Purpose: Felbamate (FBM) is a new antiepileptic drug (AED) that is oft
en effective in seizure disorders refractory to other treatments; its
use has been greatly restricted after cases of aplastic anemia were re
ported. To elucidate the putative association between FBM and aplastic
anemia, we made a detailed evaluation of the first 31 reports. Method
s: Hematologic review according to the criteria of the International A
granulocytosis and Aplastic Anemia Study (IAAAS) confirmed 23 cases (7
4%) as aplastic anemia; FBM was judged to be the only plausible cause
for three; confounding (mostly by other drugs) was considered possible
, but FBM remained the most likely cause for 11; and there was at leas
t one other plausible cause for 9.Results: Using a denominator from sa
les data of 110,000 persons exposed and a numerator of the cases for w
hich FBM was considered the only plausible cause, we established a low
er limit of incidence of 27 cases of aplastic anemia per million users
as compared with the general population rate of 2.0 per million per y
ear. With all confirmed cases used as the numerator, the upper limit o
f incidence was 209 per million. The ''most probable'' incidence was e
stimated to be 127 per million. Conclusions: Intensive, systematic inv
estigation can maximize the utility of case reports for assessing risk
s of newly released drugs. The present evaluation confirmed an associa
tion between FBM and aplastic anemia; however, confounding was signifi
cant for most cases and there was a tenfold range in the ''best case''
and ''worst case'' incidence estimates among users.