Cerebral edema leading to decompressive craniectomy: An assessment of the preceding clinical and neuromonitoring trends

Cerebral edema leading to decompressive craniectomy: An assessment of the preceding clinical and neuromonitoring trends

Strege, Rainer J

The aim of this study was to examine the pre-operative clinical and neuromonitoring courses in patients with a decompressive craniectomy to assess and to compare clinical and neuromonitoring signs indicating extensive cerebral edema. We conducted a retrospective analysis of the clinical signs and courses of simultaneous monitoring of intracranial pressure (ICP) and cerebral oxygenation (P^sub ti^O^sub 2^) in 26 consecutive patients who were sedated and treated with a decompressive craniectomy due to extensive cerebral edema after aneurysmal subarachnoid hemorrhage (SAH) (n = 20) or severe head injury (SHI) (n = 6). Pathological monitoring trends always preceded clinical deterioration. In 18 of 26 patients extensive cerebral edema was indicated solely by increasing ICP >20 mmHg or decreasing P^sub ti^O^sub 2^

Keywords: Aneurysmal subarachnoid hemorrhage; brain tissue partial pressure of oxygen; decompressive craniectomy; intracranial pressure; neuromonitoring; severe head injury


Continuous monitoring of intracranial pressure (ICP) has become an established part in the management of patients after severe head injury1-4. Monitoring of intraparenchymal partial pressure of oxygen (P^sub ti^O^sub 2^) has also gained increasing significance in the recent years with the main focus on patients with severe head injury (SHI)5-9.

So far, however, there have been only a few published reports with mention of combined ICP and P^sub ti^O^sub 2^ measurement in patients with aneurysmal subarachnoid hemorrhage (SAH)7,10,11. Dings et al.7 reported a clinical experience with P^sub ti^O^sub 2^ monitoring in 43 patients with severe SAH and 57 patients with SHI as long as ICP monitoring was indicated. Khaldi et al.11 examined the correlation between brain tissue oxygen and cerebral NO metabolites in patients with severe SAH in which a microdialysis probe was used, along with a multi-parameter sensor to measure NO metabolites. Gupta et al.10 has recently reported about P^sub ti^O^sub 2^ monitoring in comparison with PET scans in patients in whom also invasive monitoring was applied to obtain measurements of mean arterial blood pressure, central venous pressure, and intracranial pressure as well as continuous jugular bulb oximetry.

ICP and P^sub ti^O^sub 2^ monitoring is aimed towards early recognition of secondary insults prior to the occurrence of clinical signs. To further investigate the time courses of these specific monitoring techniques and their relationship with clinical signs of neurologic deterioration we examined clinical and neuromonitoring courses in SAH and SHI patients with extensive cerebral edema, who were ultimately treated with a decompressive craniectomy. The aim was to assess the utility of combined ICP and P^sub ti^O^sub 2^ monitoring in such patients.



We reviewed the charts of 20 SAH and six SHI patients, all of whom had both ICP and P^sub ti^O^sub 2^ monitoring probes placed until they went on to have a secondary decompressive hemicraniectomy during their intensive care unit (ICU) course. This series includes all patients who were consecutively treated under these criteria in our department between October 1996, when we started to use P^sub ti^O^sub 2^ monitoring, and July 2001. The rationale for neuromonitoring of these patients was their critical condition and risk to develop severe cerebral edema. One-hundred-and-eighteen other SAH and SHI patients who were treated with craniectomy during this period were excluded from our study because they had no or only ICP monitoring before surgical decompression.

The mean age of the SAH patients was 52 + or – 17.3 years (range 35 to 69), and the SHI patients’ mean age was 35 + or -20.7 years (range 9 to 58). The male:female ratio was 7:13 in the SAH patients and 2:1 in the SHI patients. The average Hunt & Hess score was 3.6. Relevant patient data are presented in Tables 1 and 2.

In keeping with generally accepted recommendations 18 of 20 SAH patients were operated on within 24 h after admission to our hospital, the remaining two patients were treated interventionally with GDC-coils within 48 h. All patients were sedated, intubated and mechanically ventilated.


ICP was measured with an intraparenchymal sensor (Camino V420, Camino Laboratories, San Diego, CA, USA). Brain partial pressure of oxygen (P^sub ti^O2) was measured with a Licox microcatheter (GMS, Kiel-Mielkendorf, Germany). ICP and P^sub ti^O2 sensors were placed on the site of injury or aneurysm, or in the right frontal area in cases with diffuse injury or anterior communicating artery aneurysms, in keeping with recommendations from the Houston group13. Both probes were introduced separately via a bolt inserted into a twist drill burrhole. We avoided placing the probes in contused brain or close to an intracranial contusion or hematoma.

ICP and P^sub ti^O2 were recorded continuously and documented by the nursing staff hourly. Mean arterial blood pressure (MAP), paO2 and paCO2 were monitored and cerebral perfusion pressure (CPP) calculated. Arterial blood samples were regularly taken for blood gas analysis. Pupils were controlled at 1-h intervals. In 17 SAH patients daily transcranial Doppler sonography (TCD) examinations of the middle (MCA) and anterior cerebral artery (ACA) were performed and documented. Mean cerebral blood flow velocities (CBFV) >120 cm sec^sup -1^ were considered vasospasm.


Management of SHI patients was performed according to general head injury treatment guidelines14. Standard treatment for both SHI and SAH patients included maintaining normothermia, paO2 > 100 mmHg, paCO2 at 35-38 mmHg, and Hb above 10 g dl^sup -1^. Cerebral perfusion pressure (CPP) was kept above 70 mmHg by reducing elevated ICP and pharmacologically raising MAP with vasopressor agents. Pathological thresholds were defined as ICP values above 20 mmHg and P^sub ti^O2 below 10 mmHg. Intracranial hypertension (ICP > 20 mmHg) was treated by head elevation, moderate hyperventilation (paCO2 32-35 mmHg), cerebrospinal fluid (CSF) drainage, and mannitol infusion. Barbiturates were used as a second tier therapy option in otherwise uncontrollable intracranial hypertension in six of 20 SAH and three of six SHI patients.

If first and second tier treatments failed because of refractory intracranial hypertension or cerebral hypoxia (P^sub ti^O2


The mean monitoring time before the craniectomy was 75 + or -53 h (median: 72, range: 2-163 h) for the SAH patients, and 35 + or -31 h (median: 12, range: 6-88 h) for the SHI patients. Pathological monitoring trends (ICP >20 mmHg and/or P^sub ti^O2

Pathological monitoring trends always preceded clinical signs of deterioration (Tables 1 and 2). In 18 of 26 cases impending extensive cerebral edema was indicated solely by increasing ICP (> 20 mmHg) or decreasing P^sub ti^O2 (

In 12 of 20 SAH patients decrease of P^sub ti^O2 was the first warning sign, in nine patients this occurred clearly before an ICP increase (Table 7), in three patients decrease of P^sub ti^O2 occurred simultaneously with ICP increase. In those nine patients P^sub ti^O2 decreases occurred 13.4 + or -11.6 h (median 11 h, range: 3-38 h; 95% confidence interval (Cl); 4.7-21.3 h) before intracranial hypertension. Figure 2A,B illustrate the exemplary clinical course of SAH patient 11 with left-sided MCA aneurysm before craniectomy.

In contrast to these findings, ICP increases were the first pathological monitoring signs in all six SHI patients (Table 2). In three patients P^sub ti^O2 decreased with a latency, but only in one patient P^sub ti^O2 fell below 10mmHg before surgery. In another SHI patient P^sub ti^O2 never exceeded 10mmHg during a short monitoring course of only 6 h.

Cerebral vasospasm (CBFV >120 cm sec^sup -1^) was observed at the time of pathological monitoring trends in 15 of 17 SAH patients in which CBFV was monitored, and falling P^sub ti^O2, the first warning sign of impending cerebral ischemia in 12 SAH patients coincided with increasing vasospasm in all cases (n = 11 ) also examined by TCD (Table 7).

CT imaging immediately before craniectomy showed circumscribed hypodensities identified by a neuroradiologist and consistent with cerebral infarcts in 14 of 20 SAH patients. In the remaining SAH patients and in all SHI patients CT imaging showed diffuse cerebral swelling/edema but no territorial infarcts.

The decompressive craniectomy was started 14.2 + or – 15.4 h (median: 11, range: 0.5-63 h, 95% Cl; 7.5-20.9 h) after the first warning sign in the SAH patients, and 10.6 + or – 12.2h (median: 7.5, range: 0.5-34 h, 95% Cl; 0.8-11.4 h) in the SHI patients. Extensive brain swelling was confirmed in every case during the operation.


We present this series in which we have analyzed our first five-year experience with combined ICP and PtiO2 measurements to evaluate the utility of such monitoring. Because a common endpoint was needed the monitoring courses before a decompressive craniectomy were analyzed. We chose this endpoint for several reasons: it is often the ‘last’ surgical procedure, which is generally performed after failure of best medical therapy or if further medical therapy is thought not to be successful; the monitoring courses before this procedure are probably the most critical, hence may most appropriately outline the comparison between monitored values and clinical signs.

Although many other SHI and SAH patients during the same time period had combined ICP and P^sub ti^O2 monitoring but no surgical decompression or surgical decompression without previous ICP and P^sub ti^O2 monitoring and who could have potentially served as a comparison group, we feel that due to the lack of a common endpoint, comparison of their clinical and monitoring courses would not add further information to answering our question and confound this analysis. Furthermore, we did not evaluate the surgical decompression as therapeutic procedure itself.


Although retrospective and based on hourly charted values our series clearly shows that pathological monitoring trends preceded clinical deterioration in all 26 cases, and we had not expected to find P^sub ti^O2 being such an early marker in SAH patients. CT scanning confirmed diffuse cerebral edema or infarction as the suspected cause for these trends in all cases and we note that clinical signs of deterioration were only present in eight cases at the time of imaging. These findings confirm our hypothesis that ICP and P^sub ti^O2 monitoring enables recognition of secondary insults prior to the occurrence of clinical signs, which offers a wider window for therapeutic opportunities and interventions such as surgical decompression or angioplasty, for example.

Mindful of some special features in pediatric SHI patients compared with adults, the utility and the consequences of ICP monitoring were nicely presented in a series of 303 consecutively monitored children: 35 children (10%) had surgical procedures performed for raised ICP detected by monitoring and 238 (78%) had nonsurgical measures to control raised ICP16.

Although cerebral partial pressure of oxygen and intracranial pressure are different parameters both can indicate situations of reduced cerebral perfusion, edema or infarction, and raised intracranial pressure1. Our findings may indicate that cerebral hypoxia, possibly due to cerebral vasospasm, precedes the development of intracranial hypertension in SAH patients. In SHI patients cerebral hypoxia and intracranial hypertension appear to coincide suggesting a more rapid and concomitant pathogenesis rather than slowly developing edema which would explain the delay in P^sub ti^O2 decrease in the SHI patients.

Cerebral hyperemia which has been reported to contribute to intracranial hypertension may also account for this finding in the SHI patients. Martin et al.17 described three distinct phases in the post-injury course in 125 patients with SHI:

1. Hypoperfusion during the first 24 h.

2. Hyperemia on post-injury days 1 to 3.

3. Vasospasm during days 4 to 14.

This observation has also been made by Unterberg et al.18 who reported in a series of 53 SHI patients that post-traumatic intracranial hypertension was transient, although most pronounced during postinjury days 1 to 3, when swelling and edema were maximal. All six SHI patients included in our study developed pathological ICP values in the ‘hyperemia’ phase which led to a decompressive craniectomy.

Our data show that cerebral vasospasm and decrease of P^sub ti^O2 indicating possible cerebral hypoxia coincided in SAH patients. Similar to this finding Dings et al.7 presented a patient with bilateral P^sub ti^O2 monitoring after SAH in which severe left-sided vasospasm correlated with increasing left hemispheric hypodensity on serial CT scanning and ipsilateral decrease of P^sub ti^O2 on the left side. Khaldi et al.11 found a decline in tissue NO production accompanied by low P^sub ti^O2 during periods of ischemia, brain infarction, and severe clinical signs of vasospasm. In a controlled study in which P^sub ti^O2 was measured in SAH and non-SAH patients Hoffmann et al.19 showed that in SAH patients decreases of P^sub ti^O2 related to the severity of the bleed; P^sub ti^O2 was significantly lower in aneurysm patients after hemorrhage compared to the control group without SAH.

Local versus global measurements and thresholds

Apart from the pathological conditions which may cause ICP to increase or P^sub ti^O2 to decrease the monitored variables themselves may have contributed to our findings. While P^sub ti^O2 measurement is regional, ICP monitoring is global. Dings et al.7 reported that the PO2-sensitive surface of the probe is 7.1 mm^sup 2^ and its sampling area varies between 7.1 mm^sup 2^ in normal tissue and 23 mm^sup 2^ in damaged tissue if the area related to the probe insertion trauma does not exceed 500 [mu]m. They also stressed that the probe location may affect P^sub ti^O2 readings depending on the insertion depth, which has also been stated by van den Brink et al.5 who observed differences in absolute P^sub ti^O2 values when two probes were used. They concluded that cerebral oxygenation is heterogenous and emphasized the utility of trend evaluation.

There is some discussion about the ischemic PtiO2 threshold and when treatment schemes and strategies should be modified or changed. Most authors agree that for the Licox probe the range between 15 and 10 mmHg indicates a condition, which should prompt correction, and any value below 10 mmHg indicates critical hypoxia5,7-9,13,20. This threshold may be different when a different sensor, e.g. the Paratrend sensor is used, with different insertion depths, sample area and nonpolarographic measurement10,21.

In keeping with this consensus we chose 10 mmHg for our analysis. If we had chosen 15 mmHg as the pathological threshold, critical PtiO2 trends in 12 SAH patients would have started 4.2 h earlier on average. Further analysis using this threshold, however, would have been confounded by inclusion of reversible transient decreases in the 10-15 mmHg PtiO2 ‘gray zone’. In keeping with data from other investigations progressive PtiO2 decrease below 10 mmHg correlated with signs of ischemia on CT7. Gopinath et al.13 reported a decrease of PtiO2 to 0 mmHg, remaining at that level with severe global ischemia. Van Santbrink et al.9 showed a case of brain herniation and a simultaneous PtiO2 decrease to 0 mmHg9. Dings et al.7 presented a case with severe vasospasm after SAH and simultaneous development of a hypodense area (infarction) and decreasing PtiO2 values to less than 10 mmHg and later on near to zero.

These cases and the experience gained from our series support a strategy which should aim at investigating any progressive PtiO2 decline as soon as possible.


Due to our retrospective study design we cannot answer the question whether and to which degree patients experienced a benefit from this type of monitoring or whether their outcome would have been the same without it. Any outcome analysis would have been particularly hindered by two facts. First, the definite time point of surgical decompression was not completely standardized but at the responsible physician’s discretion. Second, the neuromonitoring probes were removed in most patients during surgical decompression and the neuromonitoring was discontinued on at least the decompressed side. We have, however, demonstrated that in SAH patients in particular, PtiO2 monitoring indicated neurological worsening before clinical signs, and a positive effect can at least be postulated. Concerning craniectomy to decompress the acutely injured brain, Coplin recently stressed the importance of early decompression and cited the positive effects on ICP22. We think that only a prospective and randomized study can clarify the real effect of decompressive craniectomy on the outcome of SHI or SAH patients with combined ICP and PtiO2 monitoring.

Our analysis is derived from hourly charted data which does not reveal all details that one encounters with analysis of continuously collected data. Hourly charting of monitored values is a very common clinical practice and probably more frequent than continuous trend display at the bedside. We feel that it provides an appropriate summary of the monitoring course and provides the data which is usually available to the clinician who gets called to assess a monitoring course that shows increasingly critical values and trends.


This series shows the utility of combined ICP and PtiO2 monitoring in patients who develop extensive cerebral edema. Pathological monitoring trends indicate deterioration prior to clinical signs which offers a wider window for therapeutic opportunities and surgical interventions. It appears that PtiO2 monitoring is particularly valuable after aneurysmal SAH and a valuable adjunct to routine ICP monitoring and CT imaging.


1 Alvarez del Castillo M. Monitoring neurologic patients in intensive care. Curr Opin Crit Care 2001; 7: 49-60

2 Becker DP, Miller JD, Ward JD, Greenberg RP, Young HF, Sakalas R. The outcome from severe head injury with early diagnosis and intensive management. J Neurosurg 1977; 47: 491-502

3 Lang EW, Chesnut RM. Intracranial Pressure. Monitoring and Management. Neurosurg Clin N Am 1994; 5: 573-605

4 Piper I, Barnes A, Smith D, Dunn L. The Camino intracranial pressure sensor: Is it optimal technology? An internal audit with a review of current intracranial pressure monitoring technologies. Neurosurgery 2001; 49: 11 58-1164

5 van den Brink WA, van Santbrink H, Avezaat CJ, Hogesteeger C, Jansen W, Kloos LM, Vermeulen J, Mass Al. Monitoring brain oxygen tension in severe head injury: The Rotterdam experience. Acta Neurochir Suppl (Wien) 1998; 71: 1 90-194

6 Carmona Suazo JA, Maas Al, van den Brink WA, van Santbrink H, Steyerberg EW, Avezaat CJ. CO2 reactivity and brain oxygen pressure monitoring in severe head injury. Crit Care Med 2000; 28: 3268-3274

7 Dings J, Meixensberger J, Jager A, Roosen K. Clinical experience with 118 brain tissue oxygen partial pressure catheter probes. Neurosurgery 1998; 43: 1082-1095

8 Kiening KL, Unterberg AW, Bardt TF, Schneider GH, Lanksch WR. Monitoring of cerebral oxygenation in patients with severe head injuries: Brain tissue PO2 versus jugular vein oxygen saturation. J Neurosurg 1996; 85: 751-757

9 van Santbrink H, Maas Al, Avezaat CJ. Continuous monitoring of partial pressure of brain tissue oxygen in patients with severe head injury. Neurosurgery 1996; 38: 21-31

10 Gupta AK, Hutchinson PJ, Fryer T, Al-Rawi PG, Parry DA, Minhas PS, Kett-White R, Kirkpatrick PJ, Mathews JC, Downey S, Aigbirhio F, Clark J, Pickard JD, Menon DK. Measurement of brain tissue oxygenation performed using positron emission tomography scanning to validate a novel monitoring method. J Neurosurg 2002; 96: 263-268

11 Khaldi A, Zauner A, Reinert M, Woodward JJ, Bullock MR. Measurement of nitric oxide and brain tissue oxygen tension in patients after severe subarachnoid hemorrhage. Neurosurgery 2001; 49: 33-38

12 Hunt WE, Hess RM. Surgical risk as related to time of intervention in the repair of intracranial aneurysms. J Neurosurg 1968; 28: 14-20

13 Gopinath SP, Valadka AB, Uzura M, Robertson CS. Comparison of jugular venous oxygen saturation and brain tissue PO2 as monitors of cerebral ischemia after head injury. Crit Care Med 1999; 27: 2337-2345

14 Bullock R, Chesnut RM, Clifton G, Ghajar J, Marion DW, Narayan RK, Newell DW, Pitts LH, Rosner MJ, Wilberger JE. Guidelines for the Management of Severe Head Injury, New York, NY: Brain Trauma Foundation, 1996

15 Lanksch W, Meese W, Kazner E. CT findings in closed head injuries with special reference to contusions. In: Lanksch W, Kazner E, eds. Cranial Computerized Tomography, Springer, Berlin, 1976: pp. 318-328

16 Pople IK, Muhlbauer MS, Sanford RA, Kirk E. Results and complications of intracranial pressure monitoring in 303 children. Pediatr Neurosurg 1995; 23: 64-67

17 Martin NA, Patwardhan RV, Alexander MJ, Africk CZ, Lee JH, Shalmon E, Hovda DA, Becker DP. Characterization of cerebral hemodynamic phases following severe head trauma: Hypoperfusion, hyperemia, and vasospasm. J Neurosurg 1997; 87: 9-19

18 Unterberg A, Kiening K, Schmiedek P, Lanksch W. Long-term observations of intracranial pressure after severe head injury. The phenomenon of secondary rise of intracranial pressure. Neurosurgery 1993; 32: 17-23

19 Huffman WE, Wheeler P, Edelman G, Charbel FT, Torres NJ, Ausman Jl. Hypoxic brain tissue following subarachnoid hemorrhage. Anesthesiology 2000; 92: 442-446 20 Sarrafzadeh AS, Kiening KL, Bardt TF, Schneider GH, Unterberg AW, Lanksch WR. Cerebral oxygenation in contusioned vs. nonlesioned brain tissue: Monitoring of PtiO2 with Licox and Paratrend. Acta Neurochir Suppl (Wien) 1998; 71: 186-189

21 Doppenberg EM, Zauner A, Watson JC, Bullock R. Determination of the ischemic threshold for brain oxygen tension. Acta Neurochir Suppl (Wien) 1998; 71: 166-169

22 Coplin WM. Intracranial pressure and surgical decompression for traumatic brain injury: Biological rationale and protocol for a randomized clinical trial. Neurol Res 2001; 23: 277-290

Rainer J. Strege, Erhard W. Lang, Andreas M. Stark, Heike Scheffner, Michael J. Fritsch, Harald Barth and H. Maximilian Mehdorn

Department of Neurosurgery, University Hospital Kiel, Kiel, Germany

Correspondence and reprint requests to: Dr Rainer J Strege, Department of Neurosurgery, Klinikum Plau am See, 19395 Plau am See, Germany. [rainer_strege@gmx.de] Accepted for publication March 2003.

Copyright Forefront Publishing Group Jul 2003

Provided by ProQuest Information and Learning Company. All rights Reserved