Esophageal Hypervigilance and Symptom Specific Anxiety in Patients with Eosinophilic Esophagitis

Abstract

Background and Aims:

Patient symptom reporting often does not correlate with pathophysiological markers of esophageal disease, including eosinophilic esophagitis (EoE). Esophageal hypervigilance and symptom-specific anxiety are emerging as important considerations in understanding symptom reporting. As such, we aimed to conduct the first study of these constructs in EoE.

Methods:

A retrospective review of an EoE patient registry was conducted and included: 1) eosinophils per high power field (EOS/HPF from EGD biopsy: proximal, distal), 2) endoscopic reference score (EREFS), 3) distal distensibility plateau (FLIP), 4) Brief Esophageal Dysphagia Questionnaire (BEDQ), 5) Visual Dysphagia Question of EoE Activity Index (EEsAI-VDQ), 6) Northwestern Esophageal Quality of Life scale (NEQOL), 7) Esophageal Hypervigilance and Anxiety Scale (EHAS). Correlational and regression analyses evaluated relationships of hypervigilance and anxiety with BEDQ, EEsAI-VDQ, and NEQOL when controlling for histology and endoscopic severity.

Results:

103 patients had complete data: 69.9% male, average age of 40.66 (13.85) years. 41% had elevated dysphagia and 46% elevated hypervigilance and anxiety. Esophageal symptom-specific anxiety emerged as the most important predictor of BEDQ severity (44.8% of the variance), EEsAI-VDQ severity (26%), and poor HRQoL (55.3%). Hypervigilance was also important, to a lesser extent. Pathophysiological variables did not significantly predict symptoms or HRQoL. Recent food impaction may predict symptom-specific anxiety, while PPI use may reduce hypervigilance.

Conclusion:

Hypervigilance and symptom-specific anxiety are important for our understanding of self-reported patient outcomes in EoE. These processes outweigh endoscopic and histologic markers of EoE disease activity across dysphagia, difficulty eating, and HRQoL. Clinicians should assess hypervigilance and anxiety, especially in patients with refractory symptoms and poor HRQoL.

Graphical Abstract

Lay Summary:

The amount of attention a patient’s brain pays to symptoms of eosinophilic esophagitis, and how upsetting symptoms are, may be more important in explaining their severity than medical testing.

Introduction

Eosinophilic esophagitis (EoE) is a chronic esophageal disease of increasing global incidence and prevalence^{1, 2}. The prevailing theory of its etiology is a combination of immunologic and allergic processes that cause inflammation and fibrostenotic changes to the esophageal epithelium and subepithelium^{3, 4} with consistent sex-based differences (males > females). Clinical management includes dietary elimination of trigger foods, swallowed topical corticosteroids, proton pump inhibitors (PPI), or a combination of these⁵. Adult patients with EoE most often present with symptoms of dysphagia to solids and food impaction, but may also report heartburn, chest pain, and reflux. Unfortunately, dysphagia symptoms often do not correlate well with endoscopic or histological disease activity^{6, 7}. In a recent study, nondysphagia symptoms, including heartburn and reflux, did improve with histological response to treatment⁸. However, these data are preliminary and need replication.

Esophageal hypervigilance and symptom-specific anxiety are two related cognitive-affective processes that are emerging as important considerations for patient reported outcomes across esophageal diseases^{9, 10}. Hypervigilance, or the tendency to overly focus attention on physical sensations in the esophagus, and anxiety related to the presence, or possibility, of symptoms may explain dysphagia symptoms better than physiological data often used to assess esophageal disease. In a 2020 study, we found hypervigilance and anxiety had a two-fold higher predictive relationship of dysphagia symptoms than the presence of a major motor disorder per the Chicago Classification version 3.0⁹ and manometric data was not associated with dysphagia symptoms when controlling for hypervigilance and anxiety. A separate 2021 study found hypervigilance to esophageal sensations is persistent across the spectrum of reflux disorders¹⁰.

Very recently, the psychosocial effects of EoE in adult patients have been recognized in the research literature, including how these may relate to reported symptoms and treatment response. However, very few studies exist as of 2021. Adults with EoE with more anxiety related to having a choking episode tend to report more severe dysphagia and difficulty swallowing different foods^{11, 12}. Generalized, or non-disease specific anxiety, is also associated with worse EoE symptoms¹³. However, all of these studies utilized either a generic measure of anxiety or a measure of EoE-specific quality of life, neither of which elucidate the role of hypervigilance. Health related quality of life (HRQoL) is consistently degraded in patients with more severe EoE symptoms. Disease-specific measures are preferred in EoE, as these give more accurate representations to the impacts of the illness than generic measures¹⁴. Preliminary studies suggest esophageal disease specific HRQoL is associated with hypervigilance and anxiety in other populations.

No study to date assesses the presence of hypervigilance and symptom-specific anxiety in patients with EoE. As such, we aim to evaluate their relationship with 1) self-reported dysphagia and difficulty eating foods, 2) endoscopic disease severity, and 3) disease specific HRQoL. We will also assess potential predictive relationships of hypervigilance and anxiety for self-reported symptoms and HRQoL.

Materials and Methods:

A retrospective review of an EoE patient registry was performed. Data were prospectively collected as part of standard-of-care practices in adults aged 18 and up who visited a university-based gastroenterology clinic for esophagogastroduodenoscopy (EGD) between March 2019 and November 2020. The EoE diagnosis was defined by current clinical guidelines including an appropriate clinical history (i.e., dysphagia) and ≥ 15 eosinophils per high powered field on histological review⁵. Patients were included if they completed the following questionnaires immediately prior to endoscopy assessing esophageal hypervigilance and anxiety, dysphagia severity, difficulty swallowing different foods, and HRQoL:

Esophageal Hypervigilance and Anxiety Scale (EHAS)¹⁵:

The EHAS is a 15-item self-report measure that assesses esophageal hypervigilance and esophageal symptom-specific anxiety over the preceding one month. Items are rated on a 5-point Likert scale (scored 0-4). The items are summed to yield a total EHAS score, ranging from 0 – 60, with greater scores indicating greater esophageal hypervigilance and esophageal symptom-specific anxiety and a total score > 21 denotes elevated symptoms. Subscales for hypervigilance (scores 0-36) and symptom-specific anxiety (scores 0-24) are also calculated. The EHAS is validated in English, Spanish¹⁶, and French-speaking patients¹⁷.

Brief Esophageal Dysphagia Questionnaire (BEDQ)¹⁸:

The BEDQ is an 8-item self-report measure assessing the frequency and severity of dysphagia and odynophagia over the preceding 14-days. Items are rated on a 6-point Likert scale questions (scored 0-5). The items are summed to yield scores ranging from 0 (asymptomatic) – 40, with greater scores indicating greater dysphagia severity. A cut point of ≥5 is considered elevated dysphagia symptoms. Two additional items assess whether a person has had a food impaction, either self-limited (resolved within 30 minutes) or necessitating an ER visit, and are coded yes=1, no=0. Because of the low frequency of food impactions, items were pooled into a single “Food Impaction” Y/N item. The BEDQ is validated in English, Spanish¹⁹, and French²⁰ speaking patients.

Eosinophilic Esophagitis Symptom Activity Index (EEsAI)²¹:

The EEsAI is a validated measure of EoE symptom severity. For this study, only the Visual Dysphagia Question (VDQ) was given for brevity versus the full EEsAI. The VDQ evaluates the difficulty swallowing eight different types of foods over the last 7 days: solid meat, soft foods, dry rice, ground meat, fresh white untoasted bread, grits/porridge/rice pudding, raw/fibrous foods, and French fries. Each food was rated on a 4-point Likert scale (0=None, 1=mild, 2=moderate, 3=severe). Higher scores indicated greater difficulty swallowing foods and range from 0 to 24.

Northwestern Esophageal Quality of Life Scale (NEQOL)²²:

The NEQOL is a 14-item scale that measures HRQoL as it relates to esophageal symptoms that can be used across conditions. Items are rated on a 5-point Likert scale (0-4; Not at All True to Very True). Items on the NEQOL evaluate social function, emotional distress, eating impact, sleep, and financial burden. The scale is validated in patients with esophageal diseases, including EoE.

Endoscopic Assessment of Disease

Patients underwent EGD to assess EoE disease activity. During endoscopy, 4-quardant esophageal mucosal biopsies were obtained from the distal and proximal esophagus at 5 and 15-cm above the squamocolumnar junction, respectively. Histologic evaluation of biopsy specimens was performed by local pathologists with expertise in gastrointestinal pathology. The EoE Endoscopic Reference Score (EREFS) system²³ was used to rate macroscopic aspects of esophageal mucosa and scored as follows: edema (0=distinct vascularity, 1=decreased), rings (0=none, 1=mild (ridges), 2=moderate (distinct rings), 3=severe (could not pass scope), exudates (0=none, 1=mild (<10% surface area), 2=severe (>10% surface area), longitudinal furrows (0=none, 1=mild, 2=severe), and stricture (0=none, 1=present). A total EREFS score was calculated by summing the 5 items, with higher scores equating to more severe mucosal changes. A score of < 3 is considered endoscopic remission.

A subset of patients also underwent endoluminal functional luminal imaging probe (FLIP) immediately following their EGD. FLIP simultaneously measures the luminal cross-sectional area (~diameter) and its related pressure in response to controlled volumetric distension. The distensibility plateau was measured as the narrowest, fixed diameter (after excluding esophageal contractions) of the distal esophageal body in response to increasing FLIP volume and pressure²⁴.

Ethical Approval

The study protocol was approved by the Northwestern University Institutional Review Board; a minimal risk exemption was granted, and thus patient informed consent was not obtained.

Statistical Analyses

Data were exported from the REDCap database to SPSS version 27 for Macintosh (Chicago, IL) and assessed for completeness. Total and subscale scores were computed for each of the questionnaires. Tests for normal distribution were performed to determine the need for non-parametric tests (Skewness/Kurtosis >2.0 or < −2.0). Proximal and distal eosinophils per high powered field (EOS/HPF) were non-normally distributed. Categorical variables are presented as percentage (frequency) and continuous variables are presented as mean (SD) or Median (Interquartile Range). Participants were coded PRE/POST COVID-19 based on a cutoff date of March 11, 2020 and differences between the groups were evaluated with independent samples t-Test or chi-square to ensure no pandemic-related effects on the data existed.

Differences between male and female participants for all questionnaires and endoscopic data were measured via independent samples t-Test or Mann-Whitney U (EOS/HPF only). Pearson’s correlations identified significant relationships between the EHAS, BEDQ, EEsAI, NEQOL, EREFS, and EndoFLIP data; EOS/HPF was evaluated using Spearman’s Rho. Scatterplots with linear regression curve and mean confidence bands present correlational data. An additional independent samples t-Test assessed differences in the four self-report measures for patients with and without an esophageal stricture.

Due to multicollinearity between symptom-specific anxiety and hypervigilance (r=.690), anxiety and hypervigilance were evaluated separately via hierarchical linear regression to determine possible predictive relationships for self-reported symptom severity and HRQoL. Prior to the final models being generated, gender and age were tested as potential confounds and did not significantly alter the percentage variance or standardized Beta weights for the EHAS scores and physiological data. For regression 1a and 1b (Criterion: Dysphagia severity), EHAS anxiety (1a) or hypervigilance (1b) subscales were entered at step 1, EREFS total and stricture presence (No=0, Yes=1) was entered at step 2, EOS/HPF (proximal and distal) at step 3, and FLIP DP at step 4. For regression 2a and 2b (Criterion: Difficulty Swallowing Foods), EHAS anxiety (2a) or hypervigilance (2b) subscales were entered at step 1, EREFS total score and stricture was entered at step 2, EOS/HPF at step 3, and FLIP DP at step 4. For regression 3a and 3b (Criterion: HRQoL), EHAS anxiety (3a) and hypervigilance (3b) subscales were entered at step 1, EREFS score was entered at step 2, BEDQ at step 3, and EEsAI-VDQ at step 4. Probability statistic (F) for each variable set is reported with Adjusted R² for step 1, also converted to percentage variance, and R² change, converted to percentage variance for steps 2-4. Unstandardized beta (B) with Standard Error (SE), Standardized beta coefficient (β) are reported for each variable at each step of the regression. Statistical significance was set to P<.05 for all analyses.

Results:

Study Sample

One hundred and three patients with endoscopically and histologically confirmed EoE who had complete data were identified. The sample comprised 8% of the unique EoE patient population seen in the outpatient clinic from January 2019 to March 2021 (n=1,387). Study participants were predominantly male (69.9%) with an average age of 40.66 (13.85) years, Range: 18-73; they did not differ significantly from the total clinical population for gender (60% male, p=.053) or age (42.30(13.81), Range: 18-90; p=0.25). Clinical characteristics of the sample are presented in Table 1. Forty-one percent reported elevated dysphagia (BEDQ ≥ 5) and 46% had elevated hypervigilance and anxiety (EHAS > 21). Age was not associated with EHAS score (r= −.015, p= .878) nor were there differences between males and females (p= .371).

Table 1.

Clinical Characteristics of Study Sample

	N=103
	Questionnaires
	Mean (SD) or %(N)	Range
BEDQ: Dysphagia Symptoms	5.27 (6.55)	0-30
BEDQ: Any Food Impaction in past 14 days	10.7% (11)	-
BEDQ: Percentage with Elevated Symptoms (Cut point = 5 or more)	40.8% (42)	-
EEsAI: Difficulty Swallowing Foods	5.41 (5.83)	0-24
EHAS: Hypervigilance	9.85 (5.94)	0-22
EHAS: Anxiety	12.05 (8.31)	0-30
EHAS: Total	21.90 (13.15)	0-52
EHAS: Percentage with Elevated Symptoms (Cut point = 22 or more)	45.6% (47)	-
NEQOL	42.19 (13.27)	0-56
	Endoscopic and Histologic Findings
EREFS Score	2.95 (1.64)	0-7
Percentage in Endoscopic Remission (EREFS < 3)	57.8% (59)	-
Percentage with EREFS Positive
Edema	58.3% (60)	-
Rings	67.0% (69)	-
Exudates	31.1% (32)	-
Furrows	58.3% (60)	-
Stricture	48.5% (50)
Dilation Performed During EGD	49.5% (51)	-
Histology: EOS/HPF, Proximal*	10.5	0-55.25
Histology: EOS/HPF, Distal*	25.0	0-50
Percentage with Active EoE (EOS/HPF > 15)	61.3% (46)	-
FLIP: Distensibility Plateau (Distal)	17.88 (2.83)	10-21

Notes:

^*EOS/HPF are reported as Median and Interquartile Range. Food impaction includes self-limited (30 minutes or less) or impactions requiring emergency room visit.

Most patients (73%) presented for EGD with dysphagia as their primary reason for testing. Very few reported heartburn/reflux (4.9%) or chest pain (1.9%) as their chief complaint, and the remaining 12% were for follow up related to treatment (e.g., dietary, topical steroids). Approximately three-quarters (79.6%) underwent FLIP at the time of EGD. The majority were taking a PPI (72.8%), the most common being 40mg of omeprazole (54.4%) on a once-daily schedule (53.4%). A small proportion (10.7%) reported having a food impaction in the last 2 weeks, and 49.5% underwent esophageal dilation during their EGD. Males had more severe esophageal rings per the EREFS scoring system (p<.001). No other differences existed between genders.

Impact of COVID-19 Pandemic

Since the present study assesses hypervigilance and anxiety, we considered possible effects of the COVID-19 pandemic, which has had substantial consequences for mental health²⁵, on the study data. The majority of EGDs (72.8%) were performed prior to March 11, 2020 when clinical operations at Northwestern Medicine were put on hold due to the pandemic. No significant differences existed between PRE/POST COVID groups for BEDQ (p=.124), EEsAI-VDQ (p=.247), NEQOL (p=.735), EHAS: Hypervigilance (p=.170), or EHAS: Anxiety (p=.986). No differences existed by gender (p=.071) or PPI use (p=.515). As such, the sample was pooled regardless of timing of the EGD.

Self-Reported EoE Symptoms and Endoscopic Severity

Patients who reported more difficulty swallowing different types of foods on the EEsAI-VDQ also had significantly more dysphagia symptoms on the BEDQ (r= 0.729, p< .001). No relationships existed between the EREFS score and dysphagia severity (r= −0.052, p=.604) or difficulty swallowing foods (r= −0.087, p=.388). Proximal and distal EOS/HPF, as well as distal distensibility plateau, were not significantly correlated with EEsAI-VDQ or BEDQ scores (all p >.05); patients who had active EoE (EOS/HPF > 15) also did not differ for dysphagia symptoms (p=.651) and difficulty eating foods (p=.217). Patients with a stricture found during EGD reported more difficulties with eating different types of foods than those without a stricture (p=.024) but did not differ for dysphagia symptoms (p=.057). Those with at least one food impaction in the last 2 weeks had considerably greater difficulty swallowing foods (p=.004); however, food impaction occurrence was not associated with any endoscopic severity variable (all p > .10).

Esophageal Hypervigilance and Anxiety, Endoscopic Severity, and Self-Reported EoE Symptoms

Patients who experienced higher levels of esophageal hypervigilance and anxiety also reported significantly more symptoms of dysphagia (Figure 1). Similar associations also existed between EHAS scores and difficulty swallowing different types of foods. No relationships existed between EHAS scores and EREFS total score, EOS/HPF, Active EoE (EOS/HPF > 15), or distal DP. Patients with a stricture found during EGD had similar hypervigilance to esophageal symptoms (Mean(SD)= 10.96(5.84) vs. 9.22(5.88), p=.144) and symptom-anxiety (12.64(8.44) vs. 12.06(8.28), p=.731) to those who did not have a stricture. However, those who reported at least one food impaction in the last 2 weeks reported more symptom-specific anxiety (17.64(9.07) vs. 11.38(8.00), p=.0.17); they did not differ on hypervigilance (p=.405).

Figure 1.

Pearson’s Correlation Between Total Esophageal Hypervigilance and Anxiety and Dysphagia Symptoms and Difficulty Eating Foods

Esophageal Hypervigilance and Anxiety, Self-Reported EoE Symptoms, and HRQoL

Patients who reported higher levels of esophageal hypervigilance and anxiety also experienced significantly poorer HRQoL (Figure 2). Both dysphagia severity (r= −.592, p<.001) and difficulty swallowing foods (r= −.498, p<.001) were associated with lower HRQoL but reporting a food impaction was not (p=.106). Endoscopic severity scores did not correlate with HRQoL and those with a stricture scored similarly to those without (41.29(12.93) vs. 43.56(12.42), p=.379).

Figure 2.

Pearson’s Correlation Between Total Esophageal Hypervigilance and Anxiety and HRQoL

Predictors of Dysphagia Symptoms

Esophageal symptom-specific anxiety and hypervigilance were entered into separate hierarchical linear regression models with endoscopic severity, as measured by EREFS and stricture presence, EOS/HPF and distal DP to determine to what extent these variables may predict dysphagia severity in EoE patients (Table 2).

Table 2.

Hierarchical Regression Analyses for Dysphagia Symptom Severity and EHAS Scores

		Univariate Statistics			Multivariate Statistics
					Unstandardized Coefficient		Standardized Coefficient
Model	Measurement	F	R2adj	ΔR 2	B	SE	β	P
1	EHAS: Anxiety	48.88	.448	.457	.526	.075	.676	<.001
2	EREFS Score	3.05	.467	.028				.086
	EHAS: Anxiety				.527	.074	.677	<.001
	EREFS				−.645	.369	−.166	.559
3	EOS/HPF	.005	.448	.000				.995
	EHAS: Anxiety				.527	.076	.678	<.001
	EREFS				−.638	.384	−.164	.103
	EOS/HPF: Proximal				.002	.026	.016	.933
	EOS/HPF: Distal				−.003	.028	−.019	.921
4	FLIP DP	.542	.443	.005				.464
	EHAS: Anxiety				.522	.076	.672	<.001
	EREFS				−.638	.394	−.149	.146
	EOS/HPF: Proximal				.003	.026	.024	.901
	EOS/HPF: Distal				−.001	.029	−.004	.982
	FLIP DP				.181	.245	.077	.464
Model	Measurement	F	R2adj	ΔR 2	B	SE	β	P
1	EHAS: Hypervigilance	21.11	.254	.267	.538	.117	.517	<.001
2	EREFS Score	2.27	.270	.028				.138
	EHAS: Hypervigilance				.540	.116	.518	<.001
	EREFS				−.651	.432	−.168	.138
3	EOS/HPF	.067	.245	.002				.935
	EHAS: Hypervigilance				.540	.119	.519	<.001
	EREFS				−.641	.449	−.165	.159
	EOS/HPF: Proximal				.011	.030	.080	.719
	EOS/HPF: Distal				−.009	.033	−.062	.782
4	FLIP DP	1.23	.249	.016				.273
	EHAS: Hypervigilance				.540	.119	.519	<.001
	EREFS				−.540	.457	−.139	.243
	EOS/HPF: Proximal				.013	.030	.093	.674
	EOS/HPF: Distal				−.006	.033	−.038	.868
	FLIP DP				.315	.284	.134	.273

Symptom Specific Anxiety:

The univariate model for anxiety was significant (F=48.882, p<.001); EREFS score and stricture presence (F=3.050, p=.086), EOS/HPF (F=0.005, p=.995), and distal DP (F=0.543, p=.464) were not. For every 1 standard deviation (SD) increase in symptom-anxiety, dysphagia severity increases by 0.68 SDs (β=0.676). Symptom-specific anxiety remained the only significant predictors of dysphagia symptoms across the 4 multivariate models when accounting for histology and FLIP with little change in its standardized beta weight observed (β=0.672).

Hypervigilance:

The univariate model for hypervigilance was significant (F=21.11, p<.001); EREFS and stricture (F=2.268, p=.138), EOS/HPF (F=.067, p=.935) and distal DP (F=1.229, p=.273) were not. For every 1 SD increase in hypervigilance, dysphagia severity increased .52 SD (β=0.517). Hypervigilance remained the only significant predictor of dysphagia across the 4 multivariate models when accounting for the other variables (all p<.001) and little change in its standardized beta weight (β=0.519).

Predictors of Difficulty Swallowing Different Foods

Esophageal hypervigilance and anxiety were next entered into separate hierarchical linear regression models with endoscopic severity, as measured by EREFS and stricture presence, EOS/HPF, and distal DP to determine to what extent these variables may predict difficulty swallowing foods in EoE patients (Table 3).

Table 3.

Hierarchical Regression Analyses for Difficulty Eating Foods and EHAS Scores

		Univariate Statistics			Multivariate Statistics
					Unstandardized Coefficient		Standardized Coefficient
Model	Measurement	F	R2adj	ΔR 2	B	SE	β	P
1	EHAS: Anxiety	20.63	.260	.273	.314	.069	.522	<.001
2	EREFS Score + Stricture	.685	.251	.018				.508
	EHAS: Anxiety				.308	.070	.512	<.001
	EREFS				−.330	.349	−.109	.350
	Stricture Present				.794	1.20	.077	.511
3	EOS/HPF	.277	.230	.008				.759
	EHAS: Anxiety				.312	.071	.518	<.001
	EREFS				−.276	.362	−.091	.449
	Stricture Present				.927	1.23	.090	.455
	EOS/HPF: Proximal				.011	.024	.107	.642
	EOS/HPF: Distal				−.019	.027	−.163	.488
4	FLIP DP	1.16	.232	.016				.286
	EHAS: Anxiety				.306	.071	.508	<.001
	EREFS				−.197	.368	−.065	.595
	Stricture Present				.931	1.23	.090	.453
	EOS/HPF: Proximal				.013	.024	.120	.602
	EOS/HPF: Distal				−.016	.027	−.136	.563
	FLIP DP				.244	.226	.136	.286
Model	Measurement	F	R2adj	ΔR 2	B	SE	β	P
1	EHAS: Hypervigilance	7.36	.102	.118	.280	.103	.344	.009
2	EREFS Score + Stricture	.781	.095	.025				.463
	EHAS: Hypervigilance				.273	.104	.335	.011
	EREFS				−.359	.384	−.119	.355
	Stricture Present				1.05	.384	.102	.798
3	EOS/HPF	.308	.071	.010				.736
	EHAS: Hypervigilance				.283	.106	.348	.010
	EREFS				−.301	.397	−.100	.452
	Stricture Present				1.20	1.35	.117	.892
	EOS/HPF: Proximal				.017	.027	.160	.635
	EOS/HPF: Distal				−.023	.029	−.202	.438
4	FLIP DP	1.67	.083	.027				.203
	EHAS: Hypervigilance				.282	.106	.347	.010
	EREFS				−.197	.403	−.065	.627
	Stricture Present				1.20	1.34	.116	.377
	EOS/HPF: Proximal				.019	.026	.177	.485
	EOS/HPF: Distal				−.019	.029	−.168	.518
	FLIP DP				.318	.246	.177	.203

Symptom-Specific Anxiety:

The univariate model for anxiety was significant (F=20.627, p<.001); EREFS and having a stricture (F=0.685, p=.508), EOS/HPF (F=0.277, p=.759), and distal DP (F=1.164, p=.286) were not. For every 1 SD increase in symptom-specific anxiety a patient’s difficulty swallowing foods increased by 0.52 SD (β=0.522). Hypervigilance remained a significant predictor when controlling for the other variables in all the multivariate regression models with little change in beta weight (β=0.508).

Hypervigilance:

The univariate model for hypervigilance was significant (F=7.364, p=.009). EREFS and having a stricture (F=0.781, p=.463), EOS/HPF (F=.308, p=.736), and distal DP (F=1.666, p=.203) were not. For every 1 SD increase in hypervigilance, a patient’s difficulty swallowing foods increased by 0.34 SD (β=0.344). Hypervigilance remained the only significant predictor when controlling for other variables across all multivariate models with little change in beta weight (β=0.347).

Predictors of HRQoL

Esophageal hypervigilance and anxiety were entered into separate hierarchical linear regression models with endoscopic severity, as measured by EREFS, and self-reported dysphagia (BEDQ) and difficulty swallowing foods (EEsAI-VDQ) to determine to what extent these variables may predict HRQoL (Table 4).

Table 4.

Hierarchical Regression Analyses for Health-Related Quality of Life and EHAS Scores

		Univariate Statistics			Multivariate Statistics
					Unstandardized Coefficient		Standardized Coefficient
Model	Measurement	F	R2adj	ΔR 2	B	SE	β	P
1	EHAS: Anxiety	122.1	.553	.557	−1.19	.108	−.746	<.001
2	EREFS Score	.362	.550	.002				.549
	EHAS: Anxiety				−1.20	.108	−.747	<.001
	EREFS				.329	.547	.041	.549
3	BEDQ	6.74	.575	.029				.011
	EHAS: Anxiety				−.987	.132	−.617	<.001
	EREFS				.212	.533	.026	.692
	BEDQ				−.430	−.215	−.215	.011
4	EEsAI	.000	.571	.000				.996
	EHAS: Anxiety				−.987	.135	−.617	<.001
	EREFS				.212	.538	.026	.695
	BEDQ				−.430	.210	−.215	.044
	EEsAI				−.001	.226	.000	.996
Model	Measurement	F	R2adj	ΔR 2	B	SE	β	P
1	EHAS: Hypervigilance	38.9	.279	.287	−1.19	.191	−.535	<.001
2	EREFS Score	.568	.276	.004				.453
	EHAS: Hypervigilance				−1.20	.192	−.539	<.001
	EREFS				.523	.694	.065	.453
3	BEDQ	25.8	.425	.151				<.001
	EHAS: Hypervigilance				−.772	.191	−.346	<.001
	EREFS				.217	.622	.027	.728
	BEDQ				−.871	.172	−.435	<.001
4	FLIP DP	1.93	.430	.011				.168
	EHAS: Hypervigilance				−.788	.190	−.353	<.001
	EREFS				.158	.620	.020	.800
	BEDQ				−.634	.242	−.316	.010
	EEsAI				−.356	.256	−.157	.168

Symptom-Specific Anxiety:

The univariate model for anxiety was significant (F=122.091, p<.001) as was the model for dysphagia symptom severity (F=6.738, p=.011). For every 1 SD increase in symptom-specific anxiety, a patient’s HRQoL decreased by 0.75 SD (β= −0.746). Dysphagia severity was significant but with a much smaller beta weight (β= −0.215). EREFS score (F=0.362, p=.549) and difficulty swallowing foods (F=0.001, p=.996) were not significant. Anxiety and dysphagia severity remained significant predictors of HRQoL across all 4 multivariate models, where EREFS and EEsAI-VDQ scores were not. The relationship between BEDQ and HRQoL did not change, however that of symptom-specific anxiety became somewhat smaller (β= −0.617).

Hypervigilance:

The univariate model for hypervigilance was significant (F=38.993, p<.001) as was the model for dysphagia symptom severity (F=25.800, p<.001). For every 1 SD increase in hypervigilance, a patient’s HRQoL decreased by 0.54 SD (β= −0.535). Dysphagia severity was significant with comparable beta weight (β= −0.435). EREFS (F=0.568, p=.453) and difficulty swallowing foods (F=1.933, p=.168) were not significant. Hypervigilance and dysphagia severity remained significant predictors of poorer HRQoL across all 4 multivariate models, where EREFS and EEsAI-VDQ scores were not. The relationship between BEDQ and HRQoL became smaller (β= −0.316) as did the relationship for hypervigilance (β= −0.353) when considering EEsAI and EREFS.

Predictors of Esophageal Hypervigilance and Anxiety

Lastly, we attempted to elucidate some reasons why an EoE patient may experience greater levels of hypervigilance and symptom-specific anxiety. In addition to the increased symptom-specific anxiety, but not hypervigilance, in patients who reported a food impaction in the last 2 weeks, those who were taking a PPI reported significantly less hypervigilance (No: 11.89(5.63) vs. Yes: 9.09(5.91), p=.033); frequency of use was not significant (p=.065). A simple linear regression model shows having a food impaction is a significant predictor of higher anxiety (F=5.840, p=.017) and accounts for 5.5% of the variance (β= 0.234). Taking a PPI is a significant predictor of less hypervigilance (F=4.687, p=.033), accounting for 4.4% of the variance (β= −0.211).

Discussion:

This is the first study to evaluate hypervigilance and anxiety as they relate to esophageal symptoms in a cohort of patients with EoE. We found symptom-specific anxiety and, to a lesser extent hypervigilance, are important considerations when interpreting self-reported dysphagia symptoms and difficulty with swallowing various foods. In fact, symptom-specific anxiety and hypervigilance were the only predictors of increased dysphagia symptoms when accounting for endoscopic and histologic severity. Consistent with a prospective, multicenter study the EREFS total score did not predict dysphagia symptoms²⁶ and the presence of a stricture, while associated with difficulty eating, did not remain a significant predictor of these challenges. Other typical histological benchmarks (EOS/HPF) and a common FLIP measure (distal distensibility plateau) demonstrated no significant relationship with reported symptoms.

Prior research finds mixed influences of esophageal symptom specific anxiety and hypervigilance and patient self-report of dysphagia. In a large cohort undergoing high resolution manometry (HRM), pooled hypervigilance and anxiety predicted 12% of the variance in dysphagia severity. Because these constructs were not assessed separately, we cannot make a direct comparison to the HRM sample of whether hypervigilance or anxiety was the more important factor⁹. In a 2021 study, Guadagnoli et al. found hypervigilance was not only persistent across gastroesophageal reflux disease (GERD) phenotypes, but significantly predicted GERD symptom severity when controlling for symptom-specific anxiety¹⁰. While hypervigilance and anxiety are related, they represent distinct constructs, and it is feasible a patient could be hypervigilant to their esophageal sensations without developing significant anxiety, and vice versa. Additionally, patients who demonstrate very low hypervigilance or anxiety to symptoms may be at risk for under-reporting disease activity or not seeking appropriate medical treatment.

Food impaction may be a unique reason for heightened symptom-specific anxiety in EoE, as impactions are not typically seen in other esophageal conditions. However, patients with a recent food impaction were not statistically more hypervigilant and the reason is unclear. One explanation may be due to the small number of patients who reported an impaction as those who did experience this symptom had higher hypervigilance, on average, than those who did not. With a larger sample of patients with an impaction a significant difference may be detected, which would track with what is typically observed in clinic visits. Alternatively, PPI use is associated with less hypervigilance, albeit a small percentage of its variance, likely by reducing reflux symptoms occurring in some EoE patients⁸. It is plausible there are differences in the cognitive-affective processes of EoE patients compared to those with other chronic esophageal diseases due to food impaction experiences. Further research is necessary.

Both symptom-specific anxiety and hypervigilance were important contributors to patient difficulties with swallowing different foods. In prior studies evaluating disease-specific quality of life, anxiety related to choking episodes was consistently associated with greater dysphagia severity in both English and Spanish speaking patients^{11, 27-31}. Specifically, patients reporting greater scores on the EEsAI-VDQ had poorer disease-specific HRQoL¹², as measured by the EoE-QOL-A³⁰, explaining 31% of the variance in choking and illness-related anxiety. This, combined with our findings that experiencing a food impaction in the past 2 weeks may result in more anxiety, suggests prior food impactions that patients found particularly painful, or distressing, may contribute to heightened fears about food getting stuck again. Future studies on the association between EoE choking episodes or food impactions and anxiety, including post-traumatic stress, are warranted.

Mental health is only very recently investigated in adult patients with EoE. A scoping review of anxiety and depression in adults with EoE in 2019³² found only 11 studies in adult patients, with 6 of these measuring anxiety via a measure of HRQoL (the EoE-QOL-A) versus a psychiatric questionnaire, such as the Hospital and Anxiety Scale (HADS). In the only cross-sectional study prior to 2019 specifically measuring anxiety via the HADS, 31.1% of 170 patients scored above the clinical cutoff for anxiety¹³. In 2020, de Rooj and colleagues found no differences between EoE patients and population norms for anxiety and depression measured by the HADS. However, more patients had high levels of anxiety (24%) than depression (10%). Moreover, 36% of the sample scored above the 80^th percentile on the SLC-90-R Global Severity Index, which measures multiple mental health domains including anxiety, phobic anxiety, and obsessive-compulsive behaviors³³. A retrospective review in the same year found identical rates of anxiety as measured by the HADS (23%) in 883 patients with EoE³⁴, further suggesting anxiety is an important mental health consideration in this population. Comparatively, a 2019 retrospective chart review identified only 9.3% of 34 patients had an ICD-9 or ICD-10 diagnosis of an anxiety disorder. These findings suggest two possible scenarios: 1) approximately one-quarter of EoE patients are experiencing anxiety, especially related to their disease and its symptoms, but only a fraction are reporting this to their physicians or 2) anxiety related to EoE is disease specific and, thus, will elevate a generic measure but not meet the criteria for a psychiatric diagnosis.

The psychosocial burden of EoE likely starts at the onset of symptoms, which can predate diagnosis sometimes by years. Qualitative interviews with 20 adult EoE patients found 90% reported feeling worried or anxious about symptoms, 60% were irritated or frustrated, and 30% felt their life revolved around EoE prior to being diagnosed³⁵. Patients in this cohort had an average delay of 11.1 years in diagnosis with 70% stating that obtaining a diagnosis was “a great” or “very great” burden. After diagnosis, rates of worry or anxiety dropped from 90% to 65%, but twice as many people (60%) felt their life revolved around their EoE, suggesting an increase in hypervigilance to their condition. This may be reflected in the use of elimination diets as treatment by 80%, which can be difficult to follow and are associated with anxiety³⁶. Additional studies are needed to elucidate how these early experiences, as well as treatment choices, may affect esophageal hypervigilance and symptom-specific anxiety in EoE patients.

By far, HRQoL was most affected by symptom-specific anxiety and hypervigilance in this cohort of EoE patients, explaining three-quarters of HRQoL variance when controlling for endoscopic severity and symptoms. Dysphagia severity was also important, but only accounted for one-fifth of HRQoL score. Unlike the prior Swiss cohort study, difficulty swallowing foods did not predict HRQoL in this sample. We used a broader esophageal specific, versus EoE-specific, measure of HRQoL which may explain this difference. Unlike the EoE-QOL-A, the NEQOL used here does not assess specific anxieties about choking. Rather, items assess worry about something being seriously wrong with their esophagus, including developing esophageal cancer, worry eating will make them sick, and a general feeling of worry related to their esophageal condition²². The present study should be repeated using the EoE-QOL-A as this is the most specific to the effects of living with EoE.

There are some limitations to this study to consider. While data was prospectively collected as part of a larger esophageal registry, it was retrospectively reviewed. The data were collected from a single university-based center specializing in EoE, are cross-sectional in nature, and only assess no more than 4 weeks prior to their collection, making broader associations difficult. The sample was predominantly male; however, the gender ratio matches existing literature on prevalence of EoE in males and females. No racial or ethnic data was collected, so we cannot assess if these demographics are important considerations when understanding hypervigilance and symptom-specific anxiety in EoE. Questionnaires were completed prior to EGD/FLIP procedures, which may amplify anxiety and increase EHAS scores. No post-procedure repeated measures were done to determine these effects, which would be an important point for future studies. The sample size allowed for adequately powered regression analyses (β=0.855 to 0.884), however prospective replication is needed. There is also a chance some of these findings may be explained by unidentified dysmotility (e.g., ineffective esophageal motility) as would be assessed via high resolution manometry (HRM). Only 20% of the present sample underwent HRM at the time of assessment, precluding this evaluation which should be addressed in future studies.

Conclusions

Hypervigilance to, and anxiety about, EoE symptoms are important clinical considerations when evaluating patient outcomes that should be routinely evaluated. The EHAS-Short Form is a 7-item validated measure that serves this purpose³⁷. For patients with enduring dysphagia despite histological and endoscopic remission, or patients who have difficulty reintroducing foods after initial use of an elimination diet, evaluation of hypervigilance and anxiety is critical before using other treatments. What starts as a protective and instinctual response to perceived threat, pain in the esophagus, can become a conditioned attentional response in the gut-brain system. Both hypervigilance and anxiety are amenable to cognitive behavioral therapies (CBT) validated in other digestive conditions^{38, 39}. In the absence of access to Psychogastroenterology practitioners, clinicians can teach basic relaxation strategies (e.g., diaphragmatic breathing) to reduce anxiety. Several mobile applications exist for anxiety disorders, but their efficacy are mixed and some may not be based in CBT strategies⁴⁰ and none are specific to esophageal disease. Accordingly, these should be recommended with caution.

What You Need to Know:

BACKGROUND AND CONTEXT:

Hypervigilance to, and anxiety about, esophageal symptoms may explain discrepancies between self-reported symptoms and physiological data in EoE. No study evaluates if this is the case.

NEW FINDINGS:

Many EoE patients have elevated hypervigilance and symptom anxiety. No EoE physiological markers predicted dysphagia or quality of life, while anxiety predicted 44% of dysphagia and 57% of quality-of-life scores.

LIMITATIONS:

Data were retrospectively reviewed and do not include detailed demographic information.

IMPACT:

These findings may explain some gaps in patient symptom reporting and standard EoE testing, especially for those with normal physiology but persisting symptoms or reduced quality of life.