Gastrointestinal-Specific symptom anxiety in patients with gastroparesis: Relationships to symptom severity and quality of life

Abstract

Background:

Gastrointestinal (GI)-specific anxiety has been identified as a treatment target in irritable bowel syndrome. However, GI-specific anxiety has been understudied in other GI functional/motility disorders. Among adults with gastroparesis, we aimed to: (1) initially validate a measure of GI-specific anxiety, the Visceral Sensitivity Index (VSI); and (2) evaluate the relationship between GI-specific anxiety and gastroparesis symptom severity and quality of life, compared to measures of anxiety, depression, and somatization.

Methods:

Consecutive adult patients (N = 100) with gastroparesis presenting for initial consultation completed a series of self-report measures including the VSI. We conducted a confirmatory factor analysis of the VSI one-factor structure and tested internal consistency and convergent validity. We then performed hierarchical linear regression analyses to explore associations between VSI and gastroparesis symptom severity and overall quality of life.

Key Results:

Confirmatory factor analysis revealed that the original VSI one-factor structure overall fit well [χ²(90) = 220.1, p < 0.0001; SRMR = 0.08; RMSEA = 0.12; CFI = 0.96]. The VSI also had excellent internal consistency (α = 0.99) and convergent validity (r = 0.29–0.56; all p < 0.01). Higher GI-specific anxiety was significantly associated with greater gastroparesis symptom severity, including nausea/vomiting, fullness/satiety, and upper abdominal pain scores beyond depression, anxiety, or somatization (all p = <0.01–0.01). Additionally, higher GI-specific anxiety was significantly associated with lower mental health-related quality of life, beyond gastroparesis symptom severity, depression, anxiety, or somatization (p = 0.01).

Conclusions & Inferences:

The VSI is an adequate measure of GI-specific anxiety in patients with gastroparesis. Higher GI-specific anxiety was associated with increased patient-reported gastroparesis symptom severity and decreased quality of life, beyond depression/anxiety.

1 |. INTRODUCTION

Gastroparesis is a motility disorder characterized by delayed gastric emptying in the absence of mechanical obstruction along with symptoms of early satiation, postprandial fullness, nausea, vomiting, bloating/distension, and/or upper abdominal pain.¹ Current treatment guidelines focus primarily on dietary modifications and promoting gastric emptying through prokinetic agents.¹ However, it remains controversial whether the degree of gastric emptying correlates with symptom severity.² Given recent evidence of their similar clinical and pathologic features,³ gastroparesis and functional dyspepsia are thought to lie along a spectrum, with complex mechanisms that contribute to bidirectional communication between the gastrointestinal (GI) tract and the central nervous system.⁴

Symptom-specific anxiety has been identified as a key mechanism contributing to maintenance of disorders of gut-brain interaction (also known as functional GI disorders). Specifically, a model of GI-specific anxiety has been evaluated in irritable bowel syndrome (IBS)^5,6 and more recently been applied to other disorders including functional abdominal pain⁷ and functional dyspepsia.^7,8 In this model, certain stimuli (e.g., situations, foods) become associated with aversive experiences around GI symptoms over time, resulting in fear of those stimuli with attempts to either avoid or be cautious around them.⁹ In IBS, heightened GI-specific anxiety has been associated with increased IBS symptoms both cross-sectionally^10,11 and longitudinally,¹² and consistent evidence shows that GI-specific anxiety is reduced by behavioral treatment on reduced GI symptom severity.^5,6,13,14 Preliminary evidence also shows that GI-specific anxiety decreases in behavioral treatments for functional dyspepsia.⁷ It is hypothesized that GI-specific anxiety is associated with conditioned psychophysiological arousal which in turn contributes to GI symptom maintenance.¹⁵ However, to our knowledge, no measure of GI-specific anxiety has been evaluated in gastroparesis, a disorder that can have significant symptoms that impair quality of life.

The Visceral Sensitivity Index (VSI) was developed as a measure of GI-specific anxiety for IBS by assessing five domains specific to GI-specific anxiety (worry, fear, vigilance, sensitivity, and avoidance).¹⁶ While the measure creators indicate that the VSI can be used across the GI symptom spectrum, it has not yet been validated in GI functional/motility disorders outside of IBS.¹⁶ Thus, among adults with gastroparesis, we aimed to: (1) determine the initial validity of the VSI as a measure of GI-symptom-specific anxiety in patients with gastroparesis, and (2) evaluate the relationship between GI-specific anxiety and gastroparesis symptom severity, compared to measures of anxiety, depression, and somatization. Based on evidence showing that GI-specific anxiety is a key mechanism contributing to IBS maintenance,^5,6 we hypothesized that higher levels of GI-specific anxiety would be significantly related to greater gastroparesis symptom severity. We also explored the relationship between GI-specific anxiety with specific gastroparesis symptom severity subscales and overall quality of life scores.

2 |. MATERIALS AND METHODS

2.1 |. Subjects

Consecutive patients age ≥ 18 years who presented to Temple University Hospital for evaluation of gastroparesis were prospectively surveyed from January 2021 to December 2021 (N = 171) via an IRB-approved protocol (Temple University Hospital IRB #22347). For this study, we included patients with a clinical history consistent with gastroparesis and evidence of delayed gastric emptying on gastric emptying scintigraphy, defined as retention of >60% at 2 hours and/or > 10% at 4 h of a radiolabeled meal (liquid egg white (Eggbeaters^®) radiolabeled with 99 m-technetium) (N = 101). Gastric emptying scintigraphy was performed using a methodology previously described using the EggBeaters meal with 4-hour imaging.^17–19 Patients were excluded if they completed <75% of the items in the VSI (n = 1).

2.2 |. Self-report measures

GI-specific anxiety was measured using the Visceral Sensitivity Index (VSI),¹⁶ a 15-item questionnaire previously validated in adults with IBS,¹⁶ and in individuals with eating disorders.²⁰ Items are scored from 0 (“strongly disagree”) to 5 (“strongly agree”), with total scores ranging from 0 to 75. Anxiety and depression were measured using the Hospital Anxiety and Depression Scale (HADS), with total scores ranging from 0 to 21.²¹ Somatization was measured using the Patient Health Questionnaire 15 (PHQ-15), with total score ranging from 0 to 24.²² Avoidant/restrictive food intake was measured (given theoretical relevance to GI-specific anxiety for convergent validity) using the Nine-Item ARFID Screen (NIAS), with three subscales of Picky, Appetite, and Fear, each with scores ranging from 0 to 15.²³

Outcome measures included gastroparesis symptom severity and quality of life. Gastroparesis symptom severity was measured using the Patient Assessment of Upper GI Symptom Severity (PAGI-SYM) which contains the Gastroparesis Cardinal Symptom Index (GCSI; average of PAGI-SYM nausea/vomiting, fullness/satiety, and bloating/distension subscales).^24,25 Additionally, we assessed the upper abdominal pain PAGI-SYM subscale. Quality of life was measured using the Short Form 8 (SF-8), which is an 8-item questionnaire derived from the Short Form 36 (SF-36)^26,27 with a Mental Component Score (MCS) and Physical Component Score (PCS).²⁷

2.3 |. Statistical analysis

Data were compiled into Microsoft Excel and assessed for completeness. For all self-report surveys, if 75% or more items within a scale were complete, missing values were imputed by taking the average of the other items within the scale (i.e., within a subscale or within the total scale if the measure did not have a subscale). Of note, included patients completed 99 ± 2% of the VSI. Tests for normal distribution were performed to determine the need for nonparametric tests (skewness/kurtosis >2.0 or < −2.0). Statistical analysis was performed using R (version 4.1.2).

For Aim 1, we conducted a confirmatory factor analysis (CFA) of the original VSI one-factor solution, evaluated convergent validity with theoretically similar measures (HADS-A, PHQ-15, NIAS subscales) with Spearman’s rho, and calculated internal consistency for reliability. For the CFA, we used the Lavaan package in R²⁸ with robust weighted least squares mean-and variance-adjusted estimator (WLSMV) and specified items as ordinal.²⁹ We examined the fit of the original one-factor solution using comparative fit index (CFI) ≥0.95 for global model fit, root mean square error of approximation (RMSEA) ≤0.06, and standardized root mean square residual (SRMR) ≤0.08.^30,31 If there were any items with insufficient factor loadings (<0.60) or low inter-item correlations (<0.30), we ran a second model with those items removed,^32,33 but we prioritized global model fit cutoffs over factor loading and inter-item correlation cutoffs.

For Aim 2, we used a hierarchical linear regression to assess the incremental validity of the VSI score in cross-sectionally predicting GCSI total scores. For step 1, covariates including sex, age, and 4-hour gastric retention were included. Additionally, measures of depression (HADS-D), anxiety (HADS-A), and somatization (PHQ-15) were included in step 1, as these have previously been reported to be common psychological comorbidities in patients with gastroparesis.³⁴ The VSI was included in step 2 to determine whether GI-specific symptom anxiety was associated with GI symptom severity beyond covariates. Additional exploratory hierarchical linear regression models were run for each PAGI-SYM subscale (nausea/vomiting, fullness/satiety, bloating/distension) as well as upper abdominal pain to delineate how GI-specific anxiety was associated with specific gastroparesis symptoms. Similar exploratory hierarchical linear regression approaches were pursued to examine the impact of VSI on the MCS and the PCS of the SF-8, respectively.

3 |. RESULTS

A total of 100 patients with gastroparesis were studied; all completed at least 75% of the VSI and had delayed gastric emptying on gastric emptying scintigraphy. Cohort demographics are in Table 1.

TABLE 1.

Sample characteristics of 100 patients with gastroparesis.

	M (SD), range or n (%)
Age (years)	46.1 (17.3), 18–87
Sex
Female	83 (83%)
Male	17 (17%)
Race
African American	11 (11%)
Native American	0
Asian	1 (1%)
Pacific Islander	1 (1%)
White	78 (78%)
Other	3 (3%)
Unreported	6 (6%)
Ethnicity
Hispanic/Latino	6 (6%)
Non-Hispanic/Latino	94 (94%)
BMI Category
Underweight (<18.5 kg/m2)	15 (15%)
Normal weight (18.5–24.9 kg/m2)	33 (33%)
Overweight (25.0–29.9 kg/m2)	26 (26%)
Obesity (≥ 30.0 kg/m2)	26 (26%)
Etiology of Gastroparesis
Diabetic	39 (39%)
Idiopathic	18 (18%)
Postsurgical	5 (5%)
Other/Unknown	35 (35%)
Self-Report Measures
Visceral Sensitivity Index (VSI)	44.3 (19.7), 0–75
Gastroparesis Cardinal Symptom Index (GCSI)	3.1 (1.1), 0–4.8a
Short Form 8 Mental Component Score (MCS)	37.9 (12.0), 14.1–66.6
Short Form 8 Physical Component Score (PCS)	32.9 (10.9), 11.6–56.8
Patient Health Questionnaire (PHQ-15)	14.0 (5.4), 2–26
Hospital Anxiety and Depression Scale (HADS) – Anxiety	10.8 (2.1), 5–17
Hospital Anxiety and Depression Scale (HADS) – Depression	7.4 (3.7), 0–15
Nine Item ARFID Screen (NIAS) – Picky	5.7 (4.4), 0–14
Nine Item ARFID Screen (NIAS) – Appetite	9.3 (4.6), 0–15
Nine Item ARFID Screen (NIAS) – Fear	10.1 (4.0), 0–15

^aGCSI assesses symptoms related to gastroparesis in the 2 weeks prior to assessment. One patient reported a GCSI score of 0 as the patient was experiencing a symptom-free period. However, this patient reported episodes of intractable nausea and vomiting over the last several years thought to be due to gastroparesis.

3.1 |. VSI Validation

Confirmatory factor analysis revealed that the original VSI one-factor structure (all items loading onto one factor for a total score) overall fit well [χ²(90) = 220.1, p < 0.0001; SRMR = 0.08; RMSEA = 0.12; CFI = 0.96]; RMSEA did not meet our cutoff of ≤0.06, but the model as a whole had a good fit. Item 5 (“I often fear that I won’t be able to have a normal bowel movement”) and item 9 (“When I enter a place I haven’t been before, one of the first things I do is to look for a bathroom”) did not meet the ≥0.60 cutoff for factor loadings (0.53 and 0.59, respectively), however, items 5 and 9 both met the ≥0.30 cutoff for inter-item correlations (0.49 and 0.53, respectively). For sensitivity, we also ran a second model with items 5 and 9 removed and model fit was similar [χ²(65) = 187.5, p < 0.0001; SRMR = 0.08; RMSEA = 0.14; CFI = 0.96] (Table 2). Because we prioritized global model fit statistics over factor loading and inter-item correlations cutoffs, we did not remove items 5 or 9 from subsequent analyses.

TABLE 2.

Visceral Sensitivity Index items and standardized confirmatory factor analysis loadings.

		Factor loading	Inter-item r
1	I worry that whenever I eat during the day, bloating and distension in my belly will get worse.	0.73	0.62
2	I get anxious when I go to a new restaurant.	0.75	0.65
3	I often worry about problems in my belly.	0.88	0.78
4	I have a difficult time enjoying myself because I cannot get my mind off of discomfort in my belly.	0.82	0.74
5	I often fear that I will not be able to have a normal bowel movement.	0.53	0.49
6	Because of fear of developing abdominal discomfort, I seldom try new foods.	0.75	0.66
7	No matter what I eat, I will probably feel uncomfortable	0.71	0.65
8	As soon as I feel abdominal discomfort I begin to worry and feel anxious.	0.86	0.81
9	When I enter a place I have not been before, one of the first things I do is to look for a bathroom.	0.59	0.53
10	I am constantly aware of the feelings I have in my belly.	0.88	0.78
11	I often feel discomfort in my belly could be a sign of a serious illness.	0.83	0.74
12	As soon as I awake, I worry that I will have discomfort in my belly during the day.	0.82	0.75
13	When I feel discomfort in my belly, it frightens me.	0.89	0.80
14	In stressful situations, my belly bothers me a lot.	0.74	0.65
15	I constantly think about what is happening inside my belly.	0.83	0.71

The VSI also had excellent internal consistency (α = 0.94) and convergent validity (r = 0.29–0.56; all p < 0.01; Table 3).

TABLE 3.

CONVERGENT VALIDITY MEASURES.

CONVERGENT VALIDITY MEASURE	CORRELATION WITH VSI
PATIENT HEALTH QUESTIONNAIRE (PHQ-15)a	0.36***
HOSPITAL ANXIETY AND DEPRESSION SCALE (HADS) – ANXIETYb	0.34***
HOSPITAL ANXIETY AND DEPRESSION SCALE (HADS) – DEPRESSIONa	0.47***
NINE ITEM ARFID SCREEN (NIAS) – PICKYa	0.30**
NINE ITEM ARFID SCREEN (NIAS) – APPETITEa	0.42***
NINE ITEM ARFID SCREEN (NIAS) – FEARa	0.57***

Note:

^***p < 0.001,

^**p < 0.01,

^*p < 0.05.

^aPEARSON’S R CORRELATION.

^bSPEARMAN’S P CORRELATION.

3.2 |. Association of VSI with gastroparesis symptom severity

In our cohort, the mean VSI score was 44.3 ± 19.7 and the mean GCSI score was 3.1 ± 1.1. There was a moderate correlation between total VSI score and GCSI score (Figure 1, ρ = 0.51; p < 0.001).

FIGURE 1.

Scatterplot showing relations of VSI score and GCSI score. Spearman correlation coefficient of 0.51; p < 0.001.

Note: VSI: Visceral Sensitivity Index; GCSI: Gastroparesis Cardinal Symptom Index

We then assessed the incremental validity of the VSI in relation to the GCSI total score. Table 4 provides results from hierarchical regression analyses examining whether VSI scores were significantly associated with GCSI, after adjusting for covariates. In step 1 (which excluded the VSI), higher PHQ-15 and higher HADS-D scores were significantly associated with higher total GCSI scores. In step 2 (in which the VSI was added), higher VSI scores were significantly associated with higher total GCSI scores above and beyond other covariates (PHQ-15, HADS-A, HADS-D) as hypothesized (p < 0.01). In fact, in step 2, only the VSI and the PHQ-15, not the HADS-D, were significantly associated with total GCSI. We also explored hierarchical regression models with the PAGI-SYM subscales—similar results were found for the nausea/vomiting subscale, fullness/satiety subscale, and upper abdominal pain subscale, but not the bloating/distension subscale. Similar results were seen in hierarchical regression analyses using the PHQ-12 rather than the PHQ-15 which excludes three items focused on GI-specific symptoms (“stomach pain,” “constipation, loose bowels, or diarrhea,” “nausea, gas, or indigestion”).³⁵

TABLE 4.

Hierarchical regression models of GI-specific anxiety predicting gastroparesis symptom severity.

	GCSI Total	PAGI-SYM Nausea/Vomiting	PAGI-SYM Fullness/Satiety	PAGI-SYM Bloating/Distension	PAGI-SYM Upper Abdominal Pain
Model 1
Age	−0.01	−0.02*	0.00	0.00	0.01
Sex	0.37	0.43	0.27	0.47	0.35
4-HR GR	0.00	0.01	0.00	0.00	0.00
HADS-D	0.06*	0.12**	0.06	0.00	0.07
HADS-A	0.03	0.03	0.02	0.05	−0.04
PHQ-15	0.09***	0.09**	0.08**	0.09**	0.11***
Adjusted R2	0.34***	0.27***	0.17**	0.10*	0.23**
Model 2
Age	−0.01	−0.01	0.00	0.00	0.01
Sex	0.19	0.22	0.05	0.45	0.08
4-HR GR	0.00	0.01	0.00	−0.01	0.00
HADS-D	0.03	0.08	0.02	0.00	0.03
HADS-A	−0.01	−0.01	−0.02	0.04	−0.09
PHQ-15	0.08***	0.07*	0.07**	0.09**	0.10***
VSI	0.02**	0.02*	0.02**	0.00	0.03**
Adjusted R2	0.39***	0.31***	0.23***	0.09*	0.29***
ΔR2	0.05**	0.04*	0.06**	0.00	0.07**

Note: Hypothesis testing was conducted with the GCSI-Total Score model. All other models are exploratory. Values in the table are beta values. Beta values for sex are based on coding for female as 0 and male as 1.

^*p < 0.05,

^**p < 0.01,

^***p < 0.001.

Abbreviations: 4-HR GR, 4-hour gastric retention; GCSI, Gastroparesis Cardinal Symptom Index; HADS-A, Hospital Anxiety and Depression Scale - Anxiety; HADS-D, Hospital Anxiety and Depression Scale - Depression; PAGI-SYM, Patient Assessment of Upper Gastrointestinal Symptom Severity Index; PHQ-15, Patient Health Questionnaire - 15; VSI, Visceral Sensitivity Index.

3.3 |. Association of VSI with quality of life

We also explored the relationship between VSI score and both mental health-related and physical health-related quality of life (SF-8 MCS and PCS scores) using hierarchical regression analyses (Table 5). For mental health-related quality of life, step 1 showed that higher HADS-D, higher HADS-A, younger age, and female sex were all associated with lower MCS (poorer mental-related quality of life) scores. When the VSI was added in step 2, higher VSI scores were significantly associated with lower MCS scores, above and beyond other variables (p < 0.01). Additionally, in step 2, age, sex, and gastric retention no longer had significant relations to MCS scores, but higher HADS-D score was still significantly related to lower MCS score. For physical health-related quality of life, step 1 showed that greater gastric retention, higher GCSI, and higher PHQ-15 were significantly associated with lower PCS (poorer physical-related quality of life). The addition of the VSI in step 2 did not explain a statistically significant amount of variance in PCS.

TABLE 5.

Hierarchical regression models of GI-specific anxiety predicting quality of life scores.

	SF-8 MCS	SF-8 PCS
Model 1
Age	0.11*	0.03
Sex	−5.64*	4.02
4-HR GR	0.09*	−0.10*
GCSI	0.18	−3.33**
HADS-D	−1.81***	−0.42
HADS-A	−0.87*	0.43
PHQ-15	−0.20	−0.63**
Adjusted R2	0.53***	0.34***
Model 2
Age	0.10	0.02
Sex	−4.27	4.76
4-HR GR	0.09	−0.10*
GCSI	0.99	−2.89*
HADS-D	−1.58***	−0.29
HADS-A	−0.58	0.59
PHQ-15	−0.18	−0.62**
VSI	−0.15**	−0.08
Adjusted R2	0.56***	0.35***
ΔR2	0.04**	0.01

Note: Values in the table are beta values. Beta values for sex are based on coding for female as 0 and male as 1.

^*p < 0.05,

^**p < 0.01,

^***p < 0.001.

Abbreviations: 4-HR GR, 4-hour gastric retention; GCSI, Gastroparesis Cardinal Symptom Index; HADS-A, Hospital Anxiety and Depression Scale - Anxiety; HADS-D, Hospital Anxiety and Depression Scale - Depression; MCS, Mental Component Summary; PCS, Physical Component Summary; PHQ-15, Patient Health Questionnaire - 15; SF-8, Short Form 8; VSI, Visceral Sensitivity Index.

4 |. DISCUSSION

The VSI was designed to measure GI-specific anxiety in patients with IBS.¹⁶ Data are lacking on GI-specific anxiety, including potential clinical utility, in other GI functional/motility disorders including gastroparesis. In this study, we provide the first validation data to our knowledge of the VSI among adults with gastroparesis. We found that a one-factor model had adequate fit with good internal consistency and convergent validity, suggesting that all VSI items are appropriate for use with individuals with gastroparesis with one total score. Additionally, we found that GI-specific anxiety was independently associated with gastroparesis symptom severity beyond somatization, trait anxiety, and trait depression. Our finding suggests that the VSI is an overall appropriate measure of GI-specific anxiety in adults with gastroparesis and that the GI-specific anxiety is uniquely related to gastroparesis symptom severity and quality of life.

The one-factor solution for the VSI fit well, suggesting that while the VSI was originally validated in IBS,¹⁶ its use is appropriate in adults with gastroparesis. We identified two items that had lower factor loadings on the one-factor solution, both of which are the only items specific to defecation symptoms (item 5 “I often fear that I won’t be able to have a normal bowel movement” and item 9 “When I enter a place I haven’t been before, one of the first things I do is to look for a bathroom”). However, both items 5 and 9 correlated sufficiently with other items and their removal did not notably change model fit. It is possible that their inter-item correlations were sufficient because of high rates of bowel symptoms among individuals with gastroparesis.³⁶ Future studies could examine the impact of comorbid IBS on GI-symptom anxiety in gastroparesis (e.g., through analysis with both lower and upper GI symptom measures). In addition, future studies could use mixed methods approaches including use of patient focus groups to identify possible modifications to the VSI to best reflect the experiences of those with gastroparesis. Finally, while the original one-factor solution fit well in our sample, a three-factor solution for the VSI was recently identified using exploratory factor analysis³⁷ which could be explored in future larger samples. In sum, the originally created VSI appears appropriate for use in adults with gastroparesis as a measure of GI-specific anxiety, but future research is needed to understand if any modifications to the VSI would improve its utility.

The full VSI also performed well with other validity parameters—specifically with internal consistency and convergent validity. Related to convergent validity, higher VSI scores were significantly associated with higher trait anxiety (HADS-A), similar to studies in IBS (r = 0.34–0.73)^16,38 and a study of the VSI in patients with eating disorders (r = 0.41). We also found positive relationships between the VSI and somatization, depression, and symptoms of avoidant/restrictive food intake. The strongest association (r = 0.57) was with the NIAS—Fear subscale which assesses food avoidance (e.g., delaying eating, restricted food variety) due to fear about aversive consequences of eating (e.g., vomiting). As the VSI only has two items about fear around food, the relationship could suggest that both food-related and non-food-related GI-specific anxiety are significantly related and relevant in gastroparesis, but further research is needed to evaluate this relationship as well as divergent validity of the VSI in gastroparesis.

We found that the VSI is likely a clinically relevant measure in gastroparesis beyond trait anxiety. First, our sample average (44.3) was higher but within one standard deviation of VSI scores reported in the original IBS validation study (Mean = 36.5, SD = 18.5)¹⁶ and a study validating the VSI in patients with eating disorders (Mean = 43.8, SD = 17.9).²⁰ Interestingly, trait anxiety (HADS-A) was not a significant predictor of gastroparesis symptom severity either with or without GI-specific anxiety (VSI) included in our models. Previous studies have found higher trait anxiety in abdominal pain predominant gastroparesis,³⁹ idiopathic gastroparesis,⁴⁰ and general gastroparesis.⁴¹ However, these studies only found associations (and not causation) between higher GCSI scores and higher anxiety scores, which is perhaps not surprising given that more severe disease would theoretically be more worrisome for patients. Our study builds on this association by finding correlation between GI-specific anxiety, a specific form of anxiety, and gastroparesis symptom severity, with GI-specific anxiety explaining an additional 5% of variance in gastroparesis symptom severity in hierarchical linear regression beyond trait anxiety and somatization. This finding is consistent with studies on GI-specific anxiety in patients with IBS, which have shown that VSI is a strong independent predictor of IBS symptom severity¹⁶ and is associated with symptom severity even while controlling for trait anxiety and somatization.³⁷ Path analyses in patients with IBS have shown that GI-specific anxiety is a mediator of the relationship between both neuroticism and anxiety symptoms and IBS symptom severity.³⁸ Further studies are needed to understand the relationship between GI-specific anxiety and trait anxiety in gastroparesis, but the fact that GI-specific anxiety is a predictor of mental health-related quality of life at least suggests it is an important measure in patients with gastroparesis.

Given the cross-sectional nature of our study, we are not able to make interpretations about directionality—that is, it is possible GI-specific anxiety contributes to the maintenance of gastroparesis symptoms, that gastroparesis symptoms lead to heightened GI-specific anxiety, and/or a bidirectional relationship exists. GI-specific anxiety has been a consistent mediator in brain-gut behavior therapies for IBS^{5,6,13,42–44} and for pediatric functional abdominal pain/dyspepsia.^42,45 That is, reductions in GI-specific anxiety explain how brain-gut behavior therapies (compared to control conditions) reduce GI symptom severity, with reductions in GI-specific anxiety occurring before reductions in symptom severity.^5,6,42 Exposure-based cognitive-behavioral therapy for IBS also in particular seems to be most effective for individuals who have a high level of avoidance around their GI symptoms before treatment (which a few items in the VSI similarly capture).⁵ While randomized trials of brain-gut behavior therapies for gastroparesis in adults have yet to be conducted, there is promising evidence for protocols showing reductions in GI-specific anxiety measured using the VSI in open pilot designs.⁴⁶

We believe our findings supporting preliminary validation of the VSI are an important step to show that the VSI can be used to evaluate GI-specific anxiety in gastroparesis. We also showed promise for the clinical utility of the VSI—that higher GI-specific anxiety above and beyond trait anxiety and somatization was related to greater gastroparesis symptom severity and poorer mental health-related quality of life. These findings suggest there is a relationship between greater GI-specific anxiety, greater gastroparesis symptom severity, and impaired quality of life. Future longitudinal research is needed to understand the nature of the relationship, including whether targeting treatment to GI-specific anxiety may improve gastroparesis outcomes.

Key points.

• GI-specific symptom anxiety is a clinically relevant measure in gastroparesis.

• Higher GI-specific symptom anxiety is associated with increased patient-reported gastroparesis symptom severity and decreased quality of life, beyond depression, anxiety, or somatization.