Introduction 

   Rebleeding and vasospasm are leading causes of mortality and morbidity after SAH.¹⁸⁾¹⁹⁾ Outcome following SAH has improved with advances in neurosurgical management including early surgery and the prevention of vasospasm. 
   Hydrocephalus is still a common sequela of aneurysmal SAH. Depending on the diagnostic criteria used and the time period surveyed, the incidence of post-SAH hydrocephalus has been reported to range from 6 to 67%.^{1)8)25)28)34)41)} 
   Although the neurological conditions of many patients with shunt-dependent hydrocephalus have been improved by a shunt operation, it has a potential risk of infection. Clinically, the patients who develop hydrocephalus after SAH have a worse prognosis than the others.^5)34)39) 
   There has been numerous studies to reduce shuntdependent hydrocephalus. Numerous factors have been reported to be potentially associated with the occurrence of hydrocephalus after SAH in the literature. Some of these factors include female sex, advanced age, hypertension, hyponatremia, Glasgow Coma Scale(GCS) score, the amount of subarachnoid blood visualized on computed topographic(CT) scan(Fisher's grade), IVH, vasospasm, location of aneurysm, rebleeding, and use of antifibrinolytic agent, focal ischemia.^13)34)38) 
   According to international cooperative study on timing of aneurysm surgery, the removal of cisternal blood clot can reduce the incidence of chronic hydrocephalus in the early operation group.¹⁸⁾¹⁹⁾ But, several reports about 'no significant difference' in the risk of hydrocephalus between coil embolization group and operation group have engendered a lot of controversy.¹⁴⁾²⁷⁾ 
   In this study, we analyzed statistically numerous factors to find the improvable factors. 

Materials and Methods 

   This study is based on 475 aneurysmal SAH, who were admitted to our hospital between January 1996 and January 2005. 
   Exclusion criteria included : 1) angiographically nonaneurysmal SAH; 2) the patients who did not undergo treatment(coil embolization, clipping, shunt only); 3) follow up loss within 3 months. 
   The patients were divided into two groups: shunt group (105 patients) as a treatment for shunt-dependent hydrocephalus and shunt-free group (370 patients). 
   Hydrocephalus was defined as development of ventricular enlargement on the basis of the third ventricular width and periventricular low density on CT scan at any time after admission, and as clinical manifestation such as mental deterioration, memory impairment, gait disturbance and urinary incontinence. Shunt-dependent hydrocephalus was defined as the patients who underwent a shunt operation for the relief of the clinical symptom in the patients met both radiological and clinical criteria. 
   Treatment modality was divided into surgical clipping and endovascular coil embolization. In surgical cases, clipping was performed by using the pterional approach, blood clots were removed under wide opening of sylvian cistern or basal cistern. We did not use thrombolytic agents into the subarachnoid space during an operation. 
   In the timing of surgery, we compared early operation (within 3days after onset) with delayed operation (after 4days from onset). 
   Furthermore, we studied the duration of drainage, daily amount of CSF drainage and total amount of CSF drainage in cases of EVD and lumbar drainage. 
   The location of ruptured aneurysm was divided into 5 groups; anterior communicating aneurysm (A-com), anterior cerebral artery aneurysm (ACA), middle meningeal artery (MCA), internal carotid artery (ICA), vertebrobasilar aneurysm. 
   The outcome was evaluated using the Glasgow outcome scale(GOS) after 3 months from onset. 
   Statistical analysis was performed using Chi-square test and multivariate regression. [SPSS 10.0] P value of less than 0.05 was considered significant. 

Results 

1. AgeandSex 
   Age of patients ranged from 13 to 84 years with a mean of 53.4 years. 10.3 percent(%) of patients under 40 years needed the shunt operation and 20.8%, 31.7%, 28.6% of fifties, sixties, over 70 years, each. The incidence of shuntdependent hydrocephalus increased as the age of patients increased (p<0.05) (Table 1). 
   There were 161 men and 314 women. There was no statistical significance (Table 1). 

2. Hypertension 
   One hundred ninety eight patients had a medical history of hypertension. Fifty two patients (26.3%) of those had more incidence compared with 53 patients (19.1%) of 277 patients. There was no statistical significance (Table 1). 

3. Hunt-Hess grade, Fisher grade and Outcome 
   The risk of shunt operation tended to increase as Hunt-Hess grade increased (p<0.001) (Table 1). 
   The incidence of shunt-dependent hydrocephalus was 12.5% of Fisher grade 1, 13.0% of Fisher grade 2, 22.4% of fisher grade 3 and 26.8% of Fisher grade 4. The incidence of shunt operation tended to increase as Fisher grade increased. But, there was no statistical significance (Table 1). 
   Forty nine patients expired, the mortality rate was 10.3%. Analyzing the patients except the expired patients, the shunt group had a lower GCS score, compared with the shunt-free group (p<0.01) (Table 1). 

4. IVH, rebleeding and vasospasm 
   IVH was combined in 179 patients (37.7%). Sixty two patients (34.6%) of those underwent shunt operation, compared with 14.5% of the patients without IVH. There was a significant difference (p<0.001) (Table 2). 
   Rebleeding before treatment occurred in 39 patients (8.2%). Sixteen patients(41.0%) of those underwent a shunt operation. There was a statistical difference, compared with 20.4% of the patients without rebleeding. (p<0.01) (Table 2) 
   Vasospasm occurred in 189 patients (39.8%). Fifty one patients (27.0%) of those underwent a shunt operation, compared with 18.9% of the patients without vasospasm. There was a statistical significance (p<0.05) (Table 2). 

5. Location of ruptured aneurysm 
   Patients with posterior circulation aneurysm had a significantly higher incidence (52.4%) of hydrocephalus, while patients with a middle cerebral aneurysm had a significantly lower incidence(11.4%) of hydrocephalus (p<0.001) (Table 3). 

6. Treatment method and timing of operation 
   Aneurysmal clipping was performed in 379 patients (79.8%). Coil embolization was performed in 92 patients (19.4%). In four patients, the treatment of the aneurysm itself was not performed, but only shunt operation was done. In clipping group, 86 patients(22.7%) underwent shunt operation and 15 patients (16.3%) of coil embolization group did. There was no statistical significance (Table 3). 
   In aneurysmal clipping cases, early operation showed significantly low incidence (18.3%), compared with 29.0% of delayed operation cases (p<0.05) (Table 3). 

7. Acute hydrocephalus, EVD and lumbar drainage 
   Seventy six patients(16.0%) presented acute hydrocephalus, 58 patients(76.3%) of those underwent a shunt operation. There was a statistical significance, compared with 11.8% of the patients without acute hydrocephalus (p<0.001) (Table 4). 
   EVD was performed in 71 patients(14.9%). Fifty one patients(71.8%) of those underwent a shunt operation, compared with 54 patients(13.4%) of the no EVD group. There was a statistical significance (p<0.001) (Table 4). 
   In study of the duration of EVD, the longer the duration was, the more the incidence of shunt-dependent hydrocephalus was developed (p<0.05). There was a statistical significance between the daily amount of EVD and the incidence of shunt-dependent hydrocephalus (p<0.01) (Table 5). 
   Lumbar drainage was performed in 90 patients(18.9%). Twenty eight patients(31.1%) of those underwent shunt operation, compared with 77 patients(20.0%) of the others. There was a statistical significance (p<0.05) (Table 4). 
   There was a statistical significance between the duration, total amount of lumbar drainage and the incidence of shuntdependent hydrocephalus (p<0.05) (Table 5). 

8. Multivariate analysis : relative risk among risk factors 
   We found the following were significantly related to shunt-dependent hydrocephalus : increasing age, high Hunt-Hess grade, high Fisher grade, IVH, vasospasm, posterior circulation site of aneurysm, EVD, lumbar drainage, acute hydrocephalus. The odd ratio in analysis of binary factors related to shunt-dependent hydrocephalus are shown in Table 6. 
   Multivariate analysis was performed using above variables to predict shunt-dependent hydrocephalus. There was a statistical significance between acute hydrocephalus, lumbar drainage, posterior circulation and the incidence of shuntdependent hydrocephalus (Table 7). 

Discussion 

   After the Bagley's first report of the hydrocephalus developed after SAH in 1928,³⁾ there has been several literatures about hydrocephalus as a common sequela of SAH. The frequency of hydrocephalus following SAH has been variously reported: 6%~67%.^{1)8)25)28)34)41)} The differences between studies most likely resulted from a difference in criteria used to diagnose hydrocephalus.⁴⁾ 
   Kostelijanetz attributed 90% of the intracranial pressure (ICP) increase after SAH to an increase in CSF outflow resistance.²⁶⁾ Fuhrmeister et al, reported CSF outflow resistance to be elevated up to threefold in patient with SAH.¹¹⁾ Although the CSF hydrodynamics become normalized in the majority of patients, physiological CSF outflow resistance may not occur until 40 to 50 days after SAH. In case of arachnoid scarring and blockage of arachnoid granulation, however, CSF outflow resistance can remain elevated and a state of chronic hydrocephalus may develop.¹⁴⁾ 
   Foltz and Ward divided the ventriculomegaly after SAH in two stages (acute, chronic).¹⁰⁾ Recently, Vale et al, divided hydrocephalus that appears after SAH into three stages : acute (0~3 days after SAH), subacute (4~13 days), chronic (≥14 days).³⁹⁾ Several theories have been proposed to explain the pathogenesis of hydrocephalus. There was a difference in the pathogenesis of acute and chronic hydrocephalus. In acute hydrocephalus, the theory is that the subarachnoid blood interferes with CSF circulation at the sylvian aqueduct,⁶⁾¹⁵⁾ foramen of Monro, basal cistern,⁶⁾ or diffuse throughout the subarachnoid space^1)7)12) and that both blood and protein inhibit the bulk flow of CSF at the arachnoid granulation.⁹⁾ In chronic hydrocephalus, the CSF flow is permanently impeded or absorption is permanently reduced due to fibrin deposition.³⁵⁾³⁷⁾ 
   Hydrocephalus may present the symptoms associated increased ICP clinically, and deteriorate the patient's condition gradually to need CSF shunting. The CSF shunt operation has complications, such as infection and shunt malfunction up to 30% in frequency. Hydrocephalus can causes poor outcome as well as the increase of medical expenses and hospital day.^5)34)39) For reducing shunt operation as well as hydrocephalus, There has been an amount of studies about many factors associated with hydrocephalus following SAH, but it has not reached to an agreement yet, reporting quite difference in each literature.^{8)13)24)31)34)39)} 
   In our study, we found that the following factors are significantly related to hydrocephalus to shunt-dependent hydrocephalus following SAH : increasing aging, Hunt-Hess grade at admission, Fisher grade, IVH, rebleeding, vasospasm, posterior circulation site of aneurysm, treatment modality, early operation, acute hydrocephalus, EVD and lumbar drainage. Let us analyze these factors separately. 
   Graff-Radford et al. reported that older patients are more likely to have a diffuse collection of subarachnoid Blood on CT scan. Other factors that may explain the relationship of age to hydrocephalus include decreased absorption of CSF, meningeal fibrosis with aging and increased ventricular size.¹³⁾ 
   Several clinical studies, including the International Cooperative Study on the Timing of Aneurysm Surgery, have identified the level of consciousness at admission as a risk factor for the later development of hydrocephalus.¹⁸⁾¹⁹⁾ 
   In recent study, Vale et al, reported that 20% of surviving patients after aneurysmal SAH underwent V-P shunt placement secondary to chronic hydrocephalus and patients with an initial Hunt-Hess grade of 3 or higher represented almost 90% of the group that developed chronic hydrocephalus and initial Fisher grade 4 represented almost 46% of this group.³⁹⁾ In our study, 22% of patients underwent V-P shunt operation, and patients with Hunt-Fess grade 3 or higher represented 79.0% of the shunt group, Fisher grade 4, 42.8%. 
   In our study, the frequency of shunt-dependent hydrocephalus is statistically high in the group with IVH. Sheehan et al, reported that the presence of IVH correlates with hydrocephalus rather than Fisher grade.³⁴⁾ Graff- Radford et al, reported diffuse layering of subarachnoid blood or blood located strategically in the CSF pathway, such as in the ventricle or near the fourth ventricular outlet, is more likely to result in hydrocephalus.¹³⁾ 
   In our study, rebleeding group has statistical significance in frequency of shunt-dependent hydrocephalus. The deterioration of level of consciousness and increase of Fisher grade after rebleeding can explain this result. 
   This relationship of distribution of blood to shuntdependent hydrocephalus can help to explain the relation of the location aneurysm to hydrocephalus. Like our study, in several literature, the aneurysm located in posterior circulation and anterior communicating artery(A-ComA) has higher frequency of shunt-dependent hydrocephalus than other locations.^{13)25)28)32)38)} The rupture of posterior circulation aneurysm can bring out more bleeding on basal cistern or fourth ventricle, which might result in CSF flow disturbance.