Patients with an impaired reflexive cough response are at increased risk of pneumonia. This study examined the correlation between cough sensitivity to citric acid and radiographic features of airway compromise. Eighty patients referred for a radiographic assessment of swallowing at an acute hospital over an 8-month period participated in the study. Nebulised citric acid diluted in 0.9% sodium chloride was inhaled through a facemask at four concentrations to assess cough sensitivity. These data were then compared to Penetration Aspiration Scale scores based on radiographic swallowing studies. There was a statistically significant correlation between cough response/lack of response and the radiographic features of airway compromise; that is, patients who had a weak or absent response to inhalation of citric acid were also likely to aspirate silently during radiographic assessment. Sensitivity for identifying absent cough was found to be high at all 4 concentrations (0.750, 0.833, 0.941, 1.000), however specificity was consistently quite low (0.344, 0.456, 0.238, 0.078). The significant findings of this research suggest that clinicians adopting cough reflex testing into their clinical practice will have a reliable screen for silent aspiration at bedside. Clinicians will be able to identify patients who require instrumental assessment and are at high risk of pneumonia. This will likely, in turn, decrease length and cost of hospital admissions as well as decrease aspiration pneumonia related morbidities.

Cough reflex testing (CRT) is used to assess the sensory and motor components of a reflexive cough. When used as an adjunct to the clinical swallowing evaluation, it has the potential to identify individuals who are more likely to aspirate silently in the event of aspiration. It is unknown how reliable clinicians are at interpreting cough responses and the factors that influence this reliability. Therefore, this study ascertained the reliability of Speech Language Therapists in interpreting cough responses in CRT and determined how sensory perception and training influence reliability. Additionally, the study determined a relative measure of CRT interpretation validity through a comparison with ‘expert’ consensus ratings. A total of 111 clinicians completed an online survey consisting of ratings of 30 audio-, visual- and audiovisual clips over three sessions, and a final questionnaire. Reliability was substantial (κ=.76) in rating for presence of cough but only fair (κ=.25) in rating for strength of cough. Clinicians used a variety of auditory and visual cues to assist in decision making. Training did not result in higher reliability. Ratings of individual clinicians were significantly associated with consensus ratings for presence of cough (p < 0.001) and strength of cough (p < 0.05). In summary, clinicians could reliably and accurately judge presence of cough response, but were poorer at rating the strength of cough.

Silent aspiration is associated with pneumonia and mortality, and is poorly identified by traditional clinical swallowing evaluation (CSE). Currently, there is no reliable test for detecting silent aspiration during CSE. There is, however, increasing evidence for the validity of cough reflex testing (CRT) for identifying silent aspiration. This test has the potential to significantly improve clinical assessment of dysphagia. The aim of this research programme was to further investigate the validity, reliability and clinical utility of CRT for identifying patients at risk of silently aspirating. Several aspects of CRT were explored during this research programme. Two correlational studies were conducted to validate CRT for identifying silent aspiration against videofluoroscopic swallowing study (VFSS) and flexible endoscopic evaluation of swallowing (FEES). Cough reflex threshold testing was completed on 181 patients using inhaled, nebulised citric acid. Within one hour, 80 patients underwent VFSS and 101 patients underwent FEES. All tests were recorded and analysed by two researchers blind to the result of the alternate test. Significant associations between CRT result and cough response to aspiration on VFSS (p = .003) and FEES (p < .001) were identified. Sensitivity and specificity were optimised at 0.6mol/L in patients undergoing VFSS (71%, 60% respectively) and at 0.4mol/L in patients undergoing FEES (69%, 71% respectively). A concentration of 0.8mol/L had the highest odds ratio (OR) for detecting silent aspiration (8 based on VFSS, 7 based on FEES). Coughing on lower concentrations of citric acid (0.4mol/L compared with 1.2mol/L) was a better predictive measure of silent aspiration. Diminished cough strength has also been associated with aspiration and increased risk of pneumonia. Reflexive cough is our primary defensive mechanism against aspiration and a measure of reflexive cough strength therefore holds greater relevance than one of voluntary cough strength. Despite common use and clinical applicability, the reliability of subjective cough judgements has received little attention. The inter- and intra-rater reliability of subjective judgements of cough in patients following inhalation of citric acid was assessed. Forty-five speech-language therapists (SLTs) were recruited to the first study. Of these, 11 SLTs were currently using CRT in their clinical practice (experienced raters) and 34 SLTs reported no experience with CRT (inexperienced raters). Participants provided a rating of strong, weak or absent to ten video segments of cough responses elicited by inhalation of nebulised citric acid. The same video segments presented in a different sequence were re- evaluated by the same clinicians following a 15-minute break. Inter-rater reliability for experienced raters was calculated with a Fleiss’ generalised kappa of .49; intra-rater reliability was higher with a kappa of .70. Inexperienced raters showed similar reliability with kappa values for inter-rater and intra-rater reliability of .36 and .62, respectively. SLTs demonstrated only fair to moderate reliability in subjectively judging a patient’s cough response to citric acid. Experience in making cough judgements did not improve reliability significantly. In a second study, specific training in cough physiology and cough judgement was provided to 58 trained SLTs. Inter-rater reliability of subjective judgements of cough in patients following inhalation of citric acid was assessed. Participants provided a rating of present or absent, and if present then a rating of strong or weak, to ten video segments of cough responses. Inter-rater reliability for cough presence was calculated with a Fleiss’ generalised kappa of .71 and cough strength was calculated at .61. Years of clinical experience did not improve inter-rater reliability significantly. Experience in using CRT did improve inter-rater reliability. Further validity and reliability research would be beneficial for guiding clinical guidelines and training programmes. By identifying patients at risk of silent aspiration, more informed management decisions can be made that consequently lead to a reduction in preventable secondary complications such as pneumonia. The clinical utility of CRT for reducing pneumonia in acute stroke patients was assessed through a randomised, controlled trial. Three hundred and eleven patients referred for swallowing evaluation were assigned to either 1) a control group receiving standard evaluation or 2) an experimental group receiving standard evaluation with CRT. Participants in the experimental group were administered nebulised citric acid with test results contributing to clinical decisions. Outcomes for both groups were measured by pneumonia rates at three months post stroke and other clinical indices of swallowing management. Analysis of the data identified no significant differences between groups in pneumonia rate (p = .38) or mortality (p = .15). Results of CRT were shown to influence diet recommendations (p < .0001) and referrals for instrumental assessment (p <.0001). Despite differences in clinical management between groups, the end goal of reducing pneumonia in post stroke dysphagia was not achieved. Through this research, the characteristics and outcomes associated with dysphagia secondary to stroke in New Zealand were identified. Baseline characteristics of 311 patients with dysphagia following acute stroke were collected during their hospital stay and outcomes were measured at three months post stroke. Mortality rates were 16% and pneumonia rates 27%. Mean length of stay was 24 days and only 45% of patients were in their own home at three months post stroke. Pneumonia was significantly associated with mortality and increased length of stay. Only 13% of patients received referral for instrumental assessment of swallowing. These data are discussed in reference to the National Acute Stroke Services Audit 2009 and internationally published data. The outcomes for stroke patients with dysphagia in New Zealand are poor with a high risk of pneumonia and long hospital stays when compared internationally. In summary, this research programme has contributed to our understanding of the use of CRT in patients with dysphagia. The addition of a measure of reflexive cough strength may add to clinical assessment but specific training is required to reach adequate reliability. CRT results are significantly associated with aspiration response on instrumental assessment and lower concentrations of citric acid provide a better predictive measure of silent aspiration. CRT can be standardised and therefore is not as susceptible to interpretative variance that plagues much of CSE. Sensitivity and specificity values using this CRT methodology are adequate for CRT to be incorporated into clinical protocols. Inclusion of CRT alone was not shown to be sufficient to change clinical outcomes however integration of CRT into clinical pathways may prove more successful. Further research evaluating the addition of CRT to a comprehensive CSE would add greatly to the field of dysphagia assessment.