The prognosis of supratentorial haematomas is based on clinical signs
and radiological features. The role of evoked potentials has not been
evaluated systematically. In a prospective study of supratentorial hae
morrhage a number of clinical (17), radiological (3) and evoked potent
ial (2) parameters were evaluated employing univariate logistic regres
sion analysis in 69 patients and multivariate logistic regression step
down analysis in 51 patients. The outcome was graded on the basis of t
he Barthel index (BI) score at 3 months as good (PI greater than or eq
ual to 12) or poor (death or BI < 12) recovery. Employing univariate a
nalysis the significant prognostic variables were Glasgow Coma Scale,
Canadian Neurological Scale, tendon reflex, associated medical complic
ations, urinary incontinence, ventricular extension of the haematoma a
nd motor evoked potentials. Using multivariate logistic regression ana
lysis the best set of parameters in relation to outcome included Glasg
ow Coma Scale (P < 0.05), Canadian Neurological Scale (P < 0.05), tend
on reflex (P < 0.1), ventricular extent (P < 0.01) and motor evoked po
tentials (P < 0.05). From this study it is concluded that, in addition
to clinical and radiological parameters, motor evoked potentials also
have an important role in predicting outcome.