Which psychological factors exacerbate Irritable Bowel Syndrome? Development of a comprehensive model

Abstract

Objective

There is evidence that psychological factors affect the onset, severity and duration of Irritable Bowel Syndrome (IBS). However, it is not clear which psychological factors are the most important and how they interact. The aims of the current study are to identify the most important psychological factors predicting IBS symptom severity and to investigate how these psychological variables are related to each other.

Methods

Study participants were 286 IBS patients who completed a battery of psychological questionnaires including neuroticism, abuse history, life events, anxiety, somatization and catastrophizing. IBS severity measured by the IBS Severity Scale was the dependent variable. Path analysis was performed to determine the associations among the psychological variables, and IBS severity.

Results

Although the hypothesized model showed adequate fit, post hoc model modifications were performed to increase prediction. The final model was significant (Chi² = 2.2; p=0.82; RMSEA < .05) predicting 36% of variance in IBS severity. Catastrophizing (Standardized coefficient (β)=0.33; p <.001) and Somatization (β=0.20; p <.001) were the only two psychological variables directly associated with IBS severity. Anxiety had an indirect effect on IBS symptoms through catastrophizing (β=0.80; p <.001); as well as somatization (β=0.37; p <.001). Anxiety, in turn, was predicted by neuroticism (β=0.66; p<.001) and stressful life events (β=0.31; p<.001).

Conclusion

While cause-and-effect cannot be determined from these cross-sectional data, the outcomes suggest that the most fruitful approach to curb negative effects of psychological factors on IBS is to reduce catastrophizing and somatization.

Introduction

Irritable Bowel Syndrome (IBS) is a common, chronic gastrointestinal disorder, characterized by recurring episodes of abdominal pain associated with altered bowel habits¹. There is no consensus on the etiology of IBS but biological, psychological, and sociological factors are all believed to contribute to the onset, severity, and natural history of the disorder. Psychological factors appear to play particularly important roles as moderators of symptom severity, symptom persistence, decisions to seek treatment, and response to treatment^2-4, but it is not clear which psychological factors are the most important in explaining these outcomes. In a recent review 23 different psychosocial factors were identified that are associated with IBS⁵, indicating the richness of this literature and the variability in factors identified. The psychological concepts most commonly associated with IBS are listed below:

Stress

The effect of stress on IBS is almost universally recognized by clinicians and patients. IBS symptoms wax and wane with daily stress^{6, 7} and IBS patients report more lifetime stressful events than healthy controls⁸. There is particularly strong evidence for the role of early life stressors such as sexual abuse and maternal separation in IBS^9-12. Dysfunctional brain-gut interactions have been found in maternally separated rodents-an often studied model of early life stress in IBS¹³. Moreover IBS patients show greater reactivity to stress¹⁴; that is, compared to healthy controls, the same exposure to stress leads to a greater physiological gut response in IBS patients.

Personality

Certain personality traits or temperament characteristics may make one more vulnerable to the effects of stressors. A widely studied personality factor is neuroticism¹⁵, which describes people who readily experience negative affect. Subjects high in neuroticism are more reactive to stress and have stronger reactions to recurring problems. Neuroticism is one of the few personality traits that has been consistently found to be increased in IBS patients compared to controls^16-20.

Coping

The way patients cope with stress and pain mediates health outcomes²¹. One of the most robust predictors of pain intensity is a coping strategy named Pain Catastrophizing, defined as a maladaptive way of coping (or not coping) with pain by magnifying the threat or seriousness of pain and feeling helpless to do anything about it²². Catastrophizing is associated with more intense pain and greater disability in pain patients, including those who suffer from IBS^22-25.

Psychological distress refers to feeling anxious and depressed. These symptoms are more frequent and more intense in IBS patients, and they are associated with more gastrointestinal symptoms, disability and quality of life impairment^{7, 26-31}. In 30% to 90%^32-34 of IBS patients, psychological symptoms are so severe that co-morbid psychiatric disorders can be diagnosed. The association between psychological distress and IBS seems to be bidirectional in nature: psychological distress both precedes the onset of IBS^{35, 36}, and is aggravated by the challenges of managing a chronic gastrointestinal disorder³⁷.

Somatization refers to the psychological tendency to report multiple physical symptoms. Somatization is frequently seen in IBS patients³², many of whom receive diagnoses of other functional gastrointestinal disorders, chronic pain syndromes, and symptoms such as chronic fatigue, frequent urination, bad breath and heart palpitations. The overlap between IBS and these other co-morbid disorders and symptoms does not appear to be explained by a common pathophysiology³⁴. Instead people high in somatization are thought to be hypervigilant to noticing somatic sensations and to attach disease significance to these symptoms^{32, 34}.

It has been established that these psychological variables play a role in IBS, but it is essential to determine the relative strength of their contribution to the waxing and waning of IBS as this will suggest which psychological factors should be targeted in treatment. Besides the direct effects these psychological factors have on IBS outcomes, they are also associated with each other. For example, the effect of neuroticism on IBS seems to be mediated by anxiety¹⁹ and in functional dyspepsia it has been shown that psychological distress is associated with somatization³⁸. Thus, we devised a model (see Figure 1) in which these associations between psychological factors were taken into account. Specifically, we hypothesize that stressful life events, abuse history and neuroticism aggravate maladaptive coping, anxiety, and somatization, which in turn influence IBS symptom severity. The aims of the current study were to test our model, to identify the most important psychological factors predicting IBS symptom severity, and to investigate how these psychological variables are related to each other.

Figure 1.

Proposed model of psychological effects on IBS severity

Note: Direct effects between ‘fairly stable psychological factors’ and IBS severity were tested but for clarity not depicted in the current model

Methods

Study Design

This study analyzes data from a study on the pathophysiology of IBS^{39, 40}. For this study subjects were admitted to a research clinic at the University of North Carolina for a 24- to 30-h period. On the day of admission the questionnaires described below were completed. Data from all participants were used in the current analyses.

Subjects

Subjects were recruited by advertisements or physician referrals and screened by telephone. The study was approved by the institutional review board of the University of North Carolina (UNC) and all subjects provided informed consent. The study population consisted of patients with a physician diagnosis of IBS who also met Rome II or III criteria for IBS (depending on time of enrollment, 77% of patients were screened with Rome II criteria) and had current symptom activity (abdominal pain at least once a week in the past month). These subjects had no history of gastrointestinal resection (other than appendectomy or cholecystectomy), known Inflammatory Bowel Disease, coeliac disease, lactose malabsorption, heart disease, or diabetes mellitus, and they were not pregnant at the time of study.

Measures

IBS severity

The Irritable Bowel Syndrome Severity Scale (IBS-SS) is a validated questionnaire containing five questions which address severity and frequency of abdominal pain, bloating severity, dissatisfaction with bowel habits, and IBS impact on everyday activities⁴¹. All questions are scored on a scale ranging from 0–100 and contribute equally to the total score. Previous studies have established that scores below 175 represent mild IBS symptoms, 175–300 represents moderate severity, and scores above 300 represent severe IBS⁴¹.

Sexual and physical abuse

Abuse history was measured by the Sexual and Physical Abuse History Questionnaire developed for use in functional gastrointestinal disorders by Leserman and Drossman⁴². Patients are asked to indicate if they have ever been sexually abused (had someone touch their sex organs against their will, were made to touch someone else’s sex organs, were forced to have sex) or physically abused (having been hit/kicked/beaten or someone seriously threatened their life). A subject is assumed to have a history of abuse if they reported ‘yes’ to at least one of these items over their life time.

Life events

The Family Inventory of Life Events scale (FILE)⁴³ is a measure of individual stressful life events as well as life events within the immediate family (e.g., spouse, children). The FILE records 67 life events within a person’s nuclear family over the past year. Events are included that require a significant change or adjustment within the family such as ‘mother giving birth’, ‘family member assaulted or robbed’, or ‘family member quit work due to health reason’. A total score of cumulative stressful life events was obtained by counting the number of reported events.

Neuroticism

The NEO personality inventory¹⁵ consists of 5 subscales, but only the Neuroticism subscale was used in this study. Forty-eight items are included in this subscale which ask about negative emotions (e.g., susceptibility to experience anger, guilt or anxiety) and stress reactivity. All items are answered on a 5 point scale and summed to obtain a total score.

Anxiety

Anxiety was measured by the anxiety subscale of the Brief Symptom Inventory-18 (BSI-18)⁴⁴. The subscale consists of 6 items asking about anxious cognitions (e.g., “feeling fearful”) over the past 7 days. These items are rated on a 5-point scale ranging from “not at all” to “extremely”.

Somatization

Somatization was determined with the Brief Symptom Inventory-18 (BSI-18)⁴⁴. The 6-item somatization subscale asks the subject to rate how often they were bothered by somatic symptoms such as “Feeling weak in parts of your body” or “Trouble getting your breath”. The somatization subscale of the BSI contains one gastrointestinal symptom item, nausea or upset stomach; we excluded this item from the scale.

Catastrophizing

The catastrophizing scale (6 items) of the Coping Strategies Scale⁴⁵ was used. This scale reflects feelings of hopelessness and the expectation that pain will become worse. It shows good reliability and validity in a sample of pain patients⁴⁵ and has been used in studies on IBS^{24, 25}.

Data analysis

Path analysis was implemented with IBM® SPSS® AMOS 19.9, which was used to test the model shown in Figure 1. Path analysis allows for testing all associations between factors at the same time. Chi-square was used for an overall test of the model (p > .05 denotes significance). A significant model can be improved by removing non-significant pathways. We trimmed non-significant paths one by one, each time testing if cutting would create a non-significant overall model. If the overall model became non-significant the non-significant pathway would not be removed.

We fit separate models using the total IBS-SS scores as well as individual IBS-SS scores on pain frequency and disability but all three dependent variables generated largely the same significant paths, so the model for overall IBS-SS score is depicted here. Goodness of fit was determined by the Root Mean Square Error of Approximation (RMSEA) with <.05 denoting close fit.

To test the robustness of our trimmed model we tested if alternative models yielded a better goodness of fit index. Alternative models were created by substituting reciprocal direct effects for other types of paths. For example, if the final model specified Anxiety → Somatization, alternative models were tested in which this association was substituted by Somatization → Anxiety or Anxiety ↔ Somatization. With 7 variables in the model we have sufficient sample size to test the model (N should be at least 10 to 20 times the number of variables in the model)⁴⁶.

Results

Sample characteristics

A total of 286 IBS patients participated. Mean age was 34.6 years (sd=11.7; age range 18–73). Participants were predominantly female (81.5%), and race was Caucasian (71.3%), African American (19.2%), or Hispanic (3.5%). IBS severity was predominantly moderate (43%) to high (39.5%). Means and standard deviations of all variables in our model are given in Table 1.

Table 1.

Variables in the model

	M (SD) or %	Minimum and maximum score in the sample
Irritable Bowel Severity	270.0 (90.6)	50–500
Anxiety	51.7 (9.9)	38–81
Somatization	50.8 (9.0)	41–79
Catastrophizing	8.2 (7.9)	0–35
Neuroticism	63.6 (26.6)	1–99
Life events	9.3 (6.6)	0–31.4
History of abuse	54.3%	n/a

Path Model

The assumptions for path analysis were evaluated. One outlier (defined as z-scores scores in excess of 3.29⁴⁷) was found for the FILE score and one for the Catastrophizing score; these two subjects were deleted from the analysis. Normality was not violated (except for abuse which was a dichotomous variable). Missing values were replaced by maximum likelihood estimates.

The hypothesized model showed adequate fit to the data: Chi² (df=1)= 0.29; p=0.56; RMSEA < .05. Post hoc model modifications whereby non-significant associations were removed, were performed to develop a trimmed model. The final model is given in Figure 2 (Chi² (df=5)= 2.2; p=0.82; RMSEA < .05. Several alternative models –which included small model modifications- were tested and all these models became nonsignificant.

Figure 2.

Final model of psychological effects on IBS severity

Note: Numbers reflect unstandardized (standardized)coefficients. * P < .01, ** P < .001.

As can be seen in Figure 2, catastrophizing and somatization were the only two psychological variables directly associated with IBS severity. The significant positive coefficients indicate that increases in catastrophizing and somatization were associated with increases in IBS symptom severity. Additionally, anxiety had an indirect effect on IBS symptoms through catastrophizing as well as somatization -- while anxiety was in turn predicted by neuroticism and life events. Thus, the model shows that neuroticism and stressful life events increase anxiety. The effect of anxiety on IBS severity is mediated by catastrophizing and somatization.

Additional associations in the model were: (a) A positive effect of catastrophizing and life events on somatization; and (b) a negative effect of catastrophizing on anxiety. Combined these variables predicted 36% of the variance in IBS severity.

Discussion

The aim of the current study was to develop a model of psychological influences on IBS symptoms that can guide research and therapy. We observed that the two most important variables associated with IBS severity are the maladaptive coping style referred to as catastrophizing and somatization. The effects of all other psychological variables seem to be explained by these two factors. However, anxiety had an important intermediate place in the model because it was associated with both increases in catastrophizing and somatization. The role of other psychological variables such as neuroticism and life stressors was limited to being associated with increases in anxiety. Although the model may suggest cause-effect associations, we acknowledge that the current study is correlational so we cannot claim that any of the psychological factors are causally related to each other or to IBS outcomes. Prospective longitudinal studies are needed to confirm the model.

The model suggests that the most fruitful approach to curb the effects of psychological factors on IBS is to reduce catastrophizing and somatization. For catastrophizing, the common adage to ‘reduce stress’ is applicable. But rather than reducing stress by diminishing exposure to stressors (such as certain job and family responsibilities, which may be hard to avoid) the aim should be to improve coping with stress. There are existing treatment techniques, largely used within Cognitive Behavioral Therapy (CBT), that are focused on replacing maladaptive coping strategies, specifically catastrophizing, with more positive cognitions and behaviors. In fact, CBT has been found to be an effective treatment for IBS, not only reducing IBS symptoms but also reducing somatization and catastrophizing as well as improving mood^48-50. As expected from our model, most studies have found that treatment outcomes with CBT are not mediated by improvements in mood^{48, 50-52} although one study did find evidence for this⁵³. Rather, two treatment studies have confirmed the important role of coping in reducing IBS symptoms^{54, 55}. This supports our model, giving coping a more important role in predicting IBS outcomes than anxiety.

Besides a focus on improving coping, our model suggests that reducing somatization should be effective in ameliorating IBS symptoms. In our model, somatization was included as a predictor of IBS severity, but it may also be considered an outcome by itself. IBS patients can suffer from several bodily symptoms including gastrointestinal and non-gastrointestinal symptoms. Not surprisingly the same psychological factors in our model that predict IBS severity also predict somatization. These include anxiety, stressful life events, and catastrophizing. By reducing these factors, we may not only reduce IBS symptoms but also other non-gastrointestinal symptoms. Both somatization and IBS symptom severity have been found to be responsive to treatment with hypnosis, although the mechanisms by which this occurs is not clear^56-58. This suggests that hypnosis may be a particularly helpful treatment in reducing multiple symptoms. Thus, psychological treatments are available – CBT and hypnosis – which target the two psychological variables with the greatest impact on IBS symptom severity.

Of all the psychological variables in the original model (see Figure 1), only lifetime sexual/physical abuse was not related to the other psychological variables nor to IBS severity. This is a surprising finding given the large literature on the association between abuse and IBS¹¹. Most of these previous studies, however, have shown that rates of abuse are increased in IBS patients compared to healthy controls. This does not have direct relevance to our current findings in which we investigated IBS severity rather than differences between IBS and controls. Very few have studied the influence of abuse history on IBS severity. One study found that patients with a history of sexual abuse reported higher pain in general-but did not distinguish abdominal pain from other types of pain⁵⁹. Two studies found no increase in abdominal pain, constipation or diarrhea^{60, 61}, although a small decrease was found in bloating among those who have been sexually abused. Thus, there is little evidence from the existing literature that abuse history affects IBS symptom severity which is in line with our current findings. However, there is a large literature suggesting that abuse history is associated with increased anxiety^{62, 63}, so our lack of association between these variables suggests that more research is needed before excluding the role of abuse on distress and symptoms in IBS. For example, abuse severity or cumulative exposure to abuse may be more important than abuse history in predicting IBS severity.

There is one other pathway in our model that does not make intuitive sense: The reciprocal association between catastrophizing and anxiety. It can be expected that these variables are positively associated: An anxious person would be more likely to catastrophize in the face of pain; while worrying that the pain is going to become worse and there is nothing you can do, will likely increase anxiety. Indeed we found that anxiety increases catastrophizing, but catastrophizing, in turn, was associated with decreases in anxiety. It may not be immediately clear how maladaptive coping reduces anxiety until we place catastrophizing in a larger context. As proposed by Sullivan et al.²² and empirically validated among IBS patients by Lackner et al.⁶⁴, catastrophizing may not have the primary intent to reduce pain but rather to increase social support. In order to solicit social support, patients may report more pain and display more pain behaviors⁶⁴. It may be argued that soliciting social support for pain may be an effective strategy to reduce pain. If the pain becomes too intense to handle or exceeds coping abilities, patients may look to others for help; a coping strategy common among children with pain who solicit their parents for help and whose pain outcomes are dependent on parental coping strategies⁶⁵. Thus, it can be hypothesized that when IBS patients feel overwhelmed by pain they may display catastrophizing behaviors –temporarily increasing the pain- in order to elicit help and support by others. By delegating responsibility to cope with pain, distress may go down. More studies are needed to determine if catastrophizing is used to increase social support for pain or if another explanation is of more relevance.

We would like to compare our results with other studies, but surprisingly there are few studies testing the relative influence of psychological variables on IBS. There have been some studies employing structural equation modeling or path analysis to investigate the role of several physiological and psychological factors in IBS. Naliboff and colleagues⁶⁶ investigated whether health related quality of life was determined by gastrointestinal symptoms or psychological symptoms. Only one global psychological construct was investigated: overall psychological distress as measured by the Symptom Checklist 90⁶⁷. The authors concluded that psychological distress had a stronger impact on quality of life than gastrointestinal symptoms in patients with IBS. Lackner and colleagues⁵¹ investigated the pathways through which cognitive behavioral therapy (CBT) affects IBS improvement. They concluded that CBT has a direct effect on gastrointestinal symptom improvement but no significant effect on psychological distress. Of interest was their finding that psychological distress had a greater impact on IBS quality of life than did gastrointestinal symptom improvement. Garland and colleagues⁵⁵ tested therapeutic mediators of the effect of Mindfulness Training on IBS outcomes. They found that both pain sensitivity and coping (catastrophizing and reinterpretation of pain sensations) had independent effects on IBS severity and quality of life. Thus, these studies establish the importance of psychological distress and coping for IBS outcomes, but none of the models included both factors. Our current findings support and extend these observations by showing that the effects of psychological distress (as measured by anxiety) on IBS outcomes are mediated by coping.

Our study has several limitations. The psychological factors included in our model were chosen based on a literature review at the beginning of our study. We chose variables for which there was a large evidence base and did not show too much overlap with each other. This selection process can create bias. Despite our best efforts to include the most relevant psychological variables we may have left out some important and/or newer concepts such as perceived stress^{68, 69}, gastro-intestinal specific anxiety⁷⁰, anxiety sensitivity¹⁹, alexithymia^{68, 71}, or hypervigilance^{72, 73} to symptoms. Models that include these variables may show a different outcome. Therefore, studies are needed to replicate and extend our current findings. Furthermore, as mentioned before, the current retrospective dataset does not allow for cause-effect inferences between factors. Longitudinal studies are needed to test whether change in one factor is followed by change in another. In addition, our data is limited by the fact that we included a self-selected sample who all had consulted a physician for their symptoms. The decision to see a doctor may be driven by psychological factors rather than severity of the gastrointestinal symptoms^{74, 75}. This suggests that psychological factors may play a more important role in our sample. Alternatively, symptoms may interfere less with life among non-consulting patients due to reduced levels of stress, anxiety, catastrophizing etc. in this population. Therefore, our model is in need of replication in non-consulting samples before generalization to the whole population can be assumed. Moreover we acknowledge that our model is limited by its exclusive focus on psychological factors. We realize that models need to be tested that also include physiological variables such as motility and social factors such as gender and social support, but to keep the model simple these interactions have been omitted from the current analysis. We believe these limitations to the model are justified by previously cited evidence that psychological variables are the strongest moderators of symptom severity^{51, 55, 66}. The findings of our study may help devise future studies which include all biopsychosocial aspects, by providing a better understanding of the most important psychological factors to be included. A final limitation to our study is that it focuses only on current symptom severity and ignores other outcomes such as symptom persistence and responsiveness to treatment. The cross-sectional nature of the research design does not allow us to examine these delayed consequences but clearly more data is needed to determine the relative impact of psychological factors on these outcomes.

In conclusion, the present study is the first to test an overarching model explaining the role of psychological factors in IBS. It includes both past life experiences, personality and more modifiable psychological factors. Of all these factors, catastrophizing and somatization seem to be the most important in regard to IBS symptom severity and deserve particular attention in research and treatment. It needs to be acknowledged that physiological and social factors are also of importance in predicting IBS severity and interact with psychological factors to predict IBS outcomes. More studies are needed in which all variables are tested simultaneously. We hope that the current model can be a guideline for future research and treatment of IBS.