Somatosensory cortical evoked potentials (SCEPs) monitoring in neuromu
scular scoliosis surgery was evaluated in a large consecutive series o
f spinal reconstructions to define its role in the detection and preve
ntion of spinal cord injury; 141 procedures in 101 patients were evalu
ated. In 28% a reliable tracing could not be obtained. Two postoperati
ve neurologic deficits were missed by SCEP monitoring. Diagnosis (Duch
enne muscular dystrophy and polio) was the only significant factor fou
nd to correlate with a reliable tracing. The use of sublaminar wires w
as associated with a higher rate of change in tracing. Transient chang
es were common. SCEPs are unreliable and nonspecific in neuromuscular
scoliosis surgery and are not efficatious in preventing or detecting s
pinal cord injury when used alone. Adjunctive techniques using epidura
l and MEPs must be studied in these patients.