The electromyographic responses to dorsal rootlet stimulation during partial dorsal rhizotomy are inconsistent.

Concerns have arisen regarding the ability of physiologic techniques to select rootlets mediating spasticity in children undergoing partial dorsal rhizotomy (PDR) for the treatment of spastic gait from cerebral palsy. To determine whether these physiologic responses are reproducible, 60 rootlets in 6 patients were graded from 0 to 4+ according to the system first reported by Phillips and Park, and then retested and graded using a randomized, blinded paradigm. Two thirds of the rootlets had one or more grade difference between the two trials, and 25% had a grade change of 3 or more. Only 4 of 23 rootlets with a "normal' response (grade 0) on one test were also graded normal on the other test. Only 2 of 14 rootlets with a bilateral response (grade 4+) had a bilateral response on both trials. Statistical analysis demonstrated no correlation and poor agreement between trials. Regardless of whether responses are reflex in origin with technical or physiologic variability, or the consequence of stimulus spread to ventral roots, we found inconsistent responses with commonly used methods and the criteria most agreed upon. These findings suggest that currently used techniques are not reliable for rootlet "selection' and result only in a random partial rhizotomy.