`Blackwell Science, Inc.
`© International League Against Epilepsy
`
`Time Trends in Incidence, Mortality, and Case-Fatality after
`First Episode of Status Epilepticus
`
`*‡§¶Giancarlo Logroscino, *‡¶Dale C. Hesdorffer, \Gregory Cascino, **John F. Annegers, and
`*†‡¶W. Allen Hauser
`*Gertrude H. Sergievsky Center, and Department of †Neurology, Division of Epidemiology, ‡School of Public Health at
`Columbia University, New York, New York, U.S.A.; §Divisione di Neurologia, Ospedale Miulli Acquaviva, Bari, Italy;
`Departments of \Neurology and ¶Health Sciences Research, Mayo Clinic and Mayo Foundation, Rochester, Minnesota; and
`**School Of Public Health, University of Texas, Houston, Texas, U.S.A. (deceased)
`
`Summary: Purpose: Status epilepticus (SE) is a medical
`emergency associated with a high mortality. Clinical series
`have suggested that mortality after SE has decreased. No stud-
`ies have systematically examined trends in incidence, mortal-
`ity, and case fatality after SE in a well-defined population.
`Methods: All first episodes of SE receiving medical attention
`between January 1, 1935, and December 31, 1984, were ascer-
`tained through the Rochester Epidemiology Project Records-
`Linkage System and followed up until death or study
`termination (February 1, 1996). We calculated incidence rates
`in the 50-year period (1935–1984), while we considered mor-
`tality and case-fatality in the last 30-year period (1955–1984).
`Results: Incidence of SE increased over time to 18.1/100,000
`(1975 through 1984). The increase was related to an increased
`incidence in the elderly and to the advent of myoclonic SE after
`cardiac arrest, a condition not seen in the early decades. In the
`
`last decade, ∼16% of the incidence was due to myoclonic SE.
`The mortality rates increased from 3.6 per year in the decade
`1955–1965 to 4.0/100,000 per year between 1975 and 1984.
`The 30-day case-fatality (CF) was unchanged, although a trend
`toward improvement was shown after excluding myoclonic SE.
`Conclusions: Incidence and mortality rates of SE have in-
`creased in the last 30 years. Case fatality remained the same.
`The increased incidence and mortality are due to the occurrence
`in the last decade of myoclonic SE after cardiac arrest. The
`mortality in the elderly was twice that of the youngest age
`group, across all study periods. Changes in the age and cause
`distribution of SE over time are responsible for the stable sur-
`vivorship. There is improvement in survivorship in the last
`decade when myoclonic SE is excluded. Key Words: Status
`epilepticus—Mortality—Prognosis—Incidence—Case-fatality.
`
`Status epilepticus (SE) has been defined as a seizure or
`series of seizures lasting >30 min without intervening
`lucidity (1). SE is a common medical emergency asso-
`ciated with a high mortality (2–4). There are no pub-
`lished studies on time trends of incidence or mortality of
`SE. There is a perception that survivorship after SE has
`improved because of improved medical therapy. In clini-
`cal series, the reported mortality after SE has decreased
`over the past 50 years (2,5). Mortality has decreased in
`recent clinical series of children (5).
`We used the medical records linkage system of the
`Rochester Epidemiology Project to compare changes in
`incidence, case-fatality, and mortality after a first epi-
`sode of SE by using data from five sequential 10-year
`periods from 1935 through 1984.
`
`Revision accepted May 7, 2001.
`Address correspondence asnd reprint requests to Dr. W. A. Hauser at
`G.H. Sergievsky Center, 630 West168th Street, New York, NY 10032,
`U.S.A.
`
`METHODS
`The records of the Rochester Epidemiologic Project
`index all medical contacts for Rochester residents at the
`Mayo Clinic and at other medical facilities at which
`Rochester residents are likely to seek medical care. Mayo
`Clinic facilities have been the center of referral for medi-
`cal care in Rochester, Minnesota, since the beginning of
`this century. The register includes inpatient, outpatient,
`and home visits. We reviewed all records with a diag-
`nosis of acute symptomatic seizure, single unprovoked
`seizure, or epilepsy between January 1, 1935, and De-
`cember 31, 1984 (6). Moreover, we reviewed all records
`indicating a medical condition potentially linked to SE or
`seizure occurrence (stroke, metabolic encephalopathy,
`brain tumors, encephalitis, anoxic encephalopathy after
`cardiac arrest). This process permitted us to identify most
`first SE cases that came to medical attention among
`Rochester residents during the study period, all after
`1955.
`
`1031
`
`AQUESTIVE EXHIBIT 1105 Page 0001
`
`
`
`1032
`
`G. LOGROSCINO ET AL.
`
`Charts of each patient were reviewed by three of us
`(G.C., W.A.H., G.L.) to identify the occurrence of SE
`and to classify SE seizure type, duration, etiology, and to
`review the response to therapy.
`Status epilepticus seizure type
`Seizure type was categorized following criteria of the
`International League against Epilepsy (ILAE), and SE
`was identified as previously described (7,8). For the pres-
`ent analysis, we distinguished between motor (including
`generalized, secondarily generalized, and partial motor
`SE), nonmotor SE (including complex partial and ab-
`sence SE), and myoclonic SE. The aim of this categori-
`zation was to group together SE cases with a similar
`likelihood of ascertainment. Motor SE is much easier to
`identify and is more likely to come to medical attention
`than is nonmotor SE. Myoclonic SE is characterized by
`minor motor activity but was considered separately from
`motor and nonmotor SE because of the subtle character-
`istics of this seizure activity. In our classification, myo-
`clonic SE is probably equivalent
`to “subtle SE”
`described by other authors (9).
`Etiology
`Etiology was assigned into two broad categories based
`on our previous work and ILAE recommendations: acute
`symptomatic and unprovoked (including progressive
`symptomatic, remote symptomatic and idiopathic/
`cryptogenic) (5,7,8).
`The period of follow-up was terminated by death or
`the end of the study period (February 1996). We calcu-
`lated incidence for each of five 10-year intervals from
`1935 through 1984. There were several sources of medi-
`cal care in the community for whom records were not
`included in the record-linkage system prior to 1952. Be-
`cause this could have led to a systematic bias in case
`identification before that time for conditions such as SE,
`we calculated the age-specific mortality and CF only for
`the three 10-year intervals between 1955 and 1984.
`These are the decades used for analytic assessment in
`other studies of seizure disorders (10–13).
`Statistical methods
`We calculated the age-specific mortality rates as the
`number of deaths divided by the person-years at risk in
`the population. We used census data stratified by age and
`gender to determine the person-years at risk for each
`decade (14). Mortality and incidence rates were adjusted
`to the age and gender distribution of 1980 U.S. white
`population, using the direct method (15). We used pro-
`portional hazards modeling to calculate the crude and
`gender- and age-adjusted relative risk for mortality in the
`last two decades, using the 1955–1964 interval as decade
`of reference (16,17).
`Because more prompt medical treatment with antisei-
`zure medication (AED) could be responsible for changes
`
`Epilepsia, Vol. 42, No. 8, 2001
`
`in mortality (18), we compared the distribution of inter-
`val with use of AEDs of SE across decades.
`
`RESULTS
`During the period 1935–1984, we identified 356 first
`episodes of SE. We excluded from the analysis 47 cases
`younger than 5 years with fever and with no other ap-
`parent cause of SE (febrile SE). Febrile SE is an age-
`limited benign condition with no associated mortality
`(1). Therefore 309 episodes of incident afebrile SE were
`identified in the population of Rochester, Minnesota. De-
`mographics, distribution of etiologies, and crude and ad-
`justed incidence are shown in Table 1.
`Time trends in incidence
`We calculated the age-adjusted incidence of SE over
`each 10-year interval for the 50-year period (Table 1 and
`Fig. 1). In the first decade, there is probably underascer-
`tainment of cases. Therefore the incidence comparison
`with this decade would be misleading. Between 1945 and
`1954, the annual age-adjusted incidence of afebrile SE
`was 14.1/100,000. The adjusted incidence increased to
`18.1/100,000 in the last decade of the study (1975
`through 1984).
`Etiology
`The overall age-adjusted incidence of acute symptom-
`atic SE increased from 6.3/100,000 in the 1945–1954
`interval to 9.8/100,000 in the last decade (1975 through
`1984) (Fig. 1). For unprovoked SE, the adjusted inci-
`dence was stable: 7.9/100,000 in the 1945–1954 period
`and 8.3/100,000 in the 1975–1984 period.
`Age
`The age distribution of those affected by SE has
`changed over time: between 1945 and 1954, only 17% of
`subjects with SE were older than 65 years, whereas in-
`dividuals in this age range accounted for 50% of cases
`between 1975 and 1984. There was a threefold increase
`in incidence of SE in those 65 and older (Table 1).
`Seizure type
`Important changes occurred for seizure type. Better
`recognition of more subtle forms of SE could in part
`account for the increase in incidence. The incidence of
`motor SE almost doubled, increasing from 7.4 in the first
`decade (1935 through 1944) to 11.1 in the last decade
`(1975 through 1984). Nonmotor SE accounted for 3% of
`all cases identified between 1935 and 1944, 13% be-
`tween 1945 and 1954, 11% between 1955 and 1964, 15%
`between 1965 and 1974, and 13% between 1975 and
`1984 (x2 4 32; p 4 0.001). The incidence rate of non-
`motor SE increased fivefold from 0.4/100,000 in the first
`time interval (1935–1944) to 2.2/100,000 in the second
`(1945–1954) interval. The increase in incidence after
`
`AQUESTIVE EXHIBIT 1105 Page 0002
`
`
`
`MORTALITY, TIME TRENDS, AND STATUS EPILEPTICUS
`
`1033
`
`TABLE 1. Age distribution of the Rochester population, crude incidence, and cases of status
`epilepticus (SE) by decade
`Rochester population
`n (%)
`
`Incidence/100,000
`(crude)
`
`All SE
`n (%)
`
`Age by
`decade
`1935–1944
`6.5
`<65 yr
`18.9
`>65
`7.3
`Total
`8.0
`Adj I
`1945–1954
`13.1
`<65 yr
`29.2
`>65
`14.4
`Total
`14.1
`Adj I
`1955–1964
`11.7
`<65 yr
`73.5
`>65
`17.4
`Total
`15.2
`Adj I
`1965–1974
`8.6
`<65 yr
`47.8
`>65
`12.4
`Total
`11.0
`Adj I
`1975–1984
`11.7
`<65 yr
`96.1
`>65
`21
`Total
`18.1
`Adj I
`Adj I, Sex- and age-adjusted incidence in the U.S. white population, census 1980.
`
`Acute
`SE n
`
`Unprovoked
`SE n
`
`8
`0
`8
`3.5
`19
`2
`21
`6.3
`25
`14
`39
`8.5
`21
`14
`35
`5.9
`37
`27
`64
`9.8
`
`7
`3
`10
`4.5
`15
`5
`20
`7.9
`16
`12
`28
`6.7
`19
`10
`29
`5.1
`23
`32
`55
`8.3
`
`230,063 (94)
`15,906 (6)
`245,969
`
`258,452 (91)
`24,000 (9)
`285,452
`
`349,530 (94)
`35,351 (9)
`384,881
`
`466,694 (90)
`50,209 (10)
`516,903
`
`510,952 (89)
`61,374 (11)
`572,326
`
`15 (83)
`3 (17)
`18
`
`34 (83)
`7 (17)
`41
`
`41 (61)
`26 (39)
`67
`
`40 (62)
`24 (38)
`64
`
`60 (83)
`59 (50)
`119
`
`1945 was modest, reaching 2.7/100,000 in the last de-
`cade.
`The most relevant change over time was the advent of
`myoclonic SE, a SE subtype that was not present in the
`earliest decades. We identified no case of myoclonic SE
`before 1965, two cases in the 1965–1974 interval, and 18
`cases in the last decade (1975 through 1984). Myoclonic
`SE represented 16% of all cases during the 1975–1984
`interval. The adjusted incidence for myoclonic SE was
`4.3/100,000 in the last decade. The crude incidence of SE
`
`decreased from 20.8 to 17.6/100,000 after exclusion of
`myoclonic SE.
`Time trends in mortality
`We measured the 30-day age- and gender-adjusted
`mortality rate after an incident episode of SE for the last
`three decades (1955–1984) where ascertainment was
`probably thought to be homogeneous for all SE subtypes
`(Fig. 2). The mortality rate increased 10% in the 30-year
`interval from 3.6/100,000 between 1955 and 1964 to
`4.0/100,000 between 1975 and 1984 (Fig. 2).
`
`FIG. 1. Age-adjusted incidence of all status epilepticus, unpro-
`voked and acute symptomatic, during the period 1935–1984.
`
`FIG. 2. Age-adjusted mortality rates of all status epilepticus, un-
`provoked and acute symptomatic, during the period 1955–1984.
`
`Epilepsia, Vol. 42, No. 8, 2001
`AQUESTIVE EXHIBIT 1105 Page 0003
`
`
`
`1034
`
`G. LOGROSCINO ET AL.
`
`30-day case fatality
`The CF was high for those with acute symptomatic
`and low for those with unprovoked SE across all decades
`(Table 2). A Cox proportional hazard model, with age
`and time period included as covariates, showed no dif-
`ference when the CF in the decade 1975–1984 was com-
`pared with the decade 1955–1964 (RR 4 1.3; 95% CI,
`0.8–2.2).
`30-day case fatality and myoclonic SE
`CF in people with myoclonic status was 65%, more
`than twice the CF of the remaining subtypes of SE. Be-
`cause myoclonic SE did not exist in the community in
`earlier decades, we calculated CF of SE excluding the
`myoclonic cases. We thought this would provide a more
`appropriate comparison of the CF across time. After ex-
`clusion of myoclonic SE among the acute symptomatic
`cases, SE 30-day CF decreased from 37% in 1955–1964
`to 26% in the 1975–1984 decade (Table 2 and Fig. 3). In
`those older than 65 years, exclusion of myoclonic cases
`in the last decade reduced CF from 58% to 38% (Table
`2 and Fig. 4). Thus CF has decreased over time when we
`exclude myoclonic SE.
`30-Day case fatality and age
`The CF in the elderly was double that of young across
`the last three decades (Table 2).
`Therapy
`Information on interval to first treatment was available
`for 200 (91%) of the 220 subjects treated with AEDs.
`There appeared to be little change in the promptness of
`medical treatment over time for both motor and nonmo-
`tor SE across all the study period (x2 4 3.2; p 4 0.19).
`DISCUSSION
`It has been suggested that mortality after SE has de-
`clined over time, but there have been no studies of
`change in mortality rates after SE in a well-defined popu-
`lation. Disease-specific mortality rates are determined by
`
`TABLE 2. Case-fatality rate (%) after status epilepticus by
`etiology across three decades (1955–1984)
`Decades
`1955–1964
`1965–1974
`1975–1984
`All status
`25
`14
`24
`<65 yr
`17
`7
`17
`>65 yr
`38
`25
`32
`Unprovoked
`10
`0
`7
`<65 yr
`0
`0
`4
`>65 yr
`21
`0
`9
`Acute symptomatic
`37
`24
`38
`<65 yr
`27
`13
`24
`>65 yr
`58
`42
`59
`Acute symptomatic
`w/out MS
`37
`25
`26
`<65 yr
`28
`14
`18
`>65 yr
`58
`42
`38
`NS, not significant; MS, myoclonic status epilepticus.
`
`x2 (p)
`NS
`NS
`NS
`NS
`NS
`NS
`NS
`NS
`NS
`NS
`NS
`NS
`
`Epilepsia, Vol. 42, No. 8, 2001
`
`FIG. 3. Case fatality (%) after acute symptomatic status epilep-
`ticus (SE) in the period 1955–1984, with and without myoclonic
`SE after anoxic encephalopathy.
`
`incidence and CF. Our study suggests that there has been
`a modest increase in SE mortality rates in the study pe-
`riod. This increase is related to an increase in incidence
`of SE with a stable CF. The increase in incidence of SE
`occurred in all age groups but was most pronounced
`among those older than 65 years and was particularly
`evident during the period from 1975 through 1984. The
`increased incidence is due to a true overall increase of the
`incidence of SE among the elderly and related to the
`occurrence of a new condition, “myoclonic” SE, associ-
`ated with anoxic encephalopathy among survivors of car-
`diac arrest after 1965. The apparent change in incidence
`could be attributed, in part, to underascertainment of SE,
`particularly in the first decade. Even in 1935, a case of
`generalized SE would have come to medical attention. A
`case of nonmotor SE would have been less likely to be
`recognized as such and thus be less likely to receive
`prompt medical intervention. This is true for the entire
`study period but may have been a more important reason
`for underascertainment in the 1935–1944 interval, but
`less so in the last two decades. Accuracy of diagnosis
`could have improved over time because of EEG moni-
`
`FIG. 4. Age-specific (older and younger than 65 years) case
`fatality (%) after acute symptomatic status epilepticus (SE) in the
`period 1955–1984, with and without myoclonic SE after anoxic
`encephalopathy.
`
`AQUESTIVE EXHIBIT 1105 Page 0004
`
`
`
`MORTALITY, TIME TRENDS, AND STATUS EPILEPTICUS
`
`1035
`
`toring; this could be relevant especially for nonmotor SE,
`as reflected by their differential increase over time.
`The increase in mortality rate has paralleled the in-
`crease in incidence of SE. In contrast, the survivorship as
`measured by CF in the last three decades was stable.
`Cardiac arrest was probably universally fatal in all but
`last intervals of this study (19). SE after anoxic enceph-
`alopathy due to cardiac arrest did not occur in the earliest
`decades in Rochester. Improved survivorship after car-
`diac arrest is responsible for all myoclonic SE in this
`series, but the prognosis of these cases remains grave:
`two thirds in this subgroup died in the next 30 days.
`During the1975–1984 interval, myoclonic SE accounted
`for 28% of all acute symptomatic SE cases and 16% of
`all first SE cases. When CF is analyzed after exclusion of
`myoclonic SE, the CF improves over time.
`Age was also an important determinant of mortality
`and incidence: the incidence of SE among those 65 and
`older increased threefold between 1945 and 1944 and
`1975 and 1984. Because age is such an important pre-
`dictor of survivorship, the increase of the overall CF in
`the last interval is affected by the increased proportion of
`older subjects. In the elderly for all time intervals, CF
`after an incident episode of acute symptomatic SE was
`double that of the younger group. After the exclusion of
`myoclonic SE, we observed a general improvement in
`CF both in the young and in the elderly.
`The interval from onset of SE to initiation of AED
`treatment has not changed over time. Improvements in
`nonmyoclonic CF might be related to ancillary treatment
`and better management of respiratory and cardiac con-
`ditions.
`This study shows that there is an increasing incidence
`of SE due to myoclonic SE and to SE in the elderly. The
`unchanged CF is related to a new distribution of age and
`etiology of SE in the last decade.
`Acknowledgment: This study was supported in part by
`NINDS grant NS16308.
`REFERENCES
`1. Hauser W.A. Status epilepticus: epidemiologic considerations.
`Neurology 1990;40(suppl 2):9–13.
`
`2. Hesdorffer D, Logroscino G, Cascino G, et al. Incidence of status
`epilepticus in Rochester, Minnesota, 1965-84. Neurology 1998;50:
`735–41.
`3. DeLorenzo RJ, Hauser WA, Towne AR, et al. A prospective
`population-based epidemiological study of status epilepticus in
`Richmond, Virginia. Neurology 1996;46:1029–35.
`4. Logroscino G, Hesdorffer DC, Cascino G, et al. Short-term mor-
`tality after a first episode of status epilepticus. Epilepsia 1997;38:
`1344–9
`5. Philips SA, Shanahan RJ. Etiology and mortality of status epilep-
`ticus in children: a recent update. Arch Neurol 1989;46:74–6.
`6. Hauser WA, Annegers JF, Kurland LT. Prevalence of epilepsy in
`Rochester, Minnesota, 1940-1980. Epilepsia 1991;32:429–45.
`7. Commission on Epidemiology and Prognosis, International League
`Against Epilepsy. Guidelines for epidemiologic studies on epi-
`lepsy. Epilepsia 1993;34:592–6.
`8. Commission on Classification and Terminology of the Interna-
`tional League Against Epilepsy. Proposal for revised clinical and
`electroencephalographic classification of epileptic seizure. Epilep-
`sia 1981;22:489–501.
`9. Treiman D. Generalized convulsive status epilepticus in the adult.
`Epilepsia 1993;34(suppl 1):S2–11.
`10. Hesdorffer DC, Hauser WA, Annegers JF, et al. Severe uncon-
`trolled hypertension and adult-onset seizures: a case study in Roch-
`ester, Minnesota. Epilepsia 1996;37:736–41.
`11. Annegers JF, Hauser WA, Lee JR, et al. Incidence of acute symp-
`tomatic seizures in Rochester, Minnesota, 1935-1984. Epilepsia
`1995;36:327–33.
`12. Hesdorffer DC, Hauser WA, Annegers JF, et al. Dementia and
`adult onset unprovoked seizures. Neurology 1996;46:727–30.
`13. Hesdorffer DC, Hauser WA, Annegers JF, et al. Major depression
`is a risk factor for seizures in older adults. Ann Neurol 2000;47:
`246–9.
`14. Schroeder DJ, Offord KP. A SAS macro which utilizes local and
`refernce population counts appropriate for incidence, prevalence,
`and mortality rate calculations in Rochester and Olmsted County,
`Minnesota. Technical Report Series, 20. Mayo Clinic, Rochester,
`MN: 1982;6.
`15. Kleinbaum DG, Kupper LL, Morgenstern H. Epidemiologic re-
`search: principles and quantitative methods. New York: Van Nos-
`trand Reinhold, 1982:122–30.
`16. Rothman KJ. Modern epidemiology. New York: Little, Brown,
`1986:38–40.
`17. Vine M, Schoenbach V, Hulka B, et al. Atypical metaplasia and
`incidence of bronchogenic carcinoma. Am J Epidemiol 1990;131:
`781–92.
`18. Collett D. Modelling survival data in medical research. New York:
`Chapman & Hall, 1994:15–85.
`19. Goldberg RJ, Gore JM, Haffajee CI, et al. Outcome after cardiac
`arrest during acute myocardial infarction. Am J Cardiol 1987;59:
`251–5.
`
`Epilepsia, Vol. 42, No. 8, 2001
`AQUESTIVE EXHIBIT 1105 Page 0005
`
`