H reflex studies were performed in ten spastic children with cerebral
palsy who underwent partial dorsal rhizotomy (PDR). Under anesthesia p
rior to PDR, H reflex amplitude evoked by percutaneous peripheral nerv
e stimulation gradually declined in all patients and became unobtainab
le in five. Motor responses could still be evoked by apparent dorsal r
oot stimulation in these five, but since it was shown that they were M
responses and not reflexes, PDR was performed randomly. In the other
five patients, in whom H reflexes were still evokable, dorsal root sti
mulation evoked motor responses which were true reflex responses, and
PDR was performed ''selectively.'' Compared to preoperative values, po
stoperative H-max/M(max) ratios declined, but no more so in selective
than in random PDR. We conclude that current intraoperative methods fo
r selection of ''abnormal'' dorsal rootlets for section may be invalid
and may have no bearing on successful outcome, since spasticity impro
ves even with random PDR. (C) 1994 John Wiley and Sons, Inc.