IBS is associated with an increased incidence of psychological

Does Depression Influence Symptom Severity in Irritable Bowel Syndrome? Case Study of a Patient With Irritable Bowel Syndrome and Bipolar Disorder CATHERINE CRANE, BA, MARYANNE MARTIN, MA, DPHIL, DEREK JOHNSTON, MA, PHD, AND GUY M. GOODWIN, DPHIL, FRCPSYCH Objective: Irritable bowel syndrome (IBS) is frequently associated with mood disorder. However, it is typically difficult to distinguish between disturbed mood as a causal agent and disturbed mood as a consequence of the experience of IBS. This report considers the association between mood and symptom severity in a patient with diarrhea-predominant IBS and stable, rapid cycling bipolar disorder with a predominantly depressive course. Such a case provides an important opportunity to determine the direction of the relationship between mood and IBS symptom severity because the fluctuations of mood in bipolar disorder are assumed to be driven largely by biological, rather than psychosocial, processes. Methods: The study was carried out prospectively, with ratings of mood and IBS symptom severity made daily by the patient for a period of almost 12 months. Results: The patient experienced regular and substantial changes in mood as well as fluctuations in the level of IBS symptoms during the study period. Contrary to expectation, the correlation between mood and IBS symptom severity on the same day suggested that the patient experienced less severe IBS symptoms during periods of more severe depression. However, time series analysis revealed no significant association between these two processes when serial dependence within each series was controlled for. Conclusions: The unusual cooccurrence of IBS with bipolar disorder provides direct evidence to indicate that depression does not necessarily lead to an increase in the reported severity of IBS, at least in the context of bipolar disorder, and may under certain circumstances actually be associated with a reduction in the severity of IBS symptoms. Factors that might moderate the relationship between depression and symptom severity are discussed. Key words: irritable bowel syndrome, depression, mood, symptoms, longitudinal study, time-series analysis. IBS irritable bowel syndrome. INTRODUCTION IBS is associated with an increased incidence of psychological disorder in patient populations (1), and while the cause and nature of this association are a matter of discussion, several possible mechanisms, both psychological and physiological, have been proposed to account for the finding. Thus, an anxious or depressed mood may increase the severity of IBS symptoms either because individuals who are more anxious and depressed experience a given level of physical symptoms as more severe, distressing, and disabling (2) or because the physiological changes that accompany mood disorders have a direct influence on gastrointestinal function (3, 4), increasing the objective experience of IBS symptoms. Although most studies have examined the association between mood state and IBS symptom severity using between-subjects designs, these mechanisms would in fact suggest an association between mood state and IBS symptom severity within the individual. For example, although self-report measures of symptom severity cannot distinguish between the effect of mood state on physiology and on symptom perception, both the mechanisms discussed above would lead to a situation in which a worsening of mood would occur before a worsening of IBS symptoms when both were measured longitudinally. Conversely, IBS symptom severity could be causal in the relationship, with elevated anxiety and depression occurring as a consequence of the negative effects of IBS on daily life (5). Under such circumstances flare-ups of IBS symptoms would occur before worsening of mood. From the Department of Experimental Psychology (C.C., M.M.), University of Oxford; Department of Psychology (D.J.), St. Andrews University; and University Department of Psychiatry (G.M.G.), Warneford Hospital, Oxford. Address reprint requests to: Catherine Crane, Department of Psychiatry, University of Oxford, Warenford Hospital Oxford, OX3 7JX, UK. Email: catherine.crane@psychiatry.oxford.ac.uk Received October 17, 2002; revision received April 8, 2003. DOI: 10.1097/01.PSY.0000088590.01737.07 0033-3174/03/6505-0919 Copyright 2003 by the American Psychosomatic Society A third possibility is that the apparently elevated incidence of psychological disorders in patients with IBS symptoms arises as a result of ascertainment bias. In clinical case series, treatment seeking by individuals with IBS may be increased if they are also experiencing psychological distress (6). If the association between IBS and psychological disorders does arise due to differential treatment seeking rather than a direct relationship between mood and symptom severity, then within an individual the relationship between depression and symptoms might be minimal, although across a group those who are more depressed would be overrepresented among treatment seekers. Such an explanation would predict no significant relationship between changes in mood and changes in IBS symptom severity when both were observed longitudinally. We report here the case of a patient with IBS and bipolar mood disorder, who had experienced regular, cyclic fluctuations in mood for a period of years. We hypothesized that since the substantial mood changes experienced by the patient (episodes of severe depression interspersed with periods of normal functioning) were likely to be attributed primarily to biological rather than psychosocial factors, the severity of IBS symptoms would not be a major determinant of depression. In contrast, the cognitive and behavioral changes accompanying episodes of severe depression might be expected to have a profound effect on the reported severity of IBS (due to a combination of biological and psychological factors). It was hypothesized that a significant relationship between ratings of IBS symptom severity and mood would be identified and that if such a relationship were observed, time-series analysis would indicate that mood rather than symptom severity was the causal factor in the association (with changes in symptom severity occurring at the same time as or following, changes in mood). CASE REPORT The patient, a Caucasian woman, gave her informed consent, both to complete the daily ratings of mood and IBS 919

C. CRANE et al. symptom severity and to have the data concerning her mood and symptom severity presented as a case report. IBS The patient had received a clinical diagnosis of IBS in 1996 at the age of 57 years. At the time of the present study (starting December 2000), she fulfilled the Rome I criteria for IBS (7), reporting periods lasting at least 3 months during which she had experienced recurrent episodes of abdominal pain relieved by a bowel movement, in addition to experiencing loose, watery stools, more than three bowel movements per day, urgency to defecate, and abdominal fullness, bloating, and swelling (occurring on more than 25% of days or occasions during episodes of IBS). The patient had been referred to a gastroenterology outpatient clinic on two occasions: once when she first presented to her general practitioner with IBS symptoms and once following completion of the study. On this second occasion (in early 2002), the patient reported that a barium enema and colonoscopy were performed with the diagnosis of IBS reiterated. The patient used both peppermint preparations and charcoal tablets to manage IBS symptoms. Mood Disorder The onset of the patient s IBS occurred over the same interval and in relation to the same environmental stressors as the onset of her depressive symptoms. She was referred to secondary care with symptoms of anxiety and depression in early 1997. Symptoms of anxiety and depression had been present for many months. She described anxious thoughts, negative mood, and self-critical ruminations. She had early morning waking, reduced appetite, and poor concentration. A diagnosis of anxiety depression was made as an outpatient, and she was started on imipramine because of the lack of previous response to either Prothiaden or sertraline 150 mg daily. Within 4 weeks there was a dramatic improvement. Indeed, her mood had started to cycle between hypomania/ euthymia for 3 weeks and severe retarded depression for 2 to 3 weeks. The depression became so profound that she was admitted to a psychiatric unit for about 1 month and discharged home on imipramine 150 mg plus lithium 600 mg, giving blood levels of 0.5 mmol/l. When referred to one of the authors (G.M.G.) in May 1998, she was cycling between 4 to 5 weeks of marked depression and 5 to 6 weeks of euthymia or mild elation. She proved highly resistant to treatment, and over the period of the present study cycled on a regular basis despite a variety of serial adjustments in her medicines. During the present study, she was treated with lithium 600 mg and hormone replacement therapy throughout. To this was added lamotrigine (January and February 2001) then tranylcypromine 20 mg daily (March to December 2001) and in addition l-tryptophan 1 g tds (from December 2001). A diagnostic interview [the structured clinical interview for DSM-IV (8)] was used to confirm that the patient fulfilled the DSM-IV criteria for major depression during episodes of depression. The further classification of her illness by DSM-IV should probably be bipolar (not otherwise specified) because mood elevation has always been mild (never mania) and onset was associated with treatment with a tricyclic antidepressant. Measures The patient completed a daily diary, recording mood (on a 1 10 scale ranging from 1 lowest ever to 10 highest ever) and IBS symptom severity (rated from 0 no symptoms to 10 severe symptoms) over a period of nearly 12 months, from December 21, 2000 to December 14, 2001. The rating of IBS symptom severity was therefore a subjective measure, with changes potentially attributable to either an actual worsening of symptoms or a perceived worsening of symptoms or both. RESULTS Symptom Profile Figure 1 shows the time plot of the patient s mood ratings, IBS symptom ratings, and prescribed psychotropic medications during the study period. As seen in the graph, during this time the patient experienced six discrete episodes of depression. However, as would be expected, no overall trend in mood was observed. Autocorrelation (with a maximum lag of 200 days) was used to confirm regularities in the patient s mood ratings over the study period. As shown in Figure 2, mood ratings showed a regular cyclic pattern with a period of approximately 50 days. The course of the patient s IBS was also variable during the study period with episodes of remission and relapse (Figure 1) but no apparent trend in the data. Autocorrelation, again with a lag of up to 200 days, revealed no evidence of a cyclic pattern in ratings of IBS symptoms (Figure 3). It should be noted that low scores on the mood scale indicate lower mood (eg, mood rated as closer to worst mood ever) whereas low scores on the IBS scale represent a relative absence of symptoms. A simple Pearson s correlation was carried out to examine the relationship between mood score and IBS symptom severity score on the same day (a lag of 0). This revealed a positive correlation (r 0.24, N 341) between mood and symptom severity, such that a worsening of mood was associated with an improvement in IBS symptoms. Although this correlation gives an indication of the relationship between IBS symptoms and mood, tests of the statistical significance of this relationship cannot be validly assessed because serial dependence within each series may result in apparent relationships between the series that are in fact the result of within-series associations. This is particularly likely to be true of the current data due to the strong seasonal (ie, cyclical) relationships apparent in the mood data series. Partial autocorrelations were calculated to examine the form of serial dependence between successive mood ratings and successive symptom ratings, individually for each series. This analysis revealed that both the mood data and IBS data were nonrandom series that could be approximated by firstorder autoregressive relationships. The partial autocorrelation functions at 1 day lag were very large in both cases (Mood: 920

IBS AND BIPOLAR DISORDER Fig. 1. Time plot to show daily mood ratings, daily IBS ratings, and prescribed psychotropic drugs across the study period. ACF 0.91, SE 0.053; IBS: ACF 0.71, SE 0.053) and much smaller at longer lags, indicating that ratings of mood (or symptoms) on 1 day were the main predictor of ratings on the subsequent day. Indeed, the mood profile consisted of long periods during which mood was stable, with mood rated at 1 or 1.5 on 22% of days and 5 or 5.5 on 46% of days. Cross-Correlation Between Mood and IBS In order to examine the relationship between mood and IBS symptom severity, a statistical correction, referred to as prewhitening, was applied to each series. Pre-whitening eliminates autocorrelation within a series (IBS, mood) by allowing for the autocorrelation and extracting the residual series from each. This correction was necessary because statistical techniques such as correlation make an assumption of independence between separate data points, which is not fulfilled in the uncorrected series, and provides an adequate method for detecting short-term relationships between series (as might be expected if mood led to perceptual biases or had a direct impact on intestinal motility). In order to pre-whiten the mood and IBS data series, a new lagged series (with a time lag of 1 day) was created for each. A 1-day lag was used because autocorrelation indicated that both mood and IBS symptoms were first-order autoregressive series. These lagged series were then regressed on to the original series and the residuals stored as new data points. These residual series therefore represented the patient s ratings on each day with variance attributable to serial correla- Fig. 2. Autocorrelation of daily mood ratings with 200 day maximum lag. Fig. 3. Autocorrelation of daily IBS ratings with 200 day maximum lag. 921

C. CRANE et al. tion removed. Autocorrelation carried out on the residual series indicated that the correction was effective in eliminating serial correlation. Relationship Between Mood and IBS Symptom Severity Cross-correlation was used to examine the relationship between the pre-whitened mood and IBS symptom series with a maximum lag of 15 days. As shown in Figure 4, no significant relationship between mood and IBS symptom severity was identified, with the correlation coefficient functions not exceeding the confidence limits at any lag. This result suggests that substantial fluctuations in mood, including the onset and offset of episodes of major depression, were not associated with any significant changes in the reported severity of symptoms of IBS in this patient, although changes in reported symptom severity were observed over the study period. DISCUSSION Our hypothesis was that depression would be associated with a more negative perception of symptoms, with symptoms being rated as more severe and bothersome during episodes of depressed mood. Our hypothesis was not confirmed: timeseries analysis revealed no significant relationship between short-term changes in mood and short-term changes in IBS symptoms. Indeed, a simple Pearson s correlation coefficient at 0 days lag indicated a relationship between worsening of mood and an improvement in symptoms of IBS. While this relationship should be treated with caution (because it may be artifactual and result from seasonal [cyclical] variations in the mood series), examination of the plots of raw IBS data and mood data (shown in Figure 1) tentatively suggests that improvements in mood may precede a worsening of IBS symptoms. However, since the IBS autocorrelation data (Figure 3) showed no evidence of the cyclic effects observed in the mood series (Figure 2), it is suggested that any relationship between Fig. 4. Cross-correlation between pre-whitened mood ratings and pre-whitened IBS symptom severity ratings with maximum lag of 15 days. improved mood and worsening IBS is likely to be fairly weak. The findings do suggest, however, that increased depression may lead to an improvement in IBS symptoms in certain individuals. In view of the commonly held belief that depression is associated with more problematic IBS symptoms, these findings are surprising, suggesting that there is no necessary link between the experience of severely depressed mood and worsening of the symptoms of IBS. The form that IBS takes may be critical in determining the impact of mood on symptom severity. For example, diarrheapredominant IBS (experienced by this patient) may be more closely associated with anxiety states and may improve somewhat as a result of the cognitive and behavioral changes induced by depression [eg, psychomotor poverty (9) and reduced food intake], whereas constipation-predominant IBS may be worsened by depression. Indeed, research suggests that anxiety is associated with a reduction in intestinal transit time and depression with an increase (10), a finding that would be compatible with such relationships. Despite these considerations, however, it would have still been predicted that some association between depression and IBS symptoms would be observed using time-series analysis. It is possible that such a relationship exists but that during periods of severe depression the patient was preoccupied by other concerns and paid relatively less attention to gastrointestinal function or attempted fewer activities (such as shopping, socializing), rendering symptoms less disruptive and bothersome. Thus, a relationship between mood and symptoms may be most apparent during episodes of mild to moderate depression or during periods of transition. However, a previous study (11) also found no significant relationship between one off measures of depression and anxiety and mean level of symptoms recorded daily over an 8 week period in patients with IBS, and this study supports the view that there is not a relationship between daily ratings of mood and daily ratings of symptom severity. The etiology of major depression is complex, and patients with IBS may share elements of this etiology. For example, outpatients with IBS are more likely than controls to have experienced psychosocial adversity such as abuse (12) and to report high levels of health anxiety (13). Co-morbidity of IBS and major depression may reflect these antecedent factors rather than a more simple interaction between IBS and mood. Further, if interactions between mood and IBS do exist, there is likely to be considerable heterogeneity within the patient population, and the absence of a time-lagged association between mood and symptoms would not be incompatible with some such interactions. For example, long-term changes in behavior (dietary change, social withdrawal) brought about by depression or IBS might result in a nontemporal relationship between symptoms and mood, or a relationship between mood and symptom severity may be mediated by some third variable such as presence of daily hassles (14). Due to the long period of time over which records were kept, this study relied on single daily ratings of global mood and IBS symptom severity. Further research focusing on the 922

IBS AND BIPOLAR DISORDER impact of daily mood on ratings of different types of IBS symptom (eg, perception of pain, bloating, diarrhea) would be informative, particularly since research we have recently carried out suggests that these symptoms may respond differently to psychological manipulations (15). In addition, monitoring the association between mood and symptom severity in an individual with a more rapidly cycling mood disorder would provide an opportunity to examine the association between mood and symptom severity during transition periods in more detail. In our current uncertainty, single cases are still able to provide challenging findings, as the present example illustrates. The authors thank our subject for the time taken, during a difficult year, to complete our paperwork. REFERENCES 1. Creed F, Guthrie E. Psychological factors in the irritable bowel syndrome. Gut 1987;28:1307 18. 2. Drossman D. Do psychosocial factors define symptom severity and patient status in irritable bowel syndrome? Am J Med 1999;107(suppl): 41 50. 3. Ditto B, Miller S, Barr R. A one-hour active coping stressor reduces small bowel transit time in healthy young adults. Psychosom Med 1998;60: 7 10. 4. Fossey M, Lydiard R. Anxiety and the gastrointestinal system. Psychiatr Med 1990;8:175 86. 5. Corney R, Stanton R. Physical symptom severity, psychological and social dysfunction in a series of outpatients with irritable bowel syndrome. J Psychosom Res 1990;14:483 91. 6. Whitehead W, Crowell M. Psychologic considerations in the irritable bowel syndrome. Gastroenterol Clin North Am 1991;20:249 67. 7. Thompson GW, Creed F, Drossman D, Heaton K, Mazzacca G. Functional bowel disorders and functional abdominal pain. In: Drossman, D, editor. The functional gastrointestinal disorders: diagnosis, pathophysiology and treatment. A multinational consensus. Boston: Little Brown; 1994. 8. Spitzer RL, Williams JBW, Gibbon M, First M. Structured clinical interview for DSM-IV (SCID). Washington, DC: American Psychiatric Association; 1996. 9. Oettle GJ. Effect of moderate exercise on bowel habit. Gut 1991;32: 941 4. 10. Gorad DA, Gomborone JE, Libby GW, Farthing MJG. Intestinal transit in anxiety and depression. Gut 1996;39:551 5. 11. Stevens J, Wan C, Blanchard E. The short-term natural history of irritable bowel syndrome: a time series analysis. Behav Res Ther 1997;35: 319 26. 12. Talley N, Fett S, Zinsmeister A, Melton L. Gastrointestinal tract symptoms and self-reported abuse: a population-based study. Gastroenterology 1994;107:1040 9. 13. Gomborone J, Dewsnap P, Libby G, Farthing M. Abnormal illness attitudes in patients with irritable bowel syndrome. J Psychosom Res 1995;39:227 30. 14. Dancey CP, Taghavi M, Fox RJ. The relationship between daily stress and symptoms of irritable bowel: a time-series approach. J Psychosom Res 1998;44:537 45. 15. Crane C, Martin M. Attentional strategy and gastrointestinal symptoms in irritable bowel syndrome (in preparation). 923