Dysautonomia After Severe Traumatic Brain Injury

European Journal of Neurology 2010, 17: 1172–1177 doi:10.1111/j.1468-1331.2010.02989.x

Dysautonomia after severe traumatic brain injury

H. T. Hendricksa, A. H. Heerenb and P. E. Vosc aDepartment of Rehabilitation Medicine, Radboud University Medical Centre, Nijmegen, Groot Klimmendaal, Rehabilitation Centre, Arnhem, the Netherlands; bDepartment of Rehabilitation Medicine, Radboud University Medical Centre, Nijmegen, the Netherlands; and cRadboud University Medical Centre, Institute of Neurology, Nijmegen, the Netherlands

Keywords: Background: Dysautonomia after traumatic brain injury (TBI) is characterized by cohort study, diffuse episodes of increased heart rate, respiratory rate, temperature, blood pressure, muscle axonal injury, tone, decorticate or decerebrate posturing, and profuse sweating. This study addresses dysautonomia, incidence, the incidence of dysautonomia after severe TBI, the clinical variables that are asso- traumatic brain injury ciated with dysautonomia, and the functional outcome of patients with dysautonomia. Methods: A historic cohort study in patients with severe TBI [Glasgow Coma Scale Received 3 November 2009 (GCS) £ 8 on admission]. Accepted 20 January 2010 Results: Seventy-six of 119 patients survived and were eligible for follow-up. The incidence of dysautonomia was 11.8%. Episodes of dysautonomia were prevalent during a mean period of 20.1 days (range 3–68) and were often initiated by discomfort. Patients with dysautonomia showed signiﬁcant longer periods of coma (24.78 vs. 7.99 days) and mechanical ventilation (22.67 vs. 7.21 days). Dysautonomia was associated with diﬀuse axonal injury (DAI) [relative risk (RR) 20.83, CI 4.92–83.33] and the development of spasticity (RR 16.94, CI 3.96–71.42). Patients with dysautonomia experienced more secondary complications. They tended to have poorer outcome. Conclusions: Dysautonomia occurs in approximately 10% of patients surviving severe TBI and is associated with DAI and the development of spasticity at follow-up. The initiation of dysautonomia by discomfort supports the Excitatory: Inhibitory Ratio model as pathophysiological mechanism.

bility, paroxysmal autonomic instability with dystonia, Introduction hyperpyrexia associated with muscle contraction, mid- Dysautonomia after traumatic brain injury (TBI) is a brain deregulatory syndrome, and brainstem attack. syndrome characterized by episodes of autonomic Hereafter, the syndrome will be referred to as dysau- dysregulation, including an increased heart rate, respi- tonomia. ratory rate, body temperature and blood pressure, Several theories have been postulated regarding the decerebrate or decorticate posturing, an increased pathogenesis of dysautonomia [1]. The initially pro- muscle tone, and profuse sweating [1–6]. It develops posed epileptogenic etiology has been abandoned generally during the early recovery phase, continuing recently, because patients with clinical dysautonomia for days, weeks, or even months [4,6]. The reported do not exhibit epileptiform activity on EEG [8,9], and incidence of dysautonomia after TBI varies from 8 to attempts to treat dysautonomia with anti-epileptics 33% [3–6]. There are no generally accepted therapeutic have failed [10]. A second hypothesis, the disconnection strategies [7]. Long duration of dysautonomia is asso- theory, describes diﬀuse [3,8–10] or focal [6] cerebral ciated with poorer outcome [3,6]. Various terminologies damage, aﬀecting the functional interaction between have been used to describe the syndrome, such as par- cortex and hypothalamus as a cause for the autonomic oxysmal sympathetic storm, acute hypothalamic insta- hyperactivity. Cortical and subcortical, as well as anterior hypothalamus, diencephalic (thalamus or hypothalamus) and upper brainstem lesions may con- Correspondence: Henk T. Hendricks, Department of Rehabilitation tribute to dysautonomia. In his latest review, Baguley Medicine, Radboud University Medical Centre, PO 9011, 6500 HB Nijmegen, the Netherlands (tel.: +0031 243614804; fax: +0031 et al. [11] proposed the Excitatory: Inhibitory Ratio 243619839; e-mail: [email protected]). model. This model considers dysautonomia as a result

2010 The Author(s) 1172 European Journal of Neurology 2010 EFNS Dysautonomia after traumatic brain injury 1173 of lesions to central inhibitory pathways that regulate rate, respiratory rate, body temperature, and blood afferent information, leading to exorbitant autonomic pressure were derived from ICU graphics and from the reactions to various stimuli. neurological and neurosurgical charts. When symptoms In clinical practice, dysautonomia is often misinter- of dysautonomia were present, laboratory data and preted. The reported incidence rates vary considerably, reports of medical microbiology were checked to the prognostic importance is still unclear. Furthermore, exclude systemic infection. In case of rising C-reactive the suggested pathophysiological models are insuffi- protein and/or positive blood cultures simultaneously ciently supported by clinical evidence. Therefore, this with the symptoms of dysautonomia, the episode was study addresses the following research questions attributed to systemic infection. Increased muscle tone, (i) What is the incidence of dysautonomia in patients decerebrate or decorticate posturing during paroxysms with TBI? (ii) Which clinical variables are associated with were noted from medical files and daily reports of the occurrence of dysautonomia? (iii). Is the functional physiotherapists. outcome of patients who suffered from dysautonomia As independent determinants, we studied demo- worse than comparable patients without dysautonomia? graphic data and brain injury-related variables, such as admission GCS score, initial CT scans, the occurrence of diffuse axonal injury (DAI), coma duration, and Methods duration of mechanical ventilation. The co-existence of We designed a historic cohort study, in which all con- other traumatic pathology, i.e., bone fractures and secutively admitted patients (Radboud University organ ruptures, as well as secondary complications, Medical Centre, Nijmegen) with severe TBI aged 16 or such as the occurrence of systemic infection (apart from older, during the period 2002–2003, were included. the dysautonomia paroxysms), and the development of Severe head injury was defined as a history of impact to heterogenic ossifications were recorded. Furthermore, the head and a Glasgow Coma Scale (GCS) score of £ 8 we studied the development of spasticity at follow-up, on admission at the hospital. All patients with TBI apart from the dysautonomia paroxysm(s). At presenting at the emergency department of our hospital 6 months, the Extended Glasgow Outcome Scores are included in the Radboud University Nijmegen (GOSE) [16] was used as outcome measure. The GOSE Brain Injury Cohort Study (RUBICS) [12]. In RU- is an eight-point scale expressing functional outcome BICS, various clinical variables are registered obtained ranging from 1 (death) to 8 (complete recovery). GOSE from the ambulance or helicopter trauma physician, the score 2 represents vegetative state; GOSE 3, 24 h emergency department, the intensive care unit (ICU), dependency (at home); GOSE 4, patient is dependent and the neurological and neurosurgical ward. but can do without help for at least eight consecutive Patients who died were excluded. Patients with hours; GOSE 5, independence in activities of daily liv- co-existent spinal cord injury were also excluded. The ing but no resumption of former employment; GOSE 6, study was approved by the research ethics board of our reduced capacity for (former) work; and GOSE 7, university hospital. Treatment for the patients at the resumption of former employment but persisting com- Emergency Department was according to the Acute plaints interfering with daily activities. Trauma Life Support protocol [13], and treatment at The initial CT scans were classified by one investi- the ICU was based on the international guidelines for gator (PV), according to the Traumatic Coma Data- the management of patients after severe TBI [14]. bank (TCDB) criteria [17], see Table 1. DAI was Dysautonomia was defined as a simultaneous defined as a GCS of 8 or less, lasting for more than 6 h, occurrence of the following features: (i) increased heart and a CT pattern with presence of small (<15 mm), rate (>120/min), (ii) increased respiratory rate (>24/ non-expansive punctate hemorrhages at the gray/white min), (iii) raised temperature (>38.5C), (iv) increased matter interface, corpus callosum or brainstem, and systolic blood pressure (>160 mmHg), (v) increased absence of large intraparenchymal or extracerebral muscle tone (modified Ashworth score), (vi) decerebrate lesions [18]. If available, magnetic resonance imaging or decorticate posturing, and (vii) profuse sweating. (MRI) scans were also examined. Five of these seven features had to be minimally pres- The duration of coma was counted in days from ent. Furthermore, these features had to occur for at admission until a GCS of 9 or higher was reached. Days least three consecutive days. Infections or other possible of mechanical ventilation were counted from onset to causes had to be ruled out [3,15]. Medical files of all the regaining of normal ventilation (weaning process survivors were systematically reviewed by one investi- completed). gator (AH) with regard to the occurrence of dysau- Systemic infection was defined as a temperature tonomia, the onset of symptoms since trauma, and the above 38C, increased C-reactive protein and a positive duration in days. Physiological data such as the heart blood culture, not suggestive of contamination. Patients

2010 The Author(s) European Journal of Neurology 2010 EFNS European Journal of Neurology 17, 1172–1177 1174 H. T. Hendricks et al.

Table 1 Trauma coma databank criteria Table 2 Patient characteristics

Category I No visible intracranial pathology is present Age (mean/range) 36.67 (16–84) years Category II Cisterns are present with midline shift 0–5 mm and/or lesion densities with a volume not Gender male/female (%) 47/29 (61.8/38.2) exceeding 25 ml Mode of injury (%) Category III Cisterns are compresses or absent, with a midline Traffic 62 (81.6) shift 0–5 mm, no high or mixed density Violence 2 (2.6) lesions ‡ 25 ml Falls 7 (9.2) Category IV As III, but with a midline shift >5 mm Others 5 (6.6) Category V The presence of any mass lesion that is surgically Admission GCS (%) evacuated 3 49 (64.5) Category VI The presence of any mass lesion ‡ 25 ml that is not 4 5 (6.6) surgically evaluated 5 13 (17.1) 6 7 (9.2) 7 1 (1.3) 8 1 (1.3) were observed for the occurrence of heterotopic ossifi- TCDB (%) Category 1 20 (26.3) cations, defined as ectopic bone formation in joint Category 2 34 (44.7) capsules, ligaments, tendons, or muscles, causing clini- Category 3 11 (14.5) cal signs like painful swelling of joints, redness, and Category 4 4 (5.3) deceased range of motion. Category 5 4 (5.3) Spasticity was defined as any modified Ashworth Category 6 3 (3.9) Coma duration (mean/range) 10.0 (1–61) days score of 2 or more and had to be present at follow-up Mechanical ventilation (mean/range) 8.9 (0–52) days apart from the dysautonomia paroxysm(s). The occurrence of spasticity was noted from medical files and GCS, Glasgow Coma Scale; TCDB, Trauma Coma Data Bank daily reports of physiotherapists. criteria.

(range 1–22 days). Some patients had episodes every Statistical analysis day till the syndrome ceased. In other patients, the Two groups were distinguished, patients with and episodes were clustered in periods of several days. The without dysautonomia. With regard to age, coma mean total duration of the episodes was 18.8 days, duration, and days of mechanical ventilation, we used range 3–68 days. In six patients, it was noted that the the Mann–Whitney test. All other determinants were episodes of dysautonomia were reactive, e.g., after considered as present or absent and expressed in mobilization, nursing, or physical examination. contingency tables. Ordinal determinants were divided All patients with dysautonomia had an admission into two groups. As for the Glasgow Coma Score, a GCS score of 3, versus 63% in the control group. GCS of 3 was compared with GCS 4-8. The TCDB- Furthermore, the mean coma duration in the dysau- criteria were subdivided into categories I and categories tonomia group was 24.78 ± 20.83 days (range 10–52) II–VI. The Extended Glasgow Coma Outcome scale vs. 7.99 ± 8.34 days (range 1–37) in patients without was dichotomized in good outcome (categories 7–8) dysautonomia (P < 0.001, z = )3.31). Patients with versus poor outcome (categories 2–6). For each of the dysautonomia were longer mechanically ventilated; binominal determinants, relative risks (if applicable) 22.67 days ± 13.82 days (range 1–52) vs. 7.2 ±7.6 days were calculated with their 95% confidence intervals, as (range 0–34) (z = )3.69, P < 0.001). well as the corresponding positive and negative The binomial variables are listed in Table 4. All predictive values. patients with dysautonomia had initial CT abnormali- ties according to the TCDB classification, a relative risk (RR) could not be defined. The occurrence of dysau- Results tonomia showed a significant association with DAI During the study period, 119 patients were included, of (RR 20.83, CI 4.92–83.33). In three patients with dys- which 76 patients survived. Patient characteristics are autonomia, examination of available MRI scans listed in Table 2. There was no significant difference in showed punctate hemorrhages compatible with DAI in age and gender between the groups. The mean follow- (bi) frontal, paraventricular regions, in the corpus up period was 17 months, range 6–25 months. Of the callosum, and in the basal ganglia. All patients with 76 survivors, nine patients (11.8%) developed dysau- dysautonomia had co-existent bone fractures, versus 52 tonomia (Table 3). The first episode of dysautonomia controls (85.2%), a RR could not be defined. The occurred in all cases in the first weeks after trauma co-existence of systemic infections (apart from the

2010 The Author(s) European Journal of Neurology 2010 EFNS European Journal of Neurology 17, 1172–1177 Dysautonomia after traumatic brain injury 1175

Table 3 Features of the dysautonomia paroxysms and co-existing morbidity and events

Duration of Onset of dysautonomia dysautonomia Duration of Duration of Stimulus for episode Patient (days after trauma) (days) coma (days) MV (days) Complications of dysautonomia

1 10 14 10 25 Spleen rupture, pleura empyema Mobilization 2 1 68 60 10 Spleen rupture, hematothorax, decubitis Unknown 3 4 3 13 15 Pneumothorax Unknown 4 2 7 10 13 Compartment syndrome left arm Unknown 5 7 4 13 12 Rupture urinary bladder, decubitis Nursing 6 1 25 18 23 ARDS Pain stimulus 7 22 27 38 61 Spleen rupture, rupture of the liver Nursing mesenterium hematoma, pneumothorax, deep venous thrombosis right arm 8 20 6 8 37 Pneumothorax, decubitis Pain stimulus 9 4 15 15 17 ARDS Nursing

MV, mechanical ventilation; ARDS, acute respiratory distress syndrome.

Table 4 Relative risks, positive and negative predictive values for the versus one patient in the control group (RR 58.82, CI studied binomial risk factors 8.40–50.00). Other phenomena like decubitis ulcers and deep venous thrombosis were also regularly seen in the Dysautonomia Dysautonomia present absent RR (CI) dysautonomia group (see Table 3). Spasticity was more often observed in patients with dysautonomia than in GCS RR not defined patients without dysautonomia, RR 16.94 (CI 3.96– 3 9 40 PPV: 18.36% 71.42). The GOSE data are shown in Table 4. Patients 4–8 0 27 NPV: 100% TDCB RR not defined with dysautonomia tended to have a poorer outcome as Categories II–VI 9 20 PPV: 31.03% measured by the GOSE, but there was no significant Category I 0 47 NPV: 100% difference. Systemic infection RR 13.15 (CI 3.01–58.82) Present 7 9 PPV: 43.75% Discussion Absent 2 58 NPV: 96.66% DAI RR 20.83 From this study, we show that dysautonomia occurs Present 7 4 (CI 4.92–83.33) quite regularly in surviving patients after severe TBI. Absent 2 63 PPV: 63.63% We found an incidence of 11.8%. Especially of interest NPV: 96.92% was the temporal relation of the episodes of dysauton- Fractures RR not defined Present 9 52 PPV: 14.75% omia and moments of distress. It developed generally in Absent 0 15 NPV: 100% the most severely injured patients. Patients with dys- HO RR 58.82 autonomia stayed significantly longer in coma and were Present 8 1 (CI 8.40–50.00) longer mechanically ventilated. There was a significant Absent 1 66 PPV: 88.88% association of dysautonomia with the occurrence of NPV: 98.50% GOS-E RR 7.19 DAI. All patients with dysautonomia had co-existent GOSE 2–6 8 31 (CI 0.94–55.55) bone fractures. Patients with dysautonomia had suf- GOSE 7–8 1 34 PPV: 20.51% fered significantly more systemic infections, and they NPV: 97.14% developed more often heterotopic ossifications. They Spasticity RR 16.95 also developed more often spasticity at follow-up. Present 7 6 (CI 3.97–71.43) Absent 2 61 PPV: 53.84% Patients with dysautonomia tended to have a poorer NPV: 96.82% outcome as measured by the GOSE, but this was not statistically significant. GCS, Glasgow Coma Scale; TCDB, Trauma Coma Data Bank The incidence of dysautonomia of 11.8% is in criteria; DAI, diffuse axonal injury; HO, heterotopic ossification; GOS-E, Glasgow Outcome Scale extended; RR, relative risk; PPV, accordance with the results of comparable studies. positive predictive value; NPV, negative predictive value. Baguley et al. [2] reported an incidence of 8%, and Fernandez-Ortega et al. [6] an incidence of 9.3%. dysautonomia paroxysms) was particularly high (RR However, in other studies, the incidences were much 13.15, CI 3.01–58.82). Another remarkable finding was higher. Dolce et al. [4] found an incidence of 31.9% in the co-existence of heterotopic ossifications. Hetero- post-traumatic patients in a special ICU dedicated for topic ossifications were detected in eight of nine patients patients in a vegetative state. Rabinstein [5] reported an

2010 The Author(s) European Journal of Neurology 2010 EFNS European Journal of Neurology 17, 1172–1177 1176 H. T. Hendricks et al. incidence of 33%, but defined paroxysmal sympathetic the Excitatory: Inhibitory Ratio model. Unfortunately, hyperactivity differently. These authors did not utilize a because in our study insufficient imaging data were time restraint. Therefore, the broad variation in inci- generated, it is not possible to support more precisely dence of dysautonomia originates mainly from different one of the proposed theories. inclusion criteria and definitions of dysautonomia. It In our study, patients with dysautonomia tended to should be emphasized that brain-injured patients often have a poorer outcome as measured by the GOSE. This exhibit elevation of some autonomic parameters in the is in line with previous studies [2,3,6], and the notion (sub) acute phase. Baguley et al. [2] found signs of that dysautonomia develops in the most severe TBI autonomic arousal in 92% of the patients with mod- patients. erate and severe TBI during the first week of admission Despite increasing research, there is still limited at the ICU. Some of these patients with autonomic understanding of the pathophysiology of dysauto- arousal developed ultimately the distinct syndrome of nomia. Baguley et al. [11] introduced recently the dysautonomia. Hence, signs of autonomic arousal are Excitatory: Inhibitory Ratio model, considering dys- very frequently encountered in patients with severe TBI, autonomia as a result of lesions to central inhibitory whereas the incidence of the complete syndrome of structures which regulate afferent information, causing dysautonomia is much lower. Dysautonomia occurs in exorbitant autonomic reactions to painful or even approximately 10% of patients surviving severe TBI, normal stimuli. with regard to this article and previous literature. The lesions may be temporary because of local (yet In our study, dysautonomia developed generally in unknown) toxic effects [4], resolving after some time, or the most severely injured within a cohort of patients more structural, accounting for long-lasting disinhibi- with severe TBI. This is in line with findings in previous tion. In our study, episodes of dysautonomia were often patient–control studies [2,6] and a more recently pub- reactive, e.g., after mobilization, nursing, or physical lished prospective study [2], in which the authors examination. Moreover, all patients with dysautonomia showed that both patients with initial autonomic had co-existent fractures, and eight patients developed arousal and subsequent dysautonomia suffered the painful heterotopic ossifications in the first month after most severe brain injuries as measured by the GCS and trauma. Not all episodes had an evident preceding the Disability Rating Scale at week 1, and the duration stimulus; however, we assume that pain was present in of the post-traumatic amnesia. In our study, all patients all our patients with dysautonomia. Hence, these with dysautonomia had an admission GCS score of 3, observations support the Excitatory: Inhibitory Ratio and they stayed significantly longer in coma and were model. longer mechanically ventilated. Moreover, they also This study contains several limitations. First of all, presented more comorbidity. All these findings are in the retrospective character of the data collection. The accordance with the notion that dysautonomia occurs physiological data were not stored in RUBICS. Yet, we in patients with the most severe TBI. These patients are could retain them easily from the ICU files. If there also prone to develop spasticity at follow-up, as we were episodes of dysautonomia co-existent with sys- found in our study. temic infection, then the autonomic symptoms were There was also a significant association of dysau- retrospectively attributed to the infection, possibly tonomia with DAI in our study. This is in accordance false. We have defined dysautonomia according to with the observations of Baguley et al. [3] in their clinical criteria [3,15]. More recently, efforts have been patient–control study. Fernandez et al. [6] reported executed toward a revision of the diagnostic criteria more focal lesions on CT in the dysautonomia group according to a clinical algorithm [19]. These issues versus the controls. On the other hand, the recent might have led to a biased incidence of dysautonomia. prospective study of Baguley et al. [2] did not reveal On the other hand, our findings were according to significant differences regarding the initial CT data existing literature. It should be underscored that we between the patients with autonomic arousal and the have diagnosed DAI based on clinical and CT criteria, controls. However, the authors did not specifically which is inferior to a MRI diagnosis [20]. However, in address DAI in the dysautonomia versus the non- the three patients with MRI data and dysautonomia, dysautonomia group. The association of DAI with the the diagnosis of DAI was confirmed. occurrence of dysautonomia is of particular interest In conclusion, dysautonomia occurs in approxi- with respect to the pathogenesis of dysautonomia. The mately 10% of patients surviving severe TBI and is prevalence of DAI or focal lesions in regulatory auto- associated with DAI and the development of spasticity. nomic centers in the cortex and subcortex, the dien- Patients with severe TBI who developed dysautonomia cephalon (thalamus or hypothalamus) and upper tended to have a poorer outcome, but there was no brainstem supports both the disconnection theory and significant difference. Particular of interest in our study

2010 The Author(s) European Journal of Neurology 2010 EFNS European Journal of Neurology 17, 1172–1177 Dysautonomia after traumatic brain injury 1177 was the temporal relation of dysautonomia paroxysms after severe diffuse axonal head injury. Mayo Clin Proc and moments of distress. This observation supports the 1998; 73: 148–152. Excitatory: Inhibitory Ratio model. Further prospec- 10. Bullard DE. Diencephalic seizures: responsiveness to bromocriptine and morphine. Ann Neurol 1987; 21: 609– tive research should confirm our findings. 611. 11. Baguley IJ. The excitatory: inhibitory ratio model (EIR model): an integrative explanation of acute autonomic References overactivity syndromes. Med Hypotheses 2008; 70: 26–35. 1. Baguley IJ, Heriseanu RE, Cameron ID, Nott MT, Slewa- 12. Vos PE, Battistin L, Birbamer G, et al. EFNS guideline Younan S. A critical review of the pathophysiology of on mild traumatic brain injury: report of an EFNS task dysautonomia following traumatic brain injury. Neurocrit force. Eur J Neurol 2002; 9: 207–219. Care 2008; 8: 293–300. 13. American College of Surgeons. Advanced trauma life 2. Baguley IJ, Slewa-Younan S, Heriseanu RE, Nott MT, support for doctors, 7th edn. Chicago, IL: American Mudaliar Y, Nayyar V. The incidence of dysautonomia College of Surgeons, 1997. and its relationship with autonomic arousal following 14. Bratton SL, Chestnut RM, Ghajar J, et al. International traumatic brain injury. Brain Inj 2007; 21: 1175–1181. guidelines for the management of severe traumatic brain 3. Baguley IJ, Nicholls JL, Felmingham KL, Crooks J, injury. J Neurotrauma 2007; 1: S1–S95. Gurka JA, Wade LD. Dysautonomia after traumatic 15. Blackman JA, Patrick PD, Buck ML, Rust RS Jr. Par- brain injury: a forgotten syndrome? J Neurol Neurosurg oxysmal autonomic instability with dystonia after brain Psychiatry 1999; 67: 39–43. injury. Arch Neurol 2004; 61: 321–328. 4. Dolce G, Quintieri M, Leto E, et al. Dysautonomia and 16. Teasdale GM, Pettigrew LE, Wilson JT, Murray G, Jen- Clinical Outcome in Vegetative State. J Neurotrauma nett B. Analyzing outcome of treatment of severe head 2008; 25: 1079–1082. injury: a review and update on advancing the use of the 5. Rabinstein AA. Paroxysmal sympathetic hyperactivity in Glasgow Outcome Scale. J Neurotrauma 1998; 15: 587– the neurological intensive care unit. Neurol Res 2007; 29: 597. 680–682. 17. Marshall LF, Marshall SB, Klauber MR, et al. A new 6. Fernańdez-Ortega JF, Prieto-Palomino MA, Mun˜ oz- classification of head injury based on computerized Lo´pez A, Lebron-Gallardo M, Cabrera-Ortiz H, Ques- tomography. J Neurotrauma 1992; 9: 287–292. ada-Garcıá G. Prognostic influence and computed 18. Gennarelli TA, Spielman GM, Langfitt TW, et al. Influ- tomography findings in dysautonomic crises after trau- ence of the type of intracranial lesion on outcome from matic brain injury. J Trauma 2006; 61: 1129–1133. severe head injury. J Neurosurg 1982; 56: 26–32. 7. Baguley IJ, Cameron ID, Green AM, Slewa-Younan S, 19. Baguley IJ, Nott MT, Slewa-Younan S, Heriseanu RE, Marosszeky JE, Gurka JA. Pharmacological management Perkes IE. Diagnosing dysautonomia after acute trau- of dysautonomia following traumatic brain injury. Brain matic brain injury: evidence for overresponsiveness to Inj 2004; 18: 409–417. afferent stimuli. Arch Phys Med Rehabil 2009; 90: 580– 8. Do D, Sheen VL, Bromfield E. Treatment of paroxysmal 586. sympathetic storm with labetalol. J Neurol Neurosurg 20. Gentry LR. Imaging of closed head injury. Radiology Psychiatry 2000; 69: 832–833. 1994; 191: 1–17. 9. Boeve BF, Wijdicks EF, Benarroch EE, Schmidt KD. Paroxysmal sympathetic storms (‘‘diencephalic seizures’’)

2010 The Author(s) European Journal of Neurology 2010 EFNS European Journal of Neurology 17, 1172–1177