Basal Cisternostomy in Head Injury: More Questions than Answers
Correspondence Address: Source of Support: None, Conflict of Interest: None DOI: 10.4103/0028-3886.355117
Source of Support: None, Conflict of Interest: None
Keywords: Basal cisternostomy, CSF-shift edema, decompressive hemicraniectomy, glymphatic pathway, head injury, intracranial pressure, paravascular spaces
Traumatic brain injury (TBI) is a leading cause of mortality and morbidity in all age groups and causes a substantial burden on the healthcare system. The damage to the brain due to TBI may result from primary insult due to direct injury to the brain and secondary insult due to metabolic and inflammatory reactions, leading to brain edema and intracranial hypertension. Raised intracranial pressure in cases of TBI is associated with poor prognosis. Various management strategies have been advocated for treatment of raised ICP, ranging from simple maneuver of head elevation to surgical interventions in the form of decompressive craniectomy. At present, DC is recommended as a primary procedure while removing the causative mass lesion (subdural hematoma or contusion) or as a secondary procedure targeted toward decreasing the raised intracranial pressure refractory to medical therapy. DC is known to reduce intracranial pressure and duration of stay in intensive care unit (ICU) but has also been shown to have more unfavorable outcomes.
Cisternostomy has been used routinely in aneurysm surgeries and skull-base tumors. Its novel use has been suggested in cases of severe head injury in decreasing brain edema and refractory intracranial hypertension by reversal of “CSF shift.” According to the CSF-shift mechanism, traumatic subarachnoid hemorrhage leads to cisternal obstruction, resulting in elevation of cisternal pressure above brain parenchymal pressure. This creates a pressure gradient between the cistern, which is at a higher pressure, and brain parenchyma, which is at a lower pressure. This causes rapid CSF shift from the cisterns through the glymphatic pathway (constituted by paravascular Virchow–Robin spaces) into the brain parenchyma, leading to increased brain edema and clinical deterioration due to resultant intracranial hypertension. Cisternostomy involves opening the basal cisterns to near-atmospheric pressure, which results in a “back shift” of CSF through the glymphatic pathway, thereby decreasing the intrabrain pressure and allowing decreased brain swelling and improvement in clinical condition.,,,,
However, the role of this novel technique in head injury has not been established as there are very few articles in the published literature dealing with its clinical implications. Therefore, the authors conducted a clinical study with a control group to assess the role of cisternostomy in head injury.
A prospective quasi-experimental clinical study was conducted at the department of neurosurgery in our institute between November 2018 and November 2020. Institute ethical committee (IEC) approval was taken. All patients with head injuries who were candidates for the standard decompressive hemicraniectomy (DHC) as per standard norms were included. These included cases of head injury presenting with acute subdural hemorrhage (aSDH) with a maximum thickness of ≥10 mm or with a midline shift of ≥5 mm on computed tomography (CT) scan and patients with refractory intracranial hypertension, irrespective of the patient's GCS score. Comatose patients (GCS <9) with aSDH <10 mm and a midline shift of <5 mm were included if the GCS score decreased between the time of injury and hospital admission by 2 or more points or the patient initially presented with asymmetric or fixed and dilated pupils. Patients with intracerebral contusions and signs of progressive neurological deterioration referable to the lesion, medically refractory intracranial hypertension, or signs of a mass effect on CT were included. Patients with GCS scores of 6–8 and frontal or temporal contusions of >20 cc in volume with a midline shift of >5 mm or cisternal compression on CT and patients with any contusion >50 cc in volume were included in the study.
Patients with raised intracranial pressure due to chronic subdural hematoma, extradural hemorrhage, posterior fossa hemorrhage, and causes other than trauma were excluded from the study. Patients who did not consent for surgery or were considered unfit for surgery were also excluded.
At presentation to the emergency and just before the surgery, the patient was examined for Glasgow Coma Scale (GCS), pupillary abnormalities, and presence/absence of motor deficits. For all cases, noncontrast CT head was performed at presentation in the emergency and repeated 4 hours after surgery unless the patient was not stable enough to be shifted for a CT scan. Apart from these, noncontrast CT scan was repeated when it was considered necessary for patient care.
Two groups: All patients fulfilling the criteria for inclusion were included in the study and allocated to either of the two groups. The first group (DHC-BC) comprised patients under the care of the corresponding author (NG). All these patients were treated with the intention to do decompressive hemicraniectomy (DHC) with basal cisternostomy (BC). The second group (DHC group) comprised patients under the care of other neurosurgeons of the department. These patients underwent the standard decompressive hemicraniectomy (DHC) alone. Allocation of patients in either group was thus based on the call-duty roster of the neurosurgery department, thus providing a fair amount of random selection [Figure 1]. After a quick bedside preoperative anesthetic checkup and routine preoperative investigations, consent for surgery and for inclusion in the study was taken from the patient's family before the surgery. We used an operating microscope in all cases undergoing basal cisternostomy.
Surgical steps in DHC-BC group: Surgery was performed under general anesthesia with the patient in the supine position and head rotated by 10°–15° to the opposite side. A standard large question mark or reverse question mark incision is used. The skin incision starts 1 cm in front of the tragus at the zygomatic arch and extends posteriorly above the auricle, upwards over the parieto-occipital area, and forward to the frontal region to the hairline. The temporalis fascia and muscle are divided in line with the skin incision. The temporalis muscle is reflected anteriorly and inferiorly with the cutaneous flap. The pericranium is elevated of the skull and reflected anteriorly. The first burr hole was made and opening intracranial pressure (ICP) was measured from a noneloquent noncontused area through a small durotomy made under this burr hole. A frontal burr hole just anterior to the coronal suture about 1.5–2 cm lateral to the midline was usually used for this. We used the Codman ICP system with an intraparenchymal catheter. Once the intraparenchymal ICP was taken, the catheter was removed so that the rest of the procedure could be performed smoothly. The remaining burr holes were made and a fronto-temporoparietal bone flap measuring at least 10 cm × 15 cm was elevated. Then, drilling of the sphenoid ridge was done using a drill burr until the orbito-meningeal arteries and lateral edge of superior orbital fissure. Next, the orbito-meningeal band was cut and the frontal lobe was elevated off the anterior clinoid process.
At the frontal base slightly cranial to the orbital rim, a 6–7-cm-long incision was made on the dura. Using a spatula, the frontal lobe was then retracted and the ipsilateral optic nerve and the ipsilateral internal carotid artery were identified [Video 1]. This was followed by the opening of interoptic, optico-carotid, and lateral carotid cisterns. The Liliequist membrane was identified through the optico-carotid window or lateral carotid window. The pre-pontine cistern was accessed after penetration of the Liliequist membrane and the basilar artery was seen [Video 1]. In this study, in case the posterior clinoid process obstructed the view of the pre-pontine cistern, it was not drilled and the procedure was stopped after performing a “partial cisternostomy.” Irrigation of the cisternal spaces was performed using normal saline for clearing out subarachnoid hematoma, blood products, and debris. After performing these steps of basal cisternostomy, the dura was opened in a C-shaped manner and then cuts were made on the dura radially. The remaining underlying subdural hematoma or significant contusions were managed as per the standard norms.,
A cisternal drain was kept in the basal cisterns after achieving hemostasis [Video 1]. This cisternal drain was exited from an opening separate from the main incision. For this, we employed the standard malleable silicone drain used as an external ventricular drain (Surgiwear). Through this cisternal drain, the cisterns were thoroughly irrigated once again and this cisternal drain was left in situ after the surgery for draining the cisternal fluid. Draining the cisternal fluid through the drain is believed to clear the cisterns and bring the cisternal pressure down. The intraparenchymal catheter, used at the beginning of the surgery was re-inserted in the same noncontused noneloquent area. The exit of this catheter was made from an opening, separate from the main incision. Then, standard lax duraplasty was done with autologous peri-cranium, temporalis fascia, or G-patch (Johnson and Johnson). Temporalis muscle was sutured followed by closure of skin in two layers (galea and skin). A subcutaneous pouch was then made in the abdominal wall for placing the bone flap.
Surgical steps in the DHC group: All the steps up to elevation of the bone flap were the same as in the DHC-BC group patients. After elevating the bone flap, the dura was opened in a C-shaped manner and then cuts were made on the dura radially. Underlying subdural hematoma or significant contusions were managed as per the standard norms., Again, the intraparenchymal catheter, used at the beginning of the surgery, was re-inserted in the same noncontused noneloquent area. The exit of this catheter was made from an opening, separate from the main incision. This was followed by closure as in the DHC-BC group.
Postoperative monitoring: After the surgery, patients in both the groups were shifted to the intensive care unit (ICU), and hourly ICP monitoring along with neurological examination (GCS, pupillary, and motor examination) was done for up to 72 hours. In the DHC-BC group, the cisternal drain was connected to a CSF bag (Surgiwear) for continuous drainage. The ICP catheter and the cisternal drain were kept in situ for 72 hours. These were removed earlier in case of an uncooperative patient or technical problem with the monitoring or in case of death. Apart from this, the patients received standard ventilator and medical management as deemed fit by the operating surgeon's team.
The response of surgery in decreasing ICP was monitored by calculating the number of hours with ICP >20 mm Hg in the first 24 hours after surgery and by calculating the intracranial hypertension index using the following formula:
Intracranial hypertension index = (Number of end-hourly measurements of intracranial pressure of more than 20 mm Hg/total number of measurements) × 100
Clinical outcome assessment: Clinical outcome was assessed by comparing the total number of days in ICU, total duration of hospital stay, 30-day mortality rate, mean GCS at discharge, and GOS-E at one month. Patients with GOS-E ≥5 were considered to have a favorable outcome.
Statistical analysis: Statistical analysis was performed using IBM Statistical Package for Social sciences (SPSS version 25.0, SPSS, Inc., Chicago, IL, USA). We assessed the distribution of variables by Shapiro Wilk's test. Univariate comparative analysis was done between two groups using a t test (for normally distributed variables) and Mann–Whitney (for non-normal distribution) was applied. For categorical variables, Chi-square test or Fisher's exact test was performed. Univariate comparative analysis was done between two groups using a t test according to the underlying distribution for continuous variables. For categorical variables, Chi-square test was performed. P < 0.05 was considered significant.
During the study period, 659 patients were admitted, of which 177 underwent surgical intervention. Of these 177 patients, 40 patients who were candidates for decompressive hemicraniectomy for traumatic brain injury were included in this study [Figure 1]. All 40 patients consented for surgery and for enrollment into the study. A total of 40 patients (27 males and 13 females) with a mean age of 40.65 ± 15.7 years (range: 3–72 years) fulfilled the inclusion criteria and were included in the study. The most common mode of injury was road traffic accidents (25; 62.5%) followed by fall from height (11; 27.5%), and 4 patients (10.0%) presented with unknown mechanism of injury. Overall, 18 (45.0%) patients presented with severe head injury (GCS: <8), 15 (37.5%) with moderate head injury (GCS: 9–13) and 7 (17.5%) with mild head injury (GCS: 14–15).
Group allocation: Nine patients were allocated to the DHC-BC group (six males and three females) and 31 patients to the DHC group (21 males and 10 females) [Figure 1]. Both groups were comparable in terms of baseline clinical characteristics such as age, gender, preoperative GCS, motor score, severity of head injury, pupillary asymmetry, focal motor deficits and radiological diagnoses, midline shift (in mm), cisternal effacement, and Rotterdam score on preoperative CT scan [Table 1].
The mean preoperative GCS was 7.9 ± 3.1 in the DHC-BC group and 8.0 ± 3.4 in the DHC group, while pupillary asymmetry was present in two patients in the DHC-BC group and 10 patients in the DHC group. Focal deficit contralateral to the lesion was present in 33.3% of patients in the DHC-BC group and in 38.7% of patients in the DHC group [Table 1]. On preoperative CT scan, the mean midline shift was 8.6 ± 1.5 in the DHC-BC group and 7.6 ± 2.4 in the DHC group. Furthermore, 44.4% of patients in the DHC-BC group had complete basal cisternal effacement compared to 22.2% in the DHC group. The mean Rotterdam score in the DHC-BC group was 3.4 ± 0.5 and 3.2 ± 0.7 in the DHC group [Table 1].
ICP Monitoring: In the DHC-BC group, there was a significant difference between the opening (25.7 ± 10.5) and closing parenchymal pressures (11.3 ± 5.9); t (9) = 3.515; P = 0.008. Similarly, in the DHC group, a significant difference was noted between the opening (25.4 ± 12.2) and closing parenchymal pressures (5.3 ± 3.5); t (16) = 6.840; P = 0.000 [Figure 2].
The mean closing pressure in the DHC group was significantly lower than that in the DHC-BC group (P = 0.003), even though the opening pressures in both groups were comparable (P = 0.945). The mean drop in ICP in the DHC-BC group (n = 9) was 14.4 ± 11.5 while that in the DHC group (n = 16) was 18.9 ± 12.4 (P = 0.359) [Table 2].
In the DHC-BC group, opening and closing ICP were measured for all nine patients. However, postoperative ICP monitoring was done for 24 hours in seven patients, 48 hours in six patients, and 72 hours in three patients. In the DHC group, opening and closing ICP were measured for 16 out of 31 patients. However, postoperative ICP monitoring was done for 24 hours in 10 patients, 48 hours in seven patients, and 72 hours in six patients.
The average total number of hours of ICP >20 mm Hg in the DHC-BC group was 4.6 ± 5.2, compared to 3.8 ± 4.4 in the DHC group (P = 0.738). On comparing the average intracranial hypertension index, it was 19.1 ± 16.9 in the DHC-BC group and 13.9 ± 18.5 in the DHC group (P = 0.564). The mean intracranial pressure over the initial 72 hours was 11.9 ± 2.1 in the DHC-BC group and 11.7 ± 1.5 in the DHC group (P = 0.549). The difference between the DHC-BC and DHC groups in terms of these postoperative ICP parameters was not statistically significant [Table 2]. The variation in the mean of hourly intracranial pressure monitoring in DHC-BC and DHC group patients is depicted in [Figure 3]. It is clearly seen that the trend of parenchymal pressures in both the groups parallel each other and the hourly averages was mostly around 12 mm Hg.
Clinical outcome: Average number of days of stay in the ICU and hospital were lower for the DHC-BC group (7.0 ± 6.1 and 15.0 ± 20.2, respectively) as compared to those in the DHC group (10.5 ± 9.3 and 19.3 ± 13.9, respectively). The 30-day mortality rate was 66.6% in the DHC-BC group and 32.2% in the DHC group (P = 0.067). The mean GCS at discharge was better in the DHC-BC group (11.7 ± 2.9) compared to 10.5 ± 3.7 in the DHC group (P = 0.615), while 11.1% of patients in the DHC-BC group had a favorable outcome (GOS-E at 1 month ≥ 5) compared to 9.7% patients in DHC group (P = 0.903). The difference in all these clinical outcome parameters in the two groups was not statistically significant [Table 3].
Complications: One case in the DHC-BC group underwent secondary DHC-BC on the opposite side as she had refractory intracranial hypertension after first DHC-BC on left side. None of the patients in the DHC group underwent another procedure for intracranial hypertension. Another patient in the DHC-BC group had an ICA injury intraoperatively, which resulted in a large hemispheric infarct. The mean intraoperative blood loss was higher for the DHC-BC group (983.3 ± 676.4) compared to the DHC group (667.4 ± 297.5) (P = 0.094) [Table 3]. All patients operated during the COVID-19 pandemic tested negative for COVID-19.
Effect on ICP parameters: In our study, both the groups (i.e., those who underwent DHC alone and those who underwent DHC-BC) had a decrease in ICP following surgery, with both the results reaching statistical significance. When the ICP at the end of surgery was compared among the two groups, it was found to be significantly less for the patients undergoing DHC alone (5.3 ± 3.5) compared to those undergoing DHC with BC (11.3 ± 5.9) (P = 0.003). Moreover, the mean drop in ICP was greater for the DHC group (18.9 ± 12.4) compared to the DHC-BC group (14.4 ± 11.5). However, this result was not statistically significant (P = 0.359). It should be noted that these ICP values were measured at the end of the surgery when the patient was under the influence of general anesthesia and this could have influenced the results.
Further, in the postoperative period, patients undergoing DHC with BC had a higher number of hours with ICP >20 mmHg in the first 24 hours postoperatively and a higher intracranial hypertension index compared to patients who underwent DHC alone. However, both these results were not significant statistically. The mean intracranial pressure over the initial 72 hours was comparable in both groups (P = 0.549).
Giammatei et al. observed that patients who underwent adjuvant cisternostomy presented lower mean ICP values over 72 hours and had a lesser need for osmotherapy compared to patients undergoing decompressive craniectomy alone. In their study, the hourly averages in patients undergoing adjuvant cisternostomy always remained lower than the threshold of 15 mm Hg.
Effect on clinical parameters: Giammatei et al. found that patients undergoing cisternostomy along with decompressive craniectomy spent lesser days on ventilator, had lesser stay in ICU, lower mortality, better GCS at discharge, and better GOS-E at 6 months. In the study by Cherian et al., it was found that patients undergoing basal cisternostomy in addition to decompressive hemicraniectomy fared better than those undergoing decompressive hemicraniectomy alone in terms of decrease in days on ventilator and mortality and improved GOS at 6 weeks. It was further noted by Cherian et al. that patients who underwent basal cisternostomy alone had further improvement in clinical outcomes compared to the other two groups. Parthiban et al. retrospectively studied 40 head-injury patients who underwent basal cisternostomy. They observed satisfactory results with basal cisternostomy in severe-head-injury patients with a favorable outcome of 77.8% in the BC group (basal cisternostomy alone) and 72.7% in the DC + BC group (decompressive craniectomy with basal cisternostomy) with an overall mortality of 6.8% in the severe traumatic brain injury group. However, they did not compare their results with patients undergoing decompressive craniectomy alone.,
In the present study, we observed that adding basal cisternostomy to decompressive hemicraniectomy seemed to decrease the hospital and ICU stay with better GCS at discharge. However, at the same time, there appeared to be higher mortality and fewer patients with a favorable outcome at 1 month. However, none of these results in our study were statistically significant. It is possible that BC improves the outcome, but the exact mechanism through which it helps is not clear. Contrary to what was previously believed, our findings suggest that BC probably does not act through ICP reduction as in our study, the ICP parameters were found to be better in the DHC group compared to the DHC-BC group. One possible mechanism that makes cisternostomy work could be better circulation through the glymphatic pathway following cisternostomy, thereby allowing for a better supply of nutrients to the brain parenchyma and drainage of harmful debris. This mechanism could be more important than reduction in ICP.
Major complications: One of our patients who underwent basal cisternostomy required a contralateral procedure for refractory intracranial hypertension in the postoperative period, while none of the patients in the DHC group underwent a contralateral procedure for refractory intracranial hypertension. Another patient in the cisternostomy group suffered an injury to the internal carotid artery. Such complications are more likely to occur in patients undergoing cisternostomy as the procedure involves manipulation of major neurovascular structures to reach the basal cisterns. In our patient, the faulty tip of forceps caused injury to the ICA while inserting cisternal drain while performing cisternostomy during night hours. Surgery for head injury is often done at odd hours in emergency operation theater with limited staff and often poor instruments. This can result in major complications like the one we had in one patient. This aspect needs to be looked into in centers planning to take up cisternostomies, which require adequate support and instruments, including an operating microscope. Giammatei et al. mentioned that one patient in the DC group developed enlargement of the hemorrhagic contusion following decompressive craniectomy, which necessitated surgical removal of the contusion. Two patients who underwent cisternostomy developed subcutaneous hematoma, which required surgical evacuation. In our study, the intraoperative blood loss was found to be higher in the DHC-BC group.
Surgical advances in head injury have been stagnated since the introduction of decompressive craniectomy almost a century back. Therefore, the prospect of cisternostomy decreasing ICP and improving outcomes in head injury seems very appealing. However, in our early experience, the addition of cisternostomy to decompressive craniectomy had no extra advantage with regard to decreasing ICP, morbidity, or mortality and seems to have a higher likelihood of encountering major complications and greater blood loss intraoperatively. This procedure needs great expertise and has a steep learning curve. However, as the number of patients in the present study was small, a larger study is required to conclude the final role of cisternostomy.
Limitations of the study: The main limitations of the study are small sample size, nonrandomization, and unequal number of patients in the two groups. Due to the limited number of patients, the results of the study should be taken with caution and further clinical studies with larger cohorts are needed to determine the role of cisternostomy in head injury. The opening and closing ICP values were measured during surgery under the effect of various anesthetic agents, while postoperative parenchymal pressure values were measured after the anesthesia had been stopped. Thus, the effect of anesthetic agents on parenchymal pressure should be considered. Due to various reasons, ICP monitoring could not be done in all patients for 72 hours, which might have introduced a bias in the measurement of ICP values.
Our preliminary single-center study did not show any benefit of adding cisternostomy to decompressive hemicraniectomy in patients with head injuries. A larger multicenter clinical trial, preferably a randomized control trial is required to answer this very pertinent question.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
[Figure 1], [Figure 2], [Figure 3]
[Table 1], [Table 2], [Table 3]