- Email: [email protected]
Rapid Aspiration Screening for Suspected Stroke
Accepted Manuscript The Rapid Aspiration Screening in Suspected Stroke Part 1: Development and Validation Stephanie K. Daniels, PhD, Shweta Pathak, MPH, John C. Rosenbek, PhD, Robert O. Morgan, PhD, Jane A. Anderson, PhD PII:
S0003-9993(16)30076-4
DOI:
10.1016/j.apmr.2016.03.025
Reference:
YAPMR 56523
To appear in:
ARCHIVES OF PHYSICAL MEDICINE AND REHABILITATION
Received Date: 7 March 2016 Accepted Date: 29 March 2016
Please cite this article as: Daniels SK, Pathak S, Rosenbek JC, Morgan RO, Anderson JA, The Rapid Aspiration Screening in Suspected Stroke Part 1: Development and Validation, ARCHIVES OF PHYSICAL MEDICINE AND REHABILITATION (2016), doi: 10.1016/j.apmr.2016.03.025. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT Aspiration screening for suspected stroke The Rapid Aspiration Screening in Suspected Stroke Part 1: Development and Validation Stephanie K. Daniels, PhD,1,2 Shweta Pathak, MPH,3 John C. Rosenbek, PhD,4 Robert O.
1
RI PT
Morgan, PhD,3 Jane A. Anderson, PhD5,6
Research Service Line, Michael E. DeBakey VA Medical Center, Houston, TX; 2Department of
Communication Sciences and Disorders, University of Houston, Houston, TX; 3School of Public
SC
Health, University of Texas Health Sciences Center, Houston, TX; 4Department of
M AN U
Communication Sciences and Disorders, University of Florida, Gainesville, FL; 5Health Services Research and Development Center of Excellence, Michael E. DeBakey VA Medical Center, Houston, TX; 6Department of Neurology, Baylor College of Medicine, Houston TX; The authors would like to acknowledge the contribution of the research coordinators, speech pathologists, and neurologists who facilitated subject recruitment and data collection and also the
TE D
nurses who participated in the study. The project was supported by a Merit Award (1I0RX00121) from Rehabilitation Research & Development of the VA Office of Research and Development Service. The views expressed in this article are those of the authors and do not
EP
necessarily represent the views of the Department of Veterans Affairs or the University of Texas
AC C
School of Public Health. The full study protocol can be accessed at the Department of Veterans Affairs.
Device Status: The manuscript submitted does not contain information about medical device(s). SKD, SP, JCR, and ROM received salary support from the Department of Veterans Affairs-Merit Award (1I0RX00121). JAA has no conflict of interest.
ACCEPTED MANUSCRIPT Aspiration screening for suspected stroke For Correspondence and Reprints: Stephanie K. Daniels, PhD, Research Service (151), Michael E. DeBakey VA Medical Center, 2002 Holcombe Blvd, Houston, TX 77030, Phone: 713-794-
AC C
EP
TE D
M AN U
SC
RI PT
7152, FAX: 713-794-7786, Email: [email protected]
ACCEPTED MANUSCRIPT Aspiration screening for suspected stroke 1
The Rapid Aspiration Screening in Suspected Stroke
2
Part 1: Development and Validation Abstract
4
OBJECTIVES: To develop and validate a nurse-administered screening tool to identify
5
aspiration risk in patients with suspected stroke.
6
DESIGN: Validity study comparing evidence-based swallowing screening items with
7
videofluoroscopic swallowing study (VFSS).
8
SETTINGS: The study was completed in a certified primary stroke center in a major
9
metropolitan medical facility.
M AN U
SC
RI PT
3
PARTICIPANTS: Consecutive patients (N=250) admitted with suspected stroke.
11
INTERVENTIONS: Patients were administered evidence-based swallowing screening items by
12
nurses. A VFSS was completed within 2 hours of swallowing screening.
13
MAIN OUTCOME MEASURE: Validity relative to identifying VFSS determined aspiration for
14
each screening item and for various combinations of items.
15
RESULTS: Aspiration was identified in 29 of the 250 participants. Logistic regression revealed
16
age (p=0.012), dysarthria (p=0.001), abnormal volitional cough (p=0.030), and signs related to
17
trial water swallow (p=0.021) to be significantly associated with aspiration. Validity was then
18
determined based on the best combination of significant items for predicting aspiration. Results
19
revealed that age >70, or dysarthria, or signs related to trial water swallow (i.e., cough, throat
20
clear, wet vocal quality, and inability to continuously swallow 90ml water) yielded 93%
21
sensitivity and 98% negative predictive value.
AC C
EP
TE D
10
1
ACCEPTED MANUSCRIPT Aspiration screening for suspected stroke CONCLUSIONS: The final validated tool, Rapid Aspiration Screening for Suspected Stroke
23
(RAS3), is a valid nurse-administered tool to detect risk of aspiration in patients presenting with
24
suspected stroke.
25
Key Words: Deglutition Disorders, Stroke, Screening, Validity
AC C
EP
TE D
M AN U
SC
RI PT
22
2
ACCEPTED MANUSCRIPT Aspiration screening for suspected stroke List of Abbreviations
27
AUC
area under curve
28
CI
confidence interval
29
ED
emergency department
30
LAR
legally authorized representative
31
NPV
negative predictive value
32
NIHSS
National Institutes of Health Stroke Scale
33
PPV
positive predictive value
34
RAS3
Rapid Aspiration Screening for Suspected Stroke
35
ROC
receiver operating characteristic
36
RN
registered nurse
37
SLP
speech-language pathologist
38
SSI
swallowing screening item
39
SST
swallowing screening tool
40
TIA
transient ischemic attack
41
TVF
true vocal fold
42
VFSS
videofluoroscopic swallowing study
43
WS
water swallow
AC C
EP
TE D
M AN U
SC
RI PT
26
3
ACCEPTED MANUSCRIPT Aspiration screening for suspected stroke Early detection of aspiration risk in patients with suspected stroke is critical for immediate
45
intervention to reduce mortality, morbidity, and healthcare costs.1-3 These findings have led to
46
the recommendation of screening swallowing, regardless of stroke severity, in individuals
47
presenting with suspected stroke.4
RI PT
44
48
Swallowing screening tools (SSTs) administered by frontline healthcare providers such as
50
registered nurses (RNs) have been developed in individuals with stroke;5-8 however, all are
51
validated in patients with confirmed stroke, not individuals presenting with suspected stroke.
52
This is a limitation as screening is completed early in the care process, either in the emergency
53
department (ED) or upon hospital admission, frequently before confirmation of stroke diagnosis.
54
Moreover, SST validation has been limited to studies in which individuals were referred for
55
swallowing evaluation,7,8 or had specific National Institutes of Health Stroke Scale (NIHSS)
56
scores.6 These factors limit generalization of SSTs to real-world application.
M AN U
TE D
57
SC
49
Finally, most stroke SSTs have focused on identification of dysphagia, not aspiration.6,8 While
59
the incidence of dysphagia in stroke is high,9 not all patients who have dysphagia aspirate.
60
During the acute stroke workup, the focus should be on rapidly determining aspiration risk for
61
patient safety in order to quickly provide oral medication and address patient needs such as fluid
62
intake. After the emergent stroke workup, formal dysphagia evaluation by a speech-language
63
pathologist (SLP) is necessary to fully evaluate oropharyngeal swallowing and determine if an
64
instrumental swallowing evaluation is warranted. The aim of this study was to develop a valid
65
nurse-administered screening tool for the rapid identification of aspiration risk in patients
66
presenting with suspected stroke. The a priori hypothesis was that a combination of non-
AC C
EP
58
4
ACCEPTED MANUSCRIPT Aspiration screening for suspected stroke 67
swallowing and swallowing screening items would provide the highest validity for identification
68
of aspiration risk.
69
RI PT
70
Methods
72
Participants
73
Medical records of patients consecutively admitted for stroke workup to the Michael E. DeBakey
74
Veterans Affairs Medical Center (MEDVAMC) or who had a suspected stroke during
75
hospitalization between October 2012 and November 2014 were reviewed for eligibility. A
76
clinical, non-validated swallowing screening was completed in the ED as part of the hospital’s
77
stroke protocol.10 Patients were approached for participation regardless of the clinical screening
78
results. Consented participants met all the inclusion/exclusion criteria with a total of 250 patients
79
submitted for data analyses (Figure 1). RNs were recruited from MEDVAMC stroke wards to
80
administer and interpret swallowing screening items (SSIs) on consented participants (see Part 2
81
for RN details). The determination of sample size was based on the existing literature on
82
prevalence of aspiration.11-13 The study was approved by the Institutional Review Board at the
83
Baylor College of Medicine and the Research and Development Committee at MEDVAMC in
84
Houston, TX, and all individuals provided written consent prior to participation.
M AN U
TE D
EP
AC C
85
SC
71
86
Screening items
87
SSIs were obtained from a systematic literature review.14 Items were submitted to 10 content
88
experts (SLPs and RNs who had greater than 10 years of experience working with stroke and/or
89
dysphagia in VAMCs) who rated the items according to usefulness in detecting aspiration and
5
ACCEPTED MANUSCRIPT Aspiration screening for suspected stroke feasibility for nursing implementation. Content expert agreement was determined utilizing
91
content validity index methodology.15 The final SSIs evaluated for validity were six non-
92
swallowing items: male, age >70 years, lethargy, dysarthria, wet vocal quality, abnormal
93
volitional cough and five swallowing items: cough after swallowing, throat clear after
94
swallowing, wet vocal quality after swallowing, inability to continuously drink 90ml water
95
without stopping, wet vocal quality after 1 minute. Most are similar to items on available SSTs.5-
96
8
97
criteria assigned by the content experts.
RI PT
90
99
M AN U
98
SC
Other items from the literature review were not part of the tested SSIs as they did not meet
All non-swallowing items were assessed regardless of present or absent score. Non-swallowing items were followed by administration of at least one water swallow trial. The swallowing
101
portion of the screening consisted of three water swallow trials (5ml twice, 90ml) administered
102
via a cup or straw. If any sign of aspiration risk (cough, throat clear, wet voice) was present after
103
either 5ml water swallow, the screening was stopped, and subsequent water swallow trials were
104
not administered. If no aspiration sign was present for either 5ml trial, the 90ml trial was
105
administered, and the patient was assessed for ability to continuously drink and vocal quality
106
after 1 minute in addition to the same signs of aspiration risk identified on smaller water
107
volumes.
EP
AC C
108
TE D
100
109
Screeners were blinded to the clinical ED screening results. They were instructed to score items
110
as yes (present), no (not present), undecided, or no/limited response (Appendix 1; see Part 2 for
111
details on training). For scoring of abnormal volitional cough, a weak or breathy cough was
112
differentiated from other vocal responses (throat clear, verbalizes the word “cough”, or
6
ACCEPTED MANUSCRIPT Aspiration screening for suspected stroke production any other response that was not an actual cough) in order to determine differences in
114
relationship to aspiration. Because individual screenings had variable numbers of swallowing
115
items depending on whether screening was stopped, the water swallowing items were combined
116
to a single dichotomous indicator coded “1” if any swallowing items were positive and “0”
117
otherwise.
118
RI PT
113
Videofluoroscopic swallowing study (VFSS)
120
A VFSS, which served as the reference standard for screening item comparison, was completed
121
within 2 hours following the screening. Lateral radiographic views were obtained with the
122
fluoroscopic tube focused on the oral cavity and cervical esophagus. Samples were video-
123
recorded for later analysis. Participants completed two non-cued trials of self-administered
124
Varibar thin liquid bariuma (5ml, self-regulated single cup sip), and 5ml of Varibar barium
125
pudding.a A single trial of mastication and ingestion of 1/2 barium-coated cookie and sequential
126
swallowing of 90ml thin liquid barium were also completed. The SLP facilitated bolus
127
administration with hand-over-hand assistance if independent bolus delivery was impaired. The
128
VFSS began with the 5ml thin liquid volume and advanced accordingly unless a participant
129
demonstrated significant aspiration that could not be eliminated with therapeutic intervention.
M AN U
TE D
EP
AC C
130
SC
119
131
The presence and absence of aspiration was determined from the VFSS for each subject. The
132
individuals analyzing the VFSS were blinded to the clinical ED and experimental swallowing
133
screening results. Aspiration was identified when the bolus passed inferior to the true vocal folds
134
(TVFs). A single instance of the bolus inferior to the TVFs was used to identify aspiration unless
135
tracheal penetration was observed on the first swallow without subsequent evidence of
7
ACCEPTED MANUSCRIPT Aspiration screening for suspected stroke 136
aspiration. For these cases, aspiration was classified as absent unless the aspiration was so severe
137
that it warranted implementation of therapeutic intervention.
138
Inter-and intra-rater reliability measures for the presence or absence of aspiration on VFSS in a
140
subset of 20% of randomly selected participants was obtained using Cohen’s kappa. Reliability
141
was completed by two expert SLPs (SKD, JCR) each with extensive research experience in
142
examining VFSS. After initial training to criteria, the two blinded raters completed separate,
143
independent analysis of each VFSS with inter- and intra-rater reliability of k=1.00. One of these
144
two individuals completed ratings for the remaining VFSS.
M AN U
SC
RI PT
139
145
Statistical analyses
147
Demographic data were summarized using standard descriptive statistics. Logistic regression
148
was used to identify SSIs that showed statistically significant higher odds for aspiration.
149
Subsequently, sensitivity, specificity, negative predictive value (NPV), and positive predictive
150
value (PPV) relative to identifying VFSS determined aspiration were estimated for each item as
151
well as for the combination of items that were identified as statistically significant in predicting
152
aspiration by the logistic model. Receiver operating characteristic (ROC) curves were evaluated
153
to determine the performance of the models in predicting aspiration. Analyses were completed
154
using Stata 13.b
EP
AC C
155
TE D
146
156
Results
157
Mean age of the sample (N=250) was 64±9.5 years with 243 males. Discharge diagnoses for the
158
sample were confirmed stroke (n=156; 140-neuroimaging, 16 physician), transient ischemic
8
ACCEPTED MANUSCRIPT Aspiration screening for suspected stroke attack (TIA, n=49), and non-stroke/TIA diagnosis (n=45). The prevalence of aspiration on VFSS
160
was 12% (n=29). Prior stroke history was significantly higher in the aspiration group compared
161
to the non-aspiration group. The two aspirators without a discharge diagnosis of stroke/TIA had
162
a previous history of stroke, and five of the six aspirators with a TIA discharge diagnosis had a
163
previous history of stroke. None of these individuals had a previous history of dysphagia
164
associated with the prior stroke. No significant differences between aspirators and non-aspirators
165
were identified for other demographic data (Table 1). Mean time from admission to screening
166
was 43.0 ±25.8 hours. Screening and VFSS were completed within a 2 hour window (M=0.49±
167
0.24 hours). Screening duration as measured from RN introduction through completion of
168
assessment of wet voice after the 90ml bolus ingestion, or earlier if terminated due to positive
169
water swallow finding was 5.22±1.69 min. No adverse events, i.e., pneumonia, were evident
170
following testing.
M AN U
SC
RI PT
159
TE D
171
Only 1.6% (n=4) responses were undecided on the outcome of dysarthria and 0.8% (n=2) on the
173
combined water swallow item. Approximately 3.2% (8 of 250) patients demonstrated “other
174
vocal response” for cough (4 of 8 aspirated), which along with the undecided responses was
175
recoded as positive to evaluate validity indicators in order to maintain a conservative approach to
176
screening for aspiration. The final tally of positive and negative screens for all SSIs is included
177
in Table 4.
AC C
178
EP
172
179
Reliability
180
The inter-rater reliability between RNs and SLPs using Cohen’s kappa was measured for all the
181
SSIs and showed a range of 0.39 for lethargy to 1.0 for 1 minute wet voice. See Part 2 for
9
ACCEPTED MANUSCRIPT Aspiration screening for suspected stroke 182
expanded reliability results. Inter- and intra-rater reliability for identification of aspiration on
183
VFSS was k=1.00.
184
Validity
186
Logistic regression revealed age (p=0.012), dysarthria (p=0.001), abnormal volitional cough
187
(p=0.030), and the combined positive water swallow finding (p=0.021) to be significantly
188
associated with aspiration. Sensitivity, specificity, NPV, and PPV were estimated for the
189
combination of these items. Abnormal volitional cough was dropped from the final set of SSIs as
190
it did not contribute to improving validity after the rest of the items had been included (Table 3).
191
Additionally, both lethargy and 1 minute wet voice were excluded from our final set of items due
192
to their very small sample sizes (Table 4). The final validated tool, the Rapid Aspiration
193
Screening for Suspected Stroke (RAS3), contained the following items: 1) age >70, 2) dysarthria
194
(regardless of onset), 3) combined positive water swallow finding (cough, throat clear, wet voice,
195
inability to continuously drink 90ml) and yielded sensitivity = 93%, specificity = 43%, PPV =
196
18%, and NPV = 98% (Table 3, Appendix 2). Based on these results, the presence of any item
197
would indicate a failed screen without need to administer subsequent items. ROC curves (Figure
198
2) using predicted values (logits) from the logistic model without age and the logistic model with
199
age showed no statistical difference in the area under the ROC curve (p=0.301).
201
SC
M AN U
TE D
EP
AC C
200
RI PT
185
202
Discussion
203
Screening is an essential first step in the provision of individualized, patient-centered stroke care
204
to rapidly determine which patients presenting with suspected stroke are at risk of aspiration and
10
ACCEPTED MANUSCRIPT Aspiration screening for suspected stroke should remain non-oral with expedited SLP consultation and which patients are not at risk of
206
aspiration and can resume oral intake. Through rigorous research using evidence-based SSIs, the
207
objectives of our study were met by creating a valid nurse-administered aspiration screening tool
208
for patients presenting with suspected stroke. The RAS3 provides high sensitivity (0.93) and high
209
NPV (0.98) ensuring that aspiration risk is identified in the majority of patients yielding the
210
potential to reduce the occurrence of aspiration-associated complications. Specificity for the
211
aspiration screening tool was low but well within a range that is applied to most screenings with
212
the design to err in over- versus under-identification of an impairment to prevent negative health
213
outcomes.16
M AN U
SC
RI PT
205
214
To our knowledge, the RAS3 is the first aspiration screening tool validated in individuals
216
presenting with suspected stroke, as opposed to persons with confirmed stroke or specific stroke
217
severity level. As screening is frequently completed at the first point of contact in the ED, or
218
soon after admission, confirmation of stroke is routinely unknown. The final discharge diagnosis
219
in 45 of our participants was non-stroke/TIA; however, it was still part of the differential
220
diagnosis during screening and VFSS, thus they were not excluded from analysis. Moreover,
221
healthcare professionals have a tendency to selectively screen individuals based on stroke
222
severity and to not perform swallowing screening with low NIHSS scores.17,18 The median
223
NIHSS score is 3 in population-based studies.19 Our sample is representative of this mild stroke
224
trend as the average NIHSS for our cohort was 2.6. Our data, however, indicate that aspiration
225
can occur in individuals with a mild stroke. Indeed, pneumonia is not uncommon in unscreened
226
patients with minimal stroke-related deficits,17 thus supporting a screening tool validated in all
227
individuals for whom stroke/TIA is part of the differential diagnosis.
AC C
EP
TE D
215
11
ACCEPTED MANUSCRIPT Aspiration screening for suspected stroke 228
The RAS3 is composed of only three items: age >70, dysarthria, and positive response to water
230
swallowing with the two non-swallowing items administered first. Based on our findings, a
231
positive response on any RAS3 item indicates a failed screen. The entire screening takes less than
232
5 minutes given the average time of 5.22 minutes for completion of the more extensive set of
233
evaluated items. The brevity for completing the RAS3 will promote adoption by nurses and
234
supports recent research suggesting feasibility of screening swallowing in the ED.10 While
235
previous history of stroke was significantly greater in aspirators versus non-aspirators, it was not
236
analyzed as a screening item due to potential unreliability in a patient’s ability to provide this
237
information as well as the associated time delay in retrieving this information from medical
238
records.
M AN U
SC
RI PT
229
239
Notably, the area under ROC curve shows nearly equivalent performance in both logistic
241
regression models without and with age. Because increased age is associated with sensorimotor
242
changes and decreased physiological reserve both of which can increase risk of aspiration,17 the
243
identification of age >70 informs our choice to include it as a valid screening item. Moreover,
244
aging is inherently associated with processes than can increase the risk of pneumonia
245
development.21 However, the decision to include age >70 in the final set of SSIs can be based on
246
the acceptable trade-offs between true positive and false positive rates and the needs of the
247
healthcare environment.
EP
AC C
248
TE D
240
249
Dysarthria was strongly associated with aspiration in our study which supports previous
250
research.9,22 Although the incidence of chronic dysarthria was low, aspiration was evident in
12
ACCEPTED MANUSCRIPT Aspiration screening for suspected stroke 251
66% of the patients who presented with it suggesting that individuals with either chronic or new
252
onset of dysarthria are at increased risk of aspiration.
253
Most valid SSTs include a water swallow component as does the RAS3. Water protocols vary
255
and may include multiple trials of small volumes or a single 90ml volume. We selected a
256
progressive swallowing protocol as a safety measure as many stroke patients are unaware or
257
unconcerned by their dysphagia, do not monitor rate or volume of ingestion, and continue to
258
swallow even with overt signs such as coughing.23,24 These signs of aspiration risk related to the
259
water swallow are supported by the literature9,22 and are typically used in most screening tools.
M AN U
SC
RI PT
254
260
Abnormal volitional cough demonstrated a high likelihood of aspiration risk but had the broadest
262
standard error (Table 2) thus making it less reliable as a predictor. Notably, it did not improve
263
the validity measures (Table 3) in our sample when combined with other items. It may be that
264
our relatively small number of patients with aspiration (N=29, or 12% of our sample) limited the
265
sensitivity of our validity endpoints to the predictive information shown in our logistic model for
266
abnormal volitional cough. The low incidence of aspiration is likely related to the fact that
267
patients in the sample included individuals without confirmed stroke as well as patients with
268
mild stroke. Aspiration incidence is stroke is generally reported in patients with confirmed
269
stroke.9 To our knowledge, the incidence of aspiration has not been published in patients with
270
suspected stroke.
EP
AC C
271
TE D
261
272
It is also important to note that lethargy was present in only two patients, both of whom
273
aspirated. Even though lethargy was excluded due to low occurrence, we strongly suggest that
13
ACCEPTED MANUSCRIPT Aspiration screening for suspected stroke 274
individuals with any reduction in alertness remain non-oral with no screening of swallowing
275
until evidence of sustained wakefulness.
276
While the RAS3 identifies aspiration risk, dysphagia is common following stroke and should be
278
formally evaluated after the initial stroke workup. While not all individuals who aspirate develop
279
pneumonia, early identification of individuals at risk of aspiration ensures implementation of
280
safety measures until SLP evaluation is completed. Individuals without risk of aspiration are
281
cleared of any immediate swallowing safety concern and can start oral intake. Dysphagia is more
282
than aspiration. Even in the absence of aspiration, swallowing efficiency and pleasure in eating
283
can be affected by dysphagia. Given the complexities of swallowing and the need to evaluate
284
medical history, oral mechanism functioning, and swallowing of various consistencies, formal
285
dysphagia evaluation by an SLP is warranted in this high incidence population.
M AN U
SC
RI PT
277
TE D
286
Limitations
288
Limitations include the number of individuals who were excluded from participation. In the first
289
stage of development of a stroke swallowing screening tool, it is important to have stringent
290
inclusion/exclusion criteria to ensure that stroke-related aspiration is being measured. Stroke,
291
however, rarely occurs in isolation. The next step is to determine the generalizability of the RAS3
292
to all patients presenting with suspected stroke, including those who have co-morbid conditions
293
that may produce aspiration. Also important is extending these findings outside VAMCs in order
294
to include a greater number of females in the sample.
AC C
EP
287
295
14
ACCEPTED MANUSCRIPT Aspiration screening for suspected stroke Inability to obtain screening in the ED was an additional limitation. Given the focused stroke
297
workup in the ED, it was not feasible to consent participants as well as administer the SSIs and
298
obtain the VFSS within two hours. As patients with stroke can improve or decline rapidly,
299
obtaining a narrow timeline between screening and VFSS was critical for validity. Future studies
300
will evaluate RN reliability administering the RAS3 in the ED.
301
SC
302
RI PT
296
Conclusions
304
Screening for aspiration risk is a well-established best practice in providing safe, high quality
305
care for patients with suspected stroke. Early screening has the greatest potential to prevent
306
unsafe administration of oral intake to patients who are at risk of aspiration and expedites referral
307
to speech pathology for evaluation to reduce aspiration associated morbidity. The RAS3 provides
308
a valid method to determine aspiration risk in patients presenting with suspected stroke. High
309
sensitivity and NPV ensure that aspiration risk is identified in the majority of patients thus
310
reducing the occurrence of aspiration associated complications.
AC C
EP
TE D
M AN U
303
15
ACCEPTED MANUSCRIPT Aspiration screening for suspected stroke 311
Suppliers
312
a. Varibar barium sulfate Bracco Diagnostics Inc, 259 Prospect Plains Road, Building H, Monroe Township, NJ
314
08831
315 316
b. Stata Statistical Software: Release 13 (2013).
RI PT
313
StataCorp LP. 4905 Lakeway Drive, College Station, TX 77845-4512, USA
AC C
EP
TE D
M AN U
SC
317
16
ACCEPTED MANUSCRIPT Aspiration screening for suspected stroke 318
References
319
1.
protocols prevent pneumonia. Stroke 2005;36:1972-1976. 2.
insufficient evidence for guidelines. Dysphagia 2000;15:19-30.
322
3.
acute phase. Outcome and economic implications. Stroke 1993;24:1823-1827.
324 325
Odderson IR, McKenzie BS. A model for management of patients with stroke during the
4.
SC
323
Martino R, Pron G, Diamant N. Screening for oropharyngeal dysphagia in stroke:
RI PT
320 321
Hinchey JA, Shephard T, Furie K, Smith D, Wang D, Tonn S. Formal dysphagia screening
Jauch EC, Saver JL, Adams HP, Bruno A, Connors JJ, Demaerschalk BM, et al Guidelines for the early management of patients with acute ischemic stroke. A guideline for healthcare
327
professionals from the American Heart Association/American Stroke Association. Stroke
328
2013;44:870-947.
329
5.
M AN U
326
Edmiaston J, Connor LT, Steger-May K, Ford AL. A simple bedside stroke dysphagia screen, validated against videofluoroscopy, detects dysphagia and aspiration with high
331
sensitivity. J Stroke Cerebrovasc Dis 2014;23:712-716.
332
6.
TE D
330
Martino R, Silver F, Teasell R., Bayley M, Nicholson G, Streiner DL, Diamont NE. The Toronto Bedside Swallowing Screening Test (TOR-BSST): development and validation of
334
a dysphagia screening tool for patients with stroke. Stroke 2009;40:555-561. 7.
2008;23:244-250.
336 337
Suiter DB, Leder SB. Clinical utility of the 3 ounce water swallow test. Dysphagia
AC C
335
EP
333
8.
Trapl M, Enderle P, Nowotny M, Teuschl Y, Matz K, Dachenhausen A, et al. Dysphagia
338
bedside screening for acute-stroke patients: the gugging swallowing screen. Stroke
339
2007;38:2948-2952.
17
ACCEPTED MANUSCRIPT Aspiration screening for suspected stroke 9.
patients with acute stroke. Arch Phys Med Rehabil 1998;79:14-19.
341
10.
11.
stroke. Does dysphagia matter? Stroke 1996;27:1200-1204.
347 348
13.
14.
15.
16.
Streiner DL. Diagnosing tests: using and misusing diagnostic and screening tests J Personality Assessment 2003; 81:209-219.
355 356
Polit DF, Beck CT. Nursing Research: Generating and Assessing Evidence for Nursing Practice. 8th Ed. New York: Wolters Kluwer; 2008.
353 354
Daniels SK, Anderson JA, Willson PC. Valid items for screening dysphagia risk in patients with stroke: a systematic review. Stroke 2012;43:892-897.
351 352
Daniels SK, Brailey K, Priestly DH, Herrington LR, Weisberg LA, Foundas AL. Aspiration in patients with acute stroke. Arch Phys Med Rehabil 1998;79:14-19.
349 350
Smithard DG, O'Neill PA, Parks C, Morris J. Complications and outcome after acute
SC
12.
M AN U
346
Mann G, Hankey GJ, Cameron D. Swallowing function after stroke: prognosis and prognostic factors at 6 months. Stroke 1999;30:744-748.
345
RI PT
the emergency department. Nursing Research and Practice 2013;304190.
343 344
Daniels SK, Anderson JA, Petersen NJ. Implementation of stroke dysphagia screening in
TE D
342
Daniels SK, Brailey K, Priestly DH, Harrington LR, Weisberg LA, et al. Aspiration in
17.
EP
340
Lakshminarayan K, Tsai AW, Tong X, Vazquez G, Peacock JM, George MG, et al. Utility of dysphagia screening results in predicting poststroke pneumonia. Stroke 2010;41:2849-
358
2854.
359
18.
AC C
357
Masrur S, Smith EE, Saver JL, Reeves MJ, Bhatt DL, Zhao X, et al. Dysphagia screening
360
and hospital-acquired pneumonia in patients with acute ischemic stroke: findings from Get
361
with the Guidelines-Stroke. J Stroke Cerebrovasc Dis 2013;22:e301-e309.
18
ACCEPTED MANUSCRIPT Aspiration screening for suspected stroke 362
19.
Reeves M, Khoury J, Alwell K, Moonmaw C, Flaherty M, Woo D, et al. Distribution of
363
National Institutes of Health Stroke Scale in the Cincinnati/Northern Kentucky Stroke
364
Study. Stroke 2013;44:3211-3213. 20.
Nicosa MA, Hind JA, Roecker EB, Carnes M, Doyle J, Dengel GA, et al. Age effects on
RI PT
365 366
the temporal evolution of isometric and swallowing pressures. J Gerontol A Biol Sci Med
367
Sci 2000;55:M634-M640. 21.
Chumbler NR, Williams LS, Wells CK, Lo AC, Nadeau S, Peixoto AJ. Derivation and
SC
368
validation of a clinical system for predicting pneumonia in acute stroke. Neuroepidemiol
370
2010;34:193-199.
371
22.
M AN U
369
McCullough GH, Rosenbek JC, Wertz RT, McCoy S, Mann G, McCullough K. Utility of
372
clinical swallowing examination measures for detecting aspiration post-stroke. J Speech
373
Lang Hear Res 2005;48:1280-1293. 23.
Parker C, Power M, Hamdy S, Bowen A, Tyrrell P, Thompson DG. Awareness of
TE D
374 375
dysphagia by patients following stroke predicts swallowing performance. Dysphagia
376
2004;19:28-35. 24.
Schroeder MF, Daniels SK, McClain M, Corey DM, Foundas AL. (2006). Clinical and
EP
377
cognitive predictors of swallowing recovery in stroke. J Rehabil Res Develop 2006;43:301-
379
310.
380
AC C
378
19
ACCEPTED MANUSCRIPT Aspiration screening for suspected stroke Figure Legend
382
Figure 1
Study recruitment and testing flow diagram
383
Figure 2
Comparison of ROC curves using the logistic regression models with and without
384
age in predicting aspiration.
AC C
EP
TE D
M AN U
SC
RI PT
381
20
ACCEPTED MANUSCRIPT
Table 1
Demographic data Total (N=250)
Aspirators (n=29)
Age, yr, mean (SD) Male, n (%)
64 (9.5) 243 (97)
66.8 (8.3) 29 (100)
NonAspirators (n=221) 63.9 (9.6) 214 (97)
Caucasian Race, n (%)
130 (52)
16 (55)
114 (52)
Previous stroke history, n (%) NIHSS score, mean (SD) Mild (0-5), n (%) Moderate (6-13), n (%) Severe (>13), n (%) Discharge Diagnosis, n (%) Imaging confirmed stroke Physician confirmed stroke Physician confirmed TIA Non-stroke/TIA
60 (24)
17 (59)
43 (19.5)
2.6 (2.9)
3.8 (3.9)
2.5 (2.7)
219 (88) 30 (12)
23 (79) 5 (17)
Stroke Type, n (%)
16 (6)
1 (3.5)
15 (7)
6 (21)
43 (19.5)
2 (7)
43 (19.5)
Total (N=140)
Aspirators (n=20)
NonAspirators (n=120)
RHD LHD Bilateral Posterior circulation *p-value<0.05
51 (36) 50 (36) 4 (3) 35 (25)
AC C
137 (98) 3 (2)
20 (100) 0 (0)
5 (25) 7 (35) 0 (0) 8 (40)
0.082 0.417
SC 120 (54)
45 (18)
Ischemic Hemorrhagic Stroke location, n (%)
t-test Fisher’s Exact Pearson Chi-sq Pearson chi-sq t-test
0.716
<0.001* 0.091
0 (0)
20 (69)
TE D
49 (20)
1 (3.5)
p-value
196 (87) 25 (11)
M AN U
140 (56)
EP
Imaging confirmed stroke (only)
1 (<1)
Test type
RI PT
Variable
Fisher’s Exact
0.318
Fisher’s Exact
0.627
Fisher’s Exact
0.379
117 (97.5) 3 (2.5)
46 (38) 43 (36) 4 (3) 27 (22.5)
ACCEPTED MANUSCRIPT
Table 2
Logistic regression results Odds Ratio
SE
z
p-value
Lower Limit Upper Limit
3.396
1.653 2.512 0.012*
1.308
8.815
Dysarthria
5.114
2.410 3.463 0.001**
2.031
12.878
Non-swallow Wet Voice
2.191
2.524 0.681
0.229
20.964
Abnormal Cough
4.094
2.664 2.166 0.030*
1.143
14.656
Water Swallow
3.034
1.454 2.316 0.021*
1.186
7.760
SC
0.496
RI PT
Age >70
N=250; Pseudo R-square = 0.233
AC C
EP
TE D
M AN U
*p-value <0.05; **p-value <0.01
ACCEPTED MANUSCRIPT
Validity indicators for item combinations with significant odds ratios
Items
Negative Screen
VFSS (n=250) Abnormal
25
4
Normal
112
109
Abnormal
27
2
Normal
126
95
Normal
27
2
Abnormal
129
92
Dysarthria OR WS
Age >70 OR Dysarthria OR WS
Specificity (95% CI)
86.2 (68.3-96.1)
49.3 (42.6-56.1)
18.2 (12.2-25.7)
96.5 (91.2-99.0)
93.1 (77.2-99.2)
43.0 (36.4-49.8)
17.6 (12.0-24.6)
97.9 (92.7-99.7)
17.3 (11.7-24.2)
97.9 (92.7-99.7)
93.1 (77.2-99.2)
41.6 (35.1-48.4)
AC C
EP
TE D
M AN U
Abnormal Cough OR Age >70 OR Dysarthria OR WS WS=water swallow
Sensitivity (95% CI)
PPV (95% CI)
RI PT
Positive Screen
SC
Table 3
NPV (95% CI)
ACCEPTED MANUSCRIPT
Table 4
Validity indicators for individual SSIs tested Positive Screen
Negative Screen
VFSS
Dysarthriaa
Non-swallow Wet Voice
Abnormal Cough
Combined Water Swallowa
Normal
38
183
Abnormal
1
28
Normal
1
220
Abnormal
18
11
Normal
43
178
Abnormal
3
26
Normal
3
Abnormal
8
Normal
9
Abnormal
21
Normal
89
132
Abnormal
9
20
Normal
17
204
Abnormal
12
17
Normal
62
159
Abnormal
8
21
Normal
15
206
Abnormal
2
13
Normal
13
158
Abnormal
0
15
Individual Water Swallow Items
AC C
Throat Clear
EP
Cough
Wet Voice
Inability to Drink 90mlc
250
37.9
250
3.5
250
Specificity
218
PPV
NPV
RI PT
18
Sensitivity
62.1
M AN U
Lethargy
11
82.8
22.4
91.0
99.5
50.0
88.7
80.5
29.5
94.2
250
10.3
98.6
50.0
89.3
250
27.6
95.9
47.1
91.0
250
72.4
59.7
19.1
94.3
250
31.0
92.3
34.6
91.1
250
41.4
71.9
16.2
90.3
250
27.6
93.2
34.8
90.7
186
13.3
92.4
13.3
92.4
21
212 8
TE D
Age >70a
Abnormal
n
SC
Items
Wet Voice after 1 minuteb,c Normal 2 166 183 0.0 98.8 0.0 91.7 a=Items in validated RAS3; b=Not part of final combined water swallow; c=n< 250 due to missing information on sequential swallows with a positive preceding WS item
ACCEPTED MANUSCRIPT
Figure 1
RI PT
Potentially eligible participants with suspected stroke n=1042
Eligible participants n=406
M AN U
SC
Excluded (n=636) Non-stroke neurological disease (n=202) Head/neck structural changes (n=111) Dysphagia unrelated to stroke (n=88) TIA/stroke ruled out n= (51) Not competent/no legally authorized representative (n=47) Medically unstable (n=46) Repeat stroke, previously participated (n=43) >3 weeks from symptom onset (n=25) >5 days from admission (n=16) Child bearing potential female (n=4) Deceased (n=3)
TE D
No Swallowing Screening Test (n=148) Declined consent (n=79) Discharged prior to approaching (n=58) Consented but discharged, medical procedure, unstable (n=11)
Swallowing Screening Test participants n=258
RAS3 negative n=100
AC C
EP
RAS3 positive n=158
VFSS completed n=158
Excluded from analyses (n=5) b
Aspiration present (n=27) Aspiration absent (n=126)
No VFSS n=2a
VFSS completed n=98
Excluded from analyses (n=1) b
Aspiration present (n=2) Aspiration absent (n=95)
Note: a=Reasons for no VFSS: C-spine surgery evident on x-ray (n=1), NPO for surgery (n=1); b=Reasons for exclusion from VFSS analyses: Dysphagia history prior to stroke (n=2), Equipment failure (n=1), C-spine damage (n=3)
ACCEPTED MANUSCRIPT
SC
0.50 0.00
0.25
M AN U
0.25 0.00
0.50 1-Specificity
Without age AUC= 0.77(0.048)* Reference
EP
TE D
*Area Under ROC Curve with Standard Error (SE)
AC C
Sensitivity
0.75
RI PT
1.00
Figure 2
0.75
1.00
With age AUC=0.80(0.043)*
ACCEPTED MANUSCRIPT
Appendix 1. Tested Swallowing Screening Items
AC C
EP
TE D
M AN U
SC
RI PT
Non-Swallowing Section 1. Male Yes No 2. Age >70 Yes No Ask patient his/her age. Confirm in medical records 3. Lethargy (Drowsy) Yes No Undecided Unarousable by minor stimulation (requires verbalization or intermittent tactile stimulation-touching/shaking to respond). If evident, score as YES. If awake and alert (regardless of orientation or correctness), score as NO. 4. Dysarthria (Slurred Speech) Yes No Undecided No/Limited Response Ask patient “Tell me, what happened to bring you here.” If speech is slurred or not precise, score as YES. If speech is clear and precise (regardless of content), score as NO. If the person makes no verbal response or response is too limited to judge, score as NO/LIMITED RESPONSE. If still unsure of how to score, score as UNDECIDED. 5. Wet Voice Yes No Undecided No/Limited Response Tell patient to “say AHHHH.” If the voice is wet or gurgly sounding, score as YES. If not, score as NO. If no vocal response, score as NO RESPONSE 6. Abnormal Volitional Cough* Yes No Undecided No Response Other Vocal Response Tell patient to “Cough as strongly as you can, like this…,” demonstrate an example of a strong, forceful cough. Score as YES if the cough is weak or breathy. Score as NO if the person produces a STRONG, FORCEFUL cough. Score as NO RESPONSE if no vocal response. Score as OTHER VOCAL RESPONSE if person says the word “Cough”, produces a throat clear or some other vocal response that is not an actual cough. Swallowing Section 1. 5 ml via cup. Tell patient to “Drink the small amount of water in one swallow.” After swallowing, have patient say “AHHHH”. Listen for presence of WET GURGLY VOICE when saying AHHHH, or COUGH or AUDIBLE THROAT CLEAR. Score as YES if any are present following swallow. A. Cough after Swallow Yes No Undecided B. Audible Throat Clear after Swallow Yes No Undecided C. Wet Voice with AHHHHH Yes No Undecided No Response If any item is YES, STOP screening. DO NOT administer any more water. 2. 5 ml via cup. If none are YES on the first swallow, repeat the 5 ml water swallow using same instructions. A. Cough after Swallow Yes No Undecided B. Audible Throat Clear after Swallow Yes No Undecided C. Wet Voice with AHHHHH Yes No Undecided No Response If any item is YES, STOP screening. DO NOT administer any more water. 3. 90 ml via cup. Tell patient “Drink this entire amount of water without stopping, Keep the cup up to your lips and keep drinking, swallow after swallow, until the water is gone or I tell you to stop.” After drinking have the patient say “AHHHH”. Listen for presence of WET GURGLY VOICE when saying AHHHH, or COUGH or THROAT CLEAR. Score as YES if any are present during or after the swallow. If the patient cannot continuously drink the water, even with encouragement, score as YES. A. Cough after Swallow Yes No Undecided B. Audible Throat Clear after Swallow Yes No Undecided C. Wet Voice with AHHHHH Yes No Undecided No Response D. Unable to Continuously Drink Yes No Undecided E. Wet Voice with AHHHHH after 1 min Yes No Undecided No Response
ACCEPTED MANUSCRIPT
Appendix 2. Rapid Aspiration Screening for Suspected Stroke (RAS3) If the patient has reduced level of consciousness cannot have the head of the bed elevated, or is currently receiving tube feeding, DONOT screen swallowing; patient remains nothing by mouth.
Non-Swallowing Section 1. Age >70*
Yes
No
RI PT
If any item below is YES, the screening is positive. STOP screening. Patient remains NPO and Speech Pathology is consulted
2. Dysarthria (Slurred Speech) Yes No Ask patient “Tell me, what happened to bring you here.” Prompt for more speech if limited verbalization. If speech is slurred or not precise, score as YES. If speech is clear and precise (regardless of content), score as NO. If undecided, score as YES. If either item is YES, STOP screening. DO NOT administer any water.
EP
TE D
M AN U
SC
Swallowing Section Need syringes to measure water, cup, and straw 1. 5ml via cup. Tell patient to “Drink the small amount of water in one swallow.” After swallowing, have patient say “AHHHH”. Listen for presence of WET GURGLY VOICE when saying AHHHH, or COUGH or AUDIBLE THROAT CLEAR. Score as YES if any item is present following swallow. A. Cough after Swallow Yes No B. Audible Throat Clear after Swallow Yes No C. Wet Voice with AHHHHH Yes No If any item is YES, STOP screening. DO NOT administer any more water. 2. 5ml via cup. Repeat the 5 ml water swallow using same instructions. A. Cough after Swallow Yes No B. Audible Throat Clear after Swallow Yes No C. Wet Voice with AHHHHH Yes No If any item is YES, STOP screening. DO NOT administer any more water. 3. 90ml via cup. Tell patient “Drink this entire amount of water without stopping. Keep the cup up to your lips and keep drinking, swallow after swallow, until the water is gone or I tell you to stop.” After drinking have the patient say “AHHHH”. Encourage patient to keep swallowing if is pausing between swallows is evident. Listen for presence of WET GURGLY VOICE when saying AHHHH, or COUGH or THROAT CLEAR. Stop screening and remove cup if coughing, throat clearing, wet voice, and/or inability to continuous drink during swallowing. Score as YES if any item is present during or after the swallow. If the patient cannot continuously drink the water, even with encouragement, score as YES. A. Cough after Swallow Yes No B. Audible Throat Clear after Swallow Yes No C. Wet Voice with AHHHHH Yes No D. Unable to Continuously Drink Yes No
AC C
*Denotes item than can be omitted from RAS3
S0003-9993(16)30076-4
DOI:
10.1016/j.apmr.2016.03.025
Reference:
YAPMR 56523
To appear in:
ARCHIVES OF PHYSICAL MEDICINE AND REHABILITATION
Received Date: 7 March 2016 Accepted Date: 29 March 2016
Please cite this article as: Daniels SK, Pathak S, Rosenbek JC, Morgan RO, Anderson JA, The Rapid Aspiration Screening in Suspected Stroke Part 1: Development and Validation, ARCHIVES OF PHYSICAL MEDICINE AND REHABILITATION (2016), doi: 10.1016/j.apmr.2016.03.025. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT Aspiration screening for suspected stroke The Rapid Aspiration Screening in Suspected Stroke Part 1: Development and Validation Stephanie K. Daniels, PhD,1,2 Shweta Pathak, MPH,3 John C. Rosenbek, PhD,4 Robert O.
1
RI PT
Morgan, PhD,3 Jane A. Anderson, PhD5,6
Research Service Line, Michael E. DeBakey VA Medical Center, Houston, TX; 2Department of
Communication Sciences and Disorders, University of Houston, Houston, TX; 3School of Public
SC
Health, University of Texas Health Sciences Center, Houston, TX; 4Department of
M AN U
Communication Sciences and Disorders, University of Florida, Gainesville, FL; 5Health Services Research and Development Center of Excellence, Michael E. DeBakey VA Medical Center, Houston, TX; 6Department of Neurology, Baylor College of Medicine, Houston TX; The authors would like to acknowledge the contribution of the research coordinators, speech pathologists, and neurologists who facilitated subject recruitment and data collection and also the
TE D
nurses who participated in the study. The project was supported by a Merit Award (1I0RX00121) from Rehabilitation Research & Development of the VA Office of Research and Development Service. The views expressed in this article are those of the authors and do not
EP
necessarily represent the views of the Department of Veterans Affairs or the University of Texas
AC C
School of Public Health. The full study protocol can be accessed at the Department of Veterans Affairs.
Device Status: The manuscript submitted does not contain information about medical device(s). SKD, SP, JCR, and ROM received salary support from the Department of Veterans Affairs-Merit Award (1I0RX00121). JAA has no conflict of interest.
ACCEPTED MANUSCRIPT Aspiration screening for suspected stroke For Correspondence and Reprints: Stephanie K. Daniels, PhD, Research Service (151), Michael E. DeBakey VA Medical Center, 2002 Holcombe Blvd, Houston, TX 77030, Phone: 713-794-
AC C
EP
TE D
M AN U
SC
RI PT
7152, FAX: 713-794-7786, Email: [email protected]
ACCEPTED MANUSCRIPT Aspiration screening for suspected stroke 1
The Rapid Aspiration Screening in Suspected Stroke
2
Part 1: Development and Validation Abstract
4
OBJECTIVES: To develop and validate a nurse-administered screening tool to identify
5
aspiration risk in patients with suspected stroke.
6
DESIGN: Validity study comparing evidence-based swallowing screening items with
7
videofluoroscopic swallowing study (VFSS).
8
SETTINGS: The study was completed in a certified primary stroke center in a major
9
metropolitan medical facility.
M AN U
SC
RI PT
3
PARTICIPANTS: Consecutive patients (N=250) admitted with suspected stroke.
11
INTERVENTIONS: Patients were administered evidence-based swallowing screening items by
12
nurses. A VFSS was completed within 2 hours of swallowing screening.
13
MAIN OUTCOME MEASURE: Validity relative to identifying VFSS determined aspiration for
14
each screening item and for various combinations of items.
15
RESULTS: Aspiration was identified in 29 of the 250 participants. Logistic regression revealed
16
age (p=0.012), dysarthria (p=0.001), abnormal volitional cough (p=0.030), and signs related to
17
trial water swallow (p=0.021) to be significantly associated with aspiration. Validity was then
18
determined based on the best combination of significant items for predicting aspiration. Results
19
revealed that age >70, or dysarthria, or signs related to trial water swallow (i.e., cough, throat
20
clear, wet vocal quality, and inability to continuously swallow 90ml water) yielded 93%
21
sensitivity and 98% negative predictive value.
AC C
EP
TE D
10
1
ACCEPTED MANUSCRIPT Aspiration screening for suspected stroke CONCLUSIONS: The final validated tool, Rapid Aspiration Screening for Suspected Stroke
23
(RAS3), is a valid nurse-administered tool to detect risk of aspiration in patients presenting with
24
suspected stroke.
25
Key Words: Deglutition Disorders, Stroke, Screening, Validity
AC C
EP
TE D
M AN U
SC
RI PT
22
2
ACCEPTED MANUSCRIPT Aspiration screening for suspected stroke List of Abbreviations
27
AUC
area under curve
28
CI
confidence interval
29
ED
emergency department
30
LAR
legally authorized representative
31
NPV
negative predictive value
32
NIHSS
National Institutes of Health Stroke Scale
33
PPV
positive predictive value
34
RAS3
Rapid Aspiration Screening for Suspected Stroke
35
ROC
receiver operating characteristic
36
RN
registered nurse
37
SLP
speech-language pathologist
38
SSI
swallowing screening item
39
SST
swallowing screening tool
40
TIA
transient ischemic attack
41
TVF
true vocal fold
42
VFSS
videofluoroscopic swallowing study
43
WS
water swallow
AC C
EP
TE D
M AN U
SC
RI PT
26
3
ACCEPTED MANUSCRIPT Aspiration screening for suspected stroke Early detection of aspiration risk in patients with suspected stroke is critical for immediate
45
intervention to reduce mortality, morbidity, and healthcare costs.1-3 These findings have led to
46
the recommendation of screening swallowing, regardless of stroke severity, in individuals
47
presenting with suspected stroke.4
RI PT
44
48
Swallowing screening tools (SSTs) administered by frontline healthcare providers such as
50
registered nurses (RNs) have been developed in individuals with stroke;5-8 however, all are
51
validated in patients with confirmed stroke, not individuals presenting with suspected stroke.
52
This is a limitation as screening is completed early in the care process, either in the emergency
53
department (ED) or upon hospital admission, frequently before confirmation of stroke diagnosis.
54
Moreover, SST validation has been limited to studies in which individuals were referred for
55
swallowing evaluation,7,8 or had specific National Institutes of Health Stroke Scale (NIHSS)
56
scores.6 These factors limit generalization of SSTs to real-world application.
M AN U
TE D
57
SC
49
Finally, most stroke SSTs have focused on identification of dysphagia, not aspiration.6,8 While
59
the incidence of dysphagia in stroke is high,9 not all patients who have dysphagia aspirate.
60
During the acute stroke workup, the focus should be on rapidly determining aspiration risk for
61
patient safety in order to quickly provide oral medication and address patient needs such as fluid
62
intake. After the emergent stroke workup, formal dysphagia evaluation by a speech-language
63
pathologist (SLP) is necessary to fully evaluate oropharyngeal swallowing and determine if an
64
instrumental swallowing evaluation is warranted. The aim of this study was to develop a valid
65
nurse-administered screening tool for the rapid identification of aspiration risk in patients
66
presenting with suspected stroke. The a priori hypothesis was that a combination of non-
AC C
EP
58
4
ACCEPTED MANUSCRIPT Aspiration screening for suspected stroke 67
swallowing and swallowing screening items would provide the highest validity for identification
68
of aspiration risk.
69
RI PT
70
Methods
72
Participants
73
Medical records of patients consecutively admitted for stroke workup to the Michael E. DeBakey
74
Veterans Affairs Medical Center (MEDVAMC) or who had a suspected stroke during
75
hospitalization between October 2012 and November 2014 were reviewed for eligibility. A
76
clinical, non-validated swallowing screening was completed in the ED as part of the hospital’s
77
stroke protocol.10 Patients were approached for participation regardless of the clinical screening
78
results. Consented participants met all the inclusion/exclusion criteria with a total of 250 patients
79
submitted for data analyses (Figure 1). RNs were recruited from MEDVAMC stroke wards to
80
administer and interpret swallowing screening items (SSIs) on consented participants (see Part 2
81
for RN details). The determination of sample size was based on the existing literature on
82
prevalence of aspiration.11-13 The study was approved by the Institutional Review Board at the
83
Baylor College of Medicine and the Research and Development Committee at MEDVAMC in
84
Houston, TX, and all individuals provided written consent prior to participation.
M AN U
TE D
EP
AC C
85
SC
71
86
Screening items
87
SSIs were obtained from a systematic literature review.14 Items were submitted to 10 content
88
experts (SLPs and RNs who had greater than 10 years of experience working with stroke and/or
89
dysphagia in VAMCs) who rated the items according to usefulness in detecting aspiration and
5
ACCEPTED MANUSCRIPT Aspiration screening for suspected stroke feasibility for nursing implementation. Content expert agreement was determined utilizing
91
content validity index methodology.15 The final SSIs evaluated for validity were six non-
92
swallowing items: male, age >70 years, lethargy, dysarthria, wet vocal quality, abnormal
93
volitional cough and five swallowing items: cough after swallowing, throat clear after
94
swallowing, wet vocal quality after swallowing, inability to continuously drink 90ml water
95
without stopping, wet vocal quality after 1 minute. Most are similar to items on available SSTs.5-
96
8
97
criteria assigned by the content experts.
RI PT
90
99
M AN U
98
SC
Other items from the literature review were not part of the tested SSIs as they did not meet
All non-swallowing items were assessed regardless of present or absent score. Non-swallowing items were followed by administration of at least one water swallow trial. The swallowing
101
portion of the screening consisted of three water swallow trials (5ml twice, 90ml) administered
102
via a cup or straw. If any sign of aspiration risk (cough, throat clear, wet voice) was present after
103
either 5ml water swallow, the screening was stopped, and subsequent water swallow trials were
104
not administered. If no aspiration sign was present for either 5ml trial, the 90ml trial was
105
administered, and the patient was assessed for ability to continuously drink and vocal quality
106
after 1 minute in addition to the same signs of aspiration risk identified on smaller water
107
volumes.
EP
AC C
108
TE D
100
109
Screeners were blinded to the clinical ED screening results. They were instructed to score items
110
as yes (present), no (not present), undecided, or no/limited response (Appendix 1; see Part 2 for
111
details on training). For scoring of abnormal volitional cough, a weak or breathy cough was
112
differentiated from other vocal responses (throat clear, verbalizes the word “cough”, or
6
ACCEPTED MANUSCRIPT Aspiration screening for suspected stroke production any other response that was not an actual cough) in order to determine differences in
114
relationship to aspiration. Because individual screenings had variable numbers of swallowing
115
items depending on whether screening was stopped, the water swallowing items were combined
116
to a single dichotomous indicator coded “1” if any swallowing items were positive and “0”
117
otherwise.
118
RI PT
113
Videofluoroscopic swallowing study (VFSS)
120
A VFSS, which served as the reference standard for screening item comparison, was completed
121
within 2 hours following the screening. Lateral radiographic views were obtained with the
122
fluoroscopic tube focused on the oral cavity and cervical esophagus. Samples were video-
123
recorded for later analysis. Participants completed two non-cued trials of self-administered
124
Varibar thin liquid bariuma (5ml, self-regulated single cup sip), and 5ml of Varibar barium
125
pudding.a A single trial of mastication and ingestion of 1/2 barium-coated cookie and sequential
126
swallowing of 90ml thin liquid barium were also completed. The SLP facilitated bolus
127
administration with hand-over-hand assistance if independent bolus delivery was impaired. The
128
VFSS began with the 5ml thin liquid volume and advanced accordingly unless a participant
129
demonstrated significant aspiration that could not be eliminated with therapeutic intervention.
M AN U
TE D
EP
AC C
130
SC
119
131
The presence and absence of aspiration was determined from the VFSS for each subject. The
132
individuals analyzing the VFSS were blinded to the clinical ED and experimental swallowing
133
screening results. Aspiration was identified when the bolus passed inferior to the true vocal folds
134
(TVFs). A single instance of the bolus inferior to the TVFs was used to identify aspiration unless
135
tracheal penetration was observed on the first swallow without subsequent evidence of
7
ACCEPTED MANUSCRIPT Aspiration screening for suspected stroke 136
aspiration. For these cases, aspiration was classified as absent unless the aspiration was so severe
137
that it warranted implementation of therapeutic intervention.
138
Inter-and intra-rater reliability measures for the presence or absence of aspiration on VFSS in a
140
subset of 20% of randomly selected participants was obtained using Cohen’s kappa. Reliability
141
was completed by two expert SLPs (SKD, JCR) each with extensive research experience in
142
examining VFSS. After initial training to criteria, the two blinded raters completed separate,
143
independent analysis of each VFSS with inter- and intra-rater reliability of k=1.00. One of these
144
two individuals completed ratings for the remaining VFSS.
M AN U
SC
RI PT
139
145
Statistical analyses
147
Demographic data were summarized using standard descriptive statistics. Logistic regression
148
was used to identify SSIs that showed statistically significant higher odds for aspiration.
149
Subsequently, sensitivity, specificity, negative predictive value (NPV), and positive predictive
150
value (PPV) relative to identifying VFSS determined aspiration were estimated for each item as
151
well as for the combination of items that were identified as statistically significant in predicting
152
aspiration by the logistic model. Receiver operating characteristic (ROC) curves were evaluated
153
to determine the performance of the models in predicting aspiration. Analyses were completed
154
using Stata 13.b
EP
AC C
155
TE D
146
156
Results
157
Mean age of the sample (N=250) was 64±9.5 years with 243 males. Discharge diagnoses for the
158
sample were confirmed stroke (n=156; 140-neuroimaging, 16 physician), transient ischemic
8
ACCEPTED MANUSCRIPT Aspiration screening for suspected stroke attack (TIA, n=49), and non-stroke/TIA diagnosis (n=45). The prevalence of aspiration on VFSS
160
was 12% (n=29). Prior stroke history was significantly higher in the aspiration group compared
161
to the non-aspiration group. The two aspirators without a discharge diagnosis of stroke/TIA had
162
a previous history of stroke, and five of the six aspirators with a TIA discharge diagnosis had a
163
previous history of stroke. None of these individuals had a previous history of dysphagia
164
associated with the prior stroke. No significant differences between aspirators and non-aspirators
165
were identified for other demographic data (Table 1). Mean time from admission to screening
166
was 43.0 ±25.8 hours. Screening and VFSS were completed within a 2 hour window (M=0.49±
167
0.24 hours). Screening duration as measured from RN introduction through completion of
168
assessment of wet voice after the 90ml bolus ingestion, or earlier if terminated due to positive
169
water swallow finding was 5.22±1.69 min. No adverse events, i.e., pneumonia, were evident
170
following testing.
M AN U
SC
RI PT
159
TE D
171
Only 1.6% (n=4) responses were undecided on the outcome of dysarthria and 0.8% (n=2) on the
173
combined water swallow item. Approximately 3.2% (8 of 250) patients demonstrated “other
174
vocal response” for cough (4 of 8 aspirated), which along with the undecided responses was
175
recoded as positive to evaluate validity indicators in order to maintain a conservative approach to
176
screening for aspiration. The final tally of positive and negative screens for all SSIs is included
177
in Table 4.
AC C
178
EP
172
179
Reliability
180
The inter-rater reliability between RNs and SLPs using Cohen’s kappa was measured for all the
181
SSIs and showed a range of 0.39 for lethargy to 1.0 for 1 minute wet voice. See Part 2 for
9
ACCEPTED MANUSCRIPT Aspiration screening for suspected stroke 182
expanded reliability results. Inter- and intra-rater reliability for identification of aspiration on
183
VFSS was k=1.00.
184
Validity
186
Logistic regression revealed age (p=0.012), dysarthria (p=0.001), abnormal volitional cough
187
(p=0.030), and the combined positive water swallow finding (p=0.021) to be significantly
188
associated with aspiration. Sensitivity, specificity, NPV, and PPV were estimated for the
189
combination of these items. Abnormal volitional cough was dropped from the final set of SSIs as
190
it did not contribute to improving validity after the rest of the items had been included (Table 3).
191
Additionally, both lethargy and 1 minute wet voice were excluded from our final set of items due
192
to their very small sample sizes (Table 4). The final validated tool, the Rapid Aspiration
193
Screening for Suspected Stroke (RAS3), contained the following items: 1) age >70, 2) dysarthria
194
(regardless of onset), 3) combined positive water swallow finding (cough, throat clear, wet voice,
195
inability to continuously drink 90ml) and yielded sensitivity = 93%, specificity = 43%, PPV =
196
18%, and NPV = 98% (Table 3, Appendix 2). Based on these results, the presence of any item
197
would indicate a failed screen without need to administer subsequent items. ROC curves (Figure
198
2) using predicted values (logits) from the logistic model without age and the logistic model with
199
age showed no statistical difference in the area under the ROC curve (p=0.301).
201
SC
M AN U
TE D
EP
AC C
200
RI PT
185
202
Discussion
203
Screening is an essential first step in the provision of individualized, patient-centered stroke care
204
to rapidly determine which patients presenting with suspected stroke are at risk of aspiration and
10
ACCEPTED MANUSCRIPT Aspiration screening for suspected stroke should remain non-oral with expedited SLP consultation and which patients are not at risk of
206
aspiration and can resume oral intake. Through rigorous research using evidence-based SSIs, the
207
objectives of our study were met by creating a valid nurse-administered aspiration screening tool
208
for patients presenting with suspected stroke. The RAS3 provides high sensitivity (0.93) and high
209
NPV (0.98) ensuring that aspiration risk is identified in the majority of patients yielding the
210
potential to reduce the occurrence of aspiration-associated complications. Specificity for the
211
aspiration screening tool was low but well within a range that is applied to most screenings with
212
the design to err in over- versus under-identification of an impairment to prevent negative health
213
outcomes.16
M AN U
SC
RI PT
205
214
To our knowledge, the RAS3 is the first aspiration screening tool validated in individuals
216
presenting with suspected stroke, as opposed to persons with confirmed stroke or specific stroke
217
severity level. As screening is frequently completed at the first point of contact in the ED, or
218
soon after admission, confirmation of stroke is routinely unknown. The final discharge diagnosis
219
in 45 of our participants was non-stroke/TIA; however, it was still part of the differential
220
diagnosis during screening and VFSS, thus they were not excluded from analysis. Moreover,
221
healthcare professionals have a tendency to selectively screen individuals based on stroke
222
severity and to not perform swallowing screening with low NIHSS scores.17,18 The median
223
NIHSS score is 3 in population-based studies.19 Our sample is representative of this mild stroke
224
trend as the average NIHSS for our cohort was 2.6. Our data, however, indicate that aspiration
225
can occur in individuals with a mild stroke. Indeed, pneumonia is not uncommon in unscreened
226
patients with minimal stroke-related deficits,17 thus supporting a screening tool validated in all
227
individuals for whom stroke/TIA is part of the differential diagnosis.
AC C
EP
TE D
215
11
ACCEPTED MANUSCRIPT Aspiration screening for suspected stroke 228
The RAS3 is composed of only three items: age >70, dysarthria, and positive response to water
230
swallowing with the two non-swallowing items administered first. Based on our findings, a
231
positive response on any RAS3 item indicates a failed screen. The entire screening takes less than
232
5 minutes given the average time of 5.22 minutes for completion of the more extensive set of
233
evaluated items. The brevity for completing the RAS3 will promote adoption by nurses and
234
supports recent research suggesting feasibility of screening swallowing in the ED.10 While
235
previous history of stroke was significantly greater in aspirators versus non-aspirators, it was not
236
analyzed as a screening item due to potential unreliability in a patient’s ability to provide this
237
information as well as the associated time delay in retrieving this information from medical
238
records.
M AN U
SC
RI PT
229
239
Notably, the area under ROC curve shows nearly equivalent performance in both logistic
241
regression models without and with age. Because increased age is associated with sensorimotor
242
changes and decreased physiological reserve both of which can increase risk of aspiration,17 the
243
identification of age >70 informs our choice to include it as a valid screening item. Moreover,
244
aging is inherently associated with processes than can increase the risk of pneumonia
245
development.21 However, the decision to include age >70 in the final set of SSIs can be based on
246
the acceptable trade-offs between true positive and false positive rates and the needs of the
247
healthcare environment.
EP
AC C
248
TE D
240
249
Dysarthria was strongly associated with aspiration in our study which supports previous
250
research.9,22 Although the incidence of chronic dysarthria was low, aspiration was evident in
12
ACCEPTED MANUSCRIPT Aspiration screening for suspected stroke 251
66% of the patients who presented with it suggesting that individuals with either chronic or new
252
onset of dysarthria are at increased risk of aspiration.
253
Most valid SSTs include a water swallow component as does the RAS3. Water protocols vary
255
and may include multiple trials of small volumes or a single 90ml volume. We selected a
256
progressive swallowing protocol as a safety measure as many stroke patients are unaware or
257
unconcerned by their dysphagia, do not monitor rate or volume of ingestion, and continue to
258
swallow even with overt signs such as coughing.23,24 These signs of aspiration risk related to the
259
water swallow are supported by the literature9,22 and are typically used in most screening tools.
M AN U
SC
RI PT
254
260
Abnormal volitional cough demonstrated a high likelihood of aspiration risk but had the broadest
262
standard error (Table 2) thus making it less reliable as a predictor. Notably, it did not improve
263
the validity measures (Table 3) in our sample when combined with other items. It may be that
264
our relatively small number of patients with aspiration (N=29, or 12% of our sample) limited the
265
sensitivity of our validity endpoints to the predictive information shown in our logistic model for
266
abnormal volitional cough. The low incidence of aspiration is likely related to the fact that
267
patients in the sample included individuals without confirmed stroke as well as patients with
268
mild stroke. Aspiration incidence is stroke is generally reported in patients with confirmed
269
stroke.9 To our knowledge, the incidence of aspiration has not been published in patients with
270
suspected stroke.
EP
AC C
271
TE D
261
272
It is also important to note that lethargy was present in only two patients, both of whom
273
aspirated. Even though lethargy was excluded due to low occurrence, we strongly suggest that
13
ACCEPTED MANUSCRIPT Aspiration screening for suspected stroke 274
individuals with any reduction in alertness remain non-oral with no screening of swallowing
275
until evidence of sustained wakefulness.
276
While the RAS3 identifies aspiration risk, dysphagia is common following stroke and should be
278
formally evaluated after the initial stroke workup. While not all individuals who aspirate develop
279
pneumonia, early identification of individuals at risk of aspiration ensures implementation of
280
safety measures until SLP evaluation is completed. Individuals without risk of aspiration are
281
cleared of any immediate swallowing safety concern and can start oral intake. Dysphagia is more
282
than aspiration. Even in the absence of aspiration, swallowing efficiency and pleasure in eating
283
can be affected by dysphagia. Given the complexities of swallowing and the need to evaluate
284
medical history, oral mechanism functioning, and swallowing of various consistencies, formal
285
dysphagia evaluation by an SLP is warranted in this high incidence population.
M AN U
SC
RI PT
277
TE D
286
Limitations
288
Limitations include the number of individuals who were excluded from participation. In the first
289
stage of development of a stroke swallowing screening tool, it is important to have stringent
290
inclusion/exclusion criteria to ensure that stroke-related aspiration is being measured. Stroke,
291
however, rarely occurs in isolation. The next step is to determine the generalizability of the RAS3
292
to all patients presenting with suspected stroke, including those who have co-morbid conditions
293
that may produce aspiration. Also important is extending these findings outside VAMCs in order
294
to include a greater number of females in the sample.
AC C
EP
287
295
14
ACCEPTED MANUSCRIPT Aspiration screening for suspected stroke Inability to obtain screening in the ED was an additional limitation. Given the focused stroke
297
workup in the ED, it was not feasible to consent participants as well as administer the SSIs and
298
obtain the VFSS within two hours. As patients with stroke can improve or decline rapidly,
299
obtaining a narrow timeline between screening and VFSS was critical for validity. Future studies
300
will evaluate RN reliability administering the RAS3 in the ED.
301
SC
302
RI PT
296
Conclusions
304
Screening for aspiration risk is a well-established best practice in providing safe, high quality
305
care for patients with suspected stroke. Early screening has the greatest potential to prevent
306
unsafe administration of oral intake to patients who are at risk of aspiration and expedites referral
307
to speech pathology for evaluation to reduce aspiration associated morbidity. The RAS3 provides
308
a valid method to determine aspiration risk in patients presenting with suspected stroke. High
309
sensitivity and NPV ensure that aspiration risk is identified in the majority of patients thus
310
reducing the occurrence of aspiration associated complications.
AC C
EP
TE D
M AN U
303
15
ACCEPTED MANUSCRIPT Aspiration screening for suspected stroke 311
Suppliers
312
a. Varibar barium sulfate Bracco Diagnostics Inc, 259 Prospect Plains Road, Building H, Monroe Township, NJ
314
08831
315 316
b. Stata Statistical Software: Release 13 (2013).
RI PT
313
StataCorp LP. 4905 Lakeway Drive, College Station, TX 77845-4512, USA
AC C
EP
TE D
M AN U
SC
317
16
ACCEPTED MANUSCRIPT Aspiration screening for suspected stroke 318
References
319
1.
protocols prevent pneumonia. Stroke 2005;36:1972-1976. 2.
insufficient evidence for guidelines. Dysphagia 2000;15:19-30.
322
3.
acute phase. Outcome and economic implications. Stroke 1993;24:1823-1827.
324 325
Odderson IR, McKenzie BS. A model for management of patients with stroke during the
4.
SC
323
Martino R, Pron G, Diamant N. Screening for oropharyngeal dysphagia in stroke:
RI PT
320 321
Hinchey JA, Shephard T, Furie K, Smith D, Wang D, Tonn S. Formal dysphagia screening
Jauch EC, Saver JL, Adams HP, Bruno A, Connors JJ, Demaerschalk BM, et al Guidelines for the early management of patients with acute ischemic stroke. A guideline for healthcare
327
professionals from the American Heart Association/American Stroke Association. Stroke
328
2013;44:870-947.
329
5.
M AN U
326
Edmiaston J, Connor LT, Steger-May K, Ford AL. A simple bedside stroke dysphagia screen, validated against videofluoroscopy, detects dysphagia and aspiration with high
331
sensitivity. J Stroke Cerebrovasc Dis 2014;23:712-716.
332
6.
TE D
330
Martino R, Silver F, Teasell R., Bayley M, Nicholson G, Streiner DL, Diamont NE. The Toronto Bedside Swallowing Screening Test (TOR-BSST): development and validation of
334
a dysphagia screening tool for patients with stroke. Stroke 2009;40:555-561. 7.
2008;23:244-250.
336 337
Suiter DB, Leder SB. Clinical utility of the 3 ounce water swallow test. Dysphagia
AC C
335
EP
333
8.
Trapl M, Enderle P, Nowotny M, Teuschl Y, Matz K, Dachenhausen A, et al. Dysphagia
338
bedside screening for acute-stroke patients: the gugging swallowing screen. Stroke
339
2007;38:2948-2952.
17
ACCEPTED MANUSCRIPT Aspiration screening for suspected stroke 9.
patients with acute stroke. Arch Phys Med Rehabil 1998;79:14-19.
341
10.
11.
stroke. Does dysphagia matter? Stroke 1996;27:1200-1204.
347 348
13.
14.
15.
16.
Streiner DL. Diagnosing tests: using and misusing diagnostic and screening tests J Personality Assessment 2003; 81:209-219.
355 356
Polit DF, Beck CT. Nursing Research: Generating and Assessing Evidence for Nursing Practice. 8th Ed. New York: Wolters Kluwer; 2008.
353 354
Daniels SK, Anderson JA, Willson PC. Valid items for screening dysphagia risk in patients with stroke: a systematic review. Stroke 2012;43:892-897.
351 352
Daniels SK, Brailey K, Priestly DH, Herrington LR, Weisberg LA, Foundas AL. Aspiration in patients with acute stroke. Arch Phys Med Rehabil 1998;79:14-19.
349 350
Smithard DG, O'Neill PA, Parks C, Morris J. Complications and outcome after acute
SC
12.
M AN U
346
Mann G, Hankey GJ, Cameron D. Swallowing function after stroke: prognosis and prognostic factors at 6 months. Stroke 1999;30:744-748.
345
RI PT
the emergency department. Nursing Research and Practice 2013;304190.
343 344
Daniels SK, Anderson JA, Petersen NJ. Implementation of stroke dysphagia screening in
TE D
342
Daniels SK, Brailey K, Priestly DH, Harrington LR, Weisberg LA, et al. Aspiration in
17.
EP
340
Lakshminarayan K, Tsai AW, Tong X, Vazquez G, Peacock JM, George MG, et al. Utility of dysphagia screening results in predicting poststroke pneumonia. Stroke 2010;41:2849-
358
2854.
359
18.
AC C
357
Masrur S, Smith EE, Saver JL, Reeves MJ, Bhatt DL, Zhao X, et al. Dysphagia screening
360
and hospital-acquired pneumonia in patients with acute ischemic stroke: findings from Get
361
with the Guidelines-Stroke. J Stroke Cerebrovasc Dis 2013;22:e301-e309.
18
ACCEPTED MANUSCRIPT Aspiration screening for suspected stroke 362
19.
Reeves M, Khoury J, Alwell K, Moonmaw C, Flaherty M, Woo D, et al. Distribution of
363
National Institutes of Health Stroke Scale in the Cincinnati/Northern Kentucky Stroke
364
Study. Stroke 2013;44:3211-3213. 20.
Nicosa MA, Hind JA, Roecker EB, Carnes M, Doyle J, Dengel GA, et al. Age effects on
RI PT
365 366
the temporal evolution of isometric and swallowing pressures. J Gerontol A Biol Sci Med
367
Sci 2000;55:M634-M640. 21.
Chumbler NR, Williams LS, Wells CK, Lo AC, Nadeau S, Peixoto AJ. Derivation and
SC
368
validation of a clinical system for predicting pneumonia in acute stroke. Neuroepidemiol
370
2010;34:193-199.
371
22.
M AN U
369
McCullough GH, Rosenbek JC, Wertz RT, McCoy S, Mann G, McCullough K. Utility of
372
clinical swallowing examination measures for detecting aspiration post-stroke. J Speech
373
Lang Hear Res 2005;48:1280-1293. 23.
Parker C, Power M, Hamdy S, Bowen A, Tyrrell P, Thompson DG. Awareness of
TE D
374 375
dysphagia by patients following stroke predicts swallowing performance. Dysphagia
376
2004;19:28-35. 24.
Schroeder MF, Daniels SK, McClain M, Corey DM, Foundas AL. (2006). Clinical and
EP
377
cognitive predictors of swallowing recovery in stroke. J Rehabil Res Develop 2006;43:301-
379
310.
380
AC C
378
19
ACCEPTED MANUSCRIPT Aspiration screening for suspected stroke Figure Legend
382
Figure 1
Study recruitment and testing flow diagram
383
Figure 2
Comparison of ROC curves using the logistic regression models with and without
384
age in predicting aspiration.
AC C
EP
TE D
M AN U
SC
RI PT
381
20
ACCEPTED MANUSCRIPT
Table 1
Demographic data Total (N=250)
Aspirators (n=29)
Age, yr, mean (SD) Male, n (%)
64 (9.5) 243 (97)
66.8 (8.3) 29 (100)
NonAspirators (n=221) 63.9 (9.6) 214 (97)
Caucasian Race, n (%)
130 (52)
16 (55)
114 (52)
Previous stroke history, n (%) NIHSS score, mean (SD) Mild (0-5), n (%) Moderate (6-13), n (%) Severe (>13), n (%) Discharge Diagnosis, n (%) Imaging confirmed stroke Physician confirmed stroke Physician confirmed TIA Non-stroke/TIA
60 (24)
17 (59)
43 (19.5)
2.6 (2.9)
3.8 (3.9)
2.5 (2.7)
219 (88) 30 (12)
23 (79) 5 (17)
Stroke Type, n (%)
16 (6)
1 (3.5)
15 (7)
6 (21)
43 (19.5)
2 (7)
43 (19.5)
Total (N=140)
Aspirators (n=20)
NonAspirators (n=120)
RHD LHD Bilateral Posterior circulation *p-value<0.05
51 (36) 50 (36) 4 (3) 35 (25)
AC C
137 (98) 3 (2)
20 (100) 0 (0)
5 (25) 7 (35) 0 (0) 8 (40)
0.082 0.417
SC 120 (54)
45 (18)
Ischemic Hemorrhagic Stroke location, n (%)
t-test Fisher’s Exact Pearson Chi-sq Pearson chi-sq t-test
0.716
<0.001* 0.091
0 (0)
20 (69)
TE D
49 (20)
1 (3.5)
p-value
196 (87) 25 (11)
M AN U
140 (56)
EP
Imaging confirmed stroke (only)
1 (<1)
Test type
RI PT
Variable
Fisher’s Exact
0.318
Fisher’s Exact
0.627
Fisher’s Exact
0.379
117 (97.5) 3 (2.5)
46 (38) 43 (36) 4 (3) 27 (22.5)
ACCEPTED MANUSCRIPT
Table 2
Logistic regression results Odds Ratio
SE
z
p-value
Lower Limit Upper Limit
3.396
1.653 2.512 0.012*
1.308
8.815
Dysarthria
5.114
2.410 3.463 0.001**
2.031
12.878
Non-swallow Wet Voice
2.191
2.524 0.681
0.229
20.964
Abnormal Cough
4.094
2.664 2.166 0.030*
1.143
14.656
Water Swallow
3.034
1.454 2.316 0.021*
1.186
7.760
SC
0.496
RI PT
Age >70
N=250; Pseudo R-square = 0.233
AC C
EP
TE D
M AN U
*p-value <0.05; **p-value <0.01
ACCEPTED MANUSCRIPT
Validity indicators for item combinations with significant odds ratios
Items
Negative Screen
VFSS (n=250) Abnormal
25
4
Normal
112
109
Abnormal
27
2
Normal
126
95
Normal
27
2
Abnormal
129
92
Dysarthria OR WS
Age >70 OR Dysarthria OR WS
Specificity (95% CI)
86.2 (68.3-96.1)
49.3 (42.6-56.1)
18.2 (12.2-25.7)
96.5 (91.2-99.0)
93.1 (77.2-99.2)
43.0 (36.4-49.8)
17.6 (12.0-24.6)
97.9 (92.7-99.7)
17.3 (11.7-24.2)
97.9 (92.7-99.7)
93.1 (77.2-99.2)
41.6 (35.1-48.4)
AC C
EP
TE D
M AN U
Abnormal Cough OR Age >70 OR Dysarthria OR WS WS=water swallow
Sensitivity (95% CI)
PPV (95% CI)
RI PT
Positive Screen
SC
Table 3
NPV (95% CI)
ACCEPTED MANUSCRIPT
Table 4
Validity indicators for individual SSIs tested Positive Screen
Negative Screen
VFSS
Dysarthriaa
Non-swallow Wet Voice
Abnormal Cough
Combined Water Swallowa
Normal
38
183
Abnormal
1
28
Normal
1
220
Abnormal
18
11
Normal
43
178
Abnormal
3
26
Normal
3
Abnormal
8
Normal
9
Abnormal
21
Normal
89
132
Abnormal
9
20
Normal
17
204
Abnormal
12
17
Normal
62
159
Abnormal
8
21
Normal
15
206
Abnormal
2
13
Normal
13
158
Abnormal
0
15
Individual Water Swallow Items
AC C
Throat Clear
EP
Cough
Wet Voice
Inability to Drink 90mlc
250
37.9
250
3.5
250
Specificity
218
PPV
NPV
RI PT
18
Sensitivity
62.1
M AN U
Lethargy
11
82.8
22.4
91.0
99.5
50.0
88.7
80.5
29.5
94.2
250
10.3
98.6
50.0
89.3
250
27.6
95.9
47.1
91.0
250
72.4
59.7
19.1
94.3
250
31.0
92.3
34.6
91.1
250
41.4
71.9
16.2
90.3
250
27.6
93.2
34.8
90.7
186
13.3
92.4
13.3
92.4
21
212 8
TE D
Age >70a
Abnormal
n
SC
Items
Wet Voice after 1 minuteb,c Normal 2 166 183 0.0 98.8 0.0 91.7 a=Items in validated RAS3; b=Not part of final combined water swallow; c=n< 250 due to missing information on sequential swallows with a positive preceding WS item
ACCEPTED MANUSCRIPT
Figure 1
RI PT
Potentially eligible participants with suspected stroke n=1042
Eligible participants n=406
M AN U
SC
Excluded (n=636) Non-stroke neurological disease (n=202) Head/neck structural changes (n=111) Dysphagia unrelated to stroke (n=88) TIA/stroke ruled out n= (51) Not competent/no legally authorized representative (n=47) Medically unstable (n=46) Repeat stroke, previously participated (n=43) >3 weeks from symptom onset (n=25) >5 days from admission (n=16) Child bearing potential female (n=4) Deceased (n=3)
TE D
No Swallowing Screening Test (n=148) Declined consent (n=79) Discharged prior to approaching (n=58) Consented but discharged, medical procedure, unstable (n=11)
Swallowing Screening Test participants n=258
RAS3 negative n=100
AC C
EP
RAS3 positive n=158
VFSS completed n=158
Excluded from analyses (n=5) b
Aspiration present (n=27) Aspiration absent (n=126)
No VFSS n=2a
VFSS completed n=98
Excluded from analyses (n=1) b
Aspiration present (n=2) Aspiration absent (n=95)
Note: a=Reasons for no VFSS: C-spine surgery evident on x-ray (n=1), NPO for surgery (n=1); b=Reasons for exclusion from VFSS analyses: Dysphagia history prior to stroke (n=2), Equipment failure (n=1), C-spine damage (n=3)
ACCEPTED MANUSCRIPT
SC
0.50 0.00
0.25
M AN U
0.25 0.00
0.50 1-Specificity
Without age AUC= 0.77(0.048)* Reference
EP
TE D
*Area Under ROC Curve with Standard Error (SE)
AC C
Sensitivity
0.75
RI PT
1.00
Figure 2
0.75
1.00
With age AUC=0.80(0.043)*
ACCEPTED MANUSCRIPT
Appendix 1. Tested Swallowing Screening Items
AC C
EP
TE D
M AN U
SC
RI PT
Non-Swallowing Section 1. Male Yes No 2. Age >70 Yes No Ask patient his/her age. Confirm in medical records 3. Lethargy (Drowsy) Yes No Undecided Unarousable by minor stimulation (requires verbalization or intermittent tactile stimulation-touching/shaking to respond). If evident, score as YES. If awake and alert (regardless of orientation or correctness), score as NO. 4. Dysarthria (Slurred Speech) Yes No Undecided No/Limited Response Ask patient “Tell me, what happened to bring you here.” If speech is slurred or not precise, score as YES. If speech is clear and precise (regardless of content), score as NO. If the person makes no verbal response or response is too limited to judge, score as NO/LIMITED RESPONSE. If still unsure of how to score, score as UNDECIDED. 5. Wet Voice Yes No Undecided No/Limited Response Tell patient to “say AHHHH.” If the voice is wet or gurgly sounding, score as YES. If not, score as NO. If no vocal response, score as NO RESPONSE 6. Abnormal Volitional Cough* Yes No Undecided No Response Other Vocal Response Tell patient to “Cough as strongly as you can, like this…,” demonstrate an example of a strong, forceful cough. Score as YES if the cough is weak or breathy. Score as NO if the person produces a STRONG, FORCEFUL cough. Score as NO RESPONSE if no vocal response. Score as OTHER VOCAL RESPONSE if person says the word “Cough”, produces a throat clear or some other vocal response that is not an actual cough. Swallowing Section 1. 5 ml via cup. Tell patient to “Drink the small amount of water in one swallow.” After swallowing, have patient say “AHHHH”. Listen for presence of WET GURGLY VOICE when saying AHHHH, or COUGH or AUDIBLE THROAT CLEAR. Score as YES if any are present following swallow. A. Cough after Swallow Yes No Undecided B. Audible Throat Clear after Swallow Yes No Undecided C. Wet Voice with AHHHHH Yes No Undecided No Response If any item is YES, STOP screening. DO NOT administer any more water. 2. 5 ml via cup. If none are YES on the first swallow, repeat the 5 ml water swallow using same instructions. A. Cough after Swallow Yes No Undecided B. Audible Throat Clear after Swallow Yes No Undecided C. Wet Voice with AHHHHH Yes No Undecided No Response If any item is YES, STOP screening. DO NOT administer any more water. 3. 90 ml via cup. Tell patient “Drink this entire amount of water without stopping, Keep the cup up to your lips and keep drinking, swallow after swallow, until the water is gone or I tell you to stop.” After drinking have the patient say “AHHHH”. Listen for presence of WET GURGLY VOICE when saying AHHHH, or COUGH or THROAT CLEAR. Score as YES if any are present during or after the swallow. If the patient cannot continuously drink the water, even with encouragement, score as YES. A. Cough after Swallow Yes No Undecided B. Audible Throat Clear after Swallow Yes No Undecided C. Wet Voice with AHHHHH Yes No Undecided No Response D. Unable to Continuously Drink Yes No Undecided E. Wet Voice with AHHHHH after 1 min Yes No Undecided No Response
ACCEPTED MANUSCRIPT
Appendix 2. Rapid Aspiration Screening for Suspected Stroke (RAS3) If the patient has reduced level of consciousness cannot have the head of the bed elevated, or is currently receiving tube feeding, DONOT screen swallowing; patient remains nothing by mouth.
Non-Swallowing Section 1. Age >70*
Yes
No
RI PT
If any item below is YES, the screening is positive. STOP screening. Patient remains NPO and Speech Pathology is consulted
2. Dysarthria (Slurred Speech) Yes No Ask patient “Tell me, what happened to bring you here.” Prompt for more speech if limited verbalization. If speech is slurred or not precise, score as YES. If speech is clear and precise (regardless of content), score as NO. If undecided, score as YES. If either item is YES, STOP screening. DO NOT administer any water.
EP
TE D
M AN U
SC
Swallowing Section Need syringes to measure water, cup, and straw 1. 5ml via cup. Tell patient to “Drink the small amount of water in one swallow.” After swallowing, have patient say “AHHHH”. Listen for presence of WET GURGLY VOICE when saying AHHHH, or COUGH or AUDIBLE THROAT CLEAR. Score as YES if any item is present following swallow. A. Cough after Swallow Yes No B. Audible Throat Clear after Swallow Yes No C. Wet Voice with AHHHHH Yes No If any item is YES, STOP screening. DO NOT administer any more water. 2. 5ml via cup. Repeat the 5 ml water swallow using same instructions. A. Cough after Swallow Yes No B. Audible Throat Clear after Swallow Yes No C. Wet Voice with AHHHHH Yes No If any item is YES, STOP screening. DO NOT administer any more water. 3. 90ml via cup. Tell patient “Drink this entire amount of water without stopping. Keep the cup up to your lips and keep drinking, swallow after swallow, until the water is gone or I tell you to stop.” After drinking have the patient say “AHHHH”. Encourage patient to keep swallowing if is pausing between swallows is evident. Listen for presence of WET GURGLY VOICE when saying AHHHH, or COUGH or THROAT CLEAR. Stop screening and remove cup if coughing, throat clearing, wet voice, and/or inability to continuous drink during swallowing. Score as YES if any item is present during or after the swallow. If the patient cannot continuously drink the water, even with encouragement, score as YES. A. Cough after Swallow Yes No B. Audible Throat Clear after Swallow Yes No C. Wet Voice with AHHHHH Yes No D. Unable to Continuously Drink Yes No
AC C
*Denotes item than can be omitted from RAS3