SHORT COMMUNICATION
Statins are associated with decreased mortality risk after status epilepticus A. Sierra-Marcosa, V. Alvareza, M. Faouzib, B. Burnandb and A. O. Rossettia a
Department of Clinical Neurosciences, Centre Hospitalier Universitaire Vaudois (CHUV), and University of Lausanne, Lausanne; and Institute of Social and Preventive Medicine, Centre Hospitalier Universitaire Vaudois (CHUV), and University of Lausanne, Lausanne, Switzerland
b
EUROPEAN JOURNAL OF NEUROLOGY
Keywords:
epilepsy, hydroxymethylCoA reductase, outcome, status epilepticus, treatment Received 15 August 2013 Accepted 24 February 2014 European Journal of Neurology 2015, 22: 402–405 doi:10.1111/ene.12428
Background and purpose: Statins display anti-inflammatory and anti-epileptogenic properties in animal models, and may reduce the epilepsy risk in elderly humans; however, a possible modulating role on outcome in patients with status epilepticus (SE) has not been assessed. Methods: This cohort study was based on a prospective registry including all consecutive adults with incident SE treated in our center between April 2006 and September 2012. SE outcome was categorized at hospital discharge into ‘return to baseline’, ‘new disability’ and ‘mortality’. The role of potential predictors, including statins treatment on admission, was evaluated using a multinomial logistic regression model. Results: Amongst 427 patients identified, information on statins was available in 413 (97%). Mean age was 60.9 (17.8) years; 201 (49%) were women; 211 (51%) had a potentially fatal SE etiology; and 191 (46%) experienced generalized-convulsive or non-convulsive SE in coma. Statins (simvastatin, atorvastatin or pravastatin) were prescribed prior to admission in 76 (18%) subjects, mostly elderly. Whilst 208 (50.4%) patients returned to baseline, 58 (14%) died. After adjustment for established SE outcome predictors (age, etiology, SE severity score), statins correlated significantly with lower mortality (relative risk ratio 0.38, P = 0.046). Conclusion: This study suggests for the first time that exposure to statins before an SE episode is related to its outcome, involving a possible anti-epileptogenic role. Other studies are needed to confirm this intriguing finding.
Introduction Status epilepticus (SE) represents a condition with considerable morbidity and mortality [1]. The role of medications other than anti-epileptic drugs (AEDs) on SE outcome has received scarce attention. Statins are competitive inhibitors of the HMG-CoA-reductase, the key enzyme in cholesterol biosynthesis. Recently, their potential anti-inflammatory and anti-excitotoxic effects have been described in animal models of different neurological processes [2]. In epilepsy in particular, statins decrease seizure severity and related hippocampal cell death in rodents [3,4]. Despite their Correspondence: Dr Andrea O. Rossetti, Service de Neurologie, CHUV-BH07, CH-1011 Lausanne, Switzerland (tel.: +41 21 314 1220; fax: +41 21 314 1290; e-mail: [email protected]).
402
wide prescription, the impact upon SE prognosis in humans has not been studied.
Patients and methods This cohort study, fully approved by our ethics committee, prospectively enrolled consecutive adults with incident SE episodes treated in our academic center between 1 April 2006 and 1 September 2012 [5]. Subjects 30 min (until 2008) and >5 min (since 2008) [6]. Seizure semiology was classified as simple-partial,
© 2014 The Author(s) European Journal of Neurology © 2014 EAN
SE AND STATINES
absence, generalized-myoclonic, complex-partial, generalized-convulsive or non-convulsive SE in coma. Etiologies were further classified as ‘potentially fatal’ if potentially leading to death independently of SE [7–9]. The validated prognostic score STESS (Status Epilepticus Severity Score) [10] was prospectively categorized as 0–2 (favorable) versus 3–6 (unfavorable outcome). At hospital discharge, three outcomes were assessed: return to baseline conditions, survival with new disability, or death. The information about diagnostic and therapeutic procedures has been described previously [5]. Data analysis was performed using STATA 12.1 (College Station, TX, USA). Data were summarized as mean (SD) for continuous variables and numbers (percentages) for categorical data. A multinomial logistic regression was performed to assess the association of each potential predictor (age, gender, SE semiology, previous seizures, fatal etiology, STESS, SE duration, history of cerebrovascular disease, and previous statins treatment) to SE outcome. The strength of the associations was measured using the relative risk ratio (RRR), where the ‘return to baseline’ group was the reference. Significant predictors were used in a backward procedure to fit a multivariate model.
Results From a total of 515 episodes recorded, 427 incident SE patients were identified. In 413 of them (97%), data regarding statin treatment on admission were available: mean age was 60.9 (17.8) years; 201 (49%) were women; 172 (42%) had a previous diagnosis of epilepsy; 249 (60%) presented an acute symptomatic etiology and 211 (51%) a potentially fatal
15
10 mg
12 11
20 mg 8
40 mg
7
80 mg
4
1 1 0 Simvastatine (Total 29p)
1 0 Atorvastatine (Total 23p)
Undetermined dose
1 0
0 0
Pravastatine (Total 9p)
Figure 1 Type and dose of statins on admission (p, patients). Fifteen subjects received an unidentified statin.
© 2014 The Author(s) European Journal of Neurology © 2014 EAN
403
etiology. SE duration was between 5 and 29 min in 37 (9%) patients. Five (1%) experienced absence, one (0.2%) myoclonic, 81 (20%) simple-partial, 135 (33%) complex-partial, 170 (41%) generalized-convulsive and 21 (5%) non-convulsive SE in coma. Simvastatin, atorvastatin and pravastatin were prescribed before admission in 76 (18%) subjects (Fig. 1). Regarding outcome, 208 (50%) patients returned to baseline and 58 (14%) died. Table 1 illustrates univariable analyses of potential outcome predictors: older age, previous seizures, a potentially fatal etiology and STESS ≥ 3 were significantly related to new disability and mortality, P < 0.001; whilst gender, SE semiology and SE duration of the episodes were not. Previous statins treatment was not associated with outcome in univariable analyses. Older patients were clearly more likely to receive statins (P < 0.001): after adjustment for age, previous seizures, fatal etiology, STESS and history of cerebrovascular disease, statins correlated independently with a lower mortality (RRR 0.38, 95% CI 0.15–0.98, P = 0.046) but not disability (RRR 0.72, 95% CI 0.4–1.36, P = 0.29) (Table 2). Older age and potentially fatal etiology were independent predictors for both outcomes, and a high STESS score was independently related to mortality. Of note, there was no interaction between statins and stroke. Figure 2 illustrates the association between age and mortality.
Discussion This study provides evidence that statins are correlated with a lower risk of SE mortality after adjustment for the principal outcome predictors [1,7,8]. Only two human studies have examined the possible benefits of HMG-CoA inhibitors in modifying epilepsy risk. New-onset epilepsy risk in veterans was lowered by statins (odds ratio 0.64) [11], whereas amongst a large cohort of cardiovascular patients who received a revascularization procedure the adjusted risk ratio for epilepsy amongst statin users was 0.65 [12]. Our results expand these observations, exploring prognosis after a prolonged seizure episode. Mortality after SE mostly relates to etiology and complications and not to SE itself [1,7]; it is therefore possible that statins also impacted on the global health of acutely admitted patients. Statins were only related to mortality after correction for age. This may be regarded as a ‘negative bias’: in univariable analyses, age cancelled the effect of statins on the studied outcome. There was no notable effect on disability, which, however, is not as robust an outcome as mortality and is assessed at variable times (hospital discharge).
A. SIERRA-MARCOS ET AL.
404
Table 1 Potential predictors of disability or death after status epilepticus; univariable analyses
Total Age, mean (SD) Gender (female) Previous seizures Semiology (generalized convulsive or non-convulsive SE in coma) Potentially fatal etiology STESS, median (range) SE duration
Statins are associated with decreased mortality risk after status epilepticus A. Sierra-Marcosa, V. Alvareza, M. Faouzib, B. Burnandb and A. O. Rossettia a
Department of Clinical Neurosciences, Centre Hospitalier Universitaire Vaudois (CHUV), and University of Lausanne, Lausanne; and Institute of Social and Preventive Medicine, Centre Hospitalier Universitaire Vaudois (CHUV), and University of Lausanne, Lausanne, Switzerland
b
EUROPEAN JOURNAL OF NEUROLOGY
Keywords:
epilepsy, hydroxymethylCoA reductase, outcome, status epilepticus, treatment Received 15 August 2013 Accepted 24 February 2014 European Journal of Neurology 2015, 22: 402–405 doi:10.1111/ene.12428
Background and purpose: Statins display anti-inflammatory and anti-epileptogenic properties in animal models, and may reduce the epilepsy risk in elderly humans; however, a possible modulating role on outcome in patients with status epilepticus (SE) has not been assessed. Methods: This cohort study was based on a prospective registry including all consecutive adults with incident SE treated in our center between April 2006 and September 2012. SE outcome was categorized at hospital discharge into ‘return to baseline’, ‘new disability’ and ‘mortality’. The role of potential predictors, including statins treatment on admission, was evaluated using a multinomial logistic regression model. Results: Amongst 427 patients identified, information on statins was available in 413 (97%). Mean age was 60.9 (17.8) years; 201 (49%) were women; 211 (51%) had a potentially fatal SE etiology; and 191 (46%) experienced generalized-convulsive or non-convulsive SE in coma. Statins (simvastatin, atorvastatin or pravastatin) were prescribed prior to admission in 76 (18%) subjects, mostly elderly. Whilst 208 (50.4%) patients returned to baseline, 58 (14%) died. After adjustment for established SE outcome predictors (age, etiology, SE severity score), statins correlated significantly with lower mortality (relative risk ratio 0.38, P = 0.046). Conclusion: This study suggests for the first time that exposure to statins before an SE episode is related to its outcome, involving a possible anti-epileptogenic role. Other studies are needed to confirm this intriguing finding.
Introduction Status epilepticus (SE) represents a condition with considerable morbidity and mortality [1]. The role of medications other than anti-epileptic drugs (AEDs) on SE outcome has received scarce attention. Statins are competitive inhibitors of the HMG-CoA-reductase, the key enzyme in cholesterol biosynthesis. Recently, their potential anti-inflammatory and anti-excitotoxic effects have been described in animal models of different neurological processes [2]. In epilepsy in particular, statins decrease seizure severity and related hippocampal cell death in rodents [3,4]. Despite their Correspondence: Dr Andrea O. Rossetti, Service de Neurologie, CHUV-BH07, CH-1011 Lausanne, Switzerland (tel.: +41 21 314 1220; fax: +41 21 314 1290; e-mail: [email protected]).
402
wide prescription, the impact upon SE prognosis in humans has not been studied.
Patients and methods This cohort study, fully approved by our ethics committee, prospectively enrolled consecutive adults with incident SE episodes treated in our academic center between 1 April 2006 and 1 September 2012 [5]. Subjects 30 min (until 2008) and >5 min (since 2008) [6]. Seizure semiology was classified as simple-partial,
© 2014 The Author(s) European Journal of Neurology © 2014 EAN
SE AND STATINES
absence, generalized-myoclonic, complex-partial, generalized-convulsive or non-convulsive SE in coma. Etiologies were further classified as ‘potentially fatal’ if potentially leading to death independently of SE [7–9]. The validated prognostic score STESS (Status Epilepticus Severity Score) [10] was prospectively categorized as 0–2 (favorable) versus 3–6 (unfavorable outcome). At hospital discharge, three outcomes were assessed: return to baseline conditions, survival with new disability, or death. The information about diagnostic and therapeutic procedures has been described previously [5]. Data analysis was performed using STATA 12.1 (College Station, TX, USA). Data were summarized as mean (SD) for continuous variables and numbers (percentages) for categorical data. A multinomial logistic regression was performed to assess the association of each potential predictor (age, gender, SE semiology, previous seizures, fatal etiology, STESS, SE duration, history of cerebrovascular disease, and previous statins treatment) to SE outcome. The strength of the associations was measured using the relative risk ratio (RRR), where the ‘return to baseline’ group was the reference. Significant predictors were used in a backward procedure to fit a multivariate model.
Results From a total of 515 episodes recorded, 427 incident SE patients were identified. In 413 of them (97%), data regarding statin treatment on admission were available: mean age was 60.9 (17.8) years; 201 (49%) were women; 172 (42%) had a previous diagnosis of epilepsy; 249 (60%) presented an acute symptomatic etiology and 211 (51%) a potentially fatal
15
10 mg
12 11
20 mg 8
40 mg
7
80 mg
4
1 1 0 Simvastatine (Total 29p)
1 0 Atorvastatine (Total 23p)
Undetermined dose
1 0
0 0
Pravastatine (Total 9p)
Figure 1 Type and dose of statins on admission (p, patients). Fifteen subjects received an unidentified statin.
© 2014 The Author(s) European Journal of Neurology © 2014 EAN
403
etiology. SE duration was between 5 and 29 min in 37 (9%) patients. Five (1%) experienced absence, one (0.2%) myoclonic, 81 (20%) simple-partial, 135 (33%) complex-partial, 170 (41%) generalized-convulsive and 21 (5%) non-convulsive SE in coma. Simvastatin, atorvastatin and pravastatin were prescribed before admission in 76 (18%) subjects (Fig. 1). Regarding outcome, 208 (50%) patients returned to baseline and 58 (14%) died. Table 1 illustrates univariable analyses of potential outcome predictors: older age, previous seizures, a potentially fatal etiology and STESS ≥ 3 were significantly related to new disability and mortality, P < 0.001; whilst gender, SE semiology and SE duration of the episodes were not. Previous statins treatment was not associated with outcome in univariable analyses. Older patients were clearly more likely to receive statins (P < 0.001): after adjustment for age, previous seizures, fatal etiology, STESS and history of cerebrovascular disease, statins correlated independently with a lower mortality (RRR 0.38, 95% CI 0.15–0.98, P = 0.046) but not disability (RRR 0.72, 95% CI 0.4–1.36, P = 0.29) (Table 2). Older age and potentially fatal etiology were independent predictors for both outcomes, and a high STESS score was independently related to mortality. Of note, there was no interaction between statins and stroke. Figure 2 illustrates the association between age and mortality.
Discussion This study provides evidence that statins are correlated with a lower risk of SE mortality after adjustment for the principal outcome predictors [1,7,8]. Only two human studies have examined the possible benefits of HMG-CoA inhibitors in modifying epilepsy risk. New-onset epilepsy risk in veterans was lowered by statins (odds ratio 0.64) [11], whereas amongst a large cohort of cardiovascular patients who received a revascularization procedure the adjusted risk ratio for epilepsy amongst statin users was 0.65 [12]. Our results expand these observations, exploring prognosis after a prolonged seizure episode. Mortality after SE mostly relates to etiology and complications and not to SE itself [1,7]; it is therefore possible that statins also impacted on the global health of acutely admitted patients. Statins were only related to mortality after correction for age. This may be regarded as a ‘negative bias’: in univariable analyses, age cancelled the effect of statins on the studied outcome. There was no notable effect on disability, which, however, is not as robust an outcome as mortality and is assessed at variable times (hospital discharge).
A. SIERRA-MARCOS ET AL.
404
Table 1 Potential predictors of disability or death after status epilepticus; univariable analyses
Total Age, mean (SD) Gender (female) Previous seizures Semiology (generalized convulsive or non-convulsive SE in coma) Potentially fatal etiology STESS, median (range) SE duration
