- Email: [email protected]
Do Ratings of Swallowing Function Differ by Videofluoroscopic Rate? An Exploratory Analysis in Patients After Acute Stroke
Archives of Physical Medicine and Rehabilitation journal homepage: www.archives-pmr.org Archives of Physical Medicine and Rehabilitation 2018;-:-------
ORIGINAL RESEARCH
Do Ratings of Swallowing Function Differ by Videofluoroscopic Rate? An Exploratory Analysis in Patients After Acute Stroke Rachel W. Mulheren, PhD,a,b Alba Azola, MD,a Marlı´s Gonza´lez-Ferna´ndez, MD, PhDa From the aDepartment of Physical Medicine and Rehabilitation, Johns Hopkins University School of Medicine, Baltimore, MD; and bDepartment of Neuroscience, Johns Hopkins University School of Medicine, Baltimore, MD. Current affiliation for Mulheren, Case Western Reserve University, Cleveland, OH.
Abstract Objective: To determine differences between continuous videofluoroscopic swallow studies (VFSS) with a pulse rate and frame rate of 30 and the same swallows reduced to 15 frames per second (fps) on measures of swallowing function in patients after acute ischemic stroke. Design: Blinded comparison. Setting: Acute hospital. Participants: Patients after ischemic stroke (NZ20). Interventions: Not applicable. Main Outcome Measures: Single and sequential sips of thin liquids, single sips of nectar liquids, pudding, and cookie boluses were rated on measures of timing of swallowing events, Modified Barium Swallowing Impairment Profile component scores, and Penetration-Aspiration Scale scores. The ratings for videos at 15 fps and 30 fps were compared by Wilcoxon signed rank tests. Results: Pharyngeal transit time was longer and bolus entry into the hypopharynx was later for 30 fps than for 15 fps. Components of Oral Residue and Pharyngoesophageal Segment Opening ratings were more severe for 15 fps than 30 fps, whereas Bolus Transport and Initiation of Pharyngeal Swallow were rated as more severe for 30 fps than for 15 fps. There was no difference between 30 fps and 15 fps on the remaining measures, including Penetration-Aspiration Scale scores. Conclusion: Continuous VFSS recorded at 30 fps and their down-sampled 15 fps duplicates yielded contrasting results on certain durational and functional measures of swallowing, though not on others. VFSS should be administered continuously or at 30 pulses per second for valid assessment of swallowing while using other methods to reduce radiation exposure. Archives of Physical Medicine and Rehabilitation 2018;-:------ª 2018 by the American Congress of Rehabilitation Medicine
Swallowing function is often assessed using videofluoroscopic swallowing studies (VFSS) when dysphagia is suspected. During this procedure, participants consume multiple consistencies of liquid barium and solid foods coated with barium contrast as real-time pulses of radiation are delivered to the oral cavity, oropharynx, pharynx, and esophagus. In combination with medical history, the resulting images of swallowing anatomy and Presented to the Dysphagia Research Society, Portland, OR, March 2017. This work was supported by the National Institutes of Health through the National Institute on Deafness and Other Communication Disorders (grant no. K23 DC 011056) and the National Institute of Child Health & Human DevelopmentdNational Center for Medical Rehabilitation Research (grant no. 5T32HD007414-23). The content is solely the responsibility of the authors and does not necessarily represent the views of the National Institutes of Health.
physiology provide the basis for diagnostic and therapeutic decision making. Radiation exposure is of concern and can be reduced by limiting the duration (beam on time) of studies, setting appropriate collimation, and using pulsed fluoroscopy.1 Although these measures have been shown to reduce radiation exposure, they must be weighed against the need for obtaining appropriate clinical information. Standardized protocols for VFSS, such as the Modified Barium Swallow Impairment Profile (MBSImP) protocol,2 significantly reduce radiation exposure by reducing the overall duration of the study. Collimation to the structures of interest at the beginning of VFSS reduces radiation scatter. Fluoroscopy can be conducted in continuous or pulsed mode. During continuous mode, the X-ray beam is on throughout
0003-9993/18/$36 - see front matter ª 2018 by the American Congress of Rehabilitation Medicine https://doi.org/10.1016/j.apmr.2018.10.015
2
R.W. Mulheren et al
recording. Pulsed mode involves the delivery of short pulses of the X-ray beam and the pulse rate can be controlled by the operator. Pulsed fluoroscopy at 30 pulses per second (pps) is often referred to as continuous fluoroscopy on equipment settings. A lower pulse rate delivers less radiation but has a lower temporal resolution. For the purposes of display, fluoroscopy can be recorded and played at different frame rates (frames per second [fps]), with a convention of 30 fps. Switching from continuous fluoroscopy to pulsed fluoroscopy at 15 pps can reduce radiation dosage by 22%3; however, when performing VFSS, it is unclear whether reductions in temporal resolution may compromise imaging of discrete swallowing events. Several swallowing events, such as duration of laryngeal vestibule closure and duration of upper esophageal sphincter opening, occur in less than 1 second4 and may not be properly visualized at 15 pps or less; additionally, the presence, amount, and extent of penetration or aspiration may be difficult to assess. Although the American Speech-Language-Hearing Association does not specify an acquisition rate for VFSS, it endorses the principle of as low as reasonably achievable for radiation exposure to the extent that the dosage will not compromise the evaluation.5 To date, the literature on the topic of fluoroscopy modality and rate during swallowing evaluation is limited. The original papers validating the use of videofluoroscopy for the evaluation of swallowing recommended “imaging at a frame rate of 30/s”6(p956); lower frame rates were reported to inhibit accurate imaging of structures and swallowing kinematics.7 However, there is limited evidence to support continuous fluoroscopy acquired at 30 images per second over pulsed fluoroscopy at lower rates during the assessment of oropharyngeal swallowing. One study of infants and children known to penetrate to the level of the laryngeal vestibule reported penetration on continuous fluoroscopy (30pps) lasting only 1 frame (33.33ms) in the majority of patients, which would likely have been missed with a lower pulse rate.8 A pilot study of adult patients with various etiologies and degrees of dysphagia reported differences in the rating of physiological parameters, severity of penetration and aspiration, and therapeutic recommendations by speech-language pathologists when comparing studies conducted at 30 pps and simulations of the same studies at 15 pps.9 These 2 studies concluded that 30 pps should be used, though additional research is needed to support this evidence in other patient populations. In this study, we examined how continuous fluoroscopy and a simulation of pulsed fluoroscopy compare in terms of swallowing assessment. The aim of this study was to determine differences between VFSS acquired continuously at 30 fps and their duplicates at half the frame rate on measures of swallowing function in acute stroke patients with varying degrees of swallowing impairment. We hypothesized that a lower frame rate would result in differences in measures of timing, airway protection, and ratings of swallowing physiology and bolus transit.
List of abbreviations: CI fps ICC MBSImP
confidence interval frames per second intra-class correlation coefficient Modified Barium Swallowing Impairment Profile pps pulses per second VFSS videofluoroscopic swallowing study
Methods The protocol was approved by the institutional review board. A convenience sample of ischemic stroke patients at a neurologic care unit provided written consent to participate in the study. In all cases, VFSS were conducted as is usual and customary for clinical care by speech-language pathologists and radiologists within 72 hours of diagnostic neuroimaging. VFSS were acquired with continuous fluoroscopy (30pps) and recorded at 30 fps with a Philips MultiDiagnost Eleva with Flat Detector v3.2.a Exclusion criteria included history of swallowing dysfunction or neurologic disease that might lead to swallowing dysfunction. According to the clinical speech-language pathology reports, the clinical impression of swallowing function in this cohort ranged from within normal limits to severe dysphagia. Standardized barium contrast consistencies (Varibarb) and barium-coated solid food were used to perform the study as deemed safe by the examining clinician. The order of stimuli followed the MBSImP protocol.2 Oral and Pharyngeal Impairment scores were calculated based on ratings of included components (table 1), with a score of 0 indicating no impairment. The first swallow of the following boluses was individually rated and included in the analysis: (1) single cup sip of thin liquid; (2) sequential sips of thin liquid; (3) single cup sip of nectar liquid; (4) 5 mL pudding; (5) 1/2 cookie coated with pudding. These boluses were included as they represented a variety of consistencies. The full video sequences were clipped by bolus type for each participant. Each video clip was duplicated, and, to simulate 15 pps, every other frame was removed with VirtualDub.c Deleted frames were not replaced with duplicates in order to maintain blinding of the reduced frame rate during analysis. All video clips were de-identified, assigned random numeric identifiers, and exported to AVI format. Analysis was conducted in QuickTime Player 7d so that total duration was consistent between original and decimated videos. Two researchers who were MBSImP certified raters and had prior training and experience in timing analysis and PenetrationAspiration Scale ratings performed frame-by-frame analysis of each video clip. The 30-fps and 15-fps video clips were randomly assigned to each rater. Raters were not provided with time constraints for individual video analysis. To avoid potential bias caused by recollection of a previously rated swallow, each rater performed blinded analysis of either the 30-fps or 15-fps clips for a given participant. Both researchers determined timing of swallowing events (table 2),10-12 Penetration-Aspiration Scale score,13 and MBSImP component scores (table 1).2 Anterior-posterior views were not available for all participants, thus the MBSImP components requiring this view (Pharyngeal Contraction and Esophageal Clearance) were excluded. Additionally, Lip Closure was not studied, as the anterior oral cavity and lips were not visualized in all participants. After the initial analysis, each rater was assigned the remaining videos to establish interrater reliability. A subset of 10% of the videos analyzed in the first round were randomly selected and re-analyzed by each rater to determine intra-rater reliability.
Statistical analysis Timing was measured by counting the frames until the event of interest occurred. For the 15-fps sequences the frame count was multiplied by 2 for comparison. Measures of timing of swallowing events, Penetration-Aspiration Scale, and MBSImP for each frame rate were compared by Wilcoxon signed rank tests in separate www.archives-pmr.org
Ratings of swallowing poststroke Table 1
3
MBSImP component scores2 included in the analysis
Oral Impairment
Pharyngeal Impairment
Component Title
Scoring Range
Tongue Control During Bolus Hold
0-3
Bolus Preparation/Mastication Bolus Transport/Lingual Motion Oral Residue Initiation of Pharyngeal Swallow Total Oral Impairment Soft Palate Elevation
0-3 0-4 0-4 0-4 0-18 0-4
Laryngeal Elevation Anterior Hyoid Excursion Epiglottic Movement Laryngeal Vestibular ClosureHeight of Swallow Pharyngeal Stripping Wave Pharyngoesophageal Segment Opening Tongue Base Retraction Pharyngeal Residue Total Pharyngeal Impairment
0-3 0-2 0-2 0-2
Durational measures of swallowing events10-12
Measure
Definition
Bolus enters valleculae to LVC
Time from bolus crossing ramus of the mandible to first frame of laryngeal vestibule closure Time from bolus crossing posterior nasal spine to hyoid burst
Bolus enters upper oropharyngeal area to hyoid burst Swallow Reaction Time Bolus enters hypopharynx to hyoid burst Bolus enters upper esophageal sphincter to hyoid burst Bolus enters valleculae to UES opening
Time from bolus crossing ramus of the mandible to hyoid burst Time from bolus crossing base of valleculae to hyoid burst Time from bolus head crossing superior border of UES to hyoid burst Time from bolus head crossing ramus of mandible to first frame of UES opening Time from bolus head crossing ramus of mandible to bolus tail crossing superior border of UES
0-2 0-3
Pharyngeal Transit Time
0-4 0-4 0-26
Abbreviations: LVC, laryngeal vestibule closure; UES, upper esophageal sphincter.
analyses by bolus type. Interrater reliability was calculated between raters using 2-way random intra-class correlation coefficients (ICC), and intra-rater reliability was calculated with ICC of 10% of the videos for each rater, with a minimum acceptance rate of 80%.
Results Participants Twenty patients were included in the analysis. Eleven patients were men, with a mean age of 56.6 years. Participant demographics, location of infarct, and cumulative modified Oral and Pharyngeal Impairment MBSImP scores are listed in table 3. Despite instances of penetration and aspiration, the median Penetration-Aspiration Scale scores by bolus type were low (1 or 2).
Timing of swallowing events and Penetration-Aspiration Scale In all patients, the bolus entered the hypopharynx later for 30 fps than for 15 fps in solid swallows; this measure was later for 15 fps than for 30 fps when drinking nectar-thick liquids (table 4). For swallowing of sequential thin boluses, the bolus entered the upper esophageal sphincter later for 30 fps than for 15 fps. For pudding boluses, pharyngeal transit time was longer for 30 fps than for 15 fps. No other measures of timing were statistically significantly different between the 2 frame rates. There were no significant differences on Penetration-Aspiration Scale scores by frame rate. Average interrater reliability ICC values were 0.974 (95% confidence interval [CI] 0.968-0.98) for both timing and Penetration-Aspiration Scale ratings, and average intrarater reliability values for the 2 raters were 0.954 (95% CI 0.926-0.972) and 0.994 (95% CI 0.99-0.997). www.archives-pmr.org
Table 2
MBSImP Of the 14 MBSImP component scores, 15-fps swallows received more severe ratings of Oral Residue than for 30 fps during boluses of thin liquids by cup and pudding (table 5); similarly, Pharyngoesophageal Segment Opening was scored more severely for 15 fps than for 30 fps in sequential swallows of thin liquids. Bolus Transport/Lingual Motion and Initiation of Pharyngeal Swallow scores were more severe for 30 fps than for 15 fps during solid swallows (see table 5). Average interrater reliability ICC values were 0.86 (95% CI 0.834-1), and average intra-rater reliability ICC values for the 2 raters were 0.90 (95% CI 0.868-0.929) and 0.99 (95% CI 0.987-0.993).
Discussion The results of this work suggest that use of a lower frame rate (15fps), a proxy for 15-pps fluoroscopy, leads to different durations and ratings of swallowing events in comparison to continuous fluoroscopy at 30 pps. Thus, the diagnosis of swallowing impairments and subsequent treatment decisions may be affected.
Timing of swallowing events and Penetration-Aspiration Scale In all bolus consistencies analyzed, there was a trend for longer swallowing events in 30 fps in comparison to the same swallows at 15 fps (see table 4). Differences between the 2 frame rates may be due to events occurring during frames that were present at 30 fps but were omitted at 15 fps. Additionally, the context of each frame may assist in visual perception. Judgment of timing and clinical measures requires consideration of each frame as part of the dynamic process of swallowing rather than as an isolated slice. At
4
R.W. Mulheren et al Table 3
Participant demographics
Participant Number
Age, y
Sex
Infarct Location
Clinical Impression
MBSImP Oral Impairment Score (Modified)
MBSImP Pharyngeal Impairment Score (Modified)
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
55 57 57 43 67 79 48 51 57 70 25 38 71 60 74 55 48 55 73 48
Male Female Female Male Male Male Male Male Female Male Female Male Female Male Male Female Female Female Female Male
MCA MCA MCA MCA MCA Pons MCA Pons MCA ACA ACA, MCA MCA, PCA MCA BG MCA Pons MCA MCA MCA MCA
WNL Dysphagia Dysphagia Dysphagia WNL Dysphagia WNL Dysphagia Dysphagia Dysphagia Dysphagia WNL Dysphagia Dysphagia Dysphagia Dysphagia Dysphagia Dysphagia Dysphagia Dysphagia
11 8 10 13 8 13 7 7 12 11 7 7 10 10 11 11 13 11 14 9
4 7 5 2 1 12 1 7 10 4 4 5 8 7 10 7 8 10 14 9
Abbreviations: ACA, anterior cerebral artery; BG, basal ganglia; MCA, middle cerebral artery; WNL, within normal limits.
15 fps, the 2 frames surrounding a single frame at 30 fps are omitted, thus altering the context in which the single frame is judged. The timing measures that were significantly different for 15 and 30 fps were not uniform across bolus types. The difference in bolus properties between thin, nectar, pudding, and solid may have affected the results.14 The timing measures were not affected equally for thin liquid swallowing for a single cup sip vs a sequential swallowing task, which may be due to different bolus volumes as well as the different physiological responses observed during these tasks.15,16 This study included acute stroke patients with and without a clinical impression of dysphagia upon VFSS. Pharyngeal transit time, a measure that may predict penetration-aspiration in stroke patients, was more delayed in 30-fps than 15-fps videos.17,18 Although the current exploratory study did not have large enough numbers to stratify by dysphagia diagnosis or severity, the discrepancy between 15 and 30 fps may be due to more difficult visual perception of swallowing events and bolus transit in the presence of physiologic abnormality.
Table 4
Although Cohen et al8 reported penetration lasting only 1 frame in the majority of a pediatric sample at 30 fps that would likely have been missed at a lower pulse rate, we found no differences between 15 fps and 30 fps on Penetration-Aspiration Scale scores in adult stroke patients. This discrepancy may be due to differences between pediatric and adult swallowing, etiology of dysphagia or severity of penetration/aspiration in the current patient sample.
MBSImP The MBSImP data suggest that the videofluoroscopic frame rate affects the assessment of physiologic impairment. In other words, a swallow rated at 30 fps is perceived to have different degrees of impairment than the same swallow at 15 fps. Considering the significant contribution of swallowing assessment to the development of treatment, frame rate may directly affect clinical care. Furthermore, the validation of the MBSImP was completed at 30 fps,2 thus a lower frame rate may produce invalid results on this measure.
Wilcoxon sign rank test results comparing timing of swallowing measures for paired swallows Duration (Median, in ms)
Measure
Bolus Type
15 fps
30 fps
z Statistic*
P Value
Bolus enters upper esophageal sphincter from hyoid burst Bolus enters hypopharynx from hyoid burst
Sequential thin Nectar Solid Pudding
66.67 199.8 <.01 600.03
133.32 49.95 49.99 733.26
-2.42 2.38 -2.71 -2.29
.015 .017 .007 .022
Pharyngeal Transit Time
* Positive z statistic indicates longer durations for 15 than 30 fps; negative z statistic indicates shorter durations for 15 than 30 fps.
www.archives-pmr.org
Ratings of swallowing poststroke Table 5
5
Wilcoxon signed rank test results comparing MSBImP component scores for paired swallows Median Score
Measure
Bolus Type
15 fps
30 fps
z Statistic*
P Value
Oral Residue
Cup sip thin Pudding Solid Sequential thin Solid
2 2 0 1 1
1 1 2 0 1
2.65 2.33 -2.78 2.24 -2.45
.008 .020 .005 .025 .014
Bolus Transport/Lingual Motion Pharyngoesophageal Segment Opening Initiation of Pharyngeal Swallow
* Positive z statistic indicates higher, more severe ratings for 15 than 30 fps; negative z statistic indicates lower, less severe rating for 15 than 30 fps.
As bolus transport, initiation of pharyngeal swallow, and pharyngoesophageal segment opening received higher scores at 30 fps than at 15 fps, the degree of impairment was underestimated at 15 fps. Interestingly, oral residue was lower, or less impaired, at 30 fps than at 15 fps. However, an oral residue score of 1 is within normal limits,19 thus differences between 0 and 1 on this component are not practically significant. As with the timing of swallowing events, the difference in scores between the 15 fps and 30 fps may be explained by altered visual perception given contextual changes when 50% of frames are dropped.
Keywords
Study limitations
References
This study included a relatively small sample of acute stroke patients and may not be generalizable to people with other diseases resulting in dysphagia. In the present exploratory study, several of the comparisons were not significant; a larger sample with stratification by dysphagia severity and a higher median Penetration-Aspiration Scale score may yield different results. Bolus consistency and size were limited; thus, different results may be obtained from analysis of other bolus types. Examination of additional measures of swallowing function, such as kinematics, would supplement the reported results. Further research is warranted to determine the effect of lower pulse rates on treatment recommendations.
Conclusion In this sample, frame rate had a quantifiable effect on certain measures of swallowing function in patients poststroke. The use of pulsed fluoroscopy at 15 pps may distort swallowing measures, and further research is warranted to determine the effect on clinical judgment. The results of this study support the most commonly accepted standard of continuous fluoroscopy or 30 pps for optimal evaluation of swallowing function, though further, large-scale research is necessary.
Suppliers a. b. c. d.
MultiDiagnost Eleva with Flat Detector v3.2; Philips. Varibar; Bracco. VirtualDub; Avery Lee. QuickTime Player 7; Apple, Inc.
www.archives-pmr.org
Fluoroscopy; Pulse rate; Rehabilitation; Stroke; Swallowing
Corresponding author Rachel W. Mulheren, PhD, Case Western Reserve University, 11635 Euclid Ave., Cleveland, OH 44107. E-mail address: [email protected].
1. Mahesh M. Fluoroscopy: patient radiation exposure issues. Radiographics 2001;21:1033-45. 2. Martin-Harris B, Brodsky MB, Michel Y, et al. MBS measurement tool for swallow impairmentdMBSImp: establishing a standard. Dysphagia 2008;23:392-405. 3. Aufrichtig R, Xue P, Thomas CW, Gilmore GC, Wilson DL. Perceptual comparison of pulsed and continuous fluoroscopy. Med Phys 1994;21:245-56. 4. Molfenter SM, Steele CM. Kinematic and temporal factors associated with penetrationeaspiration in swallowing liquids. Dysphagia 2014; 29:269-76. 5. American Speech-Language-Hearing Association. Guidelines for speech-language pathologists performing videofluoroscopic swallowing studies. Available at: https://www.asha.org/policy/ GL2004-00050/. Accessed March 5, 2017. 6. Dodds WJ, Stewart ET, Logemann JA. Physiology and radiology of the normal oral and pharyngeal phases of swallowing. AJR Am J Roentgenol 1990;154:953-63. 7. Jones B, Donner MW. How I do it: examination of the patient with dysphagia. Dysphagia 1989;4:162-72. 8. Cohen MD. Can we use pulsed fluoroscopy to decrease the radiation dose during video fluoroscopic feeding studies in children? Clin Radiol 2009;64:70-3. 9. Bonilha HS, Blair J, Carnes B, et al. Preliminary investigation of the effect of pulse rate on judgments of swallowing impairment and treatment recommendations. Dysphagia 2013;28:528-38. 10. Curtis DJ, Cruess DF, Dachman AH, Maso E. Timing in the normal pharyngeal swallow. Prospective selection and evaluation of 16 normal asymptomatic patients. Invest Radiol 1984;19:523-9. 11. McConnel FM, Cerenko D, Jackson RT, Guffin TN Jr. Timing of major events of pharyngeal swallowing. Arch Otolaryngol Head Neck Surg 1988;114:1413-8. 12. Inamoto Y, Fujii N, Saitoh E, et al. Evaluation of swallowing using 320-detector-row multislice CT. Part II: kinematic analysis
6 of laryngeal closure during normal swallowing. Dysphagia 2011; 26:209-17. 13. Rosenbek JC, Robbins JA, Roecker EB, Coyle JL, Wood JL. A penetration-aspiration scale. Dysphagia 1996;11:93-8. 14. Steele CM, Alsanei WA, Ayanikalath S, et al. The influence of food texture and liquid consistency modification on swallowing physiology and function: a systematic review. Dysphagia 2015; 30:2-26. 15. Daniels SK, Foundas AL. Swallowing physiology of sequential straw drinking. Dysphagia 2001;16:176-82.
R.W. Mulheren et al 16. Chi-Fishman G, Sonies BC. Motor strategy in rapid sequential swallowing: new insights. J Speech Lang Hear Res 2000;43:1481-92. 17. Bingjie L, Tong Z, Xinting S, Jianmin X, Guijun J. Quantitative videofluoroscopic analysis of penetration-aspiration in post-stroke patients. Neurol India 2010;58:42-7. 18. Kim Y, McCullough GH. Stage transition duration in patients poststroke. Dysphagia 2007;22:299-305. 19. Martin-Harris B, McFarland D, Hill EG, et al. Respiratory-swallow training in patients with head and neck cancer. Arch Phys Med Rehabil 2015;96:885-93.
www.archives-pmr.org
ORIGINAL RESEARCH
Do Ratings of Swallowing Function Differ by Videofluoroscopic Rate? An Exploratory Analysis in Patients After Acute Stroke Rachel W. Mulheren, PhD,a,b Alba Azola, MD,a Marlı´s Gonza´lez-Ferna´ndez, MD, PhDa From the aDepartment of Physical Medicine and Rehabilitation, Johns Hopkins University School of Medicine, Baltimore, MD; and bDepartment of Neuroscience, Johns Hopkins University School of Medicine, Baltimore, MD. Current affiliation for Mulheren, Case Western Reserve University, Cleveland, OH.
Abstract Objective: To determine differences between continuous videofluoroscopic swallow studies (VFSS) with a pulse rate and frame rate of 30 and the same swallows reduced to 15 frames per second (fps) on measures of swallowing function in patients after acute ischemic stroke. Design: Blinded comparison. Setting: Acute hospital. Participants: Patients after ischemic stroke (NZ20). Interventions: Not applicable. Main Outcome Measures: Single and sequential sips of thin liquids, single sips of nectar liquids, pudding, and cookie boluses were rated on measures of timing of swallowing events, Modified Barium Swallowing Impairment Profile component scores, and Penetration-Aspiration Scale scores. The ratings for videos at 15 fps and 30 fps were compared by Wilcoxon signed rank tests. Results: Pharyngeal transit time was longer and bolus entry into the hypopharynx was later for 30 fps than for 15 fps. Components of Oral Residue and Pharyngoesophageal Segment Opening ratings were more severe for 15 fps than 30 fps, whereas Bolus Transport and Initiation of Pharyngeal Swallow were rated as more severe for 30 fps than for 15 fps. There was no difference between 30 fps and 15 fps on the remaining measures, including Penetration-Aspiration Scale scores. Conclusion: Continuous VFSS recorded at 30 fps and their down-sampled 15 fps duplicates yielded contrasting results on certain durational and functional measures of swallowing, though not on others. VFSS should be administered continuously or at 30 pulses per second for valid assessment of swallowing while using other methods to reduce radiation exposure. Archives of Physical Medicine and Rehabilitation 2018;-:------ª 2018 by the American Congress of Rehabilitation Medicine
Swallowing function is often assessed using videofluoroscopic swallowing studies (VFSS) when dysphagia is suspected. During this procedure, participants consume multiple consistencies of liquid barium and solid foods coated with barium contrast as real-time pulses of radiation are delivered to the oral cavity, oropharynx, pharynx, and esophagus. In combination with medical history, the resulting images of swallowing anatomy and Presented to the Dysphagia Research Society, Portland, OR, March 2017. This work was supported by the National Institutes of Health through the National Institute on Deafness and Other Communication Disorders (grant no. K23 DC 011056) and the National Institute of Child Health & Human DevelopmentdNational Center for Medical Rehabilitation Research (grant no. 5T32HD007414-23). The content is solely the responsibility of the authors and does not necessarily represent the views of the National Institutes of Health.
physiology provide the basis for diagnostic and therapeutic decision making. Radiation exposure is of concern and can be reduced by limiting the duration (beam on time) of studies, setting appropriate collimation, and using pulsed fluoroscopy.1 Although these measures have been shown to reduce radiation exposure, they must be weighed against the need for obtaining appropriate clinical information. Standardized protocols for VFSS, such as the Modified Barium Swallow Impairment Profile (MBSImP) protocol,2 significantly reduce radiation exposure by reducing the overall duration of the study. Collimation to the structures of interest at the beginning of VFSS reduces radiation scatter. Fluoroscopy can be conducted in continuous or pulsed mode. During continuous mode, the X-ray beam is on throughout
0003-9993/18/$36 - see front matter ª 2018 by the American Congress of Rehabilitation Medicine https://doi.org/10.1016/j.apmr.2018.10.015
2
R.W. Mulheren et al
recording. Pulsed mode involves the delivery of short pulses of the X-ray beam and the pulse rate can be controlled by the operator. Pulsed fluoroscopy at 30 pulses per second (pps) is often referred to as continuous fluoroscopy on equipment settings. A lower pulse rate delivers less radiation but has a lower temporal resolution. For the purposes of display, fluoroscopy can be recorded and played at different frame rates (frames per second [fps]), with a convention of 30 fps. Switching from continuous fluoroscopy to pulsed fluoroscopy at 15 pps can reduce radiation dosage by 22%3; however, when performing VFSS, it is unclear whether reductions in temporal resolution may compromise imaging of discrete swallowing events. Several swallowing events, such as duration of laryngeal vestibule closure and duration of upper esophageal sphincter opening, occur in less than 1 second4 and may not be properly visualized at 15 pps or less; additionally, the presence, amount, and extent of penetration or aspiration may be difficult to assess. Although the American Speech-Language-Hearing Association does not specify an acquisition rate for VFSS, it endorses the principle of as low as reasonably achievable for radiation exposure to the extent that the dosage will not compromise the evaluation.5 To date, the literature on the topic of fluoroscopy modality and rate during swallowing evaluation is limited. The original papers validating the use of videofluoroscopy for the evaluation of swallowing recommended “imaging at a frame rate of 30/s”6(p956); lower frame rates were reported to inhibit accurate imaging of structures and swallowing kinematics.7 However, there is limited evidence to support continuous fluoroscopy acquired at 30 images per second over pulsed fluoroscopy at lower rates during the assessment of oropharyngeal swallowing. One study of infants and children known to penetrate to the level of the laryngeal vestibule reported penetration on continuous fluoroscopy (30pps) lasting only 1 frame (33.33ms) in the majority of patients, which would likely have been missed with a lower pulse rate.8 A pilot study of adult patients with various etiologies and degrees of dysphagia reported differences in the rating of physiological parameters, severity of penetration and aspiration, and therapeutic recommendations by speech-language pathologists when comparing studies conducted at 30 pps and simulations of the same studies at 15 pps.9 These 2 studies concluded that 30 pps should be used, though additional research is needed to support this evidence in other patient populations. In this study, we examined how continuous fluoroscopy and a simulation of pulsed fluoroscopy compare in terms of swallowing assessment. The aim of this study was to determine differences between VFSS acquired continuously at 30 fps and their duplicates at half the frame rate on measures of swallowing function in acute stroke patients with varying degrees of swallowing impairment. We hypothesized that a lower frame rate would result in differences in measures of timing, airway protection, and ratings of swallowing physiology and bolus transit.
List of abbreviations: CI fps ICC MBSImP
confidence interval frames per second intra-class correlation coefficient Modified Barium Swallowing Impairment Profile pps pulses per second VFSS videofluoroscopic swallowing study
Methods The protocol was approved by the institutional review board. A convenience sample of ischemic stroke patients at a neurologic care unit provided written consent to participate in the study. In all cases, VFSS were conducted as is usual and customary for clinical care by speech-language pathologists and radiologists within 72 hours of diagnostic neuroimaging. VFSS were acquired with continuous fluoroscopy (30pps) and recorded at 30 fps with a Philips MultiDiagnost Eleva with Flat Detector v3.2.a Exclusion criteria included history of swallowing dysfunction or neurologic disease that might lead to swallowing dysfunction. According to the clinical speech-language pathology reports, the clinical impression of swallowing function in this cohort ranged from within normal limits to severe dysphagia. Standardized barium contrast consistencies (Varibarb) and barium-coated solid food were used to perform the study as deemed safe by the examining clinician. The order of stimuli followed the MBSImP protocol.2 Oral and Pharyngeal Impairment scores were calculated based on ratings of included components (table 1), with a score of 0 indicating no impairment. The first swallow of the following boluses was individually rated and included in the analysis: (1) single cup sip of thin liquid; (2) sequential sips of thin liquid; (3) single cup sip of nectar liquid; (4) 5 mL pudding; (5) 1/2 cookie coated with pudding. These boluses were included as they represented a variety of consistencies. The full video sequences were clipped by bolus type for each participant. Each video clip was duplicated, and, to simulate 15 pps, every other frame was removed with VirtualDub.c Deleted frames were not replaced with duplicates in order to maintain blinding of the reduced frame rate during analysis. All video clips were de-identified, assigned random numeric identifiers, and exported to AVI format. Analysis was conducted in QuickTime Player 7d so that total duration was consistent between original and decimated videos. Two researchers who were MBSImP certified raters and had prior training and experience in timing analysis and PenetrationAspiration Scale ratings performed frame-by-frame analysis of each video clip. The 30-fps and 15-fps video clips were randomly assigned to each rater. Raters were not provided with time constraints for individual video analysis. To avoid potential bias caused by recollection of a previously rated swallow, each rater performed blinded analysis of either the 30-fps or 15-fps clips for a given participant. Both researchers determined timing of swallowing events (table 2),10-12 Penetration-Aspiration Scale score,13 and MBSImP component scores (table 1).2 Anterior-posterior views were not available for all participants, thus the MBSImP components requiring this view (Pharyngeal Contraction and Esophageal Clearance) were excluded. Additionally, Lip Closure was not studied, as the anterior oral cavity and lips were not visualized in all participants. After the initial analysis, each rater was assigned the remaining videos to establish interrater reliability. A subset of 10% of the videos analyzed in the first round were randomly selected and re-analyzed by each rater to determine intra-rater reliability.
Statistical analysis Timing was measured by counting the frames until the event of interest occurred. For the 15-fps sequences the frame count was multiplied by 2 for comparison. Measures of timing of swallowing events, Penetration-Aspiration Scale, and MBSImP for each frame rate were compared by Wilcoxon signed rank tests in separate www.archives-pmr.org
Ratings of swallowing poststroke Table 1
3
MBSImP component scores2 included in the analysis
Oral Impairment
Pharyngeal Impairment
Component Title
Scoring Range
Tongue Control During Bolus Hold
0-3
Bolus Preparation/Mastication Bolus Transport/Lingual Motion Oral Residue Initiation of Pharyngeal Swallow Total Oral Impairment Soft Palate Elevation
0-3 0-4 0-4 0-4 0-18 0-4
Laryngeal Elevation Anterior Hyoid Excursion Epiglottic Movement Laryngeal Vestibular ClosureHeight of Swallow Pharyngeal Stripping Wave Pharyngoesophageal Segment Opening Tongue Base Retraction Pharyngeal Residue Total Pharyngeal Impairment
0-3 0-2 0-2 0-2
Durational measures of swallowing events10-12
Measure
Definition
Bolus enters valleculae to LVC
Time from bolus crossing ramus of the mandible to first frame of laryngeal vestibule closure Time from bolus crossing posterior nasal spine to hyoid burst
Bolus enters upper oropharyngeal area to hyoid burst Swallow Reaction Time Bolus enters hypopharynx to hyoid burst Bolus enters upper esophageal sphincter to hyoid burst Bolus enters valleculae to UES opening
Time from bolus crossing ramus of the mandible to hyoid burst Time from bolus crossing base of valleculae to hyoid burst Time from bolus head crossing superior border of UES to hyoid burst Time from bolus head crossing ramus of mandible to first frame of UES opening Time from bolus head crossing ramus of mandible to bolus tail crossing superior border of UES
0-2 0-3
Pharyngeal Transit Time
0-4 0-4 0-26
Abbreviations: LVC, laryngeal vestibule closure; UES, upper esophageal sphincter.
analyses by bolus type. Interrater reliability was calculated between raters using 2-way random intra-class correlation coefficients (ICC), and intra-rater reliability was calculated with ICC of 10% of the videos for each rater, with a minimum acceptance rate of 80%.
Results Participants Twenty patients were included in the analysis. Eleven patients were men, with a mean age of 56.6 years. Participant demographics, location of infarct, and cumulative modified Oral and Pharyngeal Impairment MBSImP scores are listed in table 3. Despite instances of penetration and aspiration, the median Penetration-Aspiration Scale scores by bolus type were low (1 or 2).
Timing of swallowing events and Penetration-Aspiration Scale In all patients, the bolus entered the hypopharynx later for 30 fps than for 15 fps in solid swallows; this measure was later for 15 fps than for 30 fps when drinking nectar-thick liquids (table 4). For swallowing of sequential thin boluses, the bolus entered the upper esophageal sphincter later for 30 fps than for 15 fps. For pudding boluses, pharyngeal transit time was longer for 30 fps than for 15 fps. No other measures of timing were statistically significantly different between the 2 frame rates. There were no significant differences on Penetration-Aspiration Scale scores by frame rate. Average interrater reliability ICC values were 0.974 (95% confidence interval [CI] 0.968-0.98) for both timing and Penetration-Aspiration Scale ratings, and average intrarater reliability values for the 2 raters were 0.954 (95% CI 0.926-0.972) and 0.994 (95% CI 0.99-0.997). www.archives-pmr.org
Table 2
MBSImP Of the 14 MBSImP component scores, 15-fps swallows received more severe ratings of Oral Residue than for 30 fps during boluses of thin liquids by cup and pudding (table 5); similarly, Pharyngoesophageal Segment Opening was scored more severely for 15 fps than for 30 fps in sequential swallows of thin liquids. Bolus Transport/Lingual Motion and Initiation of Pharyngeal Swallow scores were more severe for 30 fps than for 15 fps during solid swallows (see table 5). Average interrater reliability ICC values were 0.86 (95% CI 0.834-1), and average intra-rater reliability ICC values for the 2 raters were 0.90 (95% CI 0.868-0.929) and 0.99 (95% CI 0.987-0.993).
Discussion The results of this work suggest that use of a lower frame rate (15fps), a proxy for 15-pps fluoroscopy, leads to different durations and ratings of swallowing events in comparison to continuous fluoroscopy at 30 pps. Thus, the diagnosis of swallowing impairments and subsequent treatment decisions may be affected.
Timing of swallowing events and Penetration-Aspiration Scale In all bolus consistencies analyzed, there was a trend for longer swallowing events in 30 fps in comparison to the same swallows at 15 fps (see table 4). Differences between the 2 frame rates may be due to events occurring during frames that were present at 30 fps but were omitted at 15 fps. Additionally, the context of each frame may assist in visual perception. Judgment of timing and clinical measures requires consideration of each frame as part of the dynamic process of swallowing rather than as an isolated slice. At
4
R.W. Mulheren et al Table 3
Participant demographics
Participant Number
Age, y
Sex
Infarct Location
Clinical Impression
MBSImP Oral Impairment Score (Modified)
MBSImP Pharyngeal Impairment Score (Modified)
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
55 57 57 43 67 79 48 51 57 70 25 38 71 60 74 55 48 55 73 48
Male Female Female Male Male Male Male Male Female Male Female Male Female Male Male Female Female Female Female Male
MCA MCA MCA MCA MCA Pons MCA Pons MCA ACA ACA, MCA MCA, PCA MCA BG MCA Pons MCA MCA MCA MCA
WNL Dysphagia Dysphagia Dysphagia WNL Dysphagia WNL Dysphagia Dysphagia Dysphagia Dysphagia WNL Dysphagia Dysphagia Dysphagia Dysphagia Dysphagia Dysphagia Dysphagia Dysphagia
11 8 10 13 8 13 7 7 12 11 7 7 10 10 11 11 13 11 14 9
4 7 5 2 1 12 1 7 10 4 4 5 8 7 10 7 8 10 14 9
Abbreviations: ACA, anterior cerebral artery; BG, basal ganglia; MCA, middle cerebral artery; WNL, within normal limits.
15 fps, the 2 frames surrounding a single frame at 30 fps are omitted, thus altering the context in which the single frame is judged. The timing measures that were significantly different for 15 and 30 fps were not uniform across bolus types. The difference in bolus properties between thin, nectar, pudding, and solid may have affected the results.14 The timing measures were not affected equally for thin liquid swallowing for a single cup sip vs a sequential swallowing task, which may be due to different bolus volumes as well as the different physiological responses observed during these tasks.15,16 This study included acute stroke patients with and without a clinical impression of dysphagia upon VFSS. Pharyngeal transit time, a measure that may predict penetration-aspiration in stroke patients, was more delayed in 30-fps than 15-fps videos.17,18 Although the current exploratory study did not have large enough numbers to stratify by dysphagia diagnosis or severity, the discrepancy between 15 and 30 fps may be due to more difficult visual perception of swallowing events and bolus transit in the presence of physiologic abnormality.
Table 4
Although Cohen et al8 reported penetration lasting only 1 frame in the majority of a pediatric sample at 30 fps that would likely have been missed at a lower pulse rate, we found no differences between 15 fps and 30 fps on Penetration-Aspiration Scale scores in adult stroke patients. This discrepancy may be due to differences between pediatric and adult swallowing, etiology of dysphagia or severity of penetration/aspiration in the current patient sample.
MBSImP The MBSImP data suggest that the videofluoroscopic frame rate affects the assessment of physiologic impairment. In other words, a swallow rated at 30 fps is perceived to have different degrees of impairment than the same swallow at 15 fps. Considering the significant contribution of swallowing assessment to the development of treatment, frame rate may directly affect clinical care. Furthermore, the validation of the MBSImP was completed at 30 fps,2 thus a lower frame rate may produce invalid results on this measure.
Wilcoxon sign rank test results comparing timing of swallowing measures for paired swallows Duration (Median, in ms)
Measure
Bolus Type
15 fps
30 fps
z Statistic*
P Value
Bolus enters upper esophageal sphincter from hyoid burst Bolus enters hypopharynx from hyoid burst
Sequential thin Nectar Solid Pudding
66.67 199.8 <.01 600.03
133.32 49.95 49.99 733.26
-2.42 2.38 -2.71 -2.29
.015 .017 .007 .022
Pharyngeal Transit Time
* Positive z statistic indicates longer durations for 15 than 30 fps; negative z statistic indicates shorter durations for 15 than 30 fps.
www.archives-pmr.org
Ratings of swallowing poststroke Table 5
5
Wilcoxon signed rank test results comparing MSBImP component scores for paired swallows Median Score
Measure
Bolus Type
15 fps
30 fps
z Statistic*
P Value
Oral Residue
Cup sip thin Pudding Solid Sequential thin Solid
2 2 0 1 1
1 1 2 0 1
2.65 2.33 -2.78 2.24 -2.45
.008 .020 .005 .025 .014
Bolus Transport/Lingual Motion Pharyngoesophageal Segment Opening Initiation of Pharyngeal Swallow
* Positive z statistic indicates higher, more severe ratings for 15 than 30 fps; negative z statistic indicates lower, less severe rating for 15 than 30 fps.
As bolus transport, initiation of pharyngeal swallow, and pharyngoesophageal segment opening received higher scores at 30 fps than at 15 fps, the degree of impairment was underestimated at 15 fps. Interestingly, oral residue was lower, or less impaired, at 30 fps than at 15 fps. However, an oral residue score of 1 is within normal limits,19 thus differences between 0 and 1 on this component are not practically significant. As with the timing of swallowing events, the difference in scores between the 15 fps and 30 fps may be explained by altered visual perception given contextual changes when 50% of frames are dropped.
Keywords
Study limitations
References
This study included a relatively small sample of acute stroke patients and may not be generalizable to people with other diseases resulting in dysphagia. In the present exploratory study, several of the comparisons were not significant; a larger sample with stratification by dysphagia severity and a higher median Penetration-Aspiration Scale score may yield different results. Bolus consistency and size were limited; thus, different results may be obtained from analysis of other bolus types. Examination of additional measures of swallowing function, such as kinematics, would supplement the reported results. Further research is warranted to determine the effect of lower pulse rates on treatment recommendations.
Conclusion In this sample, frame rate had a quantifiable effect on certain measures of swallowing function in patients poststroke. The use of pulsed fluoroscopy at 15 pps may distort swallowing measures, and further research is warranted to determine the effect on clinical judgment. The results of this study support the most commonly accepted standard of continuous fluoroscopy or 30 pps for optimal evaluation of swallowing function, though further, large-scale research is necessary.
Suppliers a. b. c. d.
MultiDiagnost Eleva with Flat Detector v3.2; Philips. Varibar; Bracco. VirtualDub; Avery Lee. QuickTime Player 7; Apple, Inc.
www.archives-pmr.org
Fluoroscopy; Pulse rate; Rehabilitation; Stroke; Swallowing
Corresponding author Rachel W. Mulheren, PhD, Case Western Reserve University, 11635 Euclid Ave., Cleveland, OH 44107. E-mail address: [email protected].
1. Mahesh M. Fluoroscopy: patient radiation exposure issues. Radiographics 2001;21:1033-45. 2. Martin-Harris B, Brodsky MB, Michel Y, et al. MBS measurement tool for swallow impairmentdMBSImp: establishing a standard. Dysphagia 2008;23:392-405. 3. Aufrichtig R, Xue P, Thomas CW, Gilmore GC, Wilson DL. Perceptual comparison of pulsed and continuous fluoroscopy. Med Phys 1994;21:245-56. 4. Molfenter SM, Steele CM. Kinematic and temporal factors associated with penetrationeaspiration in swallowing liquids. Dysphagia 2014; 29:269-76. 5. American Speech-Language-Hearing Association. Guidelines for speech-language pathologists performing videofluoroscopic swallowing studies. Available at: https://www.asha.org/policy/ GL2004-00050/. Accessed March 5, 2017. 6. Dodds WJ, Stewart ET, Logemann JA. Physiology and radiology of the normal oral and pharyngeal phases of swallowing. AJR Am J Roentgenol 1990;154:953-63. 7. Jones B, Donner MW. How I do it: examination of the patient with dysphagia. Dysphagia 1989;4:162-72. 8. Cohen MD. Can we use pulsed fluoroscopy to decrease the radiation dose during video fluoroscopic feeding studies in children? Clin Radiol 2009;64:70-3. 9. Bonilha HS, Blair J, Carnes B, et al. Preliminary investigation of the effect of pulse rate on judgments of swallowing impairment and treatment recommendations. Dysphagia 2013;28:528-38. 10. Curtis DJ, Cruess DF, Dachman AH, Maso E. Timing in the normal pharyngeal swallow. Prospective selection and evaluation of 16 normal asymptomatic patients. Invest Radiol 1984;19:523-9. 11. McConnel FM, Cerenko D, Jackson RT, Guffin TN Jr. Timing of major events of pharyngeal swallowing. Arch Otolaryngol Head Neck Surg 1988;114:1413-8. 12. Inamoto Y, Fujii N, Saitoh E, et al. Evaluation of swallowing using 320-detector-row multislice CT. Part II: kinematic analysis
6 of laryngeal closure during normal swallowing. Dysphagia 2011; 26:209-17. 13. Rosenbek JC, Robbins JA, Roecker EB, Coyle JL, Wood JL. A penetration-aspiration scale. Dysphagia 1996;11:93-8. 14. Steele CM, Alsanei WA, Ayanikalath S, et al. The influence of food texture and liquid consistency modification on swallowing physiology and function: a systematic review. Dysphagia 2015; 30:2-26. 15. Daniels SK, Foundas AL. Swallowing physiology of sequential straw drinking. Dysphagia 2001;16:176-82.
R.W. Mulheren et al 16. Chi-Fishman G, Sonies BC. Motor strategy in rapid sequential swallowing: new insights. J Speech Lang Hear Res 2000;43:1481-92. 17. Bingjie L, Tong Z, Xinting S, Jianmin X, Guijun J. Quantitative videofluoroscopic analysis of penetration-aspiration in post-stroke patients. Neurol India 2010;58:42-7. 18. Kim Y, McCullough GH. Stage transition duration in patients poststroke. Dysphagia 2007;22:299-305. 19. Martin-Harris B, McFarland D, Hill EG, et al. Respiratory-swallow training in patients with head and neck cancer. Arch Phys Med Rehabil 2015;96:885-93.
www.archives-pmr.org