The Current Status of Status Epilepticus

Abstracts & Commentary

Sources: Mayer SA, et al. Arch Neurol. 2002;59(2):205-210; Claassen J, et al. Neurology. 2002;58(1):139-142; Logroscino G, et al. Neurology. 2002;58(4):537-541.

Mayer and colleagues analyzed 83 episodes of status epilepticus (SE) in 74 patients presenting to the Columbia campus of New York Presbyterian Hospital. They identified 26 (31%) episodes defined as refractory status epilepticus (RSE) because of continuous or repetitive seizures that did not respond to first- and second-line anticonvulsant drug (ACD) therapy. Moreover, 11 of the 26 patients continued to have seizures after a third ACD was given. Univariate analysis was applied to find clinical factors associated with RSE. Those were followed by entering significant variables in a backward, stepwise logistic regression model to identify RSE independent risk factors. Only nonconvulsive SE (NCSE) and focal motor seizures at onset proved to be statistically significant in this model. Although mortality was not affected, hospital length of stay and Glascow Outcome Scale (GOS) score were increased in the RSE group relative to the 57 episodes of SE responsive to 1-2 ACDs.

The same data set were evaluated by Claassen and colleagues from the standpoint of finding predictors of functional disability (16/69 nonfatal episodes of SE) and mortality (14 episodes) following SE. Hospital length of stay longer than 2 weeks and acute symptomatic seizures (defined as seizures related to a readily identifiable precipitant in a patient with no history of seizures), predicted worse GOS on hospital discharge compared to admission. Age older than 62 and, again, acute symptomatic seizures were independent predictors of mortality.

In another group looking at long-term mortality following an incident case of SE, Logroscino and colleagues found a cumulative mortality of 43% (n = 62) over 10 years among patients who had survived 30 days following SE (n = 145). The standardized mortality ratio (SMR) at 10 years was 2.8 for all 30-day survivors. The following clinical characteristics predicted a significant increase in mortality: SE duration lasting 24 hours or more, acute symptomatic etiology, and myoclonic SE.


Sorting through all the multivariate analyses presented in these and earlier investigations, Neurology Alert recognizes 2 practical considerations. The first treatment is speed: SE must be definitively treated with alacrity. The second conclusion is that the acute pathology causing SE is most likely to kill the patient, in both the short- and long term.

The Columbia data confirm the conclusion of the Veterans Affairs Cooperative Study of SE (Treiman DM, et al. N Engl J Med. 1998;339:792-798) as far as showing the increasing difficulty of stopping SE, the longer the condition is sustained. Although the duration of SE was not screened in the initial univariate analysis of factors correlated with RSE, when the Columbia investigators used the second ACD earlier than in the VA trial, the second-line agent was more effective. In addition, the fact that NCSE emerged as a significant risk factor for RSE implies a longer duration of SE, since NCSE is typically recognized much later than convulsive SE. The fact that SE continues to be a frequently fatal illness in the 21st century is disheartening. There can be no better model of neuronal excitoxicity than a condition that, by definition, involves unremitting cortical hyperexcitability. SE needs to be treated with the gravity and speed comparable to cardiopulmonary resuscitation. The critical topic for further studies is to attempt to formulate SE treatment algorithms comparable to those for Advanced Cardiac Life Support.

It has long been recognized that what causes SE is the main determinant of short-term survival. Logroscino et al have now shown that the same can be said for the influence of long-term mortality. Acute symptomatic SE (resulting from insults such as brain trauma, cerebrovascular disease, and CNS infection) has the worst outcome. Furthermore, SMR was not increased among patients with idiopathic or cryptogenic SE, implying that the SE per se does not alter long-term mortality. When neurologists are asked to judge the outcome of SE, these factors should guide discussions with medical colleagues and family members, but, as has just been argued, should not serve as an excuse for therapeutic nihilism. —Andy Dean

Dr. Dean is Assistant Professor of Neurology and Neuroscience, Director of the Epilepsy Monitoring Unit, Department of Neurology, New York Presbyterian Hospital—Cornell Campus.