AUTHOR(S): Stein, Sherman C., M.D.; Young, Gary S., M.A.; Talucci, Raymond C., M.D.; Greenbaum, Barbara H., M.D.; Ross, Steven E., M.D. Departments of Surgery/Neurosurgery (SCS), Trauma Surgery (RCT, SER), Pediatrics/HematologyOncology (BHG), and Strategic Planning (GSY), Cooper Hospital/University Medical Center, University of Medicine and Dentistry of New Jersey, Robert Wood Johnson Medical School at Camden, Camden, New Jersey Neurosurgery 30; 160-165, 1992 ABSTRACT: We reviewed the records of 253 patients with head injury who required serial computed tomographic (CT) scans; 123 (48.6%) developed delayed brain injury as evidenced by new or progressive lesions after a CT scan. An abnormality in the prothrombin time, partial thromboplastin time, or platelet count at admission was present in 55% of the patients who showed evidence of delayed injury, and only 9% of those whose subsequent CT scans were unchanged or improved from the time of admission (P < 0.001). Among patients developing delayed injury, mean prothrombin time at admission was significantly longer (14.6 vs. 12.6 s, P < 0.001) and partial thromboplastin time was significantly longer (36.9 vs. 29.2 s, P < 0.001) than patients who did not have delayed injury. If coagulation studies at admission were normal, a patient with head injury had a 31% risk of developing delayed insults. This risk rose to almost 85% if at least one clotting test at admission was abnormal (P < 0.001). We conclude that clotting studies at admission are of value in predicting the occurrence of delayed injury. If coagulopathy is discovered in the patient with head injury early follow-up CT scanning is advocated to discover progressive and new intracranial lesions that are likely to occur. KEY WORDS: Central nervous system injury; Coagulation; Computed tomography; Delayed brain injuries; Disseminated intravascular coagulopathy; Head injury The use of serial computed tomographic (CT) scanning in patients with head injury yields a surprising number of new and progressing abnormalities compared with the studies performed at admission. Approximately one-third of the normal CT scans will later demonstrate significant abnormalities (23), and 37 to 66% of the follow-up studies will show worsening of the condition at admission (8,13,21,34,38). Our own experience confirms the high frequency of such delayed cerebral injuries
Downloaded from https://academic.oup.com/neurosurgery/article-abstract/30/2/160/2752927 by University of New England user on 05 March 2018
and insults (35). Figure 1 illustrates a typical case of delayed injury. Delayed brain injury may take several forms: brain swelling (25), delayed extracerebral hematomas (23,24), ischemic (11,31,33), and hemorrhagic (11,38) lesions, as well as enlargement of previously small hematomas and contusions (31). Most of the serial CT scan studies have linked progressive lesions with poor outcome, but the significance of such lesions is not known. Although some authorities have considered delayed injury to be an important factor in neurological deterioration after head injury (8,21), others suggest it is the result, rather than a cause, of severe brain tissue destruction (36). A similarly large proportion of patients with serious head injury have been shown to have abnormal coagulation studies (19,26,30). Disseminated intravascular coagulation (DIC) has likewise been linked to poor outcome (2,9,22,29,42,45). The connection between intravascular coagulopathy and delayed lesions has been suggested (20,45), but it remains unproven. If such a connection exists, abnormal coagulation values at admission would be important predictors of later clinical and radiographic deterioration. It is the purpose of this study to determine whether such a link exists, in an effort to better understand the mechanism of delayed injury. PATIENTS AND METHODS The study population is drawn from patients with closed-head injury admitted to the South Jersey Regional Trauma Center at Cooper Hospital/University Medical Center during the years 1986 to 1988. During that 3-year period, 1018 patients were diagnosed as having moderate or severe closed-head injuries because of prolonged (more than 30 min) unconsciousness, an admission Glasgow coma scale (40) of less than 13, or focal neurological deficit. All patients had admission cranial CT scans and the following coagulation studies: prothrombin time (PT), activated partial thromboplastin time (PTT), and platelet count. Forty-eight patients were deemed brain dead or untreatable before a second CT scan could be performed. Another 20 were excluded from further study because of prolonged hypoxia, hypotension, or serious antecedent illness. Because of an admission GCS of 7 or more and a clinical recovery that was rapid and complete, 697 patients did not require a follow-up cranial CT scan. The remaining 253 patients had a second CT scan performed within 72 hours of admission. It is these patients whose records were retrospectively reviewed in our study. All CT scans were reported by members of the radiology staff, without reference to clinical or laboratory parameters. In each patient, CT scan reports were compared for signs of interval worsening, indicating possible delayed injury. Charts were reviewed to determine clotting parameters at admission. We recognize that the three coagulation studies used are not likely to vary independently. To allow for this collinearity and to assess whether the severity of coagulopathy influences the risk of developing
Redistribution of this article permitted only in accordance with the publisher’s copyright provisions.
Neurosurgery 1992-98 February 1992, Volume 30, Number 2 160 Delayed Brain Injury after Head Trauma: Significance of Coagulopathy Experimental and Clinical Study
based on the degree of abnormality in the three clotting studies (Table 6); the correlation between the coagulopathy index and the probability of developing delayed injury is illustrated in Figure 4. For this model, the regression coefficient is -1.149, χ2 is 39.2, significant at P < 0.001.
RESULTS Two hundred fifty-three patients, ranging in age from 3 to 74 years, qualified for the study. Of these, 123 (48.6%) developed delayed injury during their hospitalization. Thirty-six patients developed delayed insults after an initially normal CT study; these patients represent 34.3% of those whose first CT scan was normal. Among the 92 patients whose initial cranial CT scan showed abnormalities, 32 had new lesions on follow-up studies, and 56 cases experienced worsening of abnormalities that were evident at admission. CT findings at admission are shown in Table 1. Table 2 details progressive or new lesions documented in the group with delayed injury. At least one admission coagulation test was abnormal in 55% of the patients with delayed injuries, as opposed to 9% of those not demonstrating this finding (Table 3). A similar low incidence of coagulopathy (11%) was seen in the 697 patients not requiring a second CT scan. The association between abnormal clotting studies at admission and later development of delayed injury is highly significant (χ2 = 39.2, df = 1, P < 0.001). If a second set of coagulation studies was obtained within 12 hours of admission, the incidence of abnormalities increase to 74% of those who developed delayed cerebral insults. Table 3 also shows a breakdown of abnormalities in individual clotting tests. It is evident that PT is most frequently affected in delayed injury. However, the difference is statistically significant for all three coagulation studies. Mean values of the clotting tests in patients with and without delayed insults are given in Table 4. There are highly significant associations between both prolonged PT and PTT and the occurrence of delayed lesions (P < 0.001). The association between platelet count and delayed injury does not reach statistical significance. The logistic regression model shows a very strong correlation between PT and the probability of developing delayed insults (Fig. 2). For this model, the regression coefficient is -0.4392, χ2 is 32.3, significant at P < 0.001. PTTs show a similar, if less striking, relationship to the probability of delayed insults (Fig. 3). This model has a regression coefficient of -0.0195, χ2 of 6.6, and is significant at P < 0.01. The value of clotting studies at admission in predicting the genesis of delayed injury is as follows: Less than one-third (31.1%) of the patients with normal PT, PTT, and platelet counts develop delayed lesions. If at least one of these studies discloses abnormality, the incidence of delayed injury approaches 85% (84.8%). This difference is highly significant (P < 0.001). In the regression model, the relationship between single or multiple clotting abnormalities and the risk of delayed insults is shown in Table 5. A coagulopathy index was developed,
DISCUSSION What causes deterioration or death in the headinjured patient? Some patients have primary brain lesions that are so overwhelming they prevent any hope of recovery. These patients were excluded from our study. In some patients, it may be an unrecognized intracranial lesion that causes delayed complications. None of these patients is included in the present study because emergency CT scans were obtained at admission in every case. Another cause excluded by the entry criteria of our study is secondary brain injury resulting from overwhelming systemic damage or an antecedent disease. In the remaining patients, deterioration and poor outcome is most likely the result of secondary brain injury occurring because of intracranial insults. It is this patient group that usually requires serial CT scanning on clinical grounds and that constitutes our study group. It must be conceded that our patients do not represent all cases of head injury studied prospectively. The decision to obtain a second CT scan was made on clinical grounds, usually because the patients did not recover rapidly or completely. Our data show a highly significant association between the signs of a coagulation disorder at admission and a later development of cerebral injury in head-injured patients. Because we have not routinely tested for platelet function, fibrinogen consumption, or secondary fibrinolysis, we cannot invoke the mechanism of DIC with assurance. Nevertheless, such evidence has been obtained by others and is seen in 60% of head-injured patients (14, 19,26,30) . Several investigators have supported the concept of local intravascular coagulation and fibrinolysis in the injured brain (6,28,41). It is well documented that DIC can affect the brain (10). Both hemorrhage (7,18) and ischemia as a result of vascular occlusion (3,5,16) have been reported. A mechanism of microvascular occlusions with secondary lysis and hemorrhage into damaged cerebral tissue has been hypothesized for the traumatized brain (12,32,44). The mechanism of diffuse brain swelling seen in 28 of our patients is not obvious. Intravascular coagulopathy may disrupt the blood-brain barrier and damage vascular walls. Reperfusion of blood vessels so damaged may cause edema, perhaps in combination with hyperemic brain swelling. A detailed discussion of the causes of traumatic brain swelling is contained in the study by Lobato and coworkers (25). Mechanisms other than DIC may account for coagulopathy in head injury. Hypothalamic damage may alter levels of clotting factors (37). Clotting mechanisms may also be influenced by vasoactive peptides, such as endothelin (46) and Von Willebrand multimers (27), released in cerebral trauma. A number of studies have linked the appearance of
Downloaded from https://academic.oup.com/neurosurgery/article-abstract/30/2/160/2752927 by University of New England user on 05 March 2018
Redistribution of this article permitted only in accordance with the publisher’s copyright provisions.
delayed injury, logistic regression models were used (4) . This method involves a strategy of allocating patients into groups manifesting different degrees of coagulopathy, then regressing this binary group variable onto the odds of delayed injury being present.
California, October 20-25, 1990. Received for publication, May 10, 1990; accepted, final form, September 4, 1991. Reprint requests: Sherman C. Stein, M.D., Cooper Hospital/UMC, Three Cooper Plaza, Suite 411, Camden, NJ 08103. REFERENCES: (1-47) 1.
2.
3.
4. 5.
6.
7.
8.
9.
10.
11.
12. ACKNOWLEDGMENT The authors wish to thank the Institute of Brain Injury Research and Training at Mediplex RehabCamden for its support of our research activities. Presented in part at the 40th Annual Meeting of the Congress of Neurological Surgeons, Los Angeles,
Downloaded from https://academic.oup.com/neurosurgery/article-abstract/30/2/160/2752927 by University of New England user on 05 March 2018
13.
14.
Auer L: Disturbances of the coagulatory system in patients with severe cerebral trauma: I. Acta Neurochir (Wien) 43:51-59, 1978. Auer LM, Oh E: Disturbances of the coagulatory system in patients with severe cerebral trauma: II. Platelet Function. Acta Neurochir (Wien) 49:219-226, 1979. Auerbach P: Primary cerebral venous thrombosis in young adults: The diverse manifestations of an unrecognized disease. Ann Neurol 3:81-86, 1978. Berkson J: Application of the logistic function to bio-assay. J Am Statist Assoc 39:357-365, 1944. Buonanno FS, Cooper MR, Moody DM, Laster DW, Ball MR, Toole JF: Neuroradiologic aspects of cerebral disseminated intravascular coagulation. AJNR 1:245-250, 1980. Coccheri S, Tests C: Regional intravascular coagulation and microthrombosis in traumatic brain lacerations, in Agnoti A, Fazio C (eds): Platelet Aggregation in the Pathogenesis of Cerebrovascular Disorders. Berlin, SpringerVerlag, 1977, pp 121-128. Collins RC, Al-Mondhiry H, Chernik N, Posner JB: Neurologic manifestations of intravascular coagulation in patients with cancer. Neurology 25:795-806, 1975. Cooper PR, Maravillak, Moody S, Clark WK: Serial computerized tomographic scanning and the prognosis of severe head injury. Neurosurgery 5:566-569, 1979. Crone KC, Lee KS, Kelly DL Jr: Correlation of admission fibrin degradation products with outcome and respiratory failure in patients with severe head injury. Neurosurgery 21:523536, 1987. Davies-Jones GAB, Preston FE, Timperley WR: Neurological Complications in Clinical Haemotology. Oxford, Blackwell, 1980, pp 193-227. Diaz FG, Yoch DH Jr, Larson D, Rockswold GL: Early diagnosis of delayed posttraumatic intracerebral hematomas. J Neurosurg 50:217223, 1979. Druskin MS, Drijansky R: Afibrinogenemia with severe head trauma. JAMA 219:755-756, 1972. French B, Dablin A: The value of computerized tomography in the management of 1000 consecutive head injuries. Surg Neurol 7:171-183, 1977. Goodnight SH, Kenoyer G, Rappaport SI:
Redistribution of this article permitted only in accordance with the publisher’s copyright provisions.
coagulopathy after head trauma to poor neurological and medical outcome (2,9,22,29,30,41,45). Our own mortality data (41% of patients with delayed injury, 0% of those without) support this contention. Kaufman et al. (20) suggested a causative role for DIC in the genesis of delayed ischemic and hemorrhagic lesions and, indeed, found coagulopathy to be present in 11 of 12 delayed traumatic intracerebral hematomas (18). Other investigations on the role of coagulopathy in delayed hemorrhage after head injury have yielded inconclusive data (15,37,42,43,47). One comparison of coagulation status and CT scans found a strong predictive role for coagulopathy (45), whereas another failed to confirm an association between coagulopathy and delayed hemorrhage (1). We believe that cerebral trauma may initiate a coagulation disorder, which in turn causes both the hemorrhagic and ischemic lesions of delayed brain injury. This relationship would provide a mechanism linking coagulopathy with poor outcome in some head injuries. We have by no means proven a cause and effect relationship between coagulopathy and delayed lesions. Indeed, their relationship may be adventitious or linked to some common cause, such as the severity of the primary brain injury. Table 7 shows generally lower Glasgow coma scale scores at admission among patients who develop delayed lesions, although there is considerable overlap. These limitations do not detract from the prognostic value of changes in clotting parameters at admission. Studies to further characterize the patient population with delayed insults are underway. The frequency with which the platelet count was either elevated or low prevented clear correlation with delayed injury. Thirty-six patients had increased platelet counts at admission. Such thrombocytosis is a common concomitant of trauma (17). Elevated platelets under these conditions do not appear to be related to an ongoing disease process (39). Indeed, we have not found the thrombocytosis to be of prognostic significance in these patients, whereas thrombocytopenia was associated with delayed lesions. Patients with multisystem organ failure were not included in the study because general outcome was poor and it did not necessarily correlate with neurological deterioration. We have begun a prospective study employing more sophisticated tests of coagulation to confirm the mechanism of coagulopathy involved and to improve its power in predicting the occurrence of delayed brain injury. We believe it is important to report that abnormalities, in even routine clotting studies, should alert the clinician caring for head-injured patients to prepare for radiological and clinical deterioration.
16. 17. 18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
Downloaded from https://academic.oup.com/neurosurgery/article-abstract/30/2/160/2752927 by University of New England user on 05 March 2018
28.
29.
30.
31. 32.
33.
34.
35.
36.
37.
38.
39.
40. 41.
Noorma V, Tikk A: Cerebral intravascular coagulation and increased platelet aggregation as mechanisms of brain ischemia in patients with acute cerebrovascular disease and brain injury. Acta Neurol Scand 60 [Suppl 72]:630631, 1979. Olson JD, Kaufman HH, Moake J, O'Gorman TW, Hoots K, Wagner K, Brown CK, Gildenberg PL: The incidence and significance of hemostatic abnormalities in patients with head injuries. Neurosurgery 24:825-832, 1989. Pondaag W: Disseminated intravascular coagulation related to outcome in head injury. Acta Neurochir Suppl (Wien) 28:98-102, 1979. Pozzati E, Guiliani G, Gaist G, Piazza G, Vergoni G: Chronic expanding intracerebral hematoma. J Neurosurg 65:611-614, 1986. Preston FE, Malia RG, Sworn MJ, Timperley WR, Blackburn ER: Disseminated intravascular coagulation as a consequence of cerebral damage. J Neurol Neurosurg Psychiatry 37:241-248, 1974. Ram Z, Hadani M, Spiegelman R, Tadmor R, Shacked I: Delayed nonhemorrhagic encephalopathy following mild head trauma. J Neurosurg 71:608-610, 1989. Roberson FC, Kishore PRS, Miller JD, Lipper MH, Becker DP: The value of serial computerized tomography in the management of severe head injury. Surg Neurol 12:161167, 1979. Ross SE, Civil I, O'Malley KF: Early serial computed tomography of the brain in severe closed head injury. Presented at the 48th annual meeting of the American Association for the Surgery of Trauma. September 9-12, 1987, Montreal. Sawada Y, Sadamitsu D, Sakamoto T, Ikemura Y, Yoshioka T, Sugimoto T: Lack of correlation between delayed traumatic intracerebral haematoma and disseminated intravascular coagulation. J Neurol Neurosurg Psychiatry 47:1125-1127, 1984. Smokovitis A, Astrup T: Localization of fibrinolytic activity and inhibition of plasmin in the spinal cord of rat, guinea pig and rabbit. J Neurosurg 48:1008-1112, 1978. Statham PF, Johnston RA, MacPherson P: Delayed deterioration in patients with traumatic frontal contusions. J Neurol Neurosurg Psychiatry 52:351-354, 1989. Stewart M, Kelton J: The platelet: Quantitative and qualitative abnormalities, in Nathan D, Oski F (ed): Hematology of Infancy and Childhood . Philadelphia, WB Saunders, 1987, ed 3, pp 1398-1399. Teasdale GF, Jennett B: Assessment of coma and impaired consciousness: A practical scale. Lancet 2:81-84, 1974. Tikk A, Noormaa V: The significance of cerebral and systemic disseminated intravascular coagulation in early prognosis of
Redistribution of this article permitted only in accordance with the publisher’s copyright provisions.
15.
Defibrination after brain tissue destruction: A serious complication of head injury. N Engl J Med 290:1043-1047, 1974. Gudeman SK, Kishore PRS, Miller JD, Girvendulis AK, Lipper MH, Becker DP: The genesis and significance of delayed traumatic intracerebral hematoma. Neurosurgery 5:309313, 1979. Hart R, Hindman B: Mechanisms of perioperative infarction. Stroke 13:766-773, 1982. Heath HW, Pearson HA: Thrombocytosis in pediatric outpatients. J Pediatr 114:805-807, 1989. Kaufman HH, Moake JL, Olson JD, Miner ME, duCret RP, Pruessner JL, Gildenberg PL: Delayed and recurrent intracranial hematomas related to disseminated intravascular clotting and fibrinolysis in head injury. Neurosurgery 7:445-449, 1980. Kaufman HH, Olson JD, Makela ME, Pruessner JL, Moake JL, Miner ME, Haar FL, Gildenberg PL: Disseminated intravascular coagulation and fibrinolysis in head injury. Presented at the 50th Annual Meeting of the American Association of Neurological Surgeons, Boston. April 5-8, 1981. Kaufman HH, Hui KS, Mattson JC, Borit A, Childs TL, Hoots WK, Bernstein DP, Makela ME, Wagner KA, Kahan BD, Gildenberg PL: Clinicopathological correlations of disseminated intravascular coagulation in patients with head injury. Neurosurgery 15:3442, 1984. Kobayashi S, Nakazawa S, Otsuka T: Clinical value of serial computed tomography with severe head injury. Surg Neurol 20:25-29, 1983. Kumura E, Sato M, Fukuda A, Takemoto Y, Tanaka SH, Kohama A: Coagulation disorders following acute head injury. Acta Neurochir (Wien) 85:23-28, 1987. Lipper MH, Kishore PRS, Girevendulis AK, Miller JD, Becker DP: Delayed intracranial hematoma in patients with severe head injury. Radiology 133:645-649, 1979. Lobato RD, Sarabia R, Rivas JJ, Cordobes F, Castro S, Muñoz MJ, Cabrera A, Barcena A, Lamas E: Normal computerized tomography scans in severe head injury. J Neurosurg 65:784-789, 1986. Lobato RD, Sarabia R, Cordobes F, Rivas JJ, Adrados A, Cabrera A, Gomez P, Madera A, Lamas E: Posttraumatic cerebral hemisphere swelling. J Neurosurg 68:417-428, 1988. Miner ME, Kaufman HH, Graham SH, Haar FH, Gildenberg PL: Disseminated intravascular coagulation and fibrinolytic syndrome following head injury in children: Frequency and prognostic implications. J Pediatr 100:687-691, 1982. Muller-Berghaus G: Pathophysiology of generalized intravascular coagulation. Semin Thrombo Hemost 34:209-214, 1977.
43.
44.
45.
46.
47.
Downloaded from https://academic.oup.com/neurosurgery/article-abstract/30/2/160/2752927 by University of New England user on 05 March 2018
Redistribution of this article permitted only in accordance with the publisher’s copyright provisions.
42.
brain injury. Acta Neurochir Suppl (Wien) 28:96-97, 1979. Touho H, Hirakawa K, Hino A, Karasawa J, Ohno Y: Relationship between abnormalities of coagulation and fibrinolysis and postoperatived intracranial hemorrhage in head injury. Neurosurgery 19:523-531, 1986. Ueda S, Fujitsu K, Fujino H, Sekino T, Kuwabara T: Correlation between plasma fibirn-fibrinogen degradation product values and CT findings in head injury. J Neurol Neurosurg Psychiatry 48:58-60, 1985. van der Sande JJ, Veltcamp JJ, BoekhoutMussert RJ, Bouwhuis-Hoogerwerf ML: Head injury and coagulation disorders. J Neurosurg 49:357-365, 1978. van der Sande JJ, Veltcamp JJ, BoekhoutMussert RJ, Vielvoye GJ: Hemostasis and computed tomography in head injury. J Neurosurg 55:718-724, 1981. Yanagisarva M, Kurihara H, Kimufa S, Tomobe Y, Kobayashi M, Mitsui Y, Yazabi Y, Goto K, Masaki T: A novel potent vasoconstrictor peptide produced by vascular endothelial cells. Nature 332:411-415, 1988. Young HA, Gleave JRW, Schmidek HH, Gregory S: Delayed traumatic intracerebral hematoma: Report of 15 cases. Neurosurgery 14:22-25, 1984.
Downloaded from https://academic.oup.com/neurosurgery/article-abstract/30/2/160/2752927 by University of New England user on 05 March 2018
Redistribution of this article permitted only in accordance with the publisher’s copyright provisions.
Figure 1. This 19-year-old man was involved in an automobile accident and arrived with a GCS of 7 and normal pupils. A, cranial CT scan at admission shows moderate contusions in both hemispheres. Admission PT was 16 seconds, PTT, 40 seconds; and platelets, 160,000. Forty-eight hours later, he suddenly lost pupillary and motor responses, and ICP rose from a baseline of 12 mmHg to 55. B, a second CT scan shows how the contusions worsened.
Figure 3. Plot of a logistic regression model of the probability of delayed brain injury as a function of PTT. Association is significant (P < 0.01).
Downloaded from https://academic.oup.com/neurosurgery/article-abstract/30/2/160/2752927 by University of New England user on 05 March 2018
Redistribution of this article permitted only in accordance with the publisher’s copyright provisions.
Figure 2. Plot of a logistic regression model of the probability of delayed brain injury as a function of PT. Association is highly significant (P < 0.001).
Table 1. Initial Cranial Computed Tomographic Scan Findings in 253 Patients Who Underwent Serial Studiesa
Downloaded from https://academic.oup.com/neurosurgery/article-abstract/30/2/160/2752927 by University of New England user on 05 March 2018
Redistribution of this article permitted only in accordance with the publisher’s copyright provisions.
Figure 4. Plot of a logistic regression model of the probability of delayed brain injury as a function of the coagulopathy index. The coagulopathy index is defined in Table 5 and is an indicator of coagulopathy severity. Association is highly significant (P < 0.001).
Table 3. Coagulopathy in Head Injury
Table 4. Coagulation Values in Head Injury
Downloaded from https://academic.oup.com/neurosurgery/article-abstract/30/2/160/2752927 by University of New England user on 05 March 2018
Redistribution of this article permitted only in accordance with the publisher’s copyright provisions.
Table 2. Lesions Seen in Subsequent Computed Tomographic Scans in 123 Patients with Delayed Brain Injurya
Table 6. Coagulation Severity Indexa
Table 7. Admission Glasgow Coma Scale with and without Delayed Brain Injury
Downloaded from https://academic.oup.com/neurosurgery/article-abstract/30/2/160/2752927 by University of New England user on 05 March 2018
Redistribution of this article permitted only in accordance with the publisher’s copyright provisions.
Table 5. Logistic Regression Statistics
Downloaded from https://academic.oup.com/neurosurgery/article-abstract/30/2/160/2752927 by University of New England user on 05 March 2018
and insults (35). Figure 1 illustrates a typical case of delayed injury. Delayed brain injury may take several forms: brain swelling (25), delayed extracerebral hematomas (23,24), ischemic (11,31,33), and hemorrhagic (11,38) lesions, as well as enlargement of previously small hematomas and contusions (31). Most of the serial CT scan studies have linked progressive lesions with poor outcome, but the significance of such lesions is not known. Although some authorities have considered delayed injury to be an important factor in neurological deterioration after head injury (8,21), others suggest it is the result, rather than a cause, of severe brain tissue destruction (36). A similarly large proportion of patients with serious head injury have been shown to have abnormal coagulation studies (19,26,30). Disseminated intravascular coagulation (DIC) has likewise been linked to poor outcome (2,9,22,29,42,45). The connection between intravascular coagulopathy and delayed lesions has been suggested (20,45), but it remains unproven. If such a connection exists, abnormal coagulation values at admission would be important predictors of later clinical and radiographic deterioration. It is the purpose of this study to determine whether such a link exists, in an effort to better understand the mechanism of delayed injury. PATIENTS AND METHODS The study population is drawn from patients with closed-head injury admitted to the South Jersey Regional Trauma Center at Cooper Hospital/University Medical Center during the years 1986 to 1988. During that 3-year period, 1018 patients were diagnosed as having moderate or severe closed-head injuries because of prolonged (more than 30 min) unconsciousness, an admission Glasgow coma scale (40) of less than 13, or focal neurological deficit. All patients had admission cranial CT scans and the following coagulation studies: prothrombin time (PT), activated partial thromboplastin time (PTT), and platelet count. Forty-eight patients were deemed brain dead or untreatable before a second CT scan could be performed. Another 20 were excluded from further study because of prolonged hypoxia, hypotension, or serious antecedent illness. Because of an admission GCS of 7 or more and a clinical recovery that was rapid and complete, 697 patients did not require a follow-up cranial CT scan. The remaining 253 patients had a second CT scan performed within 72 hours of admission. It is these patients whose records were retrospectively reviewed in our study. All CT scans were reported by members of the radiology staff, without reference to clinical or laboratory parameters. In each patient, CT scan reports were compared for signs of interval worsening, indicating possible delayed injury. Charts were reviewed to determine clotting parameters at admission. We recognize that the three coagulation studies used are not likely to vary independently. To allow for this collinearity and to assess whether the severity of coagulopathy influences the risk of developing
Redistribution of this article permitted only in accordance with the publisher’s copyright provisions.
Neurosurgery 1992-98 February 1992, Volume 30, Number 2 160 Delayed Brain Injury after Head Trauma: Significance of Coagulopathy Experimental and Clinical Study
based on the degree of abnormality in the three clotting studies (Table 6); the correlation between the coagulopathy index and the probability of developing delayed injury is illustrated in Figure 4. For this model, the regression coefficient is -1.149, χ2 is 39.2, significant at P < 0.001.
RESULTS Two hundred fifty-three patients, ranging in age from 3 to 74 years, qualified for the study. Of these, 123 (48.6%) developed delayed injury during their hospitalization. Thirty-six patients developed delayed insults after an initially normal CT study; these patients represent 34.3% of those whose first CT scan was normal. Among the 92 patients whose initial cranial CT scan showed abnormalities, 32 had new lesions on follow-up studies, and 56 cases experienced worsening of abnormalities that were evident at admission. CT findings at admission are shown in Table 1. Table 2 details progressive or new lesions documented in the group with delayed injury. At least one admission coagulation test was abnormal in 55% of the patients with delayed injuries, as opposed to 9% of those not demonstrating this finding (Table 3). A similar low incidence of coagulopathy (11%) was seen in the 697 patients not requiring a second CT scan. The association between abnormal clotting studies at admission and later development of delayed injury is highly significant (χ2 = 39.2, df = 1, P < 0.001). If a second set of coagulation studies was obtained within 12 hours of admission, the incidence of abnormalities increase to 74% of those who developed delayed cerebral insults. Table 3 also shows a breakdown of abnormalities in individual clotting tests. It is evident that PT is most frequently affected in delayed injury. However, the difference is statistically significant for all three coagulation studies. Mean values of the clotting tests in patients with and without delayed insults are given in Table 4. There are highly significant associations between both prolonged PT and PTT and the occurrence of delayed lesions (P < 0.001). The association between platelet count and delayed injury does not reach statistical significance. The logistic regression model shows a very strong correlation between PT and the probability of developing delayed insults (Fig. 2). For this model, the regression coefficient is -0.4392, χ2 is 32.3, significant at P < 0.001. PTTs show a similar, if less striking, relationship to the probability of delayed insults (Fig. 3). This model has a regression coefficient of -0.0195, χ2 of 6.6, and is significant at P < 0.01. The value of clotting studies at admission in predicting the genesis of delayed injury is as follows: Less than one-third (31.1%) of the patients with normal PT, PTT, and platelet counts develop delayed lesions. If at least one of these studies discloses abnormality, the incidence of delayed injury approaches 85% (84.8%). This difference is highly significant (P < 0.001). In the regression model, the relationship between single or multiple clotting abnormalities and the risk of delayed insults is shown in Table 5. A coagulopathy index was developed,
DISCUSSION What causes deterioration or death in the headinjured patient? Some patients have primary brain lesions that are so overwhelming they prevent any hope of recovery. These patients were excluded from our study. In some patients, it may be an unrecognized intracranial lesion that causes delayed complications. None of these patients is included in the present study because emergency CT scans were obtained at admission in every case. Another cause excluded by the entry criteria of our study is secondary brain injury resulting from overwhelming systemic damage or an antecedent disease. In the remaining patients, deterioration and poor outcome is most likely the result of secondary brain injury occurring because of intracranial insults. It is this patient group that usually requires serial CT scanning on clinical grounds and that constitutes our study group. It must be conceded that our patients do not represent all cases of head injury studied prospectively. The decision to obtain a second CT scan was made on clinical grounds, usually because the patients did not recover rapidly or completely. Our data show a highly significant association between the signs of a coagulation disorder at admission and a later development of cerebral injury in head-injured patients. Because we have not routinely tested for platelet function, fibrinogen consumption, or secondary fibrinolysis, we cannot invoke the mechanism of DIC with assurance. Nevertheless, such evidence has been obtained by others and is seen in 60% of head-injured patients (14, 19,26,30) . Several investigators have supported the concept of local intravascular coagulation and fibrinolysis in the injured brain (6,28,41). It is well documented that DIC can affect the brain (10). Both hemorrhage (7,18) and ischemia as a result of vascular occlusion (3,5,16) have been reported. A mechanism of microvascular occlusions with secondary lysis and hemorrhage into damaged cerebral tissue has been hypothesized for the traumatized brain (12,32,44). The mechanism of diffuse brain swelling seen in 28 of our patients is not obvious. Intravascular coagulopathy may disrupt the blood-brain barrier and damage vascular walls. Reperfusion of blood vessels so damaged may cause edema, perhaps in combination with hyperemic brain swelling. A detailed discussion of the causes of traumatic brain swelling is contained in the study by Lobato and coworkers (25). Mechanisms other than DIC may account for coagulopathy in head injury. Hypothalamic damage may alter levels of clotting factors (37). Clotting mechanisms may also be influenced by vasoactive peptides, such as endothelin (46) and Von Willebrand multimers (27), released in cerebral trauma. A number of studies have linked the appearance of
Downloaded from https://academic.oup.com/neurosurgery/article-abstract/30/2/160/2752927 by University of New England user on 05 March 2018
Redistribution of this article permitted only in accordance with the publisher’s copyright provisions.
delayed injury, logistic regression models were used (4) . This method involves a strategy of allocating patients into groups manifesting different degrees of coagulopathy, then regressing this binary group variable onto the odds of delayed injury being present.
California, October 20-25, 1990. Received for publication, May 10, 1990; accepted, final form, September 4, 1991. Reprint requests: Sherman C. Stein, M.D., Cooper Hospital/UMC, Three Cooper Plaza, Suite 411, Camden, NJ 08103. REFERENCES: (1-47) 1.
2.
3.
4. 5.
6.
7.
8.
9.
10.
11.
12. ACKNOWLEDGMENT The authors wish to thank the Institute of Brain Injury Research and Training at Mediplex RehabCamden for its support of our research activities. Presented in part at the 40th Annual Meeting of the Congress of Neurological Surgeons, Los Angeles,
Downloaded from https://academic.oup.com/neurosurgery/article-abstract/30/2/160/2752927 by University of New England user on 05 March 2018
13.
14.
Auer L: Disturbances of the coagulatory system in patients with severe cerebral trauma: I. Acta Neurochir (Wien) 43:51-59, 1978. Auer LM, Oh E: Disturbances of the coagulatory system in patients with severe cerebral trauma: II. Platelet Function. Acta Neurochir (Wien) 49:219-226, 1979. Auerbach P: Primary cerebral venous thrombosis in young adults: The diverse manifestations of an unrecognized disease. Ann Neurol 3:81-86, 1978. Berkson J: Application of the logistic function to bio-assay. J Am Statist Assoc 39:357-365, 1944. Buonanno FS, Cooper MR, Moody DM, Laster DW, Ball MR, Toole JF: Neuroradiologic aspects of cerebral disseminated intravascular coagulation. AJNR 1:245-250, 1980. Coccheri S, Tests C: Regional intravascular coagulation and microthrombosis in traumatic brain lacerations, in Agnoti A, Fazio C (eds): Platelet Aggregation in the Pathogenesis of Cerebrovascular Disorders. Berlin, SpringerVerlag, 1977, pp 121-128. Collins RC, Al-Mondhiry H, Chernik N, Posner JB: Neurologic manifestations of intravascular coagulation in patients with cancer. Neurology 25:795-806, 1975. Cooper PR, Maravillak, Moody S, Clark WK: Serial computerized tomographic scanning and the prognosis of severe head injury. Neurosurgery 5:566-569, 1979. Crone KC, Lee KS, Kelly DL Jr: Correlation of admission fibrin degradation products with outcome and respiratory failure in patients with severe head injury. Neurosurgery 21:523536, 1987. Davies-Jones GAB, Preston FE, Timperley WR: Neurological Complications in Clinical Haemotology. Oxford, Blackwell, 1980, pp 193-227. Diaz FG, Yoch DH Jr, Larson D, Rockswold GL: Early diagnosis of delayed posttraumatic intracerebral hematomas. J Neurosurg 50:217223, 1979. Druskin MS, Drijansky R: Afibrinogenemia with severe head trauma. JAMA 219:755-756, 1972. French B, Dablin A: The value of computerized tomography in the management of 1000 consecutive head injuries. Surg Neurol 7:171-183, 1977. Goodnight SH, Kenoyer G, Rappaport SI:
Redistribution of this article permitted only in accordance with the publisher’s copyright provisions.
coagulopathy after head trauma to poor neurological and medical outcome (2,9,22,29,30,41,45). Our own mortality data (41% of patients with delayed injury, 0% of those without) support this contention. Kaufman et al. (20) suggested a causative role for DIC in the genesis of delayed ischemic and hemorrhagic lesions and, indeed, found coagulopathy to be present in 11 of 12 delayed traumatic intracerebral hematomas (18). Other investigations on the role of coagulopathy in delayed hemorrhage after head injury have yielded inconclusive data (15,37,42,43,47). One comparison of coagulation status and CT scans found a strong predictive role for coagulopathy (45), whereas another failed to confirm an association between coagulopathy and delayed hemorrhage (1). We believe that cerebral trauma may initiate a coagulation disorder, which in turn causes both the hemorrhagic and ischemic lesions of delayed brain injury. This relationship would provide a mechanism linking coagulopathy with poor outcome in some head injuries. We have by no means proven a cause and effect relationship between coagulopathy and delayed lesions. Indeed, their relationship may be adventitious or linked to some common cause, such as the severity of the primary brain injury. Table 7 shows generally lower Glasgow coma scale scores at admission among patients who develop delayed lesions, although there is considerable overlap. These limitations do not detract from the prognostic value of changes in clotting parameters at admission. Studies to further characterize the patient population with delayed insults are underway. The frequency with which the platelet count was either elevated or low prevented clear correlation with delayed injury. Thirty-six patients had increased platelet counts at admission. Such thrombocytosis is a common concomitant of trauma (17). Elevated platelets under these conditions do not appear to be related to an ongoing disease process (39). Indeed, we have not found the thrombocytosis to be of prognostic significance in these patients, whereas thrombocytopenia was associated with delayed lesions. Patients with multisystem organ failure were not included in the study because general outcome was poor and it did not necessarily correlate with neurological deterioration. We have begun a prospective study employing more sophisticated tests of coagulation to confirm the mechanism of coagulopathy involved and to improve its power in predicting the occurrence of delayed brain injury. We believe it is important to report that abnormalities, in even routine clotting studies, should alert the clinician caring for head-injured patients to prepare for radiological and clinical deterioration.
16. 17. 18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
Downloaded from https://academic.oup.com/neurosurgery/article-abstract/30/2/160/2752927 by University of New England user on 05 March 2018
28.
29.
30.
31. 32.
33.
34.
35.
36.
37.
38.
39.
40. 41.
Noorma V, Tikk A: Cerebral intravascular coagulation and increased platelet aggregation as mechanisms of brain ischemia in patients with acute cerebrovascular disease and brain injury. Acta Neurol Scand 60 [Suppl 72]:630631, 1979. Olson JD, Kaufman HH, Moake J, O'Gorman TW, Hoots K, Wagner K, Brown CK, Gildenberg PL: The incidence and significance of hemostatic abnormalities in patients with head injuries. Neurosurgery 24:825-832, 1989. Pondaag W: Disseminated intravascular coagulation related to outcome in head injury. Acta Neurochir Suppl (Wien) 28:98-102, 1979. Pozzati E, Guiliani G, Gaist G, Piazza G, Vergoni G: Chronic expanding intracerebral hematoma. J Neurosurg 65:611-614, 1986. Preston FE, Malia RG, Sworn MJ, Timperley WR, Blackburn ER: Disseminated intravascular coagulation as a consequence of cerebral damage. J Neurol Neurosurg Psychiatry 37:241-248, 1974. Ram Z, Hadani M, Spiegelman R, Tadmor R, Shacked I: Delayed nonhemorrhagic encephalopathy following mild head trauma. J Neurosurg 71:608-610, 1989. Roberson FC, Kishore PRS, Miller JD, Lipper MH, Becker DP: The value of serial computerized tomography in the management of severe head injury. Surg Neurol 12:161167, 1979. Ross SE, Civil I, O'Malley KF: Early serial computed tomography of the brain in severe closed head injury. Presented at the 48th annual meeting of the American Association for the Surgery of Trauma. September 9-12, 1987, Montreal. Sawada Y, Sadamitsu D, Sakamoto T, Ikemura Y, Yoshioka T, Sugimoto T: Lack of correlation between delayed traumatic intracerebral haematoma and disseminated intravascular coagulation. J Neurol Neurosurg Psychiatry 47:1125-1127, 1984. Smokovitis A, Astrup T: Localization of fibrinolytic activity and inhibition of plasmin in the spinal cord of rat, guinea pig and rabbit. J Neurosurg 48:1008-1112, 1978. Statham PF, Johnston RA, MacPherson P: Delayed deterioration in patients with traumatic frontal contusions. J Neurol Neurosurg Psychiatry 52:351-354, 1989. Stewart M, Kelton J: The platelet: Quantitative and qualitative abnormalities, in Nathan D, Oski F (ed): Hematology of Infancy and Childhood . Philadelphia, WB Saunders, 1987, ed 3, pp 1398-1399. Teasdale GF, Jennett B: Assessment of coma and impaired consciousness: A practical scale. Lancet 2:81-84, 1974. Tikk A, Noormaa V: The significance of cerebral and systemic disseminated intravascular coagulation in early prognosis of
Redistribution of this article permitted only in accordance with the publisher’s copyright provisions.
15.
Defibrination after brain tissue destruction: A serious complication of head injury. N Engl J Med 290:1043-1047, 1974. Gudeman SK, Kishore PRS, Miller JD, Girvendulis AK, Lipper MH, Becker DP: The genesis and significance of delayed traumatic intracerebral hematoma. Neurosurgery 5:309313, 1979. Hart R, Hindman B: Mechanisms of perioperative infarction. Stroke 13:766-773, 1982. Heath HW, Pearson HA: Thrombocytosis in pediatric outpatients. J Pediatr 114:805-807, 1989. Kaufman HH, Moake JL, Olson JD, Miner ME, duCret RP, Pruessner JL, Gildenberg PL: Delayed and recurrent intracranial hematomas related to disseminated intravascular clotting and fibrinolysis in head injury. Neurosurgery 7:445-449, 1980. Kaufman HH, Olson JD, Makela ME, Pruessner JL, Moake JL, Miner ME, Haar FL, Gildenberg PL: Disseminated intravascular coagulation and fibrinolysis in head injury. Presented at the 50th Annual Meeting of the American Association of Neurological Surgeons, Boston. April 5-8, 1981. Kaufman HH, Hui KS, Mattson JC, Borit A, Childs TL, Hoots WK, Bernstein DP, Makela ME, Wagner KA, Kahan BD, Gildenberg PL: Clinicopathological correlations of disseminated intravascular coagulation in patients with head injury. Neurosurgery 15:3442, 1984. Kobayashi S, Nakazawa S, Otsuka T: Clinical value of serial computed tomography with severe head injury. Surg Neurol 20:25-29, 1983. Kumura E, Sato M, Fukuda A, Takemoto Y, Tanaka SH, Kohama A: Coagulation disorders following acute head injury. Acta Neurochir (Wien) 85:23-28, 1987. Lipper MH, Kishore PRS, Girevendulis AK, Miller JD, Becker DP: Delayed intracranial hematoma in patients with severe head injury. Radiology 133:645-649, 1979. Lobato RD, Sarabia R, Rivas JJ, Cordobes F, Castro S, Muñoz MJ, Cabrera A, Barcena A, Lamas E: Normal computerized tomography scans in severe head injury. J Neurosurg 65:784-789, 1986. Lobato RD, Sarabia R, Cordobes F, Rivas JJ, Adrados A, Cabrera A, Gomez P, Madera A, Lamas E: Posttraumatic cerebral hemisphere swelling. J Neurosurg 68:417-428, 1988. Miner ME, Kaufman HH, Graham SH, Haar FH, Gildenberg PL: Disseminated intravascular coagulation and fibrinolytic syndrome following head injury in children: Frequency and prognostic implications. J Pediatr 100:687-691, 1982. Muller-Berghaus G: Pathophysiology of generalized intravascular coagulation. Semin Thrombo Hemost 34:209-214, 1977.
43.
44.
45.
46.
47.
Downloaded from https://academic.oup.com/neurosurgery/article-abstract/30/2/160/2752927 by University of New England user on 05 March 2018
Redistribution of this article permitted only in accordance with the publisher’s copyright provisions.
42.
brain injury. Acta Neurochir Suppl (Wien) 28:96-97, 1979. Touho H, Hirakawa K, Hino A, Karasawa J, Ohno Y: Relationship between abnormalities of coagulation and fibrinolysis and postoperatived intracranial hemorrhage in head injury. Neurosurgery 19:523-531, 1986. Ueda S, Fujitsu K, Fujino H, Sekino T, Kuwabara T: Correlation between plasma fibirn-fibrinogen degradation product values and CT findings in head injury. J Neurol Neurosurg Psychiatry 48:58-60, 1985. van der Sande JJ, Veltcamp JJ, BoekhoutMussert RJ, Bouwhuis-Hoogerwerf ML: Head injury and coagulation disorders. J Neurosurg 49:357-365, 1978. van der Sande JJ, Veltcamp JJ, BoekhoutMussert RJ, Vielvoye GJ: Hemostasis and computed tomography in head injury. J Neurosurg 55:718-724, 1981. Yanagisarva M, Kurihara H, Kimufa S, Tomobe Y, Kobayashi M, Mitsui Y, Yazabi Y, Goto K, Masaki T: A novel potent vasoconstrictor peptide produced by vascular endothelial cells. Nature 332:411-415, 1988. Young HA, Gleave JRW, Schmidek HH, Gregory S: Delayed traumatic intracerebral hematoma: Report of 15 cases. Neurosurgery 14:22-25, 1984.
Downloaded from https://academic.oup.com/neurosurgery/article-abstract/30/2/160/2752927 by University of New England user on 05 March 2018
Redistribution of this article permitted only in accordance with the publisher’s copyright provisions.
Figure 1. This 19-year-old man was involved in an automobile accident and arrived with a GCS of 7 and normal pupils. A, cranial CT scan at admission shows moderate contusions in both hemispheres. Admission PT was 16 seconds, PTT, 40 seconds; and platelets, 160,000. Forty-eight hours later, he suddenly lost pupillary and motor responses, and ICP rose from a baseline of 12 mmHg to 55. B, a second CT scan shows how the contusions worsened.
Figure 3. Plot of a logistic regression model of the probability of delayed brain injury as a function of PTT. Association is significant (P < 0.01).
Downloaded from https://academic.oup.com/neurosurgery/article-abstract/30/2/160/2752927 by University of New England user on 05 March 2018
Redistribution of this article permitted only in accordance with the publisher’s copyright provisions.
Figure 2. Plot of a logistic regression model of the probability of delayed brain injury as a function of PT. Association is highly significant (P < 0.001).
Table 1. Initial Cranial Computed Tomographic Scan Findings in 253 Patients Who Underwent Serial Studiesa
Downloaded from https://academic.oup.com/neurosurgery/article-abstract/30/2/160/2752927 by University of New England user on 05 March 2018
Redistribution of this article permitted only in accordance with the publisher’s copyright provisions.
Figure 4. Plot of a logistic regression model of the probability of delayed brain injury as a function of the coagulopathy index. The coagulopathy index is defined in Table 5 and is an indicator of coagulopathy severity. Association is highly significant (P < 0.001).
Table 3. Coagulopathy in Head Injury
Table 4. Coagulation Values in Head Injury
Downloaded from https://academic.oup.com/neurosurgery/article-abstract/30/2/160/2752927 by University of New England user on 05 March 2018
Redistribution of this article permitted only in accordance with the publisher’s copyright provisions.
Table 2. Lesions Seen in Subsequent Computed Tomographic Scans in 123 Patients with Delayed Brain Injurya
Table 6. Coagulation Severity Indexa
Table 7. Admission Glasgow Coma Scale with and without Delayed Brain Injury
Downloaded from https://academic.oup.com/neurosurgery/article-abstract/30/2/160/2752927 by University of New England user on 05 March 2018
Redistribution of this article permitted only in accordance with the publisher’s copyright provisions.
Table 5. Logistic Regression Statistics
