Multiloculated hydrocephalus (MLH) is a condition in which patients have multiple, separated abnormal cerebrospinal fluid (CSF) collections with no communication between them. Uniloculated hydrocephalus is defined as cystic dilatation of one part of the normal ventricular system such as obstruction of one of the foramina of Monroe or an isolated fourth ventricle obstruction³⁾. MLH is also referred to as multilocular hydrocephalus, multiseptate hydrocephalus, polycystic hydrocephalus, ventricular compartmentalization, intraventricular septations and multicompartmental hydrocephalus^1,26).

Etiologies of MLH include complications of neonatal meningitis and germinal matrix hemorrhage in premature neonates. Other causes include post-shunt infection, overdrainage in shunted hydrocephalus, direct parenchymal injury inflicted by a catheter tip, head trauma, and intracranial surgery^2,5,7,10,14). Predisposing factors include low birth weight, prematurity, perinatal complications, and congenital malformation²⁶⁾. Although the full pathogenesis of MLH remains unclear, it is known that inflammation leads to subependymal gliosis, which produces glial tufts and septations that occlude the normal ventricle system. Septations also develop via organization of intraventricular exudate and debris from ventriculitis^9,22-24).

Despite technical advancements in pediatric neurosurgery, neurological outcomes are poor in these patients and the approach to this pathology remains problematic. The mortality rate has decreased as a result of improved neonatal care; however, psychomotor retardation is a very serious problem in these patients. About 90% of these patients have severe cognitive deficits and require assistance with activities of daily living^6,26).

Patients with MLH have individual anatomic complexity and CSF hydrodynamics. Therefore strategies for management of MLH are diverse and creativity is often required. Multimodal surgical options exist for the treatment of MLH, including the insertion of multiple shunts into each dilated CSF cavity, fenestration of septate membranes by open craniotomy and/or endoscopy, and combined approaches. However, the optimal treatment strategy has not yet been identified. In this review we will discuss surgical options for MLH, and the pros and cons of each surgical option will be discussed based on previously published reports.

Currently, treatment options for MLH are microsurgical fenestration of septate compartments by open craniotomy or endoscopy, shunt surgery in which multiple catheters are placed in multilobulated compartments, and a combination of these modalities²⁴⁾. However, considering the complexity of the disease and the fact that a majority of patients are neonates and infants, effective management of progressive MLH is very challenging.

Generally, a single shunt is not useful in MLH and a multimodal approach is needed. In the past the traditional surgical treatment of MLH was implantation of a multiple shunt system in each CSF-filled space. However, there are several disadvantages to the multiple shunt system. First of all, multiple catheters may not adequately drain all CSF-filled cavities. Partial or complete malfunction of each catheter system can lead to aggravation of other lobulated areas. In such cases it is difficult to identify which catheter in the shunt system is problematic, making management quite complicated. The risk of shunt infection also increases as the number of operative procedures and amount of implanted hardware increases. It is therefore preferable to avoid multiple shunt catheters if possible^4,16,22).

Fenestration by open craniotomy was first described by Rhoton and Gomez²⁰⁾ in 1972. A ten-week-old baby with rapid head growth was diagnosed with progressive multi-septate hydrocephalus due to previous E. coli meningitis. The hydrocephalus was controlled with a single shunt placed via a small craniotomy and transcallosal approach after direct wide excision of multi-septated membranes through a small cortical incision.

The role of endoscopic fenestration for the treatment of compartmentalized hydrocephalus was first described by Kleinhaus et al.¹¹⁾ in 1982. Endoscopic fenestration was performed in a patient with post-infectious hydrocephalus who required bilateral shunt catheters, and communication was established between the compartmentalized cyst and the ventricular system prior to placement of the indwelling shunt system. Endoscopic surgery has recently emerged as an option in the treatment of MLH, but these techniques were previously used in only limited situations because of the high mortality and morbidity. However, since Lewis et al.¹³⁾ first published a large series of patients who underwent endoscopic surgery for loculated hydrocephalus in 1995, endoscopic procedures are now used frequently for treatment of MLH worldwide.

Although fenestration of all compartments by open craniotomy or endoscopy is successful, a shunt is almost always required. Thus the goal of open microsurgery or endoscopy is to make multiple compartments into a single CSF compartment to allow for drainage via a single shunt catheter system (Fig. 1). So far, the best choice of treatment for MLH has not been determined. In the treatment of patients with MLH, the most important goals are to relieve intracranial hypertension and decrease morbidity by minimizing the number of operations and complications associated with multiple shunt catheters regardless of what methods are used.

Since the suggestion of open craniotomy by Rhoton and Gomez in 1972, this surgical approach has increased in popularity. Nida and colleagues¹⁵⁾ reported that ten patients with MLH were treated with open craniotomy. They performed transcallosal fenestration of intraventricular septation in six patients and multiple shunt procedures in the remaining four patients. In the open craniotomy group, the shunt revision rate decreased from a median 2.75 cases per year prior to fenestration to 0.25 cases per year following fenestration. This is in comparison to a median shunt revision rate of 0.55 cases per year in the group of patients who underwent the shunt procedure only without fenestration. However, the transcallosal approach has several risks such as venous infarction resulting from sacrifice of bridging veins during retraction of the brain, disconnection syndromes due to the incised corpus callosum, and damage to the fornix leading to memory impairment¹³⁾. Another study reported surgical fenestration through open craniotomy in 33 cases of MLH. The authors also used endoscopic fenestration in relatively simple pathologies such as uniloculated hydrocephalus, while they preferred open craniotomy for fenestration of MLH. Open craniotomy was performed via a posterior parietal approach and cortisectomy was made closest to the loculated ventricle for fenestration. Nineteen of 33 patients needed additional fenestration and the remaining 14 patients did not require additional surgery during a median 3.7 years of follow-up. In 25 of 33 patients (76%), only one ventricular catheter was required, two catheters in six and three catheters in one patient. One patient did not require a shunt system²²⁾.

Since the study of endoscopic fenestration for intraventricular cysts by Kleinhaus and colleagues in 1982, there were several reports detailing the endoscopic approach for MLH until the mid 1990s^12,17,18). However, most were case reports, and long-term results were unknown. In 1995 Lewis and colleagues¹³⁾ reported the first large series of patients who underwent endoscopic treatment (34 cases in total: 13 MLH and 21 uniloculated hydrocephalus). Six (46%) of the 13 patients with MLH required one or more additional endoscopic procedures. The shunt revision rate decreased from 3.04 cases per year to 0.25 cases after endoscopic fenestration for patients with MLH. Therefore, the authors proposed that endoscopic fenestration was a safe and less invasive technique as initial treatment for MLH. Nowosławska and associates¹⁶⁾ reported that there was a significant reduction in shunt revision rate and complications, especially shunt infection, after endoscopic treatment (47 patients) compared to conventional shunt operations (80 patients) in patients with complex compartmentalized hydrocephalus. Only thirteen of 80 patients who underwent conventional shunt techniques were successfully treated with implantation of a simple shunt system, whereas the shunt system was able to be simplified in 30 of the 47 endoscopically-treated patients. In addition, patients treated with endoscopy had better postoperative clinical status than the other group. Several other studies compared the outcome of endoscopic fenestration with multiple shunt catheters (shunt surgery only), and concluded that the use of endoscopic fenestration can reduce revision rates compared to shunt surgery alone. However, the majority of these reports did not discriminate between uniloculated hydrocephalus and MLH^3,4,6,13,26).

One study compared open craniotomy with endoscopic fenestration for the treatment of MLH. Twelve of a total of 25 patients underwent open craniotomy and the remaining 13 patients underwent endoscopic fenestration. Shunt-related procedures and shunt revision rate were not significantly different between the two groups. There was an average of 5.1±4.5 cases of subsequent shunt-related interventions per year in the open craniotomy group versus 3.1±4.0 cases per year in the endoscopy group (p-value=0.25). The median shunt revision rate for the open craniotomy group was 0.74 revisions per year versus 0.5 revisions per year in the endoscopy group (p-value=0.51). Advantages of endoscopic fenestration included decreased blood loss, shorter operative time, and shorter hospital stay. However, there were some limitations to that study including the fact that there were several surgeons, it was a retrospective analysis and there was bias associated with procedure selection according to disease severity. Furthermore, predetermined specific clinical and radiological criteria were not considered when selecting surgical options¹⁾.

Because the disease is very complex and patient conditions vary significantly in MLH, it is difficult to compare the usefulness of endoscopy with open craniotomy as initial surgical management. Currently, endoscopic fenestration tends to be performed more often as initial treatment and open craniotomy may be useful in patients requiring repeat endoscopic procedures. However, this is not because endoscopic procedures are superior, but because of their minimally invasive nature and decreased surgical morbidity and mortality. Therefore, surgical options should be carefully considered depending on the individual case.