Background and Purpose The aim of the present study was to examine the value of pulse oximetry in the diagnosis of aspiration by comparing it with the gold standard, videofluoroscopy, by use of a prospective, controlled, single-blind study design.
Methods Pulse oximetry was performed simultaneously with videofluoroscopy in 54 consecutive dysphagic stroke patients. Oxygen saturation measurements were taken before the videofluoroscopic examination (baseline), on swallowing and continuously for 2 minutes after swallowing, and 10 minutes later.
Results Pulse oximetry reliably predicted aspiration or lack of it in 81.5% of cases. The predictive value of the test was low in patients aged ≥65 years and possibly those with chronic lung disease. One smoker also had a false-negative pulse oximetry result, ie, normal oxygen saturation despite radiological evidence of aspiration.
Conclusions Pulse oximetry is a reliable method of diagnosis of aspiration in most dysphagic patients. However, careful interpretation of pulse oximetry data is necessary in older subjects, possibly those with chronic pulmonary disease, and smokers. The method is noninvasive, simple, and quick, and can be used routinely in the clinical assessment of dysphagic patients.
Dysphagia is common after stroke, occurring in 25% to 42% of patients.1 2 About one third of patients with dysphagia aspirate food or fluid into their airways, and in 40% of them aspiration is silent (ie, it does not trigger a cough or cause symptoms or signs of distress).3 Detection of silent aspiration is important because aspiration often leads to such serious complications as pulmonary infections, septicemia, dehydration, and malnutrition.4
There are no reliable clinical signs of silent aspiration, and aspirating patients frequently do not complain of swallowing difficulties.4 Furthermore, bedside assessment of dysphagic patients, unless coupled with an objective method such as videofluoroscopy or nasal endoscopy,2 often fails to detect silent aspiration. Although nasal endoscopy detects silent aspiration in nearly all patients,5 the procedure requires a certain degree of expertise and is time consuming and sometimes poorly tolerated by patients. Similarly, videofluoroscopy is not suitable for routine clinical use, especially when repeated assessments of severely ill patients are required.
In recent years, pulse oximetry has been advocated as an alternative to nasal endoscopy and videofluoroscopy for the assessment of dysphagic patients. This method is based on the principle that reduced and oxygenated hemoglobin exhibit different absorption characteristics to red and infrared light emitted from a finger (or earlobe) probe. It has been suggested6 that aspiration of food or fluid into the airways causes reflex bronchoconstriction that leads to ventilation-perfusion mismatch and oxygen desaturation of arterial blood, which can be readily measured by pulse oximetry.
Pulse oximetry is noninvasive, simple, and repeatable, and it does not involve exposure to radiation. However, the value of this method in the diagnosis of silent aspiration has yet to be fully examined. In 49 patients with stroke, Zaidi et al7 studied oxygen saturation after the swallowing of 10 mL of water. They then compared the pulse oximetry results with the findings of bedside assessment by an independent speech and language therapist. The authors found a close correlation between aspiration as diagnosed clinically by the therapist and the drop in arterial blood oxygen saturation measured by pulse oximetry. However, the authors did not perform the dysphagia assessment simultaneously with the pulse oximetry. This poses a problem with the interpretation of their data, as aspiration may occur intermittently. Furthermore, the subjective clinical evaluation of dysphagia is unreliable: 42% to 60% of patients aspirating food or fluid are not recognized.3 8 To date, the reliability of pulse oximetry in the diagnosis of aspiration has not been systematically studied in a controlled trial comparing it with the gold standard, videofluoroscopy. The aim of the present study was to establish whether pulse oximetry is a reliable method for the detection of aspiration in dysphagic stroke patients.
Subjects and Methods
The study was conducted using a prospective, controlled, single-blind design. Fifty-four consecutive stroke patients with swallowing difficulties who were referred to a videofluoroscopy clinic between May and November 1996 were recruited from an inpatient and outpatient population. The diagnosis of stroke was made by a trained physician according to WHO criteria.9 No CT confirmation of the diagnosis was sought. Patients with Raynaud’s disease or severe peripheral vascular disease were excluded from the study. There were 28 men and 26 women, with a mean age of 65 (range, 28 to 86) years. Thirteen patients had a history of chronic obstructive airway disease, and 1 patient was a cigarette smoker at the time of the study.
Arterial oxygen saturation was measured with a Minola Pulsox 7 monitor (De Vilbiss Healthcare) by use of a finger probe. The test was carried out in a warm, dimly lighted room. Nail polish was removed from patients’ fingernails before the test. These precautions were followed to minimize data collection errors.10 The probe was attached to the index finger of the nondominant hand, and patients were instructed to keep that arm still during the study to prevent movement artifact. The pulse oximeter was allowed to equilibrate for 5 minutes. A baseline measurement of oxygen saturation was then recorded before the videofluoroscopic examination was made. Three further readings were then taken as follows: (1) during the act of swallowing; (2) during the 2 minutes after the test meal (the lowest reading was recorded), and (3) 10 minutes after the completion of videofluoroscopy. These measurements were repeated, with the patient ingesting thin liquid, pudding, and a biscuit impregnated with barium.
Videofluoroscopy was performed by a radiologist with the patient standing or sitting. The barium meal consisted of 150 mL liquid (single cream consistency), 3 oz mousse (puree consistency) and one half of a 2-inche (diameter) barium-impregnated shortbread biscuit. Patients were asked to feed themselves the barium meal but were assisted if necessary.
Pulse oximetry was carried out simultaneously with the radiological examination by an independent researcher. Following the completion of the test, the researcher reviewed the videofluoroscopy tapes with the radiologist to establish the exact time at which aspiration occurred during the examination. The manufacturer’s literature suggests that the error of the Minola Pulsox monitor is <2%. A drop of ≥2% in the arterial oxygen saturation was therefore considered clinically significant. The data were analyzed using κ statistics, with values of κ>0.60 accepted as evidence of good correlation.11
Videofluoroscopy confirmed that 13 men and 9 women aspirated during the study. Seven patients in each group had symptoms of coughing, choking, or splattering while eating, and 6 men and 2 women had silent aspiration. The mean ages of those who aspirated and those who did not were 72 and 61 years, respectively.
As shown in Table 1⇓, the mean arterial oxygen saturation levels during the test fell by >2% from baseline at 2 minutes after swallowing/aspiration in aspirating male patients. The maximum desaturation in aspirating females was 1.6%, which occurred at the point of aspiration. Variable improvement in oxygen saturation occurred in all subjects 10 minutes after completion of videofluoroscopy, with the least recovery occurring in dysphagic males who aspirated.
When the data were analyzed according to age, there was significantly more desaturation in aspirating male patients aged <65 years than in their older counterparts. A similar trend was observed in female aspirators, but the mean oxygen desaturation was less, and the recovery to baseline values was almost complete within 10 minutes of the completion of videofluorosopy. There was no significant desaturation in nonaspirators aged ≥65 years, irrespective of gender (Table 2⇓).
Correlation of the pulse oximetry results with the videofluoroscopic findings showed that 12 (55%) of the patients who aspirated had a significant degree of oxygen desaturation at the point of swallow/aspiration, but none of the nonaspirators desaturated by ≥2% in the swallow/aspiration stage. However, when the results of oximetry at swallow/aspiration and at 2 minutes after swallowing were combined, 16 (73%) of the aspirators could be identified by this method, and 4 (13%) of the nonaspirators also had a significant oxygen desaturation. In the study sample as a whole, 44 patients (81.5%) were accurately predicted as aspirators or nonaspirators (κ=0.61, P<.001). Interestingly, the prediction rate in male patients <65 years of age was 100%, but the rate in females of this age group was 67% (Table 3⇓). The sensitivity and specificity of pulse oximetry were 73% and 87%, respectively.
The present study has demonstrated that pulse oximetry accurately predicted aspiration (or lack of it) in 81.5% of dysphagic stroke patients. Although a drop of ≥2% in oxygen saturation on swallowing detected this complication in most patients, a higher correlation between oxygen desaturation and aspiration was present when we also took into account the lowest pulse oximetry reading during the 2 minutes after swallowing. The study also showed a lower baseline pulse oximetry reading in aspirating compared with nonaspirating patients. This could be due to a more pronounced weakness of the chest wall muscle in those who aspirated.12 Patients who aspirated were generally older than those who did not, and the lower baseline oxygen saturation in these subjects could also result from age-related changes in respiratory function.13
The predictive value of pulse oximetry was better in male patients who were <65 years of age compared with those who were older. Older subjects have a reduced pharyngeal sensitivity14 and require a larger-than-average bolus to trigger a swallow.15 It is possible that they also need a larger amount of aspirate for reflex bronchoconstriction and desaturation to occur. An unexplained observation is that female patients who aspirated had a smaller degree of oxygen desaturation than male patients.
In 4 patients who had evidence of aspiration on videofluoroscopy, oxygen desaturation was <2%. Although pulse oximetry recordings correlate closely with arterial blood gases,16 there may be a time delay between the onset of the hypoxic event and its detection by pulse oximetry.17 In this study we took the second measurement during radiographic examination 2 minutes after swallow/aspiration in order to reduce the screening time and minimize the patients’ exposure to radiation. This may have resulted in some false-negative oximetry findings, ie, normal oxygen saturation in the presence of aspiration. We recorded oxygen saturation from the index finger of the paretic hand. This procedure may have also led to false-positive results, as the changes in blood-vessel tone in the paralyzed arm could cause delay in the detection of arterial blood desaturation.18 Delayed pulse oximetry readings (eg, 5 minutes after swallow/aspiration) would have probably given a more accurate result. One of the patients in this group was a cigarette smoker whose false-negative result may have been caused by high carboxyhemoglobin concentrations.10 Desaturation during meals in the absence of aspiration has been reported in patients with chronic pulmonary disease,19 and this may also account for some of the false-positive results in this study.
In conclusion, the present study has demonstrated that pulse oximetry reliably detects aspiration in most dysphagic patients and may be used as an adjunct to the bedside assessment of dysphagia. However, careful interpretation of the pulse oximetry data is necessary in older subjects, smokers, and possibly those with chronic lung disease.
The research project leading to this publication was carried out in partial fulfillment of degree requirements (Master of Science in Rehabilitation Studies; M.J.C.) and was supported by the Stroke Association of Great Britain. The assistance of the Department of Medical Statistics, University of Southampton; the help of the radiology staff at Salisbury District and Southampton General hospitals; and the advice of Dr V.J. Singh and Jay Conway are warmly acknowledged.
- Received February 25, 1997.
- Revision received March 25, 1997.
- Accepted June 2, 1997.
- Copyright © 1997 by American Heart Association
Kidd D, Lawson J, Nesbitt R, MacMahon J. Aspiration in acute stroke: a clinical study with videofluoroscopy. QJM. 1993;86:825-829.
Logemann J. Manual for the videofluoroscopic study of swallowing. 2d ed, Austin, Tex: Pro-Ed Inc; 1983.
Horner J, Brazer SP, Massey W. Aspiration in bilateral stroke: a validation study. Neurology. 1993;43:430-433.
Zaidi NH, Smith HA, King SC, Park C, O’Neil PA, Connolly MJ. Oxygen desaturation on swallowing as a potential marker of aspiration in acute stroke. Age Ageing. 1995;24:267-270.
Ralston AC, Webb RK, Runciman WB. Potential errors in pulse oximetery. Effects of interferences, dyes, dyshaemoglobins and other pigments. Anaestheia. 1991;46:291-295.
Walshaw MJ, Pearson MG. Hypoxia in patients with acute hemiplegia. BMJ. 1984;288:15-17.
Levitzky MG. The effect of aging on the respiratory system. Physiology. 1984;27:102-107.
Servinghaus JW, Naiefeh KH. Accuracy of response of six pulse oximeters to profound hypoxia. Anaesthiology. 1987;67:551-558.
Wilkins CJ, Moores M, Hanning CD. Comparison of pulse oximeters: effects of vasoconstriction and venous engorgement. Br J Anaesth. 1989;62:439-444.