Identification of a Prodromal Period inCrohn’s Disease But Not Ulcerative ColitisMark Pimentel, M.D., Michael Chang, B.A., Evelyn J. Chow, B.A., Siamak Tabibzadeh, M.D.,Viorelia Kirit-Kiriak, M.D., Stephan R. Targan, M.D., and Henry C. Lin, M.D.GI Motility Program and Inflammatory Bowel Disease Center, Department of Medicine, Cedars-SinaiMedical Center, CSMC Burns and Allen Research Institute, Los Angeles; and School of Medicine, Universityof California, Los Angeles, Los Angeles, California

OBJECTIVES: Irritable bowel syndrome, a common gastroin-testinal diagnosis, has not been clearly studied in inflam-matory bowel disease. Some of the residual symptoms insubjects treated with Crohn’s disease and ulcerative colitisare thought to be related to irritable bowel syndrome. Theaims of this study were 1) to describe the duration andnature of complaints before the diagnosis of Crohn’s diseaseand ulcerative colitis (prodromal period), and 2) to deter-mine the role of IBS in this prodromal period.

METHODS: A total of 66 patients with confirmed inflamma-tory bowel disease were enrolled in the study. The subjectsreceived a questionnaire to ascertain the nature and durationof symptoms preceding the diagnosis of Crohn’s disease orulcerative colitis, including features described under theRome criteria for irritable bowel syndrome.

RESULTS: Of the 66 subjects analyzed, 45 had Crohn’sdisease and 21 had ulcerative colitis. The prodromal periodwas 7.76 10.7 yr for Crohn’s disease and 1.26 1.8 yr forulcerative colitis (p , 0.05). Once patients meeting theRome criteria for irritable bowel syndrome during the pro-drome were excluded, the duration of the prodromal period(non-IBS) for ulcerative colitis dropped to 0.86 1.3 yrcompared to 6.96 9.8 yr in the Crohn’s disease group (p ,0.05). The symptoms of the non-IBS prodrome in subjectswith Crohn’s disease were bloating, diarrhea, stomach pain,heartburn, fever, weight loss, and fatigue. Further analysisdemonstrated that subjects whose Crohn’s disease initiallybegan as colonic disease had a longer prodrome than withsmall bowel. In the non-IBS Crohn’s group, there was alsoa correlation between the age at the time of diagnosis andthe duration of prodrome (r5 0.67,p , 0.0001).

CONCLUSIONS: There is a significant prodromal period be-fore the time of diagnosis of Crohn’s disease that is notfound in ulcerative colitis even after exclusion of subjectswith IBS. (Am J Gastroenterol 2000;95:3458–3462. © 2000by Am. Coll. of Gastroenterology)

INTRODUCTION

Crohn’s disease (CD) and ulcerative colitis (UC) are inflam-matory conditions of the gastrointestinal tract. Althoughboth conditions result in inflammation, there are significantdifferences in their behavior and natural history to suggest adifferent pathophysiology.

Irritable bowel syndrome (IBS) is a common functionalgastrointestinal problem. The prevalence of IBS in the gen-eral population is estimated to be as high as 30% (1, 2). Inaddition, up to 50% of gastroenterologist visits are related tofunctional bowel complaints (3). Despite successful treat-ment, CD and UC patients with quiescent disease oftencontinue to have non-specific gastrointestinal symptomssuch as bloating, gas, and alteration in bowel habits. WithIBS being so common, it is believed by some that theresidual symptoms in CD and UC are related to eitherpre-existing or superimposed IBS (4). In one study, IBS isseen in up to 33% of subjects with treated UC and persistentbowel complaints (5).

Although some authors suspect that residual symptoms ininflammatory bowel disease (IBD) are related to IBS (4), nostudy has yet examined the prevalence of IBS in IBD.Additionally, the gastrointestinal complaints preceding thediagnosis of IBD have not been evaluated. In this study, weset out to compare the duration and nature of complaintsbefore the diagnosis of CD or UC and to determine the roleof IBS in this period.

MATERIALS AND METHODS

Patient PopulationA total of 76 consecutive subjects who were referred fromthe community presented to the Cedars-Sinai InflammatoryBowel Disease Center and were asked to participate in thestudy. All subjects were enrolled in the study after informedconsent was obtained and were required to have biopsy-proven CD or UC. Indeterminate cases were excluded fromanalysis. The protocol used in this study was approved bythe Institutional Review Board at Cedars-Sinai MedicalCenter, Los Angeles, CA.

THE AMERICAN JOURNAL OF GASTROENTEROLOGY Vol. 95, No. 12, 2000© 2000 by Am. Coll. of Gastroenterology ISSN 0002-9270/00/$20.00Published by Elsevier Science Inc. PII S0002-9270(00)02154-7

QuestionnaireA questionnaire was used to determine the nature and du-ration of symptoms preceding the diagnosis of CD or UC.Subjects were asked to fill out the questionnaire during theiroffice visit. The questionnaire contained three sections: de-mographics, specifics of their current state of inflammatorybowel disease (IBD), and duration and nature of symptomsbefore the diagnosis of their IBD.

The first section of the questionnaire included questionspertaining to general demographics such as age, ethnicity,and sex. The second section inquired about the specifics oftheir current IBD including the duration, where the diseasestarted initially, and the current location of the disease. Thequestionnaire also asked patients if their disease was ini-tially diagnosed by colonoscopy, upper endoscopy, upperGI x-ray with small bowel follow-through, or barium en-ema. The last section of the questionnaire related to symp-toms and complaints made by the subjects before the diag-nosis of their IBD was made. The questionnaire asked,“Before the diagnosis of your disease, how long did youhave symptoms?” This was used to determine the durationof a prodromal period. This section included an evaluationof symptoms during the prodromal period. The question-naire asked the patients to characterize the prodrome basedon 15 symptoms, which included Rome criteria (6). The 15symptoms were bloating, gas, diarrhea, constipation, stom-ach pain, heartburn, vomiting, pain at night, fever, weightloss, blood in stool, joint pain, fatigue, mucous, and ur-gency. Finally, subjects were asked to describe tests per-formed during the prodrome that failed to define objectivelyan etiology for their symptoms. For all cases, a chart reviewwas done to confirm the disease location at the time ofdiagnosis including the test performed to diagnose the IBD.This was done to confirm UC or CD and as a test of recall.

Analysis of DataThe duration of the prodrome period was compared betweensubjects with UC and CD. Using the Rome criteria, thenumber of subjects with IBS was determined. The remain-ing patients were grouped as non-IBS subjects. The durationof prodrome in these non-IBS UC and CD subjects wascompared. A comparison was then made of specific symp-toms during the prodrome in non-IBS subjects for both UCand CD. To determine recall bias, a comparison was madebetween where their investigations suggest their diseasesstarted (by chart review) and where they believed it started(on questionnaire). Another comparison of the prodromewas then made by dividing CD subjects into three groupsaccording to the disease location when first diagnosed (co-lon only, colon and small intestine, and small intestineonly). This was done as small bowel disease may be moredifficult to diagnose, leading to a longer prodrome comparedto that of colon disease as seen in UC. As the length ofprodrome would intuitively depend on the age of onset, theage of onset of non-IBS CD and UC subjects and theduration of prodrome were compared.

Statistical AnalysisThe duration of prodromal period in CD and in UC (totalgroup and non-IBS) was compared using Student’st test.The results were expressed as a mean6 standard deviation.IBS prevalence in the prodrome of patients with UC and CDwas compared by thex2 test. The prodrome duration of IBSand non-IBS subjects was compared byt test. The qualita-tive data from the survey was transferred into a binary code(0 5 no, 15 yes) and analyzed using thex2 test. Qualitativedata included symptoms experienced during the prodromalperiod in CD as compared to UC. A linear regression anal-ysis was used to assess the relationship between the durationof the prodrome and the age at time of diagnosis for non-IBSCD and UC subjects.

RESULTS

DemographicsA total of 76 consecutive IBD subjects filled out a ques-tionnaire during their office visit. Nine subjects had inde-terminate disease and one subject failed to complete thequestionnaire. Of the 66 remaining subjects, 45 had CD and21 had UC. The patient demographics were similar in bothgroups (Table 1). There were 18 male and 27 female sub-jects with CD as well as 13 male and eight female subjectswith UC who participated in this study. The average dura-tion of disease was 9.26 9.4 yr in CD subjects and 9.768.9 yr in subjects with UC. Of the 45 subjects with CD, 19(42.2%) were diagnosed by upper GI x-ray with small bowelfollow-through. UC was most commonly diagnosed bycolonoscopy (85.7%).

Duration of Prodromal Symptoms in UC and CDIn the questionnaire, subjects were asked about the durationof time that they had gastrointestinal symptoms before thediagnosis of IBD. In all, 39 of 45 CD subjects and 19 of 21UC subjects responded to this question. Based on this, theprodromal period was significantly longer for CD subjects(7.7 6 10.7 yr for CD and 1.26 1.8 yr for UC,p , 0.01).

Table 1. Patient Demographics

Crohn’sDisease(n 5 45)

UlcerativeColitis

(n 5 21) p Value

Age (yr)* 426 14 466 14 NSSex (female)† 27 8 NSDisease duration (yr)* 9.26 9.4 9.76 8.9 NSHow was it diagnosed?†

Colonoscopy 30 18 NSEGD 5SBF 19Barium enema 7

EGD 5 esophagogastroduodenoscopy; NS5 not significant; SBF5 small bowelfollow-through.

Data for age and disease duration are mean6 SD.* Analyzed by Student’st test.† Analyzed byx2 test.

3459AJG – December, 2000 Prodromal Period in Crohn’s Disease But Not UC

IBD and IBSTen of 39 (25.6%) CD subjects and four of 19 (21.1%) UCsubjects described symptoms during the prodromal periodconsistent with IBS based on the Rome criteria (Table 2).This left 29 subjects defined as non-IBS CD subjects and 15defined as non-IBS UC subjects. The prevalence of IBSduring the prodromal period was not different between UCand CD (x2 5 0.147,p 5 0.70).

After subjects with IBS symptoms during the prodromalperiod were eliminated, the prodrome duration in non-IBSUC subjects dropped to 0.86 1.3 yr. The prodrome innon-IBS CD subjects, however, remained significantly pro-longed at 6.96 9.8 yr (p , 0.05 compared to UC) (Table3). The symptoms of this non-IBS prodrome that differedsignificantly in CD subjects compared to UC subjects werebloating, diarrhea, stomach pain, heartburn, fever, weightloss, and fatigue (Table 3).

When a chart review was performed to confirm the dis-ease location at the time of diagnosis in CD, there was a100% correlation between the questionnaire responses andthe medical records (r5 1, p , 0.0001) in the 39 CDsubjects who responded to the question.

When subjects were grouped according to the diseaselocation at the time of diagnosis, the prodrome in the CDsubjects as a whole and in the non-IBS CD subjects sub-group was longer in those whose CD was initially found inthe colon than in the small intestine (Table 4). In the CDgroup as a whole, the duration of prodrome was 11.46 13.7yr in the colon-only group and 4.96 7.3 yr in the smallintestine–only group. In the non-IBS group the prodromewas 10.96 14.1 yr in the colon-only group and 5.16 7.5yr in the small intestine–only group.

As the age of onset of CD could be a determinant ofprodrome duration, the age at the time of diagnosis and theduration of prodrome were compared. A significant corre-lation between age at onset of CD in non-IBS CD subjectsand duration of prodromal period was observed (r5 0.67,p , 0.0001) (Fig. 1). In contrast, the correlation in non-IBSUC subjects was not significant (r5 0.38,p 5 0.16) (Fig. 2).

When CD subjects were asked to self-report attempts atinvestigations of their symptoms during the prodrome, themajority of subjects claimed that they had been evaluated.Of the 29 non-IBS CD subjects, 23 (79.3%) stated they hadbeen through a diagnostic workup. Of the 23 subjects, 18had upper GI x-ray with small bowel follow-through with noevidence of disease. Twelve of these 18 subjects ultimatelyhad terminal ileum involvement.

Table 2. Comparison of UC and CD Subjects With IBS Prodromal SymptomsVersusThose With Non-IBS Prodromal Symptoms

Total Group IBS Subjects Non-IBS Subjects* p Value

Crohn’s disease (n5 39) (n5 10) (n5 29)Age at diagnosis of CD (yr) 33.26 15.5 36.26 15.2 31.86 15.2 NSDuration of CD (yr) 9.86 9.8 6.86 6.1 10.16 10.2 NSDuration of prodrome (yr) 7.76 10.7 9.96 13.2 6.96 9.8 NS

Ulcerative colitis (n5 19) (n5 4) (n5 15)Age at diagnosis of UC (yr) 36.26 14.6 32.36 16.7 37.06 7.7 NSDuration of UC (yr) 10.26 9.2 4.36 2.9 11.06 9.4 NSDuration of prodrome (yr) 1.26 1.8 2.76 2.8 0.86 1.3 0.06

Data are mean6 SD, analyzed by Student’st test.* Patients whose symptoms did not meet Rome criteria for IBS were grouped as non-IBS subjects.CD 5 Crohn’s disease; IBS5 irritable bowel syndrome; NS5 not significant; UC5 ulcerative colitis.

Table 3. Duration and Symptoms of Prodrome in Non-IBS UCand CD Subjects

Crohn’sDisease(n 5 29)

UlcerativeColitis

(n 5 15) p Value

Duration of prodrome (yr)* 6.96 9.8 0.86 1.3 ,0.05Symptoms before diagnosis

[number of subjects (%)]†Bloating 10 (34.5) 0 ,0.01Gas 14 (48.3) 3 (20.0) 0.07Diarrhea 15 (51.7) 3 (20.0) ,0.05Constipation 6 (20.7) 1 (6.7) NSStomach pain 19 (65.5) 3 (20.0) ,0.01Heartburn 7 (24.1) 0 ,0.05Vomiting 5 (17.2) 0 NSPain at night 9 (31.0) 1 (6.7) NSFever 8 (27.6) 0 ,0.05Weight loss 9 (31.0) 0 ,0.05Blood in stool 6 (20.7) 6 (40.0) NSJoint pain 3 (10.3) 0 NSFatigue 12 (41.4) 0 ,0.01Mucous 2 (6.9) 1 (6.7) NSUrgency 8 (27.6) 3 (20.0) NS

Duration of prodrome is shown as mean6 SD.Subjects whose prodromal symptoms did not meet Rome criteria for IBS were

grouped as non-IBS subjects.* Analyzed by Student’st test.† Analyzed byx2 test.Abbreviations as in Table 2.

Table 4. Duration of Prodrome (yr) in Crohn’s Disease WithRespect to Disease Location When Diagnosed

ColonOnly

Colon andSmall Intestine

Small IntestineOnly

Total group (n5 17) (n5 2) (n5 18)11.46 13.7 6.06 1.4 4.96 7.3

Non-IBS (n5 9) (n5 2) (n5 17)10.96 14.1 6.06 1.4 5.16 7.5

No differences seen between all three groups.

3460 Pimentel et al. AJG – Vol. 95, No. 12, 2000

DISCUSSION

In this study, we found that even after removing subjectswith symptoms meeting the Rome criteria for IBS and CDbut not UC is preceded by a prodrome of significant dura-tion. The average duration of the prodromal period is 6.969.8 yr in non-IBS subjects with CD and is only 0.86 1.3 yrin non-IBS subjects with UC. The epidemiology of Crohn’sdisease has been extensively studied with respect to riskfactors and genetic predisposition. However, this is the first

study that has examined the period before the time ofdiagnosis of these diseases to identify a phase of prodromalsymptoms.

The residual symptoms in IBD patients with quiescentdisease has often been attributed to IBS (4, 5). Since theprevalence of IBS in the general population is as high as30% (1, 2), a similar proportion of subjects with IBD mayhave IBS as well. To study the prodromal period of IBD weneeded to address the issue of coexisting IBS. In our study,the prevalence of IBS during the prodrome was 21.1% inUC subjects and 25.6% in CD subjects. This is consistentwith the prevalence in the general population and previousliterature on the IBS prevalence in UC (5). To remove IBSas a confounding factor, we took the approach of analyzingour data after excluding patients with IBS. We found thatthe prodromal period for UC subjects was reduced from1.2 6 1.8 yr to 0.86 1.3 yr after excluding IBS subjects.For non-IBS CD subjects, the prodromal duration remainedprotracted at 6.96 9.8 yr. A prodromal period of 0.8 yr canprobably be accounted for by the usual delays in time topresentation to a physician and in completing the workup.

The symptoms that characterized the prodrome in non-IBS CD subjects were bloating, diarrhea, stomach pain,heartburn, fever, weight loss, and fatigue. What this pro-dromal period represents is unclear. A possibility is thatalthough the Rome criteria were not met by these subjects,many of the symptoms described have a functional compo-nent and could represent IBS variants. This would, however,imply a much higher rate of IBS in the prodromal period ofCD subjects than in the general population or in patientswith UC. Another possibility is that this prodrome repre-sents undiagnosed CD. Crohn’s disease is more difficult todiagnose than is UC because of its frequent involvement ofthe small bowel as compared to the easily accessible colonin UC. However, in our study there is actually a longerprodrome associated with colonic Crohn’s (although notstatistically significant). This implies that the prodrome inCrohn’s disease is not due to a delay in detecting the moreelusive small bowel Crohn’s.

The relationship of the prodrome duration to the age attime of diagnosis revealed an interesting finding. The olderthe age at time of diagnosis of CD, the longer the prodromalduration. Perhaps this in some way relates to the pathogen-esis of CD in olderversusyounger affected subjects. Thisintriguing possibility remains to be determined. Alterna-tively, these data imply that perhaps the true onset of CD orits predisposing disease occurs early in life.

As the symptoms of the prodrome of CD exist over a longperiod of time and the average duration of CD at the time ofpresentation is 9.26 9.4 yr, description of prodromal symp-toms may be affected by recall bias. However, because theUC and CD subjects in this study have similar durations ofdisease and therefore could be equally affected by recallbias, our finding of the striking differences in the duration ofprodrome in CD but not in UC cannot altogether be ex-

Figure 1. Correlation between age at time of diagnosis of Crohn’sdisease subjects without irritable bowel syndrome and duration ofprodrome.

Figure 2. Correlation between age at time of diagnosis of ulcer-ative colitis subjects without irritable bowel syndrome and durationof prodrome.

3461AJG – December, 2000 Prodromal Period in Crohn’s Disease But Not UC

plained by recall bias. When recall bias was tested in the CDsubjects, all correctly recalled the location of disease com-pared to chart review. This again implies good recall.

In summary, in this article we identify a unique prodromefor subjects with CD but not UC. Further studies are neededto identify the cause of symptoms during this prodrome.Most importantly, understanding whether this prodromerepresents a predisposing condition, an IBS variant, or justundiagnosed CD may have important treatment implications.

Reprint requests and correspondence:Mark Pimentel, M.D.,Cedars-Sinai Medical Center, 8700 Beverly Boulevard, Room7511, Los Angeles, CA 90048.

Received Sep. 20, 1999; accepted Mar. 27, 2000.

REFERENCES

1. Drossman DA, Sandler RS, McKee DL, et al. Bowel patternsamong subjects not seeking health care. Gastroenterology 1982;83:529–34.

2. Thompson WG, Heaton KW. Functional bowel disorders inapparently healthy people. Gastroenterology 1980;79:283–8.

3. Thompson WG. The functional gastrointestinal bowel disor-ders. In: Drossman DA, ed. The functional gastrointestinaldisorders. Boston: Little, Brown, 1994:117–34.

4. Bayless TM, et al. Inflammatory bowel disease and irritablebowel syndrome. Med Clin North Am 1990;74:21–8.

5. Isgar B, Harman M, Kaye MD. Symptoms of irritable bowelsyndrome in ulcerative colitis in remission. Gut 1983;24:190–2.

6. Drossman DA, Richter JE, et al., eds. Functional gastrointesti-nal disorders: Diagnosis, pathophysiology and treatment: Amultinational consensus. Boston: Little, Brown, 1994.

3462 Pimentel et al. AJG – Vol. 95, No. 12, 2000