Anxiety Moderates the Relationship Between Stress and Pain in IBS Patients: A Prospective Diary Study

Abstract

Background/Objectives: Irritable bowel syndrome (IBS) is a common disorder of brain–gut interaction characterized by abdominal pain and altered bowel habits. While stress and anxiety are known to exacerbate IBS symptoms, less is understood about how these factors interact on a daily timescale. This study aimed to clarify the relationship between daily stress and abdominal pain in IBS and to examine whether trait anxiety moderates this association. Methods: Forty-nine IBS patients completed daily assessments of stress and abdominal pain over a 14-day period. Participants rated abdominal pain three times daily and reported daily stress levels across seven life domains each evening. Trait anxiety was assessed at baseline using the STAI-T. Results: Hierarchical linear modeling was used to analyze within-person and between-person effects. An increase in between-person stress was associated with increased probability of abdominal pain among individuals with low-to-moderate trait anxiety, while this was not observed in patients with high trait anxiety. Even though within-person (day-to-day) stress variations had an impact on pain probability, the effects of between-person variations were multiple times greater. Conclusions: These findings suggest that the interplay between stress and anxiety in IBS might not be uniform. High trait anxiety may, under certain conditions, attenuate rather than amplify the link between stress and pain, possibly pointing to a more dynamic relationship.

1. Introduction

Irritable bowel syndrome (IBS) is a prevalent functional gastrointestinal disorder characterized by altered bowel habits and recurrent abdominal pain [1]. Its prevalence in developed countries ranges from 4.6% to 9.0%, depending on diagnostic criteria [2]. IBS is now widely classified as a disorder of brain–gut interaction (DGBI), involving disrupted communication between the central nervous system and the gastrointestinal (GI) tract. This dysregulation involves multiple pathways, including hypothalamic–pituitary–adrenal (HPA) axis dysfunction, impaired pain modulation, altered attention to visceral signals, and visceral hypersensitivity [3,4].

Abdominal pain is the hallmark symptom of IBS and a core contributor of functional impairment [1]. Although typically episodic and related to bowel movements, IBS-related pain shares neurobiological mechanisms with other chronic pain conditions. Central sensitization, or heightened sensitivity to nociceptive input in the central nervous system, has been proposed as one such mechanism, particularly in individuals with comorbid anxiety or mood disorders [5]. Pain is further maintained by maladaptive behavioral and cognitive patterns such as hypervigilance, avoidance, and negative affective appraisals, which can reinforce a reciprocal loop between pain and anxiety [6,7].

IBS patients frequently report stress as a trigger for symptom flare-ups. Empirical findings, however, are mixed. While some studies support a positive association between stress and IBS symptom severity [8,9,10], others have found that symptoms themselves may generate stress [11] or that the relationship between stress and pain is inconsistent or absent [12,13,14]. These discrepancies suggest that individual-level factors may moderate how stress influences symptom expression.

One such factor is anxiety, which plays a central role in the persistence and intensity of IBS symptoms. Compared to healthy individuals, IBS patients consistently report higher levels of both trait and GI-specific anxiety [15,16], and many meet the criteria for comorbid anxiety disorders [17]. Trait anxiety reflects a broad, stable tendency to respond to a wide range of situations with heightened worry and emotional reactivity [18], while GI-specific anxiety captures fears, attentional focus, and avoidance behaviors specifically related to GI sensations and contexts [19]. Both have been associated with greater pain sensitivity and symptom severity in IBS, possibly acting via different mechanisms [19,20,21,22,23,24,25]. Cognitive–affective traits such as catastrophizing and somatization further exacerbate pain responses. Catastrophizing, in particular, has been shown to mediate the relationship between worry and pain and is more common among IBS patients with higher anxiety and more severe symptoms [26,27]. At the same time, it seems that the effects of anxiety on pain are not uniformly amplifying. Several studies suggest that anxiety’s influence is highly context-dependent: it may enhance or reduce pain perception depending on the emotional and attentional relevance of the anxiety to the pain stimulus [28,29,30,31]. This nuanced relationship has direct implications for understanding daily symptom variation in IBS patients.

Despite extensive research on anxiety and stress in IBS, the daily dynamics of these processes remain under-investigated. Little is known about how momentary or sustained stress influences pain day to day, or how stable psychological traits such as anxiety might moderate this relationship. The current study addresses this gap by examining the association between daily stress and abdominal pain in individuals with IBS, with a particular focus on the moderating role of trait anxiety. Given that the daily stressors assessed in this study were broad and not specific to GI concerns, trait anxiety was selected as the most appropriate measure of general proneness to anxiety. Based on existing literature and clinical patterns, we hypothesized that higher daily stress would be associated with increased pain, particularly among those with elevated trait anxiety.

2. Results

Descriptive statistics for both between-person and within-person level measures are presented in

Table 1. Between- and within-person descriptives for trait anxiety, daily stress, and abdominal pain occurrence.

Variable	Between-Person		Within-Person
	Mean	SD	Min	Max	SD
Trait Anxiety	2.32	0.48	1.038	3.212
Daily stress	0.26	0.28	0.021	1.224	0.25
Pain occurrence	0.52	0.30	0.001	0.976	0.38

Means, standard deviations (SD), and minimum/maximum values are presented for trait anxiety, daily stress, and pain occurrence. Between-person values reflect participant-level averages across the 14-day sampling period. Within-person standard deviations represent variability in daily scores within individuals. Trait anxiety was measured once at baseline; daily stress and pain occurrence were assessed repeatedly across 14 days.

. Trait anxiety was approximately centered around the theoretical midpoint of the scale, with the expected variability between individuals. The mean pain occurrence was about 50% (proportion = 0.52), with substantial variability both between and within individuals. Daily stress levels were generally low on average but showed pronounced variability both across individuals and within days.

Models of increasing complexity were compared using likelihood ratio tests, and the results are summarized in

Table 2. Fit indices of the tested models and significance testing of the increased model complexity.

Models	df	AIC	BIC	Log-Likelihood	χ2	df(χ2)	p	Marginal R2
Model 0 (main effects)	6	2049.52	2083.05	−1018.76				0.074
Model 1 (+mean stress×anxiety)	7	2048.03	2087.15	−1017.02	3.49	1	0.06	0.105
Model 2 (+daily stress variation×anxiety)	8	2049.54	2094.24	−1016.77	0.49	1	0.48	0.105

Note: df—degrees of freedom; AIC—Akaike Information Criterion; BIC—Bayesian Information Criterion; Marginal R²—effect size of fixed effect.

. The model including only main effects demonstrated a relatively poor fit, R² = 0.074 (marginal). Adding the interaction between daily mean stress levels and anxiety resulted in a marginally significant improvement, p = 0.06, with a substantial increase in model fit, R² = 0.105 (marginal). The addition of the interaction between daily stress variability and anxiety did not significantly improve model fit.

Model 1 parameters are summarized in

Table 3. Estimated coefficients, standard errors, odds ratio, significance estimates, and variance components of best-fitting model.

	Estimate	S.E.	OR	z	p	Variance
Intercept	−0.7189	0.3828	0.4873	−1.878	0.060	3.893
Between-person stress	3.3675	1.1318	29.0069	2.975	0.005
Within-person stress	0.5284	0.2543	1.6963	2.078	0.038
Anxiety	1.7026	0.8241	5.4879	2.066	0.039
Between-person stress×Anxiety
				Conditional R2	0.542
				Marginal R2	0.105

.

The predicted odds of experiencing pain at mean anxiety levels and zero stress were 0.49, indicating a lower likelihood of pain in this context. Increases in between-person stress had the strongest effect on pain probability: a unit increase in between-person stress (from minimum to maximum) was associated with a 29-fold increase in the odds of pain. Within-person stress variation also had a significant, though smaller, effect: a unit increase (relative to the person’s mean) was associated with a 1.69-fold increase in the odds of pain. Between-person anxiety showed a slightly stronger effect, with the odds of pain increasing 5.49 times for a unit increase in anxiety. The interaction between between-person stress and anxiety was marginally significant and appeared to counterbalance the main effect of stress.

To illustrate these patterns more clearly, Figure 1 presents predicted pain probabilities from the logistic model, conditional on between-person levels of stress and anxiety.

As shown, the effect of between-person stress was most evident at lower levels of dispositional anxiety. At higher levels of dispositional anxiety, the effect of between-person stress diminished, and pain occurrence probability remained low and largely unrelated to stress.

3. Discussion

This study aimed to explore the daily relationship between stress and abdominal pain in IBS patients and examine whether trait anxiety moderates this relationship. In particular, we differentiated between the effects of long-term stress (two-week average) and day-to-day stress variation. The results revealed an association between average (between-person) stress and pain probability: for most IBS patients, higher levels of long-term stress were associated with a higher likelihood of experiencing abdominal pain.

There are two ways in which this finding can be linked to existing data. First, we can consider this two-week stress average as an indicator of chronic stress. In this sense, our finding is consistent with a large body of research showing that chronic psychosocial stress is a key contributor to the exacerbation of IBS symptoms [32]. Chronic stress is thought to sustain low-grade inflammation, alter gut motility, and sensitize visceral afferents, all of which increase symptom frequency and severity [33]. Our finding reinforces this framework and underlines the importance of chronic stress management in IBS treatment.

Second, comparing the two-week average of stress to daily fluctuations highlights the distinction between interindividual and intraindividual variability in stress levels. Our findings suggest that even though both two-week averages and daily variations of stress are significant predictors of pain occurrence, the between-person differences in perceived stress, or how much stress individuals tend to report or experience on average, plays a more significant role in explaining pain outcomes than the within-person day-to-day changes. Between-person differences in stress had a substantially greater impact on abdominal pain probability, with a magnitude more than fifteen times that of within-person daily stress variation. In other words, a patient’s typical stress level over time appears more predictive of their pain experience than daily ups and downs. This aligns with prior research indicating that dispositional factors such as coping style, cognitive appraisal, and affective traits, particularly worry and pain catastrophizing, shape how individuals process and react to stress [34,35]. These stable cognitive–emotional patterns may influence not only the intensity of perceived stress but also its physiological and behavioral consequences, thereby amplifying pain responses in patients with IBS. This perspective underscores the importance of targeting these trait-like vulnerabilities in interventions rather than focusing solely on short-term stress fluctuations. Indeed, research shows that a reduction in pain catastrophizing after cognitive-behavioral therapy (CBT) is related to improvements in symptom severity and quality of life in IBS patients [36]. Also, changes in cognitive flexibility after CBT were associated with a reduction in symptom severity, pain, and quality of life in IBS patients [37]. This supports the point that trait-level vulnerabilities such as worry, catastrophizing, coping style, and cognitive flexibility are stronger predictors of IBS pain and symptom severity than day-to-day stress fluctuations. It seems that interventions like CBT that target these dispositional cognitive–emotional processes exert significant effects on symptom improvement.

Our results indicate that the relationship between stress and abdominal pain in IBS is moderated by trait anxiety. Among individuals with low-to-moderate trait anxiety, higher average stress was strongly associated with increased pain probability. In contrast, those with high trait anxiety showed consistently elevated pain probability regardless of stress levels, and notably, when average stress was high, these individuals reported lower pain probability than those with moderate or low anxiety. This suggests that for high-anxiety patients, average stress does not predict pain in the typical way. Such a pattern challenges the common assumption that anxiety and stress always act synergistically to exacerbate pain. Instead, it points to a non-linear or conditional interaction between these psychological factors in the context of IBS.

One plausible interpretation of these findings centers on cognitive mechanisms of attention and worry. Prior research indicates that the impact of anxiety on pain perception depends on the relevance and focus of anxiety-related thoughts [30]. When anxiety is directed toward bodily symptoms (e.g., GI-specific anxiety), it tends to increase pain sensitivity. In contrast, when anxiety is focused on external or unrelated stressors, it may actually distract from somatic sensations, leading to attenuated pain perception, at least in the case of acute experimentally induced pain [28,31]. For instance, anxiety focused on a pain-irrelevant task has been shown to produce lower pain ratings [38]. Similarly, cognitive stressors such as mental arithmetic tasks, also unrelated to bodily symptoms, can lead to reduced pain perception, especially in individuals with high anxiety [39]. Some studies further suggest that the intensity of anxiety, rather than its relevance to pain, may be the critical factor, with low-to-moderate anxiety enhancing pain and high anxiety producing hypoalgesia [31]. In our study, daily stressors were assessed across a broad range of domains, including work, interpersonal relationships, finances, and health, only one of which was explicitly health related. Most of these stressors were thus likely to be pain-irrelevant. For individuals with high trait anxiety, these non-GI stressors may have engaged attentional systems in ways that diverted cognitive resources away from gut-related discomfort. Cognitive–affective models of pain propose that pain demands attention and competes with other salient stimuli [40]. When external stressors are emotionally intense or perceived as more urgent, they may override internal bodily cues, especially in individuals predisposed to chronic worry. In this way, the combination of high trait anxiety and non-pain-related stressors may paradoxically reduce pain reporting by shifting attention elsewhere.

Although some experimental findings are consistent with our results, it is important to recognize key differences in scale and design. For example, Hoeger Bement et al. [39] found that individuals with low baseline stress and anxiety tended to show increased pain following a cognitive stressor, whereas those with high baseline stress and anxiety exhibited reduced pain. This hypoalgesic effect was linked to anticipatory increases in cortisol, suggesting that HPA axis activation plays a role in modulating pain sensitivity. More broadly, IBS-related pain modulation reflects a complex neurobiological profile that extends beyond HPA dysregulation. IBS patients show evidence of central sensitization, impaired descending inhibitory control, and heightened brain responses to visceral input [4,5,17]. These alterations often co-occur with elevated trait and GI-specific anxiety, which have been consistently associated with increased pain sensitivity and diminished quality of life [15,24].

While experimental studies differ methodologically from our own—which relied on daily self-reports over a longer time frame—findings from ecological momentary assessment (EMA) studies offer a closer parallel. Several EMA studies have also reported seemingly paradoxical patterns in the stress–pain relationship. For instance, Chan et al. [9] found that daily stress predicted lower abdominal pain in a subset of IBS patients, while Engel et al. [41] observed that increases in abdominal pain were sometimes followed by decreases in nervousness. These findings align with cognitive–affective models emphasizing attentional resource shifts: when stress is unrelated to somatic symptoms, attention may be diverted away from interoceptive signals, resulting in reduced pain perception [40]. It is important to note that this does not negate the well-established link between anxiety and symptom exacerbation in IBS. Rather, it highlights that this relationship may be dynamic and context-dependent, shaped by the nature, timing, and content of stressors.

Our findings may point to a more complex, non-linear relationship between stress and pain in IBS and raise the possibility that under certain conditions, higher anxiety could buffer rather than exacerbate pain responses. This highlights the need to distinguish between types of anxiety (trait, GI-specific, situational), the nature of stressors (internal vs. external), and their relevance to pain. The nuanced interplay between stress, anxiety, and pain in IBS underscores the importance of tailored interventions that consider not only symptom severity but also individual cognitive–emotional profiles, such as CBT. Future research should explore these mechanisms using physiological stress markers (e.g., cortisol) and examine whether stressor content (e.g., pain-relevant vs. -irrelevant) predicts divergent pain outcomes in IBS.

Taken together, these findings, though exploratory, may have implications for clinical practice. They emphasize the value of addressing stable, trait-like psychological characteristics in the treatment of IBS, particularly those related to chronic stress perception. While daily stress fluctuations contribute to symptom variability, our results suggest that a patient’s typical stress level over time is far more predictive of pain occurrence than momentary changes. These results support the use of CBT, particularly approaches that target chronic cognitive and emotional patterns such as pain catastrophizing, chronic worry, emotional avoidance, and low cognitive flexibility, which is in line with previous studies [36,37].

Limitations

Several limitations should be considered when interpreting the results of this study. While the number of participants was modest (N = 49), the hierarchical linear modeling approach allowed us to capitalize on the richness of the data, with each participant contributing 14 daily stress ratings and up to 42 abdominal pain ratings. This multilevel framework enhanced statistical power by modeling within-person variability and accounting for the nested structure of repeated measurements. However, despite this analytic strength, the sample was recruited from a single clinical site and was relatively demographically homogeneous, which may introduce selection bias and limit the generalizability of findings across different age groups, cultural contexts, or healthcare settings. Additionally, the use of self-reported measures for anxiety, stress, and pain may contribute to inflated or obscured associations due to shared method variance and overlapping symptom content. For instance, constructs like anxiety and stress often involve physiological or affective descriptors—such as tension, fatigue, or somatic discomfort—that can also appear in reports of pain. This conceptual overlap may artificially strengthen observed relationships or mask more nuanced patterns [42]. Without objective or multimodal assessment, it is difficult to disentangle whether associations reflect true underlying mechanisms or artifacts of measurement.

Additionally, the assessment of daily stress was based on subjective intensity ratings of broad life domains without contextualizing the specific content or emotional relevance of each stressor. As such, we could not directly test whether stressors were pain-related or pain-irrelevant—an important distinction given our interpretation of stress-induced hypoalgesia. Furthermore, while pain was measured three times daily, stress was assessed only once per day (in the evening), making it difficult to determine the precise temporal relationship between stress exposure and pain reporting. This mismatch may lead to biased estimates due to lag effects, as the stress measurement may not reflect stress levels at the exact time pain was reported. Additionally, the daily stress measure likely relies on retrospective recall, which may smooth or obscure acute stress fluctuations influencing pain. Future studies should adopt fully synchronized momentary sampling protocols to better capture real-time stress–pain interactions and reduce potential biases from lag and recall effects. Finally, the study relied exclusively on self-report data and did not include physiological markers of stress (e.g., cortisol levels) or objective measures of autonomic reactivity, which could have clarified the underlying mechanisms driving the observed effects.

Finally, although the interaction between average stress and trait anxiety did not meet conventional thresholds but rather approached statistical significance (p = 0.06), the increase in prediction quality of the model, which includes the interaction term, appears to be substantial. Also, the observed trend is consistent with prior theoretical models and ecological findings and should be viewed as a promising basis for further research. Replication in larger and more diverse samples, using multimodal stress and symptom assessment, will be essential to clarify the robustness and generalizability of these effects.

4. Materials and Methods

4.1. Participants

Forty-nine patients diagnosed with IBS according to Rome III criteria participated in the study. They were recruited from the Department of Gastroenterology at the Clinical Hospital Centre in Rijeka. The sample included 11 men and 38 women, aged 18 to 69 years (M = 45.11, SD = 14.01). Most participants were married (53.1%), were employed (57.1%), and had completed high school (71.4%). Regarding symptom subtype, 56% of participants were unable to identify a predominant symptom pattern, while 20.4% reported diarrhea-predominant IBS and another 20.4% reported constipation-predominant IBS.

Sample size was determined using a power analysis adjusted for multilevel data structure [43,44]. Assuming an approximate between-person (trait-level) effect size of 0.50 [8,45] and a within-person (state-level) effect size of 0.30 [46,47], a power of 0.80, and a two-tailed alpha of 0.05, the minimum required sample size was estimated at 28.25 for Level 2 (between-person) and 84.07 for Level 1 (within-person). After correcting for nested data, the required number of Level 1 observations increased to 521. Our final sample met these requirements, with 49 participants at Level 2 and 686 observations at Level 1.

4.2. Measures

Trait anxiety was assessed using the Trait subscale of the State-Trait Anxiety Inventory (STAI-T) [18]. This 20-item questionnaire evaluates a person’s general tendency to perceive situations as threatening. Items are rated on a four-point Likert scale. The STAI-T has demonstrated high internal consistency (Cronbach’s α = 0.91).

Abdominal pain was measured using one item from the 8-item symptom diary [48], which was adapted from the IBS Symptom Diary [49]. The diary assessed abdominal pain, abdominal sensitivity, constipation, diarrhea, bloating, nausea, flatulence, and belching. Participants completed the diary three times daily (morning, afternoon, evening) for 14 consecutive days, rating symptom intensity on a scale from 0 (not present) to 4 (debilitating) based on the period since their previous entry. Due to low variability, the pain scale was consequently categorized in two categories (0: no pain; 1: pain)

Daily stress was evaluated using the Daily Stressful Events Scale [48], assessing stress across seven life domains: work, finances, living conditions, health, romantic relationships, friendships, and family. Each evening, participants rated the intensity of stressful events experienced that day in each category on a 5-point scale (1 = slightly stressful, 5 = extremely stressful). Reliabilities for this repeated-measure/multiple-time-point scale was estimated according to suggestions [50]. The reliability averaged over time was 0.92.

4.3. Procedure

Initially, participants completed baseline questionnaires, including the STAI-T, and received instructions for daily diary completion at home. Immediately following that, over a 14-day period, participants completed the symptom diary three times per day: within two hours of waking, in the early afternoon, and within two hours before bedtime. Daily stress ratings were recorded once per day in the evening. Participants received SMS reminders for each diary entry timepoint.

4.4. Data Analytic Approach

For each participant, we collected 14 data points for daily stress and up to 42 (3 × 14) datapoints for abdominal pain. The primary aim of the present study was to estimate the effects of daily stress and anxiety on the onset of abdominal pain. Given the nested structure of the data (i.e., repeated measurement points nested within individuals and days), analyses were conducted using a hierarchical linear modeling framework. Specifically, generalized linear mixed models (GLMMs) were fitted using the lme4 package [51] for the R statistical environment [52], using maximum likelihood estimation.

The outcome variable—abdominal pain—was binary (presence = 1, absence = 0). Thus, models were specified with a binomial distribution and a logit link function (i.e., hierarchical logistic models). Prior to analysis, all predictor variables were scaled as the mean of items measuring this trait/state. Disaggregation of within-person fluctuations from stable between-person differences of stress was achieved by computing both the individual’s mean level of stress (interindividual effects) and the person-centered daily deviations from this mean (intraindividual effects), following the approach outlined by [53]. To facilitate the interpretation of interaction effects, dispositional anxiety (measured via the STAI) was grand-mean centered. To investigate whether the effects of daily stress varied depending on levels of trait anxiety, interaction terms were included in the models. Specifically, we tested for interactions between anxiety and both the interindividual and intraindividual components of stress. Reliabilities for repeated-measure/multiple-time-point scales were estimated according to suggestions [50].

Model comparisons were conducted to assess the significance of these interaction effects. The initial model included the main effects of anxiety, average 14-day stress (i.e., between-person effects), and within-person day-to-day variations of stress (model 0); a more complex model additionally included the interaction of between-person stress and anxiety (model 1), while the most complex one additionally included the interaction between anxiety and within-person day-to-day stress variation (model 2). Model fit and significance of added parameters were evaluated through likelihood ratio tests. To more easily estimate the models’ effect size, a pseudo-R² measure for mixed-effect models [54,55].

5. Conclusions

This study offers tentative findings that indicate that the relationship between stress and abdominal pain in IBS might vary depending on trait anxiety. While daily stress was positively associated with pain among individuals with low-to-moderate anxiety, this pattern was not observed in those with high anxiety. These findings suggest that the interaction between stress and anxiety in IBS might be more complex and context-dependent than previously assumed. Although the effects were modest and not statistically robust, they highlight a potentially important direction for future research. Further studies with larger samples and more detailed measures of stress and physiology will be valuable in clarifying these associations.