Comparison Between Transcranial Color Doppler Ultrasonography and Angiography in the Conﬁrmation of Brain Death

Home , Cerebral angiography, Doppler ultrasonography, Transcranial Doppler

J. Poularas, D. Karakitsos, G. Kouraklis, A. Kostakis, E. De Groot, A. Kalogeromitros, D. Bilalis, J. Boletis, and A. Karabinis

ABSTRACT Objective. Cerebral blood flow tests have increasingly been advocated for the confirmation of brain death (BD). Angiography has been considered the gold standard in the diagnosis of BD but is invasive. We validated transcranial color Doppler ultrasonography (TCD) to confirm BD by comparing it to angiography. Patients and methods. Forty patients experienced the clinical diagnosis of brain death due to head injury in 19 cases (47.5%), cerebral hemorrhage in 11 (27.5%), subarachnoid hemorrhage in 7 (17.5%), and cerebral infarction in 3 (7.5%). Blood pressure, heart rate,

SPO2, and PCO2 were monitored throughout the study. Patients were excluded if episodes of hypoxia, arrhythmia, and hypotension occurred during examinations, or if the TCD was not technically feasible. Results. Both angiography and TCD confirmed BD in all patients. The agreement between the above methods to confirm BD was 100%. Angiography showed the absence of filling of intracranial arteries, while TCD revealed: (1) brief systolic forward flow or systolic spikes and diastolic reversed flow (50%); (2) brief systolic forward flow or systolic spikes and no diastolic flow (25%); (3) no demonstrable flow in a patient in whom flow had been clearly documented on a previous TCD examination (12.5%). Five patients required repeated TCD examinations, because of initial detection of a diastolic to-and-fro flow pattern. BD was confirmed by TCD in the above patients after 30 hours of clinical BD. Conclusion. TCD was a sensitive tool to diagnose BD, affording a reliable alternative examination to standard angiography.

INCE THE FIRST REPORTS describing irreversible a noninvasive and inexpensive examination, which has been S coma and brain death (BD),1,2 both the definition of used for the confirmation of cessation of cerebral circula- BD and its diagnostic criteria have evolved considerably,3,4 tion in brain-dead patients.11 mainly due to the advent of cardiopulmonary support for Previous studies have shown the accuracy of transcra- severely brain-injured patients. BD is defined as the irre- nial color-coded duplex ultrasound to estimate stenosis of versible cessation of all hemispheric, cerebellar, and brain stem neurological functions. In addition, an essential fea- From the Intensive Care Unit (J.P., D.K., A.Ka., A.Kar.), Geni- ture of BD is “brain tamponade,” that is, the arrest of brain matas General State Hospital of Athens, Athens, Greece; Aca- circulation caused by increased intracranial pressure.5 As a demic Medical Center (E.D.G.), Amsterdam, The Netherlands; result, cerebral blood flow (CBF) tests have increasingly and Second Department of Propedeutic Surgery (G.K., A.Ko.), been advocated for confirmation and/or estimation of University of Athens, School of Medicine, Transplantation Cen- cesstion of cerebral circulation.6–10 Such tests include (1) ter; Intensive Care Unit (D.B.), and Transplantation Center (J.B.), Laiko General Hospital, Athens, Greece. four vessel angiography, which has been considered the Address reprint requests to Andreas Karabinis, MD, PhD, gold standard for the diagnosis of BD but is invasive; (2) Assistant Professor in Intensive Care Medicine, Athens Univer- electroencephalography, which is noninvasive, however is sity, School of Medicine, Director of ICU Department, Genimatas time-consuming and unreliable in the presence of seda- General State Hospital, Mesogeion Ave 154, Athens, Greece. tives9; and (3) transcranial doppler ultrasonography (TCD), E-mail: [email protected]

Transplantation Proceedings, 38, 1213–1217 (2006) 1213 1214 POULARAS, KARAKITSOS, KOURAKLIS ET AL intracranial vessels to be similar to that of contrast angiog- vertebral artery. A Siemens DSA device (Angiotron) was used to raphy.12 In this study, we compared TCD with contrast obtain six images per second for 20 s. Two angiographic patterns angiography to validate the former method in the confir- were considered indicative of BD: (1) absent filling of the intra- mation of BD diagnosis. Furthermore, we have described a cranial arteries at the entry into the skull (at the foramen magnum novel color Doppler sonographic pattern observed in brain- in the posterior circulation and at the petrosal portion of the dead patients. carotid artery in the anterior circulation); and (2) minimal arterial opacification with absent parenchymal and venous phases. TCD examinations were performed with an HDI 3500 (ATL, PATIENTS AND METHODS Philips, Bothell) equipped with a 2-MHz transducer. Bilateral Patients middle cerebral arteries, anterior cerebral arteries, and basilar and From the 70 consecutive patients with Glasgow coma scale score internal carotid arteries were insulated. The following flow patterns less than 5 who were included in the study, 50 were diagnosed as were regarded as consistent with the TCD diagnosis of brain death: clinically brain dead. All patients were continuously monitored for brief systolic forward flow or systolic spikes and diastolic reversed systemic blood pressure, heart rate, intracranial pressure, PaO2, flow, brief systolic forward flow or systolic spikes and no diastolic and PaCO2, to maintain steady-state conditions and prevent hypo- flow, or no demonstrable flow in a patient in whom flow had been tension (systolic blood pressure Ͼ 110 mm Hg), bradycardia (heart clearly documented in a previous TCD examination.11 The above Ͼ Ͻ rate 60 beats/min), and hypoxia (SPao2 95%). For all patients, TCD findings were accepted as confirmation of BD only when they the PaCO2 was maintained at 33 to 35 mm Hg throughout the study. were found bilaterally or in at least three different arteries and Any episode of hypoxia or hypotension during angiography or TCD persisted for at least 3 minutes within the same examination. In resulted in exclusion of the patient from the study. patients with a unilateral absent temporal bone window, the finding Five patients were excluded because no flow could be detected at of typical TCD signals in three different arterial segments on one the baseline TCD examination (absence of bone window), three side was considered as confirmation. In contrast, the finding of other patients because apnea tests could not be performed due to diastolic back-and-forward flow (to-and-fro flow) was not accepted desaturation and arrhythmia, and two patients due to hypoxia as confirmation of BD.13 The examination was repeated until BD Ͻ (SPO2 95%) during the TCD measurements. In total, the study was confirmed ultrasonographically. Absence of flow in all arterial population keep 40 patients with clinical brain death. Twenty segments during the first TCD examination may be caused by total Ϯ patients were men and 20 were women of mean age 48 18.5 cerebral circulatory arrest or possibly by the absence of an ade- years. Brain death was attributed to various pathological conditions quate temporal bone window. All cases in which no flow was (Table 1). The study conformed to the principles outlined in the detected at the initial TCD examination were considered as having Declaration of Helsinki and was approved by the Institutional an “absent bone window.” In clinically brain-dead patients, the Ethics Committee. TCD was repeated if previous TCD examinations did not show a flow pattern consistent with BD. All patients were repeatedly Clinical Diagnosis of Brain Death and Confirmatory Tests evaluated more than twice during the study period. Furthermore, Clinical BD was diagnosed according to the following criteria: deep we recorded two-dimensional color mode patterns of interest irreversible coma, absence of brain stem reflexes, flat electroenceph- during each TCD examination. All observers who performed the alogram, and a positive apnea test in a normothermic nondrugged above confirmation tests were blinded to the patient’s identity. patient.4,5 According to Hellenic state law, confirmatory tests (stan- Statistical analysis was performed using Cohen’s k14 to verify dard angiography) establishing cessation of CBF are required for agreement between TCD and angiography. documentary records and to declare a patient brain dead. Following clinical diagnosis of BD by angiography and repeated bedside TCD evaluation, the family’s consent was obtained for RESULTS organ donation by a committee consisting of an intensive care consultant, a neurologist, a cardiologist, a neurosurgeon, and an For all cases, angiographic recordings took approximately anaesthesiologist. In this series, 20 patients became organ and tissue 20 minutes, while TCD recordings took approximately 30 donors. minutes. Thus, the entire study was completed within 1 Four-vessel angiography was performed by femoral catheteriza- hour for all patients. Both angiography and TCD showed tion with selective injection of iodinated contrast medium into the the arrest of brain perfusion in all patients (Fig 1). common carotid artery and vertebral artery of each side. Iodinated The only observed angiographic pattern was the absence nonionic medium (iodine concentration, 300 mg/mL) was injected of filling of intracranial arteries. The agreement between using pump at a dose of 12 mL (rate, 6 mL/s) into the common ϭ carotid artery and at a dose of 8 mL (rate of 4 mL/s) into the angiography and TCD in our series was 100% (k 1) with both investigations showing no CBF in all cases. In brain- Table 1. Cause of Brain Death in 40 Patients in the dead patients, the average time from admission to the Ϯ Present Series intensive care unit and clinical diagnosis of BD was 70.8 110 hours. Initial TCD examination of the 40 patients Diagnosis Total No. of Patients (%) clinically diagnosed as brain dead and having an adequate Head injury 19 (47.5%) bone window revealed flow patterns confirming BD in 35 Cerebral hemorrhage 11 (27.5%) patients while forward systole-diastolic flow or a diastolic Subarachnoid hemorrhage 7 (17.5%) to-and-fro flow pattern was detected in four patients with Cerebral infarct 3 (7.5%) head injuries and in one patient with brain infarcts and Total 40 (100%) anoxic brain injury. ULTRASONOGRAPHY AND ANGIOGRAPHY 1215

“beacon” signal. This signal was observed in all patients during the time of cessation of cerebral circulation according to synchronous Doppler signals in all studied intracranial vessels (Fig 2).

DISCUSSION Previous studies have clearly demonstrated that cessation of CBF conﬁrms BD.6–9,11,13 The above strategy has become more popular since the development of the modern intensive care for severely brain-injured patients and the growing demand for tissue and organ donors. In this study, we

Fig 1. Four-vessel angiography (top) and TCD (bottom) results for a brain-dead patient. Angiography shows no filling of the intracranial arteries at the entry into the skull, while TCD shows a pattern of brief systolic forward flow or systolic spikes and diastolic reverse flow.

TCD showed the following flow patterns: (1) brief systolic forward flow or systolic spikes and diastolic reversed flow (20 patients, 50%); (2) brief systolic forward flow or systolic spikes and no diastolic flow (10 patients, 25%); (3) no demonstrable flow in a patient in whom flow had been clearly documented in a previous TCD examination (five patients, 12.5%). The mean duration between the clinical examination and the first TCD session was 3 Ϯ 1.5 hours. The sensitivity and specificity of the first TCD examination were 87.5% and 100%, respectively. The five patients in whom initial TCD examination failed to confirm BD underwent a second TCD examination 10 Ϯ 5.5 hours. Thereafter, clinical BD that showed confirmation of TCD signs was seen in three patients. The sensitivity after the second TCD examination was 95%. The remaining two patients underwent a third TCD examination 30 hours after clinical diagnosis of BD that showed confirmatory TCD Fig 2. Two-dimensional color Doppler sonography results, in signs of BD. The sensitivity after the third TCD examina- the area of left middle cerebral artery, for a brain-dead patient tion was 100%. showing that systolic flow (top) and diastolic flow (middle) With respect to our results, two-dimensional color mode coexisted simultaneously in time resulting in a pulsating flash recordings demonstrated an interesting pattern as systolic (the “beacon” signal). TCD results for the same brain-dead and diastolic flow coexisted simultaneously, resulting in a patient during the same session showing a pattern of systolic pulsating flashing signal akin to that of a beacon, the spikes and diastolic reverse flow (bottom). 1216 POULARAS, KARAKITSOS, KOURAKLIS ET AL validated TCD to confirm BD compared with contrast autoregulatory “threshold” or “break point.”24 Mathemat- angiography. ical models have been generated to predict the CPP and Angiography is invasive, which could result in severe cerebral autoregulatory status of the cerebral circulation complications, such as vasospasm, subintimal injection, and from TCD FV waveform analysis. The intracranial vascular thrombosis, leading to a false image of absent flow and resistance (R) may be depicted from Hagen-Poisseuille’s 15 cerebral ischemia. As a result of its invasiveness, ethical law for steady laminar flow (F) in long cylindrical tubes: F questions arise pertaining to its use to diagnose BD, since ϭ P/(8 ␮l)/(␲r4) ϭ P/R (where P ϭ pressure; ␮ ϭ viscosity; the patient is exposed to possible complications with no l ϭ length; r ϭ radius; R ϭ resistance). According to this potential benefits. Its invasiveness may also make test concept, CBF at any time should be given by: CBF ϭ repetition troublesome in patients who are not brain dead 24 (MABP Ϫ ICP)/R ϭ CPP/R. Noninvasive calculation of during the first examination.16 CPP (nCPP) by means of TCD has previously been re- TCD confirmation of BD in children and adults has been ported based on the formula nCPP ϭ MABP ϫ FVd/FVm considered operator-dependent, not 100% specific, and ϩ 14 mm Hg (where FVd is diastolic flow and FVm is mean technically demanding. However, it is a noninvasive exam- 25 FVd ination that is inexpensive and manageable when per- flow). According to the above equations as CPP falls formed by experienced observers.17–19 approaches zero and forward movement of blood only In this study, TCD was not technically feasible in approx- occurs during systole. When CPP falls further and vascular imately 10% of clinically brain-dead patients. Our results resistance increases, a state of reverberant flow can be seen, indicate that TCD specificity to confirm the diagnosis of BD where reversed flow occurs during diastole. Once backflow was 100% similar to other reports.19–22 TCD sensitivity also during diastole equals forward flow during systole, no net increased after repeated examinations, which is in accor- flow of blood occurs and cerebral tamponade in the terri- dance with the findings of Dosemeci et al.23 Furthermore, tory of the insonated vessel is complete.17,26 Therefore, our our findings showed that diastolic to-and-fro flow pattern color mode recordings may be an image of the transitory was detected in four patients with head injuries and one “flickering” of cerebral circulation observed as autoregula- patient with anoxic brain injury. The preservation of for- tion fails and vascular resistance increases just before ward flow during diastole may be partially due to the fact cessation of CBF becomes permanent and brain tamponade that decompressive surgical procedures and/or ventricular is complete. derivations were performed in these patients. Thus, skull hermetism was affected during the progression to BD.13,23 Hadani et al20 described the preservation of approximately Technical Considerations and Study Limitations normal flow in all major intracranial vessel in one case of The major technical limitations of TCD include the exam- BD resulting from anoxic brain damage. It is presumed that ination of the posterior circulation, which can be extremely in some cases of anoxic injury, irreversible failure of all vital difficult in a bedridden patient who is intubated and an functions of the brain occurs without persistent elevation of “absent bone window.” Furthermore, observers who are intracranial pressure (ICP), causing total cerebral circula- performing TCD in brain-dead patients must be well tory arrest. However, our results showed that all patients trained and adequately experienced. In many countries, an who had obvious flow, even at a time when they were observation time between 6 and 72 hours is required after clinically brain dead, eventually developed cerebral circula- the first examination showing brain death to the declaration tory arrest. Finally, the TCD waveforms observed in our 27 brain-dead patients were similar to those reported by other of the patient as brain dead. This time could be shortened groups.11,17,19 in patients in whom both clinical evidence of brain death In this study, we also described a novel color Doppler and TCD confirmation of cerebral circulatory arrest is sonographic pattern observed in brain-dead patients. Our demonstrated. Such patients may be declared brain dead two-dimensional color mode recordings demonstrated an without a second examination after a period of observation interesting pattern as systolic and diastolic flow coexisted if appropriate legal measurements are taken. Only a few simultaneously in time, resulting in a pulsating flashing signal countries allow confirmatory tests to shorten the time to 28 akin to that of a beacon. This signal was observed in all declaration. However, patients who are clinically brain patients during the time of cessation of cerebral circulation dead and still have obvious CBF depicted by TCD pose a according to synchronous Doppler signals in all intracranial medicolegal problem. Moreover, no global consensus exists vessels. Although cerebrovascular hemodynamics are com- if confirmatory tests are really necessary in patients with the plex, cerebral perfusion pressure (CPP) can be calculated as clinical diagnosis of BD.27 the difference between the mean arterial blood pressure In conclusion, our results showed perfect agreement (MABP) and the ICP. Furthermore, assessment of cerebral between contrast angiography and TCD for the confirma- autoregulation using TCD depends on the assumption that tion of intracranial circulatory arrest in BD. Although relative changes in blood flow velocity (FV) correlate with further studies are clearly required, we suggest that TCD relative changes in CBF. The value of CPP at which was a sensitive noninvasive tool in the diagnosis of BD and autoregulation fails and FV begins to fall is called the a good candidate for the “gold standard” of diagnosis. ULTRASONOGRAPHY AND ANGIOGRAPHY 1217

REFERENCES 14. Cohen J: A coefficient of agreement for nominal scales. Educ Psychol Meas 20:27, 1960 1. Mollaret P, Goulon M: Le coma depassé. Rev Neurol 101:3, 15. Heiserman JE, Dean BL, Hodak JA, et al: Neurologic 1959 complications of cerebral angiography. AJNR Am J Neuroradiol 2. Harvard Medical School Ad Hoc Committee to Examine the 15:1401, 1994 Definition of Brain Death: A definition of irreversible coma. 16. Wijdicks EF: Determining brain death in adults. Neurology JAMA 205:337, 1968 45:1003, 1995 3. Criteria for the diagnosis of brain stem death. Review by a 17. Petty GW, Mohr JP, Pedley TA, et al: The role of transcra- working group convened by the Royal College of Physicians and nial Doppler in confirming brain death: sensitivity, specificity, and endorsed by the Conference of Medical Royal Colleges and their suggestions for performance and interpretation. Neurology 40:300, Facilties in the United Kingdom. JR Coll Physicians London 1990 29:381, 1995 18. Rodriguez RA, Cornel G, Alghofaili F, et al: Transcranial 4. Guidelines for the determination of death. Report of the Doppler during suspected brain death in children: potential limitation medical consultants on the diagnosis of death to the President’s in patients with cardiac “shunt.” Pediatr Crit Care Med 3:153, 2002 commission for the study of ethical problems in medicine and 19. Feri M, Ralli L, Felici M, et al: Transcranial Doppler and biomedical and behavioural research. JAMA 246:2184, 1981 brain death diagnosis. Crit Care Med 22:1120, 1994 5. Wijdicks EFM: The diagnosis of brain death. N Engl J Med 344:1215, 2001 20. Hadani M, Bruk B, Ram Z, et al: Application of transcranial 6. Paolin A, Manuali A, Di Paola F, et al: Reliability in diagnosis Doppler ultrasonography for the diagnosis of brain death. Inten- of brain death. Intensive Care Med 21:657, 1995 sive Care Med 25:822, 1999 7. Ducrocq X, Braun M, Debouverie M, et al: Brain death and 21. Zurynski Y, Dorsch N, Pearson I, et al: Transcranial Dopp- transcranial Doppler: experience in 130 cases of brain dead pa- ler ultrasound in brain death: experience in 140 patients. Neurol tients. J Neurol Sci 160:41, 1998 Res 13:248, 1991 8. Braun M, Ducrocq X, Huot JC, et al: Intravenous angiogra- 22. Kirkham FJ, Levin SD, Padaychee TS: Transcranial pulsed phy in brain death: report of 140 patients. Neuroradiology 39:400, Doppler ultrasound findings in brain stem death. J Neurol Neuro- 1997 surg Psychiatry 50:1504, 1987 9. Buchner H, Schuchardt V: Reliability of electroencephalo- 23. Dosemeci L, Dora B, Yilmaz M, et al: Utility of transcranial gram in the diagnosis of brain death. Eur Neurol 30:138, 1990 doppler ultrasonography for confirmatory diagnosis of brain death: 10. Munari M, Zucchetta P, Carollo C, et al: Confirmatory tests two sides of the coin. Transplantation 77:71, 2004 in the diagnosis of brain death: comparison between SPECT and 24. Aaslid R, Lundar T, Lindegaard KF, et al: Estimation of contrast angiography. Crit Care Med 33:2068, 2005 cerebral perfusion pressure and transcranial Doppler recordings. 11. Ducrocq X, Hassler W, Moritake K, et al: Consensus on In Miller JD, Teasdale GM, and Ronan JO (eds): Intracranial diagnosis of cerebral circulatory arrest using Doppler-sonography: Pressure VI. Berlin: Springer-Verlag; 1986, p 226 task force group on cerebral death of the Neurosonology Research 25. Czosnyka M, Matta BF, Smielewski P, et al: Cerebral Group of the World Federation of Neurology. J Neurol Sci perfusion pressure in head-injured patients: a non-invasive assess- 159:145, 1998 ment using transcranial Doppler ultrasonography. J Neurosurg 12. Laumgartner RW, Mattle HP, Schroth G: Assessment of 88:802, 1998 Ն50% and Ͻ50% intracranial stenoses by transcranial color-coded 26. Newell DW, Grady MS, Sirotta P: Evaluation of brain death duplex sonography. Stroke 30:87, 1999 using transcranial Doppler. Neurosurgery 24:509, 1989 13. Cabrer C, Dominguez-Roldan JM, Manyalich M, et al: 27. Wijdicks EF: Brain death worldwide: accepted fact but no Persistence of intracranial diastolic flow in transcranial Doppler global consensus in diagnostic criteria. Neurology 58:20, 2002 sonography exploration of patients in brain death. Transplant Proc 28. Haupt WF, Rudolf J: European brain death codes: a com- 35:1642, 2003 parison of national guidelines. J Neurol 246:432, 1999