Acute and chronic stress in duodenal ulcer disease

GASTROENTEROLOGY 1990;99:1628-1632 Acute and Chronic Stress in Duodenal Ulcer Disease K. zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA ELLARD, J. BEAUREPAIRE, M. JONES, D. PIPER, and C. TENNANT Departments of Medicine and Academic Psychiatry, Royal North Shore Hospital, University of Sydney. St. Leonards, Australia Acute and chronic life event stressors were objectively assessed in a sample of duodenal ulcer patients and community controls. Stress was assessed on two dimensions, “personal threat” and “goal frustration.” Chronic stressors (those of 6 months’ duration or more) involving high goal frustration were significantly and independently associated with the onset and relapse of duodenal ulcers, as were acute events of high immediate but transient personal threat. espite the view that stressful experiences play some role in the onset or relapse of duodenal ulcer (DU) disease, empirical findings generally have not supported this hypothesis (l-5). These largely negative findings may indeed reflect the fact that no such relationship exists; however, the lack of positive evidence for such a relationship may have other causes. First, little research to date has focused on the relative etiological impact of acute versus chronic life stressors in DU disease; acute stressors alone rather than ongoing chronic stressors have been the main focus of assessment. Evidence (6) suggests that indeed acute and chronic stressors may have a significantly different role in the development of duodenal ulcer disease. A matched case-control study that effectively distinguished acute and chronic stressors (6) found that whereas acute life events were not causally related to DU disease onset, chronic difficulties were. The DU patients had more than double the number of difficulties that controls did, but no particular descriptive type of difficulty (e.g., work, marriage, etc.) distinguished DU patients from controls. The mean duration of the patients’ difficulties was also greater, 6.4 years compared with 3.8 years for controls. These results suggest that life stressors must be of sufficient duration and severity to provoke the necessary chronic physiological changes that cause DU. Further research is required to replicate these findings. A second explanation for the negative findings for the role of stress in DU disease is that previous studies of life-event stressors have usually relied on selfreported life-event checklists. Such a methodology D has been widely criticized (7) for its poor reliability and validity and the inadequate sensitivity and specifity of items. The development of the Bedford College Interview Schedule of Life Events and Difficulties (LEDS) has overcome these problems [8,9). Whereas the LEDS has been used largely with psychiatric disorders (8,10-13), it has also been used in a few studies of medical disorders (6,14-18). Only one of these studies relates primarily to DU disease (61. Further studies using the LEDS are required to provide reliable and valid data on the relationship of stress to the development of DU disease. Third, it is possible that existing studies have not considered aspects of stress relevant to DU. Previous studies have assessed stressors primarily in terms of how distressing or threatening the event or difficulty was to the patient. However, they have largely ignored the issue of “goal frustration,” which seems to be of etiological significance to gastrointestinal disorders (18). Goal frustration assesses the degree to which important life goals and needs are frustrated by the occurrence of the stressor. This concept was developed to test early theories proposing that psychosomatic disease, including gastrointestinal (GI) disorders, are the result of frustration of ambition or personal needs (19,201. Craig and Brown (18) studied the concept of goal frustration in a sample of patients with a variety of upper GI diseases including DU. Stressors deemed specifically goal frustrating were associated with the onset or relapse of these mixed GI conditions. Although their study suggests a link between stress and GI disease, the conclusions that can be drawn from it are restricted by the study’s lack of distinction between acute and chronic stressors and the heterogeneity of the sample, which included several GI disorders. Abbreviations used in this paper: CL confidence interval; DU, duodenal ulcer; LEDS, Bedford College Interview Schedule of Life Events and DiiBculties; OR, odds ratio. 0 1990 by the American Gastroenterological Association 0018.5095/90/$3.00 December 1990 STRESS IN DUODENAL ULCER DISEASE 1629 In the present study, both acute and chronic stressors are assessed in DU patients using the LEDS (8,9). Acute and chronic stressors are rated for both threat and goal frustration so that both concepts can be compared in their possible relationship to the development of DU disease. Patients and Methods The Sample The patient sample comprised 81 consecutive patients diagnosed endoscopically as having an active duodenal ulcer at the time of entry to the study. The patients were matched with healthy community controls selected at random from the electoral roll for age, sex, and social status. Social status was assessed on the basis of occupation (21). Patients and controls were excluded if they had a history of gastric ulcer, other serious disease (e.g., cancer], or obvious psychiatric illness, were under 18 years or over 90 years of age, were unable to understand English, or were unwilling to give informed consent. Controls were also excluded if they had a history of peptic ulcer or of dyspepsia consistent with active or past peptic ulcer. Controls were not, however, necessarily free of all GI symptoms. Of the patients and controls who met the selection criteria, 95% and 67%, respectively, agreed to participate in the study. Measures and Analysis The method of assessment of acute and chronic stressors was the LEDS (7-9). The LEDS involves a semistructured interview that measures ongoing chronic difficulties as well as discrete acute events. Because of problems of recall, and based on our earlier findings (61, events and difficulties are assessed over a 6-month period only, the 6-month period before the onset of ulcer symptoms for patients and 6 months before the interview for controls. During the interview, each acute event (a stressor c6 months in duration) or chronic difficulty (a stressor r6 months in duration] was recorded in detail with particular attention to its social context. The time of its occurrence was noted as the number of weeks either before the onset of symptoms (for DU patients] or before the interview (for controls). Vignettes of each acute event or chronic difficulty were constructed by the interviewer and then objectively rated on a J-point scale both for severity of threat and of goal frustration by an independent rater, who was “blind” to the clinical status of the subject. Interrater reliability for these ratings has been shown to be very high (7,223. In essence, the threat rating measures how distressing or unpleasant such a person would find that event or difficulty given the detailed biographic and current social and personal circumstances as elicited at the interview. The goal frustration rating assesses the degree to which the subjects’ aims, needs, or ambitions were frustrated or obstructed by the occurrence of the event or difficulty. Again, the rating was objective in the sense that the rater was not told the subjects’ personal accounts of their emotional responses to the event or difficulty; judgments are again contextual in that they are made in terms of how a person in those biographical and environmental circumstances would be likely to respond. Both sets of ratings, furthermore, are both sensitive and specific because the rater is provided with considerable detail about the social context of each stressor before the rating is made (see Appendix). Accordingly, one threat rating was obtained for chronic difficulties and two ratings for each acute life event, either an “ immediate threat” (experienced at initial occurrence] or a “ long-term threat” (experienced 1 week later]. One goal frustration rating only was obtained for acute life events (experienced at 1 week] and for chronic difficulties, largely because most needs, ambitions, or goals are more enduring in nature. Acute events and chronic difficulties were then assessed as to the extent to which they were independent of the effects of ulcer disease. Only those events and difficulties likely to be independent of ulcer disease were included in the analyses. All events and difficulties were also coded by type into 10 different categories: personal illness, illness of another, death, role change or interaction, news or crisis, employment, finances, housing, marital (includes de facto or enduring relationship), and miscellaneous. Because this study is a matched case-control study, the data analysis reflects this paired aspect of the design throughout. Simple comparisons between patients and controls on such characteristics as age were performed by paired t tests in which normality assumption was met; otherwise, comparisons were made using the Wilcoxon signed-ranks test. Conditional logistic regression was performed to determine the subject characteristics that best differentiated patients from controls; this was appropriate because the outcome is binary (disease or control) and the data are paired (23). Analyses covered the total interview period (26 weeks before onset) and broke down this total period into two parts, less than 13 weeks and 13-26 weeks before onset. Initial multiple logistic regression analyses established independent life-event and chronic difficulty predictors of DU [Tables 1 and 2). Backward selection procedures established the best combination of independent DU predictors (Table 3). Univariate analysis was only used in analyses covering the two time periods before onset. Given the large number of tests and the lack of evidence enabling a planned logistic regression analysis, a conservative critical level of P < 0.01 was adopted. Results The Sample The 81 patients included 52 men (64% J and 29 women (36%). The mean ages of patients and controls were 58.2 years and 57.6 years, respectively. This difference just failed to reach statistical significance (t = 1.898, P = 0.06). Consequently, the final multivariate analysis controlled for age. The groups were matched for sex and social status. They did not differ in terms of marital status (McNemar’s x2 = 0.11; df = 1; P = 0.70). Acute Life Events Number and type of acute events. The total frequency of all acute life events was, paradoxically, slightly lower in duodenal ulcer patients than in 1630 GASTROENTEROLOGY ELLARD ET AL. Table I. Severity of Threat and Goal Frustration of Acute Events in Duodenal Ulcer and Control Subjects Ulcer patients n I%1 Immediate threat Severe High Moderate Absent Threat after 1 week Severe High Moderate Absent Goal frustration Severe High Moderate Absent Vol. 99, No. 6 Controls n I%1 10 (12) P OR 95% CI 0.78 0.81 1.19-3.47 31[36] 19 (23) 0.02 3.37 1.13-10.01 20 (24) 23 (28) 0.99 0.99 0.32-3.07 ll(13) 33 (41) 0.07 0.01-0.47 0.56-19.97 8 (101 10.01 3 (41 0.18 3.37 0.92 0.94 0.26-3.33 22 (27) 9 (111 21(26) 0.54 0.70 0.23-2.12 19 (23) 36 (44) 0.46 1.93 0.33-11.27 0.02-2.62 18 (22) 1 PI 3 (41 0.24 0.24 17 (21) 0.01 4.13 1.30-13.13 16 (20) 6 (71 14 (17) 0.91 1.04 0.46-2.32 27 (33) 33 (41) 0.41 0.73 0.36-1.51 NOTE. Data given as proportion of subjects with at least one such event O-26 weeks before onset. of threat at the time of occurrence zyxwvutsrqponmlkjihgfedcbaZYX (P -e 0.011, and this explained the slight excess of acute events overall. Frequency of events according to severity of threat were separately examined according to whether they occurred shortly before the interview (in the preceding 13 weeks) or less recently (in the period 13-26 weeks before the interview]. For the more immediate events (i.e., ~13 weeks], events of absent threat were significantly more frequent in controls (4O%] than in patients (5%) (OR, 0.09; 95% CI, 0.02-0.38; P = 0.001). In the period of 13-26 weeks, there was a higher rate of events of high immediate threat for DU patients than for controls, respectively. This was not statistically significant (OR, 2.42; 95% CI, 1.00-5.85; P = 0.04). Goal frustration of acute events. Only acute events with high goal frustration over the whole 26-week period significantly distinguished patients from controls, with a greater proportion of patients zyxwvutsrqpo (21%) experiencing these events than controls (7%) (see Table 1; OR, 4.13; 95% CI, 1.30-13.13; P = 0.01). Analyses of goal-frustrating events occurring within 13 weeks of onset or interview and 13-26 weeks before controls, 102 and 114, respectively. This difference was not significant [odds ratio (OR], 0.93; 95% confidence interval (CI), 0.10-8.44; P = 0.501. A similar proportion of ulcer patients and controls had at least one independent antecedent event, 57% and 59%, respectively (OR, 0.92; 95% CI, 0.53-1.61; P = 0.78). Of the 10 different categories of events, the only one that differed significantly between patients and controls was “role change” or “social interaction,” with frequencies of 16% and 30%, respectively (OR, 0.47; 95% CI, 0.22-1.01; P = 0.011. This reflected the fact that controls were more involved than DU patients in a wide variety of interpersonal situations either at home or at work (being the basis for the role change or interaction category]. Severity of threat of acute events. Results for severity of threat and goal frustration according to group are summarized in Table 1. There was a marked trend for a greater proportion of DU patients than controls to experience at least one antecedent highly threatening event (P = 0.02). The controls in contrast experienced significantly more events that were free Table 2. Severity of Threat and Goal Frustration of Chronic Goal Frustration Severe High Ulcer and Control Subjects Controls n I%J P 2 (21 2 (21 0.60 1.80 0.19-16.52 16 [ZO] 8 (10) 0.06 2.42 0.94-6.28 41(51) 27 (33) 0.01 2.44 1.19-4.98 16 (20) 12 (14) 0.26 1.61 0.69-3.73 3 (4) 0.02 10.71 1.41-81.28 9 (111 0.002 7.51 2.64-27.64 Ulcer patients n [%1 Immediate threat Severe High Moderate Absent Difficulties in Duodenal 5 (61 22 (27) OR 95% CI M oderate 23 (28) 15 (18) 0.04 2.40 1.03-5.61 A bsent 24 (30) 22 (27) 0.37 1.42 0.65-3.14 NOTE. D ata given as proportion of subjects with at least one such event O -26 weeks before onset. December 1999 STRESS IN DUODENAL ULCER DISEASE Table 3. Combined Life-Event and Chronic Difficulty Predictors Factors considered for multivariate analysis Age Absent immediate threat (life event] High immediate threat (life event] High goal frustration (chronic difficulty] OR 95% CI P 1.20 1.00-1.45 0.05 0.14 0.05-0.42 0.0004 3.40 1.44-8.06 0.005 1631 Chronic difficulty predictors. A multivariate analysis of all chronic difficulty characteristics in relation to DU indicated that only difficulties with high goal frustration were significant predictors of DU 95% CI, 1.33-9.69; P = 0.01). disease (OR, 3.59; Combined Analysis of Acute Life Events and Chronic Difficulties Acute life-event and chronic difficulty variables related to DU in the initial analyses were NOTE. Results of multivariate conditional logistic regression analyincluded in a final logistic analysis. High immediate sis. threat of acute events (P -c0.01) and acute events of no immediate threat (P c 0.01) remained as independent onset or interview showed that significant differences discriminators of DU and control subjects. There was existed only in the zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA 13 w eeks before onset, when a also a trend for high goal frustration of chronic greater proportion of patients than controls experidifficulties to be a discriminator (P = O.O3)‘(Table 3). enced events of high goal frustration (OR, 4.99; 95% CI, 1.09-22.81; P = 0.04) and where in contrast a greater proportion of controls experienced events free Discussion of goal frustration (OR, 0.43; 95% CI, 0.21-0.91; There was no significant difference in the total P = 0.03). number of life events experienced by DU patients and Acute life-event predictors. The results of mulcontrols in the present study, a finding consistent with tivariate analysis of all life-event characteristics indiearlier studies using questionnaire methods to assess cate that acute events with high immediate threat [OR, acute life-event stress (1,2,4). However, a significant 3.40; 95% CI, 1.43-8.06; P = O.Ol), and absent immediassociation between threat and DU disease was found ate threat (OR, 0.15; 95% CI, 0.05-0.42; P c 0.01) were in the present study. Duodenal ulcer patients experistatistically significant independent discriminators of enced significantly more events of high immediate DU subjects and controls. There was also a trend for threat, whereas events with no immediate threat were high goal frustration to be a discriminator, but this did more common among controls. This finding is somenot reach significance at the 0.01 level (OR, 3.41; 95% what at odds with earlier ones that threatening stresCI, 1.15-10.06; P = 0.03). sors seemed to play no role in DU disease (18,6). A possible explanation for this inconsistency was the Chronic Difficulties fact that one of these studies did not differentiate acute Number and type of chronic difficulties. from chronic stressors and involved a very small Duodenal ulcer patients experienced significantly more sample (12 DU patients) (61. independent antecedent chronic difficulties than conAcute events of high immediate threat were related trols (95 and 63, respectively; OR, 1.68; 95% CI, to DU disease in the present study. The great majority P -eO.OOl]. Similarly, a greater proportion of 1.14-2.46; zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA of duodenal ulcer patients in the present study experiulcer patients (73%) than controls (52%) experienced enced only one highly threatening event; moreover, the threat of these acute events seemed quite transient at least one independent antecedent chronic difficulty (OR, 2.89; 95% CI, 1.26-6.17; P < O.Ol]. Analysis by because longer term threats (such as 1 week later) category of chronic difficulty indicated that no particuwere not independently related to DU disease. Inilar type of chronic difficulty significantly differentitially it seems unlikely that an acute event of such ated DU patients from controls. brief impact could contribute significantly to the pathoSeverity of threat of chronic difficulties. Only physiological aetiology of DU. It is quite possible, chronic difficulties of moderate threat significantly however, that this acute yet brief “threat” may have differentiated DU patients from controls, with DU precipitated emotional distress such that those DU patients (51%) experiencing more of these difficulties patients with as yet undiagnosed disease experienced than controls (33%; P = 0.01) (Table 2). an exacerbation of ulcer pain to such a level that Severity of goal frustration. Only chronic diffimedical advice was sought and diagnoses made. This culties of high goal frustration significantly distinhypothesis remains to be tested. guished between DU patients and controls, with a The finding in the present study that DU patients greater percentage of DU patients (27%) than controls experience significantly more chronic difficulties than (11%) having this type of difficulty (Table 2). controls is consistent with results both from our earlier 3.41 1.15-10.06 0.03 1832 ELLARD ET AL. GASTROENTEROLOGY Vol. 99, No. 8 chronic duodenal ulcer: a comparison of the rating of life vents study (6) and from Salim’s animal studies (24) There by duodenal ulcer patients and community controls. Stand J was also a definite trend for chronic difficulties that Gastroenterology 1985:20:583-588. were highly goal frustrating to be related to DU 3. Stevenson DK, Nabseth DC, Masuda M, Holmes TH. Life disease, which is consistent with the findings of Craig changes and the post-operative course of duodenal ulcer paand Brown (18). This suggests that the difficulties that tients. J Hum Stress 1979;5:19-28. 4. Feldman E. Psychosomatic factors in duodenal ulcer disease. frustrate important life goals and the concomitant Brain Res Bull 1980;5:39-42. emotional responses (likely to be anger and frustra5. Wolcott DI, Wellish DK, Robertson CR, Arthur RJ. Serum tion) associated with those difficulties induce the gastrin and the family environment in duodenal ulcer disease. physiological changes associated with the onset or Psychosom Med 1981;43:501-507. relapse of DU disease. 8. Gilligan I, Fung L, Piper DW, Tennant C. Life events stress and chronic difficulties in duodenal ulcer: a case control study. J In summary, the current study adds to the existing Psychosom Res 1987;31:117-123. body of knowledge of stress and DU disease by 7. Tennant CT, Smith A, Bebbington P, Hurry J. The contextual distinguishing between acute and chronic stressors threat of life events: the concept and its reliability. Psychological and considering further aspect of stressors, such as Med 1979;9:525-528. goal frustration. The results suggest that acute tran8. Brown GW, Harris TH. The social origins of depression: a study sient threatening stressors may affect gut symptoms, of psychiatric disorder in women. London: Tavistock, 1978. 9. Brown GW, Harris TD (ed). Life events and illness. New York: and in turn illness-related behavior such as seeking Guilford, 1989. medical attention, which leads to the diagnosis of DU 10. Finlay-Jones RA, Brown GW. Types of stressful life events and disease. Chronic goal-frustrating stressors of longer onset of anxiety and depressive disorders. Psychological Med than 6 months’ duration also seem important and may 1981;11:803-815. well be linked to the underlying and enduring patho11. Leff J, Kuippers L, Berkowitz R, Vaughn C, Sturgeon D. Life events, relatives expressed emotion and maintenance neurolepphysiology of DU disease. zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA Appendix In the LEDS system, the contextual threat is a measure of how emotionally distressing or unpleasant the average person would find an event or difficulty given similar personal, biographical, and current circumstances. This rating is made for the immediate impact of the event (short-term threat] and for the implications experienced 1 week later (long-term threat). Goal frustration measures the degree to which the subject’s aims, needs, and ambitions are obstructed by the occurrence of the event or difficulty. Thus, for example, the death of a close relative seen frequently would be far more threatening (threat rating of 1 or 2) than the death of a distant relative seen infrequently (threat rating of 3). Similarly, a goal-frustration rating would be high for a divorce or separation where one particular partner was denied access to the children (goal frustration rating of 1 or 2) compared with an “uncontested separation and divorce” that did not include children (absent goal frustration rating of 4). (These are extremely abbreviated vignettes for the purpose of illustrating the ratings.] tics in schizophrenic relapse. Psycho1 Med 1983:13:799-808. 12. Costello CG. Social factors associated with depression: a retrospective community study. Psycho1 Med 1982;12:329-339. 13.Murphy E. Social origins of depression in old age. Br J Psychiatry 1982;141:135-142. 14. Connolly J. Life events before myocardial infarction. J Hum Stress 1978;2:3-17. 15. Penrose RJJ. Life events before subarachnoid haemorrhage. J Psychosom Res 1972;16:329-333. 18. Murphy E, Brown GW. Life events, psychiatric disturbance and physical illness. Br J Psychiatry 1980;138:328-338. 17. Creed F. Life events and appendicectomy. Lancet 1981;1:381385. 18. Craig TKJ. Brown GW. Goal frustration and life events in the aetiology of painful gastrointestinal disorder. J Psychosom Res 1984;128;5:411-421. 19. Alexander F. Psychosomatic medicine. New York: Norton, 1950. 20. Weisman AD. A study of the psychodynamics of duodenal ulcer exacerbations with special reference to treatment and the problems of specificity. J Psychosom Res 1958;18:2. 21. Congalton AA. Occupational status in Australia: studies in sociology. No. 3. Sydney, Australia: Sydney School of Sociology, University of New South Wales, 1983. 22. Parry G, Shapiro DA, Davies L. Reliability of life event ratings: an independent replication. Br J Clin Psycho1 1979;20:133-134. 23. Breslow NE, Day NE. Statistical methods in cancer research. The analysis of case control studies. Volume 1. Lyon: IARC, 1980. 24. Salim AS. Stress and the adrenergic hypothalamouagel pathway and the aetiology of chronic duodenal ulceration. J Psychosom Res 1987;31:231-237. References 1. Piper DW, McIntosh 2. JH, Ariotti DE, Calogiuri JV, Brown RW, Shy CM. Life events and chronic duodenal ulcer: a case control study.Gut 1981;22:1011-1017. McIntosh HH, Nasiry RW. McNeil D, Coates C, Mitchell J, Piper DW. Perception of life event stress in patients with Received January 31.1990. Accepted May 28,199O. Address requests for reprints to: Professor C. Tennant, Department of Academic Psychiatry, Royal North Shore Hospitdl, St. Leonards, New South Wales 2085, Australia.