The American Journal of Surgery (2014) -, -–-
Intracranial pressure versus cerebral perfusion pressure as a marker of outcomes in severe head injury: a prospective evaluation Efstathios Karamanos, M.D.*, Pedro G. Teixeira, M.D., Emre Sivrikoz, M.D., Stephen Varga, M.D., Konstantinos Chouliaras, M.D., Obi Okoye, M.D., Peter Hammer, M.D., FACS Division of Acute Care Surgery (Trauma, Emergency Surgery, and Surgical Critical Care), University of Southern CaliforniadKeck School of Medicine, Los Angeles County General Hospital (LAC + USC), 2051 Marengo Street, C5L100, Los Angeles, CA 90033-4525, USA
KEYWORDS: Intracranial pressure monitoring; Cerebral perfusion pressure; Severe traumatic brain injury; Outcomes; Mortality
Abstract BACKGROUND: Intracranial pressure (ICP) monitoring is a standard of care in severe traumatic brain injury when clinical features are unreliable. It remains unclear, however, whether elevated ICP or decreased cerebral perfusion pressure (CPP) predicts outcome. METHODS: This is a prospective observational study of patients sustaining severe blunt head injury, admitted to the surgical intensive care unit at the Los Angeles County and University of Southern California Medical Center between January 2010 and December 2011. The study population was stratified according to the findings of ICP and CPP. Primary outcomes were overall in-hospital mortality and mortality because of cerebral herniation. Secondary outcomes were development of complications during the hospitalization. RESULTS: A total of 216 patients met Brain Trauma Foundation guidelines for ICP monitoring. Of those, 46.8% (n 5 101) were subjected to the intervention. Sustained elevated ICP significantly increased all in-hospital mortality (adjusted odds ratio [95% confidence interval]: 3.15 [1.11, 8.91], P 5 .031) and death because of cerebral herniation (adjusted odds ratio [95% confidence interval]: 9.25 [1.19, 10.48], P 5 .035). Decreased CPP had no impact on mortality. CONCLUSIONS: A single episode of sustained increased ICP is an accurate predictor of poor outcomes. Decreased CPP did not affect survival. Ó 2014 Elsevier Inc. All rights reserved.
Traumatic brain injury (TBI) is the leading cause of death in the young adult population. Each year, 1.6 million people across the United States sustain severe traumatic brain injury (sTBI) that results in 52,000 * Corresponding author. Tel.: 11-323-696-5570; fax: 11-323-441-9001. E-mail address: [email protected] Manuscript received August 31, 2013; revised manuscript September 21, 2013 0002-9610/$ - see front matter Ó 2014 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.amjsurg.2013.10.026
deaths, and 80,000 people had a permanent neurologic disability.1 Management of these patients is targeted toward prevention of secondary brain injury that is shown to further deteriorate outcomes.2–7 The cornerstone of management in preventing secondary lesions has traditionally included monitoring of intracranial pressure (ICP) and cerebral perfusion pressure (CPP 5 mean arterial pressure – ICP) via utilization of ICP monitoring devices.
2
The American Journal of Surgery, Vol -, No -, - 2014
Nevertheless, the scientific foundation of preferential end point, ICP vs CPP, to predict outcomes and to trigger interventions remains controversial. Thus, we set out to study survival and in-hospital morbidity measures in patients having severe blunt head injury by stratifying patients according to ICP vs CPP values recorded in the intensive care unit setting. We hypothesized that elevated ICP would be a more reliable predictor of adverse outcomes in patients with blunt sTBI.
pressure on admission (hypotension; ,90 vs R90 mm Hg), international normalized ratio (,1.3 vs R1.3), ISS (%15, 16 to 24, R25), AIS score (R3 vs ,3), Marshall score (.3 vs % 3), and heart rate on admission (tachycardia; .120 vs %120 bpm). The groups were compared for differences in the underlying characteristics using Fisher exact or Pearson chi-square tests as appropriate for categorical variables and Student t test for continuous variables. To identify if ICP or CPP were independent predictors of mortality, a univariate analysis was performed dividing the study groups into alive and deceased. Differences at P less than .2 were inserted into a logistic regression for mortality along with decreased CPP and sustained elevated ICP. The same process was replicated for mortality because of cerebral herniation. Overall in-hospital mortality and mortality because of cerebral herniation were assessed for each study group using logistic regression to adjust for factors that were significant at P less than .05. Adjusted odds ratios (AORs) with 95% confidence intervals (CIs) were derived from the logistic regression. A Forest plot was displayed to examine the impact of each variable in the outcomes. The study population was divided into 4 groups according to their ICP and CPP findings (ICP normal/CPP normal, ICP normal/CPP decreased, ICP elevated/CPP normal, and ICP elevated/ CPP decreased). Differences in mortality between those groups were assessed using the Pearson chi square. Values are reported as mean 6 standard error of the mean for continuous variables and as percentages for categorical variables. All analyses were performed using the Statistical Package for Social Sciences (SPSS Windows), version 12.0 (SPSS, Inc., Chicago, IL).
Methods After approval by the Institutional Review Board, a prospective observational study was conducted of trauma patients with blunt severe TBI (Glasgow Coma Scale score % 8 and/or head Abbreviated Injury Scale [AIS] score R3) meeting the Brain Trauma Foundation inclusion criteria for ICP monitoring (Glasgow Coma Scale score % 8 after resuscitation and abnormal computed tomography scan of the head) admitted to the surgical intensive care unit at Los Angeles County and University of Southern California Medical Center (LAC 1 USC) from January 01, 2010, to December 30, 2011. The LAC 1 USC is an American College of Surgeons–verified Level 1 Trauma Center that manages an average of 5,000 trauma admissions annually. Demographic and clinical variables collected included age, gender, blood pressure on admission, Glasgow Coma Scale (GCS) score on admission, Injury Severity Score (ISS), AIS for each body region (head, chest, abdomen, and extremity), type of intracranial injury, and interventions performed. The Marshall Score for patients with TBI was calculated for each patient. The opening pressure was recorded. The study population was stratified into 2 study arms according to their ICP monitoring findings: elevated sustained ICP, defined as a single episode of ICP greater than 20 mm Hg for more than 15 minutes, and decreased CPP, defined as a single episode of CPP less than 50 mm Hg. An opening pressure greater than 20 mm Hg that lasted for more than 15 minutes was considered as an episode of elevated sustained ICP. All subsequent analyses were performed comparing these groups. Neurosurgical recommendations were reviewed for each enrolled case. The treatment of elevated ICP was collected. The number of episodes of sustained elevated ICP was also documented. Currently, LAC 1 USC Medical Center uses continuous ICP monitoring, and thus, the study captured all episodes of sustained (.15 minutes) and elevated (.20 mm Hg) ICP. Primary outcomes were overall in-hospital mortality and mortality caused by cerebral herniation. Secondary outcome was the development of complications (acute kidney injury [AKI], pneumonia, and deep vein thrombosis [DVT]).
Statistical analysis Continuous variables were dichotomized using clinically relevant cutpoints: age (%55 vs .55 years), systolic blood
Results Overall, 216 trauma patients met the inclusion criteria. A total of 53.2% (115 patients) did not receive an ICP monitoring based on physician’s discretion and were, thus, excluded. A total of 101 patients were available for further analysis. The study population was predominantly men (79.2%) with a mean age of 40.1 years. Hypotension was present in 2.0% of the patients, whereas 25.7% were tachycardic on admission; 43.6% of the patients had a Marshall score greater than 3%, and 38.6% had a GCS of 3 on admission. Patients who experienced an episode of increased sustained ICP had a significantly higher incidence of subarachnoid hemorrhage (Table 1) Table 2 shows the characteristics of the study population stratified by the presence of 1 episode of decreased CPP. The patients who experienced decreased CPP were significantly more injured (ISS R 25: 67% vs 41%, P 5 .015) and had a higher incidence of subarachnoid hemorrhage (68% vs 46%). Elevated ICP is treated according to the implemented treatment protocol at LAC 1 USC Medical Center (Table 3).
E. Karamanos et al. Table 1
ICP vs CPP in sTBI
3
Univariate analysis for the study population stratified by ICP Sustained increased ICP
Demographics Age (mean 6 SEM) Age . 55 y, % (n) Gender (male), % (n) Admission physiology Systolic blood pressure (mean 6 SEM) Hypotension, % (n) Heart rate (mean 6 SEM) Tachycardia, % (n) Respiratory rate (mean 6 SEM) Injury severity indices ISS (mean 6 SEM) ISS % 15, % (n) ISS 5 16–24, % (n) ISS R 25, % (n) Marshall score . 3 GCS 5 3, % (n) Chest AIS score R 3, % (n) Abdomen AIS score R 3, % (n) Extremities AIS score R 3, % (n) Specific head injury Brain contusion, % (n) Subdural hematoma, % (n) Subarachnoid hemorrhage, % (n) Intraparenchymal hemorrhage, % (n) Epidural hematoma, % (n) Midline shift, % (n) Loss of basal cisterns, % (n) Cerebral edema, % (n) Loss of gray/white differential, % (n) Fixed dilated pupils on admission, % (n) Admission laboratory values PTT (mean 6 SEM) PT (mean 6 SEM) INR (mean 6 SEM) INR R 1.3, % (n) Opening ICP (mean 6 SEM) Early nutrition, % (n) Decompressive craniectomy/craniotomy in the first 24 h, % (n) Decompressive craniectomy/craniotomy in the first 4 h, % (n) DVT prophylaxis in the first 24 h, % (n)
Overall (n 5 101)
Yes (n 5 64)
No (n 5 37)
P value
40.1 6 1.9 26.7 (27) 79.2 (80)
37.5 6 2.3 23.4 (15) 79.7 (51)
44.4 6 3.3 32.4 (12) 78.4 (29)
.081 .357 1.000
142.2 6 2.6 2.0 (2) 103.9 6 3.0 25.7 (26) 18.4 6 .8
141.1 6 2.9 1.6 (1) 101.9 6 4.0 23.8 (15) 18.5 6 1.1
144.1 6 5.0 2.9 (1) 107.7 6 4.5 30.6 (11) 18.1 6 1.3
.587 1.000 .359 .485 .778
25.1 6 1.2 14.1 (9) 25.0 (16) 60.9 (39) 50.0 (32) 42.2 (27) 32.8 (21) 7.8 (5) 17.2 (11)
25.5 10.8 43.2 45.9 32.4 32.4 45.9 16.2 18.9
6 1.8 (4) (16) (17) (12) (12) (17) (6) (7)
.828 .638 .076 .153 .099 .399 .207 .191 1.000
74.3 (75) 54.5 (55) 58.4 (59) 36.6 (37) 17.8 (18) 5.9 (6) 30.7 (31) 73.3 (74) 34.7 (35) 22.8 (23)
76.6 (49) 59.4 (38) 70.3 (45) 42.2 (27) 15.6 (10) 7.8 (5) 35.9 (23) 71.9 (46) 37.5 (24) 25.0 (16)
70.3 (26) 45.9 (17) 37.8 (14) 27.0 (10) 21.6 (8) 2.7 (1) 21.6 (8) 75.7 (28) 29.7 (11) 18.9 (7)
.490 .218 .002 .140 .590 .411 .180 .816 .517 .624
30.4 6 .7 15.4 6 .3 1.21 6 .03 27.7 (28) 18 ± 2 84.2 (85) 41.6 (42) 31.7 (32) 77.2 (78)
31.1 6 1.0 15.4 6 .3 1.21 6 .03 28.1 (18) 23 ± 3 82.8 (53) 46.9 (30) 32.8 (21) 81.3 (52)
29.1 6 .9 15.4 6 .5 1.22 6 .05 27.0 (10) 8±1 86.5 (32) 32.4 (12) 29.7 (11) 70.3 (26)
.173 .943 .908 1.000 ,.001 .780 .209 .826 .226
25.3 12.9 31.7 55.4 43.6 38.6 37.6 10.9 17.8
6 1.0 (13) (32) (56) (44) (39) (38) (11) (18)
Bold values were statistically significant. AIS 5 Abbreviated Injury Scale; DVT 5 deep venous thrombosis; GCS 5 Glasgow Coma Scale; ICP 5 intracranial pressure; INR 5 international normalized ratio; ISS 5 Injury Severity Score; PT 5 prothrombin time; PTT 5 partial thromboplastin time.
Decreased CPP is treated initially with bolus intravenous (IV) fluids, and if the patients do not respond, vasopressors are used. The most common therapeutic maneuver used to treat increased ICP was elevation of the head of the bed (91%), followed by the use of sedation (89%) and administration of furosemide (73.4%). Paralytics were the least common method used to treat intracranial hypertension (3%). ICP treatment options are depicted in Table 4. No difference in the treatments was noted when the group with sustained elevated ICP was compared with the group with decreased CPP (Table 5).
Sustained elevated ICP was an independent predictor both for mortality and mortality because of cerebral herniation (AOR [95% CI]: 5.87 [1.19, 28.95], adjusted P 5 .030 and AOR [95% CI]: 10.00 [2.63, 15.46], adjusted P 5 .013, respectively). Decreased CPP did not predict the outcomes (Tables 6 and 7). Sustained elevated ICP was associated with significantly higher both overall mortality and mortality caused by cerebral herniation (AOR [95% CI]: 3.15 [1.11, 8.91], adjusted P 5 .031, and AOR [95% CI]: 9.25 [1.19, 10.48], adjusted P 5 .035, respectively), after adjusting for
The American Journal of Surgery, Vol -, No -, - 2014
4 Table 2
Univariate analysis for the study population stratified by CPP Decreased CPP
Demographics Age (mean 6 SEM) Age . 55 y, % (n) Gender (male), % (n) Admission physiology Systolic blood pressure (mean 6 SEM) Hypotension, % (n) Heart rate (mean 6 SEM) Tachycardia, % (n) Respiratory rate (mean 6 SEM) Injury severity indices ISS (mean 6 SEM) ISS % 15, % (n) ISS 5 16–24, % (n) ISS R 25, % (n) Marshall score . 3 GCS 5 3, % (n) Chest AIS score R 3, % (n) Abdomen AIS score R 3, % (n) Extremities AIS score R 3, % (n) Specific head injury Brain contusion, % (n) Subdural hematoma, % (n) Subarachnoid hemorrhage, % (n) Intraparenchymal hemorrhage, % (n) Epidural hematoma, % (n) Midline shift, % (n) Loss of basal cisterns, % (n) Cerebral edema, % (n) Loss of gray/white differential, % (n) Fixed dilated pupils on admission, % (n) Admission laboratory values PTT (mean 6 SEM) PT (mean 6 SEM) INR (mean 6 SEM) INR R 1.3, % (n) Opening ICP (mean 6 SEM) Early nutrition, % (n) Decompressive craniectomy/craniotomy in the first 24 h, % (n) Decompressive craniectomy/craniotomy in the first 4 h, % (n) DVT prophylaxis in the first 24 h, % (n)
Yes (n 5 57)
No (n 5 44)
P value
37.6 6 2.9 24.6 (14) 78.9 (45)
43.3 6 2.8 29.5 (13) 79.5 (35)
.141 .653 1.000
139.7 6 2.9 0 101.6 6 4.5 23.2 (13) 17.9 6 1.2
145.4 6 4.6 4.7 (2) 107.1 6 3.9 30.2 (13) 18.8 6 1.2
.281 .106 .377 .493 .622
26.9 6 1.3 8.8 (5) 24.6 (14) 66.7 (38) 43.9 (57) 43.9 (25) 38.6 (22) 14.0 (8) 22.8 (13)
23.1 6 1.5 18.2 (8) 40.9 (18) 40.9 (18) 43.2 (19) 31.8 (14) 36.4 (16) 6.8 (3) 11.4 (5)
.056 .231 .089 .015 1.000 .303 .839 .248 .191
75.4 (43) 54.4 (31) 68.4 (39) 40.4 (23) 19.3 (11) 8.8 (5) 36.8 (21) 68.4 (39) 36.8 (21) 22.8 (13)
72.7 (32) 54.5 (24) 45.5 (20) 31.8 (14) 15.9 (7) 2.3 (1) 22.7 (10) 79.5 (35) 31.8 (14) 22.7 (10)
.820 1.000 .026 .411 .795 .228 .191 .260 .676 1.000
31.4 6 1.0 15.6 6 .4 1.23 6 .04 29.8 (17) 23 ± 3 84.2 (48) 40.4 (23) 24.6 (14) 73.7 (42)
29.0 6 .8 15.2 6 .4 1.19 6 .04 25.0 (11) 11 ± 2 84.1 (37) 43.2 (19) 40.9 (18) 81.8 (36)
.081 .495 .433 .658 .003 1.000 .840 .089 .473
Bold values were statistically significant. AIS 5 Abbreviated Injury Scale; CPP 5 cerebral perfusion pressure; DVT 5 deep venous thrombosis; GCS 5 Glasgow Coma Scale; INR 5 international normalized ratio; ISS 5 Injury Severity Score; PT 5 prothrombin time; PTT 5 partial thromboplastin time.
differences between the 2 groups. Decreased CPP had no independent impact on mortality (AOR [95% CI]: 1.78 [.59, 5.40], adjusted P 5 .308, and AOR [95% CI]: 4.27 [.80, 23.47], adjusted P 5 .094, respectively). All patients with elevated ICP (n 5 64) were subjected to treatment (Table 4). A total of 35 patients (54.70%) experienced a single episode of sustained elevated ICP, 11 patients (17.20%) experienced 2 to 4 episodes, 9 (14.05%) experienced 5 to 6 episodes, and 9 (14.05%) experienced 7 or more episodes. A single episode of sustained elevated ICP was significantly associated with higher
mortality rates (48.6% for patients with a single episode vs 16.2% for patients with no sustained elevated ICP, adjusted P 5 .007). Patients with R 7 episodes had a higher incidence of mortality compared with patients with a single episode (66.7% vs 48.6%), but no statistical significance was noticed (adjusted P 5 .435). Of 57 patients who experienced decreased CPP, 75% (43 of 57) were treated. All patients treated received IV fluids as initial treatment, whereas 27 needed vasopressors to treat the decreased CPP. Pressors were used in patients who did not respond to initial IV fluid challenge. Treatment of
E. Karamanos et al. Table 3
ICP vs CPP in sTBI
5
Treatment protocol for elevated ICP
$ Check for and relieve any constricting factors (trach tape, c-collar) that might be compressing the jugular veins and causing venous congestions. $ Elevate head of bed. $ Use appropriate sedation and analgesia. Sedation with propofol preferred but be aware of arterial hypotension. Do not exceed 5 mg/ kg/h, limit therapy to 24 h. Pain control with fentanyl. $ Mild hyperventilation (pCO2 32–36). $ Diuresis and hyperosmolar treatment. Goal serum osmolarity of 300–310 (Na 5 150–155) using either mannitol (.25–1 g/kg bolus for 20 minutes, contraindicated in bleeding and hypotension), Lasix (20 mg bolus), or hypertonic saline (3% or 5%, 250 cc in adults for 20–30 minutes, preferred if bleeding or there is hypotension). $ Drain CSF if ventriculostomy is present. $ Consider trial of medical paralysis. $ Consider barbiturate coma (Pentobarbital 10 mg/kg for 30 minutes, then 5 mg/kg/h ! 3 h, then maintenance 1 mg/kg/h) $ Consider surgical decompression. $ Check bladder pressures and consider decompression if there is significant intra-abdominal hypertension. CSF 5 cerebrospinal fluid; ICP 5 intracranial pressure.
decreased CPP did not seem to impact mortality in this study. The need for vasopressors to treat decreased CPP was associated with increased all in-hospital mortality (AOR [95% CI]: 31.97 [2.77, 46.95], adjusted P 5 .006) but had no impact on mortality because of cerebral herniation (Table 8). Overall, 8% of the population developed pneumonia, 2% developed acute kidney injury, and 2% developed deep venous thrombosis. There was no significant difference between patients who experienced sustained elevated ICP and those who did not (P 5 .958, .692, and .277 for pneumonia, AKI, and DVT, respectively). Similarly, no difference was noted between the patients who experienced an episode of decreased CPP vs those who did not (P 5 .702, .853, and .209 for pneumonia, AKI, and DVT, respectively). Patients who experienced elevated ICP were more prone to develop decreased CPP (80% developed decreased CPP vs 20% who did not). The vast majority (84%) of the patients who did not experience an episode of elevated ICP maintained a normal CPP throughout hospital stay (P , .001).
Table 4
Fig. 1 summarizes the impact of CPP, ICP, and guided therapy in all in-hospital and mortality caused by cerebral herniation. The baseline mortality for patients with normal ICP and CPP was 16%. The mortality rate remained the same when an episode of decreased CPP was added. Isolated increased ICP without evidence of decreased CPP doubled mortality (31%). When an episode of decreased CPP was added to the increased ICP, mortality increased to 45%. CPP did not affect mortality when it was not accompanied by increased ICP but had a synergistic effect when increased ICP was present (P 5 .005, Fig. 2).
Comments Previous studies have underlined the importance of ICP in predicting the outcomes of patients with sTBI. Using a logistic regression on 1,030 prospectively collected sTBI patients, Marmarou et al8 identified an ICP greater than 20 mm Hg as an independent predictor of poor outcomes.
Treatment of elevated ICP (n 5 64)
Hypothermia Elevation of the head of the bed Decompressive surgery Hypertonic saline Furosemide Sedation Pentobarbital coma Paralytics Profound hyperventilation Mannitol ICP 5 intracranial pressure.
Total number (n 5 64)
Alive (n 5 37)
Dead (n 5 27)
P value
9/64 58/64 10/64 40/64 47/64 57/64 15/64 2/64 13/64 35/64
7/37 34/37 5/37 23/37 26/37 31/37 10/37 0/37 6/37 20/37
2/27 24/27 5/27 17/27 21/27 26/27 5/27 2/27 7/27 15/27
.282 .275 .731 .948 .502 .223 .574 .174 .229 .589
(14.1%) (90.6%) (15.6%) (62.5%) (73.4%) (89.1%) (23.4%) (3.1%) (20.3%) (54.7%)
(18.9%) (91.9%) (13.5%) (62.2%) (70.3%) (83.8%) (27.0%) (.0%) (16.2%) (54.1%)
(7.4%) (88.9%) (18.5%) (63.0%) (77.8%) (96.3%) (18.5%) (7.4%) (25.9%) (55.6%)
The American Journal of Surgery, Vol -, No -, - 2014
6 Table 5
Comparison of treatment in patients with elevated ICP and decreased CPP
Hypothermia Elevation of the head of the bed Decompressive surgery Hypertonic saline Furosemide Sedation Pentobarbital coma Paralytics Profound hyperventilation Mannitol Fluids Pressors Sustained elevated ICP Decreased CPP
Sustained elevated ICP (n 5 64)
Decreased CPP (n 5 57)
P value
9/64 58/64 10/64 40/64 47/64 57/64 15/64 2/64 13/64 35/64 37/64 26/64 d 51/64
9/57 47/57 8/57 33/57 41/57 47/57 10/57 2/57 11/57 30/57 35/57 27/57 51/57 d
.790 .185 .806 .605 .853 .297 .424 .906 .889 .821 .688 .455 d d
(14.1%) (90.6%) (15.6%) (62.5%) (73.4%) (89.1%) (23.4%) (3.1%) (20.3%) (54.7%) (57.8%) (40.6%) (79.7%)
(15.8%) (82.5%) (14.0%) (57.9%) (71.9%) (82.5%) (17.5%) (3.5%) (19.3%) (52.6%) (61.4%) (47.4%) (89.5%)
CPP 5 cerebral perfusion pressure; ICP 5 intracranial pressure.
In a study by Saul and Ducker,9 127 sTBI patients with no strict treatment protocol for elevated ICP were compared with 106 similar patients who received treatment under a strict protocol. The study suggested an increase in mortality if ICP was maintained above a threshold of 15 to 25 mm Hg. Likewise, Schreiber et al10 in 2002 using 233 prospectively collected patients with ICP monitoring in place identified an ICP greater than 15 mm Hg as one of the 5 independent predictors of mortality. These studies are consistent with our finding that a single episode of sustained elevated ICP independently predicted all inhospital and herniation-induced mortality. We defined elevated ICP as ICP greater than 20 mm Hg based on the best data currently available.11 In fact, sustained elevated ICP effect on mortality was 26%. In the present study, the 2 comparison groups were similar in their underlying characteristics, and all the patients experiencing an elevated ICP were treated accordingly, suggesting that the increased ICP was not a surrogate marker for sicker patients. Despite efforts to aggressively treat these episodes, 1 single episode significantly increased mortality, suggesting that the most critical part in patient management is to prevent those episodes rather than intervene when they occur. Marshal et al12 studied a consecutive cohort of 100 patients with sTBI and Table 6
noted that signs of cerebral herniation could occur with an ICP as low as 18 mm Hg and consequently proposed an individualization of clinical management to prevent such adverse effects. Our study identified elevated ICP as an independent predictor for cerebral herniation and found no mortality because of herniation in the group that never sustained an episode of elevated ICP. Our results suggest that the presence of sustained elevated ICP significantly increases both all in-hospital mortality and mortality caused by cerebral herniation and that ICP is a better predictor of mortality compared with CPP. Our study demonstrates that ICP should be aggressively treated as opposed to CPP, which, being a derivative of mean arterial pressure and ICP, is subject to changes in the values of these variables. Decreased CPP without the presence of elevated ICP failed to independently predict mortality. The evidence supporting the clinical importance of CPP monitoring in patients with sTBI is scant and controversial. Existing evidence suggests that CPP seems to be important when physiologic parameters instead of clinical outcomes are used as dependent variables.13 Sahuquillo et al14 noted that CPP could not predict episodes of decreased brain oxygenation and that improvement in CPP did not
Independent predictors of all in-hospital mortality in patients with ICP monitoring
Step
Variable
Cumulative R2
AOR (95% CI)
Adjusted P value
1 2 3 4 5
ISS R 25 Intraparenchymal hemorrhage on CT Age . 55 y Sustained elevated ICP Fixed dilated pupils on admission
.396 .631 .676 .716 .746
24.56 9.58 12.84 5.87 5.92
,.001 .006 .019 .030 .046
(21.74, 27.75) (1.91, 48.06) (1.52, 18.37) (1.19, 28.95) (1.03, 33.33)
Other variables entered in the model were ISS , 15, ISS 5 16–24, GCS 5 3, epidural hematoma, loss of basal cisterns, presence of midline shift and subarachnoid hemorrhage on CT, early initiation of nutrition, PTT, and decreased CPP. AUC (95% CI): .958 (.924, .993), P ,.001. AOR 5 adjusted odds ratio; AUC 5 area under the curve; CI 5 confidence interval; CPP 5 cerebral perfusion pressure; CT 5 computed tomography; GCS 5 Glasgow Coma Scale; ICP 5 intracranial pressure; ISS 5 Injury Severity Score; PTT 5 partial thromboplastin time.
E. Karamanos et al. Table 7
ICP vs CPP in sTBI
7
Independent predictors of cerebral herniation in patients with ICP monitoring
Step
Variable
Cumulative R2
AOR (95% CI)
Adjusted P value
1 2 3 4 5
Midline shift on CT PTT Sustained elevated ICP ISS R 25 Intraparenchymal hemorrhage on CT
.158 .296 .347 .415 .456
4.13 1.17 10.00 11.76 4.79
.002 .016 .013 .030 .049
(2.49, (1.03, (2.63, (1.26, (2.94,
34.49) 1.33) 15.46) 19.56) 24.29)
Other variables entered in the model were ISS 5 16–24, epidural hematoma, loss of basal cisterns and presence of midline shift on CT, fixed dilated pupils on admission, decreased CPP, and decompressive craniectomy/craniotomy in the first 24 h. AUC (95% CI): .920 (.868, .973) P , .001. AOR 5 adjusted odds ratio; AUC 5 area under the curve; CI 5 confidence interval; CPP 5 cerebral perfusion pressure; CT 5 computed tomography; ICP 5 intracranial pressure; ISS 5 Injury Severity Score; PTT 5 partial thromboplastin time.
necessarily result in increased oxygen delivery to the brain, suggesting that cerebral blood flow has no association with CPP. In a study by Clifton et al,5 a CPP less than 60 mm Hg was found to be associated with worse outcomes, but a CPP greater than 70 mm Hg did not affect outcomes. Juul et al15 retrospectively analyzed a data set of 427 patients in an international, randomized, double-blinded clinical trial and found that a CPP less than 60 mm Hg was associated with higher incidence of unfavorable outcomes. A high percentage of patients with low CPP had an elevated ICP, which independently predicted poor outcomes, suggesting that CPP may have been a confounding factor in the analysis. Howels et al16 in 2005 prospectively observed the 6-month outcome of 131 sTBI patients who received either CPP- or ICP-targeted therapy and found that patients with intact autoregulation had better outcomes with CPP-targeted therapy; in contrast, patients with impaired autoregulation seemed to benefit from ICP-targeted therapy. Defective autoregulation seems to be a fairly common phenomenon after sTBI.17 Our study failed to show any survival benefit for Table 8
patients who managed to maintain a CPP greater than 50 mm Hg. When CPP was introduced into a stepwise regression model, it was rejected, suggesting that there is no independent association with overall mortality and sTBI-related mortality. Decreased CPP was found in 80% of the patients who experienced sustained elevated ICP, suggesting that the difference in mortality that was observed between the 2 groups (decreased CPP/normal CPP) should be explained by the presence of sustained elevated ICP. In our study group, there was no difference in mortality between patients who did and did not experience decreased CPP as long as ICP remained in the normal range throughout hospitalization. Contant et al18 in a study conducted in 2001 using 189 patients with sTBI found a significant increase in the incidence of adult respiratory distress syndrome if vasoactive medication was used to treat decreased CPP. Even though, in our study, we found no incidents of acute respiratory distress syndrome, the need for vasopressors to treat the decreased CPP was associated with a significant increase in all in-hospital mortality.
Outcomes of patients with ICP monitoring Sustained elevated ICP (n 5 64)
All in-hospital mortality Cerebral herniation
All in-hospital mortality Cerebral Herniation
All in-hospital mortality Cerebral herniation
All in-hospital mortality Cerebral herniation
No sustained elevated ICP (n 5 37)
OR
P value
AOR
3.77 (1.38, 10.30) 1.72 (1.45, 2.04)
.008 .003
3.15 (1.11, 8.91) 9.25 (1.19, 10.48)
Adjusted P value
27/64 (42.2%) 13/64 (20.3%) Decreased CPP (n 5 57)
6/37 (16.2%) 0/37 (.0%) No decreased CPP (n 5 44)
24/57 (42.1%) 11/57 (19.3%) Treated decreased CPP (n 5 43)
9/44 (20.5%) 2/44 (4.5%) Nontreated decreased CPP (n 5 14)
23/43 (53.5%) 11/43 (25.6%) Need for pressors to treat decreased CPP (n 5 27)
1/14 (7.1%) 0/14 (.0%) No need for pressors to treat decreased CPP (n 5 30)
OR
P value
AOR
Adjusted P value
20/27 (74.1%) 10/27 (37.0%)
4/30 (13.3%) 1/30 (3.3%)
18.57 (4.77, 72.34) 17.06 (2.01, 45.14)
,.001 .002
31.97 (2.77, 46.95) 6.59 (.99, 43.93)
.006 .051
OR
P value
AOR
2.83 (1.15, 6.97) 5.02 (1.05, 23.98)
.032 .036
1.78 (.59, 5.40) 4.27 (.80, 23.47)
.031 .035 Adjusted P value .308 .094
OR
P value
AOR
Adjusted P value
14.95 (1.79, 24.60) 1.44 (1.19, 1.74)
.004 .035
6.64 (.44, 10.15) 2.10 (.01, 5.83)
.173 .998
Bold values were statistically significant. AOR 5 adjusted odds ratio; CPP 5 cerebral perfusion pressure; ICP 5 intracranial pressure; OR 5 odds ratio; TBI 5 traumatic brain injury.
The American Journal of Surgery, Vol -, No -, - 2014
8
Figure 1
Impact of ICP device-related variables on outcomes (TBI-related mortality defined as cerebral herniation).
To the best of our knowledge, this study is the only prospective study assessing ICP and CPP as markers of outcomes. However, there are several limitations to the study. Per study design, we did not include penetrating
mechanisms of injury, which by their nature has different pathophysiology. Furthermore, we were unable to document the extended Glasgow Outcome Scale score. The present study might also not have enough patients to identify differences between the groups, and thus, a type II statistical error cannot be safely excluded. In addition, ICP monitoring devices are frequently erroneous, and elevated values might have been interpreted as normal and vice versa. There might be a possibility that the devices have not captured all the episodes. Finally, we are not able to comment on any differences in mortality for patients who sustained elevated ICP for less than 15 minutes. Further research is warranted to delineate if these different patterns of elevated ICP significantly affect mortality.
Conclusions -
Figure 2 All in-hospital mortality and TBI-related mortality according to ICP monitoring findings (TBI-related mortality defined as cerebral herniation).
A single episode of sustained increased ICP suggests a poorer outcome compared with 1 episode of decreased CPP in patients with blunt severe head injury. Increased
E. Karamanos et al.
ICP vs CPP in sTBI
sustained ICP significantly worsened outcomes; in contrast, isolated episodes of decreased CPP did not affect survival.
References 1. Ghajar J. Traumatic brain injury. Lancet 2000;356:923–9. 2. Chesnut RM, Marshall LF, Klauber MR, et al. The role of secondary brain injury in determining outcome from severe head injury. J Trauma 1993;34:216–22. 3. Jones PA, Andrews PJ, Midgley S, et al. Measuring the burden of secondary insults in head-injured patients during intensive care. J Neurosurg Anesthesiol 1994;6:4–14. 4. Struchen MA, Hannay HJ, Contant CF, et al. The relation between acute physiological variables and outcome on the Glasgow Outcome Scale and Disability Rating Scale following severe traumatic brain injury. J Neurotrauma 2001;18:115–25. 5. Clifton GL, Miller ER, Choi SC, et al. Fluid thresholds and outcome from severe brain injury. Crit Care Med 2002;30:739–45. 6. Andrews PJ, Sleeman DH, Statham PF, et al. Predicting recovery in patients suffering from traumatic brain injury by using admission variables and physiological data: a comparison between decision tree analysis and logistic regression. J Neurosurg 2002;97: 326–36. 7. Rosner MJ, Daughton S. Cerebral perfusion pressure management in head injury. J Trauma 1990;30:933–40; discussion 940–1. 8. Marmarou A, Anderson RL, Ward JD. The impact of ICP instability and hypotension on outcome in patients with severe head trauma. J Neurosurg 1991;75:159–66.
9 9. Saul TG, Ducker TB. Intracranial pressure monitoring in patients with severe head injury. Am Surg 1982;48:477–80. 10. Schreiber MA, Aoki N, Scott BG, et al. Determinants of mortality in patients with severe blunt head injury. Arch Surg 2002;137:285–90. 11. Brain Trauma Foundation, American Association of Neurological Surgeons, Congress of Neurological Surgeons, et al. Guidelines for the management of severe traumatic brain injury. VIII. Intracranial pressure thresholds. J Neurotrauma 2007;24(Suppl 1):S55–8. 12. Marshall LF, Barba D, Toole BM, et al. The oval pupil: clinical significance and relationship to intracranial hypertension. J Neurosurg 1983; 58:566–8. 13. Chan KH, Miller JD, Piper IR. Cerebral blood flow at constant cerebral perfusion pressure but changing arterial and intracranial pressure: relationship to autoregulation. J Neurosurg Anesthesiol 1992;4:188–93. 14. Sahuquillo J, Amoros S, Santos A, et al. Does an increase in cerebral perfusion pressure always mean a better oxygenated brain? A study in head-injured patients. Acta Neurochir Suppl 2000;76:457–62. 15. Juul N, Morris GF, Marshall SB, et al. Intracranial hypertension and cerebral perfusion pressure: influence on neurological deterioration and outcome in severe head injury. The Executive Committee of the International Selfotel Trial. J Neurosurg 2000;92:1–6. 16. Howells T, Elf K, Jones PA, et al. Pressure reactivity as a guide in the treatment of cerebral perfusion pressure in patients with brain trauma. J Neurosurg 2005;102:311–7. 17. Sahuquillo J, Amoros S, Santos A, et al. False autoregulation (pseudoautoregulation) in patients with severe head injury. Its importance in CPP management. Acta Neurochir Suppl 2000;76:485–90. 18. Contant CF, Valadka AB, Gopinath SP, et al. Adult respiratory distress syndrome: a complication of induced hypertension after severe head injury. J Neurosurg 2001;95:560–8.
Intracranial pressure versus cerebral perfusion pressure as a marker of outcomes in severe head injury: a prospective evaluation Efstathios Karamanos, M.D.*, Pedro G. Teixeira, M.D., Emre Sivrikoz, M.D., Stephen Varga, M.D., Konstantinos Chouliaras, M.D., Obi Okoye, M.D., Peter Hammer, M.D., FACS Division of Acute Care Surgery (Trauma, Emergency Surgery, and Surgical Critical Care), University of Southern CaliforniadKeck School of Medicine, Los Angeles County General Hospital (LAC + USC), 2051 Marengo Street, C5L100, Los Angeles, CA 90033-4525, USA
KEYWORDS: Intracranial pressure monitoring; Cerebral perfusion pressure; Severe traumatic brain injury; Outcomes; Mortality
Abstract BACKGROUND: Intracranial pressure (ICP) monitoring is a standard of care in severe traumatic brain injury when clinical features are unreliable. It remains unclear, however, whether elevated ICP or decreased cerebral perfusion pressure (CPP) predicts outcome. METHODS: This is a prospective observational study of patients sustaining severe blunt head injury, admitted to the surgical intensive care unit at the Los Angeles County and University of Southern California Medical Center between January 2010 and December 2011. The study population was stratified according to the findings of ICP and CPP. Primary outcomes were overall in-hospital mortality and mortality because of cerebral herniation. Secondary outcomes were development of complications during the hospitalization. RESULTS: A total of 216 patients met Brain Trauma Foundation guidelines for ICP monitoring. Of those, 46.8% (n 5 101) were subjected to the intervention. Sustained elevated ICP significantly increased all in-hospital mortality (adjusted odds ratio [95% confidence interval]: 3.15 [1.11, 8.91], P 5 .031) and death because of cerebral herniation (adjusted odds ratio [95% confidence interval]: 9.25 [1.19, 10.48], P 5 .035). Decreased CPP had no impact on mortality. CONCLUSIONS: A single episode of sustained increased ICP is an accurate predictor of poor outcomes. Decreased CPP did not affect survival. Ó 2014 Elsevier Inc. All rights reserved.
Traumatic brain injury (TBI) is the leading cause of death in the young adult population. Each year, 1.6 million people across the United States sustain severe traumatic brain injury (sTBI) that results in 52,000 * Corresponding author. Tel.: 11-323-696-5570; fax: 11-323-441-9001. E-mail address: [email protected] Manuscript received August 31, 2013; revised manuscript September 21, 2013 0002-9610/$ - see front matter Ó 2014 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.amjsurg.2013.10.026
deaths, and 80,000 people had a permanent neurologic disability.1 Management of these patients is targeted toward prevention of secondary brain injury that is shown to further deteriorate outcomes.2–7 The cornerstone of management in preventing secondary lesions has traditionally included monitoring of intracranial pressure (ICP) and cerebral perfusion pressure (CPP 5 mean arterial pressure – ICP) via utilization of ICP monitoring devices.
2
The American Journal of Surgery, Vol -, No -, - 2014
Nevertheless, the scientific foundation of preferential end point, ICP vs CPP, to predict outcomes and to trigger interventions remains controversial. Thus, we set out to study survival and in-hospital morbidity measures in patients having severe blunt head injury by stratifying patients according to ICP vs CPP values recorded in the intensive care unit setting. We hypothesized that elevated ICP would be a more reliable predictor of adverse outcomes in patients with blunt sTBI.
pressure on admission (hypotension; ,90 vs R90 mm Hg), international normalized ratio (,1.3 vs R1.3), ISS (%15, 16 to 24, R25), AIS score (R3 vs ,3), Marshall score (.3 vs % 3), and heart rate on admission (tachycardia; .120 vs %120 bpm). The groups were compared for differences in the underlying characteristics using Fisher exact or Pearson chi-square tests as appropriate for categorical variables and Student t test for continuous variables. To identify if ICP or CPP were independent predictors of mortality, a univariate analysis was performed dividing the study groups into alive and deceased. Differences at P less than .2 were inserted into a logistic regression for mortality along with decreased CPP and sustained elevated ICP. The same process was replicated for mortality because of cerebral herniation. Overall in-hospital mortality and mortality because of cerebral herniation were assessed for each study group using logistic regression to adjust for factors that were significant at P less than .05. Adjusted odds ratios (AORs) with 95% confidence intervals (CIs) were derived from the logistic regression. A Forest plot was displayed to examine the impact of each variable in the outcomes. The study population was divided into 4 groups according to their ICP and CPP findings (ICP normal/CPP normal, ICP normal/CPP decreased, ICP elevated/CPP normal, and ICP elevated/ CPP decreased). Differences in mortality between those groups were assessed using the Pearson chi square. Values are reported as mean 6 standard error of the mean for continuous variables and as percentages for categorical variables. All analyses were performed using the Statistical Package for Social Sciences (SPSS Windows), version 12.0 (SPSS, Inc., Chicago, IL).
Methods After approval by the Institutional Review Board, a prospective observational study was conducted of trauma patients with blunt severe TBI (Glasgow Coma Scale score % 8 and/or head Abbreviated Injury Scale [AIS] score R3) meeting the Brain Trauma Foundation inclusion criteria for ICP monitoring (Glasgow Coma Scale score % 8 after resuscitation and abnormal computed tomography scan of the head) admitted to the surgical intensive care unit at Los Angeles County and University of Southern California Medical Center (LAC 1 USC) from January 01, 2010, to December 30, 2011. The LAC 1 USC is an American College of Surgeons–verified Level 1 Trauma Center that manages an average of 5,000 trauma admissions annually. Demographic and clinical variables collected included age, gender, blood pressure on admission, Glasgow Coma Scale (GCS) score on admission, Injury Severity Score (ISS), AIS for each body region (head, chest, abdomen, and extremity), type of intracranial injury, and interventions performed. The Marshall Score for patients with TBI was calculated for each patient. The opening pressure was recorded. The study population was stratified into 2 study arms according to their ICP monitoring findings: elevated sustained ICP, defined as a single episode of ICP greater than 20 mm Hg for more than 15 minutes, and decreased CPP, defined as a single episode of CPP less than 50 mm Hg. An opening pressure greater than 20 mm Hg that lasted for more than 15 minutes was considered as an episode of elevated sustained ICP. All subsequent analyses were performed comparing these groups. Neurosurgical recommendations were reviewed for each enrolled case. The treatment of elevated ICP was collected. The number of episodes of sustained elevated ICP was also documented. Currently, LAC 1 USC Medical Center uses continuous ICP monitoring, and thus, the study captured all episodes of sustained (.15 minutes) and elevated (.20 mm Hg) ICP. Primary outcomes were overall in-hospital mortality and mortality caused by cerebral herniation. Secondary outcome was the development of complications (acute kidney injury [AKI], pneumonia, and deep vein thrombosis [DVT]).
Statistical analysis Continuous variables were dichotomized using clinically relevant cutpoints: age (%55 vs .55 years), systolic blood
Results Overall, 216 trauma patients met the inclusion criteria. A total of 53.2% (115 patients) did not receive an ICP monitoring based on physician’s discretion and were, thus, excluded. A total of 101 patients were available for further analysis. The study population was predominantly men (79.2%) with a mean age of 40.1 years. Hypotension was present in 2.0% of the patients, whereas 25.7% were tachycardic on admission; 43.6% of the patients had a Marshall score greater than 3%, and 38.6% had a GCS of 3 on admission. Patients who experienced an episode of increased sustained ICP had a significantly higher incidence of subarachnoid hemorrhage (Table 1) Table 2 shows the characteristics of the study population stratified by the presence of 1 episode of decreased CPP. The patients who experienced decreased CPP were significantly more injured (ISS R 25: 67% vs 41%, P 5 .015) and had a higher incidence of subarachnoid hemorrhage (68% vs 46%). Elevated ICP is treated according to the implemented treatment protocol at LAC 1 USC Medical Center (Table 3).
E. Karamanos et al. Table 1
ICP vs CPP in sTBI
3
Univariate analysis for the study population stratified by ICP Sustained increased ICP
Demographics Age (mean 6 SEM) Age . 55 y, % (n) Gender (male), % (n) Admission physiology Systolic blood pressure (mean 6 SEM) Hypotension, % (n) Heart rate (mean 6 SEM) Tachycardia, % (n) Respiratory rate (mean 6 SEM) Injury severity indices ISS (mean 6 SEM) ISS % 15, % (n) ISS 5 16–24, % (n) ISS R 25, % (n) Marshall score . 3 GCS 5 3, % (n) Chest AIS score R 3, % (n) Abdomen AIS score R 3, % (n) Extremities AIS score R 3, % (n) Specific head injury Brain contusion, % (n) Subdural hematoma, % (n) Subarachnoid hemorrhage, % (n) Intraparenchymal hemorrhage, % (n) Epidural hematoma, % (n) Midline shift, % (n) Loss of basal cisterns, % (n) Cerebral edema, % (n) Loss of gray/white differential, % (n) Fixed dilated pupils on admission, % (n) Admission laboratory values PTT (mean 6 SEM) PT (mean 6 SEM) INR (mean 6 SEM) INR R 1.3, % (n) Opening ICP (mean 6 SEM) Early nutrition, % (n) Decompressive craniectomy/craniotomy in the first 24 h, % (n) Decompressive craniectomy/craniotomy in the first 4 h, % (n) DVT prophylaxis in the first 24 h, % (n)
Overall (n 5 101)
Yes (n 5 64)
No (n 5 37)
P value
40.1 6 1.9 26.7 (27) 79.2 (80)
37.5 6 2.3 23.4 (15) 79.7 (51)
44.4 6 3.3 32.4 (12) 78.4 (29)
.081 .357 1.000
142.2 6 2.6 2.0 (2) 103.9 6 3.0 25.7 (26) 18.4 6 .8
141.1 6 2.9 1.6 (1) 101.9 6 4.0 23.8 (15) 18.5 6 1.1
144.1 6 5.0 2.9 (1) 107.7 6 4.5 30.6 (11) 18.1 6 1.3
.587 1.000 .359 .485 .778
25.1 6 1.2 14.1 (9) 25.0 (16) 60.9 (39) 50.0 (32) 42.2 (27) 32.8 (21) 7.8 (5) 17.2 (11)
25.5 10.8 43.2 45.9 32.4 32.4 45.9 16.2 18.9
6 1.8 (4) (16) (17) (12) (12) (17) (6) (7)
.828 .638 .076 .153 .099 .399 .207 .191 1.000
74.3 (75) 54.5 (55) 58.4 (59) 36.6 (37) 17.8 (18) 5.9 (6) 30.7 (31) 73.3 (74) 34.7 (35) 22.8 (23)
76.6 (49) 59.4 (38) 70.3 (45) 42.2 (27) 15.6 (10) 7.8 (5) 35.9 (23) 71.9 (46) 37.5 (24) 25.0 (16)
70.3 (26) 45.9 (17) 37.8 (14) 27.0 (10) 21.6 (8) 2.7 (1) 21.6 (8) 75.7 (28) 29.7 (11) 18.9 (7)
.490 .218 .002 .140 .590 .411 .180 .816 .517 .624
30.4 6 .7 15.4 6 .3 1.21 6 .03 27.7 (28) 18 ± 2 84.2 (85) 41.6 (42) 31.7 (32) 77.2 (78)
31.1 6 1.0 15.4 6 .3 1.21 6 .03 28.1 (18) 23 ± 3 82.8 (53) 46.9 (30) 32.8 (21) 81.3 (52)
29.1 6 .9 15.4 6 .5 1.22 6 .05 27.0 (10) 8±1 86.5 (32) 32.4 (12) 29.7 (11) 70.3 (26)
.173 .943 .908 1.000 ,.001 .780 .209 .826 .226
25.3 12.9 31.7 55.4 43.6 38.6 37.6 10.9 17.8
6 1.0 (13) (32) (56) (44) (39) (38) (11) (18)
Bold values were statistically significant. AIS 5 Abbreviated Injury Scale; DVT 5 deep venous thrombosis; GCS 5 Glasgow Coma Scale; ICP 5 intracranial pressure; INR 5 international normalized ratio; ISS 5 Injury Severity Score; PT 5 prothrombin time; PTT 5 partial thromboplastin time.
Decreased CPP is treated initially with bolus intravenous (IV) fluids, and if the patients do not respond, vasopressors are used. The most common therapeutic maneuver used to treat increased ICP was elevation of the head of the bed (91%), followed by the use of sedation (89%) and administration of furosemide (73.4%). Paralytics were the least common method used to treat intracranial hypertension (3%). ICP treatment options are depicted in Table 4. No difference in the treatments was noted when the group with sustained elevated ICP was compared with the group with decreased CPP (Table 5).
Sustained elevated ICP was an independent predictor both for mortality and mortality because of cerebral herniation (AOR [95% CI]: 5.87 [1.19, 28.95], adjusted P 5 .030 and AOR [95% CI]: 10.00 [2.63, 15.46], adjusted P 5 .013, respectively). Decreased CPP did not predict the outcomes (Tables 6 and 7). Sustained elevated ICP was associated with significantly higher both overall mortality and mortality caused by cerebral herniation (AOR [95% CI]: 3.15 [1.11, 8.91], adjusted P 5 .031, and AOR [95% CI]: 9.25 [1.19, 10.48], adjusted P 5 .035, respectively), after adjusting for
The American Journal of Surgery, Vol -, No -, - 2014
4 Table 2
Univariate analysis for the study population stratified by CPP Decreased CPP
Demographics Age (mean 6 SEM) Age . 55 y, % (n) Gender (male), % (n) Admission physiology Systolic blood pressure (mean 6 SEM) Hypotension, % (n) Heart rate (mean 6 SEM) Tachycardia, % (n) Respiratory rate (mean 6 SEM) Injury severity indices ISS (mean 6 SEM) ISS % 15, % (n) ISS 5 16–24, % (n) ISS R 25, % (n) Marshall score . 3 GCS 5 3, % (n) Chest AIS score R 3, % (n) Abdomen AIS score R 3, % (n) Extremities AIS score R 3, % (n) Specific head injury Brain contusion, % (n) Subdural hematoma, % (n) Subarachnoid hemorrhage, % (n) Intraparenchymal hemorrhage, % (n) Epidural hematoma, % (n) Midline shift, % (n) Loss of basal cisterns, % (n) Cerebral edema, % (n) Loss of gray/white differential, % (n) Fixed dilated pupils on admission, % (n) Admission laboratory values PTT (mean 6 SEM) PT (mean 6 SEM) INR (mean 6 SEM) INR R 1.3, % (n) Opening ICP (mean 6 SEM) Early nutrition, % (n) Decompressive craniectomy/craniotomy in the first 24 h, % (n) Decompressive craniectomy/craniotomy in the first 4 h, % (n) DVT prophylaxis in the first 24 h, % (n)
Yes (n 5 57)
No (n 5 44)
P value
37.6 6 2.9 24.6 (14) 78.9 (45)
43.3 6 2.8 29.5 (13) 79.5 (35)
.141 .653 1.000
139.7 6 2.9 0 101.6 6 4.5 23.2 (13) 17.9 6 1.2
145.4 6 4.6 4.7 (2) 107.1 6 3.9 30.2 (13) 18.8 6 1.2
.281 .106 .377 .493 .622
26.9 6 1.3 8.8 (5) 24.6 (14) 66.7 (38) 43.9 (57) 43.9 (25) 38.6 (22) 14.0 (8) 22.8 (13)
23.1 6 1.5 18.2 (8) 40.9 (18) 40.9 (18) 43.2 (19) 31.8 (14) 36.4 (16) 6.8 (3) 11.4 (5)
.056 .231 .089 .015 1.000 .303 .839 .248 .191
75.4 (43) 54.4 (31) 68.4 (39) 40.4 (23) 19.3 (11) 8.8 (5) 36.8 (21) 68.4 (39) 36.8 (21) 22.8 (13)
72.7 (32) 54.5 (24) 45.5 (20) 31.8 (14) 15.9 (7) 2.3 (1) 22.7 (10) 79.5 (35) 31.8 (14) 22.7 (10)
.820 1.000 .026 .411 .795 .228 .191 .260 .676 1.000
31.4 6 1.0 15.6 6 .4 1.23 6 .04 29.8 (17) 23 ± 3 84.2 (48) 40.4 (23) 24.6 (14) 73.7 (42)
29.0 6 .8 15.2 6 .4 1.19 6 .04 25.0 (11) 11 ± 2 84.1 (37) 43.2 (19) 40.9 (18) 81.8 (36)
.081 .495 .433 .658 .003 1.000 .840 .089 .473
Bold values were statistically significant. AIS 5 Abbreviated Injury Scale; CPP 5 cerebral perfusion pressure; DVT 5 deep venous thrombosis; GCS 5 Glasgow Coma Scale; INR 5 international normalized ratio; ISS 5 Injury Severity Score; PT 5 prothrombin time; PTT 5 partial thromboplastin time.
differences between the 2 groups. Decreased CPP had no independent impact on mortality (AOR [95% CI]: 1.78 [.59, 5.40], adjusted P 5 .308, and AOR [95% CI]: 4.27 [.80, 23.47], adjusted P 5 .094, respectively). All patients with elevated ICP (n 5 64) were subjected to treatment (Table 4). A total of 35 patients (54.70%) experienced a single episode of sustained elevated ICP, 11 patients (17.20%) experienced 2 to 4 episodes, 9 (14.05%) experienced 5 to 6 episodes, and 9 (14.05%) experienced 7 or more episodes. A single episode of sustained elevated ICP was significantly associated with higher
mortality rates (48.6% for patients with a single episode vs 16.2% for patients with no sustained elevated ICP, adjusted P 5 .007). Patients with R 7 episodes had a higher incidence of mortality compared with patients with a single episode (66.7% vs 48.6%), but no statistical significance was noticed (adjusted P 5 .435). Of 57 patients who experienced decreased CPP, 75% (43 of 57) were treated. All patients treated received IV fluids as initial treatment, whereas 27 needed vasopressors to treat the decreased CPP. Pressors were used in patients who did not respond to initial IV fluid challenge. Treatment of
E. Karamanos et al. Table 3
ICP vs CPP in sTBI
5
Treatment protocol for elevated ICP
$ Check for and relieve any constricting factors (trach tape, c-collar) that might be compressing the jugular veins and causing venous congestions. $ Elevate head of bed. $ Use appropriate sedation and analgesia. Sedation with propofol preferred but be aware of arterial hypotension. Do not exceed 5 mg/ kg/h, limit therapy to 24 h. Pain control with fentanyl. $ Mild hyperventilation (pCO2 32–36). $ Diuresis and hyperosmolar treatment. Goal serum osmolarity of 300–310 (Na 5 150–155) using either mannitol (.25–1 g/kg bolus for 20 minutes, contraindicated in bleeding and hypotension), Lasix (20 mg bolus), or hypertonic saline (3% or 5%, 250 cc in adults for 20–30 minutes, preferred if bleeding or there is hypotension). $ Drain CSF if ventriculostomy is present. $ Consider trial of medical paralysis. $ Consider barbiturate coma (Pentobarbital 10 mg/kg for 30 minutes, then 5 mg/kg/h ! 3 h, then maintenance 1 mg/kg/h) $ Consider surgical decompression. $ Check bladder pressures and consider decompression if there is significant intra-abdominal hypertension. CSF 5 cerebrospinal fluid; ICP 5 intracranial pressure.
decreased CPP did not seem to impact mortality in this study. The need for vasopressors to treat decreased CPP was associated with increased all in-hospital mortality (AOR [95% CI]: 31.97 [2.77, 46.95], adjusted P 5 .006) but had no impact on mortality because of cerebral herniation (Table 8). Overall, 8% of the population developed pneumonia, 2% developed acute kidney injury, and 2% developed deep venous thrombosis. There was no significant difference between patients who experienced sustained elevated ICP and those who did not (P 5 .958, .692, and .277 for pneumonia, AKI, and DVT, respectively). Similarly, no difference was noted between the patients who experienced an episode of decreased CPP vs those who did not (P 5 .702, .853, and .209 for pneumonia, AKI, and DVT, respectively). Patients who experienced elevated ICP were more prone to develop decreased CPP (80% developed decreased CPP vs 20% who did not). The vast majority (84%) of the patients who did not experience an episode of elevated ICP maintained a normal CPP throughout hospital stay (P , .001).
Table 4
Fig. 1 summarizes the impact of CPP, ICP, and guided therapy in all in-hospital and mortality caused by cerebral herniation. The baseline mortality for patients with normal ICP and CPP was 16%. The mortality rate remained the same when an episode of decreased CPP was added. Isolated increased ICP without evidence of decreased CPP doubled mortality (31%). When an episode of decreased CPP was added to the increased ICP, mortality increased to 45%. CPP did not affect mortality when it was not accompanied by increased ICP but had a synergistic effect when increased ICP was present (P 5 .005, Fig. 2).
Comments Previous studies have underlined the importance of ICP in predicting the outcomes of patients with sTBI. Using a logistic regression on 1,030 prospectively collected sTBI patients, Marmarou et al8 identified an ICP greater than 20 mm Hg as an independent predictor of poor outcomes.
Treatment of elevated ICP (n 5 64)
Hypothermia Elevation of the head of the bed Decompressive surgery Hypertonic saline Furosemide Sedation Pentobarbital coma Paralytics Profound hyperventilation Mannitol ICP 5 intracranial pressure.
Total number (n 5 64)
Alive (n 5 37)
Dead (n 5 27)
P value
9/64 58/64 10/64 40/64 47/64 57/64 15/64 2/64 13/64 35/64
7/37 34/37 5/37 23/37 26/37 31/37 10/37 0/37 6/37 20/37
2/27 24/27 5/27 17/27 21/27 26/27 5/27 2/27 7/27 15/27
.282 .275 .731 .948 .502 .223 .574 .174 .229 .589
(14.1%) (90.6%) (15.6%) (62.5%) (73.4%) (89.1%) (23.4%) (3.1%) (20.3%) (54.7%)
(18.9%) (91.9%) (13.5%) (62.2%) (70.3%) (83.8%) (27.0%) (.0%) (16.2%) (54.1%)
(7.4%) (88.9%) (18.5%) (63.0%) (77.8%) (96.3%) (18.5%) (7.4%) (25.9%) (55.6%)
The American Journal of Surgery, Vol -, No -, - 2014
6 Table 5
Comparison of treatment in patients with elevated ICP and decreased CPP
Hypothermia Elevation of the head of the bed Decompressive surgery Hypertonic saline Furosemide Sedation Pentobarbital coma Paralytics Profound hyperventilation Mannitol Fluids Pressors Sustained elevated ICP Decreased CPP
Sustained elevated ICP (n 5 64)
Decreased CPP (n 5 57)
P value
9/64 58/64 10/64 40/64 47/64 57/64 15/64 2/64 13/64 35/64 37/64 26/64 d 51/64
9/57 47/57 8/57 33/57 41/57 47/57 10/57 2/57 11/57 30/57 35/57 27/57 51/57 d
.790 .185 .806 .605 .853 .297 .424 .906 .889 .821 .688 .455 d d
(14.1%) (90.6%) (15.6%) (62.5%) (73.4%) (89.1%) (23.4%) (3.1%) (20.3%) (54.7%) (57.8%) (40.6%) (79.7%)
(15.8%) (82.5%) (14.0%) (57.9%) (71.9%) (82.5%) (17.5%) (3.5%) (19.3%) (52.6%) (61.4%) (47.4%) (89.5%)
CPP 5 cerebral perfusion pressure; ICP 5 intracranial pressure.
In a study by Saul and Ducker,9 127 sTBI patients with no strict treatment protocol for elevated ICP were compared with 106 similar patients who received treatment under a strict protocol. The study suggested an increase in mortality if ICP was maintained above a threshold of 15 to 25 mm Hg. Likewise, Schreiber et al10 in 2002 using 233 prospectively collected patients with ICP monitoring in place identified an ICP greater than 15 mm Hg as one of the 5 independent predictors of mortality. These studies are consistent with our finding that a single episode of sustained elevated ICP independently predicted all inhospital and herniation-induced mortality. We defined elevated ICP as ICP greater than 20 mm Hg based on the best data currently available.11 In fact, sustained elevated ICP effect on mortality was 26%. In the present study, the 2 comparison groups were similar in their underlying characteristics, and all the patients experiencing an elevated ICP were treated accordingly, suggesting that the increased ICP was not a surrogate marker for sicker patients. Despite efforts to aggressively treat these episodes, 1 single episode significantly increased mortality, suggesting that the most critical part in patient management is to prevent those episodes rather than intervene when they occur. Marshal et al12 studied a consecutive cohort of 100 patients with sTBI and Table 6
noted that signs of cerebral herniation could occur with an ICP as low as 18 mm Hg and consequently proposed an individualization of clinical management to prevent such adverse effects. Our study identified elevated ICP as an independent predictor for cerebral herniation and found no mortality because of herniation in the group that never sustained an episode of elevated ICP. Our results suggest that the presence of sustained elevated ICP significantly increases both all in-hospital mortality and mortality caused by cerebral herniation and that ICP is a better predictor of mortality compared with CPP. Our study demonstrates that ICP should be aggressively treated as opposed to CPP, which, being a derivative of mean arterial pressure and ICP, is subject to changes in the values of these variables. Decreased CPP without the presence of elevated ICP failed to independently predict mortality. The evidence supporting the clinical importance of CPP monitoring in patients with sTBI is scant and controversial. Existing evidence suggests that CPP seems to be important when physiologic parameters instead of clinical outcomes are used as dependent variables.13 Sahuquillo et al14 noted that CPP could not predict episodes of decreased brain oxygenation and that improvement in CPP did not
Independent predictors of all in-hospital mortality in patients with ICP monitoring
Step
Variable
Cumulative R2
AOR (95% CI)
Adjusted P value
1 2 3 4 5
ISS R 25 Intraparenchymal hemorrhage on CT Age . 55 y Sustained elevated ICP Fixed dilated pupils on admission
.396 .631 .676 .716 .746
24.56 9.58 12.84 5.87 5.92
,.001 .006 .019 .030 .046
(21.74, 27.75) (1.91, 48.06) (1.52, 18.37) (1.19, 28.95) (1.03, 33.33)
Other variables entered in the model were ISS , 15, ISS 5 16–24, GCS 5 3, epidural hematoma, loss of basal cisterns, presence of midline shift and subarachnoid hemorrhage on CT, early initiation of nutrition, PTT, and decreased CPP. AUC (95% CI): .958 (.924, .993), P ,.001. AOR 5 adjusted odds ratio; AUC 5 area under the curve; CI 5 confidence interval; CPP 5 cerebral perfusion pressure; CT 5 computed tomography; GCS 5 Glasgow Coma Scale; ICP 5 intracranial pressure; ISS 5 Injury Severity Score; PTT 5 partial thromboplastin time.
E. Karamanos et al. Table 7
ICP vs CPP in sTBI
7
Independent predictors of cerebral herniation in patients with ICP monitoring
Step
Variable
Cumulative R2
AOR (95% CI)
Adjusted P value
1 2 3 4 5
Midline shift on CT PTT Sustained elevated ICP ISS R 25 Intraparenchymal hemorrhage on CT
.158 .296 .347 .415 .456
4.13 1.17 10.00 11.76 4.79
.002 .016 .013 .030 .049
(2.49, (1.03, (2.63, (1.26, (2.94,
34.49) 1.33) 15.46) 19.56) 24.29)
Other variables entered in the model were ISS 5 16–24, epidural hematoma, loss of basal cisterns and presence of midline shift on CT, fixed dilated pupils on admission, decreased CPP, and decompressive craniectomy/craniotomy in the first 24 h. AUC (95% CI): .920 (.868, .973) P , .001. AOR 5 adjusted odds ratio; AUC 5 area under the curve; CI 5 confidence interval; CPP 5 cerebral perfusion pressure; CT 5 computed tomography; ICP 5 intracranial pressure; ISS 5 Injury Severity Score; PTT 5 partial thromboplastin time.
necessarily result in increased oxygen delivery to the brain, suggesting that cerebral blood flow has no association with CPP. In a study by Clifton et al,5 a CPP less than 60 mm Hg was found to be associated with worse outcomes, but a CPP greater than 70 mm Hg did not affect outcomes. Juul et al15 retrospectively analyzed a data set of 427 patients in an international, randomized, double-blinded clinical trial and found that a CPP less than 60 mm Hg was associated with higher incidence of unfavorable outcomes. A high percentage of patients with low CPP had an elevated ICP, which independently predicted poor outcomes, suggesting that CPP may have been a confounding factor in the analysis. Howels et al16 in 2005 prospectively observed the 6-month outcome of 131 sTBI patients who received either CPP- or ICP-targeted therapy and found that patients with intact autoregulation had better outcomes with CPP-targeted therapy; in contrast, patients with impaired autoregulation seemed to benefit from ICP-targeted therapy. Defective autoregulation seems to be a fairly common phenomenon after sTBI.17 Our study failed to show any survival benefit for Table 8
patients who managed to maintain a CPP greater than 50 mm Hg. When CPP was introduced into a stepwise regression model, it was rejected, suggesting that there is no independent association with overall mortality and sTBI-related mortality. Decreased CPP was found in 80% of the patients who experienced sustained elevated ICP, suggesting that the difference in mortality that was observed between the 2 groups (decreased CPP/normal CPP) should be explained by the presence of sustained elevated ICP. In our study group, there was no difference in mortality between patients who did and did not experience decreased CPP as long as ICP remained in the normal range throughout hospitalization. Contant et al18 in a study conducted in 2001 using 189 patients with sTBI found a significant increase in the incidence of adult respiratory distress syndrome if vasoactive medication was used to treat decreased CPP. Even though, in our study, we found no incidents of acute respiratory distress syndrome, the need for vasopressors to treat the decreased CPP was associated with a significant increase in all in-hospital mortality.
Outcomes of patients with ICP monitoring Sustained elevated ICP (n 5 64)
All in-hospital mortality Cerebral herniation
All in-hospital mortality Cerebral Herniation
All in-hospital mortality Cerebral herniation
All in-hospital mortality Cerebral herniation
No sustained elevated ICP (n 5 37)
OR
P value
AOR
3.77 (1.38, 10.30) 1.72 (1.45, 2.04)
.008 .003
3.15 (1.11, 8.91) 9.25 (1.19, 10.48)
Adjusted P value
27/64 (42.2%) 13/64 (20.3%) Decreased CPP (n 5 57)
6/37 (16.2%) 0/37 (.0%) No decreased CPP (n 5 44)
24/57 (42.1%) 11/57 (19.3%) Treated decreased CPP (n 5 43)
9/44 (20.5%) 2/44 (4.5%) Nontreated decreased CPP (n 5 14)
23/43 (53.5%) 11/43 (25.6%) Need for pressors to treat decreased CPP (n 5 27)
1/14 (7.1%) 0/14 (.0%) No need for pressors to treat decreased CPP (n 5 30)
OR
P value
AOR
Adjusted P value
20/27 (74.1%) 10/27 (37.0%)
4/30 (13.3%) 1/30 (3.3%)
18.57 (4.77, 72.34) 17.06 (2.01, 45.14)
,.001 .002
31.97 (2.77, 46.95) 6.59 (.99, 43.93)
.006 .051
OR
P value
AOR
2.83 (1.15, 6.97) 5.02 (1.05, 23.98)
.032 .036
1.78 (.59, 5.40) 4.27 (.80, 23.47)
.031 .035 Adjusted P value .308 .094
OR
P value
AOR
Adjusted P value
14.95 (1.79, 24.60) 1.44 (1.19, 1.74)
.004 .035
6.64 (.44, 10.15) 2.10 (.01, 5.83)
.173 .998
Bold values were statistically significant. AOR 5 adjusted odds ratio; CPP 5 cerebral perfusion pressure; ICP 5 intracranial pressure; OR 5 odds ratio; TBI 5 traumatic brain injury.
The American Journal of Surgery, Vol -, No -, - 2014
8
Figure 1
Impact of ICP device-related variables on outcomes (TBI-related mortality defined as cerebral herniation).
To the best of our knowledge, this study is the only prospective study assessing ICP and CPP as markers of outcomes. However, there are several limitations to the study. Per study design, we did not include penetrating
mechanisms of injury, which by their nature has different pathophysiology. Furthermore, we were unable to document the extended Glasgow Outcome Scale score. The present study might also not have enough patients to identify differences between the groups, and thus, a type II statistical error cannot be safely excluded. In addition, ICP monitoring devices are frequently erroneous, and elevated values might have been interpreted as normal and vice versa. There might be a possibility that the devices have not captured all the episodes. Finally, we are not able to comment on any differences in mortality for patients who sustained elevated ICP for less than 15 minutes. Further research is warranted to delineate if these different patterns of elevated ICP significantly affect mortality.
Conclusions -
Figure 2 All in-hospital mortality and TBI-related mortality according to ICP monitoring findings (TBI-related mortality defined as cerebral herniation).
A single episode of sustained increased ICP suggests a poorer outcome compared with 1 episode of decreased CPP in patients with blunt severe head injury. Increased
E. Karamanos et al.
ICP vs CPP in sTBI
sustained ICP significantly worsened outcomes; in contrast, isolated episodes of decreased CPP did not affect survival.
References 1. Ghajar J. Traumatic brain injury. Lancet 2000;356:923–9. 2. Chesnut RM, Marshall LF, Klauber MR, et al. The role of secondary brain injury in determining outcome from severe head injury. J Trauma 1993;34:216–22. 3. Jones PA, Andrews PJ, Midgley S, et al. Measuring the burden of secondary insults in head-injured patients during intensive care. J Neurosurg Anesthesiol 1994;6:4–14. 4. Struchen MA, Hannay HJ, Contant CF, et al. The relation between acute physiological variables and outcome on the Glasgow Outcome Scale and Disability Rating Scale following severe traumatic brain injury. J Neurotrauma 2001;18:115–25. 5. Clifton GL, Miller ER, Choi SC, et al. Fluid thresholds and outcome from severe brain injury. Crit Care Med 2002;30:739–45. 6. Andrews PJ, Sleeman DH, Statham PF, et al. Predicting recovery in patients suffering from traumatic brain injury by using admission variables and physiological data: a comparison between decision tree analysis and logistic regression. J Neurosurg 2002;97: 326–36. 7. Rosner MJ, Daughton S. Cerebral perfusion pressure management in head injury. J Trauma 1990;30:933–40; discussion 940–1. 8. Marmarou A, Anderson RL, Ward JD. The impact of ICP instability and hypotension on outcome in patients with severe head trauma. J Neurosurg 1991;75:159–66.
9 9. Saul TG, Ducker TB. Intracranial pressure monitoring in patients with severe head injury. Am Surg 1982;48:477–80. 10. Schreiber MA, Aoki N, Scott BG, et al. Determinants of mortality in patients with severe blunt head injury. Arch Surg 2002;137:285–90. 11. Brain Trauma Foundation, American Association of Neurological Surgeons, Congress of Neurological Surgeons, et al. Guidelines for the management of severe traumatic brain injury. VIII. Intracranial pressure thresholds. J Neurotrauma 2007;24(Suppl 1):S55–8. 12. Marshall LF, Barba D, Toole BM, et al. The oval pupil: clinical significance and relationship to intracranial hypertension. J Neurosurg 1983; 58:566–8. 13. Chan KH, Miller JD, Piper IR. Cerebral blood flow at constant cerebral perfusion pressure but changing arterial and intracranial pressure: relationship to autoregulation. J Neurosurg Anesthesiol 1992;4:188–93. 14. Sahuquillo J, Amoros S, Santos A, et al. Does an increase in cerebral perfusion pressure always mean a better oxygenated brain? A study in head-injured patients. Acta Neurochir Suppl 2000;76:457–62. 15. Juul N, Morris GF, Marshall SB, et al. Intracranial hypertension and cerebral perfusion pressure: influence on neurological deterioration and outcome in severe head injury. The Executive Committee of the International Selfotel Trial. J Neurosurg 2000;92:1–6. 16. Howells T, Elf K, Jones PA, et al. Pressure reactivity as a guide in the treatment of cerebral perfusion pressure in patients with brain trauma. J Neurosurg 2005;102:311–7. 17. Sahuquillo J, Amoros S, Santos A, et al. False autoregulation (pseudoautoregulation) in patients with severe head injury. Its importance in CPP management. Acta Neurochir Suppl 2000;76:485–90. 18. Contant CF, Valadka AB, Gopinath SP, et al. Adult respiratory distress syndrome: a complication of induced hypertension after severe head injury. J Neurosurg 2001;95:560–8.
