Risk factors for sudden unexpected death in epilepsy: a controlled prospective study based on coroners cases

Transcription

1 Seizure 2003; 12: doi: /s (02) Risk factors for sudden unexpected death in epilepsy: a controlled prospective study based on coroners cases KENNETH OPESKIN & SAMUEL F. BERKOVIC Department of Anatomical Pathology, St Vincent s Hospital, 41 Victoria Pde, Fitzroy 3065, Australia; Epilepsy Research Institute, Department of Medicine (Neurology), Austin and Repatriation Medical Centre, University of Melbourne, Heidelberg West 3081, Australia Correspondence to: Dr Kenneth Opeskin, M.D., F.R.C.P.A., Department of Anatomical Pathology, St Vincent s Hospital, 41 Victoria Pde, Fitzroy 3065, Australia. opeskik@svhm.org.au We performed a controlled prospective study of pathologically verified sudden unexpected death in epilepsy (SUDEP) in a coronial setting, to identify risk factors. We prospectively studied coronial deaths of people with epilepsy in Vic., Australia, during a 21-month period. Fifty SUDEP and 50 subjects with epilepsy who died of other causes (controls) were collected sequentially. Clinical data was obtained shortly after death from questionnaires completed by treating doctors, discussion with family members and coronial files, including police reports of death, autopsy and toxicology reports. Factors assessed were age, sex, duration of epilepsy, type of seizure(s), seizure frequency, symptomatic epilepsy, including post-traumatic epilepsy, presence of structural brain lesion, idiopathic epilepsy, mental retardation, psychiatric illness, including dementia, recent stressful life event, particular antiepileptic drugs (AEDs) and AED polytherapy, compliance with AED treatment, psychotropic drug prescription, alcohol and other substance abuse, place of death and evidence of terminal seizure. The SUDEP group was characterised by younger age and higher proportion found dead in bed and with evidence of terminal seizure compared to controls. The profile of patients at risk for SUDEP are young people with epilepsy. They are most likely to die in sleep and our data support the view that SUDEP is a seizure-related event. This, taken in conjunction with the finding that there was no increased risk associated with a particular AED in monotherapy or multiple AEDs suggests that attempts to better treat patients epilepsy with AEDs might decrease the risk of SUDEP. Although the literature suggests that SUDEP is more frequent in patients with severe epilepsy, we did not find a correlation with seizure frequency suggesting that other clinical indices may be more important BEA Trading Ltd. Published by Elsevier Science Ltd. All rights reserved. Key words: SUDEP; risk factor; epilepsy; sudden death. INTRODUCTION People with epilepsy have an increased incidence of sudden unexpected death (SUDEP) 1 6. Estimates of SUDEP incidence vary. In population-based studies, the SUDEP rate is generally between 1:500 and 1:1000 per year For those with severe epilepsy, the rate of SUDEP is approximately 1:200 per year Evidence for terminal seizure is found in a variable proportion of cases. However, the best circumstantial evidence seems to favour that most cases are seizure related 11, 21, Mechanisms postulated to play a role in SUDEP include cardiac arrhythmia precipitated by seizure discharge acting via the autonomic nervous system 31 38, respiratory arrest 39 41, neurogenic pulmonary oedema 42, 43 and asphyxiation 29, 41, 44, 45. Previous studies have suggested a number of associated factors but results are not completely consistent between studies. These include young age8, 11, 12, 14, 19, 21, 23, 24, 29, 30, 46 53, male sex 9, 13, 14, 27, 29, 46, 48, 50 52, generalised tonic clonic seizures8, 11, 23, 24, 46, 47, 50, 51, 54,low 11, 14, 47 or high seizure frequency16, 18 21, 23 27, 46, 55, 56, chronic epilepsy 11, 14, 47, 54, symptomatic epilepsy 23, 57, 58, structural brain lesion 23, 57, 58, posttraumatic epilepsy 23, 57, 58, early onset epilepsy 44, 56, mental retardation 14, 51, 54, 59 63, recent unusually stressful life event 46, particular antiepileptic drugs (AEDs) 27, 46, increasing number of AEDs /$ BEA Trading Ltd. Published by Elsevier Science Ltd. All rights reserved.

2 Risk factors for SUDEP 457 prescribed 13, 54, 56, 64, poor compliance with AED treatment 8, 11, 12, 30, 48, 50 52, 65 68, psychotropic drug prescription, including anxiolytic use 13, 50, 56, alcohol 8, 11 13, 50, 51, 67 and sleep 21, 23, 30, 46, 50, 66, Most previous studies, however, have not been controlled and the significance of such associated factors remains uncertain. We performed a prospective controlled study of risk factors for SUDEP in a coronial population in Vic., Australia. We concentrated on risk factors suggested by previous investigators. METHODS Study population The Victorian Institute of Forensic Medicine provides forensic pathology services for approximately 4 million people. The population base for this study was deaths that occurred in Vic., Australia, that were reported to the coroner and autopsied between December 1997 and August This included deaths (a) for which a death certificate was not completed as the doctor was not sure of the cause of death, (b) that were unexpected, unnatural or violent or resulted directly or indirectly from accident or injury, (c) that were of people held compulsorily in institutions as they or their families were unable to look after them, (d) of persons whose identity was unknown, and (e) that occurred during or as a result of an anaesthetic and were not due to natural causes. Deaths due to natural causes referred to deaths that resulted from inherent disease, such as ischemic heart disease as opposed to deaths due to accident, suicide or homicide. The average yearly population for Victoria was The annual mortality rate was 7.0 per 1000 population (source Australian Bureau of Statistics). Coronial autopsies accounted for 9% of the total number of deaths for Victoria over this period. Approval for this study was given by the Victorian Institute of Forensic Medicine Ethics Committee. Cases We identified cases as follows: police reports of death were searched for subjects with a history of epilepsy. We defined SUDEP according to the criteria of Nashef 71. With few exceptions, the autopsies in this study were performed by and the brains had formal neuropathologic examination by the one neuropathologist (KO). Control subjects were those with a history of epilepsy who died from other causes. Fifty SUDEP and 50 control subjects were collected sequentially and not matched in any way other than for assessment period. Assessment of clinical variables Questionnaires were sent to treating doctors requesting information shortly after the death. Information sought included age at seizure onset, cause, seizure type(s) and frequency, neurological examination, diagnostic tests, treatment and AED compliance, recent AED medication changes, general medical and psychiatric history, intellectual disability, non-aed treatment, alcohol and other substance abuse, recent life stresses and family history. The treating doctors were asked to classify seizure type as generalised tonic clonic/grand mal, partial simple or complex, myoclonic, multiple or unknown. Psychiatric history requested included detail of severe depression, desire to commit suicide, schizophrenia, violent behaviour and behavioural disorders. Doctors were asked to indicate whether compliance with AED treatment was good or poor (based on detail of medication usage and ante-mortem AED levels). Psychotropic drug prescription history requested included the use of antipsychotic drugs and antidepressants. Family members were also spoken to, requesting information on alcohol and other substance abuse, compliance with AED treatment, recent stressful life event, whether the deceased was found dead, where the body was found, whether terminal seizure was witnessed and whether there was urinary incontinence or disturbance of bed clothes (indicating terminal seizure) at the scene. When data from the treating doctor was in conflict with information from the subjects family, an assessment was made as to which was reliable. The data that was considered to be reliable was accepted. Where conflict remained the clinical variable was classified as not known. Police reports of death giving details of the circumstances surrounding the death and the scene of death, and autopsy and post-mortem toxicology reports were also examined. Evidence at the scene for unwitnessed terminal seizure was considered to include the body being found half on, half off the bed or wrapped up in sheets, urinary incontinence and bruising of lips or tongue at autopsy. The probable etiology of epilepsy was finally determined on the basis of clinical details and brain examination at autopsy. Epilepsy was classified as idiopathic epilepsy in those who had been developmentally normal, had no significant neurological abnormalities and had no known underlying cause for epilepsy. Epilepsy was classified as symptomatic epilepsy in those with known underlying cerebral, metabolic or genetic disorder and in those in whom this was suspected from the findings of intellectual deficiency and/or significant abnormal neurological abnormalities 72.

3 458 K. Opeskin & S. F. Berkovic Analysis Given the reasonably large number of subjects and the possible correlations between some variables, the question of multivariate versus univariate statistical analysis was carefully considered. Ultimately, the nature of the data, including heteroscedasticity and skew (see below), which would have resulted in very small numbers in some cells, led to the decision to use univariate analyses. Non-parametric method were used when necessary. Statistical analysis using Student s t-tests, chi-square analysis and Mann Whitney U-test was performed. Levene s test for homogeneity of variance was met for all variables with the exception of age and seizure frequency. For these two variables the assumption of equal variance was violated (F = 5.22, P = and F = 8.14, P = 0.006, respectively), so separate variance estimates were used to calculate Student s t-tests of the difference between the means of two independent groups. In the case of seizure frequency, violation of equality of variance was so marked that Mann Whitney U-test was run in order to check the findings of t-test. Regarding com- parison of numbers of subjects found dead in bed in the SUDEP and control groups for assessment of sleep as a risk factor, subjects who were already in a hospital bed were excluded as many of these found dead in bed may have been false positives. RESULTS Demography There were 4375 coronial autopsies between December 1997 and August 1999, of which 166 were epileptic deaths. These epileptic deaths included those in which there was a history of epilepsy and the people had died from accidents, such as drowning, suicide, unsuspected unnatural disease and SUDEP. Of these epileptic deaths 50 (30%) were SUDEPs. The causes of death in the control group are shown in Table 1. Using an estimate of prevalence of epilepsy in Victoria of 0.5% and an average yearly population of (source Australian Bureau of Statistics), the incidence of SUDEP was approximately 1:780 per year. Table 1: Causes of death in the control group. Cause of death Number of cases (n = 50) Natural Cardiovascular Ischemic heart disease 11 Valvular heart disease 1 Congenital heart disease 1 Other (combined ischemic and valvular disease, ruptured abdominal 4 aortic aneurysm, myocarditis, cardiomyopathy) Respiratory Pneumonia 6 Pulmonary thrombolism 2 Central nervous system Hypoxic brain damage 1 Other Combined pneumonia ischemic and valvular disease, bowel cancer and 4 pneumonia, multiorgan organ system failure, sepsis Subtotal 30 Accidental Heroin 5 Motor vehicle 2 Head injury/falls 3 Drowning 2 Fire 2 Airway obstruction 3 Subtotal 17 Suicide Combined drug toxicity 2 Multiple injuries (train) 1 Subtotal 3

4 Risk factors for SUDEP 459 The age of the SUDEP subjects ranged from 4 to 71 years and the control subjects 3 to 78 years. The SUDEP group was significantly younger than the control (mean 34.6 ± 13.0 SD vs ± 17.8 years). There were five subjects less than 18 years of age in the SUDEP group versus four in the control. In the SUDEP group, 54% were male, 46% female and in the control group 78% male, 22% female. Chi-square analysis indicated the proportion of females to males was greater in the SUDEP group compared with the control group (χ 2 = 6.54, df = 1, 0.02 >P>0.01). Of the nine traumatic or suicide deaths in the control group, six were male. Clinical features There was no difference in the average duration of epilepsy between SUDEP and control groups (mean = 16.8 ± 13.0 SD vs ± 17.3). Twenty-three of 40 SUDEPs and 20 of 46 controls had early onset epilepsy (onset 0 15 years) and 1 of 40 SUDEPs and 7 of 46 controls had onset of epilepsy after 45 years. The Table 2: Type of seizures in SUDEP and control groups. Type of Seizure SUDEPs Controls GTC Partial 3 5 Multiple GTC and partial 11 9 GTC, partial, absence 1 0 GTC, absence 3 4 GTC, partial, myoclonic, absence 1 0 GTC, myoclonic 0 1 GTC, partial, myoclonic 0 1 Absence 0 2 Unknown 1 1 GTC: generalised tonic clonic seizure. Table 3: Aetiology of epilepsy in SUDEP and control groups. Aetiology of epilepsy SUDEPs Controls ILAE classification Symptomatic/cryptogenic epilepsy Idiopathic epilepsy 5 8 Not known 11 5 Post-traumatic epilepsy Previous head trauma 8 6 No head trauma Not known 0 1 Structural brain lesion Structural lesion No lesion Not known 0 1 ILAE: International League Against Epilepsy. types of seizures (Table 2) and aetiology (Table 3) in SUDEP and control groups were similar. The mean number of seizures between SUDEP and control groups was not significantly different (mean = 20.9 ± 10.4 SD vs. 6.5 ± 1.3 SD, t = 1.37, df = 39.2, P = 0.178). The variance between the two groups was significantly different, consistent with a subgroup of the SUDEP group having an unduly large number of seizures. Four of 50 SUDEPs and 0 of 50 controls had greater than 50 seizures per year whereas 15 SUDEPs and 21 controls had less than 2 seizures per year (Table 4). The neuropsychiatric features in SUDEP and control groups are shown in Table 5. In particular, psychiatric illness was not more common in the SUDEP group (χ 2 = 2.8, df = 1, ns). Mental retardation, dementia and recent unusually stressful life event were not more common in the SUDEP group. Table 4: Frequency of seizures in SUDEP and control groups. Seizure frequency (per year) SUDEPs (n = 50) Controls (n = 50) < >50 4 ( =130, 900, 350, seizures per year) Unknown 10 8 Mean ± SD a 20.9 ± ± 1.3 a ns. Table 5: Neuropsychiatric features in SUDEP and control groups. Neuropsychiatric features SUDEPs Controls Mental retardation Yes No Not known 7 4 Psychiatric illness a Yes No Not known 5 4 Dementia Yes 0 3 No Not known 6 2 Recent stressful life event Yes No Not known 3 6 a ns.

5 460 K. Opeskin & S. F. Berkovic Table 6: Antiepileptic drug prescription in SUDEP and control groups. Number and type of AEDs SUDEPs Controls Number of AEDs No drug 7 4 One drug Two drugs Three drugs 10 4 Four drugs 3 2 Unknown 0 1 Type of AED Carbamazepine Sodium valproate Clonazepam a 9 4 Lamotrigine 9 8 Phenytoin a Phenobarbitone 2 2 Gabapentin 2 1 Vigabatrin 0 3 Diazepam 1 3 Ethosuximide 0 1 Clobazam 3 0 Topiramate 1 0 Primidone 2 2 Not known 0 1 AED: antiepileptic drug. a ns. Medication details The total number of cases receiving single, dual, triple or quadruple AED therapy is shown in Table 6. No AED or combination was more common in the SUDEP group. The proportion of SUDEP cases on single, dual, triple or quadruple AED treatment did not differ from the control sample. There was also no difference between the two groups in relation to clonazepam (χ 2 = 0.9, df = 1, ns) or phenytoin prescription (χ 2 = 0.3, df = 1, ns). Compliance with AED treatment is shown in Table 7. Poor compliance (based on treating doctors assessment of medication usage and AED levels and assessment by family members) was not more common in the SUDEP group. Psychotropic drug prescription was not more common in the SUDEP group. In particular, thioridazine use was not more common in the SUDEP group Table 7: Compliance with antiepileptic drug treatment in SUDEP and control groups. Compliance SUDEPs Controls Poor Good Not on AED treatment 4 3 Not known 3 3 AED: antiepileptic drug. Table 8: Psychotropic drug prescription in SUDEP and control groups. Number of psychotropic drugs SUDEPs Controls No drug One drug 9 12 Two drugs 3 2 Not known 3 5 Type of psychotropic drug SUDEPs Controls Thioridazine a 4 1 Sertraline 2 2 Paroxetine 2 1 Clozapine 1 0 Haloperidol 1 2 Trifluoperazine 2 0 Clomipramine 1 1 Chlorpromazine 2 1 Fluoxetine 0 3 Respiridone 0 2 Olanzipine 0 1 Doxepin 0 1 Fluphenazine 1 0 Haloperidol 0 1 Not known 5 a ns. Table 9: Alcohol and other substance abuse in SUDEP and control groups. Alcohol abuse SUDEPs Controls Yes 8 8 No Non-alcohol substance abuse Yes 5 11 No Not known 2 1 (χ 2 = 0.8, df = 1, ns). Anxiolytic use was significantly less common in the SUDEP group (2/48 cases vs. 8/45 cases in the control group) (χ 2 = 4.34, df = 1, P<0.05). Nitrazepam use was not more common in the SUDEP group (0/46 cases vs. 3/45 cases in the control group) (Table 8). Alcohol and other substance abuse was not more common in the SUDEP group (Table 9). Death details More of the SUDEP group were found dead in bed at home (χ 2 = 8.1, df = 3, 0.05 >P>0.02) and evidence of terminal seizure was significantly more common in the SUDEP group (χ 2 = 9.4, df = 3, 0.05 > P>0.02) (Table 10). Six SUDEPs were witnessed and seizure occurred in all of these. Three deaths in the control group were witnessed and a seizure was witnessed in one of these. There was no significant difference between the number of witnessed deaths,

6 Risk factors for SUDEP 461 Table 10: Numbers found dead in bed and showing evidence of terminal seizure in SUDEP and control groups. Death details SUDEPs Controls Place of death Hospital bed 2 11 Home bed a Not found in bed Not known 0 1 Evidence for terminal seizure Yes a 22 8 No Not known a More common in SUDEP group. and the number of seizures in witnessed deaths in the SUDEP group versus control group (χ 2 = 3.0, df = 1, 0.10 >P>0.05). DISCUSSION Previous data on associated-risk factors for SUDEP has come from selected series of subjects from centres for patients with severe epilepsy, epilepsy referral centres, coroners offices or larger cohort studies. To date, there have only been a small number of controlled studies looking at risk factors for SUDEP 9, 19, 27, 44, 54, 56, 70. Although controlled, these studies have had some limitations. These include: (1) virtually all were retrospective, limiting the amount/quality of information available 73, 74. (2) In some, the number of SUDEP cases was relatively small 9, 27, 54. (3) SUDEP was not defined in one study 27. (4) Autopsies were not done in some cases classified as SUDEP/probable SUDEP 19, 54, 56, 70. (5) Some, by design, looked at only a limited number of risk factors 19, 70. (6) Some studies included or excluded epilepsy cases on the basis of AED prescription (thereby possibly missing some SUDEP cases) 9, 56. In our study, we compared suggested risk factors in SUDEPs with epileptics dying from other causes in a coronial population. Such an approach has strengths and limitations compared to previous controlled studies. The strengths are that every case met rigorous criteria for SUDEP, including autopsy, the number of cases relative to previous controlled studies 9, 19, 27, 44, 56, 70 was large and data on epilepsy and a wide range of potential risk factors was collected from treating doctors and family members shortly after death. The limitations relate to how representative the non-sudep deaths are of the general population of people with epilepsy. It is likely that this population has an over-representation of severe epilepsy and psychiatric problems. Although a representative living control group would be preferable, ascertainment issues and the difficulty in collecting clinical data in a uniform manner from deceased and living subjects made such an approach impossible. Our study was not primarily an epidemiological one but our estimate of SUDEP of approximately 1:780 epileptics per year is in keeping with SUDEP rates as estimated from population-based studies 7, It is lower than rates estimated by most cohort studies 15 21, 23 26, 55 probably because our population had fewer subjects with severe epilepsy. Our estimate may still be an underestimate as it is likely some SUDEPs were not referred to the coroner. Also, once cases have been reported to the coroner, some family members object to autopsy on potential SUDEP cases and these objections may be upheld by the coroner. Our finding that the SUDEP group was significantly younger than the control is in keeping with most previous studies that indicated young adults are particularly at risk of SUDEP 8, 11, 12, 14, 19, 23, 24, 29, 30, This finding may relate to the definition of SUDEP, viz. in elderly people it may not be possible to diagnose SUDEP due to the presence of anatomic disease that could be capable of causing death. This limitation of the definition of SUDEP, however, affects all SUDEP research. Our finding of the SUDEP group being younger than the controls was not accounted for by there being a greater number of children in the SUDEP group compared with the control (five SUDEPs less than 18 years of age vs. four controls). Our finding of a slight male predominance in the SUDEP group (27 males and 23 females) is consistent with a number of studies 3, 7, 8, 21, 23, 40, 42, 44 46, although some found no sex difference 6, 75. Our control group contained more males. This bias was only partly explained by the expected male predominance in trauma and suicide deaths and we were not able to satisfactorily, completely explain this finding. Our finding of long duration of epilepsy in the SUDEP group is in keeping with most previous studies 11, 12, 14, 24, 29, 46, 47, 50, 51, but there was no difference in duration between SUDEP and control groups. This finding of no difference in duration between SUDEP and control groups is in keeping with three previous studies 9, 27, 44, but contrasts with another which reported shorter duration of epilepsy in the SUDEP compared with the non-sudep group 19. Our finding that early onset epilepsy was not increased in the SUDEP group agreed with one 9,but differed from another two which found increased risk with early onset epilepsy compared with late onset epilepsy 44, 56. None of the SUDEP cases had absence or myoclonic seizures alone, and virtually all had generalised tonic clonic seizures in agreement with previous studies 8, 11, 23, 46, 47, 50, 54, although the distribution of seizure types was not different in our control group. Most studies suggest that SUDEP occurs more commonly in those with frequent seizures16, 18 21, 23 27, 46,

7 462 K. Opeskin & S. F. Berkovic 55, 56, but some like our study find no association between SUDEP and difficult to control seizure disorders/seizure frequency 9, 27, 44, 51. The lack of association with frequent seizures may relate to an over-representation of severe epilepsy in our controls, but it is noteworthy that 30% of our SUDEP cases had less than 2 seizures per year emphasising that SUDEP can occur in subjects with rare seizures. However, a subgroup of our SUDEP cases had very high seizure frequency. We also did not confirm the finding of some studies that polytherapy with AEDs13, 54, 56, 64 (which may be a surrogate for seizure frequency) was a risk factor for SUDEP. We did not confirm previous reports of association of SUDEP with psychotropic drug prescription13, 50, 56 and did not find an association with psychiatric illness. Although possible over-representation of subjects with severe psychiatric illness in our coronial control group compared with the general epileptic population could account for our finding that psychiatric illness/psychotropic drug prescription was not increased in the SUDEP group, it is notable that our finding regarding psychotropic drug prescription was in keeping with two other controlled studies 44, 54. We did not confirm previous reports of association of SUDEP with particular epilepsies8, 11, 23, 24, 46, 47, 50, 51, 54, with mental retardation 9, 11, 13, 18, 23, 24, 50, 54, recent stressful life event 46 or alcohol abuse8, 11 13, 50, 51, 67, with particular AEDs 27, 46, increasing number of AEDs prescribed 13, 54, 56, 64 or compliance8, 11, 12, 30, 48, 50 52, A lack of adequate controls and highly selected case series in previous studies probably account for these differences 11 13, 18, 23, 24, 46, 51, 66, 67. The profile of patients at risk for SUDEP are young people with epilepsy. They are most likely to die during sleep, and SUDEP, should it occur, is a seizurerelated event. Although a small proportion of SUDEP cases had very frequent seizures, our finding that seizure frequency was overall not greater than controls suggests other indices of epilepsy control or of other clinical characteristics may be more important. Such indices might be seizure duration, nocturnal seizures or peculiarities of the autonomic nervous system leading to a fatal cardio-respiratory response even to rare seizures. This, taken in conjunction with the finding that there was no increased risk associated with a particular AED in monotherapy or multiple AEDs suggests that attempts to better treat patients epilepsy with AEDs might decrease the risk of SUDEP. ACKNOWLEDGEMENTS This study was approved by the Victorian Institute of Forensic Medicine Ethics Committee and was supported by a grant from the Epilepsy Foundation of Victoria. We thank Professor S. Cordner for support, Ms F. Allen for statistical analysis, mortuary, histology, toxicology, library and word processing staff at the Victorian Institute of Forensic Medicine and the State Coroner for making his files available. REFERENCES 1. Alstrom, C. H. A study of epilepsy in its clinical, social and genetic aspects. Acta Psychiatrica Neurologica Scandinavica 1950; 53 (Suppl.): Cockerell, O. C., Johnson, A. L., Sander, J. W., Hart, Y. M., Goodridge, D. M. and Shorvon, S. D. Mortality from epilepsy: results from a prospective population based study. Lancet 1994; 344: Hauser, W. A., Annegers, J. F. and Elveback, L. R. Mortality in patients with epilepsy. Epilepsia 1980; 21: Hendriksen, B., Juul-Jensen, P. and Lund, M. The mortality of epilepsy. In: Proceedings of the International Congress of Life Assurance Medicine (Ed. R. D. C. Brackenridge). London, Pitman, 1970: pp Johnston, S. C., Horn, J. K., Valente, J. and Simon, R. P. The role of hypoventilation in a sheep model of epileptic sudden death. Annals of Neurology 1995; 37: Zielinski, J. J. Epilepsy and mortality rate and cause of death. Epilepsia 1974; 15: Derby, L. E., Tennis, P. and Jick, H. Sudden unexplained death among subjects with refractory epilepsy. Epilepsia 1996; 37: Ficker, D. M., So, E. L., Shen, W. K. et al. Population-based study of the incidence of sudden unexplained death in epilepsy. Neurology 1998; 51: Jick, S. S., Cole, T. B., Mesher, R. A., Tennis, P. and Jick, H. Sudden unexpected death in young persons with primary epilepsy. Pharmacoepidemiology Drug Safety 1992; 1: Langan, Y., Nolan, N. and Hutchinson, M. The incidence of sudden unexpected death in epilepsy, (SUDEP) in South Dublin and Wicklow. Seizure 1998; 7: Leestma, J. E., Walczak, T., Hughes, J. R., Kalelkar, M. B. and Teas, S. S. A prospective study on sudden unexpected death in epilepsy. Annals of Neurology 1989; 26: Leestma, J. E., Kalelkar, M. B., Teas, S. S., Jay, G. W. and Hughes, J. R. Sudden unexpected death associated with seizures: analysis of 66 cases. Epilepsia 1984; 25: Tennis, P., Cole, T. B., Annegers, J. F., Leestma, J. E., McNutt, M. and Rajput, A. Cohort study of incidence of sudden unexplained death in persons with seizure disorder treated with antiepileptic drugs in Saskatchewan, Canada. Epilepsia 1995; 36: Terrence, C. F., Wisotzkey, H. M. and Perper, J. A. Unexpected unexplained death in epileptic patients. Neurology 1975; 25: Annegers, J. F., Coan, S. P., Hauser, W. A. and Leestma, J. Epilepsy, vagal nerve stimulation by the NCP system, all-cause mortality, and sudden, unexpected, unexplained death. Epilepsia 2000; 41: Dasheiff, R. M. Sudden unexpected death in epilepsy: a series from an epilepsy surgery program and speculation on the relationship to sudden cardiac death. Journal of Clinical Neurophysiology 1991; 8: Hennessy, M. J., Langan, Y., Elwes, R. D. C., Binnie, C. D., Polkey, C. E. and Nashef, L. A study of mortality after temporal lobe epilepsy surgery. Neurology 1999; 53: Klenerman, P., Sander, J. W. and Shorvon, S. D. Mortality in patients with epilepsy: a study of patients in long term

8 Risk factors for SUDEP 463 residential care. Journal of Neurology, Neurosurgery, and Psychiatry 1993; 56: Leestma, J. E., Annegers, J. F., Brodie, M. J. et al. Sudden unexplained death in epilepsy: observations from a large clinical development program. Epilepsia 1997; 38: Leppik, I. E. Tiagabine: the safety landscape. Epilepsia 1995; 36: S10 S Lip, G. Y. H. and Brodie, M. J. Sudden death in epilepsy: an avoidable outcome? Journal of the Royal Society of Medicine 1992; 85: Medical Research Council Antiepileptic Drug Withdrawal Study Group. Randomised study of antiepileptic drug withdrawal in patients in remission. Lancet 1991; 337: Nashef, L., Fish, D. R., Sander, J. W. and Shorvon, S. D. Incidence of sudden unexpected death in an adult outpatient cohort with epilepsy at a tertiary referral centre. Journal of Neurology, Neurosurgery, and Psychiatry 1995; 58: Nashef, L., Fish, D. R., Garner, S., Sander, J. W. and Shorvon, S. D. Sudden death in epilepsy: a study of incidence in a young cohort with epilepsy and learning difficulty. Epilepsia 1995; 36: Physician s Desk Reference, 50th edn. Oradell, NJ, Medical Economics Company, Sperling, M. R., O Connor, M. J., Saykin, A. J. and Plummer, C. Temporal lobectomy for refractory epilepsy. Journal of the American Medical Association 1996; 276: Timmings, P. L. Sudden unexpected death in epilepsy: a local audit. Seizure 1993; 2: Langan, Y., Nashef, L. and Sander, J. W. Sudden unexpected death in epilepsy: a series of witnessed deaths. Journal of Neurology, Neurosurgery, and Psychiatry 2000; 68: Nashef, L., Garner, S., Sander, J. W., Fish, D. R. and Shorvon, S. D. Circumstances of death in sudden death in epilepsy: interviews of bereaved relatives. Journal of Neurology, Neurosurgery, and Psychiatry 1998; 64: Schwender, L. A. and Troncoso, J. C. Evaluation of sudden death in epilepsy. American Journal of Forensic Medicine and Pathology 1986; 7: Blumhardt, L. D., Smith, P. E. and Owen, L. Electrocardiographic accompaniments of temporal lobe epileptic seizures. Lancet 1986; 1: Gilchrist, J. M. Arrythmogenic seizures: diagnosis by simultaneous EEG/ECG recording. Neurology 1985; 35: Kiok, M. C., Terrence, C. F., Fromm, G. H. and Lavine, S. Sinus arrest in epilepsy. Neurology 1986; 36: Lathers, C. M. and Schraeder, P. L. Experience-based teaching of therapeutics and clinical pharmacology of antiepileptic drugs. Sudden unexplained death in epilepsy: do antiepileptic drugs have a role? Journal of Clinical Pharmacology 1995; 35: Lathers, C. M. and Schraeder, P. L. Arrythmias associated with epileptogenic activity elicited by penicillin In: Epilepsy and Sudden Death (Eds C. M. Lathers and P. L. Schraeder). New York, Marcel Dekker, 1990: pp Lathers, C. M., Schraeder, P. L. and Weiner, F. L. Synchronization of cardiac autonomic neural discharge with epileptogenic activity: the lockstep phenomenon. Electroencephalography and Clinical Neurophysiology 1987; 67: Lathers, C. M. and Schraeder, P. L. Autonomic dysfunction in epilepsy: characterisation of autonomic cardiac neural discharge associated with pentylenetetrazol-induced epileptogenic activity. Epilepsia 1982; 23: Nei, M., Ho, R. T. and Sperling, M. R. EKG abnormalities during partial seizures in refractory epilepsy. Epilepsia 2000; 41: Bird, J. M., Dembny, K. A. T., Sandeman, D. and Butler, S. Sudden unexplained death in epilepsy: an intracranially monitored case. Epilepsia 1997; 38: S52 S Blum, A. S., Ives, J. R., Goldberger, A. L. et al. Oxygen desaturations triggered by partial seizures: implications for cardiopulmonary instability in epilepsy. Epilepsia 2000; 41: Nashef, L., Walker, F., Allen, P., Sander, J. W., Shorvon, S. D. and Fish, D. R. Apnoea and bradycardia during epileptic seizures: relation to sudden death in epilepsy. Journal of Neurology, Neurosurgery, and Psychiatry 1996; 60: Munson, J. F. Death in epilepsy. Medical Record 1910; 77: Terrence, C. E., Rao, G. R. and Pepper, J. A. Neurogenic pulmonary edema in unexpected, unexplained death in epileptic patients. Neurology 1981; 25: Kloster, R. and Engelskjon, T. Sudden unexpected death in epilepsy (SUDEP): a clinical perspective and a search for risk factors. Journal of Neurology, Neurosurgery, and Psychiatry 1999; 67: Wilson, J. Hazards of epilepsy. British Medical Journal 1978; 2: Earnest, M. P., Thomas, G. E., Eden, R. A. and Hossack, K. F. The sudden unexplained death syndrome in epilepsy: demographic clinical and postmortem features. Epilepsia 1992; 33: Hirsch, C. S. and Martin, D. L. Unexpected death in young epileptics. Neurology 1971; 21: Jay, G. W. and Leestma, J. E. Sudden death in epilepsy: a comprehensive review of the literature and proposed mechanisms. Acta Neurologica Scandinavica 1981; 63: S1 S Krohn, W. Causes of death among epileptics. Epilepsia 1963; 4: Leestma, J. E. Sudden, unexpected death associated with seizures: a pathological review. In: Epilepsy and Sudden Death (Eds C. M. Lathers and P. L. Schraeder). New York, Marcel Dekker, 1990: pp Leestma, J. E., Teas, S. S., Hughes, J. R. and Kalelkar, M. B. Sudden epilepsy deaths and the forensic pathologist. American Journal of Forensic Medicine and Pathology 1985; 6: Lund, A. and Gormsen, H. The role of antiepileptics in sudden death in epilepsy. Acta Neurologica Scandinavica 1985; 72: Thorn, M. Neuropathologic findings in postmortem studies of sudden death in epilepsy. Epilepsia 1997; 38: S32 S Walczak, T. S., Leppik, I. E., D Amelio, M. et al. Incidence and risk factors in sudden unexpected death in epilepsy. A prospective cohort study. Neurology 2001; 56: Annegers, J. F., Coan, S. P., Hauser, W. A., Leestma, J. E., Duffel1, W. and Tarver, B. Epilepsy, vagal nerve stimulation by the NCP system, mortality, and sudden, unexpected, unexplained death. Epilepsia 1998; 39: Nilsson, L., Farahmand, B. Y., Persson, P., Thiblin, I. and Tomson, T. Risk factors for sudden unexpected death in epilepsy: a case control study. Lancet 1999; 353: Annegers, J. F., Hauser, W. A. and Shirts, S. B. Heart disease mortality and morbidity in patients with epilepsy. Epilepsia 1984; 25: Harvey, A. S., Nolan, T. and Carlin, J. B. Community-based study of mortality in children with epilepsy. Epilepsia 1993; 34: Hauser, W. A. and Kurland, L. T. The epidemiology of epilepsy in Rochester, Minnesota, 1935 through Epilepsia 1975; 16: Dasheiff, R. M. and Dickinson, L. J. Sudden unexpected death of epileptic patient due to cardiac arrhythmia after seizure. Archives of Neurology 1986; 43: Sillanpaa, M. Medico-social prognosis of children with epilepsy. Epidemiological study and analysis of 245 patients. Acta Paediatrica Scandinavica 1973; 237 (Suppl.): Kurokawa, T., Fung, K. C., Hanai, T. and Goya, N. Mortality and clinical features in cases of death among epileptic children. Brain Development 1982; 4:

9 464 K. Opeskin & S. F. Berkovic 63. Lathers, C. M., Kelliher, G. J., Roberts, J. and Beasley, A. B. Nonuniform cardiac sympathetic nerve discharge. Circulation 1978; 57: Leppik, I. E., Rarick, J. O. and Walczak, Te Sudden unexpected death in epilepsy and antiepileptic drugs (abstract). Epilepsia 1998; 39: S69 S Bowerman, D. L., Levisky, J. A., Urich, R. W. and Wittenberg, P. H. Premature deaths in persons with seizure disorders subtherapeutic levels of anticonvulsant drugs in postmortem blood specimens. Journal of Forensic Science 1978; 23: Denmark, L. N. Tuberous sclerosis presenting as a suspicious sudden death a case report and review of sudden deaths in epilepsy. Medicine Science and the Law 1992; 32: Luhdorf, K., Jensen, L. K. and Plesner, A. M. Epilepsy in the elderly: life expectancy and causes of death. Acta Neurologica Scandinavica 1987; 76: Neuspiel, D. R. and Kuller, L. H. Sudden and unexpected natural death in childhood and adolescence. Journal of the American Medical Association 1985; 254: Borel, J. Sudden death in epilepsy. Annals of Medicine and Psychology (Paris) 1966; 124: George, J. R. and Davis, G. G. Comparison of anti-epileptic drug levels in different cases of sudden death. Journal of Forensic Science 1998; 43: Nashef, L. Sudden unexpected death in epilepsy: terminology and definitions. Epilepsia 1997; 38: S6 S Commission on Classification and Terminology of the International League Against Epilepsy. Proposal for revised classification of epilepsies and epileptic syndromes. Epilepsia 1989; 30: Appleton, R. E. Sudden unexpected death in epilepsy in children. Seizure 1997; 6: Coyle, H. P., Baker-Brian, N. and Brown, S. W. Coroners autopsy reporting of sudden unexplained death in epilepsy (SUDEP) in the UK. Seizure 1994; 3: Lerman, P. Seizures induced or aggravated by anticonvulsants. Epilepsia 1986; 27: