We studied 46 patients with paroxysmal dyskinesia and classified them
according to phenomenology, duration of attacks, and etiology. There w
ere 13 patients, 7 females, who had paroxysmal kinesigenic dyskinesia
(PKD), 10 with attacks lasting 5 minutes or less (short lasting) and 3
with attacks lasting longer than 5 minutes (long lasting). Twenty-six
patients, 18 females, had paroxysmal nonkinesigenic dyskinesia (PNKD)
, 9 with short-lasting and 17 with longlasting PNKD. Five patients, 3
females, had paroxysmal exertion-induced dyskinesia (FED), 3 with shor
t-lasting FED and the other 2 with long-lasting FED. In addition, ther
e was 1 patient with paroxysmal hypnogenic dyskinesia (PHD) and 1 with
paroxysmal superior oblique myokymia. Only 2 patients, 1 with PKD and
1 with PHD, had family history of paroxysmal dyskinesias. No specific
cause could be identified in 21 patients; in the other 23 patients th
e etiologies included the following: psychogenic (9 patients), cerebro
vascular diseases (4), multiple sclerosis (2), encephalitis (2), cereb
ral trauma (2), peripheral trauma (2), migraine (1), and kernicterus (
1). Nine of 10 (90%) patients with PKD improved with medications, most
ly anticonvulsants, compared with only 7 of 13 (37%) with PNKD. This n
ew classification, based chiefly on precipitating events, allowed appr
opriate categorization of the attacks in all our patients with paroxys
mal dyskinesias.