Forensic antiepileptic drug levels in autopsy cases of epilepsy

Epilepsy & Behavior 22 (2011) 778–785

Contents lists available at SciVerse ScienceDirect

Epilepsy & Behavior journal homepage: www.elsevier.com/locate/yebeh

Forensic antiepileptic drug levels in autopsy cases of epilepsy☆ C.M. Lathers a,⁎, S.A. Koehler b, 1, C.H. Wecht b, 2, P.L. Schraeder c a b c

Center for Veterinary Medicine, U.S. Food and Drug Administration, Rockville, MD, USA Office of the Coroner, Allegheny County, Pittsburgh, PA, USA Department of Neurology, Drexel University College of Medicine, Philadelphia, PA, USA

a r t i c l e

i n f o

Article history: Received 20 June 2011 Revised 3 October 2011 Accepted 8 October 2011 Available online 15 November 2011 Keywords: Forensic antiepileptic drug levels Sudden unexpected death in epilepsy Epilepsy Retrospective study Death certificates/corner autopsies

a b s t r a c t A 1-year retrospective coroner-based forensic examination of causes of death among persons with a history of epilepsy was conducted at the Allegheny County Coroner's Office to evaluate the phenomenon of sudden unexplained/unexpected death in epilepsy (SUDEP), a diagnosis of exclusion. All cases at the Coroner's Office from January 1, 2001 through December 31, 2001, were examined. Review of a total of 1200 autopsied deaths revealed 12 cases with a past medical history of seizure disorder on the death certificate, which listed seizure disorder as the immediate cause of death or contributory cause of the death. Of the 7 men with seizure disorders, 5 were categorized as definite SUDEP and 2 as possible SUDEP. Of the 5 women with seizure disorders, 2 were listed as definite SUDEP, 2 as possible, and 1 as non-SUDEP because the convulsive seizures developed from a grade II glial tumor. Postmortem findings were evaluated for 11 cases; 1 body was decomposed. Toxicological screens were carried out on blood, bile, urine, and eye fluid for all 12. Antiepileptic drug (AED) levels detected in postmortem toxicological analysis were examined. AED levels were determined in 7 cases. Four of 7 had subtherapeutic AED levels, 2 had therapeutic levels, and only 1 victim of SUDEP had levels above the therapeutic range. Five cases had no detectable AED levels. AED levels at autopsy were either absent or subtherapeutic in 9 of 10 SUDEP cases, findings consistent with the likelihood of poor AED compliance. Subtherapeutic levels of AEDs may be a risk factor for SUDEP that could contribute to increased interictal and/or ictal epileptiform activity with associated autonomic dysfunction leading to disturbance of heart rate, heart rhythm, and/or blood pressure. © 2011 Elsevier Inc. All rights reserved.

1. Introduction Individuals with epilepsy are at risk for one form of unexpected sudden death referred to as sudden unexplained, unexpected death in persons with epilepsy (SUDEP). SUDEP has been defined as a sudden, unexpected, witnessed or unwitnessed death, nontraumatic and nondrowning, in patients with epilepsy, with or without evidence of a seizure, excluding documented status epilepticus, where postmortem examination does not reveal an anatomical or toxicological cause of

☆ This article was presented as four abstracts and four poster presentations by Dr. Lathers and Dr. Koehler at the FDA Science Forum 2003 in Washington, DC, and at the Annual Meeting of the American College of Clinical Pharmacology, Orlando, FL, USA, September 2003. Data have been published, in part: Koehler SA, Schraeder PL, Lathers CM, Wecht CH. One-year postmortem forensic analysis of deaths in persons with epilepsy. In: Lathers CM, Schraeder PL, Bungo MW, Leestma JE, editors. Sudden death in epilepsy: forensic and clinical issues. Boca Raton, FL: CRC Press, Francis & Taylor Group; 2010. p. 145–59. ⁎ Corresponding author at: 115 South Manning Boulevard, Albany, NY 12203, USA. Fax: + 1 518 482 6732. E-mail address: [email protected] (C.M. Lathers). 1 Current address: Epidemiology Department, Graduate School of Public Health, 130 DeSoto Street, Pittsburgh, PA 15261, USA. 2 Current address: 1191 Penn Avenue, Pittsburgh, PA 15222, USA. 1525-5050/$ – see front matter © 2011 Elsevier Inc. All rights reserved. doi:10.1016/j.yebeh.2011.10.011

death [1]. SUDEP occurs most frequently among young individuals with a history of generalized tonic–clonic seizures. It is important to note that SUDEP as a final diagnosis of the cause of death is usually not used on death certificates of persons with epilepsy who die with no cause being found other than the diagnosis of a seizure disorder [2,3]. For in-depth discussions of SUDEP, the reader is referred to the book Sudden Death in Epilepsy: Forensic and Clinical Issues [4]. The term SUDEP refers to the sudden death of an individual with a clinical history of epilepsy for whom a postmortem exaination fails to uncover a gross anatomic, toxicological, or environmental cause of death [5]. Evidence of terminal seizure activity may not be present. In 2002, Shields at al. [6] reported that 1–2% of natural deaths certified by medical legal death investigators in the United States are attributed to epilepsy and that increased microscopic examination of the brain postmortem has allowed identification of structural changes that are not grossly evident, but are, nonetheless, representative of the underlying substrates of epileptogenic foci. Overall, Shields et al. [6] examined 70 death cases, all with known clinical history of seizures, and classified them as: (1) individuals who lacked a gross brain lesion, (2) individuals with a brain lesion demonstrable at autopsy, (3) individuals for whom neuropathological evaluation could not be done because of decomposition, and (4) individuals for whom only an external examination was done [5].

C.M. Lathers et al. / Epilepsy & Behavior 22 (2011) 778–785

One of the many unanswered question surrounding SUDEP is what role antiepileptic drugs (AEDs) play in these deaths [7]. Various authors have addressed this question and varying data have been reported. Risk–benefit profiles must always be considered when prescribing any drug. Physicians must consider the efficacy, safety, and side effect profile of a given AED when determining which drug is best for a given patient [8,9]. The issue of a possible relationship between AEDs and SUDEP can be thought of in two ways. First, whether AEDs are a benefit or a risk relative to SUDEP raises the question of the role of AED blood levels in the control of seizures and/or in the prevention of SUDEP [10]. Second, whether noncompliance with AEDs is a risk factor in SUDEP raises the question of subtherapeutic or totally absent AED levels as a risk factor for SUDEP. The major question is whether there is a connection between the two factors. A review of the published literature addressing this issue revealed that a variety of conclusions have been reached by different authors looking at discrete data samples [11–18]. The goal of this forensically based study was to examine the AED levels of individuals who died from SUDEP. To answer this question we analyzed the type and number of prescribed AEDs and the number and types of these AEDs detected with postmortem blood tests in deaths of persons with a history of seizures that were investigated by the Allegheny County Coroner's office over a 1-year period. 2. Methods The Allegheny County Coroner's Office has forensic jurisdiction to investigate all deaths within Allegheny County, located in western Pennsylvania, and encompasses a population of approximately 1.2 million. The office investigates more than 6000 cases and conducts more than 1200 autopsies annually. All deaths investigated by the office from January 1, 2001 to December 31, 2001 were reviewed. Allegheny County Coroner's Office protocol defined how this study was conducted. All cases with a history of seizure disorder indicated in the past medical history or the word seizure or phrase seizure disorder listed in Part I or Part II of the death certificate were identified through a computer analysis and hand search of the case files by the forensic epidemiologist (S.A.K.). The information available for review included the data contained within the Death Investigation Report, the death certificate, and the available medical records. Inclusion criteria included all cases for whom seizure was listed in either Part 1—Immedate Cause of Death or Part 2—Conditions Contributing to the Death of the death certificate. We began by identifying all death certificates featuring the word seizure and worked backward. One hundred percent of the cases were listed as having a diagnosis of seizures. In none of the cases with epilepsy on the death certificate was the type of seizure differentiated. The final conclusion on “yes” or “no” for SUDEP was based on postmortem data. All seizure types would be grouped together under the category “seizures on the death certificate,” reinforcing one of the most important issues in determining SUDEP as a cause of death, namely, the incomplete nature of the history of type and frequency of seizures in the victims, including those subjected to autopsy. Although presence of intractable epilepsy and history of generalized tonic-clonic seizures have been described as the most important factors in SUDEP, information relative to this differentiation was not available on the death certificates. The following epidemiological information was collected: age, sex, race, time/date last seen alive, and time/date of death. Seizure-related information collected included past medical history, a list of prescribed medications including all AEDs, and witness(es) to the event. Pathological information collected was obtained from the forensic autopsy report. Toxicological analysis was conducted on blood, bile, urine, and eye fluid recovered during the autopsy. The blood used for the toxicological analysis was collected from the heart during the autopsy. Number and levels of drugs detected in the body fluids were obtained from the Toxicological Report. The toxicological

779

analysis included the number of drugs detected, their blood concentrations, and determination of these levels as therapeutic, subtherapeutic, above the therapeutic range, or absent. The data were entered and analyzed with the Statistical Package for the Social Sciences® software (Version 11.0, SPSS, Chicago, IL, USA). Toxicological data for all deaths with a diagnosis of epilepsy were summarized. On the basis of these data, cases that met the criteria for SUDEP were identified. 3. Results A total of 12 deaths were identified in which seizure disorder was listed in the medical record or was listed as either the immediate cause of death or a contributing cause of death during the study period. The epidemiological profile is summarized in Table 1. According to information contained within the Death Investigation Report, 7 men and 5 women, that is, all cases, had a past medical diagnosis of seizure disorder. Among the 12 cases 58.3% were male, 41.7% female, and 100% were Caucasian and between 38 and 54 years of age; mean age was slightly higher for men (45.4 years, range: 38–50) than for women (44.8 years, range: 39–54). Overall, more than 90% of the deaths were unwitnessed events, with the only witnessed event being that of a 42-year-old women who was in her residence when she had a sudden seizure (specific type is not documented) and become unresponsive. She was pronounced dead at the emergency room 1 hour after the seizure. Among the men, 4 were found in their bathroom and 3 in bed. All women were found in their residence (1 on the kitchen floor, 1 on the bed, 1 in a chair, and 1 outside on the rear deck). Seizure disorder was listed as the cause of death for 7 of 8 men in our study. The eighth man was listed as the victim of an accident. How did the accidental death occur? What happened to the victim? Were there seizures associated with a car accident, or a drowning, or a fall? Unfortunately, in this retrospective study there is no way to answer this question. We can only acknowledge the issue. Deaths were most frequent in January among the men and in September for the women. According to information contained within the Death Investigation Report, all 12 cases had a diagnosis of seizures or epilepsy and were considered to have a seizure disorder. A complete postmortem examination was conducted on all but one case because of the advanced level of decomposition, which precluded an accurate autopsy. Tables 2 and 3 summarize the immediate and contributory causes of death for men and women, respectively. The data indicate that all 12 cases had a past medical history of a diagnosis of either seizures or epilepsy. Although all cases would be considered as having a Table 1 Epidemiological profile.

Total number of cases (all Caucasians) Age range (mean) Seizure event witnessed Yes No Month of death and number of victims

Past medical history of seizure disorder or epilepsy Manner of death Natural Accidental Type of postmortem examination Complete External only a

Decomposed.

Men

Women

7 38–50 (45.4)

5 39–54 (44.8)

0 7 January—3 February—1 March—1 August—1 October—1 7

1 4 June—1 July—1 August—1 September—2 5

6 1

5

6 1a

5 0

780

C.M. Lathers et al. / Epilepsy & Behavior 22 (2011) 778–785

Table 2 Immediate and contributory causes of death for men. Case

Death certificate

Part I: Immediate Cause of Death

1 2 3 4 5

Definite Definite Definite Definite Definite

Seizure Seizure Seizure Seizure Seizure

6

Possible SUDEP

Seizure disorder

7

Possible SUDEP

Dilated cardiomyopathy with arteriosclerotic cardiovascular disease, acute pneumonia, and emphysema

SUDEP SUDEP SUDEP SUDEP SUDEP

disorder disorder disorder disorder disorder

Part II: Contributory Cause of Death

Liver cirrhosis Liver cirrhosis Chronic obstructive pulmonary disease Chronic obstructive pulmonary disease Seizure disorder

history of seizure disorder, there were insufficient data to determine the specific seizure type and frequency of occurrence. As mentioned earlier, a complete postmortem examination was conducted on all but one case, in whom decomposition was advanced. Eleven of the 12 cases were identified as SUDEP or possible SUDEP. Among the 7 men, one death was deemed to be possible SUDEP as the immediate cause of death was listed as dilated cardiomyopathy with arteriosclerotic cardiovascular disease, acute pneumonia, and emphysema and the contributory cause of death was seizure disorder. A second male victim was classified as possible SUDEP because putrefaction prevented accurate toxicological analysis. One of the 5 women examined was found to have a grade II glial brain tumor. Whether or not the tumor had enough mass to result in death, which the autopsy data did not indicate to be the case, the tumor would most likely have been the etiology of the seizures, and we cannot determine if the death was from SUDEP. Nevertheless, this case has been classified as non-SUDEP on the basis that there was an underlying progressive pathological entity. Thus, of the possible 12 SUDEP cases identified initially, 11, including 7 men and 4 women, were evaluated for definite or possible SUDEP. Past medical history, prescribed medications, drugs identified by the toxicology screen, postmortem levels, and pathological factors for the men are summarized in Table 4. All 7 had a past history of seizures or epilepsy. Five (71%) had been prescribed AEDs (clonazepam, phenytoin). The death scene investigation failed to ascertain a list of prescription medications in two deaths. Toxicological analysis revealed that AEDs were detected in 57% of the cases, with phenytoin the most frequent. All detected AEDs were at subtherapeutic levels. In only one case was the toxicological analysis negative. In 3 cases, the bodies were in varying stages of decomposition. Past medical history, prescribed medications, drugs identified by the toxicology screen, postmortem levels of compounds, and pathological factors for the women are summarized in Table 5. All cases had a past medical history of seizures. Two women (40%) had been

Table 3 Immediate and contributory causes of death for women. Case

Death certificate

Part I: Immediate Cause of Death

Part II: Contributory Cause of Death

1 2 3

Definite SUDEP Definite SUDEP Possible SUDEP

— — —

4

Possible SUDEP

5

Non-SUDEP

Seizure disorder Seizure disorder Seizure disorder with arteriosclerotic cardiovascular disease Seizure disorder with hypertensive and arteriosclerotic cardiovascular disease Convulsive seizure developed from grade II glial tumor

—

—

prescribed medications for seizures (Dilantin, lamotrigine, Tegretol). The death scene investigation failed to ascertain a list of prescription medications in one death. Toxicological analysis revealed that AEDs were detected in 60% of the cases, with phenytoin the most frequent. Toxicological analysis showed that the AEDs were at therapeutic levels in 3, subtherapeutic levels in 1, and above the therapeutic level in 1. Even though all the female cases had a diagnosis of seizure disorder, no information was collected as to why three of them were not taking any AEDs. This incomplete approach to gathering postmortem data must be emphasized as a deficiency in data collection in seizure-related deaths even in highly professional departments led by medical examiners. This category of information, that is, why a person with a diagnosis of epilepsy is not on medication, should be a specific issue addressed in future studies Application of the SUDEP classification of Leestma et al. [20] revealed that of the 7 men with seizure disorders, 5 were categorized as definite SUDEP and 2 as possible SUDEP. In case 6 (Table 2), the degree of decomposition of the remains made it impossible to collect accurate autopsy data. In case 7 (male), the presence of atherosclerotic heart disease could have been a cause of death if coronary artery occlusion resulted in a fatal myocardial infarction. The mitigating factor in determination of cause of death is the issue of dilated cardiomyopathy that in and of itself could have resulted in potentially fatal heart failure or a potentially fatal cardiac arrhythmia. Of the 5 women with seizure disorders, 2 had findings of arteriosclerotic heart disease with no description of coronary artery occlusion or myocardial infarction, in which case the most likely cause of death could be SUDEP. As noted earlier, for the woman with a grade II glial tumor, there were no autopsy data indicating that the tumor had resulted in pathological changes that could cause death. It is more likely that the glioma was the cause of a seizure disorder rather than the direct cause of death. It should be noted that 2 cases, one man (case 6 with no reported AEDs) and one woman (case 3), were discovered in a stage of putrefaction. Putrefaction is a late stage of postmortem decomposition resulting from the anaerobic action of splitting of proteins by bacteria with the formation of foulsmelling, incompletely oxidized products. Putrefaction limits the amount of blood available for analysis and produces by-products that can affect the toxicological analysis. Therefore, data regarding the types of drugs including AEDs and/or their levels may not be accurate. Both of these cases were listed as possible SUDEP. In summary, in our study, for the male and female cases for whom levels were determined, most AED levels were found to be either subtherapeutic or absent, findings in keeping with most of the literature. 4. Discussion The presence of intractable epilepsy and a history of generalized tonic-clonic seizures have been described as the most important risk factors in SUDEP. Unfortunately, because of the retrospective nature of our study, these data were not available. Likewise, one must ask if there was information about the types of seizures and localization of seizure onset in the SUDEP cases described in this study, as type of seizure is a factor that has been associated with SUDEP and deserves description in any study. However, the reader must remember that the Allegheny County Coroner's Office is not a clinical facility, but rather a forensic facility, which explains why the cause of death is classified as “seizures” rather than the specific type of seizure. Forensic pathologists at the Allegheny County Coroner's Office are interested in answering two basic questions: What was the cause of death? Was the cause of death due to natural or nonnatural causes (accident, suicide, or homicide)? Pathologists and/or medical examiners are generally not concerned about the types of epilepsy that may have been involved in the death of a person as their role is not to differentiate to this level of specificity. In the future, for anyone designing a prospective study, these weaknesses of a retrospective study design can be avoided by designing the new study to ask whether intractable

C.M. Lathers et al. / Epilepsy & Behavior 22 (2011) 778–785

781

Table 4 Past medical history, prescribed medication, toxicology screen, and pathological features for men. Past medical history

Prescribed medication

Drugs identified in toxicology screen and level

Postmortem level

Epidemiological pathological features

Seizures

Dilantin

Acetone 6 mg% Alcohol 5 mg%

— —

Seizures

Prinvil Clonazepam K-Dur Oramorph Depo Medrol Proventil inhaler

Doxylamine 0.014 mg%

Above therapeutic

Benzodiazepines Nordiazepam 0.01 mg% Chlordiazepoxide 0.013 mg% Demoxepam 0.023 mg% Oxazepam 0.010 mg% Clonazepam—too low to quantitative Morphine 0.087 μg/mL Dextromethorphan—positive Ibuprofen—positive Diphenhydramine—positive Phenytoin 8.31 μg/mL

Therapeutic Subtherapeutic — Subtherapeutic — Subtherapeutic — — —

46/male/white Alive: 1/7/01 Found dead: 1/11/01, 12:30 PM Body moderately decomposed 49/male/white Alive: 1/17/01, 4:30 PM Dead: 1/18/01, 12:30 PM

Seizures

Dilantin

Deep vein thrombosis Seizures Alcoholism Seizures Alcoholism

Dilantin Keflex

None detected

—

Celexa Neurontin Phenytoin Indomethacin Unknown

Ethanol 0.04% Phenytoin 1.05 μg/mL

— Subtherapeutic

49/male/white Alive: 1/18/01, 7:30 AM Dead: 1/18/01, 4:50 PM 50/male/white Alive: 2/11/01, 1:00 AM Found Dead: 2/11/01, 8:00 PM 43/male/white Decomposed

Citalopram—positive Desmethylcitalopram—positive

— —

43/male/white Putrefaction

Ethanol 0.02% Phenytoin 2.45 μg/mL Olanzapine 0.041 mg% Sertraline 0.021 mg% Valproic acid (TDX) 8.99 μg/mL

— Subtherapeutic Above therapeutic Therapeutic Subtherapeutic

38/male/white Alive: 10/16/01, 11:00 PM Found dead: 10/16/01, 12:00 PM

HIV Seizures Chronic lung disease Bipolar depression Epilepsy Depression

Unknown

epilepsy and a history of generalized tonic-clonic seizures were problems in any given SUDEP case and identifying the types of seizures the patient had and the localization of the seizure onset. Additionally, there is a need to educate the staff in the Coroner's or Medical Examiner's Office, as per the discussion by Schraeder et al. [2,3], on the required approach to seizure-related deaths. With respect to the definition of epilepsy or seizures in this study, we use in the tables seizures and epilepsy as major terms. One could

Subtherapeutic

ask if perhaps some patients did not have a seizure disorder and only had single seizures? To address this aspect, it should be noted that the medical examiners did a detailed retrospective review of the medical history when possible. Additionally, what we do know for each victim is: 1. The word seizure was listed on the death certificate, in either Part I or Part II.

Table 5 Past medical history, prescribed medication, toxicology screen, and pathological features for women. Past medical history

Prescribed medications

Drugs identified in toxicology screen

Postmortem level

Epidemiological pathological features

Hysterectomy Seizures

Dilantin Lamotrigine

Seizures

Neurontin Tegretol Dilantin

Phenytoin 5.36 μg/mL Lamotrigine 0.79 mg% Ibuprofen 0.81 mg% Clonazepram 8.37 μg/mL

Subtherapeutic Therapeutic Subtherapeutic Above therapeutic

Amyotrophic lateral sclerosis Depression Seizures

Bupropion Ditropanxl Zanaflex Baclofen Rilutek Unknown

Bupropion 0.006 mg% Threoaminobupropion—positive

Therapeutic —

41/female/white Alive: 6/7/01, evening Found dead: 6/8/01, 3:30 PM 42/female/white Alive: 7/6/01, 12:20 AM Dead: 7/6/01, 12:21 AM Witnesses seizure 39/female/white Alive: unknown Found dead: 8/7/01, 5:37 PM Moderate putrefaction

Trazodone (0.296 mg%) Venlfaxine (0.175 mg%) O-Desmethylvenlafaxine 0.025 mg%

Therapeutic Therapeutic —

48/female/white Alive: 9/28/01, 11:30 PM Found dead: 9/29/01, 9:31 AM

Carbamazepine 4.56 μg/mL Phenytoin 14.68 μg/mL Butalbital 0.246 mg% Acetaminophen 14.05 mg%

Therapeutic Therapeutic Subtherapeutic Nearly toxic

54/female/white Alive: 9/4/01, 2:30 PM Found dead: 9/5/01, 11:32 PM

Diabetes Seizures Hypertension Asthma Hypertension Seizures TMJ Osteoporosis Cervical fusion Psych history

Prempro Gybutynin Butalbital/APAP/caffeine tabs

782

C.M. Lathers et al. / Epilepsy & Behavior 22 (2011) 778–785

2. All 12 cases had a past medical history of a diagnosis of either seizures or epilepsy. 3. All cases were considered as having a history of seizures, although the frequency and type were unknown. 4. Each victim had a history of at least one seizure. 5. Most victims were taking AEDS and postmortem AED drug levels were determined. Despite the weaknesses of a retrospective study, the data in our article are useful. To date very few studies have reported postmortem AED levels, and thus, our study adds information to the database for SUDEP. Retrospective data are useful but clearly not in any way complete. Sudden unexpected death in epilepsy is the death of a person with epilepsy that is sudden and unexpected, with no obvious circumstances other than a history of a seizure disorder, without the premorbid occurrence of status epilepticus, and with no cause of death being found on postmortem structural and toxicological examination [19,20]. The average victim, more commonly male, is a young adult in the prime of life with a mean age in the early thirties. As discussed by Leestma [21], sudden and unexpected deaths generally occur outside hospitals, although victims may be brought to an emergency room. Such deaths are usually not attended by a physician and usually are without detailed historical or medical information. They will usually be brought to the attention of a medical examiner or coroner, who is responsible for determining the cause and manner of death and for generating a death certificate before the remains are interred or otherwise disposed of [21]. These responsibilities are statutory in most jurisdictions, and how they are discharged varies widely [22]. When a coroner is the responsible party, this individual most often is elected and may or may not be a medical doctor. A local pathologist (typically forensically trained) will usually be employed to conduct an autopsy or other examination of the body to make the determination of cause and manner of death. When a medical examiner is responsible, this individual is almost always a trained forensic pathologist and will employ a staff of forensic pathologists to perform the required examination to determine the medical cause of the death, which can be labeled as a homicide (death at the hands of another), suicide (death by one's own hand), accident, natural (natural disease processes), or undetermined. The coroner or medical examiner may direct that an autopsy be performed and that other studies that may include toxicological examinations of tissues or body fluids be undertaken [22,23]. One must inquire if the death certificates are signed by any doctor or by a family physician who knew the patient, as the patient's physician may be biased in favor of declaring epilepsy as a cause of the death either to support or discount his premortem diagnosis. It depends on the presentation of the case. The majority of death certificates involving a natural death are signed by the family physician. If so, the family doctor would use a diagnosis that favors his medical thinking. If, however, the family doctor refuses to sign the death certificate or if there are any signs of trauma or suspicion of foul play, the death comes under the jurisdiction of the Coroner's or Medical Examiner's office. How often does the family doctor confirm his predeath diagnosis on the death certificate? Usually the attending doctor at the hospital will complete the death certificate and send it to the family doctor to sign. This procedure is also used if the victim is transported to the hospital where he or she dies. In the United States, the general physician is not permitted to sign the death certificate if there are signs of trauma or suspected foul play. Here again is an example of the institutional deficit in how SUDEP is documented. This is, it seems, a universal Medical Examiner's/Coroner's Office deficiency. Which cases are labeled with the designation SUDEP can be difficult and sometimes controversial [21]. At the behest of the then Burroughs Welcome Pharmaceutical Company, an attempt was made to provide some guidelines on the basis of data collected during the

development of lamotrigine [20]. In the clinical trial, which included databases that drew on studies in many countries with 4700 patients (5747 patient-years of exposure) over several years, 45 sudden and/ or unexpected deaths were discovered in all arms of the study. It is important to evaluate, if possible, if the incidence of SUDEP differed between the treated and control groups. Data were classified according to the following classification [20]. Definite SUDEP was defined as occurring when the dead person had suffered from epilepsy as defined by Gaustaut [24] and Gaustaut and Zifkin [25] and had been treated with one or more anticonvulsant agents, usually for many years, and died unexpectedly in a reasonable state of health. The fatal attack occurred during normal activities in benign circumstances with no obvious medical cause of death found after autopsy. If status epilepticus or acute neurotrauma occurred during the seizure, the case was excluded. Definite SUDEP depends on the performance of an autopsy in which other causes of death, except the history of epilepsy, are excluded. If all clinical evidence supports no other cause of death, but no autopsy is performed, the death is categorized as probable SUDEP. In cases that meet most or all of the SUDEP criteria, but the data suggest more than one possible cause of death associated with seizures, such as death while bathing or swimming or as a result of aspiration, with or without an observed seizure, the death is categorized as possible SUDEP. There have been and are inevitable issues associated with the SUDEP classifications, owing to all the information problems that are inherent to any patient-based study. Other confounding variables include the role of concurrent diseases in the victim, especially heart disease in the older victim, as well as conditions related to the common prevalence of alcohol and drug abuse. It is very important that a complete death scene investigation be conducted for any individual who apparently died suddenly and/or unexpectedly. The physical death scene environment must be documented by careful inspection of the body position or location to determine if normal activity was ongoing and was interrupted by the fatal attack. The presence of bruises on the face, head, and extremities does not necessarily mean that the victim was assaulted. Rather, if a seizure did occur, the victim may have fallen and been injured or become injured during the clonic phase of the seizure. Context and common sense are required in evaluating the scene. Furthermore, investigators should search for and take any pill bottles that are present and/or any materials that suggest illicit drug use. Although this postmortem methodology is ideal, it is not yet done routinely, as exemplified by the postmortem data obtained for the present study. Leestma's statements [21] about the inevitable issues with SUDEP classifications owing to all the information problems inherent to any patient-based study, including confounding variables such as the role of concurrent disease in the victim, are illustrated in the cases examined in our study. Our inclusion of the case with chronic lung disease may be questioned, as it could be argued that SUDEP was not really the cause of death and seizure was only a comorbidity. At this point, the authors remind readers that if the word seizure was listed in either Part I—Cause of Death or Part II—Contributing Cause of Death of the death certificate, but was not related to the actual death event, then the given patient would have been examined in our study. For example, smoking may be a contributory factor in death, but smoking the last cigarette does not in itself kill the patient. However, smoking may contribute to the overall cause of death by resulting in chronic lung disease, albeit not severe enough to be listed as the primary cause. We could not simply ignore these two male cases as the term seizure disorder appeared on Part I of the death certificate; that is, the primary cause of death was seizure disorder. There is an argument to be made that these cases be classified as SUDEP comorbid with lung disease. Thus, the causes of SUDEP and comorbidity must be considered. If a given patient has a diagnosis of seizures and has other risk factors for death, such as smoking, epilepsy severity, seizure type, presence

C.M. Lathers et al. / Epilepsy & Behavior 22 (2011) 778–785

of cardiac arrhythmias, age, gender, sleep, sleep position, supervision during sleep, and genetics [26], whether the given combination of risk factors increases the probability or likelihood of SUDEP must be evaluated. At least three categories of risk factors are operative in mechanisms underlying SUDEP [8,27]: arrhythmogenic (autonomic neural and cardiac function), respiratory and hypoxic, and psychological (stress, anxiety, depression). Arrhythmogenic risk includes the subcategories pharmacological, ionic channelopathies, and acquired heart disease. Thus, a comorbidity can increase the risk of death from a seizure disorder and yet not be the actual cause of death. Likewise, one may ask why the patient with amyotrophic lateral sclerosis who did not take any anti-seizure medication was included in our study. The answer to this question again lies in our inclusion criteria. The inclusion criteria were based on the word seizure being listed on the death certificate, in either Part I or Part II, and that is why this patient is included in our study. The question then becomes: Does the inclusion of patients with chronic lung disease or amyotrophic lateral sclerosis modify the incidence of SUDEP? However, as there is no indication in the postmortem data of respiratory failure as a cause of death, SUDEP would be a reasonable explanation for the death. Would the incidence be different if we deleted these two patients? We can certainly calculate the incidence with and without these patients. Inclusion of patients with only seizures and not lung disease or amyotrophic lateral sclerosis will not make any difference in the incidence of SUDEP. It does not matter if the values are 7 or 11 in 1200 or 10 in 1200 or 12 in 1200 when considering the incidence of SUDEP. A much larger study population with more seizure disorder cases is needed to determine the true incidence of SUDEP. This discussion of the incidence of SUDEP and which patients to include in the definition of SUDEP raises again the need to educate staff in coroner and medical examiner offices about the importance of obtaining such information in deaths of persons with a history of epilepsy [2,3]. A larger study population with more seizure disorders was accessed in the nonforensic Centers for Disease Control and Prevention mortality database [28]. The 2001 prevalence rates of seizures in the 1.23 million population of Allegheny County, Pennsylvania, are listed in Table 6. The putative incidence of SUDEP obtained in our study is for the forensically examined deaths and not for the general population. It should be noted that CDC would code the deaths of both types of victims as “seizure” deaths. In the present study we did search both Part I and Part II of the death certificate for the word seizure. In establishing the true prevalence of the phenomenon of SUDEP, it is important to look for autopsy evidence of the concurrent roles of the brain and the heart in the cause of death of a person with epilepsy who has died suddenly [29]. In addition to the need for autopsy to clarify the role of changes in the heart in SUDEP, there is a need for clinical studies to focus on the contribution of cardiac autonomic dysfunction and/or AEDs, with or without other drugs, in the development of cardiac abnormalities, to gain an understanding of the mechanism of death in persons with epilepsy [30], as disturbance of the autonomic nervous system may be a contributory cause [8]. Exactly how disruption of autonomic function contributes to the risk of death is unknown. The occurrence of seizures is often associated with minimal changes in cardiac conduction and rhythm, myocardial Table 6 Definite or possible SUDEP or other non-SUDEP causes of death. Total number/ Definite SUDEP gender as in Tables 2 and 3

Possible SUDEP

7/males

Case 6 (not on AEDs) 0 Case 7 (on AEDs) Cases 3, 4 Case 5: convulsive seizure no AEDs due to glial brain tumor, AEDs found postmortem

5/females

Cases 1–5 all on AEDs Cases 1, 2 all on AEDs

Other non-SUDEP cause of death

783

infarction, and stunned myocardium following brief convulsive seizures [31,32]. Natelson et al. [33] found irreversible pathological cardiac changes, that is, subendocardial perivascular and interstitial fibrosis, in hearts of persons with epilepsy who died suddenly and unexpectedly. Again, these unanswered questions pertinent to the role of changes in the hearts of SUDEP victims emphasize the need to educate the coroner and medical examiner office staff about the importance of obtaining such information in deaths of persons with a history of epilepsy. It must also be considered that a patient may not have epilepsy but yet die from hypoxia or respiratory apnea or from cardiac arrhythmias during which a witness observing the agonal death event may note visible seizure-like activity resulting from the lack of oxygen in the brain [34,35]. SUDEP may also be misdiagnosed. However, there was enough medical history in the medical records of the cases included in our study to eliminate agonal seizures associated with death. The role AED levels play in SUDEP can be most readily investigated by using retrospective data collected by medical examiner or coroner offices, including a detailed past medical history that includes a list of current medications, the events and/or activities just prior to death, descriptions and weights of the internal organs, and extensive toxicological analysis of the body fluids. In the future, prospective studies using coroner and medical examiner office facilities in conjunction with more detailed questions modeled on the concept of the oral autopsy technique should provide more comprehensive historical information about issues such as compliance with AED prescriptions including premorbid AED levels, details of seizure types and patterns of seizure occurrence, history of any premorbid cardiac rhythm issues, and family history of epilepsy and cardiac arrhythmias [8]. Relative to the role of AED levels in SUDEP is the question of whether ongoing monitoring of AED levels is beneficial to the patient. Jannuzzi et al. [36] addressed this question and reported there was no difference in treatment outcome between patients whose AED levels are regularly monitored and those treated only clinically. Nonetheless, they did find some benefit for AED monitoring, as 25% of the unmonitored group had subtherapeutic levels at the end of the study, whereas only 8% of those who were regularly monitored had subtherapeutic levels. Rates of adverse side effects were the same for patients whose AED levels were monitored and those only followed clinically. In contrast, Nilsson et al. [17] reported that routine therapeutic drug monitoring was associated with a lower risk of SUDEP (risk ratio [RR] = 9.5) when compared with no regular monitoring (RR = 3.7). The RR was 9.5 if carbamazepine levels were above the common target range of 20–40 μM. In those patients on carbamazepine, high levels were associated with polypharmacy, and were also associated with frequent changes in the dose prescribed. No association was found for the variation in AED levels in those individuals on phenytoin. Thus, although polypharmacy, frequent dose changes, especially with carbamazepine, and high carbamazepine levels may be associated SUDEP risk factors, these findings may merely reflect the severity of the seizure disorder. Compliance with AEDs has received some attention as a potential risk factor associated with SUDEP, resulting in two key questions: What is the role of AED levels in SUDEP? Is AED noncompliance a factor in SUDEP? Although Opsekin and Berkovic [37] found no difference between postmortem AED levels of SUDEP victims and AED levels of control cases and concluded that noncompliance is not a risk factor for SUDEP, other studies support the supposition that noncompliance is a risk factor for SUDEP. Lund and Gormsen [38] found subtherapeutic levels in half of the cases of SUDEP, suggesting that noncompliance may be a risk factor for SUDEP. They also found that some SUDEP victims had toxic concentrations of AEDs. One-third of the toxic AED concentrations were found in patients taking phenobarbital. In contrast, May et al. [39] reported that postmortem AED levels were lower than premortem AED levels for the same patients, particularly for phenobarbital and phenytoin and less so for

784

C.M. Lathers et al. / Epilepsy & Behavior 22 (2011) 778–785

carbamazepine. These authors also raised the question of whether or not low postmortem AED levels actually reflect noncompliance. George and Davis [14] studied 52 deaths from SUDEP, 8 epilepsyrelated accidental deaths, and 44 non-epilepsy-related deaths in persons with epilepsy. Subtherapeutic levels were found in 69% of the SUDEP deaths, 75% of the seizure-related accidental deaths, and 34% of the non-epilepsy-related deaths. They [14] concluded that subtherapeutic AED levels are a factor in epilepsy-related deaths other than SUDEP deaths, a premise that seems to hold across most such studies. As with all drug assays, one must consider the method used and the lower level of quantification of the assay, that is, how low a level of AED can be detected? Tomson et al. [40] compared premortem and 72-hour postmortem AED levels in rabbits. They found that carbamazepine levels remained stable, but that premortem phenytoin levels were only 35% of the immediate premortem levels. Thus, the accuracy of postmortem levels obtained must be questioned when attempting to determine if poor compliance was a factor in the SUDEP of persons taking phenytoin. This concern would not be operative in the case of carbamazepine as the assay suggested that levels of this AED remained stable. Nonetheless, the total absence of phenytoin, in contrast to subtherapeutic levels, would be consistent with noncompliance. In this study all blood samples and AED levels were obtained directly from heart chamber blood, which typically contains higher concentrations of drugs compared with peripheral blood. Therefore, individuals with subtherapeutic heart blood levels have even lower circulating blood levels. Williams et al. [41,42] studied AED levels in 1-cm hair segments. They determined coefficients of variability for SUDEP and nonSUDEP deaths and for outpatients and inpatients. They found that this coefficient was 20.9% in SUDEP and 15% in non-SUDEP deaths, and 9.6% for outpatients and 6.2% for inpatients. Thus, variability in ingestion of AEDs over time was found to be greater in SUDEP, indicating that noncompliance is a risk factor for SUDEP. Salmo and Connolly [43] studied autopsied sudden death cases in the west of Ireland. Of 103 sudden death cases examined, 22 were deemed to be SUDEP, with a mean age of 38 years. Among the SUDEP cases, 68% were male, 68% had absent or low AED levels, 88% had a history of generalized tonic–clonic seizures, 86% had pulmonary edema, and 45% were found dead in bed. Tennis et al. [44] identified, among 3688 subjects taking AEDs, 153 of 163 deaths that underwent autopsy. Of the 153 deaths, 18 were found to be definite SUDEP and 21 were probable SUDEP, yielding a minimum of 0.54 and a maximum of 1.35 SUDEP per 1000 patientyears. The incidence of SUDEP increased with male gender, number of AEDs ever prescribed, and prescription of psychotropic drugs. SUDEP incidence was highest in males with a history of treatment with three or more AEDs and four or more psychotropic drugs. Poisson regression showed a 1.7-fold increase in the risk of SUDEP for each increment in the maximum number of AEDs given, suggested to be a likely surrogate for the severity and persistence of seizures in a given patient. Opsekin and Berkovic [37] identified 50 SUDEP cases and 50 nonSUDEP-related cases of death among persons with epilepsy. Detailed clinical data were obtained from doctors, family members, and coroner files. The SUDEP group was characterized by younger age, greater likeliness of being found dead in bed, evidence of terminal seizures, and no increased risk with any particular AED in monotherapy or multiple AEDs. This suggests that aggressive attempts to control epileptic seizures might decrease the risk of SUDEP. However, no correlation with seizure frequency was found, suggesting other indices are more important. To postulate that poor adherence is a cause or possible cause of SUDEP one must have evidence from previous clinical notes indicating that there was poor AED compliance/adherence. This is important as the literature indicates that postmortem AED drug levels are

important but should be treated with caution. In our study we did, when possible, go back to the victim's primary care doctor to ask questions about compliance. AED data were based on the information obtained from hospital physicians or family physicians and/or family members of the deceased and/or the medical records on site at Allegheny County Coroner's Office. We can state the following about our study: 1. There was an extensive toxicological evaluation of the AEDs or other drugs found in the patients. 2. Doses of the AEDs prescribed were obtained if possible. 3. Seizure disorder diagnoses were obtained if possible. The information was subpoenaed from the hospital where the victim died or we collected the pill bottles at the home of the victim. The pill bottle contains the name of the physician. Sometimes we called the doctor if necessary. If the patient died in the hospital, we obtained the medical history from the next of kin. We did not analyze our data to determine on what date epilepsy was first diagnosed or whether the patient exhibited good or poor compliance with the AEDs prescribed. The preponderance of published data cited above suggests that patient noncompliance is a risk factor for SUDEP. In this study of 12 cases of SUDEP, we were able to document that the majority were prescribed AEDs. However, in only 2 cases were AED levels within the therapeutic range, and in 1 case, AED levels significantly exceeded the recommended therapeutic range One might speculate that AEDrelated cardiac toxicity was involved in the latter case of SUDEP. The lack of AEDs in postmortem blood samples for most SUDEP cases in our study supports the idea that noncompliance is a risk factor for SUDEP and may result in an increased chance of epileptiformrelated autonomic dysfunction in heart rate, heart rhythm, and/or blood pressure. The data also support the observation that patient compliance is important in the prevention of SUDEP and suggest that noncompliance could be a risk factor for SUDEP. Nonetheless, it is not known whether low or absent postmortem AED levels in these individuals were the result of noncompliance or unreliable methods of determining AEDs in postmortem blood. The reliability of the current assay method, the stability of AEDs after death, and the redistribution of drugs within the body need to be addressed. Terrence et al. [34] found that the AED levels in postmortem samples were below the lower limit of quantification of the assay method. Use of postmortem blood to estimate the person's circulating blood levels while alive is confounded by the phenomenon of postmortem redistribution and is a serious concern when evaluating information gleaned via use of forensic toxicology techniques. In our study the analyses were conducted on blood collected from the heart chambers. This blood typically contains higher concentrations of drugs then does peripheral blood. Therefore, AED levels would have been lower in the circulating blood, suggesting that individuals found to have subtherapeutic heart blood levels on death had even lower circulating blood levels while alive. Relative to changes in AEDs as a risk factor for SUDEP, there was no evidence of frequent drug changes. Our only source of information on AED changes was the medical record. We assumed no recent changes had occurred in the types or doses of AEDS prescribed in our study. In addition, our study cannot and did not address possible changes from brand to generic drugs or from generic to brand drugs or from one generic to another generic drug. The increased lung weight observed in our study has also been reported to be present in SUDEP by other investigators. Noncompliance with anticonvulsants has been suggested as a risk factor for neurogenic pulmonary edema, and the finding of pulmonary edema is thought to be associated with an adrenergic component in the death event [9,45,46]. Future studies need to have more detailed data collection, for example, seizure types, premorbid AED levels, general level of compliance, and stressful circumstances preceding

C.M. Lathers et al. / Epilepsy & Behavior 22 (2011) 778–785

death. In essence, the detailed information could be obtained with oral autopsy data [10,47]. Finally, the importance of prospective data collection must be emphasized. Those investigating a death possibly related to a history of seizures should obtain a detailed description of the events leading up to the death from the next of kin and review all medical records and a complete list of prescription medications. A global effort is needed to identify risk factors for SUDEP and to develop preventive measures [48]. In summary, the role of AEDs in causation and/or prevention of SUDEP is not clear. Each patient, each AED, and each combination of AEDs need to be examined to best individualize benefit/risk ratios. Drug therapies may increase or decrease risk for SUDEP [16,49]. The advantage of newer AEDs is less likelihood of sedation, which may minimize noncompliance. Many newer AEDs also have less frequent drug interactions, leading to improved tolerability. Compliance with AEDs and maintenance of stable therapeutic drug levels are crucial to avoid SUDEP. Although noncompliance with AEDs may be one important [46] risk factor for SUDEP [10], it is not the only risk factor to be addressed for persons at risk for SUDEP. Clinical pharmacology issues requiring evaluation postmortem include AED dose, actual AED or combination, possible drug–drug interactions, level of quantification of the assay method used postmortem, and recent changes in the AED or AED dose [50]. Acknowledgments Opinions expressed are those of the authors and do not reflect opinions or policy of the FDA. The authors thank Shaun Ladham, Leon Rozin, Abdulrezak Shakir, and Joseph Dominick for technical support in the data collection process. References [1] Nashef L, Shorvon SD. Mortality in epilepsy. Epilepsy 1997;38:1059–61. [2] Schraeder PL, Delin LK, McClelland RL, So EL. Coroner and medical examiner documentation of sudden unexplained death in epilepsy. Epilepsy Res 2006;68: 137–43. [3] Schraeder PL, Delin LK, McClelland RL, So EL. A nationwide survey of the extent of autopsy examinations in sudden unexplained death in epilepsy. Am J Forensic Med Pathol 2009;30:123–6. [4] Sudden death in epilepsy. Forensic and clinical issues. In: Lathers CM, Schraeder PL, Bungo MW, Leestma J, editors. Boca Raton, FL: CRC Press, Francis & Taylor Group; 2010. [5] Schraeder PL, So EL, Lathers CM. Forensic case identification. In: Lathers CM, Schraeder PL, Bungo MW, Leestma J, editors. Sudden death in epilepsy. Forensic and clinical issues. Boca Raton, FL: CRC Press, Francis & Taylor Group; 2010. p. 95–108. [6] Shields LB, Hunsaker III DM, Hunsaker JC, Parker Jr JC. Sudden unexpected death in epilepsy: neuropathologic findings. Am J Forensic Med Pathol 2002;23:307–14. [7] Lathers CM, Schraeder PL. Experience based teaching of therapeutics and clinical pharmacology of antiepileptic drugs. Sudden unexplained death in epilepsy: do antiepileptic drugs have a role? J Clin Pharmacol 1995;35:573–86 [quiz 586–7]. [8] Lathers CM, Schraeder PL, Bungo MW. The mystery of sudden death: mechanisms for risks. Epilepsy Behav 2008;12:3–24. [9] Lathers CM, Schraeder PL. Clinical pharmacology: drugs as a benefit and/or risk in sudden unexpected death in epilepsy. J Clin Pharmacol 2002;42:123–36. [10] Lathers CM, Schraeder PL. AED benefit/risk clinical pharmacology: Role of AEDs in cause and/or prevention of SUDEP. In: Lathers CM, Schraeder PL, Bungo MW, Leestma J, editors. Sudden death in epilepsy. Forensic and clinical issues. Boca Raton, FL: CRC Press, Francis & Taylor Group; 2010. p. 755–88. [11] Leestma JE, Walczak TW, Hughes JR, Kalelkar MB, Teas SS. A prospective study on sudden unexpected death in epilepsy. Ann Neurol 1989;26:195–203. [12] Terrence CF, Wiszotzkey HM, Perper JA. Unexpected, unexplained death in epileptic patients. Neurology 1975;25:594–8. [13] Kloster R, Torstein E. Sudden unexpected death in epilepsy (SUDEP): a clinical perspective and a search for risk factors. J Neurol Neurosurg Psychiatry 1999;67:439–44. [14] George JR, Davis GG. Comparison of anti-epileptic drug levels in different cases of sudden death. J Forensic Sci 1998;43:598–603. [15] Opeskin K, Burke MP, Cordner SM, Berkovic SF. Comparison of antiepileptic drug levels in sudden unexpected deaths in epilepsy with deaths from other causes. Epilepsia 1999;40:1795–8. [16] Walczak T. Do antiepileptic drugs play a role in sudden unexpected death in epilepsy? Drug Saf 2003;26:673–83. [17] Nilsson L, Bergman U, Diwan V, Farahmand BY, Persson PG, Tomson T. Antiepileptic drug therapy and its management in sudden unexpected death in epilepsy: a case– control study. Epilepsia 2001;42:667–73.

785

[18] Langan YL, Nashef L, Sander JW. Case–control study of SUDEP. Neurology 2005;64:1131–3. [19] Nashef L. Sudden unexplained death in epilepsy: terminology and definitions. Epilepsia 1997;38:S6–8. [20] Leestma JE, Annegars JF, Brodie MJ, et al. Sudden unexpected death in epilepsy: observations from a large clinical trial development program. Epilepsia 1997;38:47–55. [21] Leestma J. Forensic considerations and sudden unexpected death in epilepsy. In: Lathers CM, Schraeder PL, Bungo MW, Leestma J, editors. Sudden death in epilepsy. Forensic and clinical issues. Boca Raton, FL: CRC Press, Francis & Taylor Group; 2010. p. 37–56. [22] DiMaio VJ, DiMaio D. Chapter 1. Medical legal investive systems. In Forensic Pathology; 2001. p. 1–19. [23] Knight B. Forensic pathology. 2nd ed. London: Arnold; 1996. [24] Gaustaut H. Classification of the epilepsies: proposal for an international classification. Epilepsia 1969;10:14–21 [Suppl.]. [25] Gaustaut H, Zifkin BG. Classification of the epilepsies. J Clin Neurophysiol 1985;2: 313–26. [26] Walczak TS. Risk factors for sudden death in epilepsy. In: Lathers CM, Schraeder PL, Bungo MW, Leestma J, editors. Sudden death in epilepsy. Forensic and clinical issues. Boca Raton, FL: CRC Press, Francis & Taylor Group; 2010. p. 187–201. [27] Lathers CM, Schraeder PL, Bungo MW. Neurocardiologic mechanistic risk factors in sudden unexpected death in epilepsy. In: Lathers CM, Schraeder PL, Bungo MW, Leestma J, editors. Sudden death in epilepsy. Forensic and clinical issues. Boca Raton, FL: CRC Press, Francis & Taylor Group; 2010. p. 1–36. [28] Centers for Disease Control and Prevention, National Center for Health Statistics. Compresses Mortality File 1999–2006. CDC WONDER Online Database, compiled from Compressed Mortality File 1999–-2006 Series 20 No. 2L. Available at: http://wonder.cdc.gov/cmf-icd10-archive2006.html 2009 Accessed: July 27, 2011, 9:05:46 AM. [29] Lathers CM, Schraeder PL, Koehler SA, Wecht CH. Forensic postmortem examination of victims of sudden unexpected death in epilepsy. In: Lathers CM, Schraeder PL, Bungo MW, Leestma J, editors. Sudden death in epilepsy. Forensic and clinical issues. Boca Raton, FL: CRC Press, Francis & Taylor Group; 2010. p. 138–44. [30] Lathers CM, Schraeder PL. Autonomic dysfunction in epilepsy: characterization of autonomic cardiac neural discharge associated with pentylenetetrazol-induced epileptogenic activity. Epilesia 1982;23:633–47. [31] Chin PS, Branch KR, Becker KJ. Myocardial infarction following brief convulsive seizures. Neurology 2004;63:2453–4. [32] Chin PS, Branch KR, Becker KJ. Postictal neurogenic stunned myocardium. Neurology 2005;64:1977–8. [33] Natelson BH, Suarez RV, Terrence CF, Turizo R. Patients with epilepsy who die suddenly have cardiac disease. Arch Neurol 1998;55:857–60. [34] Terrence CF, Rao GR, Perper JA. Neurogenic pulmonary edema in unexpected, unexplained death of epileptic patients. Ann Neurol 1981:458–64. [35] Koehler SA, Schraeder PL, Lathers CM, Wecht CH. One-year postmortem forensic analysis of deaths in persons with epilepsy. In: Lathers CM, Schraeder PL, Bungo MW, Leestma J, editors. Sudden death in epilepsy. Forensic and clinical issues. Boca Raton, FL: CRC Press, Francis & Taylor Group; 2010. p. 145–59. [36] Jannuzzi G, Cian P, Fattore C, et al. A mulicenter randomized controlled trial on the clinical impact of therapeutic drug monitoring in patients with newly diagnosed epilepsy. Epilepsia 2000;41:222–30. [37] Opsekin K, Berkovic SF. Risk factors for sudden unexpected death in epilepsy: a controlled prospective study based on coroners cases. Seizure 2003;12:456–64. [38] Lund A, Gormsen H. The role of antiepileptics in sudden death in epilepsy. Acta Neurol Scand 1985;72:444–6. [39] May T, Jurgens U, Ambeck B, Schnabel R. Comparison between premortem and postmortem serum concentrations of phenobarbital, phenytoin, carbamazepine and its 10,11-epoxide metabolite in institutionalized patients with epilepsy. Epilepsy Res 1999;33:57–65. [40] Tomson T, Skold AC, Holmgen P, Nilsson L, Danielsson B. Postmortem changes in blood concentrations of phenytoin and carbamazepine: an experimental study. Ther Drug Monit 1998;20:309–12. [41] Williams J, Myson V, Steward S, et al. Self-discontinuation of antiepileptic medication in pregnancy: detection by hair analysis. Epilepsia 2002;43:824–31. [42] Williams J, Lawthom C, Dunstan FD, et al. Variability of antiepileptic medication taking behaviour in sudden unexplained death in epilepsy: hair analysis at autopsy. Neurol Neurosurg Psychiatry 2006;77:481–4. [43] Salmo EN, Connolly CE. Mortality in epilepsy in the west of Ireland: a 10-year review. Ir J Med Sci 2002;171:199–201. [44] Tennis P, Cole TB, Annegers JF, Leestma JE, McNutt M, Rajput A. Cohort study of incidence of sudden unexplained death in persons with seizure disorder treated with antiepileptic drugs in Saskatchewan, Canada. Epilepsia 1995;36:29–36. [45] Tomson T, Walczak M, Sillanpa M, Sander JW. Sudden unexpected death in epilepsy: a review of incidence and risk factors. Epilepsia 2005;46(Suppl. 11):54–61. [46] Hughes JR. A review of sudden unexpected death in epilepsy: prediction of patients at risk. Epilepsy Behav 2009;14:280–7. [47] Lathers CM, Schraeder PL. Verbal autopsies. Epilepsy Behav 2009;14:573–6. [48] Lathers CM. Epilepsy and sudden death: personal reflections and call for global action. Epilepsy Behav 2009;15:269–77. [49] Lathers CM, Schraeder PL, Claycamp HG. Clinical pharmacology of topiramate versus lamotrigine versus phenobarbital: comparison of efficacy and side effects using odds ratios. J Clin Pharmacol 2003;43:491–503. [50] Lathers CM. Antiepileptic drugs, compliance and SUDEP. In: Chapman D, Panelli R, Hanna J, Jeffs T, editors. Sudden unexpected death in epilepsy: continued global discussion; 2011. Available at:http://www.sudepglobalconversation.com/files/ sudden_unexpected_death_in_epilepsy_2011.pdf.