Elucidating the utility of neuro-navigation in reducing malposition rates in Ommaya reservoir placement: 23-year operative experience at the Louisiana State University
Correspondence Address: Source of Support: None, Conflict of Interest: None DOI: 10.4103/0028-3886.193798
Source of Support: None, Conflict of Interest: None
Objectives: Variability exists with the use of neuro-navigation in the placement of Ommaya reservoir. In the setting of recent healthcare reforms in the United States that are focused on cost-containment strategies, we discuss from our experience at the Louisiana State University, Shreveport, if the use of high-cost stealth-guided navigation technique reduces malposition rates over free-hand placement.
Keywords: Malposition; neuro-navigation; Ommaya reservoir; outcomes
Since its introduction in 1963 for the management of fungal meningitis, the Ommaya reservoir has been employed in the management of a wide array of pathologies including central nervous system (CNS) infections, as well as neoplastic and indolently benign cerebral lesions. In the recent years, intrathecal chemotherapy has become the standard of care for the treatment and in certain situations, prophylaxis of carcinomatous meningitis.,, While lumbar puncture is an acceptable means of initial access, several studies have shown inconsistent CNS drug levels, as well as fibrosis and scarring of the subarachnoid space with repeated application.,,
In most series, the implantation and usage of Ommaya reservoirs is associated with minimal complications, with the incidence of short-term complications requiring revisions historically ranging from 6% to 10.3%. These are most commonly attributed to catheter malposition, malfunction or hemorrhage.,,, Delayed complications in the form of reservoir exposure, meningitis (aseptic or bacterial), and direct neurotoxic effects of chemotherapeutic agents, can occur in up to 50% of these patients. Moreover, with increasing therapeutic administration of intrathecal chemoprophylaxis in the setting of hematologic malignancies, rates of device infection have increased significantly., Weiner et al., in a recently published series consisting of 28 patients who underwent reservoir placement using neuro-navigation, reported a 0% malposition rate. Despite the advent of stealth-guided neuro-navigation, Ommaya reservoir placement traditionally has been without the use of these advanced mapping techniques, because of its relative ease and low-risks associated with its placement. With the recent seismic changes in the US healthcare system that are directed on cost-containment initiatives without compromising healthcare delivery,, it is plausible that advocating the use of stealth-guided neuro-navigation for this relatively simplistic procedure would impose economic burden on the system, albeit with little benefit.
In an effort to explore the utility of neuro-navigation in the placement of Ommaya reservoirs as recently depicted by Weiner et al., we explored our 23-year operative experience at the Louisiana State University to evaluate if stealth guided neuro-navigation reduces the risk of catheter malposition as compared to the traditional placement of the reservoir without navigation. To the best of our knowledge, we present one of the largest series of Ommaya implantation.
Study protocol and patient population
A retrospective analysis of the cohort of 168 patients that underwent placement of Ommaya reservoir at the Louisiana State University Health Sciences Center, Shreveport between November 1991 through October 2014 was performed. The study protocol was approved by the Institutional Review Board for the Protection of Human Research Subjects under protocol # E08-068 and #H13-019.
Patient population and data-extraction
A total of 146 patients who underwent placement of Ommaya reservoir and met our inclusion criteria were included in the analyses. 22 patients who were lost to follow up, or had no or inadequate data on the patency of Ommaya reservoir, were excluded. Information on patient demographics, clinical and imaging characteristics for the patency of the reservoir, operative use of stealth neuro-navigation, and complications were extracted by reviewing electronic records and patient charts by two authors independently (AF and DC). Stringent quality measures were adopted to ensure the accuracy and reliability of the extracted data. Potential conflicts arising upon any variable were resolved by rechecking the data, and attaining a mutual consensus. Patients who had reservoir placement for various known indications were categorized into two groups, those who had undergone placement of Ommaya reservoir via the use of neuro-navigation and those without it.
Primary endpoint (outcome)
The primary outcome measure was to evaluate the differences in the rates of catheter malposition in patients who had placement of Ommaya reservoir with stealth guided neuro-navigation versus those without it.
Surgical technique and management
As per the protocol, all inpatients were evaluated for patency of ventricular anatomy as assessed preoperatively via plain computed tomographic (CT) scan of the head. In 48 cases, this included a high-resolution scan for the purposes of catheter placement under neuro-navigation. All patients were medicated with intravenous antibiotics within 60 minutes of the initial incision. In case of free-hand placement of the Ommaya reservoir, we used the Kocher point as reference, approximately 10 cm posterior to the nasion and 2-3 cm parasagittal. A standard horseshoe incision was centered over this point, typically over the non-dominant hemisphere. A single burr-hole was created and widened to accommodate the reservoir base. Catheter lengths were trimmed to length, based upon the preoperative localization of the foramen of Monro. The catheter was then advanced based upon anatomic landmarks of the ipsilateral medial canthus in the coronal plane and the external auditory meatus in the sagittal plane. When the advancing catheter hit the ventricular wall, it was further advanced without the stylet, for approximately 1 cm, using the standard technique for intraventricular catheter placement.
Location of the catheter tip within the ventricular space was confirmed intraoperatively by aspirating through the reservoir dome using a 25-gauge needle. The wound was copiously irrigated with antibiotic irrigation, followed by its closure. The galea was approximated with inverted, absorbable sutures and the skin was stapled. All patients underwent post-operative CT scan of the head to definitively assess the adequacy of catheter placement. Post-operative intravenous antibiotics were continued for a total of 3 post-operative doses. Surgical staples were removed between the post-operative day 10 and 14 depending on the condition of the wound.
In the cases utilizing neuronavigation, preoperative planning was performed with the Medtronic Stealth station (Medtronic Navigation, Louisville, CO) using 3-dimensional trajectory views following rigid cranial fixation. The entry point was localized to the same scalp incision and the ipsilateral foramen of Monro served as the target point. Following the craniotomy and opening of the dura, the navigation probe was then passed along the chosen trajectory and removed, with special attention to following the identical pathway. The premeasured catheter was attached to the reservoir and soft-passed into the created path.
Patients were cleared for intrathecal chemotherapy as early as post-operative day one, pending confirmation of appropriate catheter placement on post-operative CT scan images.
Comparison of categorical variables across patients who underwent Ommaya placement with and without neuro-navigation was performed via Pearson's chi-square test and Fisher's exact test, or independent samples t-test as appropriate for metric variables. For univariate analyses, alpha set at 0.01 was considered as an inclusion criteria for multivariate analyses. To elucidate the factors associated with catheter malposition in our cohort, we planned on constructing a multivariate binary logistic regression model while adjusting for patient demographics and clinical characteristics. All statistical tests were two-tailed, and a type I error less than 5% was considered statistically significant. All statistical analyses were performed using XLSTAT and SPSS version 22.0.
Overall, 146 patients that underwent implantation of Ommaya reservoirs between 1991 and late 2014 at our institute were included in the analyses. Of these, 48 (33%) patients underwent reservoir placement using stealth guided neuro-navigation, while the remaining 98 (67%) patients had free hand placement. The mean age of the cohort was 44.85 ± 15.05 years (Median: 45.58 years) and 45% were female. The major indications for placement included leukemia/lymphoma in 113 (77.4%), leptomeningeal disease in 21 (14.4%), and obstructive hydrocephalus/cyst draining in 12 (8.2%) patients. There were no statistical differences in the demographics and clinical characteristics across the two groups. A summary of patient demographics and outcomes are enumerated in [Table 1].
Twenty six (17.8%) patients experienced a complication during the course of follow up, occurring at a median interval of 594 days. Technical complications occurred in nine (6.2%) patients that included catheter malposition. 3 (6.3%) of these patients were in the neuronavigation group while 8 (8.2%) were in the non-neuronavigation group (P = 1.000). Interestingly, one patient in each of the two groups had a malpositioned catheter associated with intracerebral hemorrhage (2.1% versus 1.0%, P = 0.551). The rates of infections (8.3% versus 9.2%, P = 1.000) and hemorrhage (0% versus 3.1%; P = 0.551) were relatively lower in patients that had reservoir placement via neuronavigation in comparison to those with free hand placements.
As none of the demographic and clinical characteristics were significant in univariate analyses at alpha set at 0.01 for inclusion, we did not explore the associations in a multivariate regression model.
Ommaya reservoir implantation has historically been reserved for the management of late stage cancers involving leptomeningeal metastasis, a diagnosis shown to be universally fatal in majority of cases within an average time frame of four months. With a clinical scenario as dire as this, complication rates as high as 28% with an operative mortality up to 8.3%, could seem acceptable; however, with the increasing utilization of intrathecal therapy for adjunctive central nervous system (CNS) prophylaxis, we strive to make outcomes better.
Complication reporting in operative series involving Ommaya reservoir insertion has generally been divided into those events directly related to surgical intervention, typically termed “technical complications,” and delayed events due to repeated cerebrospinal fluid (CSF) access, including infections and catheter obstructions. The former have historically been reported in 6.5 to as many as 28% of cases,, 10, ,,, with the latter occurring in as many as 50%, depending upon the series. While catheter malposition and intracranial hemorrhage have been universally accepted as technical failures, there is some ambiguity within the literature regarding the amount of time after which a malfunctioning catheter or a positive CSF culture result would be classified as a surgical or a delayed complication.
Chamberlain et al., clearly delineated complications occurring within 30 days as technical failures in their report of 120 patients with leptomeningeal metastases, and found two patients with catheter malposition and four with infection, for an operative complication rate of 5%. In the current series, we have included the incidence of symptomatic and asymptomatic hemorrhage as a technical complication, as has previously reported by the Sandberg, Lishner and Perrin series.,, Utilizing this standard, we have demonstrated a 11% rate of technical complications, which is comparable to the previously reported rate of 9-10.3%. The breakdown in our series translates to 2.1% hemorrhage, 7.5% malposition, and 1.4% for both [Table 1].
Catheter malposition with and without navigation
Malposition of the catheter is the most consistently reported immediate complication with the placement of Ommaya reservoirs. Incidences have historically ranged from 2.7 to 30%.,,,, In [Table 2], we summarize the relevant studies with navigated placement of Ommaya reservoirs, and the rates of complications as reported in these series. Weiner et al., in a series of 28 patients with navigated Ommaya reservoir placement, reported no incidence of malposition, depicting the utility of brain navigation in the placement of these reservoirs. Takashi et al., in a larger cohort involving 85 patients, had only 1 malposition. Sampath et al., found a higher rate of mal-positioning in their series: 1 out of 18 Ommaya reservoir placement, translating to 5.5% attrition in positioning. In the present series, 7.5% of the total catheters required operative repositioning, without significant patient morbidity. In the majority of cases, revision surgery was required secondary to malfunction of the reservoir or placement of the catheter tip within the parenchyma, precluding administration of chemotherapeutic agents. However, it is not without precedent for malposition to lead to far more serious sequelae, ranging from asymptomatic intraparenchymal or intraventricular hemorrhage, seizures to even death in rare cases.
Prior to the routine usage of the CT scan images for confirmation of catheter position, the instillation of chemotherapeutic agents via poorly placed catheters led to a significant incidence of leukoencephalopathy.,, Although generally considered self-limited with either removal of the catheter or discontinuation of intrathecal chemotherapy, at least one case involving catheter placement terminating within the cerebral peduncle has been reported to result in a persistent tremor with a transient period of parkinsonian rigidity. Multiple methods have been developed over successive generations to minimize catheter malposition and have included intraoperative fluoroscopy with and without pneumoencephalography, endoscopy, ultrasound and frame-based and frameless CT-guided stereotaxy., 21, ,,,
Contemporary series utilizing stereotactic guidance have reported malposition rates in the range of 0-1.2%, as well as significantly diminished rates of other technical failures.,, Two of our patients in the current series experienced a malposition utilizing frameless stereotaxy. This was in the first years of application of the stealth technology in our hands, and has to do with the steep learning curve. In the later years, our experience with the use of neuro-navigation led to a decrease in malposition rates for Ommaya reservoir placed with frameless electromagnetic stereotaxy. The choice of operative navigation was based upon surgeon preference with CT-guided stereotaxy representing the vast minority of cases selected. This selection bias makes the determination of superiority difficult and serves as one of the limiting factor in our analyses.
Immediate postoperative intracranial hemorrhage has been reported in 1 to 2.8% of cases and ranges from an asymptomatic intraventricular hemorrhage to a life-threatening hematoma requiring surgical evacuation.,, Prognosis following these complications has been mixed with some series reporting complete resolution of symptoms , and others reporting significant mortality., Previous authors have typically attributed these complications to medical interventions causing either thrombocytopenia or coagulopathy. Chamberlain et al., reported no hemorrhagic complications in their series of 120 operations and attributed their success to “meticulous attention to preoperative coagulation parameters.” In the current series, 2.1% of patients experienced intracranial hemorrhage following their procedure, with one patient death. In one patient, there was both mal-positioning and intracerebral hemorrhage. No incidence of hemorrhage was observed in patients in the neuro-navigation group [Table 1].
Infection, including meningitis, ventriculitis or frank abscess, is second only to immediate catheter malposition, in terms of incidence, with contemporary rates reported in the range of 1.9 to 23%., 10, ,,, The offending organisms are typically gram-positives, most commonly Staphylococcus epidermidis. Contamination of the device occurs either at insertion, usually resulting in clinical infection within the first 30 days, or secondary to repeated access of the device, with one study demonstrating a 15% likelihood of infection within the first year. A breakdown in the sterile technique is typically to blame in the latter circumstances, despite the demonstrated lack of correlation between the frequency of reservoir puncture and the incidence of infection.,,
Some ambiguity exists within the literature regarding how best to define these infectious complications, with the authors being evenly divided between considering positive CSF culture alone,, or positive culture only, when the patient has associated symptoms of infection, including headache, fever or neurological deficit., Despite this, a consensus has been achieved with regard to the most appropriate treatment. Historically, the reservoir had been considered as a foreign body in cases of infection, necessitating its removal prior to CSF sterilization. However, recent literature strongly supports a trial of intravenous antibiotics to attempt clearance of infection prior to considering removal of the device,,, with several authors supporting a combination of systemic antibiotics with intrathecal antibiotic dosage., 9, ,,, 8.9% of the cases reported in the current investigation ultimately required explantation at a median interval of 190 days, which seems to compare favorably with previously published incidences of 1.7-8.2%.,,,,
Delayed catheter occlusion
Even in the circumstance of accurate initial catheter placement and successful CSF access, the incidence of subsequent malfunction is not insignificant. In the majority of cases, this complication has been limited to a one-way interruption of flow that continues to allow medication administration. It, however, limits CSF sampling. Chamberlain et al., reported a 5% incidence of delayed “unidirectional obstruction” that they attributed to unilateral ventricular collapse around the catheter. They reported that all cases allowed continuation of therapy and none required revision or removal. However, 6.5% of the patients in the series reported by Browne et al., required removal of their device secondary to complete malfunction at a range of two to forty weeks following implantation. Only one patient in the current series developed a catheter obstruction necessitating revision.
Explanations for the mechanisms leading to this complication have mostly centered on the technique of CSF access. In patients without hydrocephalus, an appropriately placed ventricular catheter is often in direct apposition with one or both ependymal surfaces of the ventricular horn. Repeated aggressive aspiration is seen as a potential source of unidirectional, and subsequently bidirectional, device failure. The technique for appropriate CSF access typically involves pumping the device once or twice to exchange the fluid in the reservoir with ventricular fluid and aspiration of volumes equal or less than the planned volume of medication to be administered. However, as a typical Ommaya reservoir is accessed anywhere from 29-86 times over the course of its meaningful usage, the possibility for technical failure is obvious.,,
The limitations governing the present single-institutional analyses are inherent to its retrospective nature, thereby posing selection bias. Not all patients that underwent Ommaya reservoir placement (total n = 168) were included in the analyses due to inadequate follow-up data, which could possibly affect the projected estimates. Although in comparison to other neurosurgical procedures, reservoir placement is a relatively simple procedure, the placement was not performed by a single surgeon, thereby posing some bias in malposition rates as confounded by surgical experience. As the learning curve for stealth-guided neuro-navigation is steep, it is plausible that the number of technical-related complications in the stealth group in the early years of stealth-navigation acquisition could possibly be attributed to the learning curve, rather purely by chance. For the years studied (1990-2014), possibility of clustering of outcomes within a particular 'year' or time period could falsely affect our estimates. Despite several limitations, our series on Ommaya reservoirs is one of the largest reported cohorts in literature.
Ommaya reservoir placement has been a safe and effective adjunct to the treatment of patients with metastatic cancer. Although placement of Ommaya reservoir is a relatively easier technique as compared to other neurosurgical procedures, various series have documented the use of stealth-guided neuro-navigation to affirm the accuracy of catheter tip placement in the ventricular system. Although we did not find any statistical difference in malposition rates between navigated and free-hand implantation of Ommaya, we believe that navigation, especially the frameless electromagnetic technique, could potentially be employed as a standard tool in making a clinical difference in a setting when learning curve has reached its plateau. With an increased cost of hospitalization associated with the use of stealth neuro-navigation, future studies are recommended to weigh the cost-benefit ratio of using the stealth-guided neuro-navigation technique for the placement of Ommaya reservoir over free-hand placement.
A portion of this manuscript was presented during the scientific session at the annual meeting of the Congress of Neurological Surgeons (CNS) held in October, 2010 at San Francisco, California.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
[Table 1], [Table 2]