Neurologic outcomes with cerebral oxygen monitoring in traumatic brain injury

Neurologic outcomes with cerebral oxygen monitoring in traumatic brain injury Mary C. McCarthy, MD,a Hugh Moncrief, MD,d Jean M. Sands, RN,e Ronald J. Markert, PhD,b Lawrence C. Hall, DO,f Ian C. Wenker, BS,c Harry L. Anderson III, MD,a A. Peter Ekeh, MD,a Mbaga S. Walusimbi, MD,a Randy J. Woods, MD,a Jonathan M. Saxe, MD,a and Kathryn M. Tchorz, MD,a Dayton, OH

Background. Optimizing cerebral oxygenation is advocated to improve outcome in head-injured patients. The purpose of this study was to compare outcomes in brain-injured patients treated with 2 types of monitors. Methods. Patients with traumatic brain injury and a Glasgow Coma Scale score<8 were identified on admission. A polarographic cerebral oxygen/pressure monitor (Licox) or fiberoptic intracranial pressure monitor (Camino) was inserted. An evidence-based algorithm for treatment was implemented. Elements from the prehospital and emergency department records and the first 10 days of intensive care unit (ICU) care were collected. Glasgow Outcome Scores (GOS) were determined every 3 months after discharge. Results. Over a 3-year period, 145 patients were entered into the study; 81 patients in the Licox group and 64 patients in the Camino group. Mortality, hospital length of stay, and ICU length of stay were equivalent in the 2 groups. More patients in the Licox group achieved a moderate/recovered GOS at 3 months than in the Camino Group (79% vs 61%; P = .09). Conclusion. Three-month GOS revealed a clinically meaningful 18% benefit in patients undergoing cerebral oxygen monitoring and optimization. Six-month outcomes were also better. Unfortunately, these important differences did not reach significance. Continued study of the benefits of cerebral oxygen monitoring is warranted. (Surgery 2009;146:585-91.) From the Division of Trauma, Critical Care and Emergency General Surgery, Department of Surgery,a Department of Internal Medicine and Orthopedics,b and Department of Emergency Medicine,c Wright State University School of Medicine, Dayton; the Neurosurgical Institute, Inc.,d and Clinical Research Center,e Miami Valley Hospital, Dayton; and Radiology Physicians, Inc.,f Dayton, OH

TRAUMATIC BRAIN INJURY (TBI) is a leading cause of death and disability in multiply-injured patients. Monitoring and optimizing cerebral oxygenation has been advocated to improve outcome in headinjured patients. Retrospective studies comparing outcomes with historical controls reported improved survival rates with treatment of cerebral hypoxia assessed by polarographic cerebral oxygen monitoring.1 Our trauma center participated in a nationwide initiative to improve care of the head-injured

Partial funding by the Adam Williams Initiative Foundation, Mission Viejo, California. Accepted for publication June 18, 2009. Reprint requests: Mary C. McCarthy, MD, One Wyoming Street, Suite 7000, Miami Valley Hospital, Dayton, OH 45409. E-mail: [email protected]. 0039-6060/$ - see front matter Ó 2009 Mosby, Inc. All rights reserved. doi:10.1016/j.surg.2009.06.059

patient,2 which afforded us the opportunity to prospectively evaluate patients monitored with a fiberoptic intracranial pressure (ICP) monitor (Camino; Camino Laboratories, San Diego, CA) and compare them with those patients monitored with a combined polarographic cerebral oxygen and pressure monitor (Licox; Integra Lifesciences, Plainsboro, NJ). PATIENTS AND METHODS Patients with severe TBI, defined as a Glasgow Coma Score #8, were identified on admission to our Level I trauma center (Miami Valley Hospital, Dayton, OH). Spiral computed tomographic (CT) scans of the brain were reviewed by the neurosurgeon on call who also decided the need for intracranial monitoring. CT findings entered into the study database were assessed retrospectively by an independent radiologist. A polarographic cerebral oxygen monitor (Licox) or fiberoptic ICP monitor (Camino) was inserted based on attending surgeon SURGERY 585

586 McCarthy et al

Surgery October 2009

Table I. Comparison of demographic and clinical characteristics in Licox and Camino groups Characteristic Age mean ± SD (yrs) Gender Male Female Race White Non-white GCS: EMS mean ± SD GCS: ED mean ± SD Lowest SaO2 mean ± SD Paralytics Emergency medical services Yes No Emergency department Yes No Pupils fixed and dilated Yes No Lowest SBP: ED mean ± SD ICP >20 mean hours ± SD ICP >25 mean hours ± SD Mannitol (maximum g/d) None 1--50 51--100 101--200 >200 ISS (mean ± SD) AIS head (mean ± SD) Hours until death (mean ± SD) Death within 48 hours Yes No Ventilator days (mean ± SD) Hours CPP <60 (mean ± SD) Pneumonia Yes No RBC transfusion Yes No Craniotomy Yes No Basal cisterns at midbrain level Open Partially closed Closed Midline shift at foramen of Monroe (cm) No shift <0.5 0.5--1.5 >1.5

Licox

nz

Camino

nz

33.0 ± 15.4

81

38.8 ± 20.7

64

74% 26%

60 21

73% 27%

47 17

85% 15% 4.00 ± 2.52 4.05 ± 1.80 96.6 ± 5.5

69 12 81 81 64

86% 14% 4.48 ± 3.16 4.58 ± 2.09 96.5 ± 16.8

55 9 64 64 48

17% 83%

12 59

16% 84%

9 49

2% 98%

1

3% 97%

1

23% 77% 107 ± 28 32.7 ± 38.5 21.5 ± 29.8

18 60 78 81 81

9% 91% 107 ± 32 39.8 ± 44.5 27 .4 ± 36.7

5 53 63 64 64

29% 30% 25% 11% 4% 27.1 ± 10.1 4.01 ± 0.60 114 ± 57

23 24 20 9 3 77 77 25

38% 17% 28% 17% 0% 26.1 ± 8.4 4.17 ± 0.64 179 ± 160

24 11 18 11 0 63 63 23

4% 96% 12.7 ± 10.3 19.5 ± 19.2

3 78 81 80

6% 94% 14.1 ± 10.3 16.0 ± 16.1

4 60 64 63

53% 47%

43 38

61% 39%

39 25

35% 65%

28 53

31% 69%

20 44

5% 95%

4 77

20% 80%

13 51

43% 43% 15%

34 34 12

42% 30% 28%

27 19 18

51% 38% 11% 0%

41 30 9 0

42 28 27 3%

27 18 17 2

P value* .06 .93

.90

.31 .11 .99 .83

1.00

.03y

.95 .30 .30 .78

.54 .12 .08 .75

.43 .24 .34

.67 .67 .004y

.10

.024y

(continued)

McCarthy et al 587

Surgery Volume 146, Number 4

Table I. (continued) Subarachnoid hemorrhage in basal cisterns Yes No Intraventricular hemorrhage Yes No Multiple parenchymal lesions Yes No

.97 80% 20

58 15

80% 20%

44 11

45% 55%

34 41

46% 54%

27 32

78% 22%

61 17

80% 20%

49 12

.96

.76

*Chi-square test for categorical characteristics; independent samples t-test for continuous characteristics. yP < .05. zNumber of patients with data available for the indicated variable.

preference. Of the 7 Miami Valley Hospital trauma attendings, 4 trauma attendings were interested in potential benefits offered by the new monitor and 3 were concerned that elevated inspired oxygen levels required to treat cerebral hypoxia would lead to increased pulmonary complications and longer intensive care unit (ICU) length of stay (LOS). The 4 neurosurgeons used Camino monitors in most postoperative patients. Monitors were inserted in the nondominant hemisphere per the preference of the neurosurgeon on call. Patients were admitted to the Trauma Service and an evidence-based algorithm for treatment based on the American Association of Neurosurgery guidelines was implemented to provide uniform care of the head-injured patient.2 In those patients in the Camino group, ICP and cerebral perfusion pressure were used to guide therapy. In patients with a Licox monitor, these values, in conjunction with the brain tissue oxygen level (PbtO2) guided treatment. Oxygenation levels >20 mmHg were targeted, optimizing FiO2 and oxygen delivery. The nurses in the Trauma Clinical Research Division collected data both prospectively and retrospectively from the patient record. Patient demographics, 225 trauma registry elements, and therapeutic interventions from the first 10 days of the ICU hospitalization were abstracted and entered into TBI-Trac (Brain Trauma Foundation). Patients able to travel returned to the neurosurgeon’s office for follow-up examination. Otherwise, phone contact was made with the patient or a family member to determine the Glasgow Outcome Score (GOS) at 3, 6, 12, and 24 months after discharge. Patients were considered lost to followup if they could not be reached after 3 attempts. Categorical variables are reported as frequencies and percents, whereas continuous variables are reported as means and standard deviations. The Chi-square test and the Fisher exact test were

used to compare the Licox group with the Camino group on categorical characteristics. The independent samples t test was used to compare the 2 groups on continuous characteristics. Inferences were made at the .05 level of significance with no correction for multiple comparisons. RESULTS From 2005 to 2008, 145 patients were entered into the study. There were 81 patients in the Licox group and 64 patients in the Camino group. Table I shows that prehospital and emergency department mean Glasgow Coma Score were equivalent between the 2 groups (P = .31 and .11, respectively). There were a greater number of patients with fixed and dilated pupils in the Licox group (23% vs 9%; P = .03). There were no differences in the occurrence of hypotension or desaturation in the emergency department. Head abbreviated injury severity scores (Licox 4.01 ± 0.60 vs Camino 4.17 ± 0.64; P = .12) and Injury Severity Scores (Licox 27.1 ± 10.1 vs Camino 26.1 ± 8.4; P = .54) were not different, reflecting similar injury severity, both of the head and the body as a whole. There were no differences between the 2 groups in the severity signs on brain CT scan: basal cistern patency, subarachnoid hemorrhage in the basal cisterns, intraventricular hemorrhage, or the presence of multiple parenchymal lesions. There was a difference in midline shift >0.5 cm at the Foramen of Monroe, which was 19% greater in the Camino group (P = .024). Over the 10-day monitoring period, Licox patients had a mean of 32.7 ± 38.5 hours with an ICP >20 mmHg, and Camino patients had a mean of 39.8 ± 44.5 hours (P = .30). Maximum mannitol dosage in 24 hours was also equivalent in the 2 groups. More patients had craniotomy in the Camino group (20% vs 5%; P = .004), due to the preference of the neurosurgeons for the Camino monitor. There were no other significant

588 McCarthy et al

differences in demographic or clinical characteristics in the 2 study groups (Table I). Hospital outcomes were compared between the 2 study groups. Table I shows that the Licox group had fewer ventilator days (12.7 vs 14.1), but the difference was not significant (P = .43). Pneumonia occurred in 53% of the Licox patients and in 61% of the Camino patients (P = .34). There was no difference in the number of patients undergoing transfusion despite a greater emphasis on the use of red cells in the patients undergoing brain tissue oxygen monitoring. Table II shows that the ICU LOS was 12.4 ± 7.7 days (mean ± SD) in the Licox group and 12.8 ± 9.9 days in the Camino group (P = .79). Hospital LOS was 22.7 ± 19.8 days in the Licox group and 21.2 ± 19.0 days in the Camino group (P = .64). In the Licox group, mortality was 31% (25/81 patients); in the Camino group, 36% (23/64 patients; P = .52). GOS were assessed at 3, 6, 12, and 24 months after discharge. At 3 months, GOS were available for 43 Licox patients and 28 Camino patients (10 patients were lost to follow-up in the Licox group and 11 patients in the Camino group; 28 patients in the Licox group and 25 patients in the Camino group had died earlier; Table II). The vegetative and severe disabled patients were combined (Licox 9/43 [21%]; Camino 11/28 [39%]) as were the moderately disabled and the recovered patients (Licox 34/43 [79%]; Camino 17/28 [61%]). This 18% difference in recovered or moderately disabled patients favoring the Licox group was not statistically different (P = .09). However, the posthoc statistical power for this comparison was only 36% (a = .05; Licox = 43 patients and Camino = 28 patients; difference between groups = 18%). For an 18% difference to reach statistical significance (P < .05), the sample sizes would have to reach Licox = 58 and Camino = 43 patients. Outcomes at 6 months were also better in the Licox group, but not significant, (total n = 58; 13% difference; P = .24). Too few patients have reached the 12 and 24 month follow-up periods to compare GOS. The craniotomy subgroup was analyzed separately to evaluate its potential as a confounding variable. Craniotomy patients did not differ from those without a craniotomy in hospital LOS, intensive care unit LOS, functional outcomes, or mortality. Similarly, the Licox and Camino patients who had a craniotomy did not differ in these outcomes. DISCUSSION Cerebral monitoring in head-injured patients has focused on the prevention of secondary injury

Surgery October 2009

to the brain owing to impaired perfusion. However, ICP monitoring does not equal cerebral oxygenation.3 There are currently 4 techniques that can be used to assess cerebral oxygenation--jugular venous oxygen saturation, positron emission tomography, near-infrared spectroscopy, and brain tissue oxygenation monitoring (PbtO2). Their strengths and weaknesses are the subject of several recent reviews.4,5 The selection among these forms of oxygenation monitoring is focused on the appropriateness of focal or global monitoring, the location of the monitor in relation to the injury, and the intermittent or continuous nature of the monitoring. The use of PbtO2, as assessed by the intraparenchymal polarographic oxygen probe, has the advantage of directly monitoring the zone of injury and thus earlier detection of perfusion abnormalities that may impact global cerebral oxygenation later. This may also allow salvation of watershed areas of perfusion. However, there is controversy regarding the appropriate placement of such monitors. In our study, the monitor was placed in the nondominant hemisphere, regardless of the site of injury. Placement in the watershed area was not confirmed by CT perfusion or microdialytic techniques.6 In 2 patients previously studied, the monitor was placed directly into the injury; thus, the oxygenation data were not useful for management. Insertion of the probe into noninjured areas yields data equivalent to global assessments of cerebral oxygenation. Consequently, close attention should be paid to the location of the catheter in relation to the injury in interpretation and use of PbtO2 results. Jugular venous oxygen saturation is representative of global cerebral oxygen metabolism, but technically it is difficult to obtain reproducible results. Cerebral tissue oxygenation values of <20 mmHg were targeted for intervention based on guidelines provided by centers that have studied the use of this monitor previously.1 PbtO2 can be increased by increasing the FiO2/PaO2 ratio. An increase in FiO2 was initially used in our patients to augment oxygen delivery.7 Additional interventions such as volume infusion, transfusion, and inotropic support directed at improving cardiac output can also be used to increase oxygen delivery. Several groups have studied the impact of increasing PbtO2 on cerebral oxygen metabolism using microdialysis. However, the results have been conflicting, with some investigators demonstrating a decrease in lactate without a fall in lactate-pyruvate

McCarthy et al 589

Surgery Volume 146, Number 4

Table II. Outcome comparison between Licox and Camino groups Outcome Mortality (hospital discharge) Alive Dead Total Outcome: 3 months Dead Vegetative Severe disability Moderate disability Recovered Lost to follow-up Outcome: 6 months Dead Vegetative Severe disability Moderate disability Recovered Lost to follow-up Outcome: 12 months Dead Vegetative Severe disability Moderate disability Recovered Lost to follow-up Outcome: 24 months Dead Vegetative Severe disability Moderate disability Recovered Lost to follow-up ICULOS (mean ± SD) days HLOS (mean ± SD) days

ny

ny

Licox

Camino

P value* .52

56 25 81

69% 31% 100%

41 23 64

64% 36% 100%

28 2 7 22 12 10

35% 3% 9% 27% 15% 12%

25 1 10 10 7 11

39% 2% 16% 16% 11% 17%

28 0 6 8 21 17

35% 0% 8% 10% 26% 21%

25 0 7 10 6 16

39% 0% 11% 16% 9% 25%

28 0 5 5 12 26

37% 0% 7% 7% 16% 34%

26 0 2 4 12 20

41% 0% 3% 6% 19% 31%

28 0 2 1 0 33

44% 0% 3% 2% 0% 52% 12.4 ± 7.7 22.7 ± 19.8

26 0 1 3 5 26

43% 0% 2% 5% 8% 43% 12.8 ± 9.9 21.2 ± 19.0

.36

.08

.77

.04

.79 .64

*Chi-square test for categorical outcomes; independent samples t test for continuous outcomes. yPercent of population available for follow-up at time indicated.

ratios, interpreted as no improvement in cerebral metabolism.8 Other groups have shown a fall in lactate-pyruvate ratios, representative of an improvement in cerebral metabolism.9 PbtO2 has also been strongly correlated with cerebral blood flow.10 A recent publication from the University of Pennsylvania showed a decrease in patient mortality from 44% to 25% using ICP and PbtO2 monitoring when compared with historical controls.1 In contrast, in our study mortality rates in the 2 study groups were not different: in the Licox group, 31%, and in the Camino group, 36% (P = .52). A study from the University of California at Davis11 showed impaired mitochondrial function in brain tissue from head-injured patients. Restoring mitochondrial function may be

as important as maintaining oxygen delivery in preserving tissue function. Some of the investigators initially expressed concerns regarding the prolonged high levels of inspired oxygen needed to consistently achieve PbtO2 levels >20 mmHg. Oxygen toxicity and prolonged ventilator dependence were anticipated. However, the ventilator days were not significantly different between the 2 groups and the incidence of pneumonia was comparable (Licox, 53% vs Camino, 61%; P = .34). Also, ICU LOS and hospital LOS did not differ between the 2 groups. Our study has several limitations. The principal limitation is sample size. The 18% benefit (i.e., a greater proportion of patients moderately disabled or recovered) for the Licox group is clinically

590 McCarthy et al

meaningful but did not reach significance (P = .09). If 15 more patients were in each group and the 18% difference in benefit sustained, statistical significance (P < .05) would have been reached. Multi-institutional data compilation or continued institutional accrual will be required to reach an adequate sample size. Second, although the 2 groups differed minimally on baseline demographic and clinical characteristics, lack of randomization may have impacted the results. Finally, more sensitive measures of outcomes such as neuropsychological testing may reveal additional differences between the 2 groups not apparent in this study. In conclusion, ICU LOS, hospital LOS, and survival were not significantly different between comparable study groups monitored with a pressure monitor or a combined polarographic oxygen and pressure monitor. However, GOS at 3 months revealed a clinically meaningful 18% greater benefit in those patients undergoing cerebral oxygen monitoring and optimization. Six-month outcomes were also better for the Licox group. Unfortunately, these clinically important differences did not reach significance. Continued study of the benefits of cerebral oxygen monitoring is warranted. The authors acknowledge the following for their assistance in the performance of this study: Brain Trauma Foundation, TBI-trac, Miami Valley Hospital Clinical Research Center, Miami Valley Hospital ICU care providers, Miami Valley Hospital Trauma Registry, Jennifer Brown, RN, Kay Lowe, RN, Colleen McCoart, RN, and Eileen Vagedes, RN.

REFERENCES 1. Stiefel MF, Spiotta A, Gracias VH, Garuffe AM, Guillamondegui O, Maloney-Wilensky E, et al. Reduced mortality rate in patients with severe traumatic brain injury treated with brain tissue oxygen monitoring. J Neurosurg 2005; 103:805-11. 2. Adam Williams Initiative Foundation. Available: http:// www.awtbii.org. Accessed September 18, 2008. 3. Stiefel MF, Udoetuk JD, Spiotta AM, Gracias VH, Goldberg A, Maloney-Wilensky E, et al. Conventional neurocritical care and cerebral oxygenation after traumatic brain injury. J Neurosurg 2006;105:568-75. 4. Rose JC, Neill TA, Hemphill JC. Continuous monitoring of the microcirculation in neurocritical care: an update on brain tissue oxygenation. Curr Opin Crit Care 2006;12:97-102. 5. Haitsma IK, Maas AIR. Monitoring cerebral oxygenation in traumatic brain injury. Prog Brain Res 2007;161:207-16. 6. Dong HK, Dunn IF, Ellegala DB, Litvack ZN. Neuromonitoring in neurological critical care. Neurocrit Care 2006; 04:83-92. 7. Menzel M, Doppenberg EMR, Zauner A, Soukup J, Reinert MM, Bullock R. Increased inspired oxygen

Surgery October 2009

8.

9.

10.

11.

concentration as a factor in improved brain tissue oxygenation and tissue lactate levels after severe human head injury. J Neurosurg 1999;91:1-10. Magnoni S, Ghisoni L, Locatelli M, Caimi M, Colombo A, Valeriani V, et al. Lack of improvement in cerebral metabolism after hyperoxia in severe head injury: a microdialysis study. J Neurosurg 2003;98:952-8. Tolias CM, Rienert M, Seiler R, Gilman D, Scharf A, Bullock MR. Normobaric hyperoxia-induced improvement in cerebral metabolism and reduction in intracranial pressure in patients with severe head injury: a prospective historical cohort-matched study. J Neurosurg 2004;101:435-44. Jaeger M, Soehle M, Schuhmann MU, Winkler D, Meixensberger J. Correlation of continuously monitored regional cerebral blood flow and brain tissue oxygen. Acta Neurochir (Wien) 2005;147:51-6. Verweij BH, Muizelaar JP, Vinas FC, Peterson PL, Xiong Y, Lee CP. Impaired cerebral mitochondrial function after traumatic brain injury in humans. J Neurosurg 2000;93: 815-20.

DISCUSSION Dr Betty J. Tsuei (Cincinnati, OH): Clinical therapy for the management of traumatic brain injury has largely been based on the gold standard of ICP monitoring. However, several studies have suggested that direct brain tissue PO2 measurements may be a useful complement in the treatment of TBI. Dr McCarthy and colleagues evaluated 145 patients with severe TBI who received either ICP or PBO2 directed therapy at the discretion of the attending surgeons. They found that patients whose treatment was based on Licox cerebral oxygen measurements showed a trend for improved functional outcomes at 3 and 6 months. I have a few questions regarding your study. Although you note that head abbreviated injury severity score was similar in both groups, the percentages of both patients in the Camino group with significant midline shift was twice that of those who received Licox monitoring and those patients were also 3-fold more likely to require a craniotomy. These findings suggest that the actual severity of injury in these 2 groups may not be equivalent. Could you comment further on that? At our institution, maintaining brain oxygenation at acceptable levels can be difficult in patients who have concomitant severe lung injury such as pulmonary contusion or adult respiratory distress syndrome. In following the algorithms used to maintain brain oxygenation, did you have patients who did not respond appropriately to simple increases in FiO2? If so, how did you manage them? As you note, some experts have suggested that brain oxygen levels should be measured in the areas adjacent to but not within the injury. In your study, monitoring devices were placed solely in the nondominant hemisphere, regardless of the site of injury. Have you looked at your data taking into account the location of Licox monitor placement with respect to the area of brain injury? Finally, this was a nonrandomized study with monitoring and therapies determined by attending

Surgery Volume 146, Number 4

preference. Do you have any plans for a randomized study, or have these findings already unified the practices of your trauma surgeons? Dr Mary C. McCarthy (Dayton, OH): First, the midline shift resulted in a higher percentage of Caminos. It was kind of the other way around, that because patients had surgery they got a Camino because the neurosurgeons preferred that monitor. It is true that then resulted in a higher proportion of patients with at least surgical injuries in that group. The therapy that we initiated was directed at trying to convince people that outcomes would not be worse in the Licox group so that we could perform a randomized study, because a number of the attendings were convinced that the high levels of FiO2 of 100% frequently and high levels of positive end-expiratory pressure required to achieve this level of cerebral oxygenation would result in more adult respiratory distress syndrome, more pulmonary failure, longer times on the ventilator, and longer hospital stays. So the ability to at least present results that showed comparable and in some cases even improved outcome were important for our background to be able to perform a randomized study. The location of the monitor is another issue of concern, and we have not yet broken that out in relationship to placement in the brain. But it is important to recognize where the monitor is, which is why I showed the CT scan on the 2nd day that shows monitor placement in the region of the brain. Because these are placed blindly, it is often difficult to know exactly where they end up, especially if there are intracerebral hematomas. Placement adjacent to the injury is the optimal zone, so you can help to save that watershed area that may be difficult to perfuse if you do not replace it. Unfortunately, we have not been able to direct it right at that zone exactly. We do hope to perform a randomized study in the future. Dr S. P. Stawicki (Columbus, OH): Did you routinely check gases? Was that actively being measured during the study? Second, were you using any of the technologies to verify the placement of the Licox in an area of perfusion---xenon CT or a comparable technique? Another question is, did you notice any differences in PBO2 quantitatively in terms of transfusions? There’s some evidence that 1 unit of blood may be beneficial, for example, and $2 units of blood may actually decrease PBO2 during transfusions. Was that noted in this study? Was there a role of hypertonic saline in your management protocol? Last, can you elaborate a little bit more about the difference in moderate recovered groups chronologically as the study progressed?

McCarthy et al 591

Dr Mary C. McCarthy (Dayton, OH): We did use routine ABG monitoring as part of the management of these patients as well as the comparative PBtO2s. We have not used xenon scans to check for placement of these catheters but almost all of the patients underwent a 2nd-day CT to evaluate the bleed. And in most of those you can see where the probe is situated. But we have not looked at that specifically as part of our data elements yet. The question regarding transfusion is a good question. There were a lot of comments from people circulating around, and the comments ran from, ‘‘I can’t believe you’re still transfusing these patients’’ to, ‘‘Why don’t you give this patient some more blood?’’ We took a moderate road, and it depended to some extent which of the neurosurgeons was on. And your question might provide a way to address that, and I’ll certainly look at that because we do have the on-call data available. During the period of the study, hypertonic saline was not used. We have since incorporated it into our protocol for management. And we have not used the Rancho scale in our patients routinely. It’s being used primarily in our rehab unit. Dr Fred A. Luchette (Maywood, IL): You didn’t give us any information on alcohol with the progression of the Glasgow Coma Score for the first 3 days. Second, please comment about associated injuries which may have driven the ventilator days more so than the head injury. Dr Mary C. McCarthy (Dayton, OH): I did not incorporate alcohol as an element of our data here, but all of these patients were severe and their Glasgow coma scale <8 was due to head injury. Most of the patients that come in with a mild head injury and small contusion with high alcohol levels, the neurosurgeons will usually delay for 24 hours before putting the monitor in. The need to continue these patients on the ventilator for other injuries was reflected by the injury severity scores, which were equal and not significantly different between the 2 groups. Dr James G. Tyburski (Detroit, MI): Do you have any data on achieving your treatment goals? That is, when you achieve the oxygenation or you achieve the superperfusion pressure, were there differences between those 2? Do you have any actual data on the treatment failures and the treatment successes in these 2 groups? Dr Mary C. McCarthy (Dayton, OH): No, not as an element. But that’s a good question, because sometimes it may be difficult. For example, the patient that had the monitor in the dead area of the brain, obviously that was not able to be used as part of the treatment goals.