Psychological Therapies in Patients with Irritable Bowel Syndrome: A Systematic Review and Meta-Analysis of Randomized Controlled Trials

Background. Irritable bowel syndrome (IBS) is a poorly understood disease with few effective treatments. Psychosocial factors are believed to contribute to the pathogenesis of IBS. Objective. To evaluate the evidence for psychological therapies in IBS treatment. Methods. We searched six medical databases through February 6, 2014, for randomized controlled trials (RCTs) of psychological therapies for the treatment of IBS. Two independent reviewers identified the RCTs, extracted the data, and assessed trial quality. We used the random-effect model to pool standardized mean difference (SMD) and 95% confidence interval (CI) across trials. Results. 15 RCTs that mostly evaluated cognitive behavioral therapy were included. Psychological therapies were associated with improvement in IBS symptoms severity scales (SMD −0.618; 95% CI: −0.853 to −0.383), IBS-Quality of Life (SMD 0.604; 95% CI: 0.440 to 0.768), and abdominal pain (SMD −0.282; 95% CI: −0.562 to −0.001). No statistically significant effect was observed on diarrhea or constipation. Limitations. The trials were at increased risk of bias and the overall sample size was small leading to imprecision. Conclusion. Psychological therapies may improve the quality of life and symptom severity in IBS. The effect size noted is moderate to large and is clinically meaningful.


Introduction
Irritable bowel syndrome (IBS) is a complex and widespread functional bowel disorder (10-20% worldwide prevalence [1]) that is not well understood. IBS typically presents as persistent diarrhea and/or constipation that is accompanied by abdominal discomfort. The symptoms as well as the underlying etiologies of IBS can vary considerably from patient to patient. Some of these causes may include diet, genetics, altered intestinal environment, and dysregulation of the enteric nervous system function. Multiple treatments targeting these possible causes have been used for several decades but have largely only been demonstrated to temporarily treat symptoms.
for several years, although definitive conclusions will require larger, higher quality studies. Cognitive behavioral therapies (CBT) and mind-body therapies (MBT) have also been studied in IBS with some studies showing preliminary efficacy [11][12][13]. Psychological therapies are potentially efficacious in treating IBS symptoms in many patients, and unlike many pharmaceutical treatments, they have minimal side effects and can be cost-effective [14][15][16][17].
To examine if psychological therapies merit incorporation in the clinical treatment of IBS, we conducted this systematic review and meta-analysis of published randomized controlled trials. To our knowledge, no existing systematic reviews with meta-analysis have addressed this question.

Methods
Investigators developed a protocol in advance to specify eligibility criteria, outcomes of interest, and analysis methods. The methodology and reporting of this systematic review comply with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA statement) [18].
Nonrandomized comparative studies and single arm studies were not included. We excluded trials that evaluated hypnotherapy because multiple systematic reviews have already summarized this evidence [8][9][10]. We also excluded non-English references.

Search Methods.
A comprehensive search of several databases from 1966 to February 6, 2014, any language, was conducted. The databases included Ovid Medline In-Process & Other Non-Indexed Citations, Ovid MEDLINE, Ovid EMBASE, Ovid PsycINFO, Ovid Cochrane Central Register of Controlled Trials, Ovid Cochrane Database of Systematic Reviews, and Scopus. The search strategy was designed and conducted by an experienced librarian with input from the study's principal investigator. Controlled vocabulary supplemented with keywords was used to search for comparative studies of psychological and mind-body interventions for irritable bowel syndrome. The actual strategy is included in the Appendix.

Study Selection.
Two independent blinded reviewers (OA and VS) assessed the eligibility of the candidate references for inclusion by screening titles and abstracts initially. The full-text publications were then retrieved and their eligibility was assessed. Disagreement between the two reviewers was resolved by meeting and establishing consensus. Interreviewer agreement was measured by using the Kappa statistic.

Data Collection
Process. Data were abstracted from each study using a standardized and piloted Microsoft Excel spreadsheet-based extraction form. Two independent blinded reviewers (OA and VS) did the abstraction in duplicate and disagreements were resolved by meeting and establishing consensus. The following data were abstracted: description of enrolled patients (inclusion criteria, age, gender, race, and previous treatment), description of received interventions and control, follow-up monitoring, and measures of outcome.

Outcomes of Interest.
The primary outcomes were the composite IBS symptoms severity scales and quality of life. Other outcomes were diarrhea, constipation, and abdominal pain.
2.6. Assessment of the Risk of Bias. Two reviewers (OA and VS) evaluated the methodological quality of the included trials. To evaluate randomized controlled trials, we used the Cochrane Risk of Bias tool [39]. We evaluated the adequacy of randomization, allocation concealment, blinding (patients, providers, data collectors, and outcome assessors), baseline imbalance, and extent of loss to follow-up. We also extracted the funding source.

Statistical Analysis.
Because the outcomes of interest were evaluated in the included trials using different scales, we estimated the standardized difference in means (SMD) to measure the difference between the intervention and control groups. SMD calculation involves standardizing the effect and expressing it in standard deviation units, to allow pooling it across trials. For each trial, we calculated the change in the studied scales before and after the intervention and compared it to the change in the control group. Then DerSimonian and Laird random-effects model was used to pool SMD across trials [40].
Inconsistency across the trials was assessed using the 2 static and Cochran's test. 2 value more than 50% was considered indicative of substantial heterogeneity that is due to real differences in protocols, trial populations, interventions, and/or outcomes. Also, Cochran's test value less than 0.05 indicates that the heterogeneity is beyond chance or random error [41]. We planned to conduct formal tests to assess potential publication bias using visual inspection of funnel plots and Egger's regression asymmetry tests but this was not possible due to the small number of trials [42].
We planned to explore possible causes of heterogeneity by conducting subgroup analyses comparing the effect size between trials that evaluated CBT versus other forms of psychotherapy, trials in which patients received prior treatment versus those who did not, trials with high risk of bias versus low risk of bias, trials in which placebo or sham therapy was  used in the control arm, and trials in which the control arm received pharmacological treatment versus those that did not. Interaction test between subgroups was done as suggested by Altman and Bland [43]. Statistical analyses were conducted using Comprehensive Meta-Analysis Version 2.2 [44].

Study Selection.
A total of 1,216 references were identified through the search strategy. Screening of titles and abstracts excluded 1,160 references ( Figure 1). Two reviewers performed the initial screening and had an interreviewer agreement of Kappa of 0.84. Reviewing the retrieved full texts by the two reviewers excluded 25 publications and the interreviewer agreement about study eligibility, Kappa static, was 0.87. The remaining 22 publications included 15 trials.

Study
Characteristics. The 15 included RCTs enrolled 1,352 patients. The follow-up period ranged from 10 weeks to 15 months. The criteria for the included patients, baseline characteristics of the included patients, and the interventions they received are detailed in Table 2.
Greene used the Latimer criteria [38] to diagnose IBS while Fernandez used the Manning criteria [45]. Seven of the fifteen trials used Rome I criteria [46] and five trials used the Rome II criteria [47]. Corney et al. used author specified criteria [22].

Risk of Bias within Trials.
All the included trials were randomized controlled trials. Assessment of risk of bias for each of the included trials is summarized in Table 1. Nine of the fifteen trials had high risk of bias. All these nine trials did not provide details about allocation concealment. Seven of them did not report the randomization method. For the remaining two trials, one of them did not report the presence of baseline imbalances and blinding, and the other one had inadequate randomization.
Five of the fifteen included trials had moderate risk of bias. Two of the five had high loss to follow-up rate (44.5% and 50.5%). The remaining three trials did not report whether allocation was concealed or not. Only one of the fifteen trials was found to have a low risk of bias.
Thirteen of the included trials were funded by a notfor-profit organization. One study did not disclose funding source and one study was funded partially by a pharmaceutical company.

Meta-Analysis.
Nine of the included trials reported change in composite IBS symptoms severity scales. Four of these trials used the Irritable Bowel Syndrome-Severity Scoring System (IBS-SSS) used by Francis et al. [48], three used the Composite Primary Symptoms Reduction (CPSR) score used by Blanchard and Schwarz [49], one used the composite Bowel Symptom Severity (BSS) score used by Spiegel et al. [50], and one used a Global GI Symptoms Severity Score [30]. Random-effects meta-analysis ( Figure 2) showed a statistically significant change in composite IBS severity scales in patients who received psychological therapy ( = 383) compared to patients receiving control ( = 297) with SMD = −0.618 SD favoring psychological therapy (95% CI from −0.853 to −0.383). Moderate heterogeneity was observed ( 2 = 48.04% and = 0.052).
Eight of the included trials reported change in quality of life scales. Three of the included trials used the Irritable Bowel Syndrome-Quality of Life (IBS-QOL) scale used by Drossman et al. [51,52], two trials used the Physical Component Score of the Short Form 36 Health Survey [53], one study used the Irritable Bowel Syndrome-Quality of Life scale used by Hahn et al. [54], one study used the GI Qualityof-Life Index used by Eypasch et al. [55], and one study used the Work and Social Adjustment Scale used by Mundt et al. [56]. Random-effects meta-analysis ( Figure 3) showed a statistically significant change in quality of life scales in patients receiving psychological therapy ( = 352) compared to patients receiving control ( = 286) with SMD = 0.604 SD favoring psychological therapy (95% CI from 0.440 to 0.768). No evidence of heterogeneity was observed ( 2 = 0.000% and = 0.512).
Eight trials reported change in abdominal pain scales ( Figure 4). There was a statistically significant change favoring psychological therapy ( = 321) compared to control ( = 271) on random-effects meta-analysis with  Intervention group (psychoeducation group) reported lower GI symptom severity score and higher HADS anxiety score and level of education Six trials reported changes in diarrhea scales and constipation scales ( Figures 5 and 6). Random-effects metaanalyses showed no statistically significant difference in the change of diarrhea scales with SMD = −0.448 SD (95% CI from −0.912 to 0.017) and constipation scales with SMD = −0.130 SD (95% CI from −0.484 to 0.224) between subjects who received psychological therapy ( = 200) and who received the control ( = 146). Strong evidence of heterogeneity was observed in both meta-analyses of the diarrhea scales ( 2 = 70.25% and = 0.005) and the constipation scales ( 2 = 50.24% and = 0.074).

Subgroup Analysis.
There was no statistically significant interaction based on the type of treatment (CBT versus other forms of psychotherapy), based on the risk of bias, or based on whether the control arm received pharmacological treatment ( > 0.05, Table 3). There were insufficient data to conduct subgroup analyses based on whether a placebo was used in the control arm or based on whether patients received prior treatment.

Discussion
We conducted a systematic review and meta-analysis of randomized controlled trial evaluating the effects of psychological therapies such as cognitive-behavioral therapy (CBT) and mind-body therapy (MBT) on IBS patients' symptoms and quality of life.

Main Findings.
Our search identified 15 trials evaluating a psychological therapy on a sample of IBS patients. We excluded trials of hypnotherapy, as three reviews have already examined its effect on IBS. Outcomes were evaluated using validated scoring scales. The body of evidence varied from small in measures of diarrhea (346 subjects) and constipation (346 subjects) to moderate for the composite IBS symptom severity scales (680 patients). The conducted meta-analyses demonstrated a statistically significant effect of psychological therapies on IBS-Quality of Life and composite IBS symptom severity scales with minimal heterogeneity. In contrast, psychological therapies had no statistically significant effect on diarrhea and constipation with evidence of heterogeneity observed. Psychological therapies had a statistically significant effect on abdominal pain; however, this inference was limited by heterogeneity. It should be noted that the studies assessing these secondary outcomes might not have been adequately powered to evaluate changes in symptoms, as they were either secondary outcomes or part of a composite scale. The pooled effect sizes ranged from 0.13 SD to 0.62 SD.
The standardized difference in the means was used to express the effect size for each outcome in standard deviation unit as each outcome was assessed using different scales. According to Cohen, a difference of 0.2 SD is considered a small difference, 0.5 SD is considered moderate, and 0.8 SD is considered large [57]. Norman found that a 0.5 SD was the minimally clinically important difference (MCID) for changes in most of health-related quality of life for chronic diseases [58]. Studies were done to validate and define the MCID for the IBS-Quality of Life scale [51,52]. The MCID was found to be 10 to 14 points which is equal to 0.5 to 0.7 standard deviation units [59]. In this meta-analysis, the psychological therapy increased the quality of life scales by 0.604 SD units (95% CI from 0.440 to 0.768) compared to control. Using the above MCID this translates to a clinically meaningful improvement in quality of life.
When Francis et al. described the IBS-Severity Scoring System (IBS-SSS), they found that a change of 50 points was sufficient to detect clinical improvement [48]. This is equal to 0.6 SD in most of the trials that used the IBS-SSS in this meta-analysis [12,28,33,34]. The psychological interventions decreased the symptom severity scales by 0.618 SD (95% CI from −0.853 to −0.383) compared to control in the current meta-analysis. This translates to a clinically meaningful change consistent with improvement in the IBS symptoms severity scales.
The meta-analysis of the abdominal pain showed a statistically significant difference. We were not able to find a study that established the MCID for abdominal pain scales in patients with IBS. The psychological therapies decreased the abdominal pain scales by 0.282 SD (95% CI from −0.562 to −0.001) and by applying Cohen's assumption [57] above this translates to a small difference. Also, this statistically significant difference was limited by the presence of heterogeneity.
The quality of evidence according to the GRADE [60] approach for the primary outcomes was low and moderate

S-CBT; 23 (7)
Standard CBT. Skills-based training program delivered to patients in 10 weekly, 1-hour sessions with weekly assignments. Six overlapping phases: (1) education of stress and IBS, (2) self-monitoring of stress associated with IBS, (3) muscle relaxation, (4) learning to identify, reevaluate, and change negatively skewed thoughts associated with IBS, (5) changing underlying "core" beliefs (e.g., perfectionism) that fuel threatening cognitions, and (6) formal training in problem-solving to strengthen the ability to cope with realistic stressors associated with IBS.

MC-CBT; 25 (5)
Minimal contact CBT. Covers the same range of procedures featured in S-CBT but relies extensively on self-study materials. Meeting for only four 60-minute clinic visits during the same period. Two 10-minute phone contacts are scheduled at weeks 3 and 7 to troubleshoot any problems. Meetings introduced material.

WLC; 27 (0)
Waiting list control. Subjects were placed on a 10-week delayed treatment waiting list, during which time they engaged in daily self-monitoring of gastrointestinal symptoms. NR.
IBS fact sheet in addition to a comprehensive self-management manual and weekly assignments. Also received a 1-hour face-to-face session with a health psychologist at the beginning of the program and two Attention control. Self-monitoring of IBS symptoms; educational material about IBS; and discussions with therapist. 10 fifty-minute sessions over 10 weeks.  for IBS symptom severity scale and IBS-Quality of Life, respectively, rated down for high and moderate risk of bias. The quality of evidence for secondary outcomes was very low, with downratings due to the overall high risk of bias, serious imprecision, and serious inconsistency in all of them. There are many limitations in this review. The trials evaluated had many methodological limitations and were generally small. Although specific details on frequency and duration of the psychological interventions were given in the trials, the efficacy of psychological therapies such as cognitive-behavioral therapy can range considerably depending on the training and experience of the therapist. The therapy is also often modified in the presence of psychological  comorbidity, which is common in IBS patients. This can result in variability in the therapy given in the different trials. There was variability in the follow-up periods as well. Publication bias is likely in a field in which evidence consists of trials with small size. Further, the efficacy of CBT or MBT could be driven primarily by improvement in general feeling of wellbeing and lower stress, and not specific improvement in IBS pathophysiology.

Strengths and
Another limitation pertains to the heterogeneity of IBS itself. IBS is now recognized as including three main subtypes: IBS-diarrhea predominant, IBS-constipation predominant, and IBS-mixed type. Many patients also develop IBS as sequelae to gastrointestinal infection or what is termed postinfectious IBS. These subtypes may have varying etiologies underlying their pathophysiology. By failing to stratify results by these subtypes, we are unable to know if there would be differences in efficacy of psychological therapies between subtypes. This would have important practical implications for the clinical incorporation of psychological therapies in the treatment of IBS. In terms of the analysis, heterogeneity that remained unexplained despite subgroup analyses lowers the confidence in the meta-analytic estimates.

Conclusion
Psychological therapies such as cognitive-behavioral therapy and mind-body therapy may help to improve gastrointestinal symptoms and quality of life in IBS patients. Although statistical significance was found in IBS measures of quality   of life and symptom severity, these results should be interpreted with caution as trials were generally of low quality. Future trials will require larger sample sizes, longer followup periods, and higher quality methodology to provide a definitive recommendation on the incorporation of psychological therapies in the treatment of IBS. However, despite these concerns, psychological therapies appear to be a safe intervention and could be a practical option for patients who fail standard medical therapy.