Endoscopic third ventriculostomy in tubercular meningitis with hydrocephalus
Background: Endoscopic third ventriculostomy (ETV) is increasingly being used as an alternative treatment in tubercular meningitis (TBM) hydrocephalus. This study is aimed to evaluate the role of ETV in TBM hydrocephalus. Materials and Methods: This is a prospective study of 59 patients with TBM and obstructive hydrocephalus. The diagnosis was confirmed by a computed tomography scan and/or magnetic resonance imaging scan preoperatively. The procedure was performed using the standard technique or water jet dissection. Results: Three (5.1%) patients had blocked stoma, 31 (53%) had associated malnutrition, and 13 (22%) had complex hydrocephalus. Clinical improvement was seen in 34 (58%) after ETV and in 47 (80%) patients after ETV with lumber peritoneal shunt. Thirteen patients with patent stoma and complex hydrocephalus did not improve after ETV alone; an additional lumber peritoneal shunt was required. Clinical outcome was significantly better in good grade. Early recovery was observed in 81%. Results of ETV were better in patients without cisternal exudates, good nutritional status, thin and identifiable floor of third ventricle compared to cases with cisternal exudates, malnourished, thick and unidentifiable floor respectively, although the difference was statistically insignificant. There was no operative death. Three patients with normal ICP did not show any improvement. The radiological recovery after 3 weeks of surgery was 52%; follow-up ranged between 7 and 54 months. Six patients developed CSF leak. Conclusion: Endoscopic third ventriculostomy was safe and effective in TBM hydrocephalus. Complex hydrocephalus and associated cerebral infarcts were the major causes of failure to improve. Good results were observed in better grades.
Keywords: Cerebrospinal fluid shunt, endoscopic third ventriculostomy, hydrocephalus, tuberculous meningitis
Hydrocephalus is one of the common complication of tubercular meningitis (TBM)  and it could be communicating, obstructive, or a combination of both.  Endoscopic third ventriculostomy (ETV) is increasingly being used as an alternative treatment for hydrocephalus in tubercular meningitis. ,,,,, Endoscopic third ventriculostomy is technically difficult in infective pathologies due to thick and opaque floor of third ventricle which increases the risk of hemorrhage and other vascular complications. The aim of this study was to evaluate the role of ETV and its safety in patients with TBM and hydrocephalus.
This is a prospective study of 59 patients with TBM and obstructive hydrocephalus who underwent ETV in a tertiary referral centre during a 4-year period, July 2005 to June 2009. The study was approved by the institutional ethics committee. Written consent was obtained from the patients or their legal representative. A detailed historical and neurological evaluation was done in all the cases. Patients with obstructive hydrocephalus with triventriculomegaly and with features of raised intracranial pressure (ICP) were included. Patients with communicating hydrocephalus were excluded. The older version of the modified Vellore grading system (Glasgow coma score was not included in this grading) for clinical assessment was used, grade I being the best and grade IV the worst.  Preoperative computed tomography (CT) scans were done in all the patients. Due to resource constraints, magnetic resonance imaging (MRI) scans could be done in 47 patients only. CT scans revealed hydrocephalus in all the patients. CT ventricugraphy was done in 12 suspected ETV failure cases.
Endoscopic third ventriculostomy was performed using the standard procedure or water jet dissection [Figure 1]. The technique of water jet dissection was used in 39 patients with thick and opaque ventricular floor.  Stoma of 5 mm or more was created in all the cases. Multiple tubercles were noted on choroid plexus, lateral ventricle, and floor of third ventricle in five cases.
Endoscopic third ventriculostomy was considered clinically successful when anterior fontanel was depressed or flushed (at least 48 hours after the surgery) to the adjoining scalp (in infants) and/or improvement of clinical features of raised ICP without radiological deterioration. Endoscopic third ventriculostomy was considered a failure when there was persistent or progressive dilatation of ventricles with clinical features of persistently raised ICP. Patent stomas were diagnosed when good amount of dye was observed in basal cisterns while the absence of dye signified block stoma [Figure 2]. Stoma patency could be diagnosed in mid-saggital MR or coronal MR. 
The presence or absence of any cisternal exudates in interpeduncular cistern in the CT or during surgery were defined as cisternal exudates present or absent respectively. Thin or thick floors were classified when basilar artery or its branches were seen or not seen respectively through the floor. The presence of inflammation and or when membrane was tough for perforation was also categorized as a thick membrane. Floor of third ventricle was classified as identifiable or unidentifiable when both mamillary bodies and infundibular recesses were recognized or when only one or no structure was seen respectively. Serum albumin, hemoglobin, and mid-arm circumference were recorded in all the cases. Patients with normal serum albumin, hemoglobin, and mid-arm circumference were labeled as having a good nutritional status. Early or delayed improvement was defined when patient recovered within 3 weeks or after 3 weeks respectively. An acute phase of disease was defined when the patient presented within 3 weeks of diagnosis of TBM while presentation after 3 weeks was categorized into the chronic phase. Intracranial pressure monitoring was done in 16 patients who did not improve or deteriorated. A decrease in size of ventricle and or disappearance of peri-ventricular ooze was considered as radiological recovery. Complex hydrocephalus  (combination of obstructive and communicating hydrocephalus) was a postoperative diagnosis performed by CT ventriculogram or mid-saggital or coronal MRI scan.
Management of patients failed to improve after ETV
Repeated lumbar punctures were done to relieve raised pressure in patients who failed to improve after surgery and in patients with cerebrospinal fluid (CSF) leak. Lumber peritoneal shunts were done in patients with CSF leak and complex hydrocephalus with patent stoma who continued to have CSF leak or raised ICP after 3-5 lumber punctures in 10 days time
Follow-up and complications
All patients were followed up clinically monthly for 6 months and 3 monthly thereafter. All patients had CT and/or MRI scans at 3, 6, and 12 months after the surgery. Complications such as infection, CSF leak, and failure of procedure were assessed.
Follow-up ranged from 7 to 54 months
The data of the present study were analyzed using EPI Info (CDC, Atlanta, USA) and to observe association between various factors of TBM in relation to outcome. To check any linearity of the survival and delayed improvement or no improvement according to Vellore grades, chi-square for trend was applied. Linear regression analysis was used to assess the relationship of various predictor variables (age, nutritional status, cisternal exudates, acute or chronic of disease, thin or thick floor type, identifiable or unidentifiable floor, and stage of disease as modified Vellore grading) with survival as outcome.
Patients' age ranged between 6 months and 76 years, and 39 patients were aged 18 years or below, including 9 infants. There were 43 males. The symptom duration ranged between 5 days and 2 months. All patients had fever and vomiting. There were 21, 27, 7, 4 patients in modified Vellore grades 1, 2, 3, and 4 respectively. Floor was thin and transparent in 17 patients and good nutritional status was observed in 28 patients. Detailed patient demography is shown in [Table 1].
Of the 59 patients, 50 (84.7%) patients survived [Table 1]. Overall improvement after ETV and LP shunt was seen in 47 (80%) patients [Table 2]. In 34 (58%) patients the clinical recovery could be attributed to the ETV procedure alone. Improvement was early in 38 (81%) patients. Two (3.4%) patients remained in status quo while one (1.7%) patient deteriorated. Patients in better grades improved early while those in poor grade showed delayed recovery [Table 3]. A significant linear trend between increasing grade and delayed improvement was observed (χ2trend = 6.756, P = 0.00934). Outcome was significantly better in good grade compared to that in poor grade (P = 0.00379). Survival after ETV in the chronic phase was significantly better compared to that in the acute phase (χ2 = 4.47; P = 0.03443, OR =0.21) [Table 1]. Results of linear regression analysis showed that only Vellore grading was significant while remaining other variables did not show any significant association with survival [Table 4].
Complex hydrocephalus was observed in 13 (22%) patients. Of the 21 patients in the acute phase, 6 (28.6%) patients had complex hydrocephalus, while among 38 patients in the chronic phase, 7 (18.4%) had complex hydrocephalus.
Management of patients failed to improve after ETV
Patent and block stomas were observed in 47 and 3 patients respectively in ETV survivors. Lumbar peritoneal shunt was performed in 13 patients with complex hydrocephalus and patent stoma with persistently raised ICP even after repeated lumbar punctures. All these patients improved after the LP shunt. Repeat ETV was successful in all the three patients with blocked stoma. Detailed outcome after ETV alone and after ETV and LP shunt is shown in [Table 2].
Operative success rate and morbidity
Endoscopic third ventriculostomy was technically successful in all the patients. There were no intraoperative significant complications or any peroperative death. Six patients developed CSF leak. The leak stopped in four patients in 7 days time. Lumber peritoneal shunt was required in two patients with complex hydrocephalus and persistent leak.
The ventricle size did not decrease within 3 weeks in 48% patients. There was no correlation between radiological and clinical recovery. In most patients clinical recovery was earlier than the radiological recovery. In some patients there was clinical recovery without any radiological recovery. On the contrary, in few patients there was radiological recovery without any clinical recovery mainly due to associated ischemic infarcts.
Intra cranial pressure monitoring
Intracranial pressure monitoring was done in 16 patients; 3 of them had normal pressure. Thirteen patients had raised ICP and all had complex hydrocephalus. Three patients with normal ICP did not show any improvement (MRI of these patients showed multiple infarcts) which implies that the failure to improve does not always mean that the ETV is malfunctioning (stoma was patent).
Overall clinical improvement occurred in 80% patients and in 58% patients, the improvement could be attributed to ETV alone. Similar were the observations in the earlier studies. , Results of ETV were better in patients without cisternal exudates, good nutritional status, thin and identifiable floor of third ventricle compared to patients with cisternal exudates, malnourished, thick and unidentifiable floor respectively, although the difference was statistically insignificant. In a larger sample size this difference may be significant. Association of poor outcome with the presence of cisternal exudates has been documented.  Good outcome (87%) was observed with a thin to transparent third ventricle floor.  Adult patients faired slightly well compared to children in this study, although the difference was statistically insignificant. Earlier studies have not seen such an age effect on the outcome. , Clinical outcome was significantly better in good grade patients compared to the poor grade ones in our series similar to the observation by Chugh et al. In our study better grade patients improved early while poor grade patients showed delayed improvement. In the series by Singh et al.,  60% had early recovery and 17% had delayed recovery. Patients in the chronic phase of disease did better compared to those in the acute phase in our study. This could be due to high incidence of complex hydrocephalus in the acute phase (28.6%) of the disease compared to that in the chronic phase (18.4%).
Complex hydrocephalus was seen in 22% in the present series, similar to our earlier observations.  The success rated in this group of patients is poor when compared to congenital hydrocephalus with aqueductal stenosis. , It has been proposed that the higher risk of failure in the TBM group could be attributed to high chances of complex pathologies. ,,, Incidence of complex hydrocephalus was more in TBM compared to that in the congenital group.  Defective absorption in complex hydrocephalus could be temporary or permanent.  Lumbar peritoneal shunt is very effective in the management of complex hydrocephalus who do not improve after ETV. 
Associated malnutrition was common in the present series. Patients with better nutritional status had good ETV success rates compared to patients with malnutrition, although this difference was statistically insignificant. Jain et al.,  in his series of patients with VP shunt, found better results in patients with good nutritional status compared to those with poor nutritional status.  The radiological recovery rate, within 3 weeks, was only 52% in this series; similar to the observations (55.6%) in the series by Singh et al. A decrease in the size of ventricle occurs very slowly after ETV.
Endoscopic third ventriculostomy is technically difficult in postinfective and posthemorrhagic hydrocephalus.  Such patients can be safely managed by water jet dissection.  This study showed that ETV to be safe and effective in patients with TBM and obstructive hydrocephalus. Similar observations were made by Figaji et al. Failure to improve after ETV does not always mean that the ETV is malfunctioning. It could be related to the associated infarcts as seen in this study. Patent stoma does not necessarily mean that the procedure is effective in reducing ICP. We have seen that 13 patients with patent stoma in our study did not improve; all of them had persistent raised ICP because of complex hydrocephalus.
The authors are thankful to Mr Arvind Kavishwar, Bio statistician, and Mr. Satish Verma for their help in statistics and providing images respectively.
[Figure 1], [Figure 2]
[Table 1], [Table 2], [Table 3], [Table 4]