Highlights
- •Older patients (≥60 yrs) represent the predominant subgroup in status epilepticus.
- •There was no age-dependent predominance for new onset status epilepticus (NOSE).
- •Older patients with NOSE had comparable outcome rates to younger patients.
- •Infections were associated with unfavourable outcome of NOSE in older patients.
Abstract
Purpose
We here evaluated (1) the differential characteristics of status epilepticus (SE) in older (≥60 years) compared to younger adults (18–59 years). In particular, we were interested in (2) the proportion and characteristics of new onset SE in patients with no history of epilepsy (NOSE) in older compared to younger adults, and (3) predictive parameters for clinical outcome in older subjects with NOSE.
Methods
We performed a monocentric retrospective analysis of all adult patients (≥18 years) admitted with SE to our tertiary care centre over a period of 10 years (2006–2015) to evaluate clinical characteristics and short-time outcome at discharge.
Results
One-hundred-thirty-five patients with SE were included in the study. Mean age at onset was 64 years (range 21–90), eighty-seven of the patients (64%) were older than 60 years. In 76 patients (56%), SE occurred as NOSE, sixty-seven percent of them were aged ≥60 years. There was no age-dependent predominance for NOSE. NOSE was not a relevant outcome predictor, especially regarding age-related subgroups. Older patients with NOSE had less frequently general tonic clonic SE (GTCSE; p = 0.001) and were more often female (p = 0.01). Regarding outcome parameters and risk factors in older patients with NOSE, unfavourable outcome was associated with infections during in-hospital treatment (0.04), extended stay in ICU (p = 0.001), and generally in hospital (p < 0.001).
Conclusion
In our cohort, older patients represented the predominant subgroup in patients with SE. Older patients suffered more often from non-convulsive semiology and had a less favourable short-time outcome. NOSE was not a predictive outcome parameter in older patients. Data suggest that avoiding infections should have a priority because higher infection rates were associated with unfavourable outcome.
Keywords
1. Introduction
Status epilepticus (SE) is a frequent and severe neurological emergency condition [
1
, 2
] with high mortality and morbidity. Mortality has been described in up to 20%, morbidity in up to 39% of cases [[3]
].Regarding prevalence and incidence of SE, there is a first peak in early childhood and a second peak in older subjects, which is often defined as ≥ 60 years [
4
, 5
]. The mean annual incidence rate is estimated to approximate 15.5/100,000 in patients between 60 and 69 years old, 21.5/100,000 in patients between 70 and 79 years old, and 24.9/100,000 in patients more than 80 years old [6
, 7
, 8
].In adults with SE, withdrawal of antiepileptic drugs (AED) or an acute symptomatic etiology, i.e. mainly acute stroke, were reported as the predominant causes [
5
, 9
]. In half of first episodes of SE, patients do not have any history of prior seizures or known epilepsy [[8]
]. These patients are at high risk for developing new-onset refractory status epilepticus (in up to 50% of the cases), resulting in a poor outcome [[10]
].To date only few studies exist, which focused on new onset SE (NOSE) in older patients. Of note, in this context “older” is not defined unequivocally. According to epidemiological data from Hesdorffer and colleagues [
[7]
], we here employ a cut-off of ≥ 60 years for “older” patients.We aimed at evaluating (1) characteristic differences of SE in older (≥60 years) compared to younger patients (18–59 years), (2) the proportion and characteristics of NOSE in older compared to younger patients, and (3) predictive parameters for clinical outcome in older people with SE, especially with NOSE.
2. Methods
2.1 Patient cohort
We included all adult patients (≥18 years) admitted with SE to the Department of Neurology, University Hospital Cologne, an academic tertiary care centre, over a period of 10 years (2006–2015). If patients had multiple episodes of SE in the observational period, only the first episode was included for data analysis. Patients were identified using the electronic database of the University Hospital Cologne with documented ICD-10 diagnoses of SE (G41.X), diagnosed both as primary or secondary diagnosis.
Patients were assigned to either the younger or older subgroups, with “older” being defined as ≥60 years, and “younger” being defined as aged between 18 and 59 years.
SE was defined according to previous recommendations [
11
, 12
] as a prolonged seizure lasting i) more than five minutes for generalized tonic-clonic seizures, and ii) more than 10 min for all other seizures, or iii) as a series of at least two seizures without regaining inbetween the previously known state of consciousness.Diagnosis of SE was made clinically if an epileptic semiology was unambiguously present. In case of unspecific signs and symptoms, clinical and ictal EEG features were required as proposed earlier [
[13]
].Termination of SE was assumed if seizure activity stopped clinically in alert patients or if the EEG was free of any evidence of ictal activity in intubated patients or those with persistent altered state of consciousness.
Semiology of SE was classified as generalized tonic clonic SE (GTCSE), if bilateral convulsions were observed. Non-convulsive SE was classified as complex partial SE, if a corresponding semiology was present, and as non-convulsive with coma, if severe loss of consciousness was the prominent symptom without evidence of generalized convulsions at the beginning with EEG features suggestive for constant ictal activity. Simple partial SE was assumed if no disturbance of consciousness was present.
Aetiology was classified according to the corresponding ILAE-definition: [
[14]
] acute symptomatic, remote symptomatic, progressive symptomatic and cryptogenic. Patients with SE and known idiopathic or cryptogenic epilepsy without evidence for an acute seizure provocation were grouped as remote symptomatic. Patients without history of epilepsy in whom no cause for SE could be found despite extensive diagnostics, were determined cryptogenic.We excluded patients with posthypoxic myoclonic encephalopathy and antibody-associated encephalitis because of different pathogenic mechanisms involved. We further excluded, as stated above, any patient with a severe but unspecific EEG pattern, without clear epileptic semiology, and without improvement under therapy (because of uncertainty of the underlying condition). Patients from other clinics were only included if transferred immediately after onset of SE because of relevant data loss regarding treatment, semiology, and initial diagnostics, as well a referral bias with a more protracted disease course could be assumed.
Adequate treatment regimens and sequences were adopted from current guidelines [
3
, 15
, 16
]. Of note, in Germany, prehospital patient care is provided by emergency doctors, so prehospital treatment decisions were at their discretion. As a consequence and due to the retrospective non-invasive design of our study, the below mentioned treatment scheme could not be adhered to all patients. Our in-house treatment protocol implemented a staged approach. Proposed drugs in step-wise order were benzodiazepines, followed by AED and by anesthetics. According to current guidelines and expert recommendations, treatment with benzodiazepines was assumed adequate if given intravenously in doses at the following minimum: midazolam 10 mg for >40 kg, Clonazepam 1 mg, Diazepam 10 mg (equivalent 0.15 mg/kg, max 10 mg/dose), Lorazepam 2 mg (equivalent 0.05 mg/kg/dose). AED doses were considered adequate if given intravenously at the following minimum: phenytoin 20 mg/kg, levetiracetam >30 mg/kg, valproate >20 mg/kg, lacosamide 200–400 mg. Of note, lacosamide was introduced in market since 2008. Anesthetics were considered adequate if applied at the following minimum: midazolam (bolus 0.2 mg/kg, i.v. infusion rate 0.1–0.5 mg/kg/h), propofol (bolus 2 mg/kg, i.v. infusion rate 4–10 mg/kg/h), thiopental (bolus: 5 mg/kg, i.v. infusion rate 3–7 mg/kg/h), phenobarbital (bolus 20 mg/kg, i.v. infusion rate max 100 mg/min), and isoflurane (endtidal concentration 1.2–5%).We further assessed the duration of the SE using <1 h, ≥1h–24 h, and ≥24 h as cut-offs.
SE was supposed to be refractory if adequate first- and second-line therapy regimens failed.
Favourable outcome was present if modified ranking scale (mRS) at discharge did not deteriorate compared to preclinical status, or if mRS was ≤2 at discharge. Fatal outcome was defined as in-hospital mortality or in case of discharge under strictly palliative supportive care with expected short-time mortality.
Similar to Rossetti and co-workers [
[17]
], we assessed the following comorbidities as potential independent predictors for clinical outcome in SE: acute large vessel ischemic stroke, acute cerebral hemorrhage, acute central nervous system infection, malignant brain tumors, AIDS with CNS complications, chronic or acute renal insufficiency requiring dialysis, systemic vasculitis, metabolic disturbances or acute intoxication, sufficient to cause coma in absence of SE, eclampsia, and intracranial tumor surgery. In addition, we evaluated dementia in older patients as a potential risk factor for SE or NOSE. Infections as relevant complications were also collected as possible independent outcome variables in SE.There are two proposed scores for poor outcome prediction in SE, defined as in-hospital death: the Epidemiology-Based Mortality Score in Status Epilepticus (EMSE), and the Status Epilepticus Severity Score (STESS) [
17
, 18
, 19
].In STESS, four parameters have to be calculated (level of consciousness, worst seizure semiology, age at onset, history of seizures). Sum scores ≥3 (STESS-3) or ≥4 (STESS-4) were shown to predict in-hospital death [
[20]
]. The EMSE uses etiology, comorbidity, age, and EEG as predictive parameters (EMSE-EACE). Cut-off for short-time mortality was shown to be ≥64 points [[18]
].This is a retrospective study not involving primary research but only routine diagnostics, all patients are reported anonymously. Therefore specific ethics committee approval was waived.
2.2 Statistics
Statistical analyses were performed using SPSS 23.0 for Windows (IBM, Armonk, New York, USA). For comparisons of categorical independent data, chi-square test or Fisher’s exact (if less than 5 items) were performed; for comparisons of independent metrical data, the t-test for unpaired variables was performed. Binary logistic regression was utilized in order to evaluate independent covariates for outcome of NOSE in older patient comparisons; categorical explanatory variables were therefore transformed into dummy variables. For STESS-3/STESS-4 and EMSE-AECE scores, we calculated sensitivities and specificities as well as positive and negative predictive values. All tests were performed two-tailed. P-values <0.05 were estimated as significant.
3. Results
We reviewed a total of 315 patient files with ICD-10 diagnosis of SE. Diagnosis of SE according to predefined study criteria could be confirmed in 181 cases. Thirty-one of these had been referred subacutely from other institutions and were therefore excluded from further analysis. From the remaining 150 cases of SE, eight patients had multiple episodes (range: 2–6 episodes). Thus, one-hundred-thirty-five patients were included finally in the study. The selection process is depicted in detail in Supplementary Fig. S1, inclusion rates per year are given in Supplementary Table S1.
In the overall group of SE, mean age at onset was 64 years (range 21–90), eighty-seven out of the 135 patients were older than 60 years (64%). In 76 patients (56%), SE occurred as NOSE. Most of them were of acute symptomatic etiology, details are given in Table 1. Comorbidities were identified in 26 patients (19%), sixty-nine percent of these comorbidities occurred in older patients (N = 18). Comorbidities were as follows: acute stroke (N = 2; in older patients: N = 1), acute cerebral hemorrhage (N = 2; N = 2), malignant brain tumor (N = 17; N = 12), acute renal failure (N = 1; N = 0), acute CNS-infection (N = 3; N = 3), severe metabolic disturbances (N = 1; N = 0). Infection as complication was seen in 54 patients with SE (40%) without significant difference between the two subgroups. Overall outcome in SE was favourable in 73/135 cases (54.1%) and fatal in 17/135 cases (12.6%).
Table 1Etiology of SE in all patients and of NOSE in older and younger patients.
| All SE (N = 135) N (%) | NOSE in older patients (N = 51) N (%) | NOSE in younger patients (N = 25) N (%) | |
|---|---|---|---|
| Symptomatic | 122 (90%) | 40 (78%) | 24 (96%) |
| Acute symptomatic | 56 (41%) | 23 (45%) | 16 (64%) |
| Drug withdrawal | 10 (7%) | – | – |
| Alcohol abuse | 13 (10%) | 4 (8%) | 7 (28%) |
| Acute CNS infection | 5 (4%) | 2 (4%) | 2 (8%) |
| Metabolic/toxic | 10 (7%) | 4 (8%) | 4 (16%) |
| Acute cerebral bleeding | 6 (4%) | 5 (10%) | 0 (0%) |
| Acute cerebral infarction | 2 (1%) | 2 (4%) | 0 (0%) |
| New cerebral neoplasm | 9 (7%) | 5 (10%) | 3 (12%) |
| Acute head injury | 1 (1%) | 1 (2%) | 0 (0%) |
| Remote symptomatic | 51 (38%) | 14 (27%) | 4 (16%) |
| Infarction in the past | 24 (18%) | 11 (21%) | 3 (12%) |
| Cerebral bleeding in the past | 5 (4%) | 2 (4%) | 0 (0%) |
| Residual infection | 1 (1%) | 0 (0%) | 0 (0%) |
| Head injury in the past | 7 (5%) | 1 (2%) | 1 (4%) |
| Perinatal encephalopathy | 3 (2%) | 0 (0%) | 0 (0%) |
| Known epilepsy | 11 (8%) | – | – |
| Progressive symptomatic | 15 (11%) | 3 (6%) | 3 (12%) |
| Known CNS tumor | 13 (10%) | 3 (6%) | 3 (12%) |
| Multiple sclerosis | 2 (1%) | 0 (0%) | 0 (0%) |
| Cryptogenic (N = 11) | 13 (10%) | 11 (22%) | 2 (8%) |
Bold values means "significant", i.e. p-value <0.05.
Several differences between older and younger patients with SE were observed. Older patients had less frequently a favourable outcome compared to younger patients (45% vs 71%, p = 0.004) and, vice versa, more frequently outcome was fatal in older patients (18% vs. 2% in younger patients, p = 0.006). In older patients, semiology of SE differed from younger patients (p = 0.002): SE manifested less frequently as GTCSE compared to younger adults (33% vs. 63%, p = 0.001), on the other side, non-convulsive SE with coma exclusively occurred in older patients (10.3% vs 0%, p = 0.02). Females were predominant in the older patient group (64% vs. 42%, p = 0.01). Therapeutic regimes did not differ between the two groups. For further details of demographic and clinical characteristics in all SE, see Table 2.
Table 2SE in older and younger patients.
| Older patients (N = 87) N(%) | Younger patients (N = 48) N(%) | P | |
|---|---|---|---|
| New onset SE | 51 (59%) | 25 (52%) | 0.48 |
| Favourable outcome | 39 (45%) | 34 (71%) | 0.004 |
| Fatal outcome | 16 (18%) | 1 (2%) | 0.006 |
| Female | 56 (64%) | 20 (42%) | 0.01 |
| Semiology | 0.002 | ||
| GTCSE | 29 (33.3%) | 30 (63%) | |
| CPSE | 41 (47.1%) | 14 (29%) | |
| CPSE with coma | 9 (10.3%) | 0 (0%) | |
| SPSE | 8 (9.2%) | 4 (8%) | |
| Etiology | 0.43 | ||
| Symptomatic | 76 (87%) | 46 (96%) | |
| Acute | 34 (39%) | 22 (46%) | |
| remote | 33 (38%) | 18 (38%) | |
| Progressive | 9 (10%) | 6 (14%) | |
| Cryptogenic | 11 (11%) | 2 (4%) | |
| Acute comorbidity | 15 (17%) | 5 (10%) | 0.32 |
| Therapy | |||
| First drug: | 0.06 | ||
| BDZ | 71 (81%) | 45 (94%) | |
| AED | 12 (14%) | 1 (2%) | |
| anesthetics | 4 (5%) | 2 (4%) | |
| Dose of first drug correct | 62 (71%) | 34 (71%) | 1 |
| Therapy sequence correct | 59 (68%) | 36 (75%) | 0.44 |
| Refractory SE | 44 (51%) | 20 (42%) | 0.37 |
| Duration of SE | 0.50 | ||
| <1 h | 23 (26%) | 17 (35%) | |
| 1 h–24 h | 35 (40%) | 15 (32%) | |
| >24 h | 29 (34%) | 16 (33%) | |
| ICU | 85 (97%) | 45 (94%) | 0.35 |
| Mean duration ICU in days [SD] | 7.5 [10.7] | 7.6 [15.2] | 0.95 |
| Intubation | 33 (38%) | 19 (40%) | 0.90 |
| Infections in the course | 39 (45%) | 15 (31%) | 0.14 |
| Mean stay hospital [SD] | 14.7 [12.5] | 12.1 [15.9] | 0.30 |
SE: status epilepticus, GTCSE: generalized tonic clonic SE; CPSE: complex partial SE; SPSE: simple partial SE; BDZ: benzodiazepine; AED: antiepileptic drugs.
Bold values means "significant", i.e. p-value <0.05.
Regarding outcome scores for short-time mortality, sensitivity of STESS-3/STESS-4 was 77%/47%, whereas specificity was 37%/74%. Positive predictive values (PPV) for STESS-3/STESS-4 were 15%/21% and negative predictive values (NPV) were 92% and 91%, respectively. For EMSE-AECE, sensitivity/specificity values for cut-off value ≥64 were 64.7%/51.7%, PPV/NPV were 16%/91%.
NOSE occurred in 56% of all patients with median age at onset of 72 years (range 23–90) without difference regarding age-related subgroups (59% of older vs. 52% of younger patients, p = 0.48). Most NOSE were of acute symptomatic origin (52%), followed in decreasing manner by remote symptomatic origin (24%), progressive etiology (7%), and cryptogenic etiology (17%). Regarding the proportion of dementia in older patients, we did not find different prevalences in those with known epilepsy and NOSE: nine out of 51 patients with NOSE (18%) had dementia, whereas 7/26 in SE with known epilepsy (27%) had, but this was without significance in group comparison (p = 0.34). For details in etiology of SE in the whole group and of NOSE see Table 1.
When comparing older and younger patients with NOSE, again, older patients were more often females (69% vs. 36%, p = 0.01) and had different semiology (p = 0.004): GTCSE was less frequent (27% vs. 68%, p = 0.001), whereas differences in non-convulsive SE with coma were no longer significant (12% vs. 0%, p = 0.07). Regarding treatment, no differences were observed. Short-time outcome in patients with NOSE was not different for the age related subgroups: a favourable outcome was seen in 39% of the older and 60% of the younger patients (p = 0.14), a fatal outcome was seen in 18% and 5%, respectively (p = 0.15). For further clinical details of NOSE in older and younger patients, see Table 3.
Table 3Differences in NOSE of older (≥60 years) compared to younger (<60 years) patients.
| Older patients N = 51 | Younger patients N = 25 | P | |
|---|---|---|---|
| Favourable outcome | 20 (39%) | 15 (60%) | 0.14 |
| Fatal outcome | 9 (18%) | 1 (4%) | 0.15 |
| Female | 35 (69%) | 9 (36%) | 0.01 |
| Semiology | 0.004 | ||
| GTCSE | 14 (27%) | 17 (68%) | |
| CPSE | 27 (53%) | 6 (24%) | |
| CPSE with coma | 6 (12%) | 0 (0%) | |
| SPSE | 4 (8%) | 2 (8%) | |
| Etiology | 0.16 | ||
| Symptomatic | 40 (78%) | 23 (92%) | |
| Acute | 24 (47%) | 16 (64%) | |
| remote | 14 (27%) | 4 (16%) | |
| Progressive | 2 (4%) | 3 (12%) | |
| Cryptogenic | 11 (22%) | 2 (8%) | |
| Acute comorbidity | 11 (22%) | 4 (16%) | 0.76 |
| Therapy | |||
| First drug: | 0.06 | ||
| BDZ | 38 (75%) | 23 (92%) | |
| AED | 9 (17%) | 0 (0%) | |
| Anesthetics | 4 (8%) | 2 (8%) | |
| Dose of first drug correct | 36 (71%) | 15 (60%) | 1 |
| Therapy: therapy sequence correct | 34 (67%) | 16 (64%) | 1 |
| Refractory SE | 25 (49%) | 12 (48%) | 1 |
| Duration of SE | 0.53 | ||
| <1 h | 11 (22%) | 8 (32%) | |
| 1 h–24 h | 21 (41%) | 7 (28%) | |
| >24 h | 19 (37%) | 10 (40%) | |
| ICU | 49 (96%) | 24 (96%) | 1 |
| Mean duration ICU [SD] | 9.4 [13.1] | 7.2 [9.9] | 0.42 |
| Intubation | 22 (43%) | 14 (56%) | 0.34 |
| Infections in the course | 28 (55%) | 8 (32%) | 0.09 |
| Mean stay hospital [SD] | 17.3 [13.8] | 12.9 [10.2] | 0.13 |
SE: status epilepticus, GTCSE: generalized tonic clonic SE; CPSE: complex partial SE; SPSE: simple partial SE; BDZ: benzodiazepine; AED: antiepileptic drugs.
Bold values means "significant", i.e. p-value <0.05.
When evaluating outcome parameters and risk factors for older patients with NOSE, infections in the course of the disease were associated with an unfavourable outcome (68%, p = 0.04). Further, an extended stay in ICU and an extended stay in hospital were associated with non-favourable outcome (p = 0.005 for ICU, p = 0.001 for general stay). All other parameters did not differ between the two subgroups. In logistic regression analysis, none of the covariates (mean duration of stay on ICU and in hospital, infections) were independently associated with clinical outcome. Details for outcome comparison of older patients with NOSE are given in Table 4.
Table 4Outcome predictors in NOSE of older patients.
| Favourable (N = 20) | Non favourable (N = 31) | P | |
|---|---|---|---|
| Female | 16 (80%) | 19 (61%) | 0.22 |
| Semiology | 0.72 | ||
| GTCSE | 7 (35%) | 7 (23%) | |
| CPSE | 9 (45%) | 18 (58%) | |
| CPSE with coma | 2 (10%) | 4 (13%) | |
| SPSE | 2 (10%) | 2 (6%) | |
| Etiology | 0.09 | ||
| Symptomatic | 16 (80%) | 24 (77%) | |
| Acute | 6 (30%) | 18 (58%) | |
| remote | 9 (45%) | 5 (16%) | |
| Progressive | 1 (5%) | 1 (3%) | |
| Cryptogenic | 4 (20%) | 7 (23%) | |
| Other severe comorbidity | 3 (15%) | 8 (26%) | 0.29 |
| Dementia present | 4 (16%) | 5 (25%) | 0.50 |
| Therapy | |||
| First drug: | 1 | ||
| BDZ | 15 (75%) | 23 (74%) | |
| AED | 4 (20%) | 9 (16%) | |
| Anesthetics | 1 (5%) | 3 (10%) | |
| Dose of first drug correct | 17 (85%) | 19 (61%) | 0.12 |
| Therapy: therapy sequence correct | 13 (65%) | 21 (68%) | 1 |
| Refractory SE | 7 (35%) | 18 (58%) | 0.15 |
| Duration of SE | 0.46 | ||
| <1 h | 6 (30%) | 5 (16%) | |
| 1h–24 h | 8 (40%) | 13 (42%) | |
| >24 h | 6 (30%) | 13 (42%) | |
| ICU | 19 (95%) | 30 (97%) | 1 |
| Mean duration ICU [SD] | 4 [4.6] | 12.8 [15.4] | 0.005 |
| Intubation | 2 (10%) | 20 (65%) | 0.27 |
| Infections in the course | 7 (35%) | 21 (68%) | 0.04 |
| Mean stay hospital [SD] | 10.6 [6.9] | 21.6 [15.5] | 0.001 |
SE: status epilepticus, GTCSE: generalized tonic clonic SE; CPSE: complex partial SE; SPSE: simple partial SE; BDZ: benzodiazepine; AED: antiepileptic drugs.
Bold values means "significant", i.e. p-value <0.05.
4. Discussion
We here present a large cohort of patients with SE from a tertiary University centre, wherein the majority of patients were ≥60 years at onset. Older patients with SE had less frequently a favourable outcome than younger patients with SE. NOSE occurred in 56% of all patients with SE without age-related predominance, but 67% of the patients with NOSE were aged ≥60 years. Older patients with NOSE had a comparable outcome to younger patients with NOSE. Infections during treatment and prolonged stay on the ICU and in hospital were risk factors for an unfavourable outcome in older patients with NOSE.
4.1 Characteristic differences between older and younger patients with SE
Older patients with SE had less frequently GTCSE and more often non-convulsive SE with coma. Furthermore, older patients had less frequently a favourable outcome and more often showed a fatal outcome. These findings are in accordance to previous results. A comparably low proportion of GTCSE in older patients was shown in a geriatric patient cohort of 63 patients aged more than 70 years, in which only 17% of patients had GTCSE [
[21]
].Regarding inhospital mortality in the overall group of SE, our mortality rates were supported by two very recent German multicenter [
[22]
] and population-based studies [[23]
] on status epilepticus. They provided a mean mortality ratio of 14–15% and increases with a refractory cause of SE. The short-time mortality of 18% in our subgroup of older patients was surprisingly low compared to previous reports, which identified an age ≥60 years as a relevant outcome predictor [[24]
]. In early epidemiological studies, mortality increased with age up to 38% in older patients (>60 years) [[25]
]. In a very recent study from the University of Salzburg, Austria, authors reported a cohort of 120 patients acutely admitted with SE and performed a group-wise comparison for older (≥60 years) and younger patients (<60 years). Authors reported a proportion of 31% of older patients with in-hospital death, which was significantly higher than in our older subgroup [[26]
]. Authors used comparable inclusion criteria for SE and the reporting institution is situated in an environment that is comparable to ours with respect to ethnicities, health system facilities, and diagnostic and therapeutic work-up. Indeed, epidemiological and clinical data of our study matched, at least to a considerable extent, that reported by the Salzburg group: the majority of patients with SE were older than 60 years (Cologne 64% vs. Salzburg 70%), GTCSE was the common seizure type in the overall patient cohort (44% vs. 53%) and female gender was predominant (56% vs. 60%). Likewise, treatment protocols used the same step-wise approach based upon current guidelines and expert recommendations. Furthermore, adherence to these regimens was similar (70% vs. 67%).Differences in short-time mortality rates between our cohort and the Salzburg cohort may derive rather from procedural than clinical differences. For instance, short-time mortality could be influenced by different local discharge policies. In our cohort, fifty-five percent of the surviving older patients were not discharged directly at home: thirty-three percent were discharged to other acute clinics, sixteen percent to rehabilitation clinics and 6% were newly transferred to a nursing home. Our discharge policy reflects the usual practice of a tertiary University clinic in an area of high population density in Germany with supply of subordinated clinics, but discharge policies could differ in other regions with different surroundings. Therefore, we cannot exclude underreporting of the subsequent clinics and hence higher fatal outcome than that reported here.
There are two outcome scores to assess short-time outcome in SE: the STESS-Score [
17
, 27
] and the EMSE-Score [[18]
], which both assess in-hospital mortality. In our cohort, a positive predictive value (PPV; i.e., a high value indicates a poor outcome) of STESS-3 respectively STESS-4 was low with 15% and 21% for all patients with SE, whereas a negative predictive value (NPV, i.e., a low value indicates good outcome) was high compared to former work (92% for STESS-3 and 91% for STESS-4). Sensitivity of STESS-3/STESS-4 in our cohort was 77%/47%, whereas specificity ranged 37%/74%. In the initial publications of STESS, PPV leveled at 25–39% and NPV at 96–100% [17
, 27
]. In an external evaluation, STESS-score ≥ 3 points (STESS-3) attained a PPV of only 48.7%, after adjusting the cut-off-level to ≥4 (STESS-4), PPV improved to 73.1% [[20]
].In our population, EMSE-AECE at 64 points had a low PPV of 16% and high NPV of 91%. In the original publication of the EMSE-AECE-Score [
[18]
], the lowest score of non-survivors was 64 points with a NPV of 100% and a PPV of 68.8%. Recently, a group from Buenos Aires performed an external evaluation of both scores. They showed in 46 patients with SE, a sensitivity and specificity of STESS-3 of 61.5% and 75.7%, respectively, and for STESS-4 a sensitivity of 96.9% and a specificity of 53.8%. Comparable to our results, EMSE-EACE with a cutoff value of 64 points had a low sensitivity of 24.2% and a high specificity of 92.3% [[28]
]. In a current study from Madzar and co-workers from the University of Erlangen, Germany [[29]
], STESS-3 in 71 episodes of refractory SE was not seen as an independent parameter associated with in-hospital mortality but as a reliable outcome predictor for long-term outcome, assessed at last available follow-up (median 12 weeks, range 6–35). In that study, only patients with refractory SE were included, but their clinical and epidemiological data matched with our cohort as well. When looking for episodes with refractory SE in our cohort, in 81 of 150 episodes patients were older than 65 years (54%).Taken together, predictive values of available scoring tests vary considerably between different institutions. On the one hand, this may result from different definitions of SE, the actual patient populations, and cut-off-values. In the original STESS publication, authors included adults and adolescents ≥16 years, and SE was defined as a seizure lasting >30 min. Cut-off age was set at 65 years [
17
, 25
]. In EMSE [[18]
], anoxic encephalopathies were included, which were excluded in STESS. However, discrepancies in outcome scores may not only be reflected by different patient populations or treatment regimes [[26]
]. Rather, discharge from hospital may not be the ideal timepoint for treatment evaluation. Choosing a more later timepoint for the (re-)assessment may result in more reliable results that are comparable across different institutions, as it was done recently for EMSE-AECE and STESS in 164 episodes of SE within a 30-day period [[19]
]. For clinical purposes, it could be of interest to evaluate favourable outcome variables as well, as the proportion of surviving patients without neurological deterioration.4.2 Characteristics of NOSE in older compared to younger patients
When focusing on NOSE, fifty-six percent of our patients had a first episode of SE, i.e. patients had no history of epilepsy. This proportion lies within the range of existing data, estimating that the proportion of NOSE amounts 50–70% [
5
, 30
]. Nevertheless, we could not reproduce early epidemiological data of increased incidence rates of SE without known epilepsy in older patients ≥60 years in our cohort [[5]
]. Proportion of NOSE in our cohort was 56% and, again, matched data from the Salzburg group [[26]
], who reported a first episode of SE without known epilepsy in 50% of their patients. Surprisingly, outcome of older patients was comparable to that of younger adults under equivalent treatment efforts. Comparing the two subgroups, two differences emerged: older patients were more often females and had less frequently GTCSE. In sum, NOSE could not be identified as outcome factor for older people. Consistently, Holtkamp and colleagues did not find a cryptogenic etiology more often in people with refractory SE [[31]
], which questions the reliability of cryptogenic etiology or NOSE as predictor for poor outcome. In early epidemiological studies [[5]
], no history of epilepsy was reported in 58% of all patients with SE and in 70% of older patients (≥60 years). Of note, in this cohort anoxic states were included as SE with highest mortality rates (>50%) and one cannot derive from that study, how many of these anoxic states were assigned to NOSE.4.3 Predictive parameter for clinical outcome in older patients with NOSE
When assessing the predictive value of clinical variables for short-time outcome in older patients with NOSE, infections as complication in the course and prolonged stay in hospital were associated with non-favourable outcome. We could not identify one of these covariates as independent predictor for clinical outcome. A close interaction between these variables is obvious: infections can cause a prolonged stay, likewise patients with a prolonged stay in hospital are at higher risk for nosocomial infections.
Taken together, older patients with NOSE had comparable outcome rates as younger patients under sufficient treatment efforts. For that reason, our data do not provide evidence for a more reluctant treatment use in older patients. Prevention of infections could be an important factor to improve outcome in older patients.
4.4 Limitations
This is a retrospective non-interventional study with all its inherent problems. First of all, our searching strategy could have been too close or strict, excluding a relevant proportion of patients. Because we only included patients via documented ICD-codes from our electronic database, there could be a relevant proportion of patients, who had SE but were documented as having seizures, which could lead to an underdetection of patients with SE. This represents a crucial limitation of our data. We further aimed to include patients with high certainty in diagnosis and, therefore, we excluded patients with insufficient or uncertained data. Second, we report from a single large University centre, which serves as a supraregional acute neurological reference centre with high expertise in management of neurointensive care patients (and patients with epilepsy). Therefore, a selection bias for more acute and severe cases in our patient cohort seems possible. In addition, we cannot exclude a selection bias due to organizational reasons which were adopted at the University Hospital of Cologne during the inclusion period of 10 years. Third, our database did not provide any information regarding middle or long-term outcome after the in-hospital treatment period. Fourth, one could assume that differences in clinical outcome at discharge between different centres could raise from different expertise of the reporting centres but we do not expect relevant differences in patient care between the recent reported cohorts from Salzburg or Erlangen and ours.
5. Conclusion
We here present outcome data regarding a large cohort of patients with SE with a state-of-the-art approach in treatment, wherein older patients represented the majority. Despite a less frequently favourable outcome in older patients for the whole cohort of patients with SE, NOSE could not be identified as a relevant outcome predictor for older patients. Rather, older patients with NOSE had comparable outcome rates if comparable treatment effort was undertaken. Avoidance of infection should have priority as infections constituted a negative outcome predictor in older patients with NOSE. In general, the evaluation of outcome predictors remains an ongoing challenge in SE. Further prospective studies are warranted to guide clinicians in their therapeutic decision making in this relevant neurological field.
Author contributions
Study was designed by MPM and CD. MPM, RDN and TK collected the data. Statistical analysis was performed by MPM. MPM, GRF and CD have written the draft and all authors have contributed substantially to the final version.
Conflicts of interest
MPM received payments for congress participation, travel expenses, lecture and manuscript preparation from UCB and EISAI.
RDN reports no conflict of interest.
TK reports no conflict of interest.
GRF received honoraria for lectures from Bayer, Boehringer Ingelheim, Sanofi, Desitin, GlaxoSmith-Kline, Lilly, and TEVA.
LB received payments for congress participation, travel expenses, and lectures from UCB, Bial and EISAI.
CD received payments for lectures from UCB.
Acknowledgement
We confirm that we have read the Journal’s position on issues involved in ethical publication and affirm that this report is consistent with those guidelines.
References
- Incidence and case-fatality rate of status epilepticus in the Canton of Geneva.Lancet. 1999; 353: 1496
- Incidence of status epilepticus in adults in Germany: a prospective, population-based study.Epilepsia. 2001; 42: 714-718
- Guidelines for the evaluation and management of status epilepticus.Neurocrit Care. 2012; 17: 3-23
- A systematic review of the epidemiology of status epilepticus.Eur J Neurol. 2004; 11: 800-810
- A prospective, population-based epidemiologic study of status epilepticus in Richmond, Virginia.Neurology. 1996; 46: 1029-1035
- Incidence of status epilepticus in French-speaking Switzerland: (EPISTAR).Neurology. 2000; 55: 693-697
- Risk of unprovoked seizure after acute symptomatic seizure: effect of status epilepticus.Ann Neurol. 1998; 44: 908-912
- Incidence and short-term prognosis of status epilepticus in adults in Bologna, Italy.Epilepsia. 2003; 44: 964-968
- Status epilepticus after stroke is associated with increased long-term case fatality.Epilepsia. 2006; 47: 2020-2026
- New-onset refractory status epilepticus: etiology, clinical features, and outcome.Neurology. 2015; 85: 1604-1613
- It's time to revise the definition of status epilepticus.Epilepsia. 1999; 40: 120-122
- A definition and classification of status epilepticus–Report of the ILAE Task Force on Classification of Status Epilepticus.Epilepsia. 2015; 56: 1515-1523
- Unified EEG terminology and criteria for nonconvulsive status epilepticus.Epilepsia. 2013; 54: 28-29
- Guidelines for epidemiologic studies on epilepsy. Commission on epidemiology and prognosis, international league against epilepsy.Epilepsia. 1993; 34: 592-596
- Evidence-based guideline: treatment of convulsive status epilepticus in children and adults: report of the guideline Committee of the American Epilepsy Society.Epilepsy Curr. 2016; 16: 48-61
- Pharmacologic treatment of status epilepticus.Expert Opin Pharmacother. 2016; 17: 513-534
- A clinical score for prognosis of status epilepticus in adults.Neurology. 2006; 66: 1736-1738
- Epidemiology-based mortality score in status epilepticus (EMSE).Neurocrit Care. 2015; 22: 273-282
- Mortality, morbidity and refractoriness prediction in status epilepticus: comparison of STESS and EMSE scores.Seizure. 2017; 46: 31-37
- Independent external validation of the status epilepticus severity score.Crit Care Med. 2013; 41: e475-479
- Risk and prognostic factors of status epilepticus in the elderly: a case-control study.Epilepsia. 2011; 52: 1849-1856
- Costs and cost-driving factors for acute treatment of adults with status epilepticus: a multicenter cohort study from Germany.Epilepsia. 2016; 57: 2056-2066
- Costs, length of stay, and mortality of super-refractory status epilepticus: a population-based study from Germany.Epilepsia. 2017;
- Epidemiology and outcomes of status epilepticus in the elderly.Int Rev Neurobiol. 2007; 81: 111-127
- Status epilepticus in children, adults, and the elderly.Epilepsia. 1992; 33: S15-25
- Status epilepticus in the elderly-A retrospective study on 120 patients.Epilepsy Res. 2016; 127: 317-323
- Status Epilepticus Severity Score (STESS): a tool to orient early treatment strategy.J Neurol. 2008; 255: 1561-1566
- Role of EMSE and STESS scores in the outcome evaluation of status epilepticus.Epilepsy Behav. 2016; 64: 140-142
- Association of seizure duration and outcome in refractory status epilepticus.J Neurol. 2016; 263: 485-491
- The outcome of tonic-clonic status epilepticus.Curr Opin Neurol. 1994; 7: 93-95
- Predictors and prognosis of refractory status epilepticus treated in a neurological intensive care unit.J Neurol Neurosurg Psychiatry. 2005; 76: 534-539
Article info
Publication history
Published online: August 17, 2017
Accepted:
August 13,
2017
Received in revised form:
August 9,
2017
Received:
May 31,
2017
Identification
Copyright
© 2017 British Epilepsy Association. Published by Elsevier Ltd.
User license
Elsevier user license | How you can reuse 
Elsevier's open access license policy
Elsevier user license
Permitted
For non-commercial purposes:
- Read, print & download
- Text & data mine
- Translate the article
Not Permitted
- Reuse portions or extracts from the article in other works
- Redistribute or republish the final article
- Sell or re-use for commercial purposes
Elsevier's open access license policy
