Abstract
Background
Previous studies of cerebrospinal fluid (CSF) shunt infection treatment have been limited in size and unable to compare patient and treatment characteristics by infecting organism. Our objective was to describe variation in patient and treatment characteristics for children with first CSF shunt infection, stratified by infecting organism subgroups outlined in the 2017 Infectious Disease Society of America’s (IDSA) guidelines.
Methods
We studied a prospective cohort of children <18 years of age undergoing treatment for first CSF shunt infection at one of 7 Hydrocephalus Clinical Research Network hospitals from April 2008 to December 2012. Differences between infecting organism subgroups were described using univariate analyses and Fisher’s exact tests.
Results
There were 145 children whose infections were diagnosed by CSF culture and addressed by IDSA guidelines, including 47 with Staphylococcus aureus, 52 with coagulase-negative Staphylococcus, 37 with Gram-negative bacilli, and 9 with Propionibacterium acnes. No differences in many patient and treatment characteristics were seen between infecting organism subgroups, including age at initial shunt, gender, race, insurance, indication for shunt, gastrostomy, tracheostomy, ultrasound, and/or endoscope use at all surgeries before infection, or numbers of revisions before infection. A larger proportion of infections were caused by Gram-negative bacilli when antibiotic-impregnated catheters were used at initial shunt placement (12 of 23, 52%) and/or subsequent revisions (11 of 23, 48%) compared with all other infections (9 of 68 [13%] and 13 of 68 [19%], respectively). No differences in reinfection were observed between infecting organism subgroups.
Conclusions
The organism profile encountered at infection differs when antibiotic-impregnated catheters are used, with a higher proportion of Gram-negative bacilli. This warrants further investigation given increasing adoption of antibiotic-impregnated catheters.
Keywords: antibiotic, antibiotic-impregnated catheter, cerebrospinal, infection, shunt
Although life-saving and the mainstay of hydrocephalus treatment [1], cerebrospinal fluid (CSF) shunts can cause new and chronic problems for children with hydrocephalus. Mechanical malfunction is frequent with 40% of shunts requiring surgical revision within 2 years [2], and CSF shunt infection rates range from 0% to 35% per surgery [3–8].
Although numerous review articles have been written [1, 9–12], until February 2017 no organization in the United States or elsewhere has published an official guideline specifically for management of CSF shunt infection [13]. The 3 largest pediatric observational studies of treatment of CSF shunt infection observed reinfection rates of 26% (18 of 70), 20% (10 of 51), and 26% (31 of 118), respectively [15–17]. Investigators have noted considerable variation in surgical and antibiotic decisions for shunt infections that might explain reinfections [6, 18–20]. However, because of the small numbers of children included, previous studies of CSF shunt infection treatment have been limited in their ability to compare patient and treatment characteristics by infecting organism.
The Hydrocephalus Clinical Research Network (HCRN) registry provides a unique opportunity to understand the current state of treatment for CSF shunt infection after creation of the 2017 Infectious Disease Society of America’s (IDSA) recommendations. We aimed to describe the variation in patient and treatment characteristics for children with first CSF shunt infection, stratified by infecting organism subgroups outlined in the 2017 IDSA recommendations for CSF shunt infection treatment, in a prospective cohort of children with CSF shunt placement recorded in the HCRN registry.
METHODS
Setting
The HCRN is a collaboration of 14 pediatric neurosurgical centers across North America, with 7 active during the study period and therefore participating in this study: Children’s Hospital of Alabama, Children’s Hospital of Pittsburgh, The Hospital for Sick Children, Primary Children’s Hospital, Seattle Children’s Hospital, Texas Children’s Hospital, and St. Louis Children’s Hospital. Hydrocephalus Clinical Research Network registry data use was approved by the HCRN and the Institutional Review Boards at the University of Utah and Seattle Children’s Hospital.
Study Population and Data Collection
Within the HCRN registry, data from each neurosurgical admission for each child admitted is collected contemporaneously. Data collection started in April 2008 and, for this study, ended on December 31, 2012, except for children at The Hospital for Sick Children who were observed until December 31, 2011. The final cohort included either children whose initial CSF shunt placement and all subsequent shunt surgeries were recorded in the HCRN registry during the study period (n = 151), or those whose shunt surgery history (initial CSF shunt placement and subsequent CSF shunt revision(s)) were able to be obtained retrospectively (n = 82) and whose confirmed first CSF shunt infection was recorded in the HCRN registry.
The HCRN consensus definition was used for CSF shunt infection [21–24]: (1) microbiological determination of bacteria present in a culture or Gram stain of CSF, wound swab, and/or pseudocyst fluid, or (2) shunt erosion (visible hardware), or (3) abdominal pseudocyst (without positive culture); or for children with ventriculoatrial shunts, (4) presence of bacteria in a blood culture. The first CSF sample for diagnosis of infection was usually obtained from needle aspiration of the shunt reservoir under sterile conditions and before antibiotic initiation. To ensure all infections were identified, all neurosurgical admissions involving 2 CSF shunt surgeries and ≥48 hours of intravenous (IV) antibiotic treatment were reviewed by T.D.S. and local HCRN staff to clarify whether infection criteria were met.
Our intent was to understand uncomplicated CSF shunt infections. To that end, we included only children whose first CSF shunt infection was diagnosed by recovery of bacteria in CSF culture. In addition, we excluded children who developed bacteremia or secondary ventriculitis that appropriately prolonged treatment (eg, recovery of the same organism from the bloodstream on day 7 of IV antibiotic treatment for a CSF infection). Finally, we limited analysis to the children whose infecting organism was addressed by the 2017 IDSA recommendations; these subgroups included Staphylococcus aureus, coagulase-negative Staphylococcus, Gram-negative bacilli, and Propionibacterium acnes.
Patient and Treatment Characteristics
Patient characteristics included demographics and risk factors before initial CSF shunt placement (Table 1) and treatment decisions at the time of CSF shunt placement and intervening revision surgeries (Table 2); these have been previously described [22, 24, 25]. Chronological age was handled both as a categorical and continuous variable [22, 24, 25]. Complex chronic condition(s) were classified as previously described [24, 26]. Of note, hospitals within the HCRN have instituted an infection prevention bundle used at the time of initial shunt placement and revision(s) since June 2007, although compliance with all aspects is variable; key aspects of each iteration have included routine intrathecal antibiotic use and routine use of antibiotic-impregnated catheters, respectively [21, 27].
Table 1.
Characteristics of Children in Study Cohort, Stratified by Organism
| Characteristics | Staphylococcus aureus (n = 47) | Coagulase-Negative Staphylococcus (n = 52) | Gram-Negative Bacilli (n = 37) | Propionibacterium acnes (n = 9) |
|---|---|---|---|---|
| Chronological age at initial shunt (weeks), median (IQR) | 13 (7–33) | 17 (2–59) | 14 (2–33) | 27 (6–96) |
| Gestational age (weeks), median (IQR)a | 33 (26–38) | 38 (27–40) | 35 (25–39) | 26 (26–36) |
| Gender, n (%) | ||||
| Male | 29 (62%) | 32 (61%) | 16 (43%) | 6 (67%) |
| Female | 18 (38%) | 20 (39%) | 21 (57%) | 3 (33%) |
| Race, n (%) | ||||
| White | 25 (53%) | 35 (67%) | 29 (78%) | 6 (67%) |
| African American | 12 (26%) | 7 (14%) | 2 (6%) | 3 (33%) |
| Unknown/Other | 10 (21%) | 9 (17%) | 6 (16%) | 0 (0%) |
| Native Hawaiian/Pacific Islander | 0 (0%) | 1 (2%) | 0 (0%) | 0 (0%) |
| Ethnicity, n (%) | ||||
| Latino | 4 (8%) | 11 (21%) | 4 (11%) | 1 (11%) |
| Non-Latino | 36 (77%) | 34 (65%) | 27 (73%) | 8 (89%) |
| Unknown | 7 (15%) | 7 (14%) | 6 (16%) | 0 (0%) |
| Insurance, n (%) | ||||
| Public | 25 (53%) | 26 (50%) | 12 (32%) | 4 (44%) |
| Private | 11 (23%) | 18 (34%) | 21 (57%) | 5 (56%) |
| Self-Pay | 5 (11%) | 2 (4%) | 1 (3%) | 0 (0%) |
| Other | 0 (0%) | 1 (2%) | 0 (0%) | 0 (0%) |
| Government (Canada only) | 6 (13%) | 5 (10%) | 3 (8%) | 0 (0%) |
| Birth weight (kg), mean (SD)a,b | 1.9 (1.2) | 2.7 (1.3) | 2.1 (1.3) | 2.9 (0.6) |
| Weight at initial shunt surgery (kg), median (IQR)a | 3.7 (2.5–11.2) | 4.4 (3.5–11.0) | 4.3 (2.6–16.0) | 6.3 (4.0–8.6) |
| Indication for Shunt Placement, n (%) | ||||
| Post-IVH due to prematurity | 19 (40%) | 14 (27%) | 14 (38%) | 3 (33%) |
| Congenital (communicating congenital, other congenital, encephalocele, craniosynostosis) | 9 (19%) | 11 (21%) | 4 (11%) | 3 (34%) |
| Myelomeningocele | 6 (13%) | 8 (15%) | 8 (21%) | 1 (11%) |
| Tumor (supratentorial, posterior fossa, midbrain) | 2 (4%) | 9 (17%) | 3 (8%) | 0 (0%) |
| Cyst (posterior fossa, intracranial) | 3 (7%) | 3 (6%) | 4 (11%) | 1 (11%) |
| Aqueductal stenosis | 2 (4%) | 3 (6%) | 0 (0%) | 0 (0%) |
| Posthead injury | 3 (7%) | 1 (2%) | 1 (3%) | 0 (0%) |
| Postinfectious | 1 (2%) | 2 (4%) | 1 (3%) | 1 (11%) |
| Other | 2 (4%) | 1 (2%) | 0 (0%) | 0 (0%) |
| Spontaneous hemorrhage | 0 (0%) | 0 (0%) | 2 (5%) | 0 (0%) |
| Complex Chronic Conditions, n (%) | ||||
| None (excepting hydrocephalus) | 20 (42%) | 35 (67%) | 16 (43%) | 4 (44%) |
| One | 21 (45%) | 13 (25%) | 15 (41%) | 5 (56%) |
| Two or more | 6 (13%) | 4 (8%) | 6 (16%) | 0 (0%) |
| Gastrostomy, n (%) | 5 (17%) | 3 (9%) | 4 (19%) | 0 (0%) |
| Tracheostomy, n (%) | 2 (7%) | 0 (0%) | 1(5%) | 0 (0%) |
Abbreviations: IQR, interquartile range; IVH, intraventricular hemorrhage; SD, standard deviation.
aGestational age was reported for 132 children, birth weight was reported for 97 children, and weight at initial shunt surgery was reported for 96 children.
bSignificant differences seen between groups (P < .10) in Fisher’s exact tests, F-tests, or Kruskal-Wallis tests.
Table 2.
Treatment Characteristics of Children in Study Cohort at the Time of Initial Cerebrospinal Fluid (CSF) Shunt Placement, Shunt Revision(s), and Initial CSF Shunt Infection, Stratified by Organism
| Treatment Characteristic | Staphylococcus aureus | Coagulase-Negative Staphylococcus | Gram-Negative Bacilli | Propionibacterium acnes |
|---|---|---|---|---|
| Initial CSF shunt placement a | n = 34 | n = 37 | n = 23 | n = 2b |
| Shunt Location (Proximal-Distal), n (%)b | ||||
| Ventriculoperitoneal | 34 (100%) | 34 (92%) | 21 (92%) | 1 (50%) |
| Cystoperitoneal | 0 (0%) | 3 (8%) | 0 (0%) | 0 (0%) |
| Complex:peritoneal | 0 (0%) | 0 (0%) | 1 (4%) | 0 (0%) |
| Subdural-peritoneal | 0 (0%) | 0 (0%) | 0 (0%) | 1 (50%) |
| Subdural-atrial | 0 (0%) | 0 (0%) | 1 (4%) | 0 (0%) |
| Ultrasound use, n (%) | 11 (32%) | 15 (41%) | 7 (30%) | 0 (0%) |
| Endoscope use, n (%) | 12 (35%) | 7 (19%) | 7 (30%) | 1 (50%) |
| Antibiotic-impregnated catheter component use, n (%)b | 6 (18%) | 3 (8%) | 12 (52%) | 0 (0%) |
| CSF Shunt Revision(s) Before Initial Infection a | ||||
| Revisions Before Initial Infection, n (%) | ||||
| Zero | 23 (49%) | 31 (60%) | 26 (70%) | 2 (22%) |
| One | 10 (21%) | 10 (19%) | 4 (11%) | 1 (11%) |
| Two or more | 14 (30%) | 11 (21%) | 7 (19%) | 2 (67%) |
| Shunt Location (Proximal-Distal), n (%)a,b | ||||
| Ventriculoperitoneal | 33 (97%) | 34 (92%) | 21 (92%) | 1 (50%) |
| Cystoperitoneal | 0 (0%) | 3 (8%) | 0 (0%) | 0 (0%) |
| Complex:peritoneal | 1 (3%) | 0 (0%) | 1 (4%) | 0 (0%) |
| Subdural-atrial | 0 (0%) | 0 (0%) | 1 (4%) | 0 (0%) |
| Subdural-peritoneal | 0 (0%) | 0 (0%) | 0 (0%) | 1 (50%) |
| Ultrasound use, n (%)a | 11 (32%) | 12 (32%) | 8 (35%) | 0 (0%) |
| Endoscope use, n (%)a | 11 (32%) | 6 (16%) | 6 (26%) | 1 (50%) |
| Antibiotic-impregnated catheter component use, n (%)a,b | 9 (26%) | 4 (11%) | 11 (48%) | 0 (0%) |
| Initial CSF Shunt Infection | n = 47 | n = 52 | n = 37 | n = 9 |
| Concordant IV antibiotic use, n (%) | 47 (100%) | 51 (98%) | 35 (95%) | 9 (100%) |
| and duration (days), mean (SD)b | 19 (12) | 15 (10) | 23 (13) | 19 (30) |
| Concordant and appropriate IV antibiotic use, n (%)b | 46 (98%) | 49 (94%) | 29 (78%) | 5 (56%) |
| and duration (days), mean (SD)b | 13 (8) | 14 (9) | 21 (15) | 29 (37) |
| Broad spectrum IV antibiotic use, n (%) | 46 (98%) | 52 (100%) | 37 (100%) | 9 (100%) |
| and duration (days), median (IQR)b | 4 (3–5) | 3 (2–6) | 16 (11–28) | 5 (1–18) |
| Use of IT antibiotics, n (%) | 33 (70%) | 38 (73%) | 22 (59%) | 6 (67%) |
| Use of rifampin, n (%)b | 18 (38%) | 14 (27%) | 0 (0%) | 1 (11%) |
| Surgical approach to treatment, n (%)b | ||||
| Full removal and EVD placement | 44 (94%) | 41 (78%) | 28 (76%) | 5 (56%) |
| Externalization | 0 (0%) | 1 (2%) | 1 (3%) | 1 (11%) |
| Failed externalization | 2 (4%) | 3 (6%) | 1 (3%) | 0 (0%) |
| One or more components revised | 0 (0%) | 3 (6%) | 0 (0%) | 2 (22%) |
| Shunt removed after infection treatment | 1 (2%) | 3 (6%) | 4 (10%) | 0 (0%) |
| Shunt left in place | 0 (0%) | 1 (2%) | 2 (5%) | 0 (0%) |
| Died during treatment | 0 (0%) | 0 (0%) | 1 (3%) | 1 (11%) |
| Time from first positive CSF culture to first infection surgery (days), median (IQR)c | 0.5 (−0.1 to 0.7) | 0.94 (−0.1 to 0.8) | 0.1 (−0.2 to 0.8) | −0.1 (−0.6 to 0.1) |
| Time between infection surgeries (days), median (IQR)b | 18 (12–21) | 14 (11–18) | 24 (16–30) | 15 (8–16) |
| Time from first negative CSF culture to final infection surgery (days), median (IQR)b | 13 (9–18) | 10 (9–15) | 18 (13–24) | 13 (7–16) |
| Complications, n (%)c | ||||
| Any medical | 4 (16%) | 2 (8%) | 5 (25%) | 1 (20%) |
| Any neurological/surgical | 9 (36%) | 7 (29%) | 6 (30%) | 0 (0%) |
| Shunt Location (Proximal-Distal), n (%) | ||||
| Ventriculoperitoneal | 39 (85%) | 41 (84%) | 30 (94%) | 5 (62%) |
| Ventriculoatrial | 3 (7%) | 2 (4%) | 1 (3%) | 1 (13%) |
| Complex:peritoneal | 1 (2%) | 1 (2%) | 1 (3%) | 0 (0%) |
| Cystoperitoneal | 0 (0%) | 3 (6%) | 0 (0%) | 0 (0%) |
| Ventriculopleural | 1 (2%) | 0 (0%) | 0 (0%) | 1 (13%) |
| Subdural-peritoneal | 0 (0%) | 1 (2%) | 0 (0%) | 1 (12%) |
| Complex:atrial | 1 (2%) | 1 (2%) | 0 (0%) | 0 (0%) |
| Lumbar-peritoneal | 1 (2%) | 0 (0%) | 0 (0%) | 0 (0%) |
| Ultrasound use, n (%)b | 13 (28%) | 13 (27%) | 16 (53%) | 1 (13%) |
| Endoscope use, n (%) | 20 (43%) | 15 (31%) | 6 (20%) | 4 (50%) |
| Antibiotic-impregnated catheter component use, n (%) | 14 (30%) | 11 (23%) | 12 (40%) | 2 (25%) |
| Interval CSF Shunt Revision(s) | ||||
| Revisions After Infection and Before Reinfection, n (%) | ||||
| Zero | 29 (62%) | 31 (60%) | 24 (65%) | 7 (78%) |
| One | 11 (23%) | 14 (27%) | 8 (22%) | 1 (11%) |
| Two or more | 7 (15%) | 7 (13%) | 5 (13%) | 1 (11%) |
| Last known shunt location, n (%)b | ||||
| Ventriculoperitoneal | 39 (85%) | 41 (84%) | 27 (84%) | 5 (63%) |
| Ventriculoatrial | 3 (7%) | 4 (8%) | 1 (3%) | 1 (12%) |
| Complex:peritoneal | 2 (4%) | 0 (0%) | 4 (13%) | 0 (0%) |
| Cystoperitoneal | 0 (0%) | 1 (2%) | 0 (0%) | 0 (0%) |
| Ventriculopleural | 1 (2%) | 0 (0%) | 0 (0%) | 1 (12%) |
| Subdural-peritoneal | 0 (0%) | 1 (3%) | 0 (0%) | 1 (13%) |
| Complex:atrial | 0 (0%) | 1 (2%) | 0 (0%) | 0 (0%) |
| IV ventricle-peritoneal | 0 (0%) | 1 (3%) | 0 (0%) | 0 (0%) |
| Cystoatrial | 1 (2%) | 0 (0%) | 0 (0%) | 0 (0%) |
| Ultrasound use, n (%) | 10 (22%) | 12 (25%) | 13 (43%) | 1 (13%) |
| Endoscope use, n (%) | 19 (41%) | 13 (27%) | 7 (23%) | 3 (38%) |
| Antibiotic-impregnated catheter component use, n (%) | 19 (41%) | 10 (20%) | 12 (40%) | 3 (38%) |
| Reinfection | 6 (13%) | 6 (12%) | 3 (8%) | 3 (33%) |
Abbreviations: EVD, external ventricular drain; IQR, interquartile range; IT, intrathecal; IV, intravenous; SD, standard deviation.
aReported only for the prospective cohort of 96 children.
bSignificant differences seen between groups (P < .10) in Fisher’s exact tests, F-tests, or Kruskal-Wallis tests; P acnes excluded from comparisons at initial shunt placement and CSF shunt revisions before infection due to small sample size.
cTime from first positive culture to first infection surgery is reported for 141 children who had surgery; time between infection surgeries is reported for 127 children who had 2 surgeries; time from first negative culture to final infection surgery is reported for 124 children who has a negative culture and a second surgery; and complications reported for 74 children in the cohort.
Factors considered during the treatment of first CSF shunt infection (Table 2) included IV antibiotic selection and duration [24], characterized after review by 2 authors (T.D.S. and M.P.K.) as follows: (1) concordant, defined by the start of IV antibiotic(s) to which the organism(s) grown in CSF culture was sensitive; (2) concordant and appropriate, defined by the start of concordant IV antibiotic at a dose appropriate for weight and with adequate central nervous system penetration; and (3) broad spectrum, defined by the start of IV antibiotic(s) whose spectrum of activity included Gram-positive and Gram-negative organisms. When not specifically provided, susceptibilities were anticipated when possible based on the organism (eg, a methicillin-susceptible S aureus isolate was presumed to be susceptible to cefazolin even if not tested directly against it). Drug levels were not available for review in determination of concordant and appropriate IV antibiotic use. 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 because susceptibility data were not available. Surgical approach to treatment for CSF shunt infection was defined as either (1) total shunt removal with external ventricular drain (EVD) placement followed by new shunt insertion once the CSF was sterile; (2) distal shunt externalization followed by shunt replacement (externalization); (3) distal shunt externalization followed by EVD placement followed by new shunt insertion (failed externalization); or more rarely, (4) shunt revision, (5) shunt removal without replacement, and (6) no surgical treatment (shunt left in place) [20]. Duration of time between infection surgeries denotes the time span between the first and final surgical approaches to infection treatment (eg, shunt removal and EVD placement as a first surgery and shunt replacement as a second surgery). We also examined (1) duration of time from first positive culture to first infection surgery and (2) the duration of time from first persistently negative culture to final surgical approach to infection. Finally, we considered complications after first shunt infection, which were classified as either medical (eg, urinary tract infection, sepsis, cardiac arrest, pressure sores, deep venous thrombosis, pulmonary embolus, ascites, hyponatremia) or neurological/surgical (eg, CSF leak, wound infection, seizure, postoperative hemorrhage, over-/underdrainage, intracranial fluid collection).
Factors considered during the treatment of first CSF shunt infection and interval CSF shunt revision(s) included shunt location (ie, proximal-distal), ultrasound use, endoscope use, and antibiotic-impregnated catheter use (Table 2). We also considered reinfection, which was defined using the HCRN consensus definition of infection above and for which the duration of follow-up for reinfection was through December 31, 2013, resulting in a minimum follow-up duration of 1 year.
Microbiological and diagnostic factors in CSF shunt infection that we considered (Table 3) included the following: the presence of intermittent negative CSF cultures (defined as 2 or more positive CSF cultures with the same organism and at least 1 intervening negative culture), presence of secondary ventriculitis (recovery of a different organism from that recovered initially in CSF culture), bacteremia (defined as any organism recovered in blood culture, except for children with ventriculoatrial shunts diagnosed by blood culture), the absence of subsequent negative CSF cultures, duration of positive CSF cultures, and, where available within 48 hours of infection diagnosis, first CSF studies including Gram stain, white and red blood cell counts, glucose, and protein [23].
Table 3.
Microbiological and Diagnostic Factors of Children in Study Cohort at the Time of Initial Cerebrospinal Fluid Shunt Infection, Stratified by Organism
| Factor | Staphylococcus aureus (n = 52) | Coagulase-Negative Staphylococcus (n = 47) | Gram-Negative Bacilli (n = 37) | Propionibacterium acnes (n = 9) |
|---|---|---|---|---|
| Intermittent negative CSF cultures, n (%) | 9 (19%) | 7 (14%) | 5 (14%) | 0 (0%) |
| Secondary ventriculitis | 2 (4%) | 5 (10%) | 5 (14%) | 0 (0%) |
| Bacteremiaa | 7 (15%) | 0 (0%) | 5 (14%) | 1 (11%) |
| No subsequent negative CSF cultures, n (%)a | 0 (0%) | 2 (4%) | 2 (22%) | 0 (0%) |
| Duration of positive CSF cultures, median (IQR)b | 2.7 (1.8–5.5) | 2.4 (1.0–3.5) | 3.8 (1.3–8.0) | 2.1 (0.9–4.3) |
| Initial positive CSF Gram stain, n (%)a,c | 31 (66%) | 31 (60%) | 16 (43%) | 1 (12%) |
| Initial CSF white blood cell count, median (IQR)a,c | 55 (9–331) | 86 (14–299) | 163 (15–777) | 2 (1–37) |
| Initial CSF red blood cell count, median (IQR)a,c | 39 (4–405) | 29 (3–662) | 537 (45–3689) | 146 (3–3377) |
| Initial CSF glucose, median (IQR)a,c | 50 (28–58) | 41 (21–60) | 22 (20–49) | 65 (51–66) |
| Initial CSF protein, median (IQR)a,c | 103 (21–407) | 122 (30–447) | 235 (62–561) | 24 (11–123) |
Abbreviations: CSF, cerebrospinal fluid; IQR, interquartile range.
aSignificant differences seen between groups (P < .10) in Fisher’s exact tests, F-tests, or Kruskal-Wallis tests.
bReported for 139 patients who had more than 1 CSF culture.
cCSF Gram stain available for all but 1 P acnes infection, CSF cell counts are available for 140 infections; CSF chemistries are available for 136 infections.
Statistical Analyses
Patient characteristics, surgical and treatment decisions at the time of initial shunt placement and subsequent revision surgeries, and microbiological and diagnostic factors in CSF shunt infection were summarized descriptively by infecting organism, as defined in the 2017 IDSA recommendations (S aureus, coagulase-negative Staphylococcus, Gram-negative bacilli, P acnes). Frequencies and percentages are reported for binary and categorical measures, and infecting organism subgroups were compared using Fisher’s exact test. Mean and standard deviation or median and interquartile range are reported for symmetric and asymmetric continuous variables, respectively; differences between infecting organism subgroups were tested using the F-test. The Kruskal-Wallis test was used for variables with significant distributional asymmetry. Observed rates of reinfection are reported for those patients whose shunts were in place after infection, by the infecting organism, and compared across infecting organisms using Fisher’s exact test. We report significant differences between groups when P < .10, chosen a priori to balance type I and type II error given the small subgroups. All statistical analyses were performed using SAS (version 9.2; SAS Institute, Cary, NC).
RESULTS
Of 3131 children in the HCRN registry during the study period, 233 had an initial CSF shunt infection, and 170 of these first CSF shunt infections were diagnosed by CSF culture (Figure 1). Eight infections were excluded from further consideration because treatment was appropriately prolonged due to bacteremia or secondary ventriculitis. Of the remaining 162 CSF shunt infections, another 17 (10%) were excluded because they were not addressed by the 2017 IDSA recommendations (including 6 with Enterococcus, 7 with other Gram-positive organisms, 3 with more than 1 infecting organism, and 1 with fungal infection). Therefore, the study population included 145 children whose infections were diagnosed by CSF culture and addressed by IDSA recommendations, including 47 (32%) with S aureus, 52 (36%) with coagulase-negative Staphylococcus, 37 (26%) with Gram-negative bacilli (including 9 with Pseudomonas), and 9 (6%) with P acnes.
Figure 1.
Study cohort.
Patient differences between infecting organism subgroups were observed for birth weight (Table 1). No additional differences in patient characteristics were seen between infecting organism subgroups, including age at initial shunt, gestational age, gender, race, ethnicity, insurance, weight at initial surgery, indication for shunt placement, complex chronic conditions, gastrostomy, and tracheostomy.
The proportion of infections due to Gram-negative bacilli was higher when antibiotic-impregnated catheters were used at initial shunt placement (12 with prior antibiotic-impregnated catheter use out of 23 Gram-negative bacilli infections, 52%) and/or at subsequent revisions (11 of 23, 48%) compared with all other infections (9 of 73 [12%] and 24 of 73 [33%], respectively) (Table 2). No additional differences were seen in many treatment characteristics between infecting organism subgroups at initial CSF shunt placement and CSF shunt revision(s) before infection, including ultrasound and/or endoscope use, or in numbers of revisions before infection or surgical approach to infection treatment. Differences were observed between infecting organism subgroups at initial CSF shunt placement and CSF shunt revision(s) before infection for shunt location.
Differences were observed between infecting organism subgroups in infection treatment, including duration of concordant IV antibiotics, use and duration of concordant and appropriate IV antibiotics, duration of broad-spectrum IV antibiotics, use of rifampin, and ultrasound use. In general, and consistent with 2017 IDSA recommendations, more prolonged treatment was observed for Gram-negative bacilli infections. However, significantly less use of concordant and appropriate IV antibiotics for Gram-negative bacilli and P acnes infections was observed. No differences in reinfection were observed between infecting organism subgroups.
Differences between subgroups in CSF indices were observed and included Gram stain, white and red blood cell counts, glucose, and protein (Table 3). Lack of subsequent negative cultures and bacteremia was more commonly observed in infections with S aureus and Gram-negative bacilli. No other differences in microbiological and diagnostic characteristics were seen between infecting organism subgroups, including intermittent negative cultures, secondary ventriculitis, and duration of positive cultures.
Given the intriguing finding of differences in organisms after use of antibiotic-impregnated catheter, we have provided the antibiotic susceptibility profiles for all infection cases, divided into those with and without prior antibiotic catheter use (Figure 2).
Figure 2.
Antibiotic susceptibility profiles for all infection cases, with and without prior antibiotic-impregnated catheter (AIC) use. Lighter color denotes either equally split between sensitive and not sensitive, or that sensitivity was rarely tested.
DISCUSSION
This prospective cohort of 145 children whose infections were diagnosed by CSF culture and addressed by the 2017 IDSA recommendations allows us to make the most powerful comparisons between infecting organisms in pediatric CSF shunt infection to date. Few differences in patient and treatment characteristics were seen between infecting organism subgroups (S aureus, coagulase-negative Staphylococcus, Gram-negative bacilli, and P acnes). However, Gram-negative bacilli accounted for a larger proportion of infections when antibiotic-impregnated catheters had been used at initial shunt placement and/or subsequent revisions compared with all other infecting organisms. Expected differences were observed between infecting organism subgroups in infection treatment, with more prolonged treatment of Gram-negative bacilli infections.
The use of antibiotic-impregnated catheters is an increasingly widespread but controversial technique that involves the long-term use of tubing impregnated with minocycline or clindamycin combined with rifampin for the prevention of CSF shunt infections [28–36]. The only antibiotic-impregnated catheters available to HCRN sites are the clindamycin combined with rifampin type that provides predominantly Gram-positive coverage. Numerous small retrospective cohort studies have demonstrated no difference in infection rates between antibiotic-impregnated and conventional catheters [29–32, 35]. However, the largest existing observational study in 2010 in the United Kingdom demonstrated a reduction in infection rate (7.9% versus 5.2% infection rate per CSF shunt surgery), although authors observed wide variation between centers and “large overlap… in the results from different centers indicating uncertainty in the treatment effect estimates” [33]. The first systematic literature review in 2011 that pooled results from 5 pediatric studies also found a reduction in infection rate (11.2% versus 5.0% infection rate per CSF shunt surgery) [36]. More recently, a 2014 meta-analysis that pooled results for 6 studies found a reduction from 8.6% to 5.5% infection rate per CSF shunt surgery [35]. Authors concluded “the use of antibiotic impregnated shunt tubing may be associated with a lower risk of shunt infection” with a level III strength assigned to the recommendation, meaning an “unclear degree of clinical certainty” [35, 37]. Most recently, in February of 2017, IDSA published guidelines for shunt infections and recommended the use of antimicrobial-impregnated CSF shunts with a strong rating for the quality of the evidence and a moderate grade for the recommendation [13].
Although most studies of antibiotic-impregnated catheters have focused on infection rates, work examining the medium- and long-term outcomes after use of antibiotic-impregnated catheters is limited. Long-term use of antibiotic-impregnated catheters may have unintended consequences. One large single-center study reported a change in the infecting organism after introduction of antibiotic-impregnated catheters to include a lower proportion of coagulase-negative Staphylococcus but a higher proportion of the other infecting organism subgroups [34]. Other studies have reported a change in the resistance patterns of organisms causing CSF shunt infection after antibiotic-impregnated catheter use [33, 38]. This study’s finding among children with first CSF shunt infections of a higher proportion of infections due to Gram-negative bacilli infections with antibiotic-impregnated catheter use is new. Although antibiotic-impregnated catheters target Gram-positive organisms, our findings may simply represent a relative decrease in these infections (ie, a relative increase in Gram-negative infections among all infections). However, increasing proportions of Gram-negative infections may be concerning given that they require prolonged treatment [13]. Further investigation into the medium- and long-term outcomes including infecting organisms and antibiotic resistance after use of antibiotic-impregnated catheters is needed given that new guidelines recommend their routine use [13].
This work has several limitations. The conduct of this study at HCRN centers adhering to an infection prevention bundle [21, 27] reduces generalizability of findings outside the HCRN. However, the HCRN prevention bundle has been widely disseminated and implemented in the past 5 years. In addition, the multi-institutional nature of this study gives it greater generalizability than previous studies. Additional aspects of management of CSF shunt infection—including but not limited to the number, frequency, and method of CSF cultures obtained through EVDs, flushing of EVDs, and replacement of EVDs—is not standardized in the HCRN and was subject to variation within and between participating centers. Medical and surgical treatment decisions are likely tied to both patient characteristics and to infection and reinfection risk, and observational studies are limited in their ability to disentangle these complicated and multidimensional associations. Our definition of infection was developed by consensus within the HCRN and does not permit easy comparison to infection rates at non-HCRN centers or those used by hospital infection control groups. For neurosurgical procedures, the assumption is that patients return to the same center for care and therefore that subsequent infections would be captured in our data. Drug levels were not available for review in determination of concordant and appropriate IV antibiotic use.
CONCLUSIONS
Despite these limitations, our finding that a larger proportion of shunt infections are caused by Gram-negative bacilli when antibiotic-impregnated catheters are used warrants further investigation given increasing adoption of antibiotic-impregnated catheters.
Notes
Acknowledgments. We thank the contributing children and families at all participating centers. We also thank Stephan John Nemeth IV, Gabriel Finn Nemeth, and Daschel Simon Nemeth for support and valuable feedback.
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 funding sponsors. The study funders played no role in (1) study design; (2) the collection, analysis, and interpretation of data; (3) the writing of the report; and (4) the decision to submit the paper for publication.
Finanical support. T. D. S., the Nemeth family, and K. B. W. were supported by Award K23NS062900 from the National Institute of Neurological Disorders and Stroke (NINDS), Seattle Children’s Center for Clinical and Translational Research, and Clinical and Translational Science Awards Grant Number ULI RR025014 from the National Center for Research Resources, a component of the National Institutes of Health. D. D. L. was supported by grants from NINDS, Patient Centered Outcomes Research Institute, the Hydrocephalus Association, Rudy Schulte, and research funding through Medtronic and Karl Storz. The Hydrocephalus Clinical Research Network has been funded by private philanthropy, NINDS Grant 1RC1NS068943-01, a grant from The Gerber Foundation (1692–3638), Patient-Centered Outcomes Research Institute (CER-1403–13857), and the Hydrocephalus Association.
Potential conflicts of interest. All authors: No reported conflicts of interest. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest.
References
- 1. Kestle JR. Pediatric hydrocephalus: current management. Neurol Clin 2003; 21:883–95, vii. [DOI] [PubMed] [Google Scholar]
- 2. Kulkarni AV, Riva-Cambrin J, Butler J, et al. Outcomes of CSF shunting in children: comparison of Hydrocephalus Clinical Research Network cohort with historical controls: clinical article. J Neurosurg Pediatr 2013; 12:334–8. [DOI] [PubMed] [Google Scholar]
- 3. Piatt JH Jr, Carlson CV. A search for determinants of cerebrospinal fluid shunt survival: retrospective analysis of a 14-year institutional experience. Pediatr Neurosurg 1993; 19:233–41; discussion 242. [DOI] [PubMed] [Google Scholar]
- 4. Kulkarni AV, Drake JM, Lamberti-Pasculli M. Cerebrospinal fluid shunt infection: a prospective study of risk factors. J Neurosurg 2001; 94:195–201. [DOI] [PubMed] [Google Scholar]
- 5. McGirt MJ, Leveque JC, Wellons JC 3rd, et al. Cerebrospinal fluid shunt survival and etiology of failures: a seven-year institutional experience. Pediatr Neurosurg 2002; 36:248–55. [DOI] [PubMed] [Google Scholar]
- 6. 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] [PubMed] [Google Scholar]
- 7. Ronan A, Hogg GG, Klug GL. Cerebrospinal fluid shunt infections in children. Pediatr Infect Dis J 1995; 14:782–6. [DOI] [PubMed] [Google Scholar]
- 8. Cochrane DD, Kestle J. Ventricular shunting for hydrocephalus in children: patients, procedures, surgeons and institutions in English Canada, 1989-2001. Eur J Pediatr Surg 2002; 12:S6–11. [DOI] [PubMed] [Google Scholar]
- 9. Gardner P, Leipzig T, Phillips P. Infections of central nervous system shunts. Med Clin North Am 1985; 69:297–314. [PubMed] [Google Scholar]
- 10. Gardner P, Leipzig TJ, Sadigh M. Infections of mechanical cerebrospinal fluid shunts. Curr Clin Top Infect Dis 1988; 9:185–214. [PubMed] [Google Scholar]
- 11. Kanev PM, Sheehan JM. Reflections on shunt infection. Pediatr Neurosurg 2003; 39:285–90. [DOI] [PubMed] [Google Scholar]
- 12. Tamber MS, Klimo P Jr, Mazzola CA, Flannery AM. Pediatric hydrocephalus: systematic literature review and evidence-based guidelines. Part 8: management of cerebrospinal fluid shunt infection. J Neurosurg Pediatr 2014; 14:60–71. [DOI] [PubMed] [Google Scholar]
- 13. Tunkel AR, Hasbun R, Bhimraj A, et al. 2017 Infectious Diseases Society of America’s Clinical Practice Guidelines for healthcare-associated ventriculitis and meningitis. Clin Infect Dis 2017; doi: 10.1093/cid/ciw861. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Tunkel AR, Hartman BJ, Kaplan SL, et al. Practice guidelines for the management of bacterial meningitis. Clin Infect Dis 2004; 39:1267–84. [DOI] [PubMed] [Google Scholar]
- 15. Kestle JR, Garton HJ, Whitehead WE, et al. Management of shunt infections: a multicenter pilot study. J Neurosurg 2006; 105:177–81. [DOI] [PubMed] [Google Scholar]
- 16. Kulkarni AV, Rabin D, Lamberti-Pasculli M, Drake JM. Repeat cerebrospinal fluid shunt infection in children. Pediatr Neurosurg 2001; 35:66–71. [DOI] [PubMed] [Google Scholar]
- 17. Simon TD, Mayer-Hamblett N, Whitlock KB, et al. Few patient, treatment, and diagnostic or microbiological factors, except complications and intermittent negative cerebrospinal fluid (CSF) cultures during first CSF shunt infection, are associated with reinfection. J Pediatric Infect Dis Soc 2014; 3:15–22. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. 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] [PubMed] [Google Scholar]
- 19. 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] [PubMed] [Google Scholar]
- 20. Simon TD, Hall M, Dean JM, et al. Reinfection following initial cerebrospinal fluid shunt infection. J Neurosurg Pediatr 2010; 6:277–85. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. 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] [PMC free article] [PubMed] [Google Scholar]
- 22. 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] [PMC free article] [PubMed] [Google Scholar]
- 23. 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] [PMC free article] [PubMed] [Google Scholar]
- 24. Simon TD, Kronman MP, Whitlock KB, et al. Variability in management of first cerebrospinal fluid shunt infection: a prospective multi-institutional observational cohort study. J Pediatr 2016; 179:185–191.e2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. 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] [PMC free article] [PubMed] [Google Scholar]
- 26. 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 Pediatr 2008; 1:131–7. [DOI] [PubMed] [Google Scholar]
- 27. Kestle JR, Holubkov R, Douglas Cochrane D, et al. A new Hydrocephalus Clinical Research Network protocol to reduce cerebrospinal fluid shunt infection. J Neurosurg Pediatr 2016; 391–6. [DOI] [PubMed] [Google Scholar]
- 28. Sciubba DM, Stuart RM, McGirt MJ, 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] [PubMed] [Google Scholar]
- 29. Aryan HE, Meltzer HS, Park MS, et al. Initial experience with antibiotic-impregnated silicone catheters for shunting of cerebrospinal fluid in children. Childs Nerv Syst 2005; 21:56–61. [DOI] [PubMed] [Google Scholar]
- 30. Kan P, Kestle J. Lack of efficacy of antibiotic-impregnated shunt systems in preventing shunt infections in children. Childs Nerv Syst 2007; 23:773–7. [DOI] [PubMed] [Google Scholar]
- 31. Eymann R, Chehab S, Strowitzki M, et al. Clinical and economic consequences of antibiotic-impregnated cerebrospinal fluid shunt catheters. J Neurosurg Pediatr 2008; 1:444–50. [DOI] [PubMed] [Google Scholar]
- 32. Hayhurst C, Cooke R, Williams D, et al. The impact of antibiotic-impregnated catheters on shunt infection in children and neonates. Childs Nerv Syst 2008; 24:557–62. [DOI] [PubMed] [Google Scholar]
- 33. Kandasamy J, Dwan K, Hartley JC, et al. Antibiotic-impregnated ventriculoperitoneal shunts–a multi-centre British paediatric neurosurgery group (BPNG) study using historical controls. Childs Nerv Syst 2011; 27:575–81. [DOI] [PubMed] [Google Scholar]
- 34. James G, Hartley JC, Morgan RD, Ternier J. Effect of introduction of antibiotic-impregnated shunt catheters on cerebrospinal fluid shunt infection in children: a large single-center retrospective study. J Neurosurg Pediatr 2014; 13:101–6. [DOI] [PubMed] [Google Scholar]
- 35. Klimo P Jr, Thompson CJ, Baird LC, Flannery AM. Pediatric hydrocephalus: systematic literature review and evidence-based guidelines. Part 7: antibiotic-impregnated shunt systems versus conventional shunts in children: a systematic review and meta-analysis. J Neurosurg Pediatr 2014; 14:53–9. [DOI] [PubMed] [Google Scholar]
- 36. Parker SL, Anderson WN, Lilienfeld S, et al. Cerebrospinal shunt infection in patients receiving antibiotic-impregnated versus standard shunts. J Neurosurg Pediatr 2011; 8:259–65. [DOI] [PubMed] [Google Scholar]
- 37. Robinson S. Editorial: reducing the risk of shunt malfunction: beacons of light. J Neurosurg Pediatr 2016; 17:379–81. [DOI] [PubMed] [Google Scholar]
- 38. Adams DJ, Rajnik M. Microbiology and treatment of cerebrospinal fluid shunt infections in children. Curr Infect Dis Rep 2014; 16:427. [DOI] [PubMed] [Google Scholar]