Abstract
Objective
To quantify the extent to which revision(s) of cerebrospinal fluid (CSF) shunt are associated with increased risk of CSF shunt infection, after adjusting for patient factors that may contribute to infection risk.
Study design
We used the HCRN registry to assemble a large prospective six center cohort of 1,036 children undergoing initial CSF shunt placement between April 2008 and January 2012. The primary outcome of interest was first CSF shunt infection. Data for initial CSF shunt placement and all subsequent CSF shunt revisions prior to first CSF shunt infection, where applicable, were obtained. The risk of first infection was estimated using a multivariable Cox proportional hazard model accounting for patient characteristics and CSF shunt revisions, and is reported using hazard ratios (HR) with 95% confidence intervals (CI).
Results
Of the 102 children who developed first infection within 12 months of placement, 33 (32%) followed one or more CSF shunt revisions. Baseline factors independently associated with risk of first infection included: gastrostomy tube (HR 2.0, 95% CI: 1.1, 3.3), age 6–12 months (HR 0.3, 95% CI: 0.1, 0.8), and prior neurosurgery (HR 0.4, 95% CI: 0.2, 0.9). After controlling for baseline factors, infection risk was most significantly associated with the need for revision (1 revision vs. none, HR 3.9, 95% CI: 2.2, 6.5; ≥2 revisions, HR 13.0, 95% CI: 6.5, 24.9).
Conclusions
This study quantifies the elevated risk of infection associated with shunt revisions observed in clinical practice. To reduce risk of infection risk further work should optimize revision procedures.
Keywords: cerebrospinal fluid, infection, revision, shunt, ventriculoperitoneal
Although placement of cerebrospinal fluid (CSF) shunts successfully treats hydrocephalus, it also frequently leads to new medical and surgical problems. CSF shunt failure is common and necessitates subsequent CSF shunt revision surgery.[1–3] CSF shunt infections are a substantial problem for children with hydrocephalus, their parents, and their caregivers. Infections number up to 2,400 admissions and 59,000 hospital days each year in the United States.[4] Despite aggressive treatment that typically involves two surgeries as well as a prolonged inpatient stay to receive intravenous antibiotics,[5] per patient re-infection rates are high at 20 to 25%.[6, 7]
Because of the high morbidity associated with CSF shunt infection, it is critical for families and care providers to understand whether certain children undergoing CSF shunt placement are at highest risk for subsequent infection. The association of both patient risk factors and factors at the time of initial CSF shunt placement with first CSF shunt infection is complicated when intervening CSF shunt revisions are required. Additionally, risk of infection is not constant, declining with time following a revision.[8] Interval CSF shunt revision(s)[9] and previous infection[6, 7, 10] have been associated with increased odds of infection. However, the relative contribution of each intervening revision surgery, and its relationship to baseline risk of infection for a given child, was elucidated only recently that assembled a large and detailed retrospective cohort of children at a single center undergoing initial CSF shunt placement. Young age and intervening revision procedures were found to be associated with infection risk, and underlying condition of myelomeningocele demonstrated a protective effect.[11] However, this study was limited by its retrospective design and conduct at a single center.[11]
The object of this study was to quantify the extent to which CSF shunt revision(s) are associated with increased risk of CSF shunt infection, after adjusting for patient factors that may contribute to infection risk, in a large current multicenter prospective cohort of patients undergoing initial CSF shunt placement. We also considered the surgeon as well as medical and surgical decisions at the time of last surgery (either initial shunt placement or revision surgery) as potential confounders.
Methods
The Hydrocephalus Clinical Research Network (HCRN) is a collaboration of pediatric neurosurgical centers conducting systematic investigations in the management of pediatric hydrocephalus.[12–14] The HCRN has maintained a registry with prospectively-collected data about neurosurgical procedures at seven centers across North America that first started in 2008. The seven hospitals contributing to the HCRN registry include: Children’s Hospital of Alabama and University of Alabama, Children’s Hospital of Pittsburgh, Hospital for Sick Children, Primary Children’s Medical Center, Seattle Children’s Hospital, Texas Children’s Hospital, and St. Louis Children’s Hospital/Washington University; all except Pittsburgh’s contributed data to this study. The HCRN Data Coordinating Center is housed in Salt Lake City. Use of HCRN registry data was approved by the HCRN and the Institutional Review Boards at the University of Utah and Seattle Children’s Hospital.
The HCRN enrolls all children less than 18 years of age in the network registry who receive care at any of the network hospitals and who undergo CSF shunt surgery. Data collection started in April 2008 and, for this study, ended for all procedures occurring on or before December 31, 2011. We assembled a final prospective cohort of 1036 children whose initial CSF shunt placement was recorded in the HCRN registry during the study period. All children in the cohort underwent initial CSF shunt placement, and additionally may have undergone CSF shunt revision(s) and/or infection prior to the end of the study period (Figure 1).
Figure 1.
Timeline of events for each patient in cohort. Shading indicates event did not occur for all children.
Within the HCRN registry, data from each neurosurgical admission for each cohort member is collected prospectively. Each member was followed for infection until the end of study follow-up on 12/31/11, and participants were censored prior to the end of follow-up in the event of death (n=50), a move out of the network (n=34), or shunt removal (n=12). The date a child developed infection, died, moved, or had their CSF shunt removed was the date of censoring. If the date of death was not available (n=5), we used day after last procedure in registry as censoring date. No patient withdrew consent during the follow-up period.
Primary Outcome Variable
First CSF shunt infection was defined by either microbiological determination of presence of bacteria in a culture or Gram stain of CSF, wound swab, and/or pseudocyst fluid or shunt erosion (visible hardware) or abdominal pseudocyst (even without positive culture). For children with ventriculoatrial shunts, presence of bacteria in a blood culture was diagnostic of infection. This definition of CSF shunt infection was previously developed by consensus among HCRN neurosurgeons.[11, 12, 15] The first CSF sample for diagnosis of infection usually was obtained from needle aspiration of the shunt reservoir under sterile conditions outside the operating room at a bedside “shunt tap”, usually prior to the initiation of antibiotic treatment. In all other neurosurgical admissions that involved two CSF shunt surgeries (such as removal and external ventricular drain (EVD) placement followed by new shunt placement) as well as 48 hours or more of intravenous (IV) antibiotic treatment, author TS and the local research coordinator reviewed the case in detail to review whether cases met infection criteria.
Predictor variable
The main predictor variable of interest was CSF shunt revision, defined as an operative neurosurgical interventional procedure performed on the CSF shunt. In cases where a staged revision was performed in two or more procedures (such as externalization of a shunt followed several days later by re-internalization), the staged revision was handled as a single event and tested as an independent risk factor for infection. The time period between staged procedures was considered non-contributory to infection risk, and details of medical and surgical decisions at the time of only the last neurosurgical procedure were considered in analyses. Any surgical procedure associated with infection treatment was considered the first CSF shunt infection rather than CSF shunt revisions for this study.
Potential confounding variables
Potential confounding factors included patient risk factors such as demographics, risk factors prior to initial CSF shunt placement, and risk factors at the time of CSF shunt placement as shown in Table I and Figure 1. Prior neurosurgeries included endoscopic third ventriculostomy and/or subgaleal and/or reservoir as determined by chart review; and complex chronic condition(s) (CCCs) were classified as previously described.[4]
Table 1.
Characteristics of the study cohort
| Prospective Cohort (n=1036)
|
|
|---|---|
| Demographics | |
| Sex, n (%) | |
| Male | 582 (56%) |
| Female | 454 (44%) |
| Race, n (%) | |
| White | 655 (63%) |
| Other/Unknown | 199 (19%) |
| African American | 152 (15%) |
| Asian | 20 (2%) |
| AI/AN/NH/PI | 10 (1%) |
| Insurance status, n (%) | |
| Public (Medicaid, Medicare) | 423 (41%) |
| Private | 409 (39%) |
| Self-Pay | 36 (4%) |
| Other (e.g. military) | 21 (2%) |
| Government | 147 (14%) |
| Risk factors prior to initial surgery | |
| Birth weight (kg),* median (IQR) | 3.0 (1.6, 3.4) |
| Gestational age (weeks),* median (IQR) | 37 (31, 40) |
| Length of hospitalization preceding initial shunt placement (days),* median (IQR) | 2 (0, 14) |
| Prior neurosurgery, n (%) | 131 (13%) |
| Risk factors at initial CSF placement | |
| Chronologic age, n (%) | |
| 0 to < 6 months | 577 (56%) |
| 6 to < 12 months | 110 (11%) |
| 1 to < 2 years | 71 (7%) |
| 2 to < 9 years | 165 (16%) |
| 9 to 18 years | 113 (11%) |
| Indication for shunt placement, n (%) | |
| Post-IVH due to prematurity | 226 (22%) |
| Myelomeningocele | 164 (16%) |
| Posterior fossa tumor | 109 (11%) |
| Aqueductal stenosis | 84 (8%) |
| Communicating congenital | 80 (8%) |
| Supratentorial tumor | 67 (6%) |
| Post-head injury | 49 (5%) |
| Other intracranial cyst | 44 (4%) |
| Posterior fossa cyst | 40 (4%) |
| Spontaneous ICH/IVH/SAH | 39 (4%) |
| Post-infectious | 38 (4%) |
| Other | 34 (3%) |
| Other congenital | 20 (2%) |
| Midbrain tumor/lesion | 16 (2%) |
| Encephalocele | 15 (1%) |
| Craniosynostosis | 11 (1%) |
| Weight at surgery (kg), median (IQR) | 6.0 (3.3, 13.4) |
| Complex chronic conditions, n (%) | |
| None (excepting hydrocephalus) | 727 (70%) |
| One | 247 (24%) |
| Two or more | 62 (6%) |
| Gastrostomy, n (%) | 100 (10%) |
| Tracheostomy, n (%) | 29 (3%) |
| CSF shunt procedure(s) within 12 months, n (%) | |
| One CSF shunt revision | 166 (16%) |
| Two or more CSF shunt revisions | 99 (10%) |
| CSF shunt infection | 102 (10%)
|
Data was available for birthweight among 689 children, gestational age and post-conceptional age for 908 children, and length of stay for 983 children
CSF = cerebrospinal, IQR = interquartile range, IVH = intraventricular hemorrhage, ICH= intracerebral hemorrhage, SAH = subarachnoid hemorrhage
In addition to revision procedures and patient factors, we also report the association of surgeon factors and medical and surgical decisions at preceding CSF shunt surgeries (initial CSF shunt placement and/or CSF shunt revision) with CSF shunt infection. Surgeon factors include attending surgeon factors (including surgeon, surgeon’s experience defined as years since fellowship, surgeon’s volume as number of initial CSF shunt placements per year) and resident surgeon factors (month of year reflecting their experience). Medical and surgical decisions we considered are those shown in (available at www.jpeds.com). Complex proximal shunt location was designated when two or more catheters were present, and distal shunt location included peritoneal, atrial, and other.
Statistical Analyses
Cohort characteristics at baseline were summarized overall and by subsequent infection status. Binary and categorical variables were described using frequencies and percents, and continuous variables were described using means and standard deviations (SD) or medians and interquartile ranges (IQR) for duration of follow-up, time to infection, and other skewed variables. Univariate survival analyses were performed to test the associations of each cohort characteristic with infection.
Similarly, the associations of surgeon factors and medical and surgical decisions at preceding CSF shunt surgeries with CSF shunt infection were described using hazard ratios (HR), a type of rate ratio estimate, and 95% profile likelihood confidence intervals (CI).
Cox proportional hazard models were used to calculate HR and to test for the association between CSF shunt revision(s) and first CSF shunt infection while controlling for possible confounding baseline characteristics.[16, 17] CSF shunt revision(s) were categorized as none, one, and two or more; and were treated as a time-dependent covariate, because an individual child’s revision status could change during the study period. For each day on which an infection occurred, this time-dependent model compared the infection hazard among children who had not experienced a shunt revision to date to those who had experienced one or two or more.
Baseline characteristics included a priori in the multivariable model were age, sex, and indication for shunt placement. Additional baseline characteristics shown in Table I were evaluated for inclusion in the model using stepwise regression methodology.[18, 19] Using this approach, the model is initiated with a forward selection step in which a single variable is added to the model, and each forward selection step may be followed by one or more backward elimination steps in each of which a single variable may be removed from the model.[20] Criterion for entry into the model was p ≤ 0.05; criterion for removal from the model was p > 0.05. Strict criteria were used to reduce error associated with multiple testing and identification of non-significant correlated risk factors in the multivariable regression. Hazard ratios from the final multivariable models are presented with 95% profile likelihood confidence intervals (CI).
Sensitivity analyses were performed to investigate whether surgeon factors and medical and surgical decisions (either at initial placement or revision) may mediate the relationship between revisions and infection. We decided a priori to test factors of most interest, which included antibiotic impregnated shunt tubing use, non-peritoneal distal shunt location, complex shunt, staged revision, neuroendoscope use, and ultrasound use; as well as surgeon experience as it was significant in univariate analysis. These factors were evaluated in the final multivariable model as an additional time-dependent factor to determine whether the association between revisions and infections was altered by inclusion of these additional factors. In additional sensitivity analyses, we also considered site as a main effect and tested models without prior neurosurgery due to its association with intraventricular hemorrhage (IVH) in this dataset.
All analyses were performed using SAS (version 9.2, SAS Institute, Cary, NC).
Results
Baseline patient-level risk factors for the 1,036 children in the cohort are shown in Table I. The cohort had a median age of 19 weeks (IQR 4, 123), and indication for CSF shunt placement was distributed between post-intraventricular hemorrhage (IVH) due to prematurity (22%), myelomeningocele (16%), posterior fossa tumor (11%), aqueductal stenosis (8%), and other etiologies. Overall 112 (11%) children developed first CSF shunt infection. The majority (102, 91%) of first CSF shunt infection occurred within 12 months of initial placement. During the same time period, CSF shunt revisions occurred in 265 (26%) of the entire cohort and 33 (32%) of those who developed infection.
Many infection prevention practices were standardized within the network,[12] including use of prophylactic antibiotics intravenously and intraventricularly. (Data not shown) However, we did observe differences between sites in some surgical practices, such as in type of shunt brand and use of neuroendoscope and ultrasound. (Data not shown)
There were 112 (11%) children who developed infection over the course of the entire study. Table II presents the baseline patient-level risk factors by infection status, as well as unadjusted associations with risk of CSF shunt infection. In univariate survival analyses, factors that demonstrated a significant association with infection risk included self-pay insurance, cardiac complex chronic condition, presence of a gastrostomy tube, and subsequent CSF shunt revisions. Older gestational age was associated with less infection risk.
Table 2.
Association of characteristics in the cohort with first CSF shunt infection
| First CSF shunt Infection (n=112) | No CSF shunt infection (n=924) | Unadjusted Hazard Ratio (95% CI)* | |
|---|---|---|---|
|
|
|||
| Demographics | |||
| Sex, n (%) | |||
| Male | 61 (54%) | 521 (56%) | 0.9 (0.6, 1.3) |
| Female | 51 (46%) | 403 (44%) | Referent |
| Race, n (%) | |||
| White | 65 (58%) | 590 (64%) | Referent |
| Other/Unknown | 23 (21%) | 176 (19%) | 1.2 (0.8, 2.0) |
| African American | 22 (20%) | 130 (14%) | 1.5 (0.9, 2.3) |
| Asian | 1 (1%) | 19 (3%) | 0.5 (0.0, 2.3) |
| AI/AN/NH/PI | 1 (1%) | 9 (1%) | 1.0 (0.1, 4.3) |
| Ethnicity, n (%) | |||
| Latino | 18 (16%) | 127 (14%) | 1.3 (0.7, 2.0) |
| Not Latino | 72 (64%) | 629 (68%) | Referent |
| Unknown | 22 (20%) | 168 (18%) | 1.2 (0.7, 1.8) |
| Insurance status, n (%) | |||
| Public (Medicaid, Medicare) | 46 (41%) | 377 (41%) | 1.1 (0.7, 1.8) |
| Private | 39 (35%) | 370 (40%) | Referent |
| Self-Pay | 9 (8%) | 27 (3%) | 2.9 (1.3, 5.6) |
| Other (e.g. military) | 2 (2%) | 19 (2%) | 1.1 (0.2, 3.5) |
| Government | 16 (14%) | 131 (14%) | 1.2 (0.7, 2.1) |
| Risk factors prior to initial surgery | |||
| Birth weight (kg), median (IQR) | 2.7 (1.0, 3.2) | 3.0 (1.7, 3.4) | 0.8 (0.7, 1.02) |
| Gestational age (weeks), median (IQR) | 36 (26, 39) | 38 (32, 40) | 0.95 (0.92, 0.98) |
| Length of hospitalization preceding initial shunt placement (days), median (IQR) | 4 (1, 21) | 2 (0, 14) | 1.0 (0.99, 1.01) |
| Prior surgery (neurosurgical), n (%) | 13 (12%) | 118 (13%) | 0.9 (0.5, 1.6) |
| Risk factors at initial CSF placement | |||
| Chronologic age, n (%) | |||
| 0 to < 6 months | 73 (65%) | 504 (55%) | 1.0 (0.6, 2.0) |
| 6 to < 12 months | 10 (9%) | 100 (11%) | 0.7 (0.3, 1.6) |
| 1 to < 2 years | 7 (6%) | 64 (7%) | 1.4 (0.6, 3.1) |
| 2 to < 9 years | 10 (9%) | 155 (17%) | 0.7 (0.3, 1.5) |
| 9 to 18 years | 12 (11%) | 101 (11%) | Referent |
| Post-conceptional age (weeks), median (IQR) | 2 (−3, 31) | 9 (−1, 70) | 1.0 (0.99, 1.01) |
| Indication for shunt placement, n (%) | |||
| Post-IVH due to prematurity | 36 (32%) | 190 (21%) | 1.5 (0.8, 3.4) |
| Tumor (posterior fossa, supratentorial, midbrain) | 15 (13%) | 177 (19%) | 0.8 (0.3, 1.9) |
| Myelomeningocele | 15 (13%) | 149 (16%) | 0.8 (0.4, 1.9) |
| Congenital (communicating, other, encephalocele, craniosynostosis) | 13 (12%) | 113 (12%) | 0.9 (0.4, 2.2) |
| Cyst (posterior fossa, intracranial) | 12 (11%) | 72 (8%) | 1.4 (0.6, 3.4) |
| Aqueductal stenosis | 9 (8%) | 75 (8%) | Referent |
| Post-head injury | 5 (4%) | 44 (5%) | 0.9 (0.3, 2.7) |
| Spontaneous ICH/IVH/SAH | 2 (2%) | 37 (4%) | 0.5 (0.1, 2.0) |
| Post-infectious | 3 (3%) | 35 (4%) | 0.7 (0.2, 2.3) |
| Other | 2 (2%) | 32 (3%) | 0.5 (0.1, 2.0) |
| Weight at surgery (kg), median (IQR) | 4.1 (2.9, 9.7) | 6.3 (3.4, 13.9) | 1.0 (0.99, 1.01) |
| Complex chronic conditions, n (%) | |||
| None (excepting hydrocephalus) | 72 (64%) | 655 (71%) | Referent |
| One | 30 (27%) | 217 (23%) | 1.3 (0.8, 1.9) |
| Two or more | 10 (9%) | 52 (6%) | 1.8 (0.9, 3.3) |
| Neuromuscular CCC, n (%) | 15 (13%) | 100 (11%) | 1.3 (0.7, 2.1) |
| Cardiac CCC, n (%) | 14 (13%) | 69 (7%) | 1.9 (1.1, 3.3) |
| Respiratory CCC, n (%) | 11 (10%) | 78 (8%) | 1.2 (0.6, 2.1) |
| Renal CCC, n (%) | 2 (2%) | 7 (1%) | 2.1 (0.3, 6.6) |
| Gastrointestinal CCC, n (%) | 4 (4%) | 12 (1%) | 2.4 (0.7, 5.8) |
| Hematologic CCC, n (%) | 0 (0%) | 1 (0%) | n/a |
| Metabolic CCC, n (%) | 2 (2%) | 5 (1%) | 2.8 (0.5, 8.7) |
| Congenital/genetic CCC, n (%) | 6 (5%) | 44 (5%) | 1.2 (0.5, 2.4) |
| Malignancy CCC, n (%) | 1 (1%) | 26 (3%) | 0.3 (0.0, 1.4) |
| Gastrostomy, n (%) | 19 (17%) | 81 (9%) | 2.0 (1.2, 3.2) |
| Tracheostomy, n (%) | 5 (4%) | 24 (3%) | 1.7 (0.6, 3.7) |
| CSF shunt revision(s) within 12 months, n (%) | |||
| No CSF shunt revisions | 79 (71%) | 692 (75%) | Referent |
| One CSF shunt revision | 19 (17%) | 147 (16%) | 3.6 (2.1, 5.9) |
| Two or more CSF shunt revisions | 14 (12%) | 85 (9%) | 11.9 (6.1, 21.6) |
|
|
|||
CSF = cerebrospinal, IQR = interquartile range, IVH = intraventricular hemorrhage, ICH= intracerebral hemorrhage, SAH = subarachnoid hemorrhage, CCC = complex chronic condition
Association between revisions and infection risk
Our multivariable model of infection risk demonstrates the HRs for first CSF shunt infection for each of the explanatory variables while controlling for the other explanatory variables (Table III). Gastrostomy tube at initial CSF shunt placement (HR 2.0, 95% CI: 1.1, 3.3) was significantly associated with an increased risk of infection, and age 6–12 months (HR 0.3, 95% CI: 0.1, 0.8) and prior neurosurgery (HR 0.4, 95% CI: 0.2, 0.9) were significantly associated with a decreased risk of infection. When age was handled as a continuous variable, it was not associated with infection.
Table 3.
Results from multivariable Cox proportional hazard model for the association between revisions and first CSF shunt infection, adjusting for baseline characteristics and confounders*
| Adjusted Hazard Ratio (95% CI)
|
|
|---|---|
| Revision procedure | |
| Two or more | 13.0 (6.5, 24.9) |
| One | 3.9 (2.2, 6.4) |
| None | Referent |
| Chronologic age at initial shunt placement | |
| 0 to < 6 months | 0.7 (0.3, 1.7) |
| 6 to < 12 months | 0.3 (0.1, 0.8) |
| 1 to < 2 years | 0.8 (0.3, 1.9) |
| 2 to < 9 years | 0.6 (0.2, 1.3) |
| 9 to 18 years | Referent |
| Female sex | 1.1 (0.8, 1.7) |
| Indication for initial shunt placement | |
| Myelomeningocele | 0.7 (0.3, 1.8) |
| Post-IVH due to prematurity | 1.9 (0.9, 4.2) |
| Post-head injury | 0.7 (0.2, 2.2) |
| Tumor (posterior fossa, midbrain, supratentorial) | 0.7 (0.3, 1.8) |
| Cyst (other intracranial, posterior fossa) | 1.6 (0.7, 3.9) |
| Spontaneous ICH/IVH/SAH | 0.5 (0.1, 2.0) |
| Post-infectious | 0.6 (0.1, 1.9) |
| Congenital (communicating, other, encephalocele, craniosynostosis) | 0.9 (0.4, 2.2) |
| Other | 0.5 (0.1, 2.4) |
| Aqueductal stenosis | Referent |
| Gastrostomy tube | 2.0 (1.1, 3.3) |
| Prior neurosurgery |
0.4 (0.2, 0.9)
|
Baseline characteristics included a priori were age, sex, and indication for shunt placement. Additional baseline characteristics shown in Table I were evaluated for inclusion in the model using stepwise regression methodology. Criterion for entry into the model was p ≤ 0.05; criterion for removal from the model was p > 0.05.
IVH = intraventricular hemorrhage, ICH= intracerebral hemorrhage, SAH = subarachnoid hemorrhage
Infection risk was significantly related to revision after controlling for age, sex, and indication for shunt placement. Infection risk was more than three times greater among those with a revision compared with those without a revision (HR 3.9, 95% CI: 2.2, 6.5), and the risk was 13 times greater among those with two or more revisions compared with none (HR 13.0, 95% CI: 6.5, 24.9).
Sensitivity analyses
Sensitivity analyses were performed to determine whether the association between revisions and infections were confounded by surgeon factors and medical and surgical decisions. Only a few surgeon factors, medical and surgical decisions at the preceding surgeries (initial CSF shunt placement and/or CSF shunt revision) were associated with infection hazard in univariate survival analysis; these included less surgeon experience (lower risk) and complex shunt (higher risk). Medical and surgical decisions of interest, as well as surgeon’s experience, were added to the multivariable Cox proportional hazards model in Table III to evaluate whether the association between revisions and infection remained significant. Surgeon experience for those under 5 years remained independently significant with HR 0.5 (95% CI: 0.2, 0.8); however, its addition did not change the HR estimates or significance for revisions substantially. None of the other medical and surgical factors of interest, nor inclusion of site, either demonstrated a significant association with infection hazard or changed the HR estimates or significance for revisions. Exclusion of prior neurosurgery changed the HR estimates for other variables only by making gastrostomy insignificant.
Discussion
In this large prospective multicenter cohort of children undergoing initial CSF shunt placement, 11% developed a first CSF shunt infection. We developed robust models that quantified the relative contributions to the risk of infection of baseline patient factors (including indication for CSF shunt placement) as well as each revision surgery. Baseline patient factors independently associated with first infection included presence of a gastrostomy tube (increased risk) and prior neurosurgery (decreased risk). However, it is revision surgeries that are most significantly associated with infection risk, and this risk increased dramatically with subsequent revisions. Surgeon, site, and medical and surgical decisions involved in shunt surgeries, both at initial placement and revision surgery, did not alter the relationship between revisions and infection. This study decisively quantifies this dramatic and clinically important risk of infection associated with revisions that has been observed in clinical practice.
This study augments a similar analysis performed in a single center retrospective cohort.[11] In the current study, we found that shunt placement at 6–12 months of age was protective. In our previous work, age 0 to <6 months at CSF shunt placement was significantly associated with an increased risk of infection (HR 2.4, 95% CI: 1.02, 6.7). The association of age with infection risk appears as inconsistent across ours and other studies.[21, 22]
In this larger study, we found that presence of a gastrostomy tube was associated with infection risk. We did not assess gastrostomy risk in our previous analysis. Physically, it makes some sense that the presence of a gastrostomy tubecould confer infection risk when a ventriculoperitoneal shunt is present. However, it may also be a marker for medical complexity. Unfortunately, we do not have data on the timing of gastrostomy tube placement in this cohort, limiting our ability to understand the nature of its timing with infection risk. Presence of a gastrostomy tube represents an important risk factor to consider in clinical practice as well as future studies.
In the current study, we found that prior neurosurgery was protective. Although we considered prior neurosurgery in previous work and it was not associated with infection, it was assessed from International Classification of Disease, Ninth Revision, Clinical Modification codes rather than chart review and, unlike this study, included consideration of external ventricular devices (EVDs). In our previous work, myelomeningocele was significantly associated with a decreased risk of infection (HR 0.4, 95% CI: 0.2, 0.8); we wondered if the routine use of exposure to antibiotics intravenously seen in children with myelomeningocele might explain the protective effect in the single center study. The protective effect of myelomeningocele was not replicated in this larger cohort. One possible explanation for the protective effect of prior neurosurgery in this study, which we did not test, again could be prior exposure to antibiotics intravenously.
None of the associations of baseline factors in this or the previous study was as strong as performance of CSF shunt revision. Taking both studies together, one CSF shunt revision was associated with 3 to 4 fold higher hazard of infection whereas two or more CSF shunt revisions were associated with 6 to 13 fold higher infection hazard. Our finding of each subsequent CSF shunt revision increasing infection risk for a patient also is consistent with previous work.[9] Each surgery likely represents an opportunity to introduce new organisms into the CSF and onto shunt hardware. Each surgical exposure of the shunt apparatus may represent an opportunity to further colonize a shunt apparatus in a sterile setting, with an additive effect over time. Future research efforts should focus on modalities to reduce microbial exposure during the peri-operative period, in order to optimize revision procedures and reduce risk of subsequent infection.
Conclusions about medical and surgical decisions, as well as surgeon’s volumes and experience, can be limited as they vary between and are highly confounded by surgeon and study site.[23] However, many medical and surgical decisions to prevent infection, including prophylactic use of antibiotics,[24, 25] site preparation including hair clipping,[26, 27] type of skin cleanser,[26, 28] double gloving,[25] prophylactic intraventricular antibiotic use[28–31], and use of a wound dressing were standardized within the HCRN.[12] By reducing background variation, this standardization permitted us to more concisely address baseline risk factors associated with infection risk. Areas of ongoing controversy in the prevention of CSF shunt infection include use of antibiotic impregnated shunt tubing,[32–35] distal shunt location,[22, 36–40] and use of neuroendoscope.[10, 25] In this study, surgeon, study site, or medical and surgical decision did not substantially contribute to infection risk beyond CSF shunt revision itself.
This work has several limitations. The conduct of this study at a group of centers adhering to a common checklist of medical and surgical practices to prevent shunt infection reduces generalizability of findings outside of the HCRN. The small number of surgeons and hospitals limits the ability to systematically study surgeon and hospital effects on patient outcome. However, the multi-institutional nature of this study gives it greater generalizability than previous studies. Certain patient risk factors may not have been captured in this analysis. Missing data on gestational age and birth weight limits our ability to draw conclusions about these patient factors, and we did not collect data on use of antibiotic-impregnated EVD or postoperative wound care practices. Outcomes were all collected passively. Our definition of infection was developed by consensus within the HCRN and does not permit easy comparison with infection rates at non-HCRN sites or those used by hospital infection control groups. For neurosurgical procedures, the assumption is that patients return to the same center for care. Death as an outcome depends on the neurosurgical center staff knowledge of a child’s demise. If children moved to a different locations, they were lost to follow-up.
Families of and care providers for children with recurrent CSF shunt revisions should be aware of their increased odds of CSF shunt infection. Further research is needed to optimize strategies for revision procedures and to reduce risk of subsequent infection.
Acknowledgments
We thank the contributing children and families and their pediatric neurosurgeons at all participating centers. Finally, we thank Stephen and Gabriel Finn Nemeth for support and valuable feedback.
The HCRN has been funded by private philanthropy and National Institute of Neurological Disorders And Stroke (1RC1NS068943-01). T.S. is supported by the National Institute of Neurological Disorders And Stroke (K23NS062900), the Children’s Hospital Association via the Pediatric Research in Inpatient Setting Network Executive Council, Seattle Children’s Center for Clinical and Translational Research, and the National Center for Research Resources, a component of the National Institutes of Health (CTSA ULI RR025014). Its contents are solely the responsibility of the authors and do not necessarily represent the official view of the National Center for Research Resources or National Institutes of Health.
Abbreviations
- CSF
cerebrospinal fluid
- CCC
complex chronic conditions
- EVD
external ventricular drain
- HR
hazard ratios
- HCRN
Hydrocephalus Clinical Research Network
- IQR
interquartile range
- IT
intrathecal
- IV
intravenous
- IVH
intraventricular hemorrhage
Appendix
Additional members of HCRN include:
Amita Bey (Children’s Hospital of Alabama and University of Alabama); Arlene Luther (Children’s Hospital of Pittsburgh); Lindsay O’Connor and Megha Agrawal (Hospital for Sick Children); Tracey Habrock-Bach (Primary Children’s Medical Center); Amy Anderson (Seattle Children’s Hospital); Sheila Ryan (Texas Children’s Hospital); and Deanna Mercer (Washington University).
HCRN Data Coordinating Center--Nichol Nunn and Jeff Yearley.
HCRN Data Coordinating Center Statistical Staff--Michelle Miskin and Lydia Dong.
Footnotes
The authors declare no conflicts of interest.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
References
- 1.Browd SR, Gottfried ON, Ragel BT, Kestle JR. Failure of cerebrospinal fluid shunts: part II: overdrainage, loculation, and abdominal complications. Pediatr Neurol. 2006;34:171–6. doi: 10.1016/j.pediatrneurol.2005.05.021. [DOI] [PubMed] [Google Scholar]
- 2.Browd SR, Ragel BT, Gottfried ON, Kestle JR. Failure of cerebrospinal fluid shunts: part I: Obstruction and mechanical failure. Pediatr Neurol. 2006;34:83–92. doi: 10.1016/j.pediatrneurol.2005.05.020. [DOI] [PubMed] [Google Scholar]
- 3.Garton HJ, Kestle JR, Drake JM. Predicting shunt failure on the basis of clinical symptoms and signs in children. J Neurosurg. 2001;94:202–10. doi: 10.3171/jns.2001.94.2.0202. [DOI] [PubMed] [Google Scholar]
- 4.Simon TD, Riva-Cambrin J, Srivastava R, Bratton SL, Dean JM, Kestle JR. 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]
- 5.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]
- 6.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]
- 7.Kestle JR, Garton HJ, Whitehead WE, Drake JM, Kulkarni AV, Cochrane DD, et al. Management of shunt infections: a multicenter pilot study. J Neurosurg. 2006;105:177–81. doi: 10.3171/ped.2006.105.3.177. [DOI] [PubMed] [Google Scholar]
- 8.McGirt MJ, Leveque JC, Wellons JC, 3rd, Villavicencio AT, Hopkins JS, Fuchs HE, et al. Cerebrospinal fluid shunt survival and etiology of failures: a seven-year institutional experience. Pediatr Neurosurg. 2002;36:248–55. doi: 10.1159/000058428. [DOI] [PubMed] [Google Scholar]
- 9.Simon TD, Hall M, Riva-Cambrin J, Albert JE, Jeffries HE, Lafleur B, 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]
- 10.McGirt MJ, Zaas A, Fuchs HE, George TM, Kaye K, Sexton DJ. 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]
- 11.Simon TD, Whitlock KB, Riva-Cambrin J, Kestle JR, Rosenfeld M, Dean JM, 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]
- 12.Kestle JR, Riva-Cambrin J, Wellons JC, 3rd, Kulkarni AV, Whitehead WE, Walker ML, 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]
- 13.Riva-Cambrin J, Shannon CN, Holubkov R, Whitehead WE, Kulkarni AV, Drake J, et al. Center effect and other factors influencing temporization and shunting of cerebrospinal fluid in preterm infants with intraventricular hemorrhage. J Neurosurg Pediatr. 2012;9:473–81. doi: 10.3171/2012.1.PEDS11292. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Wellons JC, Shannon CN, Kulkarni AV, Simon TD, Riva-Cambrin J, Whitehead WE, et al. A multicenter retrospective comparison of conversion from temporary to permanent cerebrospinal fluid diversion in very low birth weight infants with posthemorrhagic hydrocephalus. J Neurosurg Pediatr. 2009;4:50–5. doi: 10.3171/2009.2.PEDS08400. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Tuan TJ, Thorell EA, Hamblett NM, Kestle JR, Rosenfeld M, Simon TD. 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]
- 16.Allison MA, Daley MF, Crane LA, Barrow J, Beaty BL, Allred N, et al. Influenza vaccine effectiveness in healthy 6- to 21-month-old children during the 2003–2004 season. J Pediatr. 2006;149:755–62. doi: 10.1016/j.jpeds.2006.06.036. [DOI] [PubMed] [Google Scholar]
- 17.Allison PA. Survival analysis using the SAS system; a practical guide. Cary, North Carolina: SAS Institute, Inc; 1995. [Google Scholar]
- 18.Agresti A. Categorical Data Analysis. 2. New York: John Wiley & Sons; 2002. [Google Scholar]
- 19.Hosmer DW, Jr, Lemeshow S. Applied Logistic Regression. 2. New York: John Wiley & Sons; 2000. [Google Scholar]
- 20.SAS Institute I. SAS Help and Documentation. Cary, NC: SAS Institute, Inc; 2009. The PHREG Procedure: MODEL Statement. [Google Scholar]
- 21.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]
- 22.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;136:1–7. doi: 10.1007/BF01411427. [DOI] [PubMed] [Google Scholar]
- 23.Simon T, Whitlock K, Riva-Cambrin J, Kestle J, Rosenfeld M, Dean J, et al. Association of intraventricular hemorrhage secondary to prematurity with cerebrospinal fluid shunt surgery in the first year following initial shunt placement. Journal of Neurosurgery: Pediatrics. 2012;9:54–63. doi: 10.3171/2011.10.PEDS11307. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Classen DC, Evans RS, Pestotnik SL, Horn SD, Menlove RL, Burke JP. The timing of prophylactic administration of antibiotics and the risk of surgical-wound infection. N Engl J Med. 1992;326:281–6. doi: 10.1056/NEJM199201303260501. [DOI] [PubMed] [Google Scholar]
- 25.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]
- 26.Kestle JR, Hoffman HJ, Soloniuk D, Humphreys RP, Drake JM, Hendrick EB. A concerted effort to prevent shunt infection. Childs Nerv Syst. 1993;9:163–5. doi: 10.1007/BF00272269. [DOI] [PubMed] [Google Scholar]
- 27.Horgan MA, Piatt JH., Jr Shaving of the scalp may increase the rate of infection in CSF shunt surgery. Pediatr Neurosurg. 1997;26:180–4. doi: 10.1159/000121187. [DOI] [PubMed] [Google Scholar]
- 28.Shurtleff DB, Stuntz JT, Hayden PW. Experience with 1201 cerebrospinal fluid shunt procedures. Pediatr Neurosci. 1985;12:49–57. doi: 10.1159/000120218. [DOI] [PubMed] [Google Scholar]
- 29.Ragel BT, Browd SR, Schmidt RH. Surgical shunt infection: significant reduction when using intraventricular and systemic antibiotic agents. J Neurosurg. 2006;105:242–7. doi: 10.3171/jns.2006.105.2.242. [DOI] [PubMed] [Google Scholar]
- 30.Lambert M, MacKinnon AE, Vaishnav A. Comparison of two methods of prophylaxis against CSF shunt infection. Z Kinderchir. 1984;39 (Suppl 2):109–10. doi: 10.1055/s-2008-1044298. [DOI] [PubMed] [Google Scholar]
- 31.Quigley MR, Reigel DH, Kortyna R. Cerebrospinal fluid shunt infections. Report of 41 cases and a critical review of the literature. Pediatr Neurosci. 1989;15:111–20. [PubMed] [Google Scholar]
- 32.Sciubba DM, Stuart RM, McGirt MJ, Woodworth GF, Samdani A, Carson B, et al. Effect of antibiotic-impregnated shunt catheters in decreasing the incidence of shunt infection in the treatment of hydrocephalus. J Neurosurg. 2005;103:131–6. doi: 10.3171/ped.2005.103.2.0131. [DOI] [PubMed] [Google Scholar]
- 33.Aryan HE, Meltzer HS, Park MS, Bennett RL, Jandial R, Levy ML. Initial experience with antibiotic-impregnated silicone catheters for shunting of cerebrospinal fluid in children. Childs Nerv Syst. 2005;21:56–61. doi: 10.1007/s00381-004-1052-x. [DOI] [PubMed] [Google Scholar]
- 34.Govender ST, Nathoo N, van Dellen JR. Evaluation of an antibiotic-impregnated shunt system for the treatment of hydrocephalus. J Neurosurg. 2003;99:831–9. doi: 10.3171/jns.2003.99.5.0831. [DOI] [PubMed] [Google Scholar]
- 35.Kan P, Kestle J. Lack of efficacy of antibiotic-impregnated shunt systems in preventing shunt infections in children. Childs Nerv Syst. 2007 doi: 10.1007/s00381-007-0296-7. [DOI] [PubMed] [Google Scholar]
- 36.Griebel R, Khan M, Tan L. CSF shunt complications: an analysis of contributory factors. Childs Nerv Syst. 1985;1:77–80. doi: 10.1007/BF00706686. [DOI] [PubMed] [Google Scholar]
- 37.Davis SE, Levy ML, McComb JG, Masri-Lavine L. Does age or other factors influence the incidence of ventriculoperitoneal shunt infections? Pediatr Neurosurg. 1999;30:253–7. doi: 10.1159/000028806. [DOI] [PubMed] [Google Scholar]
- 38.George R, Leibrock L, Epstein M. Long-term analysis of cerebrospinal fluid shunt infections. A 25-year experience. J Neurosurg. 1979;51:804–11. doi: 10.3171/jns.1979.51.6.0804. [DOI] [PubMed] [Google Scholar]
- 39.Kontny U, Hofling B, Gutjahr P, Voth D, Schwarz M, Schmitt HJ. CSF shunt infections in children. Infection. 1993;21:89–92. doi: 10.1007/BF01710738. [DOI] [PubMed] [Google Scholar]
- 40.Schoenbaum SC, Gardner P, Shillito J. Infections of cerebrospinal fluid shunts: epidemiology, clinical manifestations, and therapy. J Infect Dis. 1975;131:543–52. doi: 10.1093/infdis/131.5.543. [DOI] [PubMed] [Google Scholar]