Profiling the Swallowing Safety and Efficiency of Patients Presenting With Congestive Heart Failure

ABSTRACT

Purpose

Existing literature on dysphagia in patients with congestive heart failure (CHF) largely relies on patient‐reported outcomes with limited instrumental data. This study aimed to objectively profile swallowing function in CHF patients using videofluroscopic swallow studies (VFSS) analyzed with the ASPEKT (Analysis of Swallowing Physiology: Events, Kinematics, and Timing) method.

Methods

We conducted a retrospective study of patients admitted with CHF at St. Michael's Hospital (2021–2023). VFSS of liquid boluses was rated with penetration‐aspiration scale (PAS) and ASPEKT pixel‐based measures by blinded raters. Logistic regression models identified predictors of swallowing safety and efficiency.

Results

Twenty‐five patients met inclusion criteria. Atypical swallowing safety was observed in 67% of patients with HFpEF and in 90% of patients with HFrEF (p = 0.39). Incomplete laryngeal vestibule closure (OR = 23.75, 95% CI [7.342, 76.829]) and thin liquids (OR = 5.578, 95% CI [1.477, 21.072]) were significantly associated with unsafe swallows, while a greater number of swallows was protective (OR = 1.657, 95% CI [1.085, 2.532]). Larger pharyngeal areas at maximum constriction were associated with increased residue (OR = 76.136, 95% CI [23.055, 251.421]), whereas thin (OR = 0.494, 95% CI [0.376, 0.649]) and mildly thick liquids (OR = 0.612, 95% CI [0.459, 0.818]) with reduced residue.

Conclusion

CHF may impair swallowing safety and efficiency, potentially through mechanisms similar to other pulmonary conditions. Prospective studies incorporating respiratory monitoring and stratifying by CHF severity are needed to clarify physiologic pathways and clinical risks.

1. Introduction

Congestive heart failure (CHF) is a highly prevalent condition in North America and remains a leading cause of hospitalization among older adults [1]. This condition is frequently attributed to myocardial dysfunction, either through weakened systolic contraction or impaired diastolic relaxation, which ultimately compromises the heart's ability to maintain sufficient cardiac output to meet the metabolic demands of the body [2]. As a result, systemic oxygen delivery is reduced, and elevated pressures within pulmonary circulation can lead to fluid accumulation in the lungs [3]. This fluid overload, known as pulmonary congestion, disrupts normal gas exchange and contributes to clinical presentations such as dyspnea, tachypnea, hypoxemia, and, in more severe cases, pulmonary edema [4].

Normal swallowing is closely coordinated with breathing to protect the airway. It involves a brief respiratory pause known as swallow apnea, typically during mid‐exhalation [5]. In healthy individuals, the expiratory–swallow–expiratory pattern ensures that the airway is closed during the swallow and followed by exhalation, helping to clear any residual material and reduce aspiration risk [6]. Adequate lung volume during the swallow response supports this protective mechanism by allowing a stable respiratory pattern and sufficient expiratory flow [7]. However, in individuals with CHF, the respiratory impairments such as dyspnea and tachypnea can reduce lung volumes and disrupt normal timing [2]. Shallow and rapid breathing may lead to insufficient lung inflation, shortening of swallow apnea, or cause swallows to occur during inhalation rather than exhalation, thereby increasing the risk of laryngeal penetration and aspiration [8].

Previous studies have shown that dysphagia is prevalent among older adults with heart failure [9, 10, 11, 12]. A prospective cohort study utilizing functional oral intake scale (FOIS) scores and clinical swallowing assessments performed by trained therapists within 72 h of admission demonstrated that 46.2% of hospitalized older adults with heart failure developed dysphagia, which was independently associated with slower functional recovery and increased 1‐year mortality [10]. In another study, researchers used the FOIS to assess swallowing in patients with heart failure and designated a FOIS score of 5 or below, indicating the need for a modified diet or tube feeding, as a study‐specific marker of clinically significant swallowing difficulty [12]. Patients who met this criterion experienced longer hospital stays and were more likely to be discharged to care facilities rather than returning to their prior residence. In an earlier investigation, swallowing in patients with acute heart failure was assessed using the FOIS and bedside screening tools, including the repetitive saliva swallowing test and the water swallowing test, with 36.1% of patients found to have dysphagia based on a FOIS score ≤ 5, and 65% identified as being at nutritional risk [11].

Despite growing recognition that many patients with CHF experience swallowing difficulties, few studies have employed instrumental evaluations to objectively assess swallowing function in this population [9, 13]. The videofluoroscopic swallow study (VFSS) remains the gold standard for evaluating the physiology and safety of swallowing. To address this gap, the present study uses the ASPEKT (Analysis of Swallowing Physiology: Events, Kinematics, and Timing) method to objectively characterize swallowing physiology on VFSS and identify patterns of impairment in individuals with CHF [14, 15]. We hypothesize that patients with CHF will exhibit a pattern of biomechanical impairments that contribute to unsafe swallowing. Given that respiratory dysfunction can disrupt the coordination between breathing and swallowing, individuals with CHF may be at increased risk for airway invasion during swallowing. Therefore, we hypothesize that impairments associated with respiratory compromise secondary to CHF will be present in this sample [7].

2. Methods

2.1. Research Participants and Imaging Procedures

Ethics approval was obtained from the St. Michael's Hospital Research Ethics Board (REB 23‐220), and all procedures were conducted according to the Declaration of Helsinki. A retrospective chart review was conducted to extract records of all patients admitted for congestive heart failure at the St. Michael's Hospital Cardiology Unit from January 2021 to December 2023. A waiver of consent was provided by the REB; therefore, informed consent was not collected from participants given the nature of the study. Inclusion criteria were (a) Diagnosis of congestive heart failure noted in clinical reports, (b) Videofluoroscopic swallowing study conducted between 01/01/2021 and 12/30/2023 at St. Michael's Hospital. Exclusion criteria included all patients who underwent any surgical treatment (cardiothoracic surgery, coronary artery bypass graft, mini‐mitral, etc.), presented with neurological‐related incidents (e.g., stroke, TBI) within 5 years prior to CHF‐related admission, and the standard of care VFSS, and/or head and neck related malignancies.

For patients who met the inclusion criteria, the following variables were extracted: New York Heart Association (NYHA) classification and CHF Type (HFpEF, HFrEF), sex, age, number of hospital admissions, and date of videofluoroscopy (VFSS). Additionally, the standard of care videofluoroscopy exam images were located and extracted. Standard of care VFSS imaging at our institution was captured during hospital admission using a fluoroscopic C‐arm unit and a TIMS DICOM recording system at 30 images per second. Boluses presented during these exams were standardized using institutionally approved recipes mixed with EZ‐HD and Polibar (Bracco) according to a barium calculator that matched consistencies with International dysphagia diet standardization initiative (IDDSI) levels ranging from thin to extremely thick [16]. Only liquid swallows were used in this study and included bolus volumes of 5 mL (teaspoon), single cup sip, or continuous (sequential) cup sips.

2.2. Clinical Classification of CHF

CHF severity was characterized using two common classification methods: the New York Heart Association (NYHA) functional classification and ejection fraction‐based subtypes. The NYHA classification ranges from Class I (no symptoms with ordinary activity) to Class IV (symptoms at rest), reflecting functional limitations and disease progression [17]. The ejection fraction, defined as the percentage of blood ejected from the left ventricle with each heartbeat, was used to identify the type of heart failure. Heart failure with reduced ejection fraction (HFrEF) was defined as a left ventricular ejection fraction of 40% or less, consistent with impaired pumping function. Heart failure with preserved ejection fraction (HFpEF) was defined as an ejection fraction of 50% or greater and is typically linked to stiffening of the heart muscle and impaired filling during relaxation [4].

2.3. Swallowing Measurement Procedure

Trained raters performed independent and blinded ratings to derive PAS scores, timing of airway invasion relative to the swallow, along with all ASPEKT method parameters for timing and pixel‐based parameters [14]. A subset of the entire dataset (20%) was rated in duplicate by a second rater for reliability.

2.4. Statistical Analyses

Descriptive statistics were used to summarize participant demographics and PAS scores. Existing normative frequency classifications were derived and utilized for analyses based on the healthy reference values and thresholds reported by Steele and colleagues [18]. Univariate analysis was performed using swallow safety and swallow efficiency as outcome measures to identify significant variables to be included as predictor variables in logistic regression models. A logistic regression model was fitted to the data to explore the components contributing to swallowing safety. To account for multiple swallows included per individual patient, random intercepts were included according to patient identity. The regression was fitted as a function of the following independent variables: bolus consistency, number of swallows, degree of LVC, and duration of LVC (frames). A mixed‐effects beta regression model was fitted to the data to explore the component contributing to swallowing efficiency. Again, random intercepts were included according to patient identity to account for multiple swallows. The regression was fitted as a function of the following independent variables: bolus consistency, bolus volume, pharyngeal area at maximum pharyngeal constriction (%C2–C4), and pharyngeal area at rest (%C2–C4). All data analysis was completed in the R statistical computing environment [18].

3. Results

A total of 25 patients met the inclusionary criteria, and 169 bolus level clips were analyzed in this cohort. Demographic and heart failure classification information is outlined in Table 1. Swallows were grouped based on their IDDSI levels [thin (43%), mildly thick (24%), moderately thick (14%), and extremely thick (20%)] and bolus volumes. The number of swallows, degree of LVC closure, and pharyngeal area at maximum pharyngeal constriction are detailed in Table 2. The number of typical and atypical parameters per IDDSI level and volume is also outlined in Table 2.

TABLE 1.

Demographics and heart failure characteristics.

Characteristic	N (%)
Age, mean (sd.)	80.2 ± 7
Sex
Male	15 (60)
Female	10 (40)
CHF type
HFpEF	15 (60)
HFrEF	10 (40)
NYHA
1	4 (16)
2	4 (16)
3	14 (56)
4	3 (12)
BMI	28.82 ± 6.8

TABLE 2.

Number of swallows, LVC closure degree, and PhAMPC stratified by IDDSI type.

Bolus trial (IDDSI)	Number of swallows				LVC closure degree			Pharyngeal area at maximum constriction (%C2–C4)
	Mean ± SD	n	Typical, n (%)	Atypical, n (%)	n	Typical, n (%)	Atypical, n (%)	Mean ± SD	n	Typical, n (%)	Atypical, n (%)
Thin ‐ IDDSI 0
Teaspoon	1.57 ± 0.9	28	18 (64)	10 (36)	28	18 (64)	10 (36)	8.61 ± 7.0	25	5 (20)	20 (80)
Cup sip	2.15 ± 1.3	40	15 (38)	25 (62)	40	13 (33)	27 (67)	13.7 ± 11.2	38	4 (11)	34 (89)
Consecutive cup sips	4 ± 1.2	4	0 (0)	4 (100)	4	2 (50)	2 (50)	27.3 ± 12.5	3	0 (0)	3 (100)
Mildly Thick ‐ IDDSI 2
Teaspoon	1.73 ± 1.3	15	9 (60)	6 (40)	15	5 (33)	10 (67)	14.0 ± 9.4	15	0 (0)	15 (100)
Cup sip	2.00 ± 1.0	24	10 (42)	14 (48)	24	8 (33)	16 (67)	16.2 ± 10.1	24	1 (4)	23 (96)
Consecutive cup sips	2.50 ± 0.7	2	0 (0)	2 (100)	2	0 (0)	2 (100)	14.6 ± 3.2	2	0 (0)	2 (100)
Moderately Thick ‐ IDDSI 3
Teaspoon	1.62 ± 0.7	13	6 (46)	7 (44)	13	4 (31)	9 (69)	19.9 ± 10	13	0 (0)	13 (100)
Cup sip	2.5 ± 1.5	10	2 (20)	8 (80)	10	2 (20)	8 (80)	21.0 ± 10	10	0 (0)	10 (100)
Extremely Thick ‐ IDDSI 4
Teaspoon	2.00 ± 1.0	33	11 (33)	22 (67)	33	16 (48)	15 (42)	17.8 ± 10.9	32	1 (3)	31 (97)

All swallows were classified based on the PAS scale score obtained (range from 1 to 8), with a categorization of typical scores (PAS 1‐2) compared to atypical scores (PAS 3‐8). The count of typical and atypical PAS scores is shown in Figure 1, with the majority of swallows demonstrating PAS scores ≥ 3. Of the 66 swallows that were classified as atypical, airway penetration was identified in 54 swallows (82%), and aspiration was identified in 12 swallows (18%).

FIGURE 1.

PAS score categorized by typical and atypical values compared to reference values.

Atypical swallowing safety was seen in 67% of patients with HFpEF and in 90% of patients with HFrEF (p = 0.39). When categorized by NYHA score, 75% of patients with less severe disease (scores 1 or 2) had atypical PAS scores, with airway invasion identified in 6 patients and aspiration identified in 2 patients. Among patients with more severe disease (scores 3 or 4), 76% had an atypical PAS score, with airway invasion identified in 12 patients and aspiration identified in 4 patients (p = 0.999). All aspiration events, regardless of heart failure or NYHA classification, occurred with a thin bolus consistency. Additionally, all patients, regardless of heart failure or NYHA classification, had atypical swallowing efficiency values as measured by total residue on VFSS.

Univariate analyses were performed to determine significant predictor variables of swallowing safety and swallowing efficiency. Bolus consistency, number of swallows, and degree of LVC were significantly associated with PAS score. Bolus consistency and pharyngeal area at maximum pharyngeal constriction were significantly associated with total residue. These variables, in addition to other physiologically relevant swallowing parameters, were included in further logistic regression models.

The results of the logistic regression model predicting swallowing safety are shown in Table 3. Incomplete laryngeal vestibule closure was a significant predictor for an atypical swallow (OR = 23.75, 95% CI [7.342, 76.829]). Consistency of the bolus was also significantly related to swallowing safety, with a significantly greater odds of atypical swallowing noted for thin liquids (OR = 5.578, 95% CI [1.477, 21.072]). A greater number of swallows appeared to be a protective measure against airway invasion (OR = 1.657, 95% CI [1.085, 2.532]). There was no significant association between atypical swallowing safety and duration of laryngeal vestibule closure.

TABLE 3.

Logistic regression model predicting swallowing safety.

Term	Estimate	OR	95% CI	p
Intercept	−4.718	0.009	[0.002, 0.054]	< 0.001*
Consistency: extremely thick	Ref.	Ref.	Ref.	Ref.
Consistency: thin	1.719	5.578	[1.477, 21.072]	0.011*
Consistency: mildly thick	0.731	2.078	[0.557, 7.755]	0.276
Consistency: moderately thick	0.835	2.305	[0.558, 9.514]	0.248
Number of swallows	0.505	1.657	[1.085, 2.532]	0.019*
Degree of LVC (complete)	Ref.	Ref.	Ref.	Ref.
Degree of LVC (incomplete)	3.168	23.75	[7.342, 76.829]	< 0.001*
Duration of LVC (frames)	−0.001	0.999	[0.986, 1.012]	0.904

^*Denotes a statistically significant predictor.

The results of the beta regression model predicting swallowing efficiency are shown in Table 4. Pharyngeal area at maximum constriction was a significant predictor of swallowing efficiency, with larger areas at maximum constriction corresponding to a less efficient swallow (OR = 76.136, 95% CI [23.055, 251.421]). Bolus consistency also significantly affected swallowing efficiency, with thin (OR = 0.494, 95% CI [0.376, 0.649]) and mildly thick (OR = 0.612, 95% CI [0.459, 0.818]) liquids associated with decreased total residue.

TABLE 4.

Logistic regression model predicting swallowing efficiency.

Term	Estimate	OR	95% CI	p
Intercept	−2.845	0.058	[0.036, 0.094]	< 0.001*
Consistency: extremely thick	Ref.	Ref.	Ref.	Ref.
Consistency: thin	−0.705	0.494	[0.376, 0.649]	< 0.001*
Consistency: mildly thick	−0.49	0.612	[0.459, 0.818]	< 0.001*
Consistency: moderately thick	−0.24	0.786	[0.591, 1.047]	0.1
Volume: teaspoon	Ref.	Ref.	Ref.	Ref.
Volume: cup sip	0.183	1.2	[0.949, 1.519]	0.128
Volume: consecutive	−0.117	0.89	[0.500, 1.585]	0.692
PhAMPC (%C2–C4)	4.333	76.136	[23.055, 251.421]	< 0.001*
PhAR (%C2–C4)	0.0489	1.05	[0.706, 1.562]	0.809

^*Denotes a statistically significant predictor.

4. Discussion

This study aimed to characterize swallowing physiology in patients with congestive heart failure (CHF) using instrumental videofluoroscopic swallow studies (VFSS) and the ASPEKT method, a standard operating procedure for quantifying swallowing physiology. Although existing studies have demonstrated a correlation between CHF and dysphagia, published work utilizes methods of assessment that are limited to evaluating oral intake without instrumental assessments (such as VFSS or FEES) that would allow for a detailed analysis of bolus movement through oral and pharyngeal structures. In this analysis, the hypothesis was that patients with CHF would exhibit impaired swallowing safety and efficiency due to the physiological connections between swallowing and breathing and the known respiratory dysfunction in this patient group [3]. The study's results support this hypothesis, where impairments in both swallowing functions were noted, particularly with atypical PAS scores (≥ 3) observed in the majority of participants, including those with milder CHF classifications. These findings suggest that swallowing impairment may be underrecognized in this population.

Incomplete laryngeal vestibule closure (LVC) was the strongest predictor of aspiration and penetration events, with an odds ratio of 23.75. This supports the notion that compromised airway closure is a critical physiological marker of unsafe swallowing, particularly in populations with impaired respiratory function. The high frequency of incomplete LVC across IDDSI levels may reflect a biomechanical vulnerability in this CHF cohort, though further research is needed to confirm this pattern. Additionally, while LVC duration did not emerge as a significant predictor, closure completeness appeared to be more clinically impactful to overall swallowing safety. Interestingly, the increased number of swallows observed was associated with unsafe swallows, which may instead reflect compensatory mechanisms in this patient group. Alternatively, increased swallows could reflect a more cautious, paced approach to eating in response to pathophysiological mechanisms.

Swallowing safety was most compromised with thin liquids in this cohort. Thin liquid swallows demand precise timing and robust airway protection, which may be compromised by CHF‐related respiratory impairments such as dyspnea and tachypnea. Dyspnea and tachypnea are known to reduce lung volumes and disrupt the preferred expiratory–swallow–expiratory pattern of coordination [7, 8]. This incoordination between the respiratory phase and swallowing events may predispose individuals to swallow during or immediately before inspiration, increasing the risk of bolus airway invasion. Our findings indicated that thin liquids were associated with over a five‐fold increased odds of unsafe swallowing, reinforcing the vulnerability of this group to airway invasion with fast‐flowing boluses. Similar vulnerabilities to thin liquids and other consistencies have been reported in other pulmonary populations. In COPD, thin liquid swallows were associated with delayed and shortened laryngeal vestibule closure as well as more frequent incomplete closure, reducing airway protection and increasing the risk of silent aspiration [19]. Additional evidence showed that under low lung volume conditions, thin liquid swallows in individuals with COPD resulted in prolonged swallow durations and disrupted respiratory–swallow timing [20]. Disrupted respiratory–swallow coordination was also observed in COPD, with a greater tendency to initiate swallows during inhalation and to inhale immediately after swallowing [21]. In post‐lung transplant patients, VFSS revealed that over 80% of patients exhibited airway invasion postoperatively, including 45% with aspiration and 39% with laryngeal penetration [22]. A separate prospective study using FEES found that 60% of lung transplant recipients demonstrated laryngeal penetration and 40% aspirated thin liquids, with 72% of patients showing no overt response, consistent with silent aspiration [23]. In a case series of patients with idiopathic pulmonary fibrosis (IPF), videofluoroscopic swallow studies showed that 30% exhibited laryngeal penetration and 10% aspirated thin liquids silently [24]. A separate study reported that 91.1% of patients with IPF demonstrated reduced swallowing speed for liquids, with these impairments significantly associated with lower pulmonary function and nutritional status [25].

Along with safety concerns, all participants demonstrated atypical swallowing efficiency compared to healthy normative values, with total pharyngeal residue observed regardless of CHF severity. Larger pharyngeal areas at maximum constriction (PhAMPC) were significantly associated with inefficient swallows, possibly indicating poor pharyngeal pressure generation and/or incomplete constrictor muscle contraction as a primary biomechanical contributor to residue accumulation in this cohort. In contrast, pharyngeal area at rest (PhAR) was not a significant predictor for efficiency concerns, indicating that the efficiency deficits stem more from impaired muscle contraction during the swallow compared to structural changes like generalized muscle changes associated with age‐related sarcopenia. Prior research has demonstrated similar coordination and pressure impairments in older adults, which may be compounded by chronic cardiopulmonary limitations in this group [26, 27]. Similar inefficient swallowing has been reported in patients with COPD and chronic respiratory conditions. Patients with these conditions are found to have increased pharyngeal residue and reduced pharyngeal constriction, likely attributable to disrupted breathing‐swallowing coordination [19, 21, 28, 29]. Furthermore, residue was less pronounced with IDDSI 0 (thin) and IDDSI 2 (mildly thick) consistencies–an expected finding based on previous literature. In both healthy patients and those with swallowing impairments, including pharyngeal weakness, computational modeling and clinical studies have shown that thick or highly viscous liquids are associated with increased pharyngeal residue. Quantitative videofluoroscopic studies have specifically confirmed that liquids prepared with starch‐based thickeners result in greater pharyngeal residue [30, 31, 32, 33].

This study was limited by its retrospective design and relatively small sample size across different CHF severity levels, which may affect generalizability. Respiratory phase data were not captured, preventing direct analysis of respiratory‐swallow coordination. Additionally, while bolus consistency levels were standardized according to IDDSI guidelines, volume across these consistencies was not controlled for in the statistical models. As such, the influence of bolus volume on swallow timing and airway protection could not be isolated and accounted for. Another limitation involves the distinction between statistical and clinical significance: while several findings showed statistical significance, including the strong association between incomplete laryngeal vestibule closure and airway invasion, clinical implications of these physiological impairments (e.g., aspiration pneumonia, prolonged hospitalizations, or readmissions) were not directly assessed.

In summary, this study suggests that CHF contributes to swallowing impairment through mechanisms similar to those observed in other pulmonary patient populations. Reduced cardiopulmonary reserve, tachypnea, low lung volumes, and pulmonary congestion may disrupt swallow‐breathing coordination, increasing aspiration risk. Although respiratory‐swallow phase patterns were not directly measured, the high rates of airway invasion and inefficiency observed underscore the need for prospective studies incorporating respiratory monitoring during VFSS and linking physiologic findings with clinical outcomes. Future research should also examine the impact of CHF severity, including NYHA class and ejection fraction, on swallowing physiology to better define risk profiles in this population.

Funding

The authors have nothing to report.

Conflicts of Interest

The authors declare no conflicts of interest.

Data Availability Statement

The de‐identified data that support the findings of this study are available from the corresponding author upon reasonable request.