentration ratio and prolonged postantibiotic effect. In contrast, the activity of β-lactam antibiotics is dependent on the time their concentrations exceed the minimum inhibitory concentration for the infecting organism. Further study is required to determine whether these pharmacodynamic properties apply in cerebrospinal fluid (CSF) because most studies concerning meningitis and ventriculitis are not specifically designed to compare the concentration dependency versus time dependency of the agent being investigated.2

Treatment of CNS infections will often include a combination of intravenous antibiotics and/or shunt removal if that is the source of infection. In addition, aminoglycosides and vancomycin are sometimes administered directly into the CSF in this patient population because intravenous doses needed to reach adequate CSF levels could potentially be toxic.2 Antibiotics can be delivered directly into CSF either by lumbar intrathecal administration or by intraventricular injection. In comparison with intravenously administered antibiotic therapy, intrathecal and intraventricular administration can result in higher peak CSF concentrations and maintenance of therapeutic CSF concentrations for 24 hours.2

The most common organism responsible for CNS infections is coagulase-negative staphylococcus, which accounts for one third to one half of infections in adults and children.3-5 One third of reported shunt infection cases are caused by Staphylococcus aureus, with the remaining infections caused by a variety of organisms, including aerobic gram-negative bacilli. Both S. aureus and aerobic gram-negative bacilli, such as Pseudomonas aeruginosa, seem to be more frequent pathogens when the infection arises from surgical site infections.5 When shunt infections arise from nosocomial acquisition, gram-negative bacilli and multiple drug-resistant organisms are the more common pathogens.5

Multiple drug-resistant gram-negative infections caused by P. aeruginosa, Acinetobacter baumannii, and extended-spectrum β-lactamase (ESBL) Klebsiella pneumoniae have been reported in intensive care units throughout the world.6 The emergence of these resistant strains is a major concern worldwide because therapeutic treatment options are limited.Because of limited treatment options for multiple drug-resistant organisms, clinicians must optimize antibiotic choices based on susceptibility data and appropriate dosing. Once the appropriate antibiotic is chosen, obtaining adequate drug delivery to the site of infection is another key factor to eradicating the infection.

Extended-spectrum β-lactamases have emerged as an important mechanism of resistance in gram-negative bacteria. To make matters worse, ESBL-producing organisms often possess resistance to other antibiotic groups, such as aminoglycosides and fluoroquinolones, leaving clinicians with very few therapeutic options.7 Delay in appropriate therapy for infections with ESBL producers not only prolongs hospital stay but is associated with increased mortality.7

Colistin is a polymyxin antibiotic that has activity against gram-negative bacteria including many drug-resistant strains such as Acinetobacter species, P. aeruginosa, Klebsiella species (ESBL-producing strains), and Enterobacter species. The use of intravenous colistin is avoided because of reports of common and serious nephrotoxicity and neurotoxicity associated with the drug.8 Traditionally, intravenous use of this agent is avoided and used as more of a last-line option because of the adverse effect profile.

Back to Top | Article Outline


A 24-year-old man was involved in a high-speed motor vehicle accident on May 2, 2005. He was ejected from the vehicle, found seizing, and intubated by ground crew at the site. The patient was airlifted to Brigham and Women's Hospital. The patient's Glasgow Coma Scale was 3, and systolic blood pressure was in the 70s. He received intravenous crystalloid resuscitation during the flight. The patient had a subdural hematoma and underwent emergent craniotomy and craniectomy. He also sustained a right femur fracture and underwent placement of right femoral nail after initial external fixation of right femur fracture. The fixation was complicated by an infected fluid collection positive for methicillin-resistant S. aureus on May 15, 2005; and the patient was started on vancomycin 1 g every 12 hours. The patient had persistent fevers during his hospital course. Workup included multiple cultures of blood, sputum, and urine. Chest radiographs were obtained, and all came back negative for signs of infection. No sources of fevers were found, although he did receive an empirical 14-day course of levofloxacin for suspected sinusitis. Patient was recovering from his injuries and following simple commands before his discharge to a rehabilitation hospital on May 28, 2005. A 28-day course of vancomycin 1 g every 12 hours was to be completed at the rehabilitation hospital.

On June 15, 2005, the patient was readmitted back to Brigham and Women's Hospital with altered mental status, hydrocephalus, high fevers (maximum temperature of 104°F) and leucocytosis [white blood cell (WBC) count of 10,000 cells/μL]. Initial computed tomography with contrast revealed no evidence of intracranial infection, the electroencephalogram showed no abnormalities, and the CSF profile did not indicate meningitis. The cause of his fevers was unclear at this point. Vancomycin 1 g every 12 hours, ceftazidime 2 g every 8 hours, and metronidazole 500 mg every 8 hours were started empirically; and the patient received ventilatory management via his tracheostomy. By hospital day 2, the patient achieved defervescence. Cerebrospinal fluid, blood, urine, and line cultures were negative; and the right leg showed no evidence of postoperative infection.

On June 19, 2005, an external ventricular drainage (EVD) device was placed for increased ventricle size and draining from the incision site. On May 23, 2005, four days after placement of the EVD, the patient was afebrile and showed neurological improvement, although not to baseline. The peripheral WBC count was 11,000 cells/μL. At that time, the decision was made to replace the bone flap and place a ventriculoperitoneal shunt (VPS).

On June 24, 2005, patient underwent bone flap replacement and VPS placement without perioperative complications. Patient was still receiving vancomycin 1 g every 12 hours, ceftazidime 2 g every 8 hours, and metronidazole 500 mg every 8 hours at the time of surgery. Patient was stable after procedure. However, the WBC count increased to 18,000 cells/μL on postoperative day 1 (June 25, 2006); and the patient subsequently began to develop fevers with temperatures of 104°F. Perioperative cultures grew rare coagulase-negative staphylococcus. Repeated CSF sampling on June 26, 2005, grew pan-resistant ESBL-producing K. pneumoniae as well as rare coagulase-negative staphylococcus (refer to Table 1). The ESBL was detected on Vitek and confirmed using the double-disk test. The organism was resistant to all conventional antibiotics and sensitive only to imipenem/meropenem and nitrofurantoin. Based on cultures and sensitivities, the patient was started on intravenous meropenem 2 g every 8 hours on June 26, 2006, and continued on vancomycin 1 g every 12 hours.



On June 28, 2005, magnetic resonance imaging revealed a large collection of fluid around the bone flap. On June 28, 2005, two days after the initiation of the intravenous meropenem, the patient failed to achieve defervescence and clinically improve. On the same day, the decision was made to add intrathecal colistin 10 mg daily. The intrathecal colistin was scheduled for a 7-day course, and vancomycin 1 g every 12 hours and meropenem 2 g every 8 hours were scheduled for 28-day courses. On June 29, 2005, the bone flap and VPS were removed; and an EVD was placed.

No adverse reactions related to the colistin were observed throughout the course of therapy. On July 4, 2005, after the 7-day course of intrathecal colistin 10 mg daily was completed, 8 CSF surveillance cultures were obtained and were documented to be sterile (refer to Table 1). On July 12, 2005, the patient was afebrile and stable from a neurological perspective. On July 14, 2005, the patient was transferred to the step-down unit.

Back to Top | Article Outline


The most common adverse effects of intravenous colistin therapy are nephrotoxicity and neurotoxicity. Nephrotoxicity mainly includes acute tubular necrosis manifested as decreased creatinine clearance and increased serum urea and creatinine levels. Early trials reported high incidence of nephrotoxicity at around 20% of patients.9-11 However, recent data report an incidence of nephrotoxicity of about 14% to 18% of patients.10,11 Possible explanations of this decrease in incidence between old and new data may include improvements in supportive treatments offered to critically ill patients. It is recommended that the dosing interval of colistin should be adjusted for renal insufficiency.12

Neurological toxicity is associated with dizziness, weakness, facial and peripheral paresthesia, vertigo, visual disturbances, confusion, ataxia, and neuromuscular blockade, which can lead to respiratory failure or apnea. The incidence of colistin-associated neurotoxicity reported in earlier literature was about 7%, with paresthesias constituting the main neurotoxic adverse event.9 Both renal and neurological toxicities are considered to be dose dependent and usually reversible after early discontinuation of therapy with the drug.

The administration of intrathecal colistin in this patient cleared the multi-drug-resistant ESBL-producing K. pneumoniae and produced no adverse effects that were noted by laboratory data or by the clinicians.

Previously reported cases state using intrathecal colistin at a 5-mg dose for 14 to 21 days for ventriculitis and shunt infections.13-16 In these previously reported cases, intrathecal colistin was used for resistant P. aeruginosa and carbapenem-resistant A. baumannii. In this case, it was decided to use the 10 mg dose of intrathecal colistin for a shorter duration of therapy. This decision was made based on the pharmacodynamic properties of colistin. Colistin exhibits concentration-dependent killing. When antibiotics that exhibit concentration-dependent killing are given at higher doses, they provide more rapid and complete killing by achieving a greater peak level. In this case, the CSF cultures were negative after 4 days of intrathecal colistin therapy; and the clinicians chose to discontinue therapy after 7 days of treatment.

Back to Top | Article Outline


To our knowledge, this is the first report of successful use of intrathecal colistin in the treatment of a CNS infection caused by ESBL-producing K. pneumoniae. Published data support the use of intrathecal colistin for treatment of ventriculitis and shunt infections caused by resistant P. aeruginosa and carbapenem-resistant A. baumannii.13-16 The ultimate cure of the infection is not known. There were 3 major interventions that may have led to cure of the infection in this patient either single-handedly or collectively: (1) high-dose meropenem 2 g every 8 hours started on June 26, 2005; (2) intrathecal colistin 10 mg daily started on June 28, 2005; and (3) removal of the bone flap and VPS on June 29, 2006. Based on published literature and the presented case report, the authors conclude that intrathecal colistin can be a viable therapeutic option in patients with ESBL-producing K. pneumoniae in the CSF.

Back to Top | Article Outline


The authors thank Paul M. Szumita, PharmD, BCPS, Clinical Pharmacy Practice Manager of the Brigham and Women's Hospital, for his advice and comments regarding the therapeutic intervention.

Back to Top | Article Outline


1. Kearney B, Aweeka F. The penetration of anti-infectives into the central nervous system. Neurol Clin. 1999;17:883-900.
2. Lustar I, McCracken GH, Friedland I. Antibiotic pharmacodynamics in cerebrospinal fluid. Clin Infect Dis. 1998;27:1117-1129.
3. Davis SE, Levy ML, McComb JG, et al. Does age or other factors influence the incidence of ventriculoperitoneal shunt infections? Pediatr Neurosurg. 1999;30:253-257.
4. George R, Leibrock L, Epstein M. Long-term analysis of cerebrospinal fluid shunt infections. A 25-year experience. J Neurosurg. 1979;51:804-811.
5. Schoenbaum SC, Gardner P, Shillito J. Infections of cerebrospinal fluid shunts: epidemiology, clinical manifestations, and therapy. J Infect Dis. 1975;131:543-552.
6. Berlina D, Llop E, Fort MB, et al. Use of colistin in the treatment of multiple-drug-resistant gram-negative infections. Am J Health Syst Pharm. 2005;62:39-47.
7. Weinbren MJ, Borthwick MA. Rapid detection of extended-spectrum β-lactamase (ESBL)-producing organisms in blood culture. J Antimicrob Chemother. 2005;55:131-132.
8. Falagas M, Kasiakou S. Colistin: the revival of polymyxins for the management of multidrug-resistant gram-negative bacterial infections. Clin Infect Dis. 2005;40:1333-1340.
9. Koch-Weser J, Sidel VW, Federman EB, et al. Adverse effects of sodium colistimethate: manifestations and specific reaction rates during 317 courses of therapy. Ann Intern Med. 1970;72:857-868.
10. Price DJ, Graham DI. Effects of large doses of colistin sulphomethate sodium on renal dysfunction. Br Med J. 1970;4:525-527.
11. Randall RE, Bridi GS, Setter JG, et al. Recovery from colisthemate nephrotoxicity. Ann Intern Med. 1970;73:491-492.
12. Aronoff GR, Berns JS, Brier ME, et al. Drug Prescribing in Renal Failure. Dosing Guidelines for Adults. 4th ed. Philadelphia, PA: American College of Physicians; 1999.
13. Fernandez-Viladrich P, Corbella X, Corral L, et al. Successful treatment of ventriculitis due to carbapenem-resistant Acinetobacter baumannii with intraventricular colistin sulfomethate sodium. Clin Infect Dis. 1999;28:916-917.
14. Vasen W, Desmery P, Di Martineo A. Intrathecal use of colistin. J Clin Microbiol. 2000;38:3523. [letter].
15. Gump WC, Walsh JW. Intrathecal colistin for treatment of highly resistant Pseudomonas ventriculitis: case report and review of the literature. J Neurosurg. 2005;102:915-917.
16. Quinn AL, Parada JP, Belmares J, et al. Intrathecal colistin and sterilization of resistant Pseudomonas aeruginosa shunt infection. Ann Pharmacother. 2005;39:949-952.
© 2007 Lippincott Williams & Wilkins, Inc.