Abstract
OBJECTIVE
The failure-free survival of ventriculoperitoneal shunts (VPSs) following externalization for distal catheter infection or malfunction has not been adequately explored. Conversion to a ventriculoatrial shunt (VAS) may allow earlier reinternalization in lieu of waiting for the peritoneum to be suitable for reimplantation. This option is tempered by historical concerns regarding high rates of VAS failure, and the risks of rare complications are rampant.
METHODS
In this retrospective cohort study, all patients undergoing externalization of a VPS at a single institution between 2005 and 2020 were grouped according to the new distal catheter terminus location at the time of reinternalization (VPS vs VAS). The primary outcomes were failure-free shunt survival and duration of shunt externalization. Secondary outcomes included early (< 6 months) shunt failure.
RESULTS
Among 36 patients, 43 shunt externalization procedures were performed. Shunts were reinternalized as VPSs in 25 cases and VASs in 18 cases. The median failure-free survival was 1002 (interquartile range [IQR] 161–3449) days for VPSs and 1163 (IQR 360–2927) days for VASs. There was no significant difference in shunt survival according to the new distal catheter terminus (log-rank, p = 0.73). Conversion to a VAS was not associated with shorter duration of shunt externalization (Wilcoxon rank-sum, p = 0.64); the median duration was 7 (IQR 5–11) days for VPSs and 8 (IQR 6–15) days for VASs. No rare complications occurred in the VAS group.
CONCLUSIONS
Shunt failure-free survival rates following externalization are similar to published survival rates for nonexternalized shunts. There was no significant difference in survival between reinternalized VPSs and VASs. Although the VAS was not associated with a shortened duration of externalization, this finding is confounded by strong institutional preference for the VPS over the VAS. Early conversion to the VAS may be a viable treatment option in light of reassuring modern VAS survival data.
Keywords: CSF diversion, externalize, shunt externalization, ventriculoatrial shunt, ventriculoperitoneal shunt, hydrocephalus
Abdominal complications of CSF shunting have existed since the development of the modern implantable ventriculoperitoneal shunt (VPS) in the mid-20th century.1 These complications typically result from mechanical failure or infection, or risk thereof, of the VPS system.
In 1985, McLaurin and Frame first described a “shunt externalization” procedure used in the management of shunt infections.2 Shunt externalization is a temporizing procedure in which the distal catheter is removed from the abdomen at a site along the catheter’s subcutaneous course at any point distal to the valve and allowed to drain to an external collection system. Therefore, continuous CSF diversion continues while the abdominal pathology is treated. As opposed to shunt removal and external ventricular drain (EVD) insertion, however, shunt externalization avoids the need to access or remove the proximal catheter or valve. This procedure has become a core component in the management of shunt infections,3 abdominal pseudocysts,4 and other abdominal complications.
In clinical practice, patients with externalized shunts who have complicated abdominal infections or hostile peritoneal environments often require protracted hospitalizations for shunt externalization and antibiotic therapy before the shunt is safely replaced or reinternalized into the abdomen.5 In the event that the abdomen is deemed unsuitable for continued CSF absorption, alternative distal catheter termini can be considered, including the right atrium. However, because of reports of myriad complications of ventriculoatrial shunts (VASs), by the 1970s the peritoneum was accepted as the far superior option, an opinion still widely held today.6
Recent evidence indicates that the shunt survival of VASs may be similar to that of VPSs,7 suggesting that VAS insertion may be a more viable reinternalization terminus after shunt externalization than current opinion holds. Moreover, VASs could feasibly be reinternalized without needing to wait for complete resolution of abdominal pathology. We hypothesized that after VPS externalization, conversion/reimplantation as a VAS is not associated with an increased failure rate and may allow for a shorter duration of externalization.
Methods
Study Design
A single-institution retrospective cohort study design was employed to test the study hypothesis. The study was approved by the local institutional review board (IRB no. 191689).
Study Population
Our institution maintains a database of pediatric (age < 18 years) patients who have undergone CSF diversion procedures, including CSF shunting, endoscopic third ventriculostomy, or revision thereof, from 2005 to the present. The electronic medical records of all patients included in this database were screened via free-text query for the terms “externalize” and “externalization,” which are standard institutional language in operative/procedure notes specifically referring to the explant of the distal VPS catheter to continue drainage externally. Records were reviewed to identify patients who underwent shunt externalization (rather than shunt removal and EVD placement). Patients undergoing externalization of a VAS or VPS were excluded.
Institutional Practice and Indications for Externalization
At our institution, shunt externalization is reserved for specific circumstances. While there have been subtle changes over the 15-year study period related to an increased focus on antibiotic stewardship and multidisciplinary care involving pediatric infectious disease specialists, the following practice paradigms have remained consistent over the last decade.
Importantly, shunt externalization is not considered if the patient has clear signs or symptoms of meningitis, cranial wound breakdown, or positive CSF Gram stain or cultures. In these cases, our institutional practice consists of removing the entire shunt and placing an EVD. The clinical presentation of a patient who may be a candidate for externalization is varied but broadly includes concern for infection or risk of infection distal to the valve. Typical indications include 1) symptomatic abdominal pseudocyst, 2) peritonitis (e.g., ruptured appendicitis), 3) catheter migration into a contaminated space or such that the shunt function is suboptimal (externalization indicated if emergently required), or 4) iatrogenic peritoneal contamination (liver transplant, bowel perforation during cholecystectomy, etc.). In these cases, our practice is to sample CSF via the shunt reservoir and then externalize the shunt as urgently as the patient’s neurological status dictates in the event of a distal malfunction related to pseudocyst or catheter migration. In the event of peritonitis or iatrogenic contamination, the shunt is externalized urgently to reduce the chance of ascending contamination. Broadspectrum antibiotics are initiated after CSF is sampled. If CSF cultures are positive, we proceed with shunt removal and EVD placement. Following externalization, repeat CSF cultures are routinely obtained on postoperative day 1 to monitor the CSF profile for signs of infection. If cultures of the spinal fluid show a positive result in a delayed fashion, the externalized shunt is removed and an EVD is inserted.
If the CSF cultures remain negative, the antibiotic regimen is tailored to the abdominal pathology. In the case of a sterile abdominal pseudocyst, antibiotics are narrowed to single-agent coverage with cefazolin when cultures are finalized as negative and discontinued when the shunt is reinternalized. Reinternalization may be as early as 3–4 days (based on negative culture finalization). However, a history of numerous abdominal operations or a large, loculated pseudocyst may prompt reimaging to assess for resolution prior to reimplantation. The medical center’s infectious disease service is typically not involved for cases of sterile abdominal pseudocyst.
In the other commonly occurring cases of peritonitis or iatrogenic peritoneal contamination, antibiotic management is based on the abdominal pathology, extent of contamination, and success, or lack thereof, of source control. In these cases, the pediatric infectious disease and pediatric general surgery teams assist in determining the need for additional abdominal imaging and the duration of antibiotics and advise on the feasibility of shunt reinternalization into the peritoneal cavity.
Traditionally, there has been an institutional bias to prioritize replacing the distal catheter into the peritoneum rather than the atrium. Therefore, in some cases the duration of externalization may have been extended to await recovery of the peritoneal cavity. Thus, some patients may have been otherwise medically ready for replacement of the distal catheter into the atrium, but the peritoneum may not have recovered. If the general surgery and/or infectious disease teams suggest that the peritoneum will not be sterile or hospitable for distal catheter reimplantation for an extended period of time, conversion to a VAS is discussed with these teams and the patient/family.
Externalization and Reinternalization Procedure Details
At our institution, shunt externalization is typically performed in the operating room (OR), except in extenuating circumstances when a patient is too unstable to transport to the OR or when a bedside procedure is expected to allow for more expedient treatment. Shunt externalizations were performed by one of five board-certified pediatric neurosurgeons with the assistance of neurosurgical residents, unless a nonpediatric neurosurgeon was on call and a procedure was performed overnight. Procedural details are similar to those described elsewhere.8 The shunt catheter is externalized distal to the valve mechanism, usually through a small (approximately 1 cm) skin incision below the clavicle. The distal portion of the catheter is pulled up and cut, allowing the aspiration of fluid from the distal catheter. This fluid and the distal portion of the catheter are sent for culture. The proximal, remaining portion of the externalized distal catheter is then connected to a sterile CSF collection system.
Shunt reimplantation was performed exclusively by pediatric neurosurgeons; assistance from pediatric general surgeons and/or vascular surgeons was sought at the discretion of the treating neurosurgeon. At the time of reinternalization, only the distal catheter was replaced unless there was a specific concern requiring removal of the entire prior shunt system and placement of a new system. From a technical standpoint, reinternalization was done by exposing the distal portion of the valve and disconnecting the distal catheter. At this point, a member of the anesthesia team pulled the distal catheter out from the externalization site. A new catheter was then tunneled to the neck or abdomen site of access to either the internal jugular/subclavian vein or peritoneal cavity, respectively. The catheter was then delivered to its final position (with fluoroscopic confirmation in a VAS and direct or laparoscopic visualization in a VPS) and connected to the valve.
Data Collection
Demographic, clinical, procedural, and outcome data were collected manually via review of the electronic medical record. Demographic variables included age (in days), sex, history of necrotizing enterocolitis, and any prior abdominal operations. The etiology of hydrocephalus was categorized using the nomenclature of the Hydrocephalus Research Network.9 The dates of shunt insertion, most recent shunt revision, and most recent revision of the distal catheter (i.e., excluding revisions only of the proximal catheter or valve) and the number of prior externalizations were recorded.
Indications for externalizations were recorded from the medical record. These were subsequently grouped into six categories. Abdominal pseudocysts were diagnosed by physical examination plus either abdominal ultrasound or CT of the abdomen and pelvis. Abdominal infections included abscess, ruptured appendix, intestinal perforation, and any other noniatrogenic peritoneal infectious process. Iatrogenic contaminations consisted of any presumed or confirmed infectious process in the peritoneal cavity following an abdominal procedure. Patients with distal catheter malfunction in the absence of kinking, discontinuity, abdominal infection, or pseudocyst were considered to have an untenable peritoneal environment. Distal catheters migrating outside of the peritoneal cavity (e.g., into the thorax, bowel, or abdominal wall) were categorized as catheter migrations. Indications that did not meet any of these criteria were recorded as “other.”
Culture data included speciated CSF cultures, distal catheter cultures (including distal catheter fluid aspirate and/or culture of the catheter tip), and any other culture data (i.e., blood, urine, sputum, tracheal aspirate, etc.). Antibiotic treatment plans were determined by review of notes and medication administration records. Infectious disease service consultation was recorded.
To address the study hypothesis, two primary outcomes were studied: shunt survival (freedom from failure) and duration of externalization. Shunt survival was defined as the time from shunt reinternalization to subsequent presentation with symptoms of shunt failure requiring revision, measured in days. Duration of externalization was defined as the interval between externalization and reinternalization in the OR, measured in days. The two study groups were defined according to the new distal catheter terminus utilized at the time of reimplantation (peritoneum vs atrium). A secondary shunt survival outcome of failure within 6 months of reimplantation was also recorded.
Statistical Methods
A threshold of statistical significance of α = 0.05 was defined a priori. Continuous variables are listed by median and interquartile range (IQR) unless stated otherwise. Differences in shunt survival were assessed using a log-rank test. The duration of shunt externalization was compared between the two groups using a Wilcoxon rank-sum test. Univariate analyses used Fisher exact and Wilcoxon rank-sum tests to assess for differences. Due to sample size limitations, no multivariable regression was performed. Analysis was performed in Stata/IC version 16.1 (StataCorp).
Results
Study Population
From our institutional database of 627 pediatric patients with surgically treated hydrocephalus, 38 patients were identified who underwent a shunt externalization procedure. From these, 1 patient with an externalized ventriculopleural shunt was excluded. Among the 37 remaining patients, 44 shunt externalizations were performed during the years 2005 to 2020. One patient died prior to reinternalization and was thus excluded from outcome analysis, leaving a final sample size of 43 cases. At the time of reinternalization, 25 distal catheters were inserted into the peritoneal space and 18 into the right atrium. The median follow-up duration after externalization was 533 days. The demographics of the study population are summarized in Table 1.
TABLE 1.
Demographics of the study population
| Factor | Distal Catheter Location | p Value | |
|---|---|---|---|
| Peritoneum | Atrium | ||
| No. of cases | 25 | 18 | |
| Age, mos | 50.00 (6.00–114.00) | 114.50 (38.00–173.00) | 0.092 |
| Sex | 0.54 | ||
| Female | 12 (48%) | 11 (61%) | |
| Male | 13 (52%) | 7 (39%) | |
| Necrotizing enterocolitis | 1 (4%) | 1 (6%) | >0.99 |
| No. of prior abdominal ops | 2.00 (1.00–4.00) | 3.50 (1.00–4.00) | 0.31 |
| Days since previous abdominal op | 312.00 (32.00–882.00) | 85.00 (12.00–1095.00) | 0.84 |
Values are presented as number (%) of cases or median (IQR) unless otherwise indicated.
Shunt History
The hydrocephalus and CSF shunt history of the study population is listed in Table 2. No significant differences existed between the two groups as defined by reimplantation terminus. Myelomeningocele was the most common etiology of hydrocephalus, followed by intraventricular hemorrhage of prematurity.
TABLE 2.
Previous CSF shunt history
| Factor | Distal Catheter Location | p Value | |
|---|---|---|---|
| Peritoneum | Atrium | ||
| No. of cases | 25 | 18 | |
| Hydrocephalus etiology | 0.89 | ||
| Post-IVH | 7 (28%) | 5 (28%) | |
| Myelomeningocele | 8 (32%) | 8 (44%) | |
| Post–head injury | 2 (8%) | 0 (0%) | |
| Encephalocele | 1 (4%) | 0 (0%) | |
| Posterior fossa cyst (DWM, variants) | 1 (4%) | 2 (11%) | |
| Large interhemispheric arachnoid cyst | 1 (4%) | 0 (0%) | |
| Communicating congenital hydrocephalus | 4 (16%) | 2 (11%) | |
| Other | 1 (4%) | 1 (6%) | |
| Age of distal catheter (days) | 358.00 (37.00–1273.00) | 608.00 (16.00–2053.00) | 0.68 |
| No. of prior shunt revisions | 2.00 (0.00–4.00) | 2.50 (0.00–4.00) | 0.92 |
| No. of prior externalizations | >0.99 | ||
| 0 | 20 (80%) | 16 (89%) | |
| 1 | 3 (12%) | 2 (11%) | |
| 2 | 1 (4%) | 0 (0%) | |
| 3 | 1 (4%) | 0 (0%) | |
DWM = Dandy-Walker malformation; IVH = intraventricular hemorrhage.
Values are presented as number (%) of cases or median (IQR) unless otherwise indicated.
Externalization Data
Externalization procedural data as well as antibiotic and culture data are depicted in Tables 3 and 4, respectively. There were no major differences between the two study groups among all variables considered. Five patients in this cohort had multiple shunt externalizations: 1 patient had 4 externalizations and 4 patients had 2 externalizations each, resulting in a total of 7 nonindex externalizations.
TABLE 3.
Shunt externalization data
| Factor | Distal Catheter Location | p Value | |
|---|---|---|---|
| Peritoneum | Atrium | ||
| No. of cases | 25 | 18 | |
| Indication for externalization | 0.11 | ||
| Pseudocyst | 11 (44%) | 9 (50%) | |
| Abdominal infection* | 4 (16%) | 0 (0%) | |
| Iatrogenic peritoneal cavity contamination† | 3 (12%) | 5 (28%) | |
| Untenable peritoneal environment‡ | 1 (4%) | 3 (17%) | |
| Catheter migration§ | 4 (16%) | 0 (0%) | |
| Other¶ | 2 (8%) | 1 (6%) | |
| Bedside externalization | 2 (8%) | 1 (6%) | >0.99 |
| Simultaneous procedure** | 7 (28%) | 3 (17%) | 0.47 |
| Entire shunt system removal prior to internalization | 3 (12%) | 1 (6%) | 0.63 |
| Duration of externalization (days) | 7 (5–11) | 8 (6–15) | 0.64 |
| Pediatric general surgery assistance w/ reinternalization | 17 (68%) | 11 (61%) | 0.75 |
| Laparoscopic distal catheter replacement | 15 (60%) | NA | |
NA = not applicable.
Values are presented as number (%) of cases or median (IQR) unless otherwise indicated.
Two patients had abdominal abscesses, 1 had peritoneal infection secondary to severe pyelonephritis, and 1 had a ruptured appendix.
Peritoneal cavity contamination noted during/after urologic procedures (n = 3) and general surgery procedures (n = 5).
Distal catheter malfunctions without evidence of catheter kinking/discontinuity, abdominal pseudocyst, or peritonitis/infection. One patient was treated with broad-spectrum antibiotics under the assumption that a low-grade, culture-negative abdominal infection had caused multiple distal failures; this catheter was replaced into the peritoneum.
Migration out of peritoneal cavity into abdominal wall (n = 1), migration of catheter tip into thoracic cavity (n = 2), and erosion of distal catheter into rectum (n = 1).
Abdominal wound dehiscence and presumed distal catheter infection (n = 2) and large ovarian cyst initially thought to represent pseudocyst on imaging workup (n = 1).
Any general or urological surgical procedures performed during the same anesthetic exposure in the OR as the shunt externalization procedure.
TABLE 4.
Infectious workup, culture data, and antibiotic therapy
| Factor | Distal Catheter Location | p Value | |
|---|---|---|---|
| Peritoneum | Atrium | ||
| No. of cases | 25 | 18 | |
| Culture data | |||
| Positive CSF cultures | 1 (4%) | 1 (6%) | >0.99 |
| Corynebacterium accolens | 1 (100%) | 0 (0%) | |
| Mixed gram-positive species | 0 (0%) | 1 (100%) | |
| Positive distal catheter cultures | 4 (20%) | 4 (31%) | 0.68 |
| Mixed coagulase-negative staphylococcus | 0 (0%) | 1 (25%) | |
| Escherichia coli | 0 (0%) | 2 (50%) | |
| Enterococcus faecalis | 1 (25%) | 0 (0%) | |
| Enterobacter cloacae | 1 (25%) | 0 (0%) | |
| Haemophilus influenzae | 1 (25%) | 0 (0%) | |
| Trichophyton tonsurans | 1 (25%) | 0 (0%) | |
| Candida albicans | 0 (0%) | 1 (25%) | |
| Positive cultures (other) | 3 (13%)* | 4 (22%)† | 0.44 |
| Antibiotics started w/in 24 hrs | 23 (92%) | 18 (100%) | 0.50 |
| Planned antibiotic stop date | 0.26 | ||
| Prior to internalization | 2 (8%) | 3 (17%) | |
| After internalization, before discharge | 13 (52%) | 7 (39%) | |
| After discharge | 5 (20%) | 8 (44%) | |
| Missing | 5 (20%) | 0 (0%) | |
| Antibiotic-impregnated catheter at reinternalization | 21 (84%) | 12 (67%) | 0.28 |
| Antibiotic duration after externalization (days) | 13.00 (7.00–21.00) (n = 23) | 16.00 (8.00–20.00) (n = 17) | 0.53 |
| Infectious disease consultation | 8 (32%) | 7 (39%) | 0.75 |
Values are presented as number (%) of cases or median (IQR) unless otherwise indicated.
Abdominal culture with Trichophyton tonsurans; blood culture with Staphylococcus capitis; urine culture with Pseudomonas aeruginosa.
Abdominal culture with Proteus mirabilis, Klebsiella pneumoniae, and Escherichia coli; urine culture with Escherichia coli, urine culture with Pseudomonas aeruginosa, orthopedic external fixation wound culture with Enterobacter cloacae.
Four patients returned to the OR for complete removal of the remaining shunt system and EVD insertion (prior to reinternalization). Two of these patients presented with abdominal pseudocysts; underwent shunt taps with initially negative CSF Gram stain and reassuring cell count, glucose, and protein; and underwent uneventful externalization. However, each of these patients’ CSF cultures became positive in a delayed fashion (postoperative day 4 following shunt externalization): one patient’s cultures speciated to Corynebacterium accolens, and the other patient had both Staphylococcus hominis and Staphylococcus capitis. The externalized shunt system was removed and an EVD was placed for each patient.
A third patient had a prolonged duration of externalization following peritoneal cavity contamination from intestinal perforation. This patient developed a fever on postoperative day 52; a CSF sample drawn from the externalized shunt grew Staphylococcus epidermidis. The fourth patient underwent externalization for an abdominal pseudocyst but later had progressively declining output from the externalized shunt. Imaging workup demonstrated that the proximal catheter and valve mechanism had retracted, and thus the shunt system was no longer draining the ventricle.
Duration of Externalization
The median duration of externalization was 7 days for shunts reinternalized to the peritoneum and 8 days for those reinternalized into the atrium (Wilcoxon rank-sum, p = 0.64; Table 3). A box-and-whisker plot depicting the distribution of externalization durations by reinternalization terminus is shown in Fig. 1.
FIG. 1.
Box-and-whisker plot of duration of shunt externalization, according to ultimate distal catheter reimplantation location.
Shunt Survival
Outcomes following shunt reinternalization are shown in Table 5. The median survival durations of shunts reinternalized to the peritoneum and atrium were 1002 and 1163 days, respectively. There was no difference in shunt survival according to the reinternalization terminus (log-rank test, p = 0.73; Fig. 2). The secondary survival outcome of early (< 6 months) shunt failure was also not significantly different between the two groups (Fisher exact test, p = 0.58).
TABLE 5.
Patient outcomes following shunt reinternalization
| Factor | Peritoneum | Atrium | p Value* |
|---|---|---|---|
| No. of cases | 25 | 18 | |
| Follow-up duration (days) | 533 (91–1412) | 522 (148–934) | 0.62 |
| Subsequent shunt failure | 14 (56%) | 8 (44%) | 0.54 |
| Failure type | 0.91 | ||
| Proximal | 2 (14%) | 2 (25%) | |
| Distal (valve/distal catheter) | 8 (57%) | 4 (50%) | |
| Infection | 2 (14%) | 1 (13%) | |
| Other‡ | 2 (14%) | 1 (13%) | |
| Infection-associated wound breakdown | 0 (0%) | 1 (100%) | 0.33 |
| Shunt failure treatment | 0.047 | ||
| Shunt revision | 6 (24%) | 6 (33%) | |
| Re-externalization | 5 (20%) | 0 (0%) | |
| Removal & EVD placement | 3 (12%) | 1 (6%) | |
| Other | 0 (0%) | 2 (11 %)§ | |
| Shunt survival (days) | 1002 (161–3449) | 1163 (172–2927) | 0.73† |
| Early (<6 mos) shunt failure | 6 (27%) | 5 (29%) | 0.58 |
Values are presented as number (%) of cases or median (IQR) unless otherwise indicated.
Continuous variables were compared using the Wilcoxon rank-sum test and categorical variables using Fisher exact tests unless otherwise indicated
Log-rank test.
One patient underwent a shunt revision at an outside hospital and failure type was not available; the second patient’s distal VAS catheter fractured and the free tip was lodged in the right ventricle; the third patient had a wound dehiscence but no infection was identified.
VAS converted to VPS at the time of shunt failure.
FIG. 2.
Kaplan-Meier survival function of time to shunt failure following reimplantation.
The etiology of subsequent shunt failures was not different between the two groups, and distal failures were the most frequent type for both. Operative shunt revision was the most common treatment approach for both groups, although 5 of 14 failed peritoneal shunts underwent repeat externalization.
Discussion
In this study of outcomes following VP shunt externalization, we confirmed our study hypothesis that reinternalization as a VAS was not associated with decreased shunt survival. Duration of externalization, however, was not improved by conversion to a VAS.
Reinternalized Shunt Survival
Shunt externalization is an uncommon but fundamental technique in the management of distal shunt catheter malfunctions and/or infections.10 The unique clinical scenario of an externalized shunt necessitates an evaluation of risk factors for failure that may be distinct from those found in the general shunt population. Potential risk factors specific to externalized shunts include the potential for infection to travel along externalized shunt hardware; prolonged hospitalization, often in an ICU setting; and an abnormal peritoneal environment (which necessitated externalization to begin with).
The survival of reinternalized shunts in our study, however, was consistent with the published survival of all-comer shunt systems. The 1-month failure rate in our population was 12%, compared to 14% in the series of 353 patients described by McGirt et al.,11 and the 1-year failure rate was 32.9%, similar to Shah et al.’s rate of 27% in their national database study of 7399 patients.12 These consistencies would seem to suggest that externalized shunts have a failure hazard similar to that of the general shunt population.
VAS Risks
Historical concerns about VASs have included distal catheter migration into pulmonary circulation, pulmonary hypertension,13 bacteremia, shunt nephritis,14 and outgrowing of distal catheter length,15 among others. In our study population, however, none of these rare complications arose. The absence of these complications may be due to a truly decreased risk as a result of surgical techniques, antibiotic-impregnated catheters, and multiple other quality improvement initiatives promulgated over recent decades.3,16,17 In our population, for example, all patients with non–antibiotic-impregnated catheters underwent reinternalization before the year 2012. Alternatively, the low rate of these complications could be a result of our small sample size and/or insufficient follow-up. Shunt nephritis, for instance, can occur many years after shunt insertion.14 Patients were followed clinically for the development of these complications, but routine echocardiograms and renal function laboratory testing were not performed unless there was a clinical suspicion of a VAS complication.
The need for catheter lengthening following VAS placement is a frequent indication for revision.7,15 Although we do not routinely perform screening aimed at diagnosing a short catheter or catheter migration in an asymptomatic patient, no patients underwent revision, specifically, for a short distal catheter or migration. Overall, we found VASs had similar survival times to VPSs, and there were no VAS-specific complications to suggest that a VAS is not a reasonable option following externalization. Additionally, VASs deserve further consideration in this population as several patients required repeat externalization.
Externalization Duration
The second component of our primary study hypothesis was that conversion to a VAS would shorten the duration of shunt externalization. Theoretical benefits of shortened externalization duration might include 1) reduced risk of infection tracking along externalized hardware and 2) reduced length of ICU or hospital stay. In our population, however, the duration of externalization was not different among patients who subsequently had a shunt infection (data not shown, Wilcoxon rank-sum test, p = 0.71), so the potential risk of seeding infection from lengthy externalization seems low.
With respect to length of stay, our hospital protocol (instituted circa 2013) is for patients with externalized shunts to be monitored in the ICU. Patients in this study spent a median of 8 (IQR 5–15) days in the ICU, and 27 of 43 cases were discharged within 1 day after shunt reinternalization. Earlier reinternalization of the shunt would permit transfer out of the ICU to a regular hospital bed—or even discharge home—if there is no indication for inpatient hospitalization to treat any underlying abdominal pathology. Again, this difference was not appreciable in our study findings.
Rather, in our experience the duration of shunt externalization is determined by the abdominal pathology and/or need for lengthy antibiotic treatment. In theory, conversion to a VAS would not require waiting for completion of antibiotic therapy for abdominal pathology, but this was not borne out in this study.
Several confounding factors may explain the absence of a timing benefit for a VAS. First, our clinical practice is likely influenced in part by the 50-year tenet of VASs being an inferior option to VPSs.6,15,18,19 As such, conversion to a VAS is considered a last resort, often after a period of observation in hopes that the peritoneum would be deemed suitable. Only 5 patients had early (< 7 days) conversion to a VAS; 3 had poor peritoneal absorption, 1 had iatrogenic peritoneal contamination, and 1 had a sterile pseudocyst. Early conversion to a VAS, then, seems to have been reserved primarily for patients with noninfectious etiology of distal failure.
On the other hand, there was no difference in the indications for shunt externalization between the two reinternalization groups overall. If VASs were considered a better option in patients requiring long-term antibiotic treatment, then we would expect VASs to be more frequent in patients with nonsterile pseudocysts, abdominal infections, and iatrogenic bowel contamination. The need for a peripherally inserted central venous catheter (PICC) for long-term antibiotic therapy could, in theory, increase the risk of VAS infection, although this risk is estimated to be low.20
The decision of when to reinternalize the shunt is multifactorial, and input is usually sought from multiple medical teams. Readiness of the peritoneal cavity for replacement of a VPS is determined largely by the general surgery team, with the input of infectious disease specialists in the setting of any infection. The intensive care team provides input regarding any comorbid conditions and/or medical clearance for a surgical procedure. Finally, day of the week/weekend, availability of surgical staff, and OR vacancies may affect the ultimate timing of reinternalization, although we were unable to determine the impact of these variables on provider decision-making via retrospective chart review.
In this study, the potential timing benefit of conversion to a VAS seems to have been negated mostly by our practice pattern of favoring reinternalization into the abdomen. Given the retrospective nature of the current study and that the timing of reinternalization was subject to local institutional bias in favor of VPSs, we believe the lack of significant difference in length of externalization between the procedures should be interpreted as an open research question, and that there remains the possibility that VASs—in the right patient—may reduce the length of externalization compared to VPSs.
Implications for Practice
The findings of this study may help to inform patients and families at the time of shunt externalization and/or reinternalization that the expected rate of shunt failure is similar to the published rate among all shunt insertions/revisions.
Our findings—and those of Rymarczuk et al.7—suggest that a VAS has similar survival as a VPS and may be a more viable option than historical dogma would imply. Even for young patients likely to experience a subsequent shunt failure, the option exists to attempt to salvage the peritoneum and convert back to a VPS at a later date.21 This option would be ideal in situations where the only indication for inpatient or ICU stay is the presence of an externalized shunt.
Study Limitations
As noted above, the main limitations of this study are the small sample size and single-center design. The small sample size limits the power of the study to detect a possible small difference in survival between the two reinternalization sites, and thus the statistical equivalence of the procedure is unable to be demonstrated. In fact, demonstrating the noninferiority of reimplantation of VASs compared to VPSs—assuming a 5-year VPS failure rate of 50% and a noninferiority margin hazard ratio of 0.9—would require randomization of 5660 patients (with α = 0.05 and β = 0.20). Given the rarity of this clinical situation, a robust noninferiority study of this size is not feasible.
Another potential weakness of this study is that multiple shunt externalizations for the same patient were included. This was done intentionally, so that estimated failure rates would be broadly applicable (rather than only applicable to index shunt externalizations). Finally, although the median follow-up period for our study was over 1.5 years, some complications of VASs can occur after many years.
Future Directions
Pathologies warranting shunt externalization are rare; hence, traditional randomized trials are not logistically feasible to address unanswered questions about externalization. Assessment of a shunt reinternalization protocol that includes consideration of early VASs may be amenable to a multicenter stepped-wedge study or other small-sample trial design. A survey of varying practice patterns among multiple providers and institutions may offer initial insight into the relative indications and contraindications for conversion to a VAS, based on expert opinion. Last, new long-term follow-up studies of VASs are needed to determine updated risks of rare complications in the modern era.
Conclusions
Following shunt externalization, there was no significant difference in survival between patients with a reinternalized VPS and those with a VAS. Current clinical decision-making regarding conversion to a VAS does not result in shortened duration of shunt externalization. Early conversion to a VAS may be a viable option in patients with abdominal pathology requiring lengthy treatment.
Acknowledgments
REDCap, used for study data management, is supported by NCATS/NIH grant UL1 TR000445. P.D.K. is supported by a training grant from the National Cancer Institute of the National Institutes of Health under award number T32CA106183.
ABBREVIATIONS
- EVD
external ventricular drain
- IQR
interquartile range
- OR
operating room
- VAS
ventriculoatrial shunt
- VPS
ventriculoperitoneal shunt
Footnotes
Disclosures
The authors report no conflict of interest concerning the materials or methods used in this study or the findings specified in this paper.
References
- 1.Davidson RI. Peritoneal bypass in the treatment of hydrocephalus: historical review and abdominal complications. J Neurol Neurosurg Psychiatry. 1976;39(7):640–646. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.McLaurin RL, Frame PT. The role of shunt externalization in the management of shunt infections. Concepts Pediatr Neurosurg. 1985;6:133–146. [Google Scholar]
- 3.Tamber MS, Klimo P Jr, Mazzola CA, Flannery AM. Pediatric hydrocephalus: systematic literature review and evidencebased guidelines. Part 8: Management of cerebrospinal fluid shunt infection. J Neurosurg Pediatr. 2014;14(suppl 1):60–71. [DOI] [PubMed] [Google Scholar]
- 4.Erwood A, Rindler RS, Motiwala M, et al. Management of sterile abdominal pseudocysts related to ventriculoperitoneal shunts. J Neurosurg Pediatr. 2020;25(1):57–61. [DOI] [PubMed] [Google Scholar]
- 5.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;64(6):e34–e65. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Vernet O, Rilliet B. Late complications of ventriculoatrial or ventriculoperitoneal shunts. Lancet. 2001;358(9293):1569–1570. [DOI] [PubMed] [Google Scholar]
- 7.Rymarczuk GN, Keating RF, Coughlin DJ, et al. A comparison of ventriculoperitoneal and ventriculoatrial shunts in a population of 544 consecutive pediatric patients. Neurosurgery. 2020;87(1):80–85. [DOI] [PubMed] [Google Scholar]
- 8.Connolly ES Jr, McKhann GM II, Huang J, et al. , eds. Fundamentals of Operative Techniques in Neurosurgery. 2nd ed. Thieme; 2010. [Google Scholar]
- 9.Riva-Cambrin J, Kestle JRW, Holubkov R, et al. Risk factors for shunt malfunction in pediatric hydrocephalus: a multicenter prospective cohort study. J Neurosurg Pediatr. 2016;17(4):382–390. [DOI] [PubMed] [Google Scholar]
- 10.Browd SR, Gottfried ON, Ragel BT, Kestle JRW. Failure of cerebrospinal fluid shunts: part II: overdrainage, loculation, and abdominal complications. Pediatr Neurol. 2006;34(3):171–176. [DOI] [PubMed] [Google Scholar]
- 11.McGirt MJ, Leveque J-C, Wellons JC III, et al. Cerebrospinal fluid shunt survival and etiology of failures: a seven-year institutional experience. Pediatr Neurosurg. 2002;36(5):248–255. [DOI] [PubMed] [Google Scholar]
- 12.Shah SS, Hall M, Berry JG, et al. A multicenter study of factors influencing cerebrospinal fluid shunt survival in infants and children. Neurosurgery. 2009;64(6):E1206. [DOI] [PubMed] [Google Scholar]
- 13.Fundar T, Langmoen IA, Hovind KH. Fatal cardiopulmonary complications in children treated with ventriculoatrial shunts. Childs Nerv Syst. 1991;7(4):215–217. [DOI] [PubMed] [Google Scholar]
- 14.Haffner D, Schindera F, Aschoff A, et al. The clinical spectrum of shunt nephritis. Nephrol Dial Transplant. 1997;12(6):1143–1148. [DOI] [PubMed] [Google Scholar]
- 15.Vernet O, Campiche R, de Tribolet N. Long-term results after ventriculoatrial shunting in children. Childs Nerv Syst. 1993;9(5):253–255. [DOI] [PubMed] [Google Scholar]
- 16.Tomei KL. The evolution of cerebrospinal fluid shunts: advances in technology and technique. Pediatr Neurosurg. 2017;52(6):369–380. [DOI] [PubMed] [Google Scholar]
- 17.Flannery AM, Duhaime A-C, Tamber MS, Kemp J. Pediatric hydrocephalus: systematic literature review and evidencebased guidelines. Part 3: Endoscopic computer-assisted electromagnetic navigation and ultrasonography as technical adjuvants for shunt placement. J Neurosurg Pediatr. 2014;14(suppl 1):24–29. [DOI] [PubMed] [Google Scholar]
- 18.Olsen L, Frykberg T. Complications in the treatment of hydrocephalus in children. A comparison of ventriculoatrial and ventriculoperitoneal shunts in a 20-year material. Acta Paediatr Scand. 1983;72(3):385–390. [DOI] [PubMed] [Google Scholar]
- 19.Drucker MH, Vanek VW, Franco AA, et al. Thromboembolic complications of ventriculoatrial shunts. Surg Neurol. 1984; 22(5):444–448. [DOI] [PubMed] [Google Scholar]
- 20.White IK, Shaikh KA, Nyarenchi OM, et al. Analysis of the potential risk of central intravenous lines and/or total parenteral nutrition with ventriculoatrial shunts. Childs Nerv Syst. 2015;31(4): 563–568. [DOI] [PubMed] [Google Scholar]
- 21.Bhasin RR, Chen MK, Pincus DW. Salvaging the “lost peritoneum” after ventriculoatrial shunt failures. Childs Nerv Syst. 2007;23(5):483–486. [DOI] [PubMed] [Google Scholar]