Abstract
Abstract
Introduction:
Ventriculoperitoneal shunts (VPSs) are the mainstay of treatment of hydrocephalus but frequently need revision. We sought to directly compare the impact of laparoscopic versus open peritoneal shunt revision on the need for subsequent VPS revisions in pediatric patients.
Materials and Methods:
A prospectively maintained, externally validated database of pediatric patients who underwent a first peritoneal VPS revision at a single center between 2008 and 2016 was reviewed. Outcomes, including subsequent revisions, shunt infections, operative time, and hospital stay between open and laparoscopic groups, were compared.
Results:
A total of 148 patients underwent a first peritoneal VPS revision during the time period—40 laparoscopically and 108 open—with no significant difference in age or gender between the groups. Operative time, length of stay after shunt revision, and shunt infection rates did not vary between laparoscopic versus open revisions. There was no significant difference between need for subsequent overall (peritoneal or ventricular) shunt revisions in the laparoscopic (20%) versus the open group (34%), P = .07. However, there were significantly fewer frequent peritoneal revisions in the laparoscopic group (3% versus 15%, P = .04).
Conclusions:
This first cohort analysis of laparoscopic versus open VPS revision in pediatric patients suggests that laparoscopic peritoneal VPS revision may reduce the rate of subsequent peritoneal revisions without increasing shunt infections or operative time in pediatric patients.
Introduction
V
Open minilaparotomy is the traditional operative approach for peritoneal catheter revision. Laparoscopy offers improved visualization of the entire abdominal cavity at the time of catheter placement as well as the potential to treat some of the causes of shunt malfunction, including adhesions, disconnected shunts, or cerebrospinal fluid (CSF) pseudocysts.11–14 For adult patients, a meta-analysis of studies comparing laparoscopic and open approaches for VPS insertion from 1995 to 2015 demonstrated that laparoscopic shunt surgery was associated with decreased operative time, decreased length of hospital stay, and decreased rates of subsequent peritoneal shunt revision. 14 Within the pediatric population, while case series have demonstrated the safety of laparoscopic VPS revision or placement, there have been no direct comparison studies between laparoscopic and open VPS revision.15–18
In this study, we compare two cohorts of patients who underwent first revision of the peritoneal portion of a VPS in an open or laparoscopic manner at a single institution. The aim of this study was to examine whether laparoscopic peritoneal revision influences the rate of subsequent need for revision or impacts on the rate of shunt infections in pediatric patients.
Materials and Methods
Ethical approval for this study was obtained for the Research Ethics Board of the Hospital for Sick Children, and analysis was conducted in accordance with the Ontario Personal Health Information Protection Act.
Patients who undergo VPS insertion or revision at our tertiary care pediatric hospital are prospectively entered into the Hydrocephalus Clinical Research Network (HCRN) database at the time of initial VPS insertion and at the time of any subsequent shunt manipulation or revision. Data from our center were obtained from this database and retrospectively reviewed. Patients who underwent peritoneal shunt revision at our single institution between January 2008 and December 2016 were identified. Details were extracted, including patient demographics, etiology, laparoscopic or open abdominal entry, perioperative and postoperative complications, shunt failures, and subsequent shunt revisions. Any details not entered available in the HCRN database were identified through electronic patient record review when necessary.
Patients were divided between those who underwent a first peritoneal VPS revision and those who underwent a second or subsequent revision. All patients had shunt failure diagnosed by a neurosurgeon on the basis of clinical assessment and radiologic evidence. All open minilaparotomy VPS revisions were performed by a pediatric neurosurgeon. All laparoscopic VPS revisions were performed jointly by a pediatric neurosurgeon and a pediatric general surgeon. The abdomen was entered via an open Hasson technique through a periumbilical incision. Pneumoperitoneum was established up to pressures of 8–12 mm of mercury (mmHg). An inspection of the abdominal cavity was undertaken, and adhesiolysis was performed when necessary. The most favorable location for the entry site of the shunt was selected, and the peritoneal end of the catheter was tunneled subcutaneously to this location. A stab incision was made and a needle followed by a peel-away sheath was inserted through the fascia using a Seldinger technique under direct visualization. The peritoneal end of the catheter was then placed in the pelvis and CSF flow was confirmed before closing the abdomen.
Our primary outcome was rate of subsequent VPS revisions, in particular subsequent peritoneal revisions. Secondary outcomes included shunt infection, operative time, length of stay, as well as perioperative and postoperative complications. All data were analyzed using SPSS software version 24 (IBM Corporation, New York, NY). Student's t-test and chi-squared analysis were used to compare groups. Statistical significance was defined by a P value <.05.
Results
First revisions
We identified 148 patients who underwent first peritoneal VPS revision: 40 patients through a laparoscopic approach and 108 patients through an open approach (Table 1). The median length of follow-up was 73 months (range: 8–103 months). Gender distribution was similar in both groups (60% male and 40% female) (Table 1). Age at first peritoneal revision was not statistically different between the groups (137 ± 67 months versus 115 ± 66 months, P = .07). Hydrocephalus etiologies included intraventricular hemorrhage secondary to prematurity, tumors, myelomeningocele, encephalocele, communicating congenital hydrocephalus, and aqueductal stenosis. These did not vary significantly between the laparoscopic or open groups (P = .83). Indications for revision included obstruction, disconnection or fracture, shunt migration, or underdrainage, and were not significantly different between the laparoscopic or open groups (P = .15). There had been concern that emergent cases may be less likely to have an available laparoscopic surgeon on hand. However, a similar number of the laparoscopic and open interventions were undertaken in an elective, emergent, and add-on capacity on the operative lists (P = .12, Table 1).
P
Demographics and indications for revision for patients who underwent laparoscopic or open peritoneal ventriculoperitoneal shunt revision.
ICH, intracerebral hemorrhage; IVH, intraventricular hemorrhage; SAH, subarachnoid hemorrhage; SD, standard deviation.
Of the 148 patients who underwent a first VPS peritoneal revision, 45 patients required a subsequent VPS revision for shunt failure, obstruction, disconnection, or infection—8 (20%) from the laparoscopic group and 37 (34%) from the open group (Table 2). The majority of revisions were indicated for revision of the ventricular portion and were required for shunt failure, obstruction, or disconnection. In total, 17 patients required peritoneal revisions (Table 2). There was a statistically significant lower incidence of subsequent peritoneal revisions in the patients whose first revision had been performed laparoscopically (n = 1, 3%) compared with open (n = 16, 15%), P = .04. Ten patients required a peritoneal revision within the first 6 months, an additional 6 patients within the first 2 years, and an additional 1 patient by the end of the follow-up period. The median time of a peritoneal revision was 2 months (range: 0–72 months) (Table 2). The 1 patient who required revision after a laparoscopic first revision underwent this at 1 month postoperatively. The median time to revision for patients who had undergone an initial open revision was 12.8 months. The incidence of postoperative shunt infections did not vary between laparoscopic (n = 3, 7%) or open (n = 10, 9%), P = .51 (Table 3).
NA, not applicable.
Perioperative complications for patients who underwent laparoscopic or open first peritoneal ventriculoperitoneal shunt revisions.
CSF, cerebrospinal fluid; DVT, deep vein thrombosis; PE, pulmonary embolism; SD, standard deviation.
There was no difference in operative time between laparoscopic (72 ± 27 minutes) or open (79 ± 31 minutes) revisions, P = .24 (Table 3) nor was there a difference in hospital length of stay following revision (6.6 days compared with 5.6 days, P = .73), respectively, for laparoscopic or open revisions. There were 13 perioperative complications in these 148 patients—3 in the laparoscopic group (8%) and 10 (9%) in the open group, P = .4 (Table 3). There were no bowel perforations, visceral injuries, intraperitoneal hemorrhages, or mortalities in either group. There were no complications related to pneumoperitoneum in the laparoscopic group. In the laparoscopic group, there was 1 patient with a ventricular hemorrhage, 1 patient with pseudomembranous collection, and 1 patient with a superficial wound infection. In the open group, there were two CSF leaks, 1 seizure, 1 patient with overdrainage, 1 patient with an intracranial fluid collection, 2 patients with transient hyponatremia, 1 patient with a urinary tract infection, and 1 with a pressure sore.
Second or subsequent revisions
We identified 88 patients who underwent second or subsequent peritoneal revisions between 2008 and 2016. Overall revision rates were not significantly different between the two groups. However, when specifically analyzing peritoneal revisions, there was a significantly greater proportion of peritoneal revisions in the open group (33%) compared with the laparoscopic group (6%), P = .05. Of note, the single revision in the laparoscopic group was in fact an entire shunt exchange that involved both the ventricular and peritoneal portions. Despite significantly longer operative times in the laparoscopic (112 ± 52 minutes) versus the open (74 ± 21 minutes) group, infection rates and hospital length of stay were similar. In these cohorts, there were no bowel perforations, visceral injuries, or intraperitoneal hemorrhages. One patient in the laparoscopic group had transient tachycardia possibly related to pneumoperitoneum. In the open group, there was one incidence of delayed bowel ischemia related to torsion of the bowel around a shunt lead point. There was one incidence of cardiorespiratory arrest and death in a patient in the open group, and an additional patient who experienced cardiorespiratory arrest with return of circulation. In the laparoscopic group, there was 1 patient with a pressure sore and 1 patient with a postoperative headache (Table 4).
Outcomes for patients who underwent a second or subsequent peritoneal ventriculoperitoneal shunt revision in a laparoscopic or open procedure.
CSF, cerebrospinal fluid; DVT, deep vein thrombosis; PE, pulmonary embolism; SD, standard deviation.
Discussion
For patients with hydrocephalus, VPSs are critical devices that often lead to a continuous relief of symptoms, but shunts can require revision or replacement secondary to malfunction or infection. We demonstrate that for pediatric patients who undergo first peritoneal VPS revision, a laparoscopic approach offers a statistically reduced incidence of subsequent peritoneal revisions without impacting on shunt infection rates or operative time. For patients undergoing second or subsequent revisions, while operative time is longer in the laparoscopic group, there is no difference in shunt infection rates and patients undergoing laparoscopic revision have a statistically lower incidence of subsequent revisions. Our experience suggests that for patients requiring a first or subsequent peritoneal VPS revision, the laparoscopic approach may improve patient outcomes.
Previous case series have demonstrated that laparoscopic VPS revision is safe in the pediatric setting.15–20 In a recent cohort study with 25 patients <1 year of age and weighing <5 kg, Soleman et al. 16 demonstrated that laparoscopy was safe in this relatively higher risk pediatric population with no postoperative complications secondary to intra-abdominal shunt placement. Another study demonstrated that even in pediatric patients with complex shunt revisions and surgical histories, there were no further shunt failures when they underwent laparoscopic revisions within 9 months of follow-up. 15 Our analysis identified minimal perioperative complications in pediatric patients who underwent laparoscopic VPS revision, in particular there were no complications that could be directly related to pneumoperitoneum, confirming the safety of the procedure even in direct comparison with patients who underwent an open procedure. In addition, given the potential for optimal visualization and potential to reduce the risk of bowel injury, we propose that the more optimal visualization achieved laparoscopically allows the surgical team to operate under direct vision and reduce the potential risk of injury. While we did not observe any immediate bowel injuries or visceral injuries in either the open or laparoscopic group, there was a delayed incidence of bowel ischemia secondary to torsion around the shunt in 1 patient in the open group that could possibly have been avoided if the revision had been performed under laparoscopic guidance.
Laparoscopic exploration permits a thorough inspection of the peritoneal cavity, lysis of adhesions or pseudomembranous cysts, as well as intentionally guided catheter placement. 21 In addition, in cases of catheter disconnection, laparoscopy enables the retrieval of this shunt, for which parents of patients undergoing VPS revision particularly advocate. 13 While laparoscopy offers these additional advantages over open placement or revision of a peritoneal VPS, our cohort analysis indicates that perhaps because of the intentional placement, laparoscopically revised VPSs importantly reduce the risk of need for subsequent (second or more) revisions. In a nation-wide cohort study in Ireland investigating shunt survival, it was demonstrated that 13% of shunts fail by 30 days, and by 1 year, 30% had failed. Following revision, subsequent shunt survival was even worse, with 21% of revisions failing at 30 days and 40% at 1 year. 5 This sets up a vicious cycle, where revisions only complicate and appear to hasten additional revisions. 22 Thus, the potential to reduce this rate of subsequent revisions to 3%, as demonstrated in this study by undertaking a first laparoscopic VPS revision, could have a significant impact on long-term outcomes for these patients.
The adult literature has indicated that laparoscopic VPS revision offers lower rates of failure, and particularly with reference to certain subsets of etiology, such as normal pressure hydrocephalus.22–24 They saw comparable operative times, no variation in complications, and lower rates of subsequent revision. 22 Some centers have advocated for laparoscopic assistance to become the gold standard within the adult VPS population.21,25 Hydrocephalus also represents a major problem in pediatrics and often occurs in the setting of complex patients with a multitude of other medical and surgical problems, thus complicating intra-abdominal catheter insertions. In this study, we have demonstrated that laparoscopic VPS revision does not have an impact on infection rates or operative time for a first revision, but has an impact on patient outcomes that are similar to some of the adult studies. Given the lack of direct comparative studies within the pediatric setting, we feel this is a valuable contribution to our understanding of VPS management in pediatric patients.
Limitations
This study is limited by its single center nature, and we look forward to future analysis of the HCRN database and other multicenter studies to evaluate whether laparoscopic VPS revision offers reduced rates of peritoneal shunt failure. However, as the first cohort study to directly compare open and laparoscopic revision, it offers a first step toward evaluating the two techniques in a comparative manner. The nonrandomized nature of this study leads to the potential risk that patients with more complex abdominal history were more likely to be offered a laparoscopic intervention with the expected increased safety of optimal visualization. However, despite this possible bias to place more complex patients in the laparoscopic group, we saw reduced risk of subsequent revisions in the laparoscopic group. A randomized trial would offset this risk of bias and may expose an increased difference between the groups. This cohort analysis is further limited by the small numbers within the individual subsets of patient etiology or indication for VPS revision. This limits additional subset analysis and makes it difficult to identify if there is a subset of patients who may achieve more benefit from laparoscopic revision, and whether there is a subset of patients who may achieve similar rates of revision or outcomes with an open procedure.
Conclusions
Laparoscopic VPS revision in pediatric patients offers a safe means of peritoneal VPS revision, which does not impact on shunt infection rates but appears to reduce the risk of subsequent need for peritoneal VPS revisions.
Footnotes
Acknowledgments
We appreciate the datapoints kindly provided by the Hydrocephalus Clinical Research Network. We acknowledge the patients whose data was analyzed, as well as the surgeons who performed the interventions.
Disclosure Statement
No competing financial interests exist.
