- Email: [email protected]
Drowning in epilepsy: A population-based case series
Epilepsy Research 145 (2018) 123–126
Contents lists available at ScienceDirect
Epilepsy Research journal homepage: www.elsevier.com/locate/epilepsyres
Drowning in epilepsy: A population-based case series a
a
a
a
T b
Eva Bain , Anne E. Keller , Jordan Ho , Robyn Whitney , Michael S. Pollanen , ⁎ Andrew S. Williamsb, Elizabeth J. Donnera, a b
The Hospital for Sick Children, Division of Neurology, 555 University Avenue, Toronto, Ontario, M5G 1X8, Canada Ontario Forensic Pathology Service & Department of Laboratory Medicine and Pathobiology, University of Toronto, Toronto, Ontario, Canada
A R T I C LE I N FO
A B S T R A C T
Keywords: Drowning Epilepsy Seizures Mortality
Objectives: The risk of drowning is reported to be 15–19 times greater in people with epilepsy compared to the general population. Despite this disproportionate burden, there is limited data about the circumstances surrounding drowning deaths in people with epilepsy. This population-based case series characterizes drowning deaths in people with epilepsy. Methods: Postmortem data from coroner-ordered autopsies conducted in Ontario between 2014 and 2016 were screened for cases of drowning in people with a history of seizures. Demographic information, epilepsy characteristics, and circumstances surrounding death were extracted from post mortem reports. The incidence of drowning in people with epilepsy was calculated using government estimates of the Ontario population and the number of people with epilepsy. Results: Twenty-five people with epilepsy drowned during the three-year study period, giving an estimated incidence of 1.5 per 10,000 epilepsy person-years (95% CI: 0.98, 2.23). Decedents were mostly young (mean age 36 years) and without physical or developmental disability. Approximately one-third had psychiatric comorbidities. Epilepsy severity ranged from well-controlled to drug refractory. Only 3 people had alcohol or illicit drugs detected on toxicological analysis. Forty-four percent of deaths were the result of an unwitnessed drowning in a bathtub. Conclusions: This population-based case series confirms people with epilepsy drown at a rate nearly ten times greater than the general population (1.5 per 10,000 epilepsy person-years compared to the estimated provincial average of 0.13 per 10,000). Drowning deaths in people with epilepsy most often occur in the bathtub. These deaths are only rarely associated with intoxication. People with epilepsy should receive counseling on the increased risk of drowning, including information regarding the significant risk associated with bathtub use, the potential protective roles of anti-epileptic drug (AED) adherence and supervision when in or around water, and the fact that all people with epilepsy remain at an increased risk of drowning regardless of their apparent seizure control.
1. Introduction People with epilepsy are at a 2–3 times greater risk of premature mortality than the general population (Thurman et al., 2017). Drowning is one cause of premature mortality in people with epilepsy and is a potentially preventable cause of death. Based on a 2008 metaanalysis, people with epilepsy have a 15–19 times greater risk of drowning than the general population (Bell et al., 2008), which is evidenced in population-based drowning cohort studies (Peden et al., 2018; Mahony et al., 2017). While the excess burden of drowning in those with epilepsy is relatively well-documented, information about the circumstances
⁎
surrounding death and characteristics of those who died is limited. The only population-based study of drownings in people with epilepsy was conducted between 1981 and 1990 in Alberta, Canada and identified 25 deaths (Ryan and Dowling, 1993). Other studies have been conducted exclusively in children (Diekema et al., 1993; Franklin et al., 2017; Kemp and Sibert, 1993; Orlowski et al., 1982; Pearn, 1977). It is unknown whether these findings describe the general and current epilepsy populations. To better inform prevention recommendations, a population-based case series was conducted to describe characteristics of epilepsy-related drownings and obtain a current estimate of the incidence of drowning in people with epilepsy.
Corresponding author. E-mail addresses: [email protected] (E. Bain), [email protected] (A.E. Keller), [email protected] (H. Jordan), [email protected] (W. Robyn), [email protected] (M.S. Pollanen), [email protected] (A.S. Williams), [email protected] (E.J. Donner). https://doi.org/10.1016/j.eplepsyres.2018.06.010 Received 1 May 2018; Received in revised form 8 June 2018; Accepted 20 June 2018 Available online 20 June 2018 0920-1211/ © 2018 Published by Elsevier B.V.
Epilepsy Research 145 (2018) 123–126
E. Bain et al.
2. Methods
Table 1 Population Characteristics.
Case summaries of all autopsies conducted by the Ontario Forensic Pathology Service in Ontario, Canada between January 1st, 2014 and December 31st, 2016 were retrospectively searched for the terms “Seiz”, “Epilep-”, and “Sz” to identify decedents with a history of epilepsy or seizures. Cases were then reviewed for evidence of drowning or potential drowning, either through the listed cause of death or a history suggesting the death occurred around water. Full post mortem reports of these cases were reviewed. Cases were excluded if the history of seizure did not meet the inclusion criteria for epilepsy, that is if there was no history of recurrent unprovoked seizures, or if there was no evidence of drowning detailed in the post mortem report. Data was abstracted from the post mortem reports, including basic demographic information, medical history, epilepsy/seizure history, the pathologist’s judgment of the causal and contributing factors to death, circumstances surrounding death, and post-mortem toxicology results. Natural water drownings were further described by water depth, as shallow when the water was less than 12 in. deep and deep when the water depth was 12 in and greater. The incidence of drowning in people with epilepsy was calculated by dividing the number of observed cases by the estimated number of epilepsy person-years during the study period. Epilepsy person-years was calculated by applying an estimate of the prevalence of epilepsy across all ages from 2010 to 2012 obtained from government survey data (Gilmour et al., 2016) to government census data of the population living in Ontario during 2014–2016 (Statistics Canada, 2016). Confidence intervals (95%) were determined using MedCalc Statistical Software version 17.9.7 (MedCalc Software bvba, Ostend, Belgium; http://www.medcalc.org; 2017), assuming the Poisson distribution. This study was approved by the Hospital for Sick Children Research Ethics Board (REB# 1000034304).
Variable
Category
N = 25
%
Sex
Male Female
14 11
56% 44%
Age
< 15 15–24 25–34 35–44 45–54 > 55
1 7 6 2 6 3
4% 28% 24% 8% 24% 12%
BMI Categorya
Underweight (< 18.5 kg/ m2) Normal (18.5–24.9 kg/m2) Overweight (25–29.9 kg/m2) Obese (> 30 kg/m2) Unknown
1
4%
11 8 3 2
44% 32% 12% 8%
Co-morbidities
Developmental Delay Psychiatric Condition Motor Disability
4 8 3
16% 32% 12%
Last Known Seizure
< 1 month > 1 month to < 2 years > 2 years Unknown
3 3 2 17
12% 12% 8% 68%
Seizure Severity as Indicated By Pathologist
“Relatively Well-Controlled”
1
4%
“Intractable”/“Brittle” No Statement Made
2 22
8% 88%
Seizure Frequency
> 1 per month Increasing in frequency “Often” Unknown
2 1 1 21
8% 4% 4% 84%
Evidence of Craniotomyb
Yes
5
21%
AED Therapy
Monotherapy Polytherapy On medication, unclear of number No AEDs Unknown
10 5 3
40% 20% 12%
1 6
4% 24%
Adherent
1
4%
Non-adherent Not reported
5 18
21% 75%
3. Results Thirty-two people with a history of seizures and a death suspicious of drowning were identified based on case history summaries and cause of death. After review of the post mortem reports, 25 cases were considered drowning in an individual with epilepsy as per the pathologist’s interpretation. Of the seven excluded cases, four had no evidence of drowning as per the pathologist’s report and three did not meet epilepsy inclusion criteria (one had provoked seizures and two had seizure histories limited to a single seizure at the time of death). There was no indication in the data reviewed that any of the cases represented intentional drowning. The estimated incidence of drowning in the general epilepsy population of Ontario is 1.5 per 10,000 people with epilepsy (95% CI: 0.98, 2.23). This was calculated using government survey data that reported a prevalence of active epilepsy of 4.0 per 1000 people (Gilmour et al., 2016), and the summed population in Ontario for each year during the study period of 2014–2016 (total 41,465,200) (Statistics Canada, 2016). Characteristics of the drowning cohort are described in Table 1. Fourteen (56%) were male. The mean age was 36 years (median age 32, range 12–68). In the 23 cases with information about body habitus three were obese and eight were overweight. About a third (n = 8) had documented psychiatric comorbidities, including schizophrenia, depression, borderline personality disorder, bipolar disorder, and anxiety disorder. Few had documented developmental delays (n = 4; 3 with global delays, 1 with a learning disability) or motor disability (n = 3). Information available from the post mortem reports on seizure frequency and severity was limited. When present, time from last known seizure ranged from 10 days to 2.5 years and seizure severity was described as “relatively well-controlled” to “intractable”. The majority of people were known to be prescribed antiepileptic drugs (AED) (n = 18) and, of these, most (n = 10) were on monotherapy. Twenty-one percent
Reported History of AED Adherencec
a One individual was a child and as such growth curves were used instead of BMI calculations. When BMI could not be calculated (n = 2), clear indications of body habitus from autopsy were used as an estimate of BMI. b The one individual who did not receive a full autopsy was not included in this calculation. c The one individual who was not on any AEDs was not included in this calculation.
(n = 5) had evidence of a prior craniotomy. Five people had a reported history of AED nonadherence. Details of the circumstances surrounding death are presented in Table 2. Eleven (44%) died in the bathtub, and 10 (40%) died in natural water settings. Most deaths were unwitnessed (n = 20, 80%). Six people had evidence supporting a seizure prior to drowning, such as a tongue bite at autopsy or a documented seizure (e.g. video footage from a security camera, witness account). In 12 cases (48%), seizures were described in the post-mortem report as the only contributor to drowning. In 5 cases (20%), seizure, along with another competing potential contributor to the drowning was described by the pathologist, including possible arrhythmia in the setting of heart disease (n = 4), and acute ethanol intoxication (n = 1). In 8 cases (32%), the postmortem report did not include an acknowledgement that a seizure could have contributed to the drowning. In 4 of these cases, no additional contributing factors to drowning were listed. In the other 4, the report listed factors that may have contributed to death, but were not 124
Epilepsy Research 145 (2018) 123–126
E. Bain et al.
Variable
Category
N = 25
%
Location
Bathtub Natural Water – Deep Natural Water – Shallow Pool or Hot Tub Bucket, Sink, or Small WaterContaining Vessel
11 9 1 2 2
44% 36% 4% 8% 8%
levels. Ethanol was tested in 20 cases and cannabis (tetrahydrocannabinol/cannabidiol) in 8 cases. Of these, 2 cases were positive for ethanol (141 mg/100 mL and 196 mg/100 mL), both considered to be in the intoxicated range for driving (80 mg/100 mL in Ontario). One of these cases was also positive for cannabis (8.3 ng/mL tetrahydrocannabinol). An additional case was positive for cannabis only (< 2 ng/mL tetrahydrocannabinol). No other illicit drugs were detected in samples tested with laboratory methods capable of detecting them.
Activity at Time of Drowning
Bathing
12
48%
4. Discussion
Swimming/Non-Motorized Boating Other
3 10
12% 40%
Witnessed
Not witnessed Not directly witnessed, someone nearby Witnessed
20 3
80% 12%
2
8%
Final Cause of Death
Drowning Undetermineda
24 1
96% 4%
Evidence of Seizure Prior to Drowning
Tongue bite at autopsy
4
16%
Video footage of seizure Decedent stated they were about to have a seizure
1 1
4% 4%
Seizures are the only listed contributor to death
12
48%
Seizure potentially led to death, potential competing factors are also present No relationship between seizures and cause of death described/ reported
5
20%
8
32%
Detectedc
8
32%
3 14 3 2
12% 56% 21% 14%
8 1
57% 7%
Twenty-five cases of drowning were identified in people with epilepsy over a three-year period in Ontario, Canada, giving an estimated incidence of 1.5 per 10,000 people with epilepsy per year (95% CI: 0.98, 2.23). About half of the deaths occurred in the bathtub. Two people tested within the impaired range for alcohol and two positive for cannabis among the cases that had been tested with laboratory methods capable of detecting these substances. People were generally young (mean age = 36 years, median age = 32 years), ambulatory, and had no documented developmental disability. About one-third had psychiatric co-morbidities, but none of the cases appeared to be the result of intentional drowning. Of those who were known to be prescribed AEDs, most were on monotherapy. The current findings are generally concordant with past studies of drowning in people with epilepsy. Ryan and Dowling (1993) found 60% (n = 15) of drownings in people with epilepsy occurred in the bathtub, which is slightly greater than the proportion reported here (44%, n = 11). Bathtubs are also cited as a very common location of drowning in studies of children with epilepsy (Diekema et al., 1993; Franklin et al., 2017; Kemp and Sibert, 1993; Pearn, 1977). Ryan and Dowling (1993) also found that deaths were typically unwitnessed, alcohol and illicit drugs were not a factor in any cases, and that some deaths can be preceded by a long seizure-free interval (15 years in a single case included in their report). The current cohort similarly showed a predominance of unwitnessed deaths, a paucity of alcohol and illicit drug intoxication, and at least one example of a long seizure free interval (2.5 years). Eight individuals were known to be overweight, and 3 to be obese. It is unclear what role, if any, body habitus may have played in drowning. It is possible that an individual with increased soft tissue may have an impaired ability to remove himself or herself from a compromising position following a seizure, which could potentially predispose them to drowning. Conversely, individuals with a higher BMI may have more adipose tissue, which could improve their buoyancy and thus their ability to stay afloat compared to individuals of a lower BMI (Byard, 2017). The proportion of cases with undetected AEDs on post mortem toxicology is higher than in past work. Three of eleven people with epilepsy (27%) in our study were prescribed AEDs that were tested and undetected compared to 1 of 23 (4%) in the series by Ryan and Dowling (1993). Ryan and Dowling also reported a high proportion of people had subtherapeutic AED levels. We included those with subtherapeutic levels in the “detected” category because evidence indicates that subtherapeutic post-mortem AED levels may not be a reliable indicator of subtherapeutic pre-mortem AED levels and thus should not be used to indicate nonadherence (Lathers et al., 2011; May et al., 1999). Consequently, it is unclear the role that subtherapeutic AED levels or nonadherence may have played in the drownings. In contrast to concerns regarding AED nonadherence and subtherapeutic AED levels, supratherapeutic AED levels may also be importantly linked to drowning. We documented one case of supratherapeutic AED levels (phenobarbital), which, while not elevated sufficiently to be the cause of death, was considered to contribute to the drowning, through central nervous system depression and impaired ability to stay afloat. It is unknown whether supratherapeutic AED levels may have been a contributing factor in other cases as well because
Table 2 Circumstances Surrounding Drowning.
Role of Seizure In Drowning as Per Pathologistb
AED Levels on Post Mortem Toxicology
d
Not Detected Not Tested or Unknown a) On AED(s) that can be tested b) On AED that cannot be tested (levetiracetam) c) Specific AED(s) unknown d) Not prescribed AED
This individual’s formal cause of death was given as “undetermined” because autopsy was limited to external examination (in accommodation with family’s wishes), however, the pathologist considered the scene and circumstances surrounding the death to be supportive of accidental drowning. b As per final cause of death, or pathologist’s summary/opinion. c Includes any individual with a trace, detected, subtherapeutic, therapeutic, or supratherapeutic AED level (see Discussion). d One individual on polytherapy had only some of the prescribed AEDs detected. This individual was included in the AED not detected group (considered a marker of non-adherence). a
sufficient to be the cause of death, including developmental delay, AED toxicity, blunt force facial trauma, and possible loss of consciousness due to trauma sustained as the individual entered the water. There were no cases where seizure was excluded as a cause of drowning, or where there was a clear cause of drowning other than seizure. There was no difference in age between bathtub drownings and natural water drownings (median age 33 vs. 31.5 years), however, bathtub drownings were more likely to be females (72% vs. 10%, p = 0.008). Body habitus was similar between drowning settings: 45% of bathtub drownings occurred in individuals who were overweight (n = 3) or obese (n = 2) while 40% of natural water drownings occurred in overweight individuals (n = 4), though none were obese. Complete body habitus information was unavailable in one natural water drowning. Toxicological analysis varied across cases. Testing of at least one AED was completed in 11 cases. Of these, 3 cases had undetected AED 125
Epilepsy Research 145 (2018) 123–126
E. Bain et al.
AEDs were often only reported as “detected” on toxicological testing reports. The role of a seizure in these drowning deaths is unclear. Beyond a few cases of witnessed or video-captured seizures, evidence of seizure may be inferred by the presence of a tongue bite identified at autopsy, which is not entirely specific. Alternatively, a seizure may occur and leave no signs at autopsy. Approximately one quarter of cases had evidence of a seizure around the time of drowning. In some cases, the potential role of seizures was further obscured by other case factors that may potentially contribute to drowning (e.g. cardiac disease, evidence of injury that may have led to loss of consciousness). The well-established psychiatric comorbidities, developmental delays, and social isolation (Camfield and Camfield, 2007) associated with epilepsy may increase the likelihood of drowning. Approximately one-third of our cohort had psychiatric comorbidities including bipolar disorder, schizophrenia, obsessive-compulsive disorder, anxiety, and others. The exact role this may have played in the events leading to drowning is not known. Drowning in people with epilepsy is markedly different from drowning in the general population. The estimated incidence of drowning reported here is approximately ten times greater than that quoted in the general population (1.5 per 10,000 people with epilepsy compared to 0.13 per 10,000 in the general population) (Lifesaving Society, 2016). In the general population, 60% of bathtub drowning deaths occur in those 65 years of age or older; in contrast, all bathtub drownings reported here occurred in those under 60. Alcohol is a significant risk factor for drowning and is implicated in approximately half of all drowning deaths in those ages 20–64 in the general population (Lifesaving Society, 2016), whereas only 2 people with epilepsy in the current study had evidence of alcohol consumption. This population-based study offers current insight into drowning in people with epilepsy, based on a comprehensive dataset collected in a province with a large population and diversity in culture, geography, and degree of urbanization. Screening of forensic autopsies is anticipated to have identified nearly all, if not all, people with epilepsy who drowned during the study period. Review of the full post mortem reports allowed for exclusion of people who suffered an alternative cause of death in the context of a water setting. Finally, in most cases, there were no other obvious potential contributors to drowning besides seizures, which allows for a clearer understanding of the role epilepsy may have played in these drownings. There are several limitations to this study. First, the study was retrospective in nature and information available was often limited, especially regarding the seizure history. A history of “epilepsy”, “seizure disorder”, or unprovoked “seizures” in the post mortem report was accepted as evidence of epilepsy, but, without a more detailed seizure history available, it cannot be certain that all subjects reported met the current clinical definition of epilepsy. Second, missed cases cannot be fully excluded. All drowning cases should be reported to the coroner and would be eligible for forensic autopsy, but it is possible that a case was missed (e.g. a drowning that was mistakenly not referred to a coroner, refused by a coroner, or not referred for autopsy). Furthermore, not all cases had toxicological testing and laboratory methods limit the ability to detect all illicit substances. It is possible illicit drugs may have been involved in additional cases. Finally, the incidence of drowning in people with epilepsy is estimated by calculation and is limited by the uncertainty of the exact number of people living in Ontario with epilepsy from 2014–2016. Underestimating the
prevalence of epilepsy would overestimate of the incidence of drowning in the epilepsy population, and conversely, overestimating the prevalence of epilepsy would underestimate the incidence of drowning in people with epilepsy. 5. Conclusion While the overall incidence of drowning in people with epilepsy is low, it represents a highly preventable cause of death. People with epilepsy should receive counseling on the increased risk of drowning, including information regarding the significant risk associated with bathtub use, the potential protective roles of supervision when in or around water and the fact that all people with epilepsy remain at an increased risk of drowning regardless of their apparent seizure control. It is not known how many and to what extent people with epilepsy receive counseling regarding drowning prevention. This is an area for future research. Improved awareness of the risk of drowning and regular education around preventative measures may help to prevent this devastating outcome of drowning in people with epilepsy. Disclosure of conflicts of interest The authors report no conflicts of interest to disclose. References Bell, G.S., Gaitatzis, A., Bell, C.L., Johnson, A.I., Sander, J.W., 2008. Drowning in people with epilepsy, how great is the risk? Neurology 71 (8), 578–582. Byard, R.W., 2017. Body mass index and saltwater drowning. Am. J. Forensic Med. Pathol. 38 (4), 304–305. Camfield, C.S., Camfield, P.R., 2007. Long‐term social outcomes for children with epilepsy. Epilepsia 48 (s9), 3–5. Diekema, D.S., Quan, L., Holt, V.L., 1993. Epilepsy as a risk factor for submersion injury in children. Pediatrics 91 (3), 612–616. Franklin, R.C., Pearn, J.H., Peden, A.E., 2017. Drowning fatalities in childhood: the role of pre-existing medical conditions. Arch Dis. Child. 102, 888–893. Gilmour, H., Ramage-Morin, P., Wong, S.L., 2016. Epilepsy in Canada: prevalence and impact. Heal Rep. 27 (9), 24–30. Kemp, A.M., Sibert, J.R., 1993. Epilepsy in children and the risk of drowning. Arch Dis. Child. 68 (5), 684–685. Lathers, C.M., Koehler, S.A., Wecht, C.H., Schraeder, P.L., 2011. Forensic antiepileptic drug levels in autopsy cases of epilepsy. Epilepsy Behav. 22 (4), 778–785. Lifesaving Society, 2016. Ontario Drowning Report 2016. (Accessed 12 July 2017). http://www.lifesaving.ca/wp-content/uploads/2013/09/Ontario-Drowning-Report2016.pdf. Mahony, A.J., Peden, A.E., Franklin, R.C., Pearn, J.H., Scarr, J., 2017. Fatal, unintentional drowning in older people: an assessment of the role of preexisting medical conditions. Healthy Aging Res. 6 (e7), 1–8. May, T., Jurgens, U., Rambeck, B., Schnabel, R., 1999. Comparison between premortem and postmortem serum concentrations of phenobarbital, phenytoin, carbamazepine, and its 10,11-epoxide metabolite in institutionalized patients with epilepsy. Epilepsy Res. 33 (1), 57–65. Orlowski, J.P., Rothner, A.D., Lueders, H., 1982. Submersion accidents in children with epilepsy. Am. J. Dis. Child. 136 (9), 777–780. Pearn, J.M., 1977. Epilepsy and drowning in childhood. BMJ 1 (6075), 1510. Peden, A.E., Franklin, R.C., Pearn, J.H., 2018. Unintentional fatal child drowning in the bath: a 12-year review (2002–2014). J. Paediatr. Child Health 54 (2), 153–159. Ryan, C.A., Dowling, G., 1993. Drowning deaths in people with epilepsy. Can. Med. Assoc. J. 148 (5), 781. Statistics Canada, 2016. Population by Year, by Province and Territory. (Accessed 3 August 2017. http://www.statcan.gc.ca/tables-tableaux/sum-som/l01/cst01/ demo02a-eng.htm. Thurman, D.J., Logroscino, G., Beghi, E., Hauser, W.A., Hesdorffer, D.C., Newton, C.R., Scorza, F.A., Sander, J.W., Tomson, T., 2017. The burden of premature mortality of epilepsy in high-income countries: a systematic review from the Mortality Task Force of the International League Against Epilepsy. Epilepsia 58 (1), 17–26.
126
Contents lists available at ScienceDirect
Epilepsy Research journal homepage: www.elsevier.com/locate/epilepsyres
Drowning in epilepsy: A population-based case series a
a
a
a
T b
Eva Bain , Anne E. Keller , Jordan Ho , Robyn Whitney , Michael S. Pollanen , ⁎ Andrew S. Williamsb, Elizabeth J. Donnera, a b
The Hospital for Sick Children, Division of Neurology, 555 University Avenue, Toronto, Ontario, M5G 1X8, Canada Ontario Forensic Pathology Service & Department of Laboratory Medicine and Pathobiology, University of Toronto, Toronto, Ontario, Canada
A R T I C LE I N FO
A B S T R A C T
Keywords: Drowning Epilepsy Seizures Mortality
Objectives: The risk of drowning is reported to be 15–19 times greater in people with epilepsy compared to the general population. Despite this disproportionate burden, there is limited data about the circumstances surrounding drowning deaths in people with epilepsy. This population-based case series characterizes drowning deaths in people with epilepsy. Methods: Postmortem data from coroner-ordered autopsies conducted in Ontario between 2014 and 2016 were screened for cases of drowning in people with a history of seizures. Demographic information, epilepsy characteristics, and circumstances surrounding death were extracted from post mortem reports. The incidence of drowning in people with epilepsy was calculated using government estimates of the Ontario population and the number of people with epilepsy. Results: Twenty-five people with epilepsy drowned during the three-year study period, giving an estimated incidence of 1.5 per 10,000 epilepsy person-years (95% CI: 0.98, 2.23). Decedents were mostly young (mean age 36 years) and without physical or developmental disability. Approximately one-third had psychiatric comorbidities. Epilepsy severity ranged from well-controlled to drug refractory. Only 3 people had alcohol or illicit drugs detected on toxicological analysis. Forty-four percent of deaths were the result of an unwitnessed drowning in a bathtub. Conclusions: This population-based case series confirms people with epilepsy drown at a rate nearly ten times greater than the general population (1.5 per 10,000 epilepsy person-years compared to the estimated provincial average of 0.13 per 10,000). Drowning deaths in people with epilepsy most often occur in the bathtub. These deaths are only rarely associated with intoxication. People with epilepsy should receive counseling on the increased risk of drowning, including information regarding the significant risk associated with bathtub use, the potential protective roles of anti-epileptic drug (AED) adherence and supervision when in or around water, and the fact that all people with epilepsy remain at an increased risk of drowning regardless of their apparent seizure control.
1. Introduction People with epilepsy are at a 2–3 times greater risk of premature mortality than the general population (Thurman et al., 2017). Drowning is one cause of premature mortality in people with epilepsy and is a potentially preventable cause of death. Based on a 2008 metaanalysis, people with epilepsy have a 15–19 times greater risk of drowning than the general population (Bell et al., 2008), which is evidenced in population-based drowning cohort studies (Peden et al., 2018; Mahony et al., 2017). While the excess burden of drowning in those with epilepsy is relatively well-documented, information about the circumstances
⁎
surrounding death and characteristics of those who died is limited. The only population-based study of drownings in people with epilepsy was conducted between 1981 and 1990 in Alberta, Canada and identified 25 deaths (Ryan and Dowling, 1993). Other studies have been conducted exclusively in children (Diekema et al., 1993; Franklin et al., 2017; Kemp and Sibert, 1993; Orlowski et al., 1982; Pearn, 1977). It is unknown whether these findings describe the general and current epilepsy populations. To better inform prevention recommendations, a population-based case series was conducted to describe characteristics of epilepsy-related drownings and obtain a current estimate of the incidence of drowning in people with epilepsy.
Corresponding author. E-mail addresses: [email protected] (E. Bain), [email protected] (A.E. Keller), [email protected] (H. Jordan), [email protected] (W. Robyn), [email protected] (M.S. Pollanen), [email protected] (A.S. Williams), [email protected] (E.J. Donner). https://doi.org/10.1016/j.eplepsyres.2018.06.010 Received 1 May 2018; Received in revised form 8 June 2018; Accepted 20 June 2018 Available online 20 June 2018 0920-1211/ © 2018 Published by Elsevier B.V.
Epilepsy Research 145 (2018) 123–126
E. Bain et al.
2. Methods
Table 1 Population Characteristics.
Case summaries of all autopsies conducted by the Ontario Forensic Pathology Service in Ontario, Canada between January 1st, 2014 and December 31st, 2016 were retrospectively searched for the terms “Seiz”, “Epilep-”, and “Sz” to identify decedents with a history of epilepsy or seizures. Cases were then reviewed for evidence of drowning or potential drowning, either through the listed cause of death or a history suggesting the death occurred around water. Full post mortem reports of these cases were reviewed. Cases were excluded if the history of seizure did not meet the inclusion criteria for epilepsy, that is if there was no history of recurrent unprovoked seizures, or if there was no evidence of drowning detailed in the post mortem report. Data was abstracted from the post mortem reports, including basic demographic information, medical history, epilepsy/seizure history, the pathologist’s judgment of the causal and contributing factors to death, circumstances surrounding death, and post-mortem toxicology results. Natural water drownings were further described by water depth, as shallow when the water was less than 12 in. deep and deep when the water depth was 12 in and greater. The incidence of drowning in people with epilepsy was calculated by dividing the number of observed cases by the estimated number of epilepsy person-years during the study period. Epilepsy person-years was calculated by applying an estimate of the prevalence of epilepsy across all ages from 2010 to 2012 obtained from government survey data (Gilmour et al., 2016) to government census data of the population living in Ontario during 2014–2016 (Statistics Canada, 2016). Confidence intervals (95%) were determined using MedCalc Statistical Software version 17.9.7 (MedCalc Software bvba, Ostend, Belgium; http://www.medcalc.org; 2017), assuming the Poisson distribution. This study was approved by the Hospital for Sick Children Research Ethics Board (REB# 1000034304).
Variable
Category
N = 25
%
Sex
Male Female
14 11
56% 44%
Age
< 15 15–24 25–34 35–44 45–54 > 55
1 7 6 2 6 3
4% 28% 24% 8% 24% 12%
BMI Categorya
Underweight (< 18.5 kg/ m2) Normal (18.5–24.9 kg/m2) Overweight (25–29.9 kg/m2) Obese (> 30 kg/m2) Unknown
1
4%
11 8 3 2
44% 32% 12% 8%
Co-morbidities
Developmental Delay Psychiatric Condition Motor Disability
4 8 3
16% 32% 12%
Last Known Seizure
< 1 month > 1 month to < 2 years > 2 years Unknown
3 3 2 17
12% 12% 8% 68%
Seizure Severity as Indicated By Pathologist
“Relatively Well-Controlled”
1
4%
“Intractable”/“Brittle” No Statement Made
2 22
8% 88%
Seizure Frequency
> 1 per month Increasing in frequency “Often” Unknown
2 1 1 21
8% 4% 4% 84%
Evidence of Craniotomyb
Yes
5
21%
AED Therapy
Monotherapy Polytherapy On medication, unclear of number No AEDs Unknown
10 5 3
40% 20% 12%
1 6
4% 24%
Adherent
1
4%
Non-adherent Not reported
5 18
21% 75%
3. Results Thirty-two people with a history of seizures and a death suspicious of drowning were identified based on case history summaries and cause of death. After review of the post mortem reports, 25 cases were considered drowning in an individual with epilepsy as per the pathologist’s interpretation. Of the seven excluded cases, four had no evidence of drowning as per the pathologist’s report and three did not meet epilepsy inclusion criteria (one had provoked seizures and two had seizure histories limited to a single seizure at the time of death). There was no indication in the data reviewed that any of the cases represented intentional drowning. The estimated incidence of drowning in the general epilepsy population of Ontario is 1.5 per 10,000 people with epilepsy (95% CI: 0.98, 2.23). This was calculated using government survey data that reported a prevalence of active epilepsy of 4.0 per 1000 people (Gilmour et al., 2016), and the summed population in Ontario for each year during the study period of 2014–2016 (total 41,465,200) (Statistics Canada, 2016). Characteristics of the drowning cohort are described in Table 1. Fourteen (56%) were male. The mean age was 36 years (median age 32, range 12–68). In the 23 cases with information about body habitus three were obese and eight were overweight. About a third (n = 8) had documented psychiatric comorbidities, including schizophrenia, depression, borderline personality disorder, bipolar disorder, and anxiety disorder. Few had documented developmental delays (n = 4; 3 with global delays, 1 with a learning disability) or motor disability (n = 3). Information available from the post mortem reports on seizure frequency and severity was limited. When present, time from last known seizure ranged from 10 days to 2.5 years and seizure severity was described as “relatively well-controlled” to “intractable”. The majority of people were known to be prescribed antiepileptic drugs (AED) (n = 18) and, of these, most (n = 10) were on monotherapy. Twenty-one percent
Reported History of AED Adherencec
a One individual was a child and as such growth curves were used instead of BMI calculations. When BMI could not be calculated (n = 2), clear indications of body habitus from autopsy were used as an estimate of BMI. b The one individual who did not receive a full autopsy was not included in this calculation. c The one individual who was not on any AEDs was not included in this calculation.
(n = 5) had evidence of a prior craniotomy. Five people had a reported history of AED nonadherence. Details of the circumstances surrounding death are presented in Table 2. Eleven (44%) died in the bathtub, and 10 (40%) died in natural water settings. Most deaths were unwitnessed (n = 20, 80%). Six people had evidence supporting a seizure prior to drowning, such as a tongue bite at autopsy or a documented seizure (e.g. video footage from a security camera, witness account). In 12 cases (48%), seizures were described in the post-mortem report as the only contributor to drowning. In 5 cases (20%), seizure, along with another competing potential contributor to the drowning was described by the pathologist, including possible arrhythmia in the setting of heart disease (n = 4), and acute ethanol intoxication (n = 1). In 8 cases (32%), the postmortem report did not include an acknowledgement that a seizure could have contributed to the drowning. In 4 of these cases, no additional contributing factors to drowning were listed. In the other 4, the report listed factors that may have contributed to death, but were not 124
Epilepsy Research 145 (2018) 123–126
E. Bain et al.
Variable
Category
N = 25
%
Location
Bathtub Natural Water – Deep Natural Water – Shallow Pool or Hot Tub Bucket, Sink, or Small WaterContaining Vessel
11 9 1 2 2
44% 36% 4% 8% 8%
levels. Ethanol was tested in 20 cases and cannabis (tetrahydrocannabinol/cannabidiol) in 8 cases. Of these, 2 cases were positive for ethanol (141 mg/100 mL and 196 mg/100 mL), both considered to be in the intoxicated range for driving (80 mg/100 mL in Ontario). One of these cases was also positive for cannabis (8.3 ng/mL tetrahydrocannabinol). An additional case was positive for cannabis only (< 2 ng/mL tetrahydrocannabinol). No other illicit drugs were detected in samples tested with laboratory methods capable of detecting them.
Activity at Time of Drowning
Bathing
12
48%
4. Discussion
Swimming/Non-Motorized Boating Other
3 10
12% 40%
Witnessed
Not witnessed Not directly witnessed, someone nearby Witnessed
20 3
80% 12%
2
8%
Final Cause of Death
Drowning Undetermineda
24 1
96% 4%
Evidence of Seizure Prior to Drowning
Tongue bite at autopsy
4
16%
Video footage of seizure Decedent stated they were about to have a seizure
1 1
4% 4%
Seizures are the only listed contributor to death
12
48%
Seizure potentially led to death, potential competing factors are also present No relationship between seizures and cause of death described/ reported
5
20%
8
32%
Detectedc
8
32%
3 14 3 2
12% 56% 21% 14%
8 1
57% 7%
Twenty-five cases of drowning were identified in people with epilepsy over a three-year period in Ontario, Canada, giving an estimated incidence of 1.5 per 10,000 people with epilepsy per year (95% CI: 0.98, 2.23). About half of the deaths occurred in the bathtub. Two people tested within the impaired range for alcohol and two positive for cannabis among the cases that had been tested with laboratory methods capable of detecting these substances. People were generally young (mean age = 36 years, median age = 32 years), ambulatory, and had no documented developmental disability. About one-third had psychiatric co-morbidities, but none of the cases appeared to be the result of intentional drowning. Of those who were known to be prescribed AEDs, most were on monotherapy. The current findings are generally concordant with past studies of drowning in people with epilepsy. Ryan and Dowling (1993) found 60% (n = 15) of drownings in people with epilepsy occurred in the bathtub, which is slightly greater than the proportion reported here (44%, n = 11). Bathtubs are also cited as a very common location of drowning in studies of children with epilepsy (Diekema et al., 1993; Franklin et al., 2017; Kemp and Sibert, 1993; Pearn, 1977). Ryan and Dowling (1993) also found that deaths were typically unwitnessed, alcohol and illicit drugs were not a factor in any cases, and that some deaths can be preceded by a long seizure-free interval (15 years in a single case included in their report). The current cohort similarly showed a predominance of unwitnessed deaths, a paucity of alcohol and illicit drug intoxication, and at least one example of a long seizure free interval (2.5 years). Eight individuals were known to be overweight, and 3 to be obese. It is unclear what role, if any, body habitus may have played in drowning. It is possible that an individual with increased soft tissue may have an impaired ability to remove himself or herself from a compromising position following a seizure, which could potentially predispose them to drowning. Conversely, individuals with a higher BMI may have more adipose tissue, which could improve their buoyancy and thus their ability to stay afloat compared to individuals of a lower BMI (Byard, 2017). The proportion of cases with undetected AEDs on post mortem toxicology is higher than in past work. Three of eleven people with epilepsy (27%) in our study were prescribed AEDs that were tested and undetected compared to 1 of 23 (4%) in the series by Ryan and Dowling (1993). Ryan and Dowling also reported a high proportion of people had subtherapeutic AED levels. We included those with subtherapeutic levels in the “detected” category because evidence indicates that subtherapeutic post-mortem AED levels may not be a reliable indicator of subtherapeutic pre-mortem AED levels and thus should not be used to indicate nonadherence (Lathers et al., 2011; May et al., 1999). Consequently, it is unclear the role that subtherapeutic AED levels or nonadherence may have played in the drownings. In contrast to concerns regarding AED nonadherence and subtherapeutic AED levels, supratherapeutic AED levels may also be importantly linked to drowning. We documented one case of supratherapeutic AED levels (phenobarbital), which, while not elevated sufficiently to be the cause of death, was considered to contribute to the drowning, through central nervous system depression and impaired ability to stay afloat. It is unknown whether supratherapeutic AED levels may have been a contributing factor in other cases as well because
Table 2 Circumstances Surrounding Drowning.
Role of Seizure In Drowning as Per Pathologistb
AED Levels on Post Mortem Toxicology
d
Not Detected Not Tested or Unknown a) On AED(s) that can be tested b) On AED that cannot be tested (levetiracetam) c) Specific AED(s) unknown d) Not prescribed AED
This individual’s formal cause of death was given as “undetermined” because autopsy was limited to external examination (in accommodation with family’s wishes), however, the pathologist considered the scene and circumstances surrounding the death to be supportive of accidental drowning. b As per final cause of death, or pathologist’s summary/opinion. c Includes any individual with a trace, detected, subtherapeutic, therapeutic, or supratherapeutic AED level (see Discussion). d One individual on polytherapy had only some of the prescribed AEDs detected. This individual was included in the AED not detected group (considered a marker of non-adherence). a
sufficient to be the cause of death, including developmental delay, AED toxicity, blunt force facial trauma, and possible loss of consciousness due to trauma sustained as the individual entered the water. There were no cases where seizure was excluded as a cause of drowning, or where there was a clear cause of drowning other than seizure. There was no difference in age between bathtub drownings and natural water drownings (median age 33 vs. 31.5 years), however, bathtub drownings were more likely to be females (72% vs. 10%, p = 0.008). Body habitus was similar between drowning settings: 45% of bathtub drownings occurred in individuals who were overweight (n = 3) or obese (n = 2) while 40% of natural water drownings occurred in overweight individuals (n = 4), though none were obese. Complete body habitus information was unavailable in one natural water drowning. Toxicological analysis varied across cases. Testing of at least one AED was completed in 11 cases. Of these, 3 cases had undetected AED 125
Epilepsy Research 145 (2018) 123–126
E. Bain et al.
AEDs were often only reported as “detected” on toxicological testing reports. The role of a seizure in these drowning deaths is unclear. Beyond a few cases of witnessed or video-captured seizures, evidence of seizure may be inferred by the presence of a tongue bite identified at autopsy, which is not entirely specific. Alternatively, a seizure may occur and leave no signs at autopsy. Approximately one quarter of cases had evidence of a seizure around the time of drowning. In some cases, the potential role of seizures was further obscured by other case factors that may potentially contribute to drowning (e.g. cardiac disease, evidence of injury that may have led to loss of consciousness). The well-established psychiatric comorbidities, developmental delays, and social isolation (Camfield and Camfield, 2007) associated with epilepsy may increase the likelihood of drowning. Approximately one-third of our cohort had psychiatric comorbidities including bipolar disorder, schizophrenia, obsessive-compulsive disorder, anxiety, and others. The exact role this may have played in the events leading to drowning is not known. Drowning in people with epilepsy is markedly different from drowning in the general population. The estimated incidence of drowning reported here is approximately ten times greater than that quoted in the general population (1.5 per 10,000 people with epilepsy compared to 0.13 per 10,000 in the general population) (Lifesaving Society, 2016). In the general population, 60% of bathtub drowning deaths occur in those 65 years of age or older; in contrast, all bathtub drownings reported here occurred in those under 60. Alcohol is a significant risk factor for drowning and is implicated in approximately half of all drowning deaths in those ages 20–64 in the general population (Lifesaving Society, 2016), whereas only 2 people with epilepsy in the current study had evidence of alcohol consumption. This population-based study offers current insight into drowning in people with epilepsy, based on a comprehensive dataset collected in a province with a large population and diversity in culture, geography, and degree of urbanization. Screening of forensic autopsies is anticipated to have identified nearly all, if not all, people with epilepsy who drowned during the study period. Review of the full post mortem reports allowed for exclusion of people who suffered an alternative cause of death in the context of a water setting. Finally, in most cases, there were no other obvious potential contributors to drowning besides seizures, which allows for a clearer understanding of the role epilepsy may have played in these drownings. There are several limitations to this study. First, the study was retrospective in nature and information available was often limited, especially regarding the seizure history. A history of “epilepsy”, “seizure disorder”, or unprovoked “seizures” in the post mortem report was accepted as evidence of epilepsy, but, without a more detailed seizure history available, it cannot be certain that all subjects reported met the current clinical definition of epilepsy. Second, missed cases cannot be fully excluded. All drowning cases should be reported to the coroner and would be eligible for forensic autopsy, but it is possible that a case was missed (e.g. a drowning that was mistakenly not referred to a coroner, refused by a coroner, or not referred for autopsy). Furthermore, not all cases had toxicological testing and laboratory methods limit the ability to detect all illicit substances. It is possible illicit drugs may have been involved in additional cases. Finally, the incidence of drowning in people with epilepsy is estimated by calculation and is limited by the uncertainty of the exact number of people living in Ontario with epilepsy from 2014–2016. Underestimating the
prevalence of epilepsy would overestimate of the incidence of drowning in the epilepsy population, and conversely, overestimating the prevalence of epilepsy would underestimate the incidence of drowning in people with epilepsy. 5. Conclusion While the overall incidence of drowning in people with epilepsy is low, it represents a highly preventable cause of death. People with epilepsy should receive counseling on the increased risk of drowning, including information regarding the significant risk associated with bathtub use, the potential protective roles of supervision when in or around water and the fact that all people with epilepsy remain at an increased risk of drowning regardless of their apparent seizure control. It is not known how many and to what extent people with epilepsy receive counseling regarding drowning prevention. This is an area for future research. Improved awareness of the risk of drowning and regular education around preventative measures may help to prevent this devastating outcome of drowning in people with epilepsy. Disclosure of conflicts of interest The authors report no conflicts of interest to disclose. References Bell, G.S., Gaitatzis, A., Bell, C.L., Johnson, A.I., Sander, J.W., 2008. Drowning in people with epilepsy, how great is the risk? Neurology 71 (8), 578–582. Byard, R.W., 2017. Body mass index and saltwater drowning. Am. J. Forensic Med. Pathol. 38 (4), 304–305. Camfield, C.S., Camfield, P.R., 2007. Long‐term social outcomes for children with epilepsy. Epilepsia 48 (s9), 3–5. Diekema, D.S., Quan, L., Holt, V.L., 1993. Epilepsy as a risk factor for submersion injury in children. Pediatrics 91 (3), 612–616. Franklin, R.C., Pearn, J.H., Peden, A.E., 2017. Drowning fatalities in childhood: the role of pre-existing medical conditions. Arch Dis. Child. 102, 888–893. Gilmour, H., Ramage-Morin, P., Wong, S.L., 2016. Epilepsy in Canada: prevalence and impact. Heal Rep. 27 (9), 24–30. Kemp, A.M., Sibert, J.R., 1993. Epilepsy in children and the risk of drowning. Arch Dis. Child. 68 (5), 684–685. Lathers, C.M., Koehler, S.A., Wecht, C.H., Schraeder, P.L., 2011. Forensic antiepileptic drug levels in autopsy cases of epilepsy. Epilepsy Behav. 22 (4), 778–785. Lifesaving Society, 2016. Ontario Drowning Report 2016. (Accessed 12 July 2017). http://www.lifesaving.ca/wp-content/uploads/2013/09/Ontario-Drowning-Report2016.pdf. Mahony, A.J., Peden, A.E., Franklin, R.C., Pearn, J.H., Scarr, J., 2017. Fatal, unintentional drowning in older people: an assessment of the role of preexisting medical conditions. Healthy Aging Res. 6 (e7), 1–8. May, T., Jurgens, U., Rambeck, B., Schnabel, R., 1999. Comparison between premortem and postmortem serum concentrations of phenobarbital, phenytoin, carbamazepine, and its 10,11-epoxide metabolite in institutionalized patients with epilepsy. Epilepsy Res. 33 (1), 57–65. Orlowski, J.P., Rothner, A.D., Lueders, H., 1982. Submersion accidents in children with epilepsy. Am. J. Dis. Child. 136 (9), 777–780. Pearn, J.M., 1977. Epilepsy and drowning in childhood. BMJ 1 (6075), 1510. Peden, A.E., Franklin, R.C., Pearn, J.H., 2018. Unintentional fatal child drowning in the bath: a 12-year review (2002–2014). J. Paediatr. Child Health 54 (2), 153–159. Ryan, C.A., Dowling, G., 1993. Drowning deaths in people with epilepsy. Can. Med. Assoc. J. 148 (5), 781. Statistics Canada, 2016. Population by Year, by Province and Territory. (Accessed 3 August 2017. http://www.statcan.gc.ca/tables-tableaux/sum-som/l01/cst01/ demo02a-eng.htm. Thurman, D.J., Logroscino, G., Beghi, E., Hauser, W.A., Hesdorffer, D.C., Newton, C.R., Scorza, F.A., Sander, J.W., Tomson, T., 2017. The burden of premature mortality of epilepsy in high-income countries: a systematic review from the Mortality Task Force of the International League Against Epilepsy. Epilepsia 58 (1), 17–26.
126