BACKGROUND: Status epilepticus (SE) is a neurological emergency with high mortality and often a poor functional outcome amongst survivors. So far, only status epilepticus severity score (STESS) is available to predict individual outcomes. STESS is based on weighted sum scores of age, type of seizure, level of consciousness and history of previous seizures. Weighting factors were based on a priori assumptions. METHODS: We tested in an explorative, hypothesis generating approach, whether epidemiological data (i.e. mortality rates) can be combined to form a score (Epidemiology-based Mortality score in SE-EMSE), and further, which combination of aetiology, age, comorbidity, EEG, duration and level of consciousness yields highest test performance. Positive and negative predictive value, and correctly classified were compared to STESS (with different cut-off levels: STESS-3, STESS-4). Score points for each parameter, e.g. age, were derived from previously published specific mortality rates. For each combination of parameters, the lowest sum-score of deceased individuals was taken as cut-off. Ninety-two consecutive non-hypoxic patients (age range 20-90 years), with various forms of SE admitted to a tertiary care neurological intensive care unit were investigated retrospectively. RESULTS: EMSE using a combination of aetiology, age, comorbidity and EEG (NPV = 100 %, PPV = 68.8 %, correctly classified 89.1 %) was superior to STESS-3 and STESS-4 (p = 0.0022 or lower). CONCLUSION: EMSE explained individual mortality in almost 90 % of cases, and performed significantly better than previous scores. This explorative study needs external prospective corroboration. EMSE may be a valuable tool for risk stratification in interventional studies in the future.
BACKGROUND:Status epilepticus (SE) is a neurological emergency with high mortality and often a poor functional outcome amongst survivors. So far, only status epilepticus severity score (STESS) is available to predict individual outcomes. STESS is based on weighted sum scores of age, type of seizure, level of consciousness and history of previous seizures. Weighting factors were based on a priori assumptions. METHODS: We tested in an explorative, hypothesis generating approach, whether epidemiological data (i.e. mortality rates) can be combined to form a score (Epidemiology-based Mortality score in SE-EMSE), and further, which combination of aetiology, age, comorbidity, EEG, duration and level of consciousness yields highest test performance. Positive and negative predictive value, and correctly classified were compared to STESS (with different cut-off levels: STESS-3, STESS-4). Score points for each parameter, e.g. age, were derived from previously published specific mortality rates. For each combination of parameters, the lowest sum-score of deceased individuals was taken as cut-off. Ninety-two consecutive non-hypoxic patients (age range 20-90 years), with various forms of SE admitted to a tertiary care neurological intensive care unit were investigated retrospectively. RESULTS: EMSE using a combination of aetiology, age, comorbidity and EEG (NPV = 100 %, PPV = 68.8 %, correctly classified 89.1 %) was superior to STESS-3 and STESS-4 (p = 0.0022 or lower). CONCLUSION: EMSE explained individual mortality in almost 90 % of cases, and performed significantly better than previous scores. This explorative study needs external prospective corroboration. EMSE may be a valuable tool for risk stratification in interventional studies in the future.
Authors: Wolfgang G Muhlhofer; Stephen Layfield; Daniel Lowenstein; Chee Paul Lin; Robert D Johnson; Shalini Saini; Jerzy P Szaflarski Journal: Epilepsia Date: 2019-04-07 Impact factor: 5.864