Long-term outcome of convulsive status epilepticus: a 10-year follow-up

This study aimed to determine the mortality, causes of death and factors affecting the outcome of convulsive status epilepticus (CSE) at 10 years.

This retrospective study consisted of 62 consecutive adult patients diagnosed with CSE at the Helsinki University Hospital (HUS) emergency department during 2002-2003. Patients were followed for up to 10 years or up to the time of death. Data on patient demographics, CSE characteristics, treatment, complications, and outcome from the time of CSE were collected. The Official Statistics of Finland provided the information on mortality and causes of death. Survival analysis was conducted using Cox proportional hazards regression analysis.

In-hospital mortality was 8.1%, and mortality was 25.8% at one year, 51.6% at five years and 64.5% at 10 years. Estimated standardized mortality ratio (SMR) was 5.3 and the deceased patients lost 20.9 potential years of life, on average. The leading causes of death were disorders of the brain or the circulatory system, epilepsy-related conditions or intracranial tumours. The univariable survival analysis demonstrated that age ≥65 (HR=2.8, p=0.001), Charlson Comorbidity Index (CCI)>0 (CCI=1-3: HR=3.0, p=0.009; CCI>3: HR=8.4, p<0.001), Status Epilepticus Severity Score (STESS)>4 (HR=5.3, p<0.001) and Epidemiology-Based Mortality Score (EMSE-EAC)>15 (HR=2.2, p=0.036) were risk factors and a Glasgow outcome scale (GOS) of 5 at discharge (HR=0.14, p=0.025) was a protective factor for survival. The multivariable analysis established STESS>4 (HR=5.0, p=0.002) and CCI>0 (CCI=1-3: HR=2.9, p=0.015;CCI>3: HR=6.3, p=0.006) as independent risk factors and GOS>3 (time-dependent) (GOS=4: HR=0.33, p=0.048;GOS=5: HR=0.13, p=0.019) as a protective factor for survival.

The rate of long-term mortality and number of potential years of life lost were high. Factors demonstrative of the overall situation of the patients, such as comorbidities, functional state after CSE and age, were significant predictors for long-term outcome.