Gastroenterology
Volume 133, Issue 2 , Pages 702-705, August 2007

In Search of Mechanisms of Change in Treatment Outcome Research: Mediators and Moderators of Psychological and Pharmacological Treatments for Irritable Bowel Syndrome

  • Jennifer S. Labus

      Affiliations

    • Corresponding Author InformationAddress requests for reprints to: Jennifer Labus, PhD, Center for Neurovisceral Sciences and Women’s Health, Neuroimaging Imaging Core, Peter V. Ueberroth Building, Room 2338C2, 10945 LeConte Avenue, Los Angeles, California 90095. fax: (310) 825-1919.

Center for Neurovisceral Sciences and Women’s Health, Department of Psychiatry and Biobehavioral Sciences, David Geffen School of Medicine at UCLA, Los Angeles, California

Article Outline

 

See “How does cognitive behavior therapy for Irritable Bowel Syndrome work? A mediational analysis of a randomized clinical trial,” by Lackner JM, Jaccard J, Krasner SS, et al, on page 433.

Irritable bowel syndrome (IBS) is a common syndrome characterized by abdominal pain and discomfort associated with altered bowel habits, and despite continuing controversy regarding the syndrome’s etiology and primary pathophysiology, there is general agreement about alterations in brain–gut interactions in symptom manifestations. Based on this concept the general target areas for gut-directed, neurologically directed pharmacologic therapy (see also, Gershon and Tack1), and psychological therapies can be identified (Figure 1).2

  • View full-size image.
  • Figure 1. 

    Targets within the brain–gut axis for the treatment of FGIs. Shown are general target areas for various therapeutic approaches to IBS symptoms. Whereas neurologically directed pharmacologic therapies aim for targets within the spinal cord, brain stem, and limbic and paralimbic regions, psychological therapies are likely to affect thoughts, belief systems, and memories processed in the prefrontal cortex. Through this effect, they may strengthen the cortico–limbic–pontine pain inhibition system and counteract central pain amplification. Modified from Mayer et al.2

Systematic reviews of the literature3, 4, 5, 6, 7 and recent treatment outcome studies8, 9, 10 indicate certain psychological treatments including hypnotherapy and cognitive–behavioral therapy (CBT) are effective in treating IBS. Published evidence demonstrates that psychological treatments are effective in relieving somatic symptoms (eg, bowel dysfunction, rectal hypersensitivity, and abdominal pain), in various adult functional gastrointestinal (GI) disorders (eg, IBS, functional abdominal pain).3, 4, 10, 11 A recent study9 also demonstrated, for the first time, the effectiveness of hypnosis in children with IBS and functional abdominal pain (FAP). Specifically, 1 year after gut-direct hypnotherapy, the relative risk ratio of being a responder (complete remission as defined by >80% improvement in pain frequency and intensity) was a remarkable 3.8 (95% confidence interval, 1.6–6.9) with the average number of patients needed to treat (NNT) with hypnotherapy for one patient to achieve improvement over the standard medical therapy (ie, education, dietary advice, fiber, pain medication, and supportive therapy) was 2. Of particular note, 96% of the patients treated with hypnotherapy were either considered cured (85% achieved complete symptom remission) or improved (11%). These exciting results extend previous findings that hypnotherapy is a beneficial treatment for relatively refractory IBS patients that provides long-lasting improvements in GI symptoms (eg, pain, bowel dysfunction, bloating) and psychological distress (eg, anxiety, depression, and quality of life in adult IBS patients).3, 5, 6, 7

The improvements achieved with psychological treatments are similar or even greater than the improvement obtained with recently developed, gut-directed pharmacologic treatments (tegaserod, alosetron), which have shown efficacy in relieving overall GI symptoms, abdominal discomfort, and normalizing bowel habits.12, 13 In the most sophisticated meta-analysis of CBT for IBS to date, efficacy data (>50% reduction of GI symptoms) from 17 studies yielded an odds ratio of 12.0 (5.6–26.0). The average number of patients needed to treat with CBT or hypnotherapy for 1 patient to achieve improvement over the wait list or other control groups (ie, symptom monitoring, attention-placebo control, medical treatment as usual) was estimated to be 2.4 In comparison, the pooled odds ratio for adequate relief of pain or global symptom improvement from 6 large randomized control trials of alosetron was 1.8 (1.6–2.1), with the NNT = 7.12 From 8 studies of tegaserod, the pooled relative risk ratio of being a responder (global relief of GI symptoms) versus a nonresponder was 1.2 (1.1–1.3), with an NNT of 17.13 For antidepressant therapy for functional GI disorders (FGIDs) in 8 studies (primarily low-dose tricyclic antidepressants [TCAs]), the summary odds ratio for improvement was 4.2 (2.3–7.0) and it was estimated that 4 (2.1–6.5) patients would need to be treated to achieve improvement over placebo treatment.14

Even though CBTs influence the same symptom outcome measures as pharmacologic treatments for IBS, a direct comparison of the magnitude of these effects is discouraged because the nonpharmacologic treatment studies have not used a control condition comparable to a medication placebo control. Implementing experimental designs with parallel treatment arms comprising an active pharmacologic treatment, medication placebo, psychological treatment, and combined medication and psychological treatment will ultimately permit more direct comparison of the magnitude of efficacy of psychological versus pharmacologic treatments in IBS (see Holroyd et al15). As has been the case in other functional pain disorders (eg, migraine, chronic tension-type headache), differences in overall efficacy between effective drug and psychological treatments may prove to be negligible16 and research may shift to developing algorithms for optimizing combinations of cognitive, behavioral and pharmacological treatments to maximize treatment effectiveness and patient satisfaction.17, 18

As we move beyond questions regarding whether cognitive and behavioral treatments work, two important questions will take precedence: How does treatment work and for whom does it work? This type of research involves identifying mediators (how does the treatment work?) and moderators (for what types of patients in what settings, under what conditions does treatment work) of treatment.19

Mediators are intervening variables that occur or change after application of a therapeutic agent or intervention and represent the potential mechanisms or causal agents by which a treatment affects outcomes.19, 20, 21, 22 Characterizing the mechanism(s) underlying cognitive, behavioral, and pharmacologic treatment effects on IBS symptoms (ie, identifying treatment mediators) is not only important for improving treatment programs by identifying the most important components, but also for improving the understanding of functional disorders and the variables associated with their course.21 In contrast with pharmacologic treatments, which are generally based on specific mechanisms and aimed at molecular targets identified in animal models, the underlying nature of the success of psychological treatments has traditionally been poorly understood. Treatment success is commonly thought to be mediated by reductions in comorbid psychological distress, rather than somatic symptoms directly. Other mechanisms of treatment success have been proposed in relation to gut-directed hypnotherapy10, 23, 24 and are readily extended to other successful CBTs, including changes in small bowel and colonic motility, alterations in central and autonomic nervous system functioning, reduction in rectal sensitivity, or improvement in psychological symptoms. In relation to psychotherapy and TCAs, Creed et al25 have suggested that observed symptomatic improvement with these treatments may be mediated by changes in depression or improved coping, or possibly by alterations in neurotransmitter systems by TCAs.25

Along this mechanistic vein, Lackner et al,26 in this issue of Gastroenterology, applied a mediation analysis to a subsample of data from a randomized clinical trial of group-based CBT for moderate to severe IBS.8 In this subsample, group-based CBT in comparison to a psychoeducational group, as an attention-based placebo control condition, reduced GI symptoms on a composite symptom measure (abdominal pain and tenderness, bowel dysfunction, nausea, and belching). Treatment gains were maintained at 3-month follow-up.8 Lackner et al’s analysis of these data provides preliminary evidence for the provocative idea that CBT exerts a direct effect on GI symptoms independent of its effects on psychological distress and health-related quality of life (HRQOL), and challenges the notion that improvement in IBS symptoms after psychological treatment results from reduction in comorbid psychological distress.26 Strikingly, the results suggest the greatest therapeutic value of CBT may lie in improving GI symptoms. These results will likely spark much debate and (hopefully) promote further investigation into the process of improvement in FGIDs through the use of statistical tools such as mediation analyses. One plausible mechanism underlying such a direct effect is shown in the brain–gut model in Figure 1. It is conceivable that strengthening of cortico–limbic–pontine inhibition circuits by CBT directly reduces central pain amplification mechanisms.

Lackner et al26 also examined potential moderators of treatment effects. Moderators are variables that significantly determine who will respond to a particular treatment and under what conditions,19, 20 and are particularly important in heterogeneous populations like those with FGIDs. Potential moderators include bowel subtype, age, gender, chronicity, comorbid psychiatric disorders, comorbid functional disorders (eg, migraine, fibromyalgia), baseline symptom severity, or even physiologic (high baseline startle), and genetic polymorphisms. This type of work is exemplified in recent work that indicates severity of illness; and impairment, age, and comorbid psychological disorders predict who will respond to pharmacologic treatment for attention deficit hyperactivity disorder, anxiety, and depression in children and adolescents.27, 28, 29 Lackner et al26 examined whether abuse history, age, gender, chronicity of disease, symptom severity, catastrophizing, bowel habit, gender, or comorbid psychiatric disorders predicted treatment response to CBT. The results of this moderator analysis suggested patients with greater disease chronicity and baseline symptom severity were likely to achieve the greatest improvements in HRQOL. The surprising lack of evidence for comorbid psychiatric disorder as predictive of treatment response to psychological treatment is consistent with previous research.25 Continued work in determining the significant moderators of drug and psychological treatments treatment is highly recommended because it will ultimately assist gastroenterologists in determining what treatments or combination of treatments will work best for particular patients.

Lackner et al26 used a structural equation modeling (SEM) framework for their mediation analysis, which is akin to estimating a series of linear regression equations simultaneously. Compared with a series of directed regressions, SEM is particularly advantageous for mediation analyses because it improves parameter estimation by using a full information estimation method, allowing bootstrapping of estimates and confidence intervals for small samples, and dealing better with missing data.30, 31, 32 SEM also provides indices of model fit, which test how consistent the observed data are with an a priori theoretical model, independent of significance testing of the path (ie, regression) coefficients.32 In a mediation analysis, indices of model fit can be considered in an exploratory fashion to inform future research. Additionally, in the presence of a moderator effect, mediation may depend on the value of the moderator. SEM can readily incorporate moderators in this framework, eschewing an awkward piecemeal approach to analysis.30, 31, 33 SEM is relatively easy to implement; there are a number of commercially available software packages.

One limitation of the analytic approach of Lackner et al,26 as well as that used by Creed et al,25 is the fact that mediation was examined using simultaneously assessed outcomes (ie, change in psychological distress, change in HRQOL, and global symptom improvement rating were assessed 2 weeks post-treatment). Simultaneously assessed outcomes do not permit the necessary condition of temporal precedence for a mediator. The need for temporal precedence in mediation analysis requires multiple assessments of potential mediators during a trial and can easily be implemented into future trial designs. Regardless, the results suggest the strong possibility of mediation and a causal sequence and a theoretical model that will need to be tested further.

A second issue relevant to experimental study design for treatment outcome research pertains to how to assess and quantify the effects of potential therapeutic agents as well as what to consider as a mediator for treatment. Current recommendations of primary endpoints for clinical trials include adequate and satisfactory relief as well as validated global symptom measures, such as the IBS Severity Scale.34, 35 Indeed, Lackner et al26 included such measures in their analysis. However, although they are considered the gold standard to assess outcomes in clinical trials, these more global assessments lack the specificity required for mechanistic studies. As such, careful selection of secondary endpoints, which generally include specific validated measures of emotional and cognitive dysfunction (eg, state or GI-specific symptom anxiety, vigilance, pain catastrophizing), is vital to ensuring the ability to more precisely test current theoretical models regarding the mechanisms of change for therapeutic agents. Not surprisingly, more specific measures of constructs such as psychological distress and GI symptoms could produce very different results. Although Lackner et al26 only used measures of phenomenologic experience (eg, self-report measures of emotional, cognitive, and physical functioning), mediators and moderators can be characterized at many levels of analysis. SEM can easily be extended to incorporate psychophysiologic (eg, heart rate variability, skin conductance response, acoustic startle response), and neuroimaging (eg, positron emission tomography, evoked-response potential, function magnetic resonance imaging) variables. SEM can integrate extrinsic variables (experimental design, gender, anxiety levels, and a variety of psychophysiological data) to more precisely model the mechanisms of therapeutic agents.

In summary, when symptomatology encompasses a wide span of symptoms (eg, pain, bowel dysfunction, psychological distress, reduced HRQOL) as it does in FGIDs, it is particularly meaningful to ask which symptoms are mechanisms versus correlates of change. It is also important to determine which patient characteristics will predict response to a particular treatment. The rather provocative results given by Lackner et al26 require further replication. Future researchers assessing mechanisms of change need to carefully consider choice and temporal assessment of mediators when designing treatment outcome research studies to allow for precise mechanistic studies. In addition, sophisticated but accessible analytic approaches are widely available to more precisely answer substantive hypotheses of interest. Researchers should capitalize on these statistical tools. Ultimately, treatment outcome research for IBS will greatly benefit from continued mechanistic studies of pharmacologic and psychological treatments. The fruits of this labor will be the ability to more efficiently target and prescribe the most useful treatment for a specific patient, resulting in improved patient outcome and overall satisfaction.

Back to Article Outline

References 

  1. Gershon MD, Tack J. The serotonin signaling system: from basic understanding to drug development for functional GI disorders. Gastroenterology. 2007;132:397–414
  2. Mayer EA, Naliboff BD, Craig AD. Neuroimaging of the brain-gut axis: from basic understanding to treatment of functional GI disorders. Gastroenterology. 2006;131:1925–1942
  3. Blanchard EB, Scharff L. Psychosocial aspects of assessment and treatment of irritable bowel syndrome in adults and recurrent abdominal pain in children. J Consult Clin Psychol. 2002;70:725–738
  4. Lackner JM, Mesmer C, Morley S, et al. Psychological treatments for irritable bowel syndrome: a systematic review and meta-analysis. J Consult Clin Psychol. 2004;72:1100–1113
  5. Wilson S, Maddison T, Roberts L, et al. Birmingham IBS Research Group Systematic review: the effectiveness of hypnotherapy in the management of irritable bowel syndrome. Aliment Pharmacol Ther. 2006;24:769–780
  6. Gholamrezaei A, Ardestani SK, Emami MH. Where does hypnotherapy stand in the management of irritable bowel syndrome? (A systematic review). J Altern Complement Med. 2006;12:517–527
  7. Whitehead WE. Hypnosis for irritable bowel syndrome: the empirical evidence of therapeutic effects. Int J Clin Exp Hypn. 2006;54:7–20
  8. Blanchard EB, Lackner JM, Sanders K, et al. A controlled evaluation of group cognitive therapy in the treatment of irritable bowel syndrome. Behav Res Ther. 2007;45:633–648
  9. Vlieger AM, Menko-Frankenhuis C, Wolfkamp SC, et al. Hypnotherapy for children with function abdominal pain or irritable bowel syndrome: a randomized controlled trial. Gastroenterology. 2007;[in Press]
  10. Tan G, Hammond DC, Joseph G. Hypnosis and irritable bowel syndrome: a review of efficacy and mechanism of action. Am J Clin Hypn. 2005;47:161–178
  11. Guthrie E, Barlow J, Fernandes L, et al. North of England IBS Research Group Changes in tolerance to rectal distension correlate with changes in psychological state in patients with severe irritable bowel syndrome. Psychosom Med. 2004;66:578–582
  12. Cremonini F, Delgado-Aros S, Camilleri M. Efficacy of alosetron in irritable bowel syndrome: a meta-analysis of randomized controlled trials. Neurogastroenterol Motil. 2003;15:79–86
  13. Evans BW, Clark WK, Moore DJ, et al. Tegaserod for the treatment of irritable bowel syndrome. Cochrane Database Syst Rev. 2004;CD003960
  14. Jackson JL, O’Malley PG, Tomkins G, et al. Treatment of functional gastrointestinal disorders with antidepressant medications: a meta-analysis. Am J Med. 2000;108:65–72
  15. Holroyd KA, O’Donnell FJ, Stensland M, et al. Management of chronic tension-type headache with tricyclic antidepressant medication, stress management therapy, and their combination: a randomized controlled trial. JAMA. 2001;285:2208–2215
  16. Holroyd KA, Penzien DB. Pharmacological versus non-pharmacological prophylaxis of recurrent migraine headache: a meta-analytic review of clinical trials. Pain. 1990;42:1–13
  17. Holroyd KA, Drew JB. Behavioral approaches to the treatment of migraine. Semin Neurol. 2006;26:199–207
  18. Holroyd KA, Lipchik GL, Penzien DB. Psychological management of recurrent headache disorders: empirical basis for clinical practice. In:  Dobson KS,  Craig KD editor. Best Practice: Developing and Promoting Empirically Supported Interventions. Newbury Park, CA: Sage; 1998;p. 193–212
  19. Kraemer HC, Wilson GT, Fairburn CG, et al. Mediators and moderators of treatment effects in randomized clinical trials. Arch Gen Psychiatry. 2002;59:877–883
  20. Baron RM, Kenny DA. The moderator-mediator variable distinction in social psychological research: conceptual, strategic, and statistical considerations. J Pers Soc Psychol. 1986;51:1173–1182
  21. Laurenceau JP, Hayes AM, Feldman GC. Some methodological and statistical issues in the study of change processes in psychotherapy. Clin Psychol Rev. 2007;in press
  22. MacKinnon DP, Fairchild AJ, Fritz MS. Mediation analysis. Annu Rev Psychol. 2007;58:593–614
  23. Simren M. Hypnosis for irritable bowel syndrome: the quest for the mechanism of action. Int J Clin Exp Hypn. 2006;54:65–84
  24. Palsson OS, Turner MJ, Johnson DA, et al. Hypnosis treatment for severe irritable bowel syndrome: investigation of mechanism and effects on symptoms. Dig Dis Sci. 2002;47:2605–2614
  25. Creed F, Guthrie E, Ratcliffe J, et al. North of England IBS Research Group Does psychological treatment help only those patients with severe irritable bowel syndrome who also have a concurrent psychiatric disorder?. Aust N Z J Psychiatry. 2005;39:807–815
  26. Lackner JM, Jaccard J, Krasner SS, et al. How does cognitive behavior therapy for irritable bowel syndrome work? (A mediational analysis of a randomized clinical trial). Gastroenterology. 2007;133:433–444
  27. Walkup JT, Labellarte MJ, Riddle MA, et al. Research Units on Pediatric Psychopharmacology Anxiety Study Group Searching for moderators and mediators of pharmacological treatment effects in children and adolescents with anxiety disorders. J Am Acad Child Adolesc Psychiatry. 2003;42:13–21
  28. Moderators and mediators of treatment response for children with attention-deficit/hyperactivity disorder: The Multimodal Treatment Study of children with Attention-deficit/hyperactivity disorder. Arch Gen Psychiatry. 1999;56:1088–1096
  29. Curry J, Rohde P, Simons A, et al. TADS Team Predictors and moderators of acute outcome in the Treatment for Adolescents with Depression Study (TADS). J Am Acad Child Adolesc Psychiatry. 2006;45:1427–1439
  30. Shrout PE, Bolger N. Mediation in experimental and nonexperimental studies: new procedures and recommendations. Psychol Methods. 2002;7:422–445
  31. MacKinnon DP, Lockwood CM, Hoffman JM, et al. A comparison of methods to test mediation and other intervening variable effects. Psychol Meth. 2002;7:83–104
  32. Jaccard J, Guilamo-Ramos V, Johansson M, et al. Multiple regression analyses in clinical child and adolescent psychology. J Clin Child Adolesc Psychol. 2006;35:456–479
  33. Edwards JR, Lambert LS. Methods for integrating moderation and mediation: a general analytical framework using moderated path analysis. Psychol Meth. 2007;12:1–22
  34. Camilleri M, Mangel AW, Fehnel SE, et al. Primary endpoints for irritable bowel syndrome trials: a review of performance of endpoints. Clin Gastroenterol Hepatol. 2007;5:534–540
  35. Committee DoTT, Irvine EJ, Whitehead WE, et al. Design of treatment trials for functional gastrointestinal disorders. Gastroenterology. 2006;130:1538–1551

 Supported in part by K08 DK071626, R24 AT002681, P50 DK064539

PII: S0016-5085(07)01226-7

doi:10.1053/j.gastro.2007.06.052

Refers to article:

  • How Does Cognitive Behavior Therapy for Irritable Bowel Syndrome Work? A Mediational Analysis of a Randomized Clinical Trial , 22 May 2007

    Jeffrey M. Lackner, James Jaccard, Susan S. Krasner, Leonard A. Katz, Gregory D. Gudleski, Edward B. Blanchard
    Gastroenterology August 2007 (Vol. 133, Issue 2, Pages 433-444)

Gastroenterology
Volume 133, Issue 2 , Pages 702-705, August 2007

Article Tools

* Image Usage & Resolution