The results are illustrated separately for all patients and for those
who fulfilled the predefined inclusion criteria, termed the per protocol group.
A, Outcomes on the modified Rankin scale at 3 months after stroke onset and
at late stage (>1 year). The range for excellent outcome on the modified Rankin
scale is 0 to 1. The percentages of good outcome (mRS 0-2) are given separately
in the Results. B, Outcomes on the Barthel Index. The range of functionally
independent outcome is 95 to 100.
Lindsberg PJ, Soinne L, Tatlisumak T, Roine RO, Kallela M, Häppölä O, Kaste M. Long-term Outcome After Intravenous Thrombolysis of Basilar Artery Occlusion. JAMA. 2004;292(15):1862-1866. doi:10.1001/jama.292.15.1862
Author Affiliations: Department of Neurology,
Helsinki University Central Hospital (Drs Lindsberg, Soinne, Tatlisumak, Roine,
Kallela, Häppölä, and Kaste); and Neuroscience Program, Biomedicum
Helsinki (Drs Lindsberg and Tatlisumak), Helsinki, Finland.
Context Basilar artery occlusion (BAO) is an infrequent disease with high morbidity
and mortality. Intra-arterial thrombolysis is advocated for treatment but
is limited to use at specialized centers.
Objective To evaluate outcomes for patients with BAO treated with intravenous
Design, Setting, and Participants During 1995 to 2003, 50 consecutive patients with angiographically proven
BAO were treated according to an institutional therapy protocol based on intravenous
thrombolysis with recombinant tissue plasminogen activator (alteplase). Patients
were treated at an urban university teaching hospital receiving all patients
with ischemic stroke who were considered for thrombolysis in a catchment area
of 1.5 million inhabitants in Helsinki, Finland.
Intervention Intravenous administration of alteplase (0.9 mg/kg) during a 1-hour
Main Outcome Measures Basilar artery recanalization determined by magnetic resonance angiography
and clinical outcomes at 3 months and at 1 year or longer determined by modified
Rankin Scale and Barthel Index scores.
Results Recanalization was studied in 43 patients and verified in 26 (52%) of
all patients. By 3 months, 20 patients (40%) had died while 11 had good outcomes
(modified Rankin Scale score, 0-2); 12 (24%) reached independence in activities
of daily living (Barthel Index score, 95-100), and 6 (16%) were severely disabled
(Barthel Index score, 0-50). In the long term (median follow-up 2.8 years),
15 patients (30%) reached good outcomes (modified Rankin Scale score, 0-2)
while 23 (46%) died.
Conclusions Intravenous administration of alteplase for patients with BAO appears
to be associated with rates of survival, recanalization, and independent functional
outcome comparable with those reported with endovascular approaches. These
data suggest that a randomized trial is needed to compare these approaches
for treatment of BAO.
Basilar artery occlusion (BAO) is an infrequent and catastrophic disease
and is the gravest form of posterior circulation strokes that account for
about 20% of ischemic strokes. Basilar artery occlusion has a dismal natural
course and the mortality rate is 85% to 95% even with anticoagulant and fibrinolytic
therapy if the artery is not recanalized.1,2 Transient
prodromal symptoms are common, but complete BAO precipitates a sudden or gradually
progressing clinical syndrome. The most devastating sequela is the locked-in
state.3 Randomized trial data on the efficacy
of recanalization therapies do not exist,4,5 and
many stroke centers have adopted ad hoc BAO thrombolysis protocols, mainly
with intra-arterial approaches.
Previous reports3,4,6,7 have
advocated thrombolytic therapy delivered with an invasive endovascular approach
to the occlusion site of the posterior circulation, but even short delays
in therapy onset have been reported to be the single most critical factor
affecting outcome.5 We report the results of
the first 50 consecutive patients with proven BAO treated according to our
institutional protocol.8 Due to unacceptable
delays, we changed our approach to treatment of BAO from intra-arterial to
noninvasive intravenous delivery of alteplase (recombinant tissue plasminogen
activator [rtPA]), a protocol used in anterior circulation strokes.9
With a catchment area of 1.5 million, the Helsinki University Central
Hospital annually admits 20 to 30 stroke emergencies with BAO. Thrombolysis
for BAO represented 20% of all strokes treated with thrombolytic therapy at
our center. The primary inclusion criteria for this study were an angiographically
verified (magnetic resonance angiography [MRA] or digital subtraction angiography)
BAO and a consistent clinical syndrome of severe posterior circulation ischemia
in a previously independently functioning person. The onset could be preceded
by nonspecific, transient prodromal symptoms before sudden dense symptoms
or gradually progressing acute phase. For patients with sudden disturbance
of consciousness and tetraparesis, the protocol allowed treatment delays up
to 12 hours, and for patients with gradually increasing symptoms such as ophthalmoplegia,
dysarthria, and bilateral weaknesses, treatment delays up to 48 hours were
permitted. The treatment time (symptom-to-needle time) was defined as the
interval between the onset of brainstem symptoms that did not resolve and
the onset of rtPA infusion. Magnetic resonance angiography was performed on
a 1.5-T imager (Siemens Vision; Siemens Medical Systems, Erlangen, Germany)
with a standard 3-dimensional time-of-flight (TOF) sequence. Raw images were
evaluated to determine vessel patency.
Exclusion criteria were widespread ischemic changes and any signs of
intracranial bleeding on computed tomography (CT). Limited or unilateral hypodensity
of the brainstem was not an exclusion criterion provided the patient still
had spontaneous breathing, whereas absence of all brainstem reflexes was an
exclusion. Because the BAO therapy protocol was an institutionally accepted
management guideline,8 formal ethics committee
approval was not applicable. Nevertheless, we obtained oral informed consent
from each patient or next of kin.
After considering our general criteria for stroke thrombolysis9 and the BAO therapy protocol,8 patients
with suspected BAO based on the initial CT were immediately treated with standard
full-dose anticoagulation with unfractionated heparin adjusted to maintain
activated partial thromboplastin time between 50 and 100 seconds. Intravenous
labetalol was given as necessary to stabilize arterial blood pressure below
185/105 mm Hg. Alteplase (0.9 mg/kg) was infused (10% as a bolus) in 1 hour.
Repeat MRA examinations were carried out the next day. If only partial vessel
recanalization was found and the patient’s condition deteriorated during
follow-up with intracranial hemorrhage ruled out, or if reocclusion took place,
intravenous abciximab (0.6 mg/h) was administered as rescue therapy (to a
total of 4 patients). Four patients with systemic hypotension were treated
with norepinephrine or dopamine. Parenteral anticoagulation was continued
until oral warfarin therapy could be administered and reached the desired
level of anticoagulation. Sedatives (diazepam or propofol) were given as required
to permit imaging.
Following thrombolysis in the emergency department, patients were transferred
to the stroke unit for individualized early rehabilitation by multidisciplinary
personnel and for workup of etiology. A 3-month outcome follow-up was performed
for survivors by a stroke neurologist and a study nurse. It included the modified
Rankin scale (mRS): no or minimal symptoms (0-1), mild to moderate disability
(2-3), severely disabled or bedridden (4-5), dead (6); and the Barthel Index
score: functional independence (95-100), moderate dependency (90-55), full
dependency (50-0). For patients institutionalized or still in a rehabilitation
hospital, scoring was performed by the treating physician and the nurse, determined
from patient records, or by contacting the patient or relatives.
In addition, the functional outcome (Barthel Index, mRS) during the
late follow-up (12 months to several years later) was obtained by a stroke
study nurse assisted by a stroke neurologist via telephone conversation with
a family member or the nursing institute. At the same time, we assessed—from
a year to several years after thrombolysis—perceived quality of life
using a modification of a standardized questionnaire that assessed the patient’s
satisfaction with 10 different aspects of life (going out, walking, activities,
physical health, personal care, happiness, communication, sleep, recreation,
and family, along with an overall average score) on a 10-point scale.10,11 (The questionnaire is available on
request.) Surviving patients or their caretakers either completed and returned
a mailed survey or answered the survey questions in a telephone conversation
with standard instructions provided by the same study nurse.
Although our patients form a prospective series based on predefined
published criteria,8 and the 3-month data were
collected prospectively, all patient records also were reexamined to ascertain
all parameters, including the type of clinical presentation, treatment delays,
possible protocol violations, hemorrhagic complications, and the most probable
etiology. Recanalization, either partial or complete, was estimated based
on TOF images performed after the thrombolysis. However, in 5 patients recanalization
was presumed to have occurred by distinctive postthrombolytic clinical improvement
followed by confirmation on MRA obtained several weeks later. If the patient
had a known cardiac embolic source such as atrial fibrillation, or patent
foramen ovale, with sudden symptom onset and the occlusion located typically
distally in the basilar artery, cardioembolic etiology was presumed. The etiology
was presumed atherothrombotic if the arterial contour was stenotic and/or
severely affected by atherosclerosis after recanalization in the absence of
Distributions of the continuous variables were assessed with the Shapiro-Wilk
test. We carried out the univariate analysis using a t test
or Mann-Whitney U test, as appropriate. Dichotomous
variables were compared with a χ2 test or 2-tailed Fisher
exact test. Association of quality-of-life scales to the degree of disability
and outcome scale was evaluated with Spearman rank correlation coefficient.
A 2-tailed P <.05 was considered statistically significant.
We used the statistical software Statistica, version 5.5 (StatSoft, Tulsa,
Of the 50 consecutive patients with BAO treated with intravenous thrombolytic
therapy from 1995 to mid-2003, 13 were women and 37 men, with a mean age of
61.9 (range, 27-84) years. The mean (SD) admission blood pressure was 151
(27) mm Hg systolic and 84 (15) mm Hg diastolic, and labetalol was administered
to 15 patients. The mean (SD) admission blood glucose concentration was 137
(39.5) mg/dL (7.6 [2.2] mmol/L) and hemoglobin, 14.1 (1.6) g/dL. The most
commonly diagnosed cardiovascular risk factors were hypertension (40%), coronary
artery disease (26%), atrial fibrillation (24%), previous transient ischemic
attack/stroke (22%), generalized atherosclerosis (8%), hemostatic abnormality
(6%), peripheral artery disease (8%), and type 1 or 2 diabetes (4%).
The mean symptom-to-needle time was 9 hours in patients classified as
having the sudden-onset type of BAO (n=26) and 17.5 hours in patients with
gradually progressing symptoms (n=24). Twenty-three patients (46%) were unconscious,
and 38 (76%) were intubated. Thirty patients (60%) needed respiratory support
to maintain adequate ventilation or for sedation, and 9 patients (18%) with
reduced consciousness received mechanical ventilation without sedation. The
major etiologic conditions were vertebrobasilar atherothrombosis (44%), cardioembolism
(32%), and dissection of the vertebral artery (14%).
Basilar artery occlusion was diagnosed using MRA in 45 patients, DSA
in 3, and by both methods in 2 patients. Recanalization was reinvestigated
with MRA (TOF) after thrombolysis (median, 1.0 days; interquartile range,
1-3 days) in 43 patients and confirmed as complete or partial in 26 of these
(60%; 52% for the whole cohort). The main reason for not repeating MRA was
an unimproved clinical condition with an obviously poor prognosis. The recanalization
rate was 60% in both cardioembolic and atherothrombotic BAO and 71% in vertebral
artery dissection. Following rtPA treatment, 12 patients had intracranial
hemorrhage; 4, multifocal hemorrhages; 1, subarachnoid; 3, intraparenchymal;
and 4, asymptomatic hemorrhagic transformations. Seven hemorrhages (14% of
all 50 patients) were symptomatic based on acute worsening of the patients’
At 3 months, 20 patients had died (40% case-fatality rate). Nine patients
(18%) died within 7 days and 6 (12%) died within 30 days. The case-fatality
rate was 12% (3/26) in patients with partial or complete recanalization and
71% (17/24) in those with no evidence of recanalization (Fisher exact test, P<.001). The case-fatality rate was also significantly
higher in patients who were unconscious on admission (13/23 vs 7/27; χ21 test,P=.03). None of the 7 symptomatic
hemorrhages occurred in patients with evidence of recanalization (Fisher exact
At 3 months, 11 patients (22%) scored 0 to 2 on mRS (good outcome),
and 12 patients (24%) were independent in activities of daily living (Barthel
Index score, 95-100) (Figure). Eight
patients (16%) were fully dependent, scoring 0 to 50 on the Barthel Index.
None of those with failed recanalization were independent at 3 months, whereas
46% of those with verified recanalization were independent (Fisher exact test;P=.008). When mRS was dichotomized to good vs poor outcomes
(0-2 vs 3-6) at 3 months, there was a trend for shorter overall symptom-to-needle
times in patients with good outcome (9.2 vs 13.0 hours; Mann-Whitney U test,P=.05), whereas no significant
differences were detected for glucose level, blood pressure, the mode of onset
(sudden or progressive), or the fraction of unconscious patients.
On retrospective analysis, 12 patients did not fulfill the predefined
inclusion or exclusion criteria of our institutional protocol. In these cases,
the treating physicians together with the relatives deliberately agreed on
thrombolysis as the last resort for survival. Two such patients did not fulfill
the inclusion criterion of prestroke functional independence and scored 3
and 4 on the mRS. Three patients’ symptom-to-needle times were too long
(exceeding the 12-hour limit by 15 to 50 minutes). In 5 patients, there were
CT exclusion criteria of widespread posterior circulation and brainstem infarcts.
Another 2 instances had both of these exclusion criteria. In the per protocol
treated patients, the case-fatality rate was 32% and the functional outcome
was better (Figure).
During the extended follow-up (median follow-up time, 2.8 years; interquartile
range, 1.1-4.2 years), the condition of 8 survivors improved as assessed by
the mRS and 5 deteriorated. Individuals with mRS scores of 4 to 2 often improved,
while those with the most severe disability at 3 months (5) did not maintain
this score (3 died and 2 improved to score 4). In the long term, 30% (15/50)
of the patients reached good outcome (mRS score, 0-2) while 46% (23/50) died.
Of patients treated per protocol, 34% (13/38) reached good outcome. Two patients
with failed recanalization eventually did reach good outcome (2 and 1 on the
mRS, 85 and 100 on the Barthel Index). Patients with good long-term recovery
were not significantly younger than those with poor outcome (mean, 59.8 vs
63.5 years; P=.20).
The quality-of-life assessment was performed in 23 of the 27 survivors
after a mean (and median) follow-up time of 2.9 years (interquartile range,
1.5-4.6 years). The range of the total scores was 20 to 99 (possible range,
0-100; best score, 100 [mean, 71.7; median, 74]). The 2 lowest scores were
20 and 37 and the rest exceeded 50. The mean scores for the specified 10 items
were above mid point (range, 5.9-8.4; SD, 1.7-3.6), and the median values
were consistently higher (range, 7-9). The total scores were strongly correlated
with the Barthel Index (R=0.65, P<.001) and with the mRS (R=−0.79, P<.001).
In this case series of 50 patients with BAO treated with intravenous
thrombolysis, the overall recanalization rate was 52%, the likelihood of good
outcome (mRS score 0-2 or Barthel Index score 95-100) was 24%, and survival
at 3 months was 60%. These outcomes are equal or more favorable than outcomes
achieved with intra-arterial endovascular thrombolytics in series of similar
size and in a meta-analysis.1,2,4,5 The
case-fatality rates in the meta-analysis summarizing data from 164 patients
and in a recent series of 83 patients, which also included distal bilateral
vertebral artery occlusions, were 67% and 60%,4,5 while
it was 40% in our series. We included many unconscious patients, those with
extensive baseline infarctions, and elderly patients. In anterior circulation
stroke, the recommendation is to treat as soon as possible with intravenous
rtPA,12,13 and the sooner rtPA
is administered the greater the benefit.14 We
cannot find an a priori reason that BAO should be treated differently, thereby
preventing the unavoidable delays incurred by invasive endovascular procedures.
Intravenous rtPA may be a reasonable alternative and an efficient way to provide
thrombolysis in this devastating disorder.
We excluded patients with widespread ischemic changes on initial CT,
typically involving several posterior circulation territories, even extensively
in some cases of protocol violation. Limited/unilateral hypodensity of the
brainstem was not an exclusion criterion if the patient still breathed spontaneously.
Although many earlier series of BAO thrombolysis excluded patients with brainstem
series had at least equally high prevalence of unconsciousness (46%), and
the majority of patients (60%) eventually needed mechanical ventilation. Thus,
despite differences in the initial diagnosis, our cohort had at least comparably
severe posterior circulation symptoms as did earlier series of BAO thrombolysis.
Physicians treating patients with stroke and reduced consciousness,
immediate need of assisted ventilation, and established brain damage may find
it difficult to justify rescue therapies without scientific proof and question
their long-term implications to the patient and his/her family. Although unconsciousness
has been noted as a predictor of death,4 unconsciousness
as the presenting symptom in our cohort did not preclude good outcome. Without
thrombolysis, unconsciousness in BAO invariably predicted a fatal outcome.15 Our data suggest that the condition of survivors
who initially stabilize at the poorest outcome (mRS score, 5) tend to either
decease or improve in the long term. Eventually, half of recanalized patients
will reach good outcome (mRS 0-2). Differences in imaging methods may compromise
the comparison of our recanalization rates to those of studies using intra-arterial
Survivors in our study reported fair quality of life, based on a simple
scale of perceived well-being.10,11 The
average scores of the equivalent items compare well with those reported in
adults with and without chronic disease.11 The
fair quality of life of survivors also supports the use of intravenous thrombolysis
in view of the higher survivorship achieved in our series.
A limitation of this study is that we lacked a comparison group to show
the safety and efficacy of our thrombolytic protocol, but nonrecanalized BAO
allowed only 4% chance of good outcome in the long term. The safety of intravenous
thrombolysis of BAO was supported by the fact that symptomatic hemorrhages
occurred only to patients without recanalization. This cohort had relatively
long symptom-to-needle times, which have recently improved by the use of CT
angiography instead of CT followed by MRA.
Endovascular techniques have been proposed as the optimal treatment
for BAO but without referring to data on the noninvasive approach.5,7 We now report the highest rates of
survival and good outcome in BAO published so far, based on use of intravenous
rtPA, which is more noninvasive, quicker, and more widely available than endovascular
techniques (which require neuroradiological interventionist service). Despite
these findings from our case series, the evidence is not sufficient to recommend
a change in treatment guidelines. A randomized multicenter trial is needed
to directly compare intra-arterial and intravenous thrombolysis in BAO. Until
results of such a study become available, centers unequipped with 24-hour
access to immediate invasive neuroradiology may consider noninvasive thrombolysis
once the diagnosis of BAO is confirmed.
Corresponding Author: Perttu J. Lindsberg,
MD, PhD, Department of Neurology, Helsinki University Central Hospital, PO
Box 340, 00029 HUS, Helsinki, Finland (firstname.lastname@example.org).
Author Contributions: Drs Lindsberg and Soinne
had full access to all of the data in the study and take responsibility for
the integrity of the data and the accuracy of the data analysis.
Study concept and design: Lindsberg, Soinne,
Acquisition of data: Lindsberg, Soinne, Tatlisumak, Roine,
Analysis and interpretation of data: Lindsberg,
Soinne, Tatlisumak, Roine, Kallela, Häppölä, Kaste.
Drafting of the manuscript: Lindsberg, Soinne,
Critical revision of the manuscript for important
intellectual content: Lindsberg, Tatlisumak, Roine, Kallela, Kallela,
Statistical analysis: Lindsberg, Soinne.
Administrative, technical, or material support: Tatlisumak,
Study supervision: Lindsberg, Kaste.