Mortality

Samden D. Lhatoo

Josemir W. Sander

Törbjorn Tomson

Introduction

The majority of patients with epilepsy have a good prognosis that allows them long periods of seizure freedom.⁹^,¹⁰^,⁴³ Despite this, it is known that the mortality rate in epilepsy is raised to two to three times that of the general population⁸^,¹⁹^,²⁸^,²⁹^,³²^,³⁹^,⁵³ and life expectancy in some of these patients is reduced.¹⁷ Studies in the last two decades have thrown considerable light on the various factors that influence mortality and also on the various causes of mortality in epilepsy in the developed world.⁸^,¹⁹^,²⁸^,²⁹^,³²^,³⁹^,⁵³ However, there are no large-scale studies of mortality in epilepsy from resource-poor countries.⁷

There are many problems inherent to the study of mortality in epilepsy, including the problems of case ascertainment. Death certificates have been found to be an unreliable source of case ascertainment,³ and prospective studies of cohorts of people with epilepsy are more likely to provide accurate data on mortality.

Several aspects of epilepsy may affect mortality:

Seizures themselves can be a cause of death, either directly as in prolonged status epilepticus or indirectly because of increased risk of accidental death, especially drowning.
Some of the risk factors for epilepsy (e.g., brain tumors, cerebrovascular disease, traumatic brain injury) are associated with increased mortality whether or not epilepsy is present.
There is an increased risk for sudden unexpected death in epilepsy (SUDEP).
The long-term use of antiepileptic drugs (AEDs) has been thought to increase the incidence of malignant neoplasia and osteoporosis, thereby potentially affecting long-term mortality rates in persons with epilepsy, especially in younger adults (ages 15 to 49 years). Although rare, other adverse effects of idiosyncratic type can also occasionally be fatal.

The objectives of this chapter are to review the evidence for, and patterns of, mortality in persons with epilepsy and to evaluate the current understanding of these factors.

Overall Mortality

The most often used and preferred statistical measure of epilepsy mortality is the standardized mortality ratio (SMR). This is the ratio of the number of observed deaths in a study population to the expected number of deaths in the age- and sex-matched general population in that time period. Older studies utilize the proportional mortality ratio (PMR), which expresses mortality as a ratio of the number of observed deaths due to a particular cause in a study population to the total number of deaths in that study population. This method is subject to more bias, depending on the kind of cohort studied, where deaths due to any one cause may be overrepresented, and is not a measure of the mortality rate in epilepsy versus the general population. SMRs among people with epilepsy have been evaluated in several population-based studies (Table 1) and vary from 1.6 to 8.8. Although these studies span several decades and countries, follow incidence or prevalence cases, and vary in terms of the age composition and characteristics of the patients studied, most are consistent in reporting a twofold to threefold increased mortality rate for people with epilepsy. Similarities between the population-based epidemiologic studies and the more selective hospital-based studies are not truly consistent with each other but probably the result of compensating biases. It should also be noted that the overall relative increased premature mortality of twofold to threefold should not be applied to all people with epilepsy; this is a summary measure that masks very important differences among those with epilepsy. The most important of these are the etiology of epilepsy, age, duration, and type of epilepsy.

Etiology of Epilepsy

The increased mortality rates of epilepsy reflect both the effects of epilepsy and also the effects of central nervous system (CNS) insults, such as brain tumors, cerebrovascular disease, and head trauma, which are the presumed cause of approximately one third of all epilepsy cases.³³ In epidemiologic studies, seizure disorders have often been classified into four broad groups according to presumed etiology:

Idiopathic (or cryptogenic) epilepsy.
Remote symptomatic epilepsy, related to CNS lesions acquired postnatally from trauma, brain tumors, cerebrovascular disease, infection, or chronic degeneration.
Acute symptomatic seizures, related to CNS insults where seizures manifest within a week of the acute insult (e.g., head injury, cerebral hemorrhage, or infarction).
Major neurologic dysfunction of uncertain cause but presumed to have been present at birth and manifested by gross neurologic deficit (spasticity, hemiparesis), or mental retardation (IQ <70).

Table 1 Overall Mortality in Epilepsy—Standardized Mortality Ratios (SMR)

Author	Country	Study design	Age	Follow-up years	SMR (95% CI)
Population-based studies
Zielinski 1974⁵³	Poland	Prevalence	All		1.8 (1.6–2.1)
Hauser 1980¹⁹	United States	Retrospective	All	33	2.3 (1.9–2.6)
Cockerell 1994⁸	United Kingdom	Prospective	All	9	2.5 (2.1–2.9)
Olafsson 1998³⁹	Iceland	Retrospective	All	30	1.6 (1.2–2.2)
Lindsten 2000²⁹	Sweden	Prospective	Adult	11	2.5 (1.2–3.2)
Lhatoo 2001²⁸	United Kingdom	Prospective	All	14	2.1 (1.8–2.4)
Camfield 2002⁶	Canada	Retrospective	Children	20	5.3 (2.29–8.32) First 10 years 8.8 (4.10–13.4) Second 10 years
Berg 2004⁴	United States	Prospective	Children	8	7.54 (4.38–12.99)
Hospital-based studies
Klenerman 1993²²	United Kingdom	Retrospective	Adults		1.9 (1.6–2.3)
Nilsson 1997³⁸	Sweden	Retrospective	Adults	>16	3.6 (3.5–3.7)
Shackleton 1999⁴⁴	Netherlands	Retrospective	All	30	3.2 (2.9–3.5)
Callenbach 2001⁵	Netherlands	Clinical series	Children	5	6.6 (2.2–15.5) Males
					7.4 (2.0–19.0) Females
CI, confidence interval; SMR, standardized mortality ratio. SMR = Observed deaths/expected deaths.

Table 2 Standardized Mortality Ratios (with 95% Confidence Intervals) According to Etiology

Country	Idiopathic	Remote symptomatic	Acute symptomatic	Neurodeficit
United States 1980¹⁹	1.8 (1.4–2.3)	2.2 (1.8–2.7)		11.0 (6.9–16.4)
United Kingdom 1994⁸	1.6 (1.0–2.4)	4.3 (3.3–5.5)		50.0 (10–146)
France 1999³²	1.5 (0.4–3.9)	6.5 (3.8–10.5)
Sweden 2000²⁹	1.1 (0.5–2.4)	3.3 (2.4–4.5)
United Kingdom 2001²⁸	1.3 (0.9–1.9)	3.7 (2.9–4.6)	3.0 (2.0–4.3)	25 (5.1–73.1)
United States 2004⁴	1.43 (0.36–5.73)	33.46 (18.53–60.43)
NB—idiopathic, idiopathic or cryptogenic or nonsymptomatic.

The level of increased premature mortality appears greatest in patients with epilepsy in association with a neurologic deficit present from birth. Such patients had a considerably high SMR of 7.0 (95% confidence interval [CI], 4.6 to 10.2) in the Rochester, Minnesota, study¹⁹ and of 25 (95% CI, 5.1 to 73.1) in the UK General Practice Study (although this latter SMR was based on only a few deaths).²⁸ The extremely high SMR in the neurologic deficit group reflects the large number of observed deaths at relatively young ages, when few deaths are expected. Mortality is also high in individuals with symptomatic epilepsy compared to those in other groups. In the Rochester study,¹⁹ persons with epilepsy as a result of postnatal CNS insults had an SMR of 2.76 (95% CI, 2.3 to 3.4). In the UK study, the SMR for postnatally acquired epilepsy was 3.7 (95% CI, 2.9 to 4.6).²⁸ The higher relative mortality in the British study may
be a consequence of the comparatively shorter mean follow-up period, as SMRs decline with duration.

Studies that present SMRs for all patients and separate SMRs for persons with idiopathic (or cryptogenic) epilepsy (epilepsy without a presumed cause) show considerably lower SMRs for patients with idiopathic epilepsy than for all patients with epilepsy. The mortality rates among those with idiopathic epilepsy from Rochester,¹⁹ Sweden,²⁹ and the UK General Practice Study²⁸ are still raised above those of the general population, but only by 50% to 60% and not significantly so in the latter two studies (Table 2). The UK General Practice Study also studied patients with acute symptomatic seizures as a separate category as defined above and found an SMR of 3.0 (95% CI, 2.0 to 4.3) in this group.²⁸

Age

The population-based studies that present age-specific SMRs are shown in Table 3. Mortality rates are increased at all ages among persons with epilepsy, but the increase is not uniform across age groups. There is a sharp decline in SMRs with age in these study populations. The higher SMRs in persons under 50 years of age are caused in part by the extremely high relative death rates for patients with epilepsy in association with neurologic deficits and the low expected mortality in these age groups. However, in the UK General Practice Study, the 50- to 59-year age group had higher mortality rates than all other age groups, and this may be explained by the high incidence of brain tumors in this population.²⁸ Head trauma in younger patients may be another contributory factor. Another factor is the increased risk for sudden death in younger adults with epilepsy; it is uncertain whether this increased risk persists among elderly patients with epilepsy. Although the SMRs decrease with age, the death rates among the elderly (>65 years of age) with epilepsy are still 50% to 60% above the rates in the general population and remain significantly elevated even in patients over the age of 80 years.¹⁹

Table 3 Standardized Mortality Ratios (SMRs) According to Age

Age (y)	SMRs (95% CI)
Age (y)	Poland	United States¹⁹	United Kingdom²⁸	Sweden²⁹
0–24		8.5 (5.4–12.9)
0–29	3.5
0–49			7.6 (4.2–12.5)
15–39				9.5 (3.1–29.4)
25–44		7.7 (5.1–11.0)
30–49	3.4
45–54		3.5 (2.0–5.7)
40–59				10.7 (6.5–17.6)
50–59	2.5		8.6 (4.7–14.1)
55–64		3.0 (2.0–4.5)
60–69	1.8
60–79				2.4 (1.6–3.8)
65–74		1.5 (1.0–2.2)
70–79			1.9 (1.2–2.8)
≥70	1.5
≥75		1.4 (1.1–1.9)
≥80			2.6 (1.8–3.6)	1.3 (0.7–2.4)
CI, confidence interval.