Comparative analysis of videofluoroscopy and pulse oximetry for aspiration identification in patients with dysphagia after stroke and non-dysphagics

Dysphagia, a condition characterized by difficulties in swallowing, frequently occurs in acute stroke due to disruption of the normal swallowing mechanism. Dysphagia is prevalent in over half of stroke survivors [1]. Stroke is a prevalent neurological disease globally, associated with a high mortality rate [2]. In Turkey specifically, vascular system diseases are the leading cause of death, and among these diseases, cerebrovascular diseases rank as the third most common cause of mortality [3].

The presence of dysphagia may lead to various health complications such as dehydration, malnutrition, weight loss, increased hospitalization, and respiratory infections namely pneumonia, [4, 5]. Consequently, early identification and intervention in swallowing disorders are crucial to prevent the aforementioned life-threatening consequences, reduce hospital stay, and minimize long-term healthcare costs [6, 7].

National and international guidelines have recently emphasized the importance of dysphagia assessment and management in acute care settings worldwide [8‐11]. Several diagnostic techniques are utilized to determine the presence of dysphagia, including bedside assessment, clinical examination, and instrumental tests. Among these techniques, the videofluoroscopic swallowing study (VFSS) is considered the gold standard for comprehensive visualization of swallowing physiology and accurate identification of aspiration [12]. However, VFSS has certain drawbacks including radiation exposure, altered food consistency and sensory feedback due to the use of contrast media, and limited availability of specialized equipment [13]. The feasibility, applicability, widespread availability, and safety of pulse oximetry for bedside assessment, complementing existing diagnostic methods were proposed in a seminal study [14].

Pulse oximetry has been shown as a potential tool in assessing dysphagia and can be utilized in conjunction with other diagnostic approaches to enhance the overall assessment process [15, 16]. Several studies have indicated that the use of pulse oximetry to monitor oxygen desaturation during and immediately after swallowing contributes to diagnosing aspiration [17‐22]. Furthermore, some authors have suggested that combining bedside assessment with pulse oximeter measurements of SpO₂ may provide valuable insights into the detection of aspiration [22]. However, some studies have indicated that pulse oximetry is not a reliable instrument to detect the risk of aspiration in dysphagia and does not support the use of pulse oximetry to detect aspiration [22‐25]. It should be noted that some studies have also taken a more cautious approach to the detection of aspiration and call for further research [23].

Numerous studies on this subject have yielded different findings might be due to the different methodological factors adopted by the studies such as the definition of desaturation criteria. In the review conducted by Wang et al. [28], the potential risk of bias associated with different pulse oximetry parameters was highlighted. In their review, Britton et al. (2017) reported that among the 10 studies included in their analysis, 4 utilized a 2% SpO₂ decrease criterion, 1 used a 3% decrease, and 1 employed a 4% decrease. Another study directly compared desaturations of 2% and 3% [25], highlighting the variability in defining desaturation criteria across these studies, with a range of 2% to 4% SpO₂ decrease considered as a potential marker. Despite some conflicting outcomes, national guidelines still recommend a decline in SpO₂ > 2% as a marker of aspiration [23].

The primary objective of this study was to investigate the relationship between aspiration and oxygen desaturation levels in dysphagic patients, compared to a healthy control group. To achieve this, peripheral capillary oxygen saturation (SpO₂) is continuously monitored using pulse oximetry before, during (simultaneously with VFSS), and after swallowing considering the consistency used during VFSS. Furthermore, the study sought to determine whether there was a relationship between oxygen desaturation and the severity of aspiration. In parallel, we assessed the accuracy of pulse oximetry in detecting aspiration in both liquid and semi-solid consistencies, employing sensitivity, specificity, and area under the curve (AUC) values. By addressing these research questions, our study aimed to provide valuable insights into the potential utility of pulse oximetry as a non-invasive tool for early aspiration detection and dysphagia management.

Forty patients with post-stroke dysphagia with a mean age of 68.4 ± 11.9 and 40 healthy controls with a mean age of 64.05 ± 10.6, a total of 80 participants, took part in the study. Twenty female (50%) and 20 male (50%) participants took part in both participant groups. Regarding the stroke subtypes, 38 (95%) were diagnosed with ischemic stroke and 2 (5%) had a hemorrhagic stroke. As for the lesion site, 18 (45%) and 22 (55%) had lesions in the right and left hemispheres, respectively. Thirty-five patients (35%) had a stroke for the first time, four patients (10%) had a stroke for the second time and one patient (2.5%) had a stroke for more than two strokes. The mean stroke severity of the patient group was 10.15 ± 4.9 according to the National Institute of Health Stroke Scale (NIHSS).

The data in Table 2 indicate that there were significant differences among the two groups on SpO₂ values in terms of different time points including during and after the swallowing in both consistencies.

A statistically significant difference was obtained between the oxygen saturation values obtained during and after swallowing in both IDDSI-4 and IDDSI-0 consistency between the patient group and the healthy control group (p = 0.006, p = 0.002, p = 0.001, p = 0.0001 < 0.01).

It was examined whether there was a difference between the patient group and the healthy control group in terms of a decrease of ≥ 2% in SpO₂ values and the results obtained are shown in Table 3.

A decrease in SpO₂ value was observed in 8 (20%) patients in the patient group in IDDSI-4 consistency, while a decrease in SpO₂ value was observed in only 1 (2.5%) patient in the control group. There was a statistically significant difference between the groups in terms of a decrease in saturation ≥ 2% in IDDSI-4 consistency (p = 0.013). In IDDSI-0 consistency, a decrease in SpO₂ value was observed in 14 (35%) patients, and no decrease in SpO₂ value was observed in any participant in the control group. There was a statistically significant difference between the groups in terms of a decrease in saturation ≥ 2% in IDDSI-0 consistency (p = 0.0001).

Whether there was a change in oxygen saturation values in the patient group at two different time points according to the presence of aspiration was analyzed descriptively and the results obtained are shown in Table 4.

In IDDSI-4 consistency, aspiration was not observed in 39 (97.5%) of 40 patients, while aspiration was observed in 1 (2.5%) patient. However, no decrease in SpO₂ value was observed in the patient with aspiration. A decrease in SpO₂ value was observed in 8 (20.5%) of 40 patients without aspiration. Before swallowing, the mean SpO₂ level of the patients was 96.31 ± 1.9, whose SpO₂ value was already decreased without aspiration and the mean SpO₂ level after swallowing was 95.74 ± 1.9.

In IDDSI-0 consistency, aspiration was observed in 23 (57.5%) of 40 patients, while aspiration was not observed in 17 (42.5%) patients. In 14 (60.9%) of 23 patients with aspiration, a decrease of ≥ 2% in SpO₂ value was found. The mean SpO₂ level of aspirating patients before swallowing was 96.04 ± 2 and the mean SpO₂ level after swallowing was 94.61 ± 2. In none of the 17 patients in whom aspiration was not observed in IDDSI Level-0 consistency, a decrease of > 2% in SpO₂ value was detected. The mean SpO₂ level of the patients without aspiration was 96.47 ± 1.5 before swallowing and 96.53 ± 1.7 after swallowing.

It was analyzed whether there was a relationship between a decrease of ≥ 2% in SpO₂ value seen in measurements and aspiration detected with VFSS. The results obtained are shown in Table 5.

Comparing the results of pulse oximetry and the VFSS findings in IDDSI-0 consistency, 14 patients (60.9%) out of 23 patients who had an aspiration finding showed significant oxygen desaturation as well (SpO₂ drops exceed 2%). Of 17 patients with no aspiration on VFSS, (100%) none had a decrease in SpO₂ of more than ≥ 2%. There was a significant relationship between a decrease in SpO₂ and aspiration on VFSS examination using the Chi-square test (p = 0.0001).

In IDDSI-4 consistency, among 39 patients without aspiration, 31 (79.5%) exhibited no oxygen desaturation, while 8 (20.5%) experienced a ≥ 2% decrease in SpO₂ without aspiration. Notably, oxygen desaturation was not observed in one patient with aspiration at this consistency. There was no significant difference between a decrease in SpO₂ and aspiration on VFSS in IDSSI-4 consistency (p = 0.613).

The sensitivity and specificity, predictive values, and area under curve (AUC) of pulse oximetry in each consistency was evaluated by VFSS compared in 40 patients and the results obtained are shown in Table 6.

For IDDSI-4 consistency, the SpO₂ desaturation in aspiration had a sensitivity of 0%, specificity of 79.5% (95%CI 63.5–90.7%), PPV of 0%, NPV of 96.9% (95%CI 96.4–97.3%) and AUC of 0.48. For IDDSI-0 consistency, the SpO₂ desaturation in aspiration had a sensitivity of 60.9% (95%CI 38.5–80.3%), specificity of 100%(95%CI 80.5–100.0%), PPV of 60.9% (95%CI 53.2–75.8%), NPV of 100% (95%CI 93.5–100%) and AUC of 0.83 (Fig. 1).

The temporal changes in SpO₂ values before, during, and after swallowing in the IDDSI-4, and IDDSI-0 consistencies were examined between the patient and control groups, and the results obtained are presented in Fig. 2.

There was a statistically significant difference in SpO₂ values before, during, and after swallowing in the dysphagic group for both IDDSI-4 and IDDSI-0 consistencies (p = 0.007, p = 0.003). The difference was attributed to the variations between SpO₂ values before and during swallowing, as well as between SpO₂ values before and after swallowing.

In the healthy control group, there was no statistically significant difference in SpO₂ values before, during, and after swallowing for the IDDSI-4 and IDDSI-0 consistency (p = 0.469, p = 0.521).

The relationship between the score obtained from the Penetration-Aspiration Scale (PAS) and the observed decrease in SpO₂ values in patients with aspiration was examined, and the results are shown in Table 7.

In the present study, efforts were made to characterize in detail the relationship between oxygen desaturation and the presence of aspiration. For this purpose, pulse oximetry was utilized in patients with dysphagia before, during (simultaneously applying VFSS), and after-swallowing in the acute post-stroke period and compared with pulse oximetry values of individuals without dysphagia. In addition, it was determined whether there was a relationship between the presence of aspiration and oxygen desaturation according to consistency. It was also investigated whether there was a relationship between oxygen desaturation according the severity of aspiration determined by PAS.

In our study, aspiration was observed during VFSS evaluation in approximately 60% of acute stroke patients with suspected dysphagia according to GUSS, oxygen desaturation was detected simultaneously in more than half of them. The finding is consistent with studies supporting a decrease in SPO₂ values during and/or after swallowing during aspiration [17‐19, 21] but contradicts the findings of studies reporting that SPO₂ may not be a reliable indicator of aspiration [20, 22, 27, 28]. As aforementioned, several studies support that there is an association between aspiration and oxygen desaturation. Measurement tools used to determine dysphagia and aspiration, such as the Volume-Viscosity Swallowing Test [29], included pulse oximetry to detect silent aspiration, which occurs in about half of the patients. They concluded that pulse oximetry increases the diagnostic sensitivity of this test. However, the review by Britton et al. concluded that those inconsistent findings may be due to methodological differences or even different definitions of desaturation, and therefore the available evidence is not sufficient to support the use of pulse oximetry in detecting aspiration [25]. One reason for the inconsistent findings might be the clinically recognized phenomenon that aspiration does not necessarily occur with each swallow [30].

Despite these challenges, accurately defining silent aspiration and penetration accurately with only screening tests remains difficult. However, the Volume-Viscosity Swallowing Test (V-VST) asserts that by integrating pulse oximetry as a complementary measure, it provides a more reliable evaluation of dysphagia. Recent findings from the V-VST study demonstrate promising results, with 83.7% sensitivity and 64.7% specificity for bolus penetration into the larynx, and 100% sensitivity and 28.8% specificity for aspiration. These findings suggest that incorporating diagnostic accuracy and provide clinicians with more robust tools for assessing dysphagia and aspiration [29, 31].

In terms of consistency, in participants who had no complaints about swallowing and were assigned to the healthy control group, none of the participants had a decrease in SPO₂ at IDDSI-0 consistency, whereas in the dysphagia group, 35% of patients had a decrease in SPO₂. Aspiration was detected in approximately 60% of all patients in the IDDSI-0 consistency. It was determined that aspiration was accompanied by a decrease in SPO₂ in approximately 60% of patients with aspiration. In IDDSI-4 consistency, oxygen desaturation was detected in only one participant in the non-dysphagic group, while oxygen desaturation was detected in approximately 1/5 of the dysphagic group despite the absence of aspiration, and no decrease in SPO₂ was detected in only one participant with dysphagia despite aspiration. The data obtained indicate that there is a relationship between oxygen desaturation and aspiration depending on the consistency. When examined according to consistency, no significant difference was observed in oxygen saturation according to consistency in those without dysphagia, while oxygen desaturation decreased more in IDDSI-0 consistency than in IDDSI-4 consistency in the dysphagia group. The mechanism of oxygen desaturation during dysphagia is different from the normal swallowing process. Most likely, the bolus entering the larynx narrows the airway by stimulating the chemical receptors in the larynx and it causes a decrease in SPO₂ [20, 32]. Therefore, aspiration of a solid bolus might be more likely to result in airway occlusion. However, some researchers theoretically argue that a decrease in SPO₂ due to the presence of a bolus at the laryngeal level does not make sense, as airway obstruction would occur at the level of the lower respiratory system or lungs [25]. In our study, the frequency of aspiration with a decrease in SPO₂ and the accompaniment of aspiration with a decrease in SPO₂ displayed different patterns in IDDSI-0 and IDDSI-4 consistencies. Thin liquids such as water are known to pose a safety issue for people with dysphagia due to their rapid flow [33, 34]. In line with this, we found more aspiration in IDDSI-0 than in IDDSI-4 consistency, and more than half of those were accompanied by oxygen desaturation. Nonetheless, it should be noted that despite the absence of aspiration, oxygen desaturation might still be observed with IDDSI-4 consistency. Conversely, there might not be observed oxygen desaturation despite the presence of aspiration. This discrepancy suggests a decrease in the sensitivity of pulse oximetry especially when assessing semi-solid consistency. This finding suggests that the compatibility of aspiration and desaturation varies depending on the consistency. Butler et al. also indicated that the aspiration risk may vary according to consistency [35].

The sensitivity rate for IDDSI-4 consistency indicates that pulse oximetry did not exhibit the capacity to correctly identify true positive cases. The lack of sensitivity implies that pulse oximetry failed to detect dysphagia in IDDSI-4 consistency. However, the specificity value in IDDSI-4 consistency indicates the pulse oximetry’s ability to accurately true negative cases, implying its proficiency in identifying non-dysphagics. The low PPV further suggests that pulse oximetry was unable to reliably predict true positive cases, while the high NPV indicates that pulse oximetry is quite accurate in correctly identifying individuals without dysphagia. The AUC value suggests a limited ability of the pulse oximetry to distinguish between positive and negative cases. The findings raise concerns about the effectiveness of pulse oximetry in accurately diagnosing individuals with dysphagia in semi-solid consistency.

The good sensitivity rate in IDDSI-0 consistency indicates that pulse oximetry can effectively capture a substantial portion of true positive cases. Moreover, the high specificity in IDDSI-0 consistency underscores the pulse oximetry’s ability to accurately identify true negative cases. Furthermore, the AUC value suggests that pulse oximetry has good discriminatory power in IDDSI-0 consistency, allowing it to effectively distinguish between cases with and without dysphagia. The combination of sensitivity, specificity, and NPV, along with the AUC value within the context of IDDSI-0 consistency, suggests a potential diagnostic capability of pulse oximetry in detecting aspiration during liquid swallowing.

Our study measured oxygen desaturation by finger pulse oximetry at three different time points: before, during, and after swallowing. We found no difference between the patient and healthy control group in SPO₂ values before-swallowing in both consistencies, but we found a significant difference at the time points including during- and after-swallowing. However, while this difference was statistically significant, the difference between the SPO₂ values measured at the time points before, during and after swallowing in the patient group was less than 1%.

One of the crucial mechanisms to protect the airway is the coordination between respiration and swallowing [36]. If the temporal coupling between respiration and swallowing is disrupted, it can cause the prolongation of swallowing latency leading to a delay in the timing of the swallowing. Patients with dysphagia may be more likely to develop aspiration pneumonia if there is a higher degree of discoordination between their respiration and swallowing cycle [37].

Park et al. indicated that stroke patients with dysphagia had a significantly longer delay in the laryngeal closure [38]. The fact that the time point after-swallowing was statistically more significant in terms of SPO₂ values than the time point during-swallowing might be due to the temporal delay in the detection of oxygen desaturation in finger pulse oximetry due to airway obstruction. In contrast, Huang et al. found that laryngeal closure duration is shorter in post-stroke dysphagic patients than the healthy controls and the delay increases the risk of penetration and aspiration [39]. When the time points were analyzed according to the consistencies, a statistically significant difference was found between dysphagic patients and healthy controls in SPO₂ values both after- and during-swallowing in IDDSI-0 consistency compared to IDDSI-4 consistency, respectively. Hiss et al. found differences in the timing of the onset of respiratory pause because of bolus consistency [40]. The authors explained this finding as being related to the flow rate of the consistencies and that physiological events, including the onset of respiratory pause, may occur more slowly in thick consistencies than in thin consistencies. In our study, the statistical difference in the time points in IDDSI-0 consistency compared to the time points in IDDSI-4 consistency may be due to the flow rate of the consistencies. However, there is an ongoing debate regarding the impact of bolus on the timing and duration of respiratory pauses [41].

We endeavor to control the potential confounding factors mentioned in the literature on whether oxygen desaturation can be used to determine dysphagia. In our study, we limited the age range and included patients with the same disease etiology such as stroke, and compared the data obtained with a comparison group of normal healthy subjects. We determined the effect of variables such as both aspiration severity and food consistency on oxygen desaturation. We used a standardized method such as IDDSI to determine food consistency. In addition, we used VFSS and PAS during swallowing and collected the data blindly. To prevent artifacts, we controlled parameters such as room temperature, the presence of nail polish, the presence of a tattoo, keeping the arm still to which, the oximeter was connected, cleaning the finger, and attaching the probe to the finger on the non-hemiplegic side. One of our limitations is that we could not assess the pulmonary function of dysphagic patients before data collection, although we excluded patients with respiratory or cardiac functions. Another limitation is that we relied on the participants' self-report of whether they had dysphagia while assigning them to the healthy control group. One of the limitations of this study, we did not include non-dysphagic post-stroke patients. Comparing three groups could have more comprehensive understanding of the discriminative ability, reliability, and specificity, and it could provide insights into the generalizability of the findings. The other limitation of our study was the inadequate number of dysphagic patients assigned all three different Penetration-Aspiration Scale (PAS) levels. Although we did not observe a statistically significant decline in oxygen saturation levels before, during, and after swallowing across the three PAS levels, we acknowledge the need for future studies to evaluate comprehensively SPO₂ levels in groups with varying aspiration severity based on PAS ratings. Particularly in silent aspiration, a decline in SPO₂ might be a crucial marker. Therefore, we might expect a higher decrease in SPO₂ in PAS 8. However, when the mean values of the SPO₂ levels of the patients in the PAS 8 group were analyzed; the decrease between before and after time points was found to be 1.5% and remained below the 2% change. Despite the limited number of patients with aspiration, the present results indicate that pulse oximetry alone may not be sufficiently effective in detecting aspiration.

Furthermore, we believe that examining SPO₂ levels concerning aspiration awareness levels would provide valuable insights. Additionally, incorporating a multimodal sensory assessment to identify the presence of hypesthesia and exploring its association with aspiration severity and the degree of SPO₂ decline might further enhance our understanding of the field.

In further studies, more reliable results can be obtained by using measurement of multimodal physiological signals with devices such as nasal airflow and nasal temperature sensors, plethysmography, electromyography, electrocardiogram, and skin conductance meter in addition to pulse oximetry for monitoring swallowing and respiratory cycle to detect dysphagia. In addition, complementary screening techniques such as measurement of laryngeal elevation and/or cervical auscultation may be included in the assessment process. These methods offer the advantages of being relatively low-cost and easily accessible.

The findings of this study highlight the importance of considering the consistency of ingested materials when utilizing pulse oximetry as a diagnostic tool for detecting aspiration in dysphagic patients. Further research and evaluation of pulse oximetry’s performance in larger patient populations are needed to validate its effectiveness in aspiration detection for both consistencies. Additionally, it is recommended to consider SPO₂ decrease exceeding 2% as a criterion for improved accuracy.

A decrease in SPO₂ may hold the potential to predict aspiration, contingent upon the consistency being considered. However, detecting aspiration alone is not sufficient. This study suggested that pulse oximetry could serve complementary tool in assessing individuals with dysphagia. However, aspiration occurring on VFSS cannot be predicted based on decrease in SpO₂ in pulse oximetry.

Taken together, oxygen saturation monitorization during and after prandial swallowing may have the potential to improve complementary assessments such as bedside swallowing assessment and provide a more comprehensive approach to aspiration detection. However, further research is needed to confirm these findings.