Abstract
Background
The relationship between first and subsequent cerebrospinal fluid (CSF) shunt infections is poorly understood. By understanding the factors associated with increased risk of reinfection, researchers may provide optimal treatment strategies at the time of first infection. The objective of this study was to describe and compare children with and without CSF shunt reinfection.
Methods
A retrospective cohort study was performed among 118 children who underwent initial CSF shunt placement and developed first CSF shunt infection. The primary outcome variable was CSF shunt reinfection. Patient risk factors and medical and surgical management of initial CSF shunt placement and first CSF shunt infection were compared between children with and without reinfection.
Results
Of 118 children with first infection, 31 (26%) developed a reinfection during the study period (overall median follow-up, 2096 days). Factors associated with reinfection in this cohort included ventriculoatrial or complex shunt at initial CSF shunt placement, complications after first CSF shunt infection, and intermittent negative CSF cultures.
Conclusions
Few patient or treatment factors were associated with reinfection. Factors associated with difficult-to-treat first CSF shunt infection, including complications after first CSF shunt infection and intermittent negative CSF cultures, were associated with reinfection. Clinicians who treat patients with unusual CSF shunts or more difficult first infections should have a high index of suspicion for reinfection after treatment is completed.
Keywords: children, infection, shunt
BACKGROUND
Infection is frequently seen in children with hydrocephalus who undergo cerebrospinal fluid (CSF) shunt placement. Reported infection rates range from 6.5% to 23.5% per patient [1–7] and occur in approximately 11% of children within 24 months of uncomplicated first CSF shunt placement [8]. Risk factors for CSF shunt infection include young age at the time of initial CSF shunt placement as well as subsequent CSF shunt revision procedures [9]. The latter have been associated with intraventricular hemorrhage secondary to prematurity [10]. Most shunt infections are reported to occur within 1 month of CSF shunt procedures [11]. A high percentage of pathogens originate from skin and include Staphylococcus epidermidis and Staphylococcus aureus; for ventriculoperitoneal shunts pathogens orginate from the enteric system with gram-negative bacilli [12, 13]. Thus, the prevailing hypothesis for CSF shunt infection is that the infecting organism is introduced during surgical manipulation of the site.
Two of the largest observational studies of treatment of CSF shunt infection to date observed reinfection rates of 26% (18 of 70) and 20% (10 of 51), respectively [14, 15]. Investigators have noted considerable variation in treatment duration and methods of medical and surgical management of shunt infections that might explain reinfections [11, 12, 16, 17]. Development of a second CSF shunt infection is a particularly critical event because it is associated with recurrent reinfections in a small number of children [18]. Therefore, we sought a better understanding of the characteristics associated with reinfection after first CSF shunt infection, because these may provide information for optimal treatment strategies at the time of first infection.
The objective of this paper was to describe and compare children with and without CSF shunt reinfection.
METHODS
We conducted a retrospective cohort study among children who received care at Primary Children's Medical Center (PCMC), a tertiary care children's hospital serving the Intermountain West (Utah, Nevada, Idaho, Colorado). The study was reviewed and approved by the Institutional Review Boards at the University of Utah and Seattle Children's Research Institute, and the research did not require informed consent.
The cohort was derived from 579 children under 18 years of age who underwent initial CSF shunt placement with a discharge date between January 1, 1997 and October 12, 2006 at PCMC, as described previously [9, 10]. Of these, 118 subsequently developed CSF shunt infection before December 12, 2006, which was confirmed by medical record review. Data from each neurosurgical admission for each cohort member up until the time of first CSF shunt infection were collected using Intermountain Healthcare's database and chart review [9, 10] (Figure 1).
Figure 1.
Data collected for each patient in the study cohort. Abbreviation: CSF, cerebrospinal fluid.
First CSF shunt infection and reinfection was defined by 1 or more of the following criteria: (1) presence of bacteria in a Gram stain or culture of CSF, wound swab, or pseudocyst fluid; (2) documentation of visible hardware; or (3) abdominal pseudocyst. In children with a ventriculoatrial shunt in place, presence of bacteria in a blood culture was also considered a CSF shunt infection [9, 19]. Organisms that grew in broth only were considered to be infections. Visible hardware documented in history but not mentioned in physical exam met infection criteria. For children who had first shunt infection before December 31, 2006, an additional chart review was performed to ascertain the outcome variable, CSF shunt reinfection, before June 28, 2010 [18].
Predictor variables of interest include patient risk factors and surgical management at the time of initial shunt placement and intervening revision surgeries as previously described [9, 10, 20]. We characterized prematurity using International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) codes for preterm by either birth weight or gestational age under 36 weeks; for seizure, we used either ICD-9-CM codes for convulsions or documentation in the medical record.
We also considered the following factors during the treatment of first CSF shunt infection: chronological age; surgical approach to infection treatment [17]; bedside location of first infection surgery; duration of time from first positive culture to first infection surgery; duration of time between infection surgeries (eg, time from externalization or full removal/external ventricular drain [EVD] placement to reinternalization or new shunt placement); duration of time from first negative culture to final infection surgery; complications after first shunt infection; and distal shunt type after first shunt infection (ie, peritoneal, atrial, or complex, meaning multiple shunts placed or any single shunt with multiple catheters joined together).
We further characterized the use and duration of intravenous (IV) antibiotics between infection surgeries in detail, after a review by our infectious disease expert (E. A. T.). Concordance of IV antibiotic use was assigned when the organism(s) grown in culture were sensitive to the antibiotic. Intravenous antibiotic use was further characterized as appropriate if its dose was appropriate for weight and, where applicable, it had adequate central nervous system (CNS) penetration. Because guidelines for most organisms as to when aminoglycoside synergy should be used are unclear and CNS penetration is poor, we characterized aminoglycoside treatment as concordant but not appropriate when the organism was susceptible. If a susceptible aminoglycoside was used as primary therapy intrathecally, it was considered concordant and appropriate. Because aminoglycoside synergy is recommended in severe enterococcal infection [21] or bacteria with intermediate resistance to the primary antibiotic, these circumstances were also considered concordant and appropriate. If only a Gram stain or visible hardware was seen, we did not characterize concordance. When susceptibilities were not provided, they were anticipated based on the organism grown. Broad-spectrum antibiotic use was defined as use of an antibiotic that was more broad than necessary given the organism's susceptibility pattern and CNS penetration. For example, for gram-positive organisms, the use of vancomycin was considered broad spectrum if the organism was susceptible to oxacillin. If anaerobic coverage was added when it was not necessary based on organism recovered, this antibiotic coverage was also considered broad spectrum. When infections were polymicrobial with bacteria of different classes, antibiotic use was characterized as broad spectrum. Gentamicin use is widespread, and we considered its use as broad spectrum unless it was used in conjunction with severe Enterococcus infection. Use of intrathecal antibiotics was handled as a dichotomous variable. Use of rifampin was considered synergistic and was not incorporated further into consideration of broad spectrum or concordant antibiotics as susceptibility data were not available.
Finally, we examined diagnostic and microbiological factors in CSF shunt infection [18]. We also considered the following factors: specific infection criteria met (see outcome definition); organism(s) recovered in CSF, blood, and wound culture(s); presence of ventriculitis (new organism grown in CSF culture) and bacteremia (new organism grown in blood culture); polymicrobial infection (defined as growth of more than 1 organism from 1 or more sources); the presence of intermittent negative CSF cultures (defined by positive CSF cultures clearing and then returning over the course of treatment); duration of positive CSF cultures; and, where available within 48 hours of infection diagnosis, first CSF shunt infection studies including Gram stain, white and red blood cell counts, glucose, and protein.
Children with first CSF shunt infection were described overall and by subsequent reinfection status using frequency and proportion for categorical variables; and mean and standard deviation, or median and interquartile range (IQR), were used for continuous variables. Risk factors for reinfection were assessed in separate unadjusted Cox proportional hazard models to ascertain risk and account for censoring. Risk factors that demonstrated statistically significant hazard ratios in unadjusted analyses were subsequently entered into a Cox proportional hazards model. No variables were placed into the multivariable model a priori. A stepwise approach was used for model selection, with entry and retention criteria of α = 0.05. To maximize the sample size on which the multivariable models were based, CSF laboratory tests such as white blood cell count and protein were excluded from model selection due to missing data. However, we also performed a sensitivity analysis in which only 80 children with complete CSF laboratory data were included. Results are presented as adjusted hazard ratios (HRs) with 95% confidence intervals (CIs). All statistical analyses were performed using SAS (version 9.2; SAS Institute, Cary, NC).
RESULTS
Of the 118 children with first CSF shunt infection in the cohort, 31 (26%) developed a reinfection during the study period. The median time between infections for the 31 children with reinfection was 66 days (IQR, 27–789 days). The overall median duration of follow-up was 2096 days (IQR, 252–3246 days), and median duration of follow-up for the group without reinfection was 2744 days (IQR, 1729–3493 days). The characteristics of the children in this cohort at the time of initial shunt placement are shown in Table 1. Children with first infection were generally under 6 months of age (70%), male (62%), and non-Latino white (80%). Treatment characteristics in this cohort are shown in Table 2. During first CSF shunt infection, most children in the cohort underwent treatment with concordant (93%), appropriate (83%), and broad-spectrum (92%) IV antibiotics for a median of 14 days (IQR, 11–21 days); as well as full shunt removal and EVD placement followed by new shunt placement (90%). Diagnostic and microbiological factors for infections in the cohort are shown in Table 3. Most infections were diagnosed by CSF culture (89%); organisms were predominantly coagulase-negative Staphylococcus (56%) and S aureus (18%).
Table 1.
Characteristics of Children With and Without Cerebrospinal Fluid (CSF) Shunt Reinfection At the Time of Initial CSF Shunt Placement
| No Reinfection (n = 87) | Reinfection (n = 31) | Entire Cohort (n = 118) | |
|---|---|---|---|
| Age at initial CSF shunt placement, n (%) | |||
| 0–30 days | 29 (33%) | 12 (39%) | 41 (35%) |
| 1–6 months | 28 (32%) | 13 (42%) | 41 (35%) |
| 6–48 months | 18 (21%) | 5 (16%) | 23 (19%) |
| ≥48 months | 12 (14%) | 1 (4%) | 13 (11%) |
| Gender, n (%) | |||
| Male | 49 (56%) | 24 (77%) | 73 (62%) |
| Female | 38 (44%) | 7 (23%) | 45 (38%) |
| Race/ethnicity, n (%) | |||
| Non-Latino white | 67 (77%) | 27 (87%) | 94 (80%) |
| Latino | 8 (9%) | 3 (10%) | 11 (9%) |
| Other/Unknown | 12 (14%) | 1 (3%) | 13 (11%) |
| Insurance, n (%) | |||
| Private | 46 (53%) | 23 (74%) | 69 (58%) |
| Medicaid | 39 (45%) | 8 (26%) | 47 (40%) |
| Self-Pay | 2 (2%) | 0 (0%) | 2 (2%) |
| Weight at surgery (kg), median (IQR) | 4.1 (3.2, 10.0) | 3.5 (2.7, 5.1) | 3.9 (3.0, 8.4) |
| Prior inpatient antibiotics, n (%) | 35 (40%) | 12 (39%) | 47 (40%) |
| Indication for shunt placement, n (%) | |||
| Post-IVH due to prematurity | 18 (21%) | 8 (26%) | 26 (22%) |
| Aqueductal stenosis | 10 (12%) | 6 (19%) | 16 (14%) |
| Cyst (posterior fossa, intracranial) | 8 (9%) | 8 (26%) | 16 (14%) |
| Myelomeningocele | 12 (14%) | 3 (10%) | 15 (13%) |
| Congenitala | 11 (13%) | 3 (10%) | 14 (12%) |
| Tumor (supratentorial, posterior fossa, midbrain) | 11 (13%) | 1 (3%) | 12 (10%) |
| Post-head injury | 5 (5%) | 1 (3%) | 6 (5%) |
| Spontaneous ICH/IVH/SAH | 5 (5%) | 0 (0%) | 5 (4%) |
| Post-infectious | 4 (5%) | 1 (3%) | 5 (4%) |
| Other | 3 (3%) | 0 (0%) | 3 (2%) |
| Complex chronic conditions, n (%) | |||
| None (excepting hydrocephalus) | 66 (76%) | 28 (90%) | 94 (80%) |
| One | 12 (14%) | 2 (7%) | 14 (12%) |
| Two or more | 9 (10%) | 1 (3%) | 10 (8%) |
| Prior surgery (any), n (%) | 26 (30%) | 9 (29%) | 35 (30%) |
| Prior surgery (neurosurgical), n (%) | 48 (55%) | 11 (36%) | 59 (50%) |
| Prematurity, n (%) | 22 (25%) | 9 (29%) | 31 (26%) |
| Seizures, n (%) | 20 (23%) | 8 (26%) | 28 (24%) |
Abbreviations: CSF, cerebrospinal fluid; IQR, interquartile range; IVH, intraventricular hemorrhage; SAH, spontaneous subarachnoid hemorrhage.
aCongenital includes communicating congenital, other congenital, encephalocele, and craniosynostosis.
Table 2.
Treatment Characteristics Seen in Children With and Without Cerebrospinal Fluid (CSF) Shunt Reinfection At the Time of Initial CSF Shunt Placement, Shunt Revision(s), and First CSF Shunt Infection
| Initial CSF Shunt Placement | No Reinfection (n = 87) | Reinfection (n = 31) | Entire Cohort (n = 118) |
|---|---|---|---|
| Shunt type, n (%) | |||
| Peritoneal | 85 (98%) | 24 (77%) | 109 (92%) |
| Atrial | 2 (2%) | 5 (16%) | 7 (6%) |
| Complex | 0 (0%) | 2 (7%) | 2 (2%) |
| CSF Shunt Revision(s) | |||
| Revisions before first infection, n (%) | |||
| Zero | 54 (62%) | 17 (55%) | 71 (60%) |
| One | 21 (24%) | 8 (26%) | 29 (25%) |
| Two or more | 12 (14%) | 6 (19%) | 18 (15%) |
| Shunt type before infection, n (%) | |||
| Peritoneal | 78 (90%) | 24 (77%) | 102 (86%) |
| Atrial | 3 (3%) | 3 (10%) | 6 (5%) |
| Complex | 6 (7%) | 4 (13%) | 10 (9%) |
| First CSF Shunt Infection | |||
| Age at first CSF shunt infection, n (%) | |||
| 0–30 days | 11 (13%) | 3 (10%) | 14 (12%) |
| 1–6 months | 27 (31%) | 20 (64%) | 47 (40%) |
| 6–48 months | 31 (35%) | 7 (23%) | 38 (32%) |
| ≥48 months | 18 (21%) | 1 (3%) | 19 (16%) |
| Concordant IV antibiotic use, n (%) and duration (days), median (IQR) | 81 (93%) | 29 (94%) | 110 (93%) |
| 13 (10, 19) | 15 (11, 23) | 14 (10, 20) | |
| Concordant and appropriate IV antibiotic use, n (%) and duration (days), median (IQR) | 75 (86%) | 23 (74%) | 98 (83%) |
| 12 (9, 18) | 14 (9, 20) | 13 (9, 19) | |
| Broad spectrum IV antibiotic use, n (%) and duration (days), median (IQR) | 79 (91%) | 30 (97%) | 109 (92%) |
| 13 (5, 18) | 15 (8, 20) | 13 (6, 19) | |
| Use of IT antibiotics, n (%) | 13 (15%) | 2 (7%) | 15 (13%) |
| Use of rifampin, n (%) | 21 (24%) | 5 (16%) | 26 (22%) |
| Surgical approach to treatment, n (%) | |||
| Full removal and EVD placement | 78 (90%) | 28 (90%) | 106 (90%) |
| Full removal | 3 (3%) | 2 (7%) | 5 (4%) |
| Externalization | 1 (1%) | 0 (0%) | 1 (1%) |
| Failed externalization | 5 (6%) | 1 (3%) | 6 (5%) |
| First surgery at bedside, n (%) | 56 (64%) | 21 (68%) | 77 (65%) |
| Time from first positive culture to first infection surgery (hours), median (IQR) for 114 infections | 5 (−3, 22) | 3 (−3, 18) | 4 (−3, 19) |
| Time between infection surgeriesa (days), median (IQR) for 111 infections | 14 (11, 21) | 15 (13, 25) | 14 (11, 21) |
| Time from first negative culture to final infection surgery (days), median (IQR) for 107 infections | 14 (10, 19) | 14 (11, 20) | 14 (10, 19) |
| Complications, n (%) | |||
| Shunt malfunction | 2 (2%) | 4 (13%) | 6 (5%) |
| Infarction/hemorrhage | 2 (2%) | 2 (7%) | 4 (3%) |
| Subdural hemorrhage | 2 (2%) | 0 (0%) | 2 (2%) |
| Abdominal abscess | 0 (0%) | 2 (7%) | 2 (2%) |
| CSF leak | 0 (0%) | 1 (3%) | 1 (1%) |
| Brain abscess | 0 (0%) | 0 (0%) | 0 (0%) |
| Shunt after infection treatment, n (%) | |||
| Peritoneal | 72 (83%) | 19 (61%) | 91 (77%) |
| Atrial | 5 (6%) | 6 (19%) | 11 (9%) |
| Complex | 3 (3%) | 6 (19%) | 9 (8%) |
| Shunt removal | 7 (8%) | 0 (0%) | 7 (6%) |
Abbreviations: CSF, cerebrospinal fluid; EVD, external ventricular drain; IQR, interquartile range; IT, intrathecal; IV, intravenous.
aTime between infection surgeries (eg, time from externalization or full removal/EVD placement to reinternalization or new shunt placement).
Table 3.
Diagnostic and Microbiological Factors in Children With and Without Cerebrospinal Fluid (CSF) Shunt Reinfection At the Time of First CSF Shunt Infection
| No Reinfection (n = 87) | Reinfection (n = 31) | Entire Cohort (n = 118) | |
|---|---|---|---|
| Specific criteria in diagnosis of infection, n (%) | |||
| CSF culture | 77 (89%) | 28 (91%) | 105 (89%) |
| Wound culture excluding CSF culture | 3 (3%) | 2 (6%) | 5 (4%) |
| Visible hardware only | 4 (5%) | 0 (0%) | 4 (3%) |
| Blood culture in VA shunt only | 2 (2%) | 1 (3%) | 3 (3%) |
| CSF Gram stain only | 1 (1%) | 0 (0%) | 1 (1%) |
| Organisms in 105 CSF cultures, n (%) | |||
| Coagulase-negative Staphylococcus | 46 (60%) | 13 (46%) | 59 (56%) |
| Staphylococcus aureus | 13 (17%) | 6 (21%) | 19 (18%) |
| Other Gram-positive organisms | 5 (6%) | 4 (14%) | 9 (9%) |
| Gram-negative organisms | 5 (6%) | 4 (14%) | 9 (9%) |
| More than 1 organism | 6 (8%) | 1 (4%) | 7 (7%) |
| Fungal | 2 (3%) | 0 (0%) | 2 (2%) |
| Ventriculitis, n (%) | 15 (17%) | 4 (13%) | 19 (16%) |
| Bacteremia, n (%) | 12 (14%) | 1 (3%) | 13 (11%) |
| Polymicrobial, n (%) | 17 (20%) | 5 (16%) | 22 (19%) |
| Intermittent negative cultures, n (%) | 6 (7%) | 8 (26%) | 14 (12%) |
| Duration of positive CSF cultures, mean (SD)a | 2.7 (2.2) | 3.6 (3.1) | 2.9 (2.5) |
| Initial positive Gram stain, n (%)a | 38 (62%) | 15 (60%) | 53 (62%) |
| Initial white blood cell count, median (IQR)a | 84 (21, 257) | 231 (45, 1171) | 103 (26, 328) |
| Initial red blood cell count, median (IQR)a | 59 (5, 607) | 140 (15, 684) | 95 (6, 643) |
| Initial glucose, median (IQR)a | 44 (24, 54) | 33 (20, 44) | 39 (20, 52) |
| Initial protein, median (IQR)a | 101 (45, 189) | 206 (88, 291) | 112 (55, 246) |
Abbreviations: CSF, cerebrospinal fluid; IQR, interquartile range; SD, standard deviation; VA, ventriculoatrial.
aDuration of positive CSF cultures available for 114 children; because only laboratory work performed within 48 hours of diagnosis were included in analysis, Gram stain was available for 86 children, initial white and red blood cell counts were available for 84 children, initial glucose was available for 81 children, and protein was available for 82 children.
Factors associated with reinfection in univariate survival analyses (P < .05) were few and included: male gender; both atrial and complex shunt; age 1–6 months at first CSF shunt infection; complications at first CSF shunt infection including malfunction, subdural hemorrhage, abdominal abscess, and CSF leak; both atrial and complex shunt after infection treatment; intermittent negative CSF cultures; higher initial white blood cell count; and higher initial protein (data not shown).
Treatment characteristics associated with reinfection included ventriculoatrial (HR, 4.0; 95% CI = 1.3, 10.0) and complex (HR, 7.7; 95% CI = 1.2, 28.1) shunt at initial CSF shunt placement and any complication after first CSF shunt infection (HR, 3.1; 95% CI = 1.2, 7.0) (Table 4). Neither antibiotic approach or duration nor surgical approach to treatment of infection were significantly associated with reinfection. Specific complications significantly associated with reinfection in the base model included abdominal abscess and CSF leak (data not shown). No differences were seen in reinfection risk based on how the diagnosis of infection was made, the infecting CSF organisms, the presence of ventriculitis or bacteremia, nor polymicrobial infections. Only intermittent negative CSF cultures (HR, 3.2; 95% CI = 1.3, 7.0) were independently associated with reinfection. Among the 13 children with intermittent negative CSF cultures, the median time between positive cultures was 3.5 days (IQR, 2–6 days), and 12 had negative CSF cultures before second shunt surgery. The 13 children with intermittent negative cultures had similar organisms to the overall cohort: 6 children were infected with S epidermidis, 4 children were infected with S aureus, 1 child was infected with Enterococcus faecalis, 1 child was infected Pseudomonas aeruginosa, and 1 child was infected with Klebsiella oxytoca. Among the 8 children with intermittent negative cultures who developed reinfection, a new organism was seen in 7 cases at the time of reinfection. In a sensitivity analysis performed on 80 children with complete CSF laboratory data, infection hazard remained increased in children with ventriculoatrial shunt at initial CSF shunt placement, ventriculoatrial and complex shunt after first CSF shunt infection, CSF leak, and for every increase of 10 white blood cell count in CSF (data not shown).
Table 4.
Results From Multivariable Cox Proportional Hazard Model For Reinfection Riska
| Adjusted Hazard Ratio (95% CI) | |
|---|---|
| Initial CSF shunt type | |
| Complex | 7.7 (1.2, 28.1) |
| Atrial | 4.0 (1.3, 10.0) |
| Peritoneal | Referent |
| Any complication after first infection | 3.1 (1.2, 7.0) |
| Intermittent negative CSF cultures | 3.2(1.3, 7.0) |
Abbreviations: CI, confidence interval; CSF, cerebrospinal fluid.
aRisk factors that demonstrated statistically significant hazard ratios in unadjusted analyses were subsequently entered into a stepwise Cox proportional hazards model (see text for details). No variables were placed into the multivariable model a priori.
DISCUSSION
An improved understanding of risk factors that underlie shunt reinfection is critical because these children experience a high degree of morbidity. In this large and detailed cohort of children with first CSF shunt infection, 26% developed reinfection during the study. Few patient, treatment, and diagnostic or microbiological factors were associated with reinfection. Intriguingly, evidence of more difficult treatment both at initial placement with non-ventriculoperitoneal shunts, and at first infection with intermittent negative CSF cultures and complications, was associated with reinfection.
The finding that difficult treatment of first CSF shunt infection is associated with reinfection is of interest, even though this study is based on a small retrospective cohort. In previous prospectively conducted studies, CSF leak has already been shown to be associated with infection [22], and this association was seen again in this work. The association of intermittent negative CSF cultures during first infection with reinfection is a new finding. It makes sense that difficulty clearing bacterial pathogens from CSF would be associated with reinfection. The presence of initial white blood cell count in the sensitivity analysis also points to a role for more established inflammation during first CSF shunt infection being associated with reinfection. These findings support the practice of regular surveillance of CSF cultures while a child is being treated for CSF shunt infection.
Medical management of CSF shunt infection, and specifically IV antibiotic use in CSF shunt infection treatment, is an area of considerable interest and variation [14, 16, 17, 22, 23]. Although IV antibiotics are a mainstay of clinical practice in conjunction with surgery [13, 24, 25], duration of IV antibiotic use varies widely [14] and depends, in part, on the surgical approach used [1, 26] and pathogen involved [25, 27]. In previous studies, observed durations range from 9 days to 17 days [14, 15, 17]. We were able to characterize IV antibiotic use and duration in great detail, including concordance with recovered organisms, appropriateness of dosing, and use of broader spectrum antibiotics. Because it is regular practice at this institution to use gentamicin, a high percentage of patients were on broad-spectrum therapy. We were unable to determine whether use of broad-spectrum therapy had any adverse effects such as antibiotic-associated diarrhea or development of drug-resistant bacterial infections or colonization; these adverse effects should be studied further. In this study, we noted no differences between those with and without reinfection in duration or spectrum of antibiotic use. Further study is needed to determine the optimal duration and spectrum of therapy.
Surgical management of CSF shunt infection includes primarily shunt removal and EVD insertion followed by new shunt placement once the CSF is sterile, and less frequently shunt externalization followed by shunt replacement [13, 17, 28–32]. Surgical management is also of considerable interest and substantial variation [14, 16, 17, 22], with shunt removal or new shunt placement ranging from 55% to 71% and externalization ranging from 12% to 37% [14, 15, 17]. The superiority of any surgical approach in preventing CSF shunt reinfection is unclear [14, 15, 17, 23]. Although we found no differences between those with and without reinfection in surgical approach, we also observed far less variation in surgical approaches than the 3 earlier studies, with 90% of the cohort undergoing shunt removal or new shunt placement.
This study includes several limitations. The main limitation is the retrospective, single center study design. The small numbers of reinfections limits our ability to find differences; as such, future work with larger and better powered multicenter cohorts is still needed. The large number of variables and low number of infections mean logistic regression models may overfit these data. Because this study relied on medical chart review, some data, such as gestational age and birth weight, were missing. In addition, interventions that were not documented on the chart were listed as not done. From 2001 to 2002, antibiotic-impregnated shunt tubing may have been used; these data are not feasible to capture, and they may have affected recovery of organisms from CSF. Cerebrospinal fluid cultures are obtained at the discretion of the neurosurgeon, and variations in neurosurgeon practice affect our ability to know the true duration of positive CSF cultures. We were also unable to capture and consider antibiotic use or shunt tapping between infection episodes in both outpatient and inpatient settings. Nonetheless, this study has the strength of a large cohort of children, and it contains rich information about both initial CSF shunt placement and first infection management.
The results of this study suggest that few patient or treatment factors, except non-ventriculoperitoneal shunts and complications during first CSF shunt infection, were associated with reinfection. This work also supports a growing body of evidence that demonstrates clinical equipoise in the treatment of CSF shunt infection, because the superiority of particular medical or surgical approaches to the treatment of CSF shunt infection has not been established. Diagnostic or microbiological factors during the first infection associated with reinfection included intermittent negative CSF cultures and high initial white blood cell count in CSF. These findings suggest that clinicians who treat patients with unusual CSF shunts or more difficult first infections should have a high index of suspicion for reinfection after treatment is completed.
Acknowledgments
We thank Neal Swensen, Jared Olson, Barb Nelson, and Ali Dowling for assistance with data collection, as well as Jeff Yearley and Rene Enriquez for assistance with data management. We also appreciate the additional support for this work provided by the Division of Inpatient Medicine at PCMC, including Gena Lattin Fletcher, Flory Nkoy, and Chris Maloney; T. D. S.'s advisory committee members Bonnie Ramsey, Nino Ramirez, Jeff Ojemann, Danielle Zerr, and Rita Mangione-Smith; and last, but most importantly, Stephan Nemeth IV and Gabriel Finn Nemeth.
Disclaimer. None of the sponsors participated in design and conduct of the study; collection, management, analysis, and interpretation of the data; or preparation, review, or approval of the manuscript. Its contents are solely the responsibility of the authors and do not necessarily represent the official view of the National Center for Research Resources (NCRR) or National Institutes of Health (NIH).
Financial support. This work was supported by a PCMC Innovative Research Grant; the Children's Health Research Center at the University of Utah; Seattle Children's Center for Clinical and Translational Research; and Clinical and Translational Science Award Grant Number ULI RR025014 from NCRR, a component of NIH. T. D. S. was supported by Award K23NS062900 from the National Institute of Neurological Disorders and Stroke; the Children's Hospital Association via the Pediatric Research in Inpatient Setting Network Executive Council; and Seattle Children's Center for Clinical and Translational Research. E. A. T. was supported by NIH/National Cancer Institute Grant 1KM1CA156723 for the Translational Comparative Effectiveness Research Scholar Program.
Potential conflicts of interest. All authors: No reported conflicts.
All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.
References
- 1.Vinchon M, Dhellemmes P. Cerebrospinal fluid shunt infection: risk factors and long-term follow-up. Childs Nerv Syst. 2006;22:692–7. doi: 10.1007/s00381-005-0037-8. [DOI] [PubMed] [Google Scholar]
- 2.Frykberg T, Olden L. Infection as a cause of peritoneal catheter dysfunction in ventriculo-peritoneal shunting in children. Z Kinderchir. 1983;38(Suppl 2):84–6. doi: 10.1055/s-2008-1063084. [DOI] [PubMed] [Google Scholar]
- 3.Amacher AL, Wellington J. Infantile hydrocephalus: long-term results of surgical therapy. Childs Brain. 1984;11:217–29. doi: 10.1159/000120180. [DOI] [PubMed] [Google Scholar]
- 4.Cochrane DD, Kestle JR. The influence of surgical operative experience on the duration of first ventriculoperitoneal shunt function and infection. Pediatr Neurosurg. 2003;38:295–301. doi: 10.1159/000070413. [DOI] [PubMed] [Google Scholar]
- 5.Kestle J, Drake J, Milner R, et al. Long-term follow-up data from the Shunt Design Trial. Pediatr Neurosurg. 2000;33:230–6. doi: 10.1159/000055960. [DOI] [PubMed] [Google Scholar]
- 6.Borgbjerg BM, Gjerris F, Albeck MJ, Borgesen SE. Risk of infection after cerebrospinal fluid shunt: an analysis of 884 first-time shunts. Acta Neurochir (Wien) 1995;136(1–2):1–7. doi: 10.1007/BF01411427. [DOI] [PubMed] [Google Scholar]
- 7.Di Rocco C, Marchese E, Velardi F. A survey of the first complication of newly implanted CSF shunt devices for the treatment of nontumoral hydrocephalus. Cooperative survey of the 1991–1992 Education Committee of the ISPN. Childs Nerv Syst. 1994;10:321–7. doi: 10.1007/BF00335171. [DOI] [PubMed] [Google Scholar]
- 8.Simon TD, Hall M, Riva-Cambrin J, et al. Infection rates following initial cerebrospinal fluid shunt placement across pediatric hospitals in the United States. Clinical article. J Neurosurg Pediatr. 2009;4:156–65. doi: 10.3171/2009.3.PEDS08215. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Simon TD, Whitlock KB, Riva-Cambrin J, et al. Revision surgeries are associated with significant increased risk of subsequent cerebrospinal fluid shunt infection. Pediatr Infect Dis J. 2012;31:551–6. doi: 10.1097/INF.0b013e31824da5bd. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Simon TD, Whitlock KB, Riva-Cambrin J, et al. Association of intraventricular hemorrhage secondary to prematurity with cerebrospinal fluid shunt surgery in the first year following initial shunt placement. J Neurosurg Pediatr. 2012;9:54–63. doi: 10.3171/2011.10.PEDS11307. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.McGirt MJ, Zaas A, Fuchs HE, et al. Risk factors for pediatric ventriculoperitoneal shunt infection and predictors of infectious pathogens. Clin Infect Dis. 2003;36:858–62. doi: 10.1086/368191. [DOI] [PubMed] [Google Scholar]
- 12.Odio C, McCracken GH, Jr, Nelson JD. CSF shunt infections in pediatrics. A seven-year experience. Am J Dis Child. 1984;138:1103–8. doi: 10.1001/archpedi.1984.02140500009004. [DOI] [PubMed] [Google Scholar]
- 13.Walters BC, Hoffman HJ, Hendrick EB, Humphreys RP. Cerebrospinal fluid shunt infection Influences on initial management and subsequent outcome. J Neurosurg. 1984;60:1014–21. doi: 10.3171/jns.1984.60.5.1014. [DOI] [PubMed] [Google Scholar]
- 14.Kestle JR, Garton HJ, Whitehead WE, et al. Management of shunt infections: a multicenter pilot study. J Neurosurg 105. 2006;(3 Suppl):177–81. doi: 10.3171/ped.2006.105.3.177. [DOI] [PubMed] [Google Scholar]
- 15.Kulkarni AV, Rabin D, Lamberti-Pasculli M, Drake JM. Repeat cerebrospinal fluid shunt infection in children. Pediatr Neurosurg. 2001;35:66–71. doi: 10.1159/000050393. [DOI] [PubMed] [Google Scholar]
- 16.Whitehead WE, Kestle JR. The treatment of cerebrospinal fluid shunt infections. Results from a practice survey of the American Society of Pediatric Neurosurgeons. Pediatr Neurosurg. 2001;35:205–10. doi: 10.1159/000050422. [DOI] [PubMed] [Google Scholar]
- 17.Simon TD, Hall M, Dean JM, et al. Reinfection following initial cerebrospinal fluid shunt infection. J Neurosurg Pediatr. 2010;6:277–85. doi: 10.3171/2010.5.PEDS09457. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Tuan TJ, Thorell EA, Hamblett NM, et al. Treatment and microbiology of repeated cerebrospinal fluid shunt infections in children. Pediatr Infect Dis J. 2011;30:731–5. doi: 10.1097/INF.0b013e318218ac0e. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Kestle JR, Riva-Cambrin J, Wellons JC, 3rd, et al. A standardized protocol to reduce cerebrospinal fluid shunt infection: the Hydrocephalus Clinical Research Network Quality Improvement Initiative. J Neurosurg Pediatr. 2011;8:22–9. doi: 10.3171/2011.4.PEDS10551. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Simon TD, Riva-Cambrin J, Srivastava R, et al. Hospital care for children with hydrocephalus in the United States: utilization, charges, comorbidities, and deaths. J Neurosurg Pediatrics. 2008;1:131–7. doi: 10.3171/PED/2008/1/2/131. [DOI] [PubMed] [Google Scholar]
- 21.Committee on Infectious Diseases, Pediatrics AAp. Section 3: Summaries of Infectious Diseases, Non-Group A or B Streptococcal and Enterococcal Infections. In: Pickering LK, editor. Red Book: 2012 Report of the Committee on Infectious Diseases. 29th. Elk Grove Village, IL: American Academy of Pediatrics; 2012. pp. 686–8. [Google Scholar]
- 22.Kulkarni AV, Drake JM, Lamberti-Pasculli M. Cerebrospinal fluid shunt infection: a prospective study of risk factors. J Neurosurg. 2001;94:195–201. doi: 10.3171/jns.2001.94.2.0195. [DOI] [PubMed] [Google Scholar]
- 23.Williams MA, McAllister JP, Walker ML, et al. Priorities for hydrocephalus research: report from a National Institutes of Health-sponsored workshop. J Neurosurg. 2007;107i(5 Suppl):345–57. doi: 10.3171/PED-07/11/345. [DOI] [PubMed] [Google Scholar]
- 24.Sells CJ, Shurtleff DB, Loeser JD. Gram-negative cerebrospinal fluid shunt-associated infections. Pediatrics. 1977;59:614–8. [PubMed] [Google Scholar]
- 25.Fan-Havard P, Nahata MC. Treatment and prevention of infections of cerebrospinal fluid shunts. Clin Pharm. 1987;6:866–80. [PubMed] [Google Scholar]
- 26.Nelson JD. Cerebrospinal fluid shunt infections. Pediatr Infect Dis. 1984;3(3 Suppl):S30–2. doi: 10.1097/00006454-198405001-00011. [DOI] [PubMed] [Google Scholar]
- 27.Younger JJ, Christensen GD, Bartley DL, et al. Coagulase-negative staphylococci isolated from cerebrospinal fluid shunts: importance of slime production, species identification, and shunt removal to clinical outcome. J Infect Dis. 1987;156:548–54. doi: 10.1093/infdis/156.4.548. [DOI] [PubMed] [Google Scholar]
- 28.Kanev PM, Sheehan JM. Reflections on shunt infection. Pediatr Neurosurg. 2003;39:285–90. doi: 10.1159/000075255. [DOI] [PubMed] [Google Scholar]
- 29.Gardner P, Leipzig T, Phillips P. Infections of central nervous system shunts. Med Clin North Am. 1985;69:297–314. [PubMed] [Google Scholar]
- 30.Gardner P, Leipzig TJ, Sadigh M. Infections of mechanical cerebrospinal fluid shunts. Curr Clin Top Infect Dis. 1988;9:185–214. [PubMed] [Google Scholar]
- 31.Morissette I, Gourdeau M, Francoeur J. CSF shunt infections: a fifteen-year experience with emphasis on management and outcome. Can J Neurol Sci. 1993;20:118–22. doi: 10.1017/s0317167100047661. [DOI] [PubMed] [Google Scholar]
- 32.Venes JL. Infections of CSF shunt and intracranial pressure monitoring devices. Infect Dis Clin North Am. 1989;3:289–99. [PubMed] [Google Scholar]