Intra-, Epidural And Intracranial Pressure Changes During Interlaminar Endoscopy, With and Without Dural Tear
Article information
Abstract
Objective
Endoscopic spine surgery implies possibly severe complications of the central nervous system, from headache to seizures and autonomic dysreflexia. These adverse events might be due to increased intracranial pressure (ICP), presumably induced by increased spinal intra-/epidural pressure caused by fluid irrigation. This study was designed to perform interlaminar endoscopic lumbar discectomy (IELD) at different irrigation fluid settings while monitoring its effect on intra-/epidural and ICPs, with and without dural tears.
Methods
Spinal intradural pressures were measured by introducing catheters through a sacral approach to human cadavers’ lumbar, thoracic, and cervical levels. Additionally, an epidural probe was placed at L3–4. ICP was measured by an intraventricular probe. IELD was performed at L3–4, and the effect of varying irrigation pressures by different endoscopic pump systems and gravity-based irrigation on intra-/epidural and ICP pressures was measured before and after durotomy at L3–4.
Results
Intradural pressure at L3–4 correlated linearly with increasing irrigation pressure, irrespective of the used pump system (median pressure increase at 100-mmHg irrigation pressure: system I: 7 mmHg, r=0.94, p=0.002; system II: 7 mmHg, r=0.89, p=0.017) or gravity (8 mmHg, r=0.93, p=0.242). This effect was also seen intradurally at the thoracic/cervical spine, epidural, and intracranial level, and was even more pronounced with the maneuver of outflow-occlusion and a dural tear present.
Conclusion
While performing IELD, pump pressures correlated linearly to intra-/epidural pressures and ICPs. Pressures did not rise to concerningly high levels without outflow-occlusion, even with increased pump pressures. In the presence of a dural tear, higher pump pressures exacerbated by occlusion may lead to deleterious intradural and ICP elevations.
INTRODUCTION
Endoscopic techniques, in contrast to traditional open surgeries, are notable for their minimally invasive nature, characterized by reduced incision sizes, minimized anatomical disruption, enhanced visualization, minimal to no bone removal, minimal blood loss, lower infection rates, reduced length of hospital stay, improvement in patient-reported clinical outcomes, and rapid return to normal activities of daily living [1,2]. The technological progression of pump systems employed in endoscopic spine surgery has markedly enhanced intraoperative visualization capabilities [2]. Initially, gravity flow systems were implemented and continue to be in widespread use [3]. However, advancements have led to the development of automated pump systems equipped with either pressure regulation or a combination of pressure and volume control mechanisms. Nevertheless, all endoscopic systems have irrigation as a potential surgical risk factor in common. Possible sequelae of irrigation, like headache or seizure [4-6], or even significant subarachnoid hemorrhage, hydrocephalus, cardiac arrhythmias, and hypotension, have been observed [7]. Additionally, an international survey among surgeons specializing in spinal endoscopy highlighted that approximately 5% of these practitioners believe they have encountered patients suffering from neurological impairments (such as nerve root damage or cauda equina syndrome) that were attributable to the irrigation process during endoscopy [7]. Not much is known about physiological intraspinal and intracranial pressures (ICPs) except that 25 mmHg was the measured epidural pressure in patients undergoing lumbar interlaminar full-endoscopic discectomy [7], and ICPs of >15 mmHg are potentially dangerous [8]. It is postulated that the aforementioned complications may occur due to irrigation fluid entering the epi- and/or intradural space, causing pressure-related symptoms in the further cranial parts of the spinal axis. To date, this hypothesis has not been investigated quantitatively. Consequently, there are neither established guidelines regarding the pressure settings when using pump systems nor recommendations for the height-above-patient when using gravity-based irrigation.
MATERIALS AND METHODS
The study was approved by the local ethics committee of Kantonale Ethikkommission Zürich (2023-01986), and conducted in accordance with the Helsinki Declaration.
1. Experimental Setup
The experiment was sequentially performed in 3 fresh frozen human cadavers that were thawed at room temperature. Before the experiments, computed tomography scans of the spine and skull were performed to rule out relevant spinal pathologies such as stenosis or intracranial bony pathologies such as basal skull fractures or tumors. The experimental setup has been previously described in detail [9] (Fig. 1).
(A) Experimental setup depicting the specimen, endoscopy tower, C-arm for location verification, and monitors showing pressure graphs. (B) Schematic illustration of the various catheters and probes in their anatomical locations.
Briefly, 3 lumbar catheters (Neuromedex Lumbalkatheter 4.5F, Neuromedex GmbH) were advanced to the levels L3–4, T8–9, and C5–6 through a durotomy in a dorsal sacral approach. A pressure transducer (Codman microsensor, Codman Specialty Surgery, Integra LifeSciences Corp.) was placed epidurally at L3–4 after introducing it through the same approach. A ventricular catheter (TraumaCath Ventricular Catheter Set, Integra LifeSciences Corp.) was introduced intracranially in the lateral ventricle [10,11]. A pressure sensor (Neurovent P, RAUMEDIC AG) was also placed intraparenchymal on the contralateral side. To simulate a physiological intradural pressure of approximately 15 cmH2O, the dura sac was filled by a feeding catheter connected to a Ringer’s solution bag at a level 15 cm above the spinal canal.
Data acquisition for intradural measurements was previously described in patients with spinal cord disorders [12-17]: Catheters were connected through separate pressure transducers (VentrEX system, Neuromedex GmbH) to 3 interface monitors (RAUMED NeuroSmart, RAUMEDIC AG). After linking the epidural pressure transducer to a Codman DirectLink ICP box, it was connected to an interface monitor (Philips X2-Pat. Interface+MX 700, Philips), and recording of the pressure signal was performed with ICM+software (Cambridge Enterprise).
2. Endoscopic Technique
Uniportal interlaminar endoscopy was performed with gravity-based irrigation and 2 different endoscopic pump systems (System I and System II). Endoscopic Pump System I is a pump system originally designed for arthroscopy, providing independent adjustment of in-line pressure and flow levels and maintaining a constant, nonpulsed pressure. Although offering the possibility of controlling outflow as well, for this study, we used the inflow-only mode of the pump. The applicable pump pressure ranges from 0–120 mmHg. In contrast, Endoscopic Pump System II is a pump system designed specifically for spine endoscopy. It is an inflow-only pump system as well, however, the applicable pump pressure ranges from 0–100 mmHg. After reaching a steady state, lumbar epidural, intraventricular, cervical intradural, thoracic intradural, and lumbar intradural probes were registered as repeated baseline measurements without water inflow after each system change. Relative pressure increases are displayed compared to these baselines.
The maneuver of “occlusion” to increase the hydrostatic pressure was performed by manually closing the irrigation outflow of the endoscope. Finally, a durotomy of 5 mm×5 mm was performed at L3–4, and all measurements described above were repeated.
3. Data Analysis
Data analysis was performed in MATLAB (Matlab 2020b, MathWorks). To test for linear correlation (Pearson, r) between applied and measured relative pressures, best-fit regression was computed based on the median measurement of the 3 specimens at each applied pressure. The level of significance was set to 0.05.
RESULTS
1. Intra- and Epidural Pressure Measurement With Increasing Irrigation Pressure
Intradural and epidural pressure measured at the level of the interlaminar approach L3–4 correlated linearly with a pump pressure increase for both endoscopic pump systems (system I: intradural r=0.94, epidural r=0.93, p=0.002; system II: intradural r=0.89, epidural r=0.89, p=0.017) and the gravity-based system, although not significantly (gravity-based system: intradural r=0.93, p=0.242 epidural r=0.93, p=0.246). The median measured intradural pressure increase at L3–4 and a maximum of 100-mmHg irrigation pressure was 7 mmHg in systems I and II and 8 mmHg in the gravity-based irrigation system. Interindividual differences were apparent, although the general trend was similar for all specimens (Fig. 2).
Intradural and epidural median pressure measured at the level of the interlaminar approach L3–4 correlated linearly (r) with a pump pressure increase (+/- outflow-occlusion) for both endoscopic pump systems and not significantly for the gravity system. With the maneuver of occlusion, spinal intra- as well as epidural pressure increased in some cases more than double compared to no outflow-occlusion.
2. Effect of Occlusion of the Outflow
Despite a linear correlation in the setting of no occlusion and even with pump pressures of 100–120 mmHg, the intra- and extradural pressure measured at L3–4 did not increase more than 20 mmHg. However, with the maneuver of occlusion, spinal intra-as well as epidural pressure increased more than double compared to no outflow-occlusion, particularly in the pump systems (Fig. 2).
3. Multilevel Measurements With Dural Tear
The above-described effects of linear relation with pump pressure increase and occlusion could be observed at all intra- and epidural levels as well as intracranially, in the presence of a dural tear at the level L3–4. The largest effect was seen epidurally, the least effect intracranially (Fig. 3). Although the median pressure increase remained below 20 mmHg at all levels in all systems, a large increase up to 80 mmHg was observed with outflow-occlusion in the presence of a dural tear.
Multilevel relative median pressure evolution at L3–4 with and without occlusion in the presence of a dural tear at L3–4 (intracranial pressure [ICP]; intradural C5–6, T8–9, L3–4; epidural L3–4).
4. Effect on ICP With and Without Dural Tear
ICP increased with the applied pump pressure. In the presence of a dural tear, median absolute intracranial pressure reached or exceeded the maximal tolerable ICP of 15 mmHg [8] with all 3 endoscopic systems (Fig. 4). The effect was more pronounced with outflow-occlusion, reaching 90 mmHg in one specimen with 120-mmHg pump pressure. Notably, absolute pressures have to be interpreted with care in a cadaveric model.
DISCUSSION
Full-endoscopic lumbar endoscopy is increasingly gaining popularity. In spinal endoscopy, irrigation pressures can reach levels that may critically elevate pressures in various spinal and/or intracranial compartments. A pump pressure of 40 mmHg has somehow become an acceptable “standard” in the spine endoscopy community, but very little is known about the actual pressure translation from the pumps to the surgery site [7]. To understand pressure variations, an experimental cadaveric model has previously been developed and validated to investigate intradural, epidural, and ICPs across the craniospinal axis during full-endoscopic lumbar discectomy [9]. This approach has been further evolved in this study, applying it to clinically relevant situations like a dural tear, since incidental durotomy is a potential threat to safe endoscopy. The model provides a framework for understanding safe irrigation parameters, thereby improving patient safety.
1. Intra- and Extradural Pressure Measurement With Increasing Irrigation Pressure With Different Irrigation Systems
Spine endoscopy may be performed with different irrigation concepts, namely gravity-based and pressure- and/or flow-controlled pump systems. There are various advantages and disadvantages of the different concepts that have been originally adopted from arthroscopy. The gravity-based irrigation system is controlled by the height of the bag of irrigation fluid in relation to the surgical level. Consequently, a fluid height of 0.5 m above the spine should equal approximately 37 mmHg of pressure (hydrostatic pressure (P)=fluid density (ρ) * gravity (g) * fluid height (h)). Bernoulli’s principle states that increased velocity of fluids in motion causes decreased pressure, although this accounts only for ideal fluid dynamics [18,19]. Based on this principle, the Venturi effect states that when fluid passes through a narrowed section of a pipe, such as the endoscopic working channel, its velocity increases while the pressure decreases [7]. Consequently, these factors may explain why the actual pressure at the surgical site can differ from the pressure expected based solely on the height of the irrigation fluid [20]. On the other hand, there are different pump systems available commercially. Similarly, pressure as indicated on the pump monitor does not necessarily correspond to the pressure applied at the surgical field. Accordingly, in our study, the measured pressure correlated with an increase in irrigation pressure in all systems but could not be translated directly to the applied pressure for the abovementioned reasons.
2. Effect of Occlusion of the Outflow
Typically, when severe bleeding interferes with the endoscopic view, one approach for bleeding control is to increase the irrigation pressure. Elevating the pressure of the irrigation can surpass the pressure from a bleeding vessel, thus enabling its visualization and cauterization. Once bleeding is under control, the pump pressure is usually reduced to its original rate. In our study, we showed a linear pressure increase with irrigation pressure increase at all levels of the spinal axis. However, it is noteworthy that even with a pump- or gravity pressure increase to 100–120 mmHg, the relative intra- and epidural pressure increase compared to the baseline did not exceed 20 mmHg as long as the outflow of the endoscopic system was kept open. During the experiments, we observed that closing the outflow of the endoscope can increase the pressure without increasing the infusion pump’s flow rate. A pressure of 25 mmHg has been measured in patients who underwent routine L4–5 and L5–S1 interlaminar full-endoscopic discectomy [7]. However, with the occlusion maneuver, especially epidural pressure at L3–4 increased more than 20 mmHg compared to the baseline, with all 3 irrigation systems at higher pump pressures.
3. Presence of Dural Tear
In the presence of a dural tear at L3–4, the measured epidural and intradural pressures at all levels increased more than 20 mmHg with occlusion, even at a pump pressure of 40 mmHg (which is commonly regarded as safe in the spine community) in systems I and II. Accompanied by the rising prevalence of spinal endoscopy comes an increasing awareness of potential procedural complications up to 10% [21,22]. depending on surgical training and expertise, as well as patient selection [2]. Dural tears are a special concern in spinal endoscopy because they are technically challenging to address due to the limited working channel of the endoscope and the scarce availability of dedicated surgical instruments [7,23]. Often, surgeons might encounter a comparably small tear that does not need to be repaired, or even a tear that may not be visible immediately due to masking by irrigation fluid. Such a tear, however, could cause serious adverse events due to an unknown increase in intradural and ICPs [7]. In this study, we were able to show that, in the presence of a dural tear, ICP correlated linearly with a pump pressure increase with endoscopic pump systems and also with a trend in the gravitybased system. An intracranial pressure of 15 mmHg is generally considered the maximal normal pressure in adults [8]. In the presence of a dural tear, absolute ICP was ≥15 mmHg even without outflow-occlusion with irrigation pressures which are generally considered safe [7]. With outflow-occlusion, ICP reached 90 mmHg in one specimen with an irrigation pressure of 120 mmHg. Similar results have been found in an animal study measuring ICP during full-endoscopic lumbar spine surgery under varying irrigation pump pressures and observing a critical increase of ICP with outflow-occlusion and intentional durotomy (reaching approximately 86 mmHg) [24].
After observing 4 of 16,725 patients undergoing percutaneous endoscopic lumbar discectomy (PELD) suffering from seizure in a clinical retrospective study [5], Choi et al. [4] investigated the underlying mechanisms by monitoring the epidural pressure at the cervical level in patients undergoing PELD with local anesthesia. They found a correlation between clinical neck pain and increased epidural cervical pressures and, therefore, associated neck pain as a prodromal sign of seizure, even though seizure did not occur in this patient cohort. These findings corroborate with our study, associating increased epidural pressures with adverse events at the intracranial level. However, we did not measure the epidural but intradural pressure directly at the cervical level. Similarly, we found a linear correlation of ICPs with pump pressure increases. This was most pronounced with the maneuver of outflow-occlusion and in the presence of a dural tear, which might explain the underlying mechanisms of patients suffering from adverse events such as seizure undergoing full-endoscopic lumbar discectomy. Case series have associated uncontrolled irrigation in the presence of a dural tear with acute hydrocephalus and intracranial hemorrhage, resulting from the sudden influx of fluid into the cranial subarachnoid space [7]. While such adverse events are rare, they highlight the potentially life-threatening consequences of uncontrolled irrigation pressure and stress the importance of systematically investigating underlying mechanisms.
4. Limitations
The insights presented here were derived from a biomechanical cadaveric model, which inherently had several limitations. First, the model does not account for cardiocirculatory and respiratory factors contributing to craniospinal cerebrospinal fluid (CSF) compartment pressures. Second, pressures built up in different compartments depend on multiple factors, including patient-related factors due to different anthropometrics, particularly body mass index, or, in the case of ICP, intracranial volume. Therefore, pressures of the intraspinal compartment differ inter-individually and have to be analyzed and interpreted as such [25]. Also, a small sample size of 3 specimens might lead to a selection bias, and general conclusions must be drawn carefully from these findings.
Furthermore, this experimental setup did not simulate other physiological features like cerebral autoregulation, cardiac-gated CSF flow, and CSF viscosity. Additionally, freezing and thawing might alter tissue conditions compared to the in vivo state. Considering these factors, absolute pressures might significantly deviate from the model. However, from a biomechanical perspective, relative pressure changes did provide useful clinically applicable information, especially considering that experiments of this kind could not be performed in patients due to ethical reasons.
CONCLUSION
Based on a previously developed cadaveric experimental model systematically exploring relative intra-, epidural, and ICP changes across the craniospinal axis induced by irrigation during lumbar spine endoscopy [9], this study investigated fluid management during clinically relevant situations like a dural tear. While all pressures correlated linearly with pump pressure increase or height in gravity-based irrigation, the increases remained shallow as long as there was no dural tear and/or no prolonged occlusion of the outflow. The model provides a framework for understanding safe irrigation parameters, thereby improving patient safety during spine endoscopy.
Supplementary Materials
Supplementary Tables 1-2 and Supplementary Fig. 1 are available at https://doi.org/10.14245/ns.2550456.228.
Absolute pressures measured at different locations in 3 specimens with increasing pressures during interlaminar endoscopy at L3–4 with intact dura
Absolute pressures measured at different locations in 3 specimens with increasing pressures during interlaminar endoscopy at L3–4 with intact dura
Multilevel relative median pressure evolution at L3–4 with and without occlusion with intact dura at L3–4 (intracranial pressure [ICP]; intradural C5–6, T8–9, L3–4; epidural L3–4).
Notes
Conflict of Interest
The first author (MF) reports being a Consultant for Medacta, Arthrex, 25Segments, Incremed, Zimmer Biomet, Zurimed and President of the Board of MovingSpine (Balgrist University Startup) and shareholder of Balgrist University Startups. VH reports being a consultant for Arthrex, Nexon and Joimax. JW is a shareholder of Balgrist University Startup (MovingSpine). JFS receives fellowship support by ETH MedLab Fellowship. All the other authors report no conflicts of interest.
Funding/Support
This study received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.
Acknowledgments
The authors would like to thank Prof. Vartan Kurtcuoglu for consulting about methodology, Tanja Walther and the team of OR-X (a swiss national research infrastructure for translational surgery) for assisting during endoscopic experiments.
Author Contribution
Conceptualization: MF, JFS, VH, AS, CMZ, JW; Data curation: JFS, MF, VH, JMS, AS, CMZ, NK, MRF, JW; Formal analysis: MF, JFS, CMZ, NK, MRF; Methodology: MF, JFS, CMZ, VH, AS; Project administration: JFS, AS; Visualization: VH, JMS, AS, JFS, CMZ, NK; Writing – original draft: JFS; Writing – review and editing: VH, JMS, NK, JW, AS, MF, CMZ, MRF, JFS.