Relationship between post-stroke dysphagia and pharyngeal sensory impairment

Dysphagia is a common symptom after stroke and occurs in up to 80% of patients in the acute phase of the disease [1]. In addition to affecting quality of life [2] and placing a significant financial burden on the health care system [3] post-stroke dysphagia (PSD) causes serious complications such as malnutrition [4] and pneumonia [5], resulting in increased mortality [6].

Instrumental procedures such as the Videofluoroscopic swallowing study (VFSS) or Flexible Endoscopic Evaluation of Swallowing (FEES) have been developed to visualize swallowing. FEES and VFSS have comparable diagnostic metrics in the detection of relevant dysphagia pathologies and are therefore both considered as diagnostic gold standards [7]. In addition to reliable detection of PSD, instrumental procedures also allow characterization of the dysphagia phenotype, i.e., the pattern of dysphagia impairment [8]. However, the use in clinical routine is determined by methodological advantages and disadvantages of both modalities. In stroke patients, FEES has the distinct advantage that the examination can be performed at the bedside even in patients with limited ability to cooperate. Further, secretion management can be assessed.

In recent decades, an extensive cortical and subcortical network involved in the central control of swallowing has increasingly come into scientific focus. Studies using voxel-based lesion symptom mapping have demonstrated that post-central lesions, i.e. the primary sensory cortex, in particular are associated with severe forms of swallowing impairment [9]. In addition, individual studies have shown a correlation between sensory pharyngeal dysfunction and aspiration [10] or PSD severity [11]. This points to a specific role of the central sensory system in the pathophysiology of PSD and suggests that the sensory system exerts a secondary modulatory effect on swallowing motor function.

In clinical practice, it is difficult to determine the extent of pharyngeal sensory dysfunction. Sensory testing is mostly conducted, if at all, using qualitative or (semi-)quantitative methods such as the FEES-based touch-technique [12]. This involves touching structures of the pharynx and larynx with the endoscope and subjectively evaluating the patient’s response. However, a quantitative tool would be favorable for accurate assessment in disease progression and for elaborate analysis in scientific studies. Therefore, a FEES-guided swallowing provocation test was developed previously, in which different volumes of liquid are administered via a tube placed in the upper third of the oropharynx and the latency of swallowing response (LSR) is determined (FEES-LSR-Test). This test was validated in healthy subjects and clearly distinguished between the physiological state and experimentally induced pharyngeal anesthesia [1]. Further, prolonged LSR was the only clinical determinant of swallowing alterations in terms of presbyphagia in a cohort of healthy subjects older than 70 years [13]. However, the FEES-LSR-Test has not yet been used and validated in stroke patients.

The aims of this study were therefore (1) to evaluate the relationship between sensory impairment and dysphagia severity in acute stroke patients, (2) to investigate which dysphagia pathologies and phenotypes are linked to sensory impairment, (3) To verify the applicability/suitability of the FEES-LSR-Test in acute stroke patients compared the conventional touch-technique, and (4) to determine the ideal test volume in the LSR procedure to recommend an abbreviated but equally valid protocol in clinical practice. To this end, a prospective FEES study was conducted in a cohort of patients with acute stroke that included sensory evaluation using the touch-technique and the FEES-LSR-Test.

The main finding of our study was that pharyngeal hypesthesia is an independent predictor of the severity of PSD and impaired secretion management. In line with this finding, previous studies have shown an association between sensory deficits and dysphagia as well [10, 11, 19, 20, 21]. In contrast to these previous studies, the present study has now succeeded in demonstrating that this association persisted after accounting for other relevant cofactors such as stroke severity. Thus, the results suggest that pharyngeal hypesthesia is a relevant factor in the development of PSD.

With regards to patterns of swallowing impairment, we found an association between pharyngeal hypesthesia and a delayed or absent swallowing reflex. However, there was no association with the extent of pharyngeal residue or premature bolus spillage. This suggests that the crucial PSD mechanism caused by sensory impairment is mediated by an absent or delayed swallowing reflex. Evidence for other secondary motor modulatory effects induced by pharyngeal hypesthesia leading to premature bolus spillage or pharyngeal residue was not found in our study. In contrast, Sulica et al. reported an increase in premature bolus spillage and pharyngeal residue caused by pharyngeal anesthesia in healthy participants [22]. However, in stroke patients decreased pharyngeal contractility, attenuated tongue retraction, hypercontractility of the upper esophageal sphincter, bulbar and pseudobulbar oral paresis, and deficits in higher cortical functions may play a more vital role in the development of pharyngeal residue and premature bolus spillage. A dysfunctional swallowing reflex seems to affect not only the severity of dysphagia in general, but also the management of secretions in particular. One plausible reason for this could be that the pooling of secretions due to pharyngeal hypesthesia is not sufficiently perceived by the patient and thus does not trigger swallowing. This may result in a risk of saliva aspiration.

The underlying mechanisms of pharyngeal hypesthesia in stroke seem to be heterogeneous and are not well understood in detail. On the one hand, central mechanisms of sensory impairment are a major contributing factor. This is illustrated by the association of particularly severe dysphagia with lesions in the post-central cortex as the primary sensory area [9]. Electrophysiological study results also provided evidence of central hypesthesia in PSD. In contrast to stroke patients without dysphagia, patients with PSD showed reduced ipsilesional activation of sensory evoked potentials to pharyngeal stimulation [23]. Also, abnormal asymmetric sensory evoked potentials were described in the majority of patients during pharyngeal electrical stimulation [24]. Furthermore, contralesionally decreased pharyngeal sensory evoked potentials were associated with increased duration of swallowing [25]. In addition to these central effects, peripheral mechanisms of pharyngeal hypesthesia have also been postulated. This includes substance P, a neuropeptide that is released into saliva by sensory nerve endings. It is assumed that the degeneration of these peripheral sensory nerves results in a decrease of substance P, which mediates dysphagia [26]. In stroke patients, a decreased salivary substance P level was found to be a predictor of reduced swallowing frequency, independent of age, stroke severity, and vigilance. In addition, a low substance P level was associated with a higher rate of pneumonia [27]. Possible reasons for peripheral deficits may include mucosal damage, e.g., due to interventions such as intubation or tube placement. In summary, studies suggest that perception and integration of sensory information is impaired at both central and peripheral levels, although the complex interactions are not yet understood in detail.

Therapeutic approaches targeting the sensory system with neurostimulation have shown promising study results. One method is pharyngeal electrical stimulation (PES), in which a sensory stimulus is used with the aim to trigger central neuroplasticity. In a multicenter randomized controlled trial, pharyngeal electrical stimulation increased decannulation rates in stroke patients by improving PSD and secretion management [28]. In proof-of-concept neuroimaging studies in healthy subjects, PES modulated the organization of the cortical swallowing network [9, 29] and led to increased sensorimotor activation during swallowing after pharyngeal anesthesia [30]. In addition, PES was shown to increase substance P level in saliva in tracheotomized stroke patients, which was a predictor of decannulation success [31]. This suggests that in addition to the central triggering of neuroplasticity, peripheral sensory mechanisms may also contribute to the therapeutic effect of PES. Furthermore, pharmaceutical sensory stimulants such as capsaicin are also used to trigger substance P release. In two recent randomized controlled trials in PSD patients, swallowing function evaluated by a water swallow test was improved [32, 33]. In addition, data from another randomized trial with a cross-over design suggest that capsaicin leads to a short-term increase in motor cortex excitability in stroke patients [23]. This again highlights the complex central and peripheral interactions in the sensory system. In contrast, a study in healthy participants suggested that capsaicin has no central modulatory effect on the swallowing network [34].

In addition to the association between PSD and sensory impairment in general, this study also compared two different assessment methods of pharyngeal hypesthesia. Our results indicate, that in principle both the touch-technique and the FEES-LSR-Test are suitable for the detection of pharyngeal hypesthesia. The advantage of the LSR approach when used in studies is that a metric result is obtained, allowing for more sophisticated analyses than with a binary or ordinal result as in the touch-technique. As in the validation study with healthy subjects [18], our study also showed a decrease in LSR with increasing trigger volume, however as expected with longer LSR compared to the healthy subjects. This shows that an increased sensory stimulus (larger trigger volume) is associated with a decreased LSR and thus suggests that LSR is a suitable marker of pharyngeal sensation in this patient population. When considering the different trigger volumes, the correlation with the touch-technique was found only with 0.3 ml and 0.4 ml trigger volumes. This suggests that pharyngeal hypesthesia in PSD can be quantified by LSR with these volumes, in contrast to the volumes of 0.2 ml and 0.5 ml. At 0.2 ml, the trigger may be too weak, so that a delayed reaction may occur even in the presence of clinically irrelevant hypesthesia. Conversely, at 0.5 ml, the trigger may be too strong so that a timely reaction may be elicited even in the presence of relevant sensory impairment. In the regression analyses, 0.4 mL showed an association with several parameters studied (severity of dysphagia, Murray secretion scale impairment, and with absent or delayed swallowing reflex). Therefore, 0.4 mL is particularly well suited to distinguish between the physiologic state and hypesthesia. Consistent with this, both the validation study and a study of healthy subjects over 70 years of age showed the largest effect size or differences in LSR at 0.4 ml to distinguish pharyngeal anesthesia and presbyphagia, respectively, from the physiologic state [3, 18]. This is further corroborated by a study in stroke patients, in which 0.4 ml trigger volume was superior to 2 ml in detecting aspirations using a swallow provocation test [35].

Another quantitative method that has been proposed to determine pharyngeal sensitivity is the so-called air-pulse method, or Flexible Endoscopic Evaluation of Swallowing with Sensory Testing (FEESST). Here, an air-pulse is applied to the anterior pharyngeal wall via a second endoscopic working channel and the pressure threshold of the air pulse is determined at which the laryngeal adductor reflex is triggered [36, 37]. However, the clinical relevance and reliability of this method is controversial. Some authors report that the results of the air-pulse method, unlike the touch-technique, cannot be associated with penetration and aspiration [12], or that the interobserver agreement is low [38]. One possible source of error is the difficulty in achieving a constant distance to the pharyngeal wall with the endoscope [12]. This problem is countered in swallow provocation tests by using different, easy-to-dose amounts as triggers (with the hypothesis that this makes the exact location of application less relevant). However, in comparison to the FEES-LSR-Test, the touch-technique in our study demonstrated to be equally valid in detecting deficits in pharyngeal sensitivity. Therefore, this method can be considered a useful and straightforward approach for clinical practice.

There are different limitations that must be considered when interpreting the results of this study. On the one hand, both test methods used rely on a motor response and cannot conclusively differentiate between an afferent and an efferent problem. Electrophysiological methods such as sensory evoked potentials are required for this purpose. When evaluating the results of the touch-technique the raters were not blinded to the FEES-results. Furthermore, individual regression analyses did not fulfill the condition of proportional odds according to the test of parallel lines, so that in these cases the results must be interpreted with caution. In addition, many patients who did not wish to participate or for whom it was not possible to inform their caregivers in time were not included in the study. Thus, a selection bias may have occurred and generalisability to other cohorts may be limited. In addition, the cross-sectional design and lack of follow-up do not allow conclusions about the temporal dynamics of reported outcomes.

Pharyngeal hypesthesia is an independent predictor of PSD severity as well as impaired secretion management. PSD caused by sensory impairment is mediated by an absent or attenuated swallowing reflex. Other PSD pathologies such as pharyngeal residue or premature bolus spillage do not appear to be related to pharyngeal hypesthesia. Both, the swallow provocation test via the LSR as a quantitative method and the touch-technique with binary or ordinal results, respectively, are suitable as assessment methods of pharyngeal hypesthesia. In the former method, 0.3 ml or especially 0.4 ml is suitable as a trigger volume.