Although the general guidelines for do-not-resuscitate (DNR) orders ap
ply to acute stroke patients, few data are available to aid decision-m
aking. With a view to developing specific guidelines for use in patien
ts with acute stroke, we decided to evaluate the clinical factors asso
ciated with DNR orders at our university teaching hospital. We prospec
tively studied 450 consecutive patients with acute hemispheric strokes
(237 men and 213 women, mean age 75 +/- 12 years). Thirty-six patient
s (8%) had intracerebral hemorrhage (ICH) and 414 (92%) had ischemic s
trokes. Overall inhospital mortality was 26%. DNR status was given to
31% of all patients at some time during their admission (83% of those
died). DNR decision-making was closely associated with the severity of
the neurologic deficit (Canadian Neurological Scale score less than o
r equal to 5); the patient's incapacity for informed DNR decision-maki
ng; age (>60 years); and devastating ICH unsuitable for surgery (p < 0
.001). Fifty-three percent of DNR orders were given on admission (firs
t 24 hours of the hospital stay), 35% during the first week of the hos
pital stay, due to brain damage, and 12% at any time between days 8 an
d 44 due to systemic complications. Once DNR status was given, 53% of
patients continued to receive normal nutrition and 60% still received
medical or surgical treatment. Although the current practice of DNR or
ders in patients with acute stroke is generally satisfactory, some cri
teria leg, age and operable ICH) need revision. Following the decision
to withhold CPR, patients with severe stroke, irreversible brain dama
ge, and/or significant comorbidities should receive DNR status wheneve
r the prognosis has become clear for physicians and family. We suggest
provisional disease-specific criteria for DNR decision-making in acut
e stroke.