Objective: To investigate the initial termination rate of status epilepticus (SE) in a large observational study and explore associated variables. Methods: Data of adults treated for SE were collected prospectively in centers in Germany, Austria, and Switzerland, during 4.5 years. Incident episodes of 1,049 patients were analyzed using uni-and multivariate statistics to determine factors predicting cessation of SE within 1 hour (for generalized convulsive SE [GCSE]) and 12 hours (for non-GCSE) of initiating treatment. Results: Median age at SE onset was 70 years; most frequent etiologies were remote (32%) and acute (31%). GCSE was documented in 43%. Median latency between SE onset and first treatment was 30 minutes in GCSE and 150 minutes in non-GCSE. The first intravenous compound was a benzodiazepine in 86% in GCSE and 73% in non-GCSE. Bolus doses of the first treatment step were lower than recommended by current guidelines in 76% of GCSE patients and 78% of non-GCSE patients. In 319 GCSE patients (70%), SE was ongoing 1 hour after initiating treatment and in 342 non-GCSE patients (58%) 12 hours after initiating treatment. Multivariate Cox regression demonstrated that use of benzodiazepines as first treatment step and a higher cumulative dose of anticonvulsants within the first period of treatment were associated with shorter time to cessation of SE for both groups. Interpretation: In clinical practice, treatment guidelines were not followed in a substantial proportion of patients. This underdosing correlated with lack of cessation of SE. Our data suggest that sufficiently dosed benzodiazepines should be used as a first treatment step.