Escolar Documentos
Profissional Documentos
Cultura Documentos
Summary
Background Swallowing dysfunction after stroke is common, but there is little reliable evidence for how the disorder
should be managed. This study compared standard low-intensity and high-intensity behavioural interventions with
usual care for dysphagia.
Methods 306 patients with clinical dysphagia admitted to hospital with acute stroke were randomly assigned to
receive usual care (n=102), prescribed by the attending physician; standard low-intensity intervention (n=102),
comprising swallowing compensation strategies and diet prescription three times weekly for up to a month; or
standard high-intensity intervention and dietary prescription (n=102), at least daily for up to a month. The primary
outcome measure was survival free of an abnormal diet at 6 months. Analysis was done by intention to treat. This
study is registered with ClinicalTrials.gov, number NCT00257764.
Findings 60 patients died and three patients were lost to follow up before the 6-month analysis. Of patients randomly
allocated usual care, 56% (57/102) survived at 6 months free of an abnormal diet compared with 67% (136/204)
allocated standard swallowing therapy (relative risk 119, 95% CI 098145). Standard swallowing therapy was
associated with a non-signicant trend toward a reduction in death (080, 0513), institutionalisation (069,
0411), and dependency (105, 0813); a signicant reduction in swallowing-related medical complications
(073, 0609), chest infection (056, 0408), and death or institutionalisation (073, 055097); and a
signicant rise in the proportion of patients regaining swallowing function (141, 103194) by 6 months.
Compared with usual care and low-intensity therapy, high-intensity therapy was associated with an increased
proportion of patients who returned to a normal diet (p=004) and recovered swallowing (p=002) by 6 months.
Interpretation These data show a consistent trend towards more favourable outcomes in dysphagic stroke patients
who are assigned a standard programme of early behavioural swallowing intervention, including active therapeutic
approaches and dietary modication.
Introduction
Swallowing disorders (dysphagia with or without
aspiration) are present in about 2750% of patients with
acute stroke,15 and are associated with an increased risk
of complications, such as aspiration pneumonia,
dehydration, and malnutrition, in a fth to a third
of patients.1,6,7 These complications are theoretically
preventable by prompt and accurate diagnosis of the
swallowing disorder and appropriate intervention. The
possible interventions include indirect behavioural
strategies (eg, modication of food consistency),8 direct
behavioural strategies (eg, stimulation of oral and
pharyngeal structures),9 and enteral feeding by means of
a nasogastric tube or by percutaneous endoscopic
gastrostomy.1013
Feeding by percutaneous endoscopic gastrostomy is
associated with an increased rate of death or a poor
outcome compared with feeding by nasogastric tube.11,12
The effectiveness of behavioural interventions for
swallowing disorders after stroke, however, is uncertain.
Kasprisin and colleagues14 did a non-randomised,
uncontrolled study of swallowing therapy in patients
with stroke and mixed neurogenic causes (n=69). The
main outcome of this study was pneumonia
development as reported from chart review. Despite the
http://neurology.thelancet.com Vol 5 January 2006
Articles
Methods
Patients
The study was undertaken in a university teaching
hospital, the Royal Perth Hospital, which provides
medical services for the eastern suburban region of
Perth, Western Australia. All patients presenting to the
Royal Perth Hospital over a 3-year period were screened
for inclusion in the study. Patients were included if a
clinical diagnosis of stroke was conrmed by the
attending clinician (GJH), according to the WHO
denition of stroke;18 if the onset of stroke was within the
previous 7 days; if the study speech pathologist (JP)
made a clinical diagnosis of swallowing difculty
(dysphagia), as measured by a score of less than 85 on
the Paramatta Hospitals assessment of dysphagia;1921 if
the patient had no history of swallowing treatment or
surgery of the head or neck; and if the patient gave
written informed consent to participate in the trial and
be followed up for the next 6 months. Patients for whom
consent could not be gathered were placed on a waiting
list, reassessed daily, and included if they met the entry
criteria and consented within 7 days of stroke onset.
Eligible patients underwent a baseline assessment of
demography, comorbidity, premorbid function, clinical
stroke syndrome, stroke pathology, stroke subtype, and
results of CT and MRI brain scans. Stroke severity was
measured according to the modied Rankin score22 and
the modied Barthel index,23 and swallowing function
32
Procedures
Randomisation was undertaken by use of a block
randomisation technique. The treatment allocation was
based on a computer-generated random numbers list
generated with the SPSS statistical package (version
9.0). The randomisation schedule was held in the trial
ofce, remote from the study environment. After clinical
assessment by the study speech pathologist (JP), eligible
patients were informed about the trial and, after
providing informed consent, were randomly assigned to
one of three treatment options by means of a telephone
call to the trial ofce by the study speech pathologist.
The three treatment groups were usual care, standard
low-intensity treatment, and standard high-intensity
treatment. Usual care (control) consisted of patient
management by the attending physicians as per usual
practice. Physicians referred their patients to the
existing hospital speech pathology service if they
considered it to be appropriate. Treatment, if offered,
consisted mainly of supervision for feeding and
precautions for safe swallowing (eg, positioning, slowed
rate of feeding). Videouoroscopic assessment of
swallowing function was undertaken only if prescribed
by the attending physician.
Standard low-intensity swallowing therapy was
composed of swallowing compensation strategies,
mainly environmental modications (eg, upright
positioning for feeding); safe swallowing advice (eg,
reduced rate of eating); and appropriate dietary
modication, under the direction of the study speech
pathologist, three times per week for a month, or for the
duration of the hospital stay (if less than a month). The
choice of specic swallowing compensation strategies
was directed by the ndings of the clinical swallowing
examination and videouoroscopy (at baseline and at
follow up, if necessary).
Standard high-intensity swallowing therapy consisted
of direct swallowing exercises (eg, effortful swallowing,
supraglottic swallow technique) and appropriate dietary
modication, under the direction of the study speech
pathologist, every working day for a month or daily for
the duration of the hospital stay (if less than a month).
The choice of specic swallowing exercises was
established by the ndings of the clinical examination
and videouoroscopy (at baseline and at follow up, if
necessary).
http://neurology.thelancet.com Vol 5 January 2006
Articles
2356 ineligible
460 not notified of study
92 discharged/outside window
595 not dysphagic
1041 non-stroke diagnosis
90 died before assessment
78 inappropriate
871 eligible
306 randomised
17 died at 6 months
2 lost to follow up
20 died at 6 months
1 lost to follow up
23 died at 6 months
Demographics
Age, years
Sex, male
Clinical syndrome
Total anterior circulation syndrome
Partial anterior circulation syndrome
Lacunar syndrome
Posterior circulation syndrome
Unknown/non-stroke
Pathology
Cerebral infarction
Cerebral haemorrhage
Unknown
Stroke severity
Barthel index 15
Barthel index 15
Stroke handicap
Rankin score 3
Rankin score 3
Length of hospital stay, days
High intensity
(n=102)
Low intensity
(n=102)
Usual care
(n=102)
698 (125)
60 (59%)
72 (124)
59 (58%)
714 (127)
59 (58%)
36 (35%)
34 (33%)
23 (23%)
8 (8%)
1 (1%)
30 (29%)
35 (34%)
25 (25%)
10 (10%)
2 (2%)
42 (41%)
30 (29%)
19 (19%)
10 (10%)
1 (1%)
93 (91%)
8 (8%)
1 (1%)
90 (88%)
10 (10%)
2 (2%)
90 (88%)
11 (11%)
1 (1%)
80 (78%)
22 (22%)
80 (78%)
22 (22%)
81 (79%)
21 (21%)
15 (15%)
87 (85%)
191 (105)
21 (21%)
81 (79%)
192 (133)
17 (17%)
85 (83%)
214 (124)
33
Articles
Usual care
ANOVA
81 (79%)
116 (63)
242 (76)
114 (65)
28 (27%)
48 (39)
160 (92)
179 (87)
51
389
20
00001
00001
0001
88 (86%)
78 (38)
248 (67)
167 (79)
Data are number (%) or mean (SD). *Data are for patients who received treatment and do not include those who received none.
Usual care
(n=102)
Standard swallowing
therapy (n=204)
57
136
119 (098145)
33
64
48
23
26
49
49
72
93
94
54
37
36
103
72
140
141 (103194)
073 (060092)
056 (041076)
080 (04913)
069 (04311)
105 (08213)
073 (055097)
097 (08311)
not due to renal insufciency or use of diuretics. Calorienitrogen decit was dened as serum albumin less
than 25 g/L or sustained ketonuria without glycosuria.
Dependency was dened as a modied Rankin scale score
of 3 or more or a Barthel index score of 15/20 or less.
Statistical analysis
The study sample size calculations were based on our
earlier nding that 87% of patients with dysphagia after
rst-ever acute stroke returned to a normal diet by
6 months.7 Because many of these patients with mild
stroke had been treated with standard behavioural
swallowing therapies, we postulated that about 70% of
patients assigned usual care in this study would have a
6-month survival period free of an abnormal diet and that
assignment to standard swallowing therapy would
Usual care
Low intensity
High intensity
57
65
71
004
33
48
64
23
26
72
49
72
44
26
44
20
17
68
36
69
49
28
50
17
19
69
36
72
002
0003
005
029
022
065
006
087
Table 4: Comparison of high-intensity versus low-intensity speech therapy versus usual care
34
Results
A total of 3227 patients with suspected strokes were
referred to the study team between May 1, 1996, and
May 31, 1999, of whom 27% were eligible for inclusion
in the study (gure 1). Written informed consent was
obtained from 306 (35%) of the 871 eligible patients.
Ineligible patients did not differ signicantly from
enrolled patients in severity of stroke or dysphagia.
Table 1 shows the baseline characteristics of the patients.
The mean time to admission was 8 h after stroke onset
and to randomisation was 27 days after stroke onset for
all treatment groups. The mean duration of hospital stay
did not differ between the three groups.
The number of swallowing therapy sessions and the
duration of sessions for patients assigned to the
standard treatment groups were signicantly greater
than for those assigned to the usual care group (table 2).
The number of days over which patients received
swallowing therapy was signicantly less in the standard
treatment groups than in the usual care group.
Follow up at 6 months was complete for 243 (99%) of
the 246 survivors. Data for the 60 patients who died and
three patients lost to follow up at 56 days, 110 days,
and 163 days after randomisation, respectively, were
http://neurology.thelancet.com Vol 5 January 2006
Articles
10
Cumulative survival
08
06
High intensity
Low intensity
Usual care
04
02
Log rank test p=046
00
0
50
100
150
42
44
40
23
32
19
14
22
8
Discussion
The results of this study do not lend support to the
primary study hypothesis that a standard programme of
swallowing therapy given early to patients with acute
stroke and maintained as required, is associated with a
signicant increase in the proportion of patients who
survive free of an abnormal diet by 6 months after
stroke. Furthermore, there was no signicant effect of
swallowing therapy on death, institutionalisation, and
dependency at 6 months after stroke. However, the study
did show a consistent, albeit non-signicant, trend in
favour of swallowing therapy versus usual care for these
outcomes, and a signicant rise in the proportion of
patients who achieved functional swallowing and a
decrease in the proportion of patients who had a
dysphagia-related medical complication (especially
aspiration pneumonia) and died or needed
institutionalisation (as a combined outcome) by
6 months after stroke. A doseresponse relation was also
apparent with a signicant trend toward improved
recovery of swallowing function and less chest infection
in patients randomly allocated intensive standard
swallowing therapy than in those allocated less intensive
therapy and usual care.
The strengths of this study are the reduced systematic
bias in treatment allocation by randomisation, reduced
observer bias in outcome assessment by the masking of
the outcome assessor to the treatment allocation, and the
minimisation of random error by the relatively large
number of patients and outcomes studied.
There are several weaknesses of this study. The rst is
that the patient and the treating therapist were aware of
the treatment allocation, which introduces the
possibility of bias in the way treatment was delivered to
certain patients and the way patients reported subjective
outcomes. The most likely systematic error was that
patients assigned usual care could have been treated
with a more standard approach by hospital therapists
because they may have indirectly learnt about the
ongoing trial and observed how other enrolled patients
were being treated, thus biasing the results toward the
null hypothesis. We believe that the patients awareness
of treatment allocation was unlikely to have biased the
results because 59% of patients discharged from
hospital were unable to identify the nature of the
swallowing treatment they received. Furthermore,
patient outcome recording bias is unlikely to have
affected rates of objective outcomes, such as medical
complications (eg, chest infection), death, and
institutionalisation. Finally, the independent outcome
assessor, who recorded patient diet progress and
swallowing function, was able to predict the treatment
allocation correctly in only 29% of patients (Kappa 032).
Another potential weakness is that, despite this being
the largest randomised trial of any behavioural
swallowing intervention in acute stroke, the numbers of
patients and outcome events were still rather small. As a
35
Articles
Articles
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
37