The current study showed that craniectomy procedure and large residual air volume are significantly associated with intracranial shift caused by postural change in postneurosurgical patients. In addition, in post-craniectomy patients, supratentorial bone defect and presence of parenchymal brain injury were associated with PBS, and a large skull defect area and a long interval from craniectomy were correlated with a larger PBS.
Previous studies using conventional horizontal imaging have reported postural brain displacements in supine versus prone or lateral positions15,16,17, whereas until recently, only few studies have evaluated intracranial positional shifts in supine versus upright posture. Although early mobilization is recommended after surgeries, safety during mobilization has not been completely validated. Because most postneurosurgical patients undergo mobilization without complications, intraoperative procedures may not affect the brain stability to a clinically significant degree in most cases. However, after craniectomy or in cases involving residual intracranial air, PBS, which is not observed from the outside of the cranium, should be considered.
PBS in post-craniectomy patients and its association with SSFS
The skull is supporting intracranial structures and is protecting them from not only mechanical damages but also posture-related physiological changes. Craniectomy partially removes this rigid barrier. Thus, the intracranial structures are more easily affected by the surrounding atmospheric pressure. The SSFS, or the syndrome of the trephined, is characterized by neurological disturbances in post-craniectomy patients that resolves after cranioplasty, and its pathophysiology is yet to be elucidated18,19,20,21,22,23,24. This condition is associated with a sunken skin flap in the skull defect region and is presumably caused by the distortion of underlying brain tissues23. Meanwhile, patients can be symptomatic even without radiological signs in supine position14,19,25. Aggravation of symptoms in upright position is a typical characteristic of SSFS26,27, and a previous case report has shown that upright imaging is effective in obtaining a diagnosis3. Although there was no statistically significant difference regarding the extent of PBS between patients with and without symptoms probably due to the small number of symptomatic individuals, it is reasonable to consider that a large PBS is responsible for orthostatic symptoms in SSFS.
This study revealed that large skull defect, long interval from craniectomy, and presence of parenchymal brain injury are associated with PBS. Previous studies have shown that these are also the risk factors of SSFS19,21,28. By contrast, other previously reported predisposing factors of SSFS, such as age25,28, sex19,28, and small relative intracranial CSF volume14 were not associated with PBS in this study. In the current study, PBS after infratentorial craniectomy was small. This supports the previous notion that infratentorial craniectomy does not require cranioplasty, and this is consistent with the fact that there is no reported case of SSFS after infratentorial craniectomy. Since only one patient had an implanted ventriculoperitoneal shunt, we could not assess the effect of CSF diversion devices, which can be a cause of SSFS based on previous reports19,21. In this patient, a relatively large PBS was observed despite a short interval from the craniectomy (7 days) and a small-sized bone defect, indicating the role of the CSF diversion device in increasing the risk of PBS (Fig. 2A).
SSFS is likely to develop when the resistance of skin flap tension and intracranial pressure have been affected by atmospheric pressure. Therefore, an increase in the elastance of the skull defect region, mainly composed of the skin flap, brain parenchyma, and CSF, can make the craniectomy area more vulnerable to internal brain shift. Factors correlated with a larger PBS contribute to a lower regional elastance. These include the removal of a large part of a rigid skull, tissue scarring with atrophy in the skin flap, which occurs if there is a long interval from craniectomy, injured brain with loss of parenchymal volume, and use of CSF diversion devices that contribute to intracranial hypotension.
PBS in postoperative pneumocephalus
Residual intracranial air was significantly associated with a large PBS. We intuitively understand that air in the restricted space moves according to positional changes. The specific gravity of air is quite lower than that of CSF. Therefore, intracranial air moves upward in upright position. When there is a large volume of air in the cranium, it can be space-occupying, which is enough to compress the adjacent brain. By contrast, the bridging veins in the convexity area are fixed to the brain surface and dura mater29. When the brain surface is compressed by air, these bridging veins might be stretched and torn in severe cases. With an accelerated motion with positional change, the tension affecting surrounding structures may increase, causing hemorrhage in the veins often located in the frontoparietal parasagittal region30,31,32. One report has shown that a torn bridging vein caused spontaneous subdural hematoma in a patient with pneumocephalus33. A certain number of cases involve postoperative ischemic or hemorrhagic complications of unknown cause after uneventful surgeries. These complications are often described as a venous cause with unvalidated etiologies. We assume that an increase in PBS can cause such complications. In patients with a large volume of residual intracranial air, rapid postural changes should be prevented, and possible stretching of the frontoparietal bridging veins while in upright position, which is not expected from examinations performed in supine position, must be considered. A study of PBS evaluated via supine versus prone MRI performed a similar assessment of the risk of bridging vein damage caused by positional change17. Major cortical and bridging veins do not show positional shifts compared with brain parenchyma.
Moreover, a related mechanism in the so-called remote hemorrhage after neurosurgical surgeries has been discussed34,35. The presence of postoperative pneumocephalus indicates that air has substituted the space previously occupied by something intraoperatively removed, such as resected mass lesion and drained CSF. One of the most convincing mechanisms is tearing or transient occlusion of stretched bridging or subcortical veins associated with CSF overdrainage and mechanical brain shift34,35,36. This is supported by the fact that surgery performed in sitting position is a risk factor of remote hemorrhage34. We believe that even patients who underwent surgeries in supine position can be affected by the same mechanical shift postoperatively due to PBSs. Although we routinely substitute artificial CSF for intracranial air before dural closure, the air cannot be removed completely. In addition, such a replacement maneuver cannot be utilized in the endoscopic endonasal approach. After endoscopic endonasal skull base surgery with massive intraoperative CSF leakage with severe pneumocephalus, the risk of PBS must be considered. Delayed mobilization might be recommended until the volume of residual air is decreased with absorption.
Upright imaging in intracranial hypotension
Previous studies have shown that pneumocephalus concomitant with CSF hypovolemia or intracranial hypotension is an important radiological finding. This condition can lead to severe post-craniotomy complications, referred to as sinking brain syndrome37 or brain sag after craniotomy38. Because upright images were not obtained immediately or earlier after surgery, the effect of postoperative intracranial hypotension could not be evaluated. However, one patient with Marfan syndrome who underwent CSF leak closure presented with remarkable radiological changes based on positions. Upright imaging revealed typical findings of intracranial hypotension39,40,41. Whether these results are solely caused by intracranial hypotension due to CSF leakage or are correlated with connective tissue weakness in Marfan syndrome has not been validated42. Both factors might have affected positional radiological changes since both entities are closely correlated43,44,45. A previous report of six patients with intracranial hypotension caused by CSF leak showed no positional changes. However, the study conducted 0.6 Tesla MRI with a low spatial resolution, and only few patients were included46. Therefore, further studies targeting patients with spontaneous intracranial hypotension or orthostatic headache must be conducted to validate whether upright imaging is effective for diagnosing this rare but underdiagnosed pathology.
The current study had several limitations. First, there was selection bias in terms of patient selection in this single-arm exploratory study, which was performed at a single institution. We could not enroll patients with severe brain damage who have restricted activities or who cannot provide informed consent for the study. Thus, baseline diseases comprised a relatively large proportion of extra-axial tumors. In addition, we excluded patients who cannot maintain a stable upright position, including those with severe orthostatic symptoms.
Second, the timing of CT scans varied as the procedures must be performed based on the limited availability of the upright CT scanner and the hospital’s clinical workflow should not be disturbed. Therefore, most patients underwent upright and routine supine CT scan on postoperative days 6 or 7. CT scan images obtained in the earlier postoperative stage must have shown more significant changes according to positions due to a large volume of residual air with less refilled CSF and undeveloped postsurgical tissue adhesion. However, earlier scans were not feasible owing to safety concerns. In addition, post-craniectomy patients underwent CT scans at different timings because procedures were clinically scheduled according to the timing of the cranioplasty. Nevertheless, variations in the intervals from the craniectomy, in turn, contributed to the sub-analysis that used this interval as an important variable.