The present study was a retrospective review of patients who had underwent BHD following diagnosis of CSDH between January 2010 and June 2014. In the surgical findings, CSDH was defined as hematoma wrapped in a thin capsule and consisting of dark reddish liquefied blood.19) A total of 123 cases of BHD were performed at our institution, of which, 93 cases were selected for the study. The remaining 30 cases were excluded because they involved a prior history of treatment for cerebral infarction or cerebral hemorrhage such as subarachnoid hemorrhage, or they received ventriculo-peritoneal shunt surgery, or due to a lack of follow-up for a minimum of 3 months. Comparisons were made after dividing the patients into three groups based on the surgical technique: single BHD without saline irrigation (Group A, n=31), single BHD without saline irrigation (Group B, n=32), and double BHD with saline irrigation (Group C, n=30). In each cases, surgical techniques were selected according to operator's preferred method.

In all patients, surgery was performed under general anesthesia. A high-speed drill was used to create the burr hole in all patients, after which, the dura was opened carefully to verify the CSDH capsule. In Group A, a single hole was drilled at the parietal boss and a 5N silicon catheter was inserted to perform the drainage, without saline irrigation. In Group B, 2 holes were drilled, one each at the parietal boss and frontal area, to each of which, a catheter was inserted in a mutually crossing direction, again without saline irrigation. In Group C, 2 holes were drilled and irrigation was performed thoroughly with saline to confirm the complete removal of hematoma. Next, 2 catheters were placed in the same manner as in Group B. In all patients, catheters were connected to drainage bags and fixed at the same height as the patient's tragus to allow continuous drainage. The catheters were removed when the patient showed improvement in neurologic symptoms, and when it was determined from computed tomography (CT) that sufficient drainage of the hematoma had occurred. CT exams were performed immediately after surgery, and 1 week and 3 months after surgery for follow up monitoring.

For comparisons of the three groups, age, gender, status of hypertension, diabetes, hyperlipidemia, chronic renal or heart insufficiency, chronic alcoholism, smoking, and coagulopathy [international normalized ratio (INR)14) >1.20], taking of anti-platelets or anti-coagulants, presence of dementia, and preoperative Glasgow Coma Scale29) were investigated via a chart review. For comparisons of clinical results from each group, hospital days, recurrence and mortality rates, and Glasgow Outcome Scale (GOS) at discharge8) were investigated. From the GOS results, good recovery and moderate disability were defined as good outcomes, whereas, severe disability, persistent vegetative state, and death were defined as poor outcomes.691629)

Recurrence in patients was evaluated by performing CTs prior to surgery and 1 day, 1 week, and 3 months after surgery. In the preoperative CT, the density of hematoma was measured, with uniform density being defined as homogenous internal architecture and mixed density being defined as heterogeneous internal architecture (Figure 1).1011) The resolution of hematoma was compared by calculating the change in thickness and midline shift of the hematoma from the area with the maximal amount of hematoma seen on CTs taken prior to and 1-day after surgery.26) The change was defined as a percentile obtained by dividing the preoperative CT value by the difference derived from subtracting the postoperative CT value from the preoperative CT value.

Recurrence of CSDH was defined as an increase in the amount of hematoma in the surgical area, change in hematoma density, or effacement of cerebral sulci seen on CT exam within 3 months after surgery.119) Revision surgery was performed when neurologic symptoms was occurred, such as mental decrease, motor weakness, and dysarthria, clearly accompanied the recurrence.

For intergroup comparisons, the distribution of the continuous data was first evaluated for normality using the Shapiro-Wilk test. The normally distributed data was presented as the mean±standard deviation and the groups were compared using analysis of variance and Tukey test as a post hoc test. The non-normally distributed data was expressed in median (Q₁ to Q₃) and this data was analyzed via the Kruskal-Wallis test followed by Bonferroni's correction. Descriptive variables were subjected to χ² analysis or Fisher's exact test, as appropriate. Unadjusted logistic models were used to identify significant variables affecting the recurrence (Table 1).

Variables with p<0.10 were included in a multivariate logistic regression with conditional backward selection model. The level of statistical significance was set at p<0.05. All analysis was performed using SPSS 21.0 (IBM Corp., Armonk, NY, USA).

The baseline patient characteristics between Groups A, B, and C were compared. Age, gender, status of hypertension, diabetes, hyperlipidemia, renal insufficiency, heart insufficiency, smoking, and coagulopathy, taking of anti-platelets or anti-coagulants, and presence of dementia were similar between the three groups (Table 2). The number of chronic alcoholics was statistically significantly higher in Group B (p<0.05). The internal architecture on preoperative CT showed no statistically significant difference between the groups (Table 2).

Postoperative clinical outcomes between the three groups were compared. When hospital days and outcome at discharge (GOS) between the three groups were compared, no statistically significant differences were found. The number of cases in which revision surgery was performed was 6 (19.4%), 6 (18.8%), and 2 (6.7%) in Groups A, B, and C, respectively. Thus, Group C had relatively fewer revision surgeries. The number of mortality cases was found to be 3 (9.7%), 1 (3.1%), and 1 (3.3%) in Groups A, B, and C, respectively (Table 3). The causes of death were founded as following: there were 2 cases of ventilator-associated pneumonia (VAP) and 1 case of catheter-induced sepsis (Group A), 1 case of VAP (Group B), and 1 case of unknown origin sepsis were investigated (Group C).

Change in hematoma thickness in Groups A, B, and C was 29.77±7.94%, 49.73±12.87%, and 75.29±4.32%, respectively, while change in midline shift was 40.81±15.47%, 51.78±10.94%, and 56.16±16.16%, respectively. Thus, Group C had the most effective for resolution of hematoma and midline shift (p<0.05). As for recurrence, there were 12 cases (38.7%) in Group A, 8 cases (25.0%) in Group B, and 3 cases (10.0%) in Group C. Thus, Group C had a statistically significantly lower recurrence rate than Group A (p<0.05) (Table 3).

The authors analyzed factors associated with the postoperative recurrence of CSDH by multivariate logistic regression analysis, and found that presence of coagulopathy, single burr hole and without saline irrigation were significantly independent risk factors associated with the recurrence of chronic subdural hematoma (p<0.05). Therefore, double BHD with saline irrigation is strong associated with reduction of postoperative recurrence (Table 4).