Evidence table for systematic reviews
Study reference Design Inclusion Criteria Quality Results Comments

Gilbert, 2003


Inclusion criteria:

Studies in Medline, published between 1970-2000, English language, about seizures and epilepsy in children and adults. Studies had to describe standar, not prolonged, interictal EEG and had to describe which findings were considered abnormal. Longitudinal studies were included when subjects had EEG and where all patients, irrespective of EEG results, were followed for at least 1 yr for the occurence of a seizure. Studies had to report data from a single EEG for each patients or present data separately on a single EEG performed before follow-up period.


Exclusion criteria:

Case-control studies and studies exclusively of neonates and infants. Studies where EEG was used for a decision to treat and there was a significant difference between recurrence in treated vs not treated patients


=25 studies, involving 4912 EEG



Search: adequate


Selection: adequate


Quality assessment of individual studies: Not done

Specificity: 0.13-0.99
X2 for heterogenity of studies was significant (P<.001)

Sensitivity: 0.2-0.91
X2 for heterogenity was significant (P<.001)

Thresholds for classifying EEG as positive accounted for 37% of the variance (suggesting wide interreader variation, what influences ability of EEG to discriminate between seizure recurrences)

Riviello, 2006

Evidence-based systematic review

Inclusion criteria:
Studies in the library of the University of Minnesota and Medline, for English-language articles from 1970 to 2005. Evidence on the assessment pf the child with status epilepticus (SE)

Exclusion criteria: Seizures occuring in neonates less than 1 month of age, seizures occuring in adults (>18), studies with <21 patients included, articles consisting of single patient case reports or small samples of unusual pathologic findings, or articles that referred specifically to febrile or refractory SE

= 25 articles (6 about EEG, including 413 EEG findings in 358 children with SE)

Search: adequate


Selection : adequate


Quality assessment of individual studies: adequate

89.3±13.6% of EEGs were abnormal.

43.1% was noted epileptiform and helped determine the nature and location of precipitating electroconvulsive events, of which:
Generalized: 16%
Focal: 19
Combined generalized and focal: 19%

Normal: 7.7%
Generalized slowing: 41%
Focal slowing: 6.3%
Electrocerebral inactivity: 1.9%)