Volume 14, Issue 3 , Pages e216-e219, March 2010
Infections associated with indwelling ventriculostomy catheters in a teaching hospital
Article Outline
- Abstract
- 1. Introduction
- 2. Patients and methods
- 3. Results
- 4. Discussion
- Acknowledgements
- References
- Copyright
Abstract
Background
Ventriculostomy-associated infections are a serious complication of external ventricular drains. The objective of this study was to analyze the clinical features of and risk factors for such infections.
Methods
We retrospectively collected demographic and clinical data on patients with indwelling ventriculostomy catheters hospitalized in a teaching hospital from July 2001 to June 2006, comparing those with and without ventriculostomy-associated infections.
Results
A total of 197 drains (2910 catheter-days) placed in 155 patients were studied. Infections developed in 28 of the 197 (14.2%) drains. The duration from insertion to infection ranged from 7 to 36 days. The cut-off point of duration from insertion to infection was 15.5 days. Re-insertion because of catheter malfunction carried a high risk of infection (p
<0.001). Patients with infections had a longer intensive care unit stay (p=0.001), longer duration of catheterization (p=0.002), and a higher incidence of concurrent sepsis (p=0.018), urinary tract infection (p=0.011) and pneumonia (p=0.004). Gram-negative bacilli were the leading pathogens (84%); Pseudomonas aeruginosa was the most common isolate. Polymicrobial infections occurred later than monomicrobial infections (p=0.003).
Conclusions
Repeated insertion and longer duration of drains are major risk factors for ventriculostomy-associated infections.
Keywords: Risk, Ventriculostomy, Indwelling catheter, Pseudomonas aeruginosa, Central nervous system infection
1. Introduction
Indwelling ventriculostomy catheters are an essential component of the management of many neurosurgical patients. The catheters are used to monitor intracranial pressure and to drain cerebrospinal fluid (CSF) in patients with traumatic brain injury and in obstructive hydrocephalus associated with tumors or hemorrhage.1, 2, 3 The use of these drains is, however, associated with a risk of life-threatening infection. The reported incidence of ventriculostomy-associated infections in various series ranges from 0% to 22%.4, 5, 6, 7, 8 Factors associated with an increased risk of infection include prolonged catheterization, underlying disease, concurrent systemic infection, catheter leakage, changing the catheter, and operative techniques.4, 5, 6, 8, 9, 10, 11 In one study, the risk of infection was found to significantly increase when drains were routinely changed, but duration of catheterization was not independently associated with an increased incidence of infection.12 Other investigators, however, have found an association between duration of catheterization and infection, recommending removal of catheters as soon as possible.7, 13 The evidence is thus conflicting,14 and is inconclusive due to the controversies and discrepancies between the various reports. Hence, we decided to review our experience with indwelling ventriculostomy drains, evaluating the risk factors for infection, particularly the duration of catheterization.
2. Patients and methods
We retrospectively reviewed the records of all patients with external ventricular drains hospitalized in Taitung Mackay Memorial Hospital between July 2001 and June 2006. The decision to place a drain was made by the neurosurgical team responsible for the management of patients with severe traumatic brain injuries, tumors, intracerebral bleeding, and other disorders.
A subcutaneous tunneled catheter was inserted from the Kocher's point in the operation room, after preparation with standard sterile techniques. The distal end of the catheter was advanced approximately 5–7cm until CSF was obtained. The catheter was then tunneled in a retrograde fashion to a distant skin exit site and was connected via a three-way stopcock to an external pressure transducer and CSF drainage system. Our patients were given prophylactic antibiotic therapy with a first-generation cephalosporin such as cefazolin 1g before surgery, followed by 1g every 6hours for 1–2 days. Thereafter, most patients were managed in the surgical intensive care unit (ICU), allowing constant monitoring of intracranial pressure and drainage of the CSF as required.
Drains were generally removed once the acute neurosurgical condition had stabilized and intracranial pressure monitoring was no longer needed. In some cases where prolonged drainage of CSF was required, the drain was left in place even after discharge from the ICU. Catheter obstruction was managed by flushing the drain with 2ml of 0.9% NaCl. There was no policy for routine catheter exchange, most replacements being made because of obstruction, accidental dislodgment or leakage, and infection.
Routine CSF cultures were not performed but were ordered if there was a clinical suspicion of meningitis or ventriculitis based on a new fever, peripheral leukocytosis, neurologic deterioration, or a change in the appearance of the CSF. Ventriculostomy-associated infection was defined as culture of a pathogen from the CSF obtained either from the catheter or by lumbar puncture. An infection was considered as occurring on the day the CSF sample was obtained. However, infections documented before ventriculostomy or up to the first day after catheter insertion were not considered to be catheter-related and were not included in this analysis.
Data collected from the records included demographic information, neurosurgical indication for ventriculostomy, acute physiology and chronic health evaluation II (APACHE II) score, length of ICU and of hospital stay, and mortality. Factors related to catheterization that were reviewed included use of prophylactic systemic antibiotics at the time of catheter placement, presence of a catheter leakage or obstruction, the number of catheters placed, and duration of catheterization, defined as the number of days from insertion until an infection was documented or until the catheter was removed if no infection occurred.
2.1. Statistical analysis
Continuous variables are expressed as mean±standard deviation (SD) and were compared using the Student's t-test. Categorical variables are expressed as absolute and relative frequencies and were compared using the Chi-square or Fisher's exact test. A p-value of ≤0.05 was considered significant. The time until infection was plotted using Kaplan–Meier curves. A receiver operating characteristic (ROC) curve was generated to compare the duration of catheterization with and without infection to detect the cut-off point for a catheter infection. Statistical analyses were performed using the Statistical Package for the Social Sciences (SPSS) version 12.0.
3. Results
Over the study period, 197 drains placed in 155 patients for a total of 2910 catheter-days were included in the analysis; 33 drains placed after infection were excluded from the analysis. The most common indication for ventriculostomy was hydrocephalus, followed by spontaneous and traumatic intracerebral hemorrhage (Table 1). The average duration of catheterization was 15.5±10.0 days (range 1–61 days). One hundred ten patients (71%) had only one catheter and 45 (29%) had two or more (up to six). The mean number of catheters per patient was thus 1.5±0.9. The most common reason for re-insertion was blocked CSF drainage. Men were more likely to require re-insertion than women (34/98, 35% vs. 11/57, 19%; p=0.042), but there was no significant difference in catheter malfunction in patients with or without intracranial hemorrhage (p=0.44).
Table 1. Clinical features of 155 patients with external ventricular drains.
| Patients, n (%) | |
|---|---|
| Gender | |
| Male | 98 (63.2) |
| Female | 57 (36.8) |
| Indication of ventriculostomy | |
| Hydrocephalus | 84 (54.2) |
| Traumatic ICH | 10 (6.5) |
| Spontaneous ICH | 47 (30.3) |
| Traumatic SAH | 4 (2.6) |
| Non-traumatic SAH | 1 (0.6) |
| Traumatic SDH | 3 (1.9) |
| Non-traumatic SDH | 3 (1.9) |
| Brain tumor | 3 (1.9) |
| Number of insertions | |
| 1 | 110 (71.0) |
| 2 | 26 (16.8) |
| 3 | 12 (7.7) |
| 4 | 4 (2.6) |
| 5 | 2 (1.3) |
| 6 | 1 (0.6) |
A total of 28 infections developed (14.2% of 197 catheters or 9.6 per 1000 catheter-days). None occurred in patients with subarachnoid or subdural hemorrhage. The mean duration from catheter insertion to infection ranged from 7 to 36 days. The infection rate of indwelling catheters according to duration of catheterization is shown in Table 2. Figure 1 shows the Kaplan–Meier analysis of re-insertion and first insertion indwelling catheters. No infections occurred in the first 6 days or later than 36 days after catheter insertion, and the mean duration of catheterization was significantly longer in those with versus those without infection (20.2±7.8 days vs.14.7±10.1 days, p=0.002). The cut-off point of duration from insertion to infection was 15.5 days (cut-off level ranged from 12.5 to 17.5 days). Significantly more infections occurred in patients who had catheter re-insertions (mean catheter number with vs. without infection 2.1±0.9 vs. 1.3±0.8; p<0.001) (Table 3). Infections occurred in only 4.5% of catheters that did not have to be changed. The incidence was significantly higher in re-inserted catheters and rose with each repeated insertion – 42.3%, 58.3%, 75.0%, and 100% for the second, third, fourth, and fifth insertion, respectively. Second and third drains in previously uninfected patients were more likely to become infected than first drains.
Table 2. Infection rate of 197 external ventricular drains by week.
| Duration (days) | No. of infected catheters | No. of inserted catheters | Infection rate (%) |
|---|---|---|---|
| 1–7 | 1 | 47 | 2.1 |
| 8–14 | 7 | 48 | 14.6 |
| 15–21 | 9 | 57 | 15.8 |
| 22–28 | 6 | 29 | 20.7 |
| >28 | 5 | 16 | 31.3 |
| Total | 28 | 197 | 14.2 |
Figure 1.
Kaplan–Meier analysis of the cumulative survival of indwelling catheters. The dashed line (---) represents re-insertion catheters and the solid line (—) represents first insertion catheters. Catheters removed because they were no longer needed before the infection occurred were treated as censored and recorded as +. No infections occurred before day 7 or after day 36. Among the re-insertion catheters, the number of infected catheters rose sharply after the second week.
Table 3. Comparison of infected and uninfected catheters with regard to patient characteristics, management, and catheter-related factors.
| Variable | Infection (n=28) | No infection (n=169) | p-Valuea |
|---|---|---|---|
| Patient age (years) | 60.8±18.3 | 57.7±20.0 | 0.42 |
| APACHE II score | 19.6±15.0 | 20.5±5.8 | 0.47 |
| Number of catheters inserted | 2.1±0.9 | 1.3±0.8 | <0.001 |
| ICU stay (days) | 31.1±23.7 | 14.4±14.6 | 0.001 |
| Catheter duration (days) | 20.2±7.8 | 14.7±10.1 | 0.002 |
| Male patients | 20 | 107 | 0.41 |
| Prophylactic antibiotics <24 hours | 18 | 115 | 0.69 |
| Catheter-related factors | |||
| Catheter obstruction | 12 | 81 | 0.62 |
| Catheter leakage | 2 | 1 | 0.009 |
| Repeat insertion | 23 | 64 | <0.001 |
| Underlying factors | |||
| Trauma | 3 | 15 | 0.75 |
| Craniotomy | 2 | 23 | 0.34 |
| Intraventricular hemorrhage | 11 | 45 | 0.17 |
| ICU stay before drain insertion | 9 | 21 | 0.007 |
| Urinary tract infection | 12 | 35 | 0.011 |
| Pneumonia | 18 | 60 | 0.004 |
| Bacteremia | 7 | 16 | 0.018 |
aStudent's t-test, Fisher's exact test, or Chi-square test. |
Prophylactic antibiotics were prescribed for all patients, but using them for a longer duration was not associated with the incidence of infection (p=0.69). Catheters with compared to those without infections were associated with a higher incidence of concurrent sepsis (25% vs. 9.5%, p=0.018), urinary tract infection (42.9% vs. 20.7%, p=0.011), and pneumonia (64.3% vs. 35.5%, p=0.004). Catheter infections were more likely associated with a longer ICU stay (31.1±23.7 days vs. 14.4±14.6 days, p=0.001), but age, APACHE II score, catheter site, and presence of catheter obstruction were not significantly different.
In the 28 episodes of infection, 37 different organisms were isolated. The infections were monomicrobial in 19 and polymicrobial in nine (Table 4). Polymicrobial infections occurred significantly later than monomicrobial infections (19.6±4.9 days vs. 12.9±5.0 days, p=0.003). Gram-negative bacilli were the most common isolates (31/37, 84%), with Pseudomonas aeruginosa found in nine cases and Acinetobacter baumannii and Stenotrophomonas maltophilia found in six cases each. Among Gram-positive infections, the two Staphylococcus aureus isolates were methicillin-resistant.
Table 4. Microorganisms isolated from 28 infected ventriculostomy catheters.
| Infections | Number |
|---|---|
| Monomicrobial infections | |
| Gram-positive | 2 |
| Methicillin-resistant Staphylococcus aureus | 2 |
| Gram-negative | 17 |
| Pseudomonas aeruginosa | 6 |
| Acinetobacter baumannii | 4 |
| Stenotrophomonas maltophilia | 3 |
| Enterobacter aerogenes | 2 |
| Klebsiella pneumoniae | 1 |
| Proteus mirabilis | 1 |
| Polymicrobial infections | 9 |
| Coagulase-negative Staphylococcus+Stenotrophomonas maltophilia | 2 |
| Acinetobacter baumannii+Pseudomonas aeruginosa | 1 |
| Acinetobacter baumannii+Enterococcus | 1 |
| Chromobacterium spp+Empedobacter brevis | 1 |
| Enterococcus+Serratia marcescens | 1 |
| Klebsiella pneumoniae+Stenotrophomonas maltophilia | 1 |
| Pseudomonas aeruginosa+Citrobacter diversus | 1 |
| Pseudomonas aeruginosa+Flavobacterium spp | 1 |
4. Discussion
The main limitation of ventriculostomy is the high risk of infection, especially with prolonged drainage. Although the infection rate in our single center study was relatively high at 14.2%, this is comparable with the infection rates found in similar studies.5, 8, 11 The higher incidence might be because of a longer duration of drainage and repeated insertion of drains because of malfunction. However, comparing studies is difficult because of varying case definitions. We defined an infection as a positive CSF culture obtained from the drain or via lumbar puncture, a definition also used by other authors.4, 5, 8, 9 Since CSF is obtained through the catheter, it is possible that some positive cultures represent colonization of the drain rather than true infection. Nonetheless, we treated these patients as ventriculostomy-associated infection because all of them had clinically relevant factors for a suspicion of meningitis.
Risk factors for ventriculostomy-associated infections include those that promote bacterial growth or bacterial access to the CSF.4 In a study in which 58 patients were considered to have infection (with 97 positive cultures), multivariate logistic regression analysis indicated that subarachnoid hemorrhage was an independent risk factor associated with infection.15 We failed to find such an association in our series, but this may be because relatively few of our patients had subarachnoid hemorrhage. The factors identified in our study as being associated with an increased risk of infection were having more than one catheter inserted (the risk increasing with each subsequent insertion), a longer duration of catheterization, and the presence of a catheter leakage. Clark et al. found a significantly higher incidence of infection in second as compared to first drains, although the incidence did not increase that much with a third catheter.16 A catheter leakage presumably provides a route by which microorganisms can migrate into the central nervous system. Our finding in this regard is similar to that of other studies.8, 17
The duration of drainage has not uniformly been found to be an independent predictor of infection; it is primarily multiple insertions that seem to increase the risk.12 In an extensive review of the literature, Lozier et al. found considerable variation in study results, but they did note an increasing incidence of infection over the first 10 days of catheterization. The data available were insufficient to assess the risk beyond 10 days.4 A significantly higher infection rate in patients with a mean catheterization duration of 11 days or longer has been reported.9, 18 No patients in our series developed infection in the first 6 days after insertion, but the incidence of infection increased significantly after 2 weeks. This contrasts with the findings of Paramore and Turner who found a progressive increase in the daily infection rate, with the maximum risk on day 6.19 The lower rate of infection in the first week at our institution may reflect catheter maintenance techniques and the use of prophylactic antibiotics. Some investigators have reported a substantially higher incidence of late infections in a population that continued to be at risk.9, 20, 21 The results from our patients suggest an increasing risk over time, especially after 2 weeks.
In our hospital, the policy in patients requiring prolonged drainage is to leave uninfected ventriculostomy catheters in place instead of routinely replacing them after a certain time. A retrospective study by Holloway et al. failed to demonstrate any reduction in the cumulative risk of infections when a routine catheter exchange was performed every 5 days,7 and there is no support in the current literature for routine catheter exchange.22 Certainly drains should be removed as soon as possible when they are no longer needed. A new catheter should be placed only if continued drainage is necessary and there is obstruction or infection in the one already in place.
Over four-fifths of the bacterial isolates in our study were Gram-negative organisms, a result consistent with that of other investigations.6, 11, 12 Traditionally, Gram-positive organisms have been the most common pathogens causing ventriculostomy-associated infections.5, 8, 19, 23 It has been suggested that trauma resulting in open head injury might predispose the patient to Gram-negative meningitis,24 but only three infections in our series were associated with trauma. The reasons for the preponderance of Gram-negative infections in our study might be prolonged ICU stay and local flora in the surgical ICU. It is hypothesized that pathogens colonizing patients gain access to the CSF by retrograde migration along the catheter.6 The use of prophylactic antibiotics has been shown not to significantly alter infection rates.11, 25 The association of prolonged ICU stay and colonization with Gram-negative organisms is reasonable, as these organisms are prevalent in the ICU environment. This effect may predispose to Gram-negative infections.
In our study, about one-fourth of patients with catheter infections had bacteremia and more than half of them had pneumonia, higher than previously reported figures.7, 26 Clark et al. found that patients with infected catheters experienced significantly higher rates of systemic infections than those who did not have ventriculostomy-associated infections.16 Patients with a prolonged ICU stay are often intubated, have central venous lines, indwelling Foley catheters, and so on. They are therefore at high risk of infection. Strict infection control and removal of unnecessary catheters is certainly warranted to reduce the risk of nosocomial infection.26
Our study has several limitations, including the single center, retrospective study design. The relatively small sample size makes it difficult to draw firm conclusions regarding associations with particular underlying conditions. Nevertheless, our study shows that the re-insertion of catheters and catheters placed for more than 15.5 days increase the risk of ventriculostomy-associated infection. The shift toward Gram-negative organisms causing such infections is important to recognize, as this will influence the choice of empiric treatment while awaiting the results of CSF cultures. While our study was not designed to assess infection prevention measures, it is reasonable to urge meticulous care in inserting and maintaining ventriculostomy catheters, using strict aseptic techniques.
Acknowledgements
We thank the staff of the intensive care unit and the infection control team of Mackay Memorial Hospital, Taitung Branch for their assistance with data collection and Miss Fang-Ju Sun of the Department of Medical Research of Mackay Memorial Hospital for her guidance in the statistical analysis.
Conflict of interest statement: We all work at Mackay Memorial Hospital. There are no potential conflicts of interest among us. The patients were diagnosed and treated at our hospital. This study did not receive any grants.
References
- . Management problems in acute hydrocephalus after subarachnoid hemorrhage. Stroke. 1989;20:747–753
- . External ventriculostomy devices EVDs. J Neurosci Nurs. 1998;30:342–344
- . Surgical treatment of spontaneous cerebellar hemorrhage. Surg Neurol. 1985;23:555–558
- . Ventriculostomy-related infections: a critical review of the literature. Neurosurgery. 2002;51:170–181
- Ventriculostomy-related infections. A prospective epidemiologic study. N Engl J Med. 1984;310:553–559
- . Ventriculitis complicating use of intraventricular catheters in adult neurosurgical patients. Clin Infect Dis. 2001;33:2028–2033
- Ventriculostomy infections: the effect of monitoring duration and catheter exchange in 584 patients. J Neurosurg. 1996;85:419–424
- . Bacterial meningitis caused by the use of ventricular or lumbar cerebrospinal fluid catheters. J Neurosurg. 2005;102:229–234
- . Risk factors for infections related to external ventricular drainage. Acta Neurochir (Wien). 2008;150:209–214
- . Comparison of rates of infection of two methods of emergency ventricular drainage. J Neurol Neurosurg Psychiatry. 1995;58:444–446
- Ventriculostomy-associated infections: incidence and risk factors. Am J Infect Control. 2005;33:137–143
- . External ventricular drain infections are independent of drain duration: an argument against elective revision. J Neurosurg. 2007;106:378–383
- . Intracranial pressure monitors. Epidemiologic study of risk factors and infections. Am J Med. 1986;80:369–376
- . Nosocomial infections in neurocritical care. Curr Neurol Neurosci Rep. 2006;6:525–530
- . Ventriculostomy-related infections in critically ill patients: a 6-year experience. J Neurosurg. 2005;103:468–472
- . Complications of intracranial pressure monitoring in trauma patients. Neurosurgery. 1989;25:20–24
- . Continuous-pressure controlled, external ventricular drainage for treatment of acute hydrocephalus—evaluation of risk factors. Neurosurgery. 1992;31:898–903
- . Bacterial ventriculitis and duration of ventriculostomy catheter insertion. J Neurosci Nurs. 1993;25:158–164
- . Relative risks of ventriculostomy infection and morbidity. Acta Neurochir (Wien). 1994;127:79–84
- . Duration of intracranial pressure monitoring does not predict daily risk of infectious complications. Neurosurgery. 1993;33:424–430
- . Infected ventriculostomy: bacteriology and treatment. Acta Neurochir (Wien). 1989;100:67–69
- . Failure of regular external ventricular drain exchange to reduce cerebrospinal fluid infection: result of a randomized controlled trial. J Neurol Neurosurg Psychiatry. 2002;73:759–761
- . Ventriculostomy-related infections—an epidemiological study. Acta Neurochir (Wien). 1986;83:20–23
- . Gram-negative bacillary meningitis in neurosurgical patients. J Neurosurg. 1983;59:634–641
- . The efficacy and cost of prophylactic and perioprocedural antibiotics in patients with external ventricular drains. Neurosurgery. 2000;47:1124–1127
- . Hospital-acquired infection surveillance in a neurosurgical intensive care unit. J Hosp Infect. 2006;64:23–29
PII: S1201-9712(09)00196-9
doi:10.1016/j.ijid.2009.04.006
Crown Copyright © 2009. Published by Elsevier Inc. All rights reserved.
Volume 14, Issue 3 , Pages e216-e219, March 2010
