Lumbar Catheter for Monitoring of Intracranial Pressure in Patients with Post-Hemorrhagic Communicating Hydrocephalus

Transcription

1 DOI 1.17/s ORIGINAL ARTICLE Lumbar Catheter for Monitoring of Intracranial Pressure in Patients with Post-Hemorrhagic Communicating Hydrocephalus Verena Speck Dimitre Staykov Hagen B. Huttner Roland Sauer Stefan Schwab Juergen Bardutzky Ó Springer Science+Business Media, LLC 21 Abstract Background We investigated the feasibility and accuracy of intracranial pressure (ICP)-measurement by lumbar drainage (LD) catheter in patients with post-hemorrhagic communicating hydrocephalus (PHCH). Methods Patients with subarachnoid hemorrhage (SAH, n = 21) or spontaneous ganglionic hemorrhage (ICH, n = 22) with ventricular involvement and the need for external ventricular drainage (EVD) due to acute hydrocephalus were included. When EVD weaning was not feasible due to persistent hydrocephalus, an additional LD was placed, after which EVD was clamped off. During this overlap period, patients underwent simultaneous pressure recording via EVD ( EVD-ICP ) and LD ( LD-ICP ). Testing included manual compression of the jugular veins and body-posture changes from supine to 3 position. After EVD removal, we evaluated sensitivity and specificity of ICP-rise >2 mmhg during continuous monitoring via LD for the detection of persistent PHCH using additional evaluation with computed tomography (CT). Results A total of 1,86 measurements were performed in 43 patients. LD-ICP was strongly correlated to EVD- ICP, with determination coefficients R 2 for the baseline measurements and each of the maneuvers ranging from.9.99, and slopes ranging Sensitivity of V. Speck (&) D. Staykov H. B. Huttner R. Sauer S. Schwab Department of Neurology, University of Erlangen, Schwabachanlage 6, 914 Erlangen, Germany verena.speck@uk-erlangen.de J. Bardutzky Department of Neurology, University of Freiburg, Breisacher Straße 64, 7916 Freiburg, Germany LD-ICP >2 mmhg for detection of persistent PHCH as compared to CT was 81% and specificity was 1%. Two patients with severe SAH developed reversible signs of herniation after gradually increasing differences between LD-ICP and EVD-ICP indicated a cranio-spinal pressure gradient, likely due to cerebrospinal fluid overdrainage via LD. Conclusion ICP measured via LD highly and reliably correlated to ICP measured via EVD in patients with PHCH. Keywords Lumbar drainage Intracranial pressure Hydrocephalus Subarachnoid hemorrhage Intracerebral hemorrhage Introduction Patients with severe stroke are at high risk for the development of secondary brain damage and increase of intracranial pressure (ICP). Secondary brain injury may be caused by local mass effect due to hematoma growth, ischemic or perihemorrhagic brain edema, or by global increase in ICP due to hydrocephalus [1 3]. Although clinical observation is an important monitoring tool, its value is usually limited in such patients due to reduction of consciousness or the need for sedation and mechanical ventilation. Therefore, continuous ICP monitoring is commonly recommended to guide therapy and prevent irreversible damage and secondary neurological deterioration [3, 4]. An external ventricular drainage (EVD) is usually considered to be the gold standard for invasive ICP monitoring []. Especially in patients with post-hemorrhagic hydrocephalus after subarachnoid (SAH) or intracerebral hemorrhage (ICH) with ventricular involvement, an EVD has the advantage of additional

2 therapeutic cerebrospinal fluid (CSF) release to decrease elevated ICP [4, 6]. However, in the past decade, another treatment alternative has emerged for patients with communicating post-hemorrhagic hydrocephalus. Lumbar drainage (LD), a simple and less invasive bedside technique, has been increasingly used to replace EVD, thereby reducing EVD duration and avoiding EVD exchange [7 9]. Moreover, additional beneficial effects of LD on the development of vasospasm after SAH [1 13] and the incidence of permanent hydrocephalus after ICH with severe ventricular involvement [7 9] have been suggested, likely by accelerating blood removal from the subarachnoid space. However, early removal of EVD in such patients treated with LD may occur at a time point at which ICP monitoring may still be necessary due to reduced consciousness or prolonged need for sedation. Theoretically, in the setting of communicating hydrocephalus, hydrostatic pressure in the lumbar cistern ( LD-ICP ) should correlate with ICP in the ventricular system ( EVD-ICP ). Therefore, the aim of this prospective study on patients with PHCH after SAH or ICH was (i) to investigate the accuracy of ICP estimation using LD by simultaneous recording of LD-ICP and EVD-ICP, and (ii) to evaluate the feasibility and the diagnostic value of continuous ICP monitoring via LD for the detection of persistent hydrocephalus using additional evaluation by computed tomography (CT). Patients and Methods Patient Selection This prospective study was approved by our local ethics committee. This study recruited patients from two clinical trials. These studies investigated the safety and feasibility of lumbar CSF drainage (i) in patients with SAH and persistent hydrocephalus and (ii) in patients with spontaneous ganglionic ICH with secondary IVH and persistent hydrocephalus [9]. Inclusion criteria for ICH patients were: (1) spontaneous hypertensive ganglionic ICH <3 ml; (2) secondary IVH; (3) acute obstructive hydrocephalus with the need for EVD at admission. Exclusion criteria for ICH patients were midline shift >. cm or parenchyma hematoma >3 ml; complete obstruction of the third or fourth ventricles with blood (LD was not placed before third and fourth ventricles were cleared from blood); international normalized ratio >1.4; coagulopathy; ICH due to trauma, tumors, or vascular malformation; infratentorial ICH; and age <18 or >8 years. Inclusion criteria for SAH patients were: (1) SAH Hunt and Hess >2 and fisher grade 3 or 4 (parenchymal hematoma <3 ml); (2) acute hydrocephalus (obstructive or communicating) with the need for EVD. Exclusion criteria for SAH patients were midline shift >. cm; complete obstruction of the third or fourth ventricles with blood at the time of LD placement; all basal cisterns filled with blood, international normalized ratio >1.4; coagulopathy; SAH due to trauma; and age <18 or >8 years. In both patient collectives, an additional LD was placed only when EVD clamping was not feasible due to increase of ICP >2 mmhg and ventricles enlargement on CT, despite communication between inner and outer CSF spaces indicating persistent communicating hydrocephalus. Thus, at the time of LD placement, further ICP monitoring and CSF drainage was indicated in all patients. Study Protocol Basic Management ICH or SAH with IVH were diagnosed by cranial CT on admission in all patients. Hydrocephalus was defined by measuring the bicaudate index, considered present if it exceeded the 9th percentile for age [14]. All patients received standard medical treatment according to the American Stroke Association guidelines [3] or SAH guidelines [4] including early intubation at Glasgow Coma Scale levels <9. In ICH patients with obstruction of the third and fourth ventricles, intraventricular fibrinolysis with recombinant tissue plasminogen activator was performed as described in detail [9] using a dose-modified protocol (1 mg every 8 h, max. 12 mg). For the detection of side effects possibly related to LD, the following procedures were performed: CSF analysis for infection (leucocytes, protein, lactate, glucose) every other day, daily clinical examination for local infection or bleeding complication at the side of LD placement, serial CT scans (based on study protocol: daily up to day 4, at day 7 and 1, and if clinically indicated) for the detection of cerebral subdural hematoma or hygroma due to CSF overdrainage, and hourly recording of pupils. Management of Extracorporeal CSF Drainage An EVD was inserted as soon as acute hydrocephalus was diagnosed. As soon as the third and fourth ventricles were cleared from blood and communication between inner and outer CSF spaces was seen on CT, the EVD was clamped under continuous ICP monitoring and removed after another 24 h when ICP remained <2 mmhg and CT showed no enlargement of the ventricles. If the attempt to clamp the EVD was unsuccessful, the existence of

3 communicating hydrocephalus was assumed and a silicone catheter (Codman lumbar drainage kit) was inserted into the subarachnoid space at the L3 to L4 level as previously described in detail [7, 8]. External CSF outflow was then continued by LD at a rate of 1 ml/h, while the EVD was clamped under continuous ICP monitoring. EVD was removed after at least 24 h when ICP remained <2 mm Hg and CT showed no enlargement of the ventricles, otherwise EVD was reopened. Extracorporeal CSF drainage was then continued for another 48 h through the LD. LD was then clamped and left closed for 24 h under continuous LD-ICP -monitoring. If enlargement of the ventricles was evident on CT, LD was reopened and attempts of LD clamping were repeated every 48 h. If no hydrocephalus was seen on CT, the lumbar catheter was removed. After five unsuccessful attempts of LD clamping, a permanent hydrocephalus was assumed and a ventriculo-peritoneal shunt was placed [7 9]. ICP Measurements Correlation between ICP Measurements Via LD and EVD For simultaneous measurement, pressure transducers for both the EVD and the LD were always positioned at the level of the Foramen of Monro. Before all measurements, both drainage systems were clamped off and zero balanced, and ICP-values via EVD ( EVD-ICP ) had to be stable for a period of at least min. Pressure measurements were recorded continuously with real-time technique and values were averaged every minute. No manipulation of the patients was performed during measurements including nursing, changing infusions or ventilation parameters. A standardized measuring procedure consisting of six different phases was performed every four hours during the overlapping period of EVD and LD. First, the patient was placed in supine position and baseline values were recorded (phase 1). A Queckenstedt maneuver was then performed by soft compression of the jugular vein until a plateau was reached without a rise of both measured pressure values for at least 3 s. EVD-ICP and LD- ICP values were noted at the plateau phase (phase 2) and 3 s after jugular compression was released (phase 3). Then, the same procedures were repeated with the patient in 3 upper body position (phase 4-6). Thus, during each measuring procedure, six EVD-/LD-ICP pairs of six different situations were available for analysis. Pulse Wave Analysis In addition to the mean pressure, amplitude (difference between diastolic minimum and systolic maximum pressures) and time from diastolic minimum to systolic maximum pressures were compared for LD-/EVD-ICP wave pairs [, 16]. LD-ICP Monitoring for Detection of Persisting Communicating Hydrocephalus We further evaluated the feasibility and the diagnostic value of continuous LD-ICP monitoring for the detection of PHCH. After clamping the LD (EVD had already been removed), the course of LD-ICP during continuous monitoring via LD was compared to the presence of ventricle enlargement on CT 24 h after clamping. Sensitivity and specificity of LD-ICP >2 mmhg for detecting hydrocephalus on CT were determined. Statistical Analyses Mean values of EVD-ICP and wave amplitudes were correlated with those of LD-ICP separately for each of the six maneuvers and each of the measurement procedures by using linear regression analysis. Agreement was assessed using the Bland and Altman method. The Shapiro Wilk test was used to prove presence of normal distribution and data was compared by the t test. In normally distributed data, mean ± standard deviation was used to summarize the variables. Otherwise, median and range were used. A value of P <. was considered to be statistically significant. Results Between March 29 and May 21, a total of 43 patients with ICH (n = 22) or SAH (n = 21) fulfilling the inclusion criteria were enrolled in the study. Mean age was 8 ± 22 years and median GCS on admission was 7 (range 3 14). In 2 patients with SAH the aneurysm was coiled, in one patient no aneurysm was found. Initial Hunt and Hess score was 2 and 3 in two patients, respectively, 4 in 7 patients, and in 1 patients. Fisher grade was 3 in 7 patients and 4 in 14 patients. In the 22 ICH patients, 18 patients were treated with intraventricular fibrinolysis (mean total dosage 4 ± 2 mg rt-pa). Parenchymal hematoma volume was 18 ± 11 ml, total IVH volume was 31 ± 22 ml. LD was inserted 62 ± 2 h after admission. At the time of LD placement, 34 patients were sedated (with midazolam and sufentanyl) and mechanically ventilated. In the remaining 9 patients (4 with SAH, with ICH) the level of consciousness was reduced, from drowsy (n = 3) to responsive only to pain stimuli (n = 6). 3 patients were still sedated and ventilated during the attempts of LD clamping while EVD had already been removed. During the overlapping period of EVD and LD (median 3 h, range h), the standardized measuring procedure was performed every 4 h (mean 7 ± 2 times per patient), resulting in a total number of simultaneous

4 pressure readings via EVD and LD of 1,86. The range of ICP values observed was between and 4 mmhg, 1% of the values were above 2 mmhg. Correlation Between EVD-ICP and LD-ICP The mean pressure values measured via LD and EVD during the six different maneuvers are shown in Fig. 1. Atotalof 31 measurements in 43 patients were performed of each maneuver. Mean baseline ICP measured via LD was essentially identical to that measured via EVD in both the 3 (7.7 ±.4 vs. 7.7 ±. mmhg) and (12.8 ±.8 vs ±.7 mmhg) body position. The Queckenstedt maneuver resulted in a simultaneous and equivalent rise in EVD-ICP and LD-ICP (mean 4.3 ±.4 vs. 4.4 ±.3 mmhg), and the release of the vein compression led to a prompt and comparable drop of EVD-ICP and LD-ICP. To account for the independency of measurements, correlation analysis of the two measurement methods was performed separately for each of the six maneuvers, and also separately for each measurement procedure, since multiple measurements (7 ± 2 measurements procedures with six maneuvers, respectively) were done in the same patient. Figure 2 shows the linear regression analysis for the 43 patients pressure pairs for each maneuver of the first measurement procedure performed after LD placement before start of CSF drainage through LD. Regression analysis showed a highly significant and robust correlation between the two methods for each maneuver, with a determination coefficient R 2 ranging between.96 and.99, and a slope ranging between (Fig. 2). Regression analysis for the 6 remaining measurement procedures revealed similar high correlations (R 2 ranging mmhg Queckenstedt veinrelease EVD-ICP LD-ICP Queckenstedt veinrelease Fig. 1 Mean pressure values for EVD-ICP and LD-ICP during different maneuvers. Total number of measurements n = 186, number of measurements per maneuver-phase n = 31. EVD external ventricular drainage, LD lumbar drainage from.9 to.99; slope ranging from.96 to 1.1, data not shown). Using the Bland Altman analysis for all pressure pairs (n = 186), differences between EVD and LD (mean.7 ±.93 mmhg) were homogenously distributed over the range of ICP values and more than 9% of the differences were scattered within the limits of conformity (mean ± 1.96 SD, Fig. 3), indicating agreement between both methods independent of the extent of ICP. Pulse Waves ICP measurements via LD and EVD showed clearly visible oscillating pulse waves in all patients. The time from diastolic minimum to systolic maximum pressure was significantly longer for lumbar pressure wave amplitudes than for EVD wave amplitudes (.4 ±.7 vs..2 ±.6 s, P <.1). Pressure wave amplitudes (difference between diastolic minimum and systolic maximum pressures), recorded by EVD, had a mean amplitude of 7.4 ± 4 mmhg, whereas the pressure waves of the lumbar drain had a mean amplitude of 2.8 ± 1. mmhg. Increases and declines of the EVD-ICP wave amplitudes during different measurement procedures led to changes in the same direction of the LD wave amplitudes. ICP Monitoring Via LD for Detection of Persistent Communicating Hydrocephalus Successful clamping of the LD was achieved in 32 of 43 patients at the first attempt. Eleven patients had at least one unsuccessful clamping attempt; six of them required a ventriculo-peritoneal shunt. In these eleven patients, a total of 42 clamping attempts failed due to ventricular widening on CT. Thus, based on CT criteria, out of a total of 79 clamping attempts, 37 were successful and 42 were not. Sensitivity of LD-ICP rise >2 mmhg for detection of hydrocephalus on CT was 34 of 42 clamping attempts (81%), whereas specificity (no false ICP alarm in case of absence of hydrocephalus on CT) was 37 of 37 clamping attempts (1%). Of note, clinical deterioration (reduction in consciousness) as a result of persistent hydrocephalus was only observed in eight of the 42 unsuccessful clamping attempts (sensitivity 19%; specificity 1%). In the six patients with the need for permanent shunt, sensitivity of LD-ICP >2 mmhg for detection of persisting hydrocephalus on CT was 1% (6/6 patients) and specificity was 1% (37/37 patients). Complications of LD Three patients had signs of meningitis (pleocytosis and elevated lactate levels) without detection of bacteria on days

5 Fig. 2 Correlation between EVD-ICP and LD-ICP of each of the six maneuverphases during the first measurement procedure at baseline, (each n = 43). The figure shows ICP measured simultaneously via EVD and LD with a linear regression line. EVD external ventricular drainage, LD lumbar drainage supine position 4 3 y =,98 x +, degree position 3 3 y =,99 x +, supine position Queckenstedt plateau-phase 4 3 degree position Queckenstedt plateau-phase y =,97 x +, y =,99 x +, supine position 3 seconds after releasing vein compression 3 3 degree position 3 seconds after releasing vein compression y =,98 x +, y =,99 x +, ,, and 9, respectively. The infection could be sufficiently treated with systemic antibiotics. No complications were observed with regard to LD placement, local bleeding, local infection, subdural cerebral hematoma, or catheter breakage. In two patients with severe aneurysmal SAH (Hunt&Hess 4 and, respectively; Fisher grade 4 each) the following phenomenon was observed: after placing LD, baseline LD- ICP and EVD-ICP were identical and maneuvers resulted in equal changes in ICP. However, over the next hours both patients gradually developed a progressive difference between LD-ICP and EVD-ICP during continuous lumbar CSF drainage with a constant rate of 1 ml/h, with a maximum pressure difference reaching 12 mmhg. The Queckenstedt maneuver was now positive (failure of lumbar pressure increase). Both patients developed transient clinical signs of herniation (unilaterally dilated pupil) and on control CT, crowding of all basilar cisterns as a sign of impending herniation was seen. Immediately after clamping off the lumbar drain, flat positioning and reopening of the EVD, the pupils returned to normal size and the pressure gradient disappeared within 2 h. Discussion Although ICP measurement through the lumbar subarachnoid cistern has already been introduced in the diagnostic of normal pressure hydrocephalus in alert patients [, 16], data on the feasibility and accuracy of continuous ICP monitoring via a lumbar catheter in neuro-critically ill patients is lacking. In this prospective study, we included neurocritically ill patients with post-hemorrhagic communicating hydrocephalus (PHCH) after SAH or ICH. At the time of LD placement, the need for further ICP monitoring and

6 Fig. 3 Bland Altman plot. Differences between the two measurement methods are plotted against the averages of the two methods. Horizontal lines are drawn at the mean difference, and at the limits of agreement, which are defined as the mean difference ± 1.96-times the standard deviation of the differences. Analysis includes all single measurements of the six maneuvers and the seven measurement procedures (n = 186). EVD external ventricular drainage, LD lumbar drainage extracorporeal CSF drainage was indicated by ICP increase >2 mmhg (measured via EVD) and ventricular enlargement on CT during EVD clamping in all patients. ICP measurement via LD at the level of Foramen of Monro continuously provided a clearly visible oscillating pressure curve synchronic to the arterial pressure curve. At baseline, i.e., after LD placement but before start of CSF drainage, ICP-LD was essentially identical to the ICP measured via EVD, with a high correlation indicated by a determination coefficient R 2 C.96 and a slope between for each maneuver. Different induced testing maneuvers resulted in equal changes in pressure values of EVD-ICP and LD-ICP. Importantly, this high correlation was observed over a wide range of ICP ( 4 mmhg). The observation that mean lumbar drain amplitudes were about 4.6 mmhg lower might be explained by the elastic nature of the spinal thecal sac as compared to the rigid cranial cavity and by the different length, thickness and rigidness of the lumbar catheter type as compared to EVD [17]. During continuous CSF drainage via LD, while EVD remained closed, EVD-ICP and LD-ICP were also highly comparable in all but two patients with severe SAH. In these two patients with transient signs of herniation, simultaneous measurements revealed a gradual increasing difference with higher EVD-ICP values, indicating the development of a progressive cranio-spinal pressure gradient as the reason for downward herniation. ICP monitoring via LD may have several advantages as compared to direct intracranial measurement via EVD or a parenchyma probe. First, LD placement can be performed at bedside within minutes and is less invasive than the placement of an intracranial device that is always associated with additional brain injury and with periprocedural risks such as parenchyma hemorrhage [18]. Second, in contrast to ICP probes, LD additionally allows therapeutic CSF release for ICP-lowering therapy [19, 2]. Furthermore, there is some evidence from recent studies that CSF drainage via LD may accelerate the subarachnoid blood removal compared to drainage via EVD, with beneficial effects against the development of vasospasm or the need for permanent shunting [7 13]. In the majority of patients with SAH or ICH and severe ventricular involvement, early EVD removal within the first days is not possible due to persisting malresorptive hydrocephalus and the need for prolonged extracorporeal CSF drainage [2, 7 9]. In this situation, LD may represent a simple and less invasive alternative to EVD for CSF drainage as suggested in recent studies [7 9]. However, since no data exists about the accuracy of LD-ICP monitoring, repeated CT scans are usually necessary to monitor hydrocephalus. Thus, especially in patients with reduced consciousness or the need for sedation with limited clinical observation, indirect ICP monitoring by the LD may be helpful, and reduce the frequency of CT scans. In this study, LD-ICP - monitoring had a 1% specificity and 81% sensitivity for detection of hydrocephalus using a cut-off value of LD-ICP >2 mmhg. Of note, in patients with need for VP shunt, sensitivity of LD-ICP >2 mmhg was 1%. In contrast, clinical observation was a poor parameter for detection of hydrocephalus within 24 h of catheter clamping (sensitivity of 19%). On the other hand, LD-ICP <2 mmhg did not necessarily imply the absence of hydrocephalus on CT, as shown in eight of 42 (19%) clamping attempts. Interestingly, none of the three involved patients with those eight clamping attempts unrecognized by LD-ICP monitoring needed a permanent shunt in the end. Moreover, all six patients fulfilling the criteria for VP shunting were detected by LD-ICP -monitoring. Thus, the use of continuous LD-ICP monitoring can be helpful in following the course of post-hemorrhagic hydrocephalus and thereby reduce the frequency of CT scans. Moreover, LD-ICP monitoring may enable earlier recognition of hydrocephalus, before CT can demonstrate ventricular enlargement, and thereby avoid secondary brain damage caused by progressive intracranial hypertension. Certainly, there are several side effects that have to be considered when using LD including lumbar epidural hematoma or abscess, cerebral subdural hematoma, or breakage of the catheter. In our prospective study, we did not observe any of these complications. A frequent

7 complication associated with prolonged LD is bacterial CSF infection. We observed evidence of meningitis in three patients (7%) that could be sufficiently treated with systemic antibiotics. This infection rate is comparable to those reported in other trials using LD in SAH or ICH patients [9, 1, 21] and comparable to that reported for EVD [22]. The most feared complication of LD is potential downward herniation of the brain as a consequence of excessive LD. In this study, two SAH patients developed substantial difference between LD-ICP and EVD- ICP. In these two patients, the amount of CSF drainage via LD probably exceeded the capacity of communication between the intracranial cisterns and the lumbar cistern. Thus, continuous therapeutic LD drainage resulted in a gradually increasing difference between EVD-ICP and LD-ICP (additionally, the Queckenstedt maneuver was positive) leading to impending herniation. Possibly, the accumulation of densely packed blood predominantly in the basal cisterns after SAH might have reduced the craniospinal communication capacity. In this situation, LD of relatively large amounts of CSF (1 ml/h) may exceed the reduced communication capacity leading to a progressive craniospinal pressure gradient and eventual herniation. Although CSF drainage via lumbar catheter may represent a promising approach in SAH to accelerate blood removal and reduce the risk of vasospasm [1 13], these two patients in the present study demonstrated that therapeutic CSF drainage via lumbar catheter is potentially harmful, especially when not performed according to strict safety criteria, i.e., reopened third and fourth ventricles, assignment of basal cisterns and exclusion of supratentorial mass effect. Therefore, simultaneous recording of LD- ICP and EVD-ICP seems reasonable and helpful when performing lumbar CSF drainage in severe SAH patients in order to early detect the development of a pressure gradient. Of note, in these two patients, LD-ICP and EVD- ICP were identical after the placement of LD before starting CSF drainage, and also early after cessation of lumbar CSF release. Thus, the issue of different pressure values only occurred during lumbar CSF overdrainage, but not when LD was closed and only used for ICP monitoring. A similar incidence of transient herniation has been reported in two retrospective studies analysing the safety of LD in SAH patients: reversible clinical signs of herniation were observed in three out of 81 patients [1], and in 1 out of 2 patients [21], respectively. However, it has to be mentioned that retrospective analysis of safety aspects, especially in experienced centers, may underestimate the spectrum and frequency of complications associated with the use of LD. This study has several limitations, mainly due to the highly selective patient collective. We only included patients with SAH or ICH <3 ml and ventricular involvement with acute hydrocephalus who required an EVD. The primary intention for LD was to continue extracorporeal CSF drainage due to communicating hydrocephalus and LD was not placed before communication of ventricular system and subarachnoid space was restored on CT. Patients with supratentorial mass effect and midline shift >. cm, patients with parenchymal hematoma size >3 ml, and SAH patients with all basal cisterns filled with blood were not included. Thus, a large part of ICH and SAH patients with clear need for ICP monitoring and extracorporal CSF drainage were excluded. Whether this technique is also feasible and accurate in such patients or patients with other underlying diseases including severe brain ischemia remains unclear and cannot be answered by this study. We defined an ICP increase >2 mmhg as cut-off for intracranial hypertension. However, clinical deterioration cannot necessarily be equated with ICP increase since neurological worsening can occur before the development of a significant ICP elevation due to critical brain shift [23]. In conclusion, LD seems to represent an accurate and less invasive option for ICP monitoring in patients suffering of PHCH after SAH and ICH. Further studies should investigate whether ICP monitoring via lumbar catheter with or without CSF release is also safe and accurate in patients with other underlying cerebral diseases. References 1. Zazulia AR, Diringer MN, Derdeyn CP, Powers WJ. Progression of mass effect after intracerebral hemorrhage. Stroke. 1999;3: Germanwala AV, Huang J, Tamargo RJ. Hydrocephalus after aneurysmal subarachnoid hemorrhage. Neurosurg Clin N Am. 21;21: Broderick J, Connolly S, Feldmann E, et al. Guidelines for the management of spontaneous intracerebral hemorrhage in adults: 27 update: a guideline from the American Heart Association/ American Stroke Association Stroke Council, High Blood Pressure Research Council, and the Quality of Care and Outcomes in Research Interdisciplinary Working Group. Stroke. 27;38: Bederson JB, Connolly ES Jr, Batjer HH, et al. Guidelines for the management of aneurysmal subarachnoid hemorrhage: a statement for healthcare professionals from a special writing group of the Stroke Council, American Heart Association. Stroke. 29; 4: Smith M. Monitoring intracranial pressure in traumatic brain injury. Anesth Analg. 28;16: Sumer MM, Acikgoz B, Akpinar G. External ventricular drainage for acute obstructive hydrocephalus developing following spontaneous intracerebral haemorrhages. Neurol Sci. 22;23: Huttner HB, Nagel S, Tognoni E, et al. Intracerebral hemorrhage with severe ventricular involvement: lumbar drainage for communicating hydrocephalus. Stroke. 27;38: Huttner HB, Schwab S, Bardutzky J. Lumbar drainage for communicating hydrocephalus after ICH with ventricular hemorrhage. Neurocrit Care. 26;:193 6.

8 9. Staykov D, Huttner HB, Struffert T, et al. Intraventricular fibrinolysis and lumbar drainage for ventricular hemorrhage. Stroke. 29;4: Klimo P Jr, Kestle JR, MacDonald JD, Schmidt RH. Marked reduction of cerebral vasospasm with lumbar drainage of cerebrospinal fluid after subarachnoid hemorrhage. J Neurosurg. 24;1: Hanggi D, Liersch J, Turowski B, Yong M, Steiger HJ. The effect of lumboventricular lavage and simultaneous low-frequency head-motion therapy after severe subarachnoid hemorrhage: results of a single center prospective Phase II trial. J Neurosurg. 28;18: Hanggi D, Eicker S, Beseoglu K, Behr J, Turowski B, Steiger HJ. A multimodal concept in patients after severe aneurysmal subarachnoid hemorrhage: results of a controlled single centre prospective randomized multimodal phase I/II trial on cerebral vasospasm. Cen Eur Neurosurg. 29;7: Kwon OY, Kim YJ, Cho CS, Lee SK, Cho MK. The Utility and Benefits of External Lumbar CSF Drainage after Endovascular Coiling on Aneurysmal Subarachnoid Hemorrhage. J Korean Neurosurg Soc. 28;43: van Gijn J, Hijdra A, Wijdicks EF, Vermeulen M, van Crevel H. Acute hydrocephalus after aneurysmal subarachnoid hemorrhage. J Neurosurg. 198;63: Lenfeldt N, Koskinen LO, Bergenheim AT, Malm J, Eklund A. CSF pressure assessed by lumbar puncture agrees with intracranial pressure. Neurology. 27;68: Eide PK, Brean A. Lumbar cerebrospinal fluid pressure waves versus intracranial pressure waves in idiopathic normal pressure hydrocephalus. Br J Neurosurg. 26;2: Tronnier V, Aschoff A, Hund E, Hampl J, Kunze S. Commercial external ventricular drainage sets: unsolved safety and handling problems. Acta Neurochir (Wien). 1991;11: Gardner PA, Engh J, Atteberry D, Moossy JJ. Hemorrhage rates after external ventricular drain placement. J Neurosurg. 29; 11: Munch EC, Bauhuf C, Horn P, Roth HR, Schmiedek P, Vajkoczy P. Therapy of malignant intracranial hypertension by controlled lumbar cerebrospinal fluid drainage. Crit Care Med. 21;29: Tuettenberg J, Czabanka M, Horn P, et al. Clinical evaluation of the safety and efficacy of lumbar cerebrospinal fluid drainage for the treatment of refractory increased intracranial pressure. J Neurosurg. 29;11: Hoekema D, Schmidt RH, Ross I. Lumbar drainage for subarachnoid hemorrhage: technical considerations and safety analysis. Neurocrit Care. 27;7: Scheithauer S, Burgel U, Ryang YM, et al. Prospective surveillance of drain associated meningitis/ventriculitis in a neurosurgery and neurological intensive care unit. J Neurol Neurosurg Psychiatry. 29;8: Frank JI. Large hemispheric infarction, deterioration, and intracranial pressure. Neurology. 199;4: