research
Finger on Read this if you want to •carry out a literature review •test a hypothesis •identify noninvasive portable solutions
Although videofluoroscopy can identify dysphagia and distinguish between penetration and aspiration on swallowing, it is not always available or appropriate. Judi Hibberd and colleagues investigate the potential of using the less invasive pulse oximetry in conjunction with cervical auscultation as a bedside screen for dysphagia.
R
isk assessment in dysphagia management considers the likelihood of the client being harmed from continuing oral intake, and the potential severity of that harm. The key to risk assessment is whether or not the client is aspirating. So, how do we find out? Videofluoroscopy is accepted as the ‘gold standard’ (Bastian, 1993). However, the procedure is invasive, labour intensive and resource consuming. Furthermore, it is not easily accessible and cannot be performed at the bedside. Other screening tools are cervical auscultation and pulse oximetry, both of which are non-invasive and readily available to most therapists working with clients with dysphagia. Cervical auscultation enables deduction of the point in the swallow where the problem occurs. Pulse oximetry allows instant and continuous monitoring of the oxygen saturation of the blood, and a drop in saturation may be an important indication of respiratory compromise. As the airway is not directly compromised either during a normal swallow or penetration,, it may be that saturation penetration saturation levels only drop during aspiration, allowing the therapist to distinguish between aspiration and penetration at the bedside. The literature is less convincing on pulse oximetry than cervical auscultation so our hypothesis was that, if pulse oximetry oximetry could be shown to be effective, effective, we could use the two tools together to create a valuable, non-invasive bedside assessment system.
A reliable tool Research into the effectiveness of cervical auscultation has firmly established it as a reliable tool. Cichero & Murdoch (1998) hypothesised that the sounds generated by pressure changes during
SPEECH & LANGUAGE THERAPY IN PRACTICE SPRING 2004
aspiration are distinct from those made during a normal swallow. Hirano et al (2000) added further weight when they found dysphagia diagnosed by auscultation demonstrated an 87.8 per cent correlation with that diagnosed by videofluoroscopy. Research into pulse oximetry is more contradictory and less conclusive. So far there is no general consensus as to the efficacy of the tool, as results and their interpretation vary widely. Collins & Bakheit (1997), using pulse oximetry in conjunction with videofluoroscopy, concluded that the pulse oximeter indicated a two per cent drop in saturation levels in aspirators. Zaidi et al (1995) also found that saturation levels in aspirators when eating and drinking decreased significantly more than in non-aspirators. Sherman et al (1999) too found it useful to a statistically significant degree, commenting that, “These encouraging study results indicate that pulse oximetry may become an essential component of the management of dysphagic patients.” Conversely,, Sellars et al (1998) found no significant Conversely correlation between oxygen saturation and aspiration. Leder (2000) did find a statistically significant difference between aspirators and non-aspirators but concluded concluded that there was no indication indication that pulse oximetry is clinically relevant. Colodny (2000) found that, while aspirators’ saturation levels drop during feeding, this was not significant and that the two per cent drop suggested by Collins & Bakheit (1997) would not be enough to predict aspiration. From this contradictory research, two points caught our attention: 1) Controls and subjects Pulse oximetry will only be effective in dysphagia screening if it can be shown conclusively that a person at no risk of aspiration does not experience a drop in saturation levels when eating / drinking. Yet, in the research designs, numbers of controls have varied from 0 (Sherman et al, 1999; Collins & Bakheit 1997) to 77 (Colodny, 2000). Leder (2000) collected all of his data from patients in an intensive care ward; his control data cannot therefore be assumed to reflect what is ‘normal’. Another difference between studies is participants’ medical status. Sellars et al (1998), who found no significant difference between saturation in aspirators and non-aspirators, noted that their subjects had respiratory difficulties. Perhaps those with respiratory difficulties have different results to those without them? 2) Establishment of mean baseline Sellars et al (1998) calculated baseline saturation measurements to one per cent and then noted deviation from that point. They concluded that a significant drop would be four per cent. This is in
research
the pulse contrast with Collins & Bakheit (1997) who suggested a two per cent drop was significant. We wondered if the lack of consensus establishment of significant percentage decrease could be due to the original assumption of the baseline level as a single point. If, on the other other hand, oxygen saturation saturation is not fixed and stable but fluctuates within a band, then the deduction of level of decrease from a single point will not encourage statistically significant or useful results. For example, a patient who is aspirating may drop out of their baseline band, while a non-dysphagic may also experience a decrease but remain within their banding level. Taking into account these methodological problems, we aimed: 1.To use pulse oximetry to investigate whether saturation levels drop in people with dysphagia. 2.To collect saturation data from non-dysphagics as a large control group to compare non-dysphagics with people with dysphagia. 3. To investigate if there is a significant difference in saturation levels between those with neurologica neurologicall and those with respiratory dysphagia. 4. To use cervical auscultation alongside pulse oximetry to detect the point during the swallow where the problem occurs. Permission was granted from two local trusts in March 2000. We categorised the 92 participants (35 male, 57 female, aged 15-99 years) into one of four groups: Group A - normal controls (n=50) Non-dysphagic volunteers from the local community with no history of neurological or respiratory problems who were non smokers or ex-smokers of five years or more. Group B - respiratory controls (n=17) Non-dysphagics with a respiratory element who were volunteers from a respiratory function test clinic at the hospital, and who had no history of any neurological condition. Eleven had chronic obstructive pulmonary disease and two had asthma. Group C - respiratory dysphagia (n=5) People with dysphagia which appeared to have a respiratory base. One had chronic obstructive pulmonary disease and one renal problems. Group D - neurological dysphagia (n=20) People with dysphagia which appeared to have a neurological base. Two had no confirmed diagnosis at referral and 18 had cerebrovascular accidents. Groups C and D were part of a clinical caseload referred to us by a medical practitioner for a swallowing assessment. For all subjects two therapists simultaneously collected data using the pulse oximeter and cervi-
cal auscultation. The same equipment was used for all data collection and the two collectors did not exchange information except for notification of when a swallow occurred during the procedure. Firstly, the finger sensor was attached and a reading taken to establish the participant’s baseline. This took two minutes timed by a stopwatch. (Participants were excluded if the finger sensor could not be attached adequately or if they were wearing nail varnish.) The pulse oximeter then indicated that the therapist could begin the clinical evaluation. Cervical auscultation involved the participant taking three sips of liquid and three teaspoons of semi-solid food. Risk was assigned based on any acoustic changes and without access to any information from the pulse oximeter. Data was included only if the therapist completed either a food or drink assessment consisting of three swallows (De Pippo et al, 1992) but the assessment was curtailed if the therapist felt that the risk level was severe before the three swallows had been completed. When the therapist heard the participant swallow she indicated this to her assistant who noted down the exact time on the stopwatch which ran continuously from baseline to completion. The therapists then downloaded the participant’s saturation levels at each of the times on the data sheet that indicated a swallow had occurred. We could therefore examine any changes at rest or during swallows. Where participants were part of the therapist’s caseload, the bedside clinical evaluation was completed by communicating the results of the assessment to the relevant multidisciplinary team.
When the therapist heard the participant partici pant swallow she indicated this to her assistant who noted down the exact time on the stopwatch which ran continuously from baseline to completion.
Data was collected over a nine month period and analysed using SPSS Version 9.
Fluctuation The first significant finding involved the normal control group. Results indicated that normals’ saturation levels did not remain static for the two minute baseline establishment period but fluctuated by up to three per cent. This fluctuation continued during eating and drinking but, significantly, never dropped out of the three per cent band. Having established a norm to a statistically significant level, we went on to investigate saturation levels in neurological dysphagics. Results showed that, although the baseline level was lower than the controls, all of the neurological dysphagic group also had a baseline saturation level within the three point band when at rest. Thus we established that, for both dysphagics and non-dysphagics at rest, oxygen saturation does not necessarily sit at one point but fluctuates by up to three per cent. When taking fluid or food, the results strongly suggested that those at severe risk of dysphagia, unlike the controls, dropped out of this baseline band. The results were not statistically significant overall for this but were significant for the first swallow of food and of liquid. This was the only unbiased sample. The lack of significance for the second and third swallows may be due to ethical considerations preventing the therapist from continuing with the assessment when auscultation indicated a severe risk. Also significant was the finding that those ascribed mild risk using cervical auscultation tended not to drop out of the three point band. This suggests that they may not require clinical compensation. The evidence for the two respiratory groups followed the same pattern of the three point band when at rest, with respiratory controls tending to have a lower baseline than normal controls and, in some cases, lower than neurological dysphagics. This suggests that previous evidence that was not statistically significant may have been adversely affected by the inclusion of control data from participants with respiratory compromise. The data for respiratory dysphagia also seems to follow a similar pattern as for neurological dysphagia but, interestingly, some participants’ saturation levels moved upwards out of their baseline band when eating/drinking. Our sample was very limited in number, but this suggests that there may be a difference between neurological and respiratory dysphagia, an idea that further research could investigate. Is it possi-
SPEECH & LANGUAGE THERAPY IN PRACTICE SPRING 2004
research
..resources...
Signalong Hands for Talking, a partnership of two media professionals who are parents of children with special needs, has produced a video to accompany Signalong’ss nursery rhyme book and CD. Signalong’ Sign A Song, see www.handsfortalking.com, www.handsfortalking.com, £14.74 inc. VAT and postage. And, if you are already thinking about next Christmas, Signalong also has a book of secular celebrations to complement ‘The Little Star’ with its nativity and carols. We Wish You A Merry Christmas, £11.75 inc. postage. www.signalong.org.uk, www.signal ong.org.uk, tel. 0870 774 3753
Computers and disability National computing and disability charity AbilityNet has released a step-by-step guide on CD-ROM to help people with disabilities access and make the most of the opportunities offered by computers. Successful computing for disabled people is £45 + VAT, tel. 01926 312847, www.abilitynet.org.uk.
Helping Children with Feelings Four new titles in Speechmark’s Helping Children with Feelings series can be used by child professionals and parents to help children aged 4-12 resolve fear, low selfesteem, loss and rage or hate. Each guidebook has an accompanying illustrated storybook (How Hattie Hated Kindness; Ruby and the Rubbish Bin; Teenie Weenie in a Too Big World; The Day the Sea Went Out and Never Came Back). £90 for all four sets, www.speechmark.net. www.speechmark.net.
Bilingual assistance A teacher support pack to boost the attainment of minority ethnic pupils in England and Wales will be released shortly. Based on research carried out in Derby, the 30 card pack recommends the use of bilingual language assistants in the classroom, introducing dual language tests, raising confidence and motivation, encouraging parental involvement, ensuring effective leadership and introducing elements of the curriculum relevant to the child’s cultural background. www.derby.ac.uk
Computer recycling Tools for Schools is a not-for-profit organisation that refurbishes surplus business computers for low-cost sale to schools to help them boost pupil computer access. www.tfs.org.uk
ble that the lack of significant difference between controls and dysphagics identified in previous research could be attributed to the inclusion of both respiratory and neurological dysphagics in the same group?
Statistically significant We therefore reached two main conclusions. Firstly, we established to a statistically significant level that, when at rest, oxygen saturation levels of all participants remained within a three point band. The sample size was such that we can extrapolate the results to suggest that all oxygen saturation levels fall within a three point band when at rest. Secondly, all control subjects, and those at mild risk of dysphagia, remained within that band whether eating and drinking or at rest. A significant number of those at severe risk of dysphagia dropped out of the band when eating and drinking. This was established to the highly statistically significant level of p<0.001. Thus our hypothesis that pulse oximetry may be able to distinguish between aspiration and penetration cannot be discounted, and we suggest that, in conjunction with cervical auscultation, it can be used as an effective and reliable screening system at the bedside. Judi Hibberd, Alison Shale, Stephanie Stephanie Bowers and Karen Miles are speech and language therapists with the University Hospitals Cov and Warwick Trust & Coventry Healthcare NHS Trust. Correspondence to: Mrs Judi Hibberd, Speech and Language Therapy Department, Gulson Hospital, Gulson Road, Coventry, CV1 2HR, tel: 024 7624 6262.
References Bastian, R.W. (1993) The videoendoscopic swallowing study, an alternative and partner to the videofluoroscopic swallowing study. Dysphagia 8: 359-367. Cichero, J.A.Y. & Murdoch, B.E. (1998) The physiological cause of swallowing sounds: answers from sounds and vocal tract acoustics. Dysphagia 13: 39-52. Collins, M.J. & Bakheit, M.D. (1997) Does pulse oximetry reliably detect aspiration in dysphagic stroke patients? Stroke 28 (9): 1773-1775.
Colodny, N. (2000) Comparison of dysphagics and non dysphagics on pulse oximetry during oral feeding. Dysphagia 15: 68-73. De Pippo, K.L., Holas, M.A., & Reding, M.J. (1992) Validation of the 3 oz water swallow test for aspiration following stroke. Arch Neurol 49: 1259 -1261. Hirano, K., Takahashi, K., Uyama, R. & Michi, K. (2000) Evaluation of cervical auscultation: the accuracy and acoustic characteristics. Abstract of presen tation at the ninth annual dyspha gia research society meeting , Georgia USA. Leder, S.B. (2000) Use of arterial oxygen saturation, heart rate and blood pressure as indirect objective physiologic markers to predict aspiration. Dysphagia 15: 201-205. Sellars, C., Dunnet, C. & Carter Carter,, R. (1998) A preliminary comparison of videofluoroscopy of swallow and pulse oximetry in the identification of aspiration in dysphagic patients. Dysphagia 13: 82 -86. Sherman, B., Nisenboum, J., Jesberger, B., Morrow, C. & Jesberger, J.A. (1999) Assessment of dysphagia with the use of the pulse oximeter. Dysphagia 14:152-156. Zaidi, N.H., Smith, A.H., King, S.C., Park, C., O’Neill, P.A. & Connelly, M.J. (1995) Oxygen desaturation on swallowing as a potential marker of aspiration in acute stroke. Age and Ageing 24: 267-270.
Acknowledgements We are grateful to Rav Jayram, our research supervisor, Bill Waine, our statistical analyst, Jan Lacey, Development Nurse, Jenni Hibberd and Russell Miles.
Reflections for research • Do I base comparisons comparisons with the ‘norm’ on fact rather than assumption? • Do I consider consider potential confounding confound ing factors when grouping subjects? • Do I seek expert expert assistance for supervision and analysis?
Finger on the pulse: dysphagia terminology update Aspiration: Entry of material below the true vocal folds.
Penetration: Entry of material into the laryngeal vestibule above the true vocal folds.
Cervical auscultation: The placement of a stethoscope on or near the cricoid cartilage to listen to the acoustic signal generated by a swallow sequence.
Pulse oximetry: The measurement of the ratio of oxygenated haemoglobin to the total haemoglobin in arterial blood.
Oxygen saturation: A measure of the amount of oxygen in the haemoglobin at a particular pressure.
SPEECH & LANGUAGE THERAPY IN PRACTICE SPRING 2004
Respiratory dysphagia: A swallowing problem with a respiratory basis, (chronic obstructive pulmonary disease; asthma) or a respiratory consequence (renal failure; cardiac compromise) rather than a neurological base.
