Dyspepsia and Irritable Bowel Syndrome After a Salmonella Gastroenteritis Outbreak: One-Year Follow-up Cohort Study

doi:10.1053/j.gastro.2005.04.012F A * d
B p l e n u f S t b a 1 v p D d o c w t s j A w w a p w s p d
F d p o c p D u a
GASTROENTEROLOGY 2005;129:98–104
yspepsia and Irritable Bowel Syndrome After a Salmonella astroenteritis Outbreak: One-Year Follow-up Cohort Study
ERMÍN MEARIN,* MARC PÉREZ–OLIVERAS,‡ ANTONIA PERELLÓ,* JAUME VINYET,‡
NABEL IBAÑEZ,‡ JORDI CODERCH,§ and MÓNICA PERONA* Institute of Functional and Motor Digestive Disorders, Centro Médico Teknon, Barcelona; and ‡ABS Torroella de Montgrí and §Direcció
’Avaluació, Informació i Recerca, Serveis de Salut Integrats Baix Empordà, Girona, Spain
t p l t h s p o
6 e s U g
b s t o t o o n i t a h p p
R p w d
c i
ackground & Aims: It has been reported that some atients develop functional digestive disorders, particu-
arly irritable bowel syndrome (IBS), after acute gastro- nteritis (AGE). However, the presence of dyspepsia has ot been specifically addressed. We prospectively eval- ated development of dyspepsia and IBS during a 1-year ollow-up in a cohort of adult patients affected by a almonella enteritidis AGE outbreak. Methods: Ques- ionnaires were sent to 1878 potential participants at aseline and 3, 6, and 12 months; 677 had experienced Salmonella enteritidis AGE on June 23, 2002, and
201 had not (randomly selected controls, matched for illage of residence, age, and sex). At 12 months, 271 atients and 335 controls returned the questionnaires. ata permitted the establishment of dyspepsia and IBS iagnosis by Rome II criteria. Results: Before the AGE utbreak, the prevalence of dyspepsia was similar in ases and controls (2.5% vs 3.8%); the prevalence of IBS as also similar (2.9% vs 2.3%). At 3, 6, and 12 months,
he prevalence of both dyspepsia and IBS had increased ignificantly in exposed compared with unexposed sub- ects. Overlap between dyspepsia and IBS was frequent. t 1 year, the relative risk for development of dyspepsia as 5.2 (95% confidence interval, 2.7–9.8) and for IBS as 7.8 (95% confidence interval, 3.1–19.7). Prolonged bdominal pain and vomiting during AGE were positive redictors of dyspepsia. No predictive factors for IBS ere found. Conclusions: Salmonella gastroenteritis is a ignificant risk factor not only for IBS but also for dys- epsia; at 1 year of follow-up, 1 in 7 and 1 in 10 subjects eveloped dyspepsia or IBS, respectively.
unctional gastrointestinal disorders (FGIDs), includ- ing functional dyspepsia and irritable bowel syn-
rome (IBS), are very frequent and imply significant ersonal, social, and economic consequences.1 The etiol- gy and pathophysiology of FGIDs are not completely lear, and no single physiologic abnormality can be im- licated as the cause of symptoms in every patient. iverse pathophysiologic mechanisms appear to contrib- te to FGIDs, including altered motility, visceral hyper-
lgesia, brain-gut disturbances, genetic and environmen-
al factors, and psychosocial upsets, among others.2 Some atients with IBS report the onset of symptoms as fol- owing an episode of acute gastroenteritis (AGE); thus, heories of infectious and inflammatory etiologies for IBS ave been proposed.3 A similar situation has been de- cribed in some cases of functional dyspepsia,4 although redisposition to persistence of dyspepsia after an episode f AGE is more controversial.5
A retrospective survey found that approximately 1 in individuals reported acute onset of IBS following an
pisode of AGE. Thus, patients attending either hospital pecialists in the United States or general practice in the nited Kingdom attributed their IBS to an attack of astroenteritis in 6% and 17%, respectively.6
Some prospective studies also found a relationship etween AGE and subsequent IBS. After an outbreak of almonellosis, 31% of patients developed new IBS symp- oms that were still present 1 year after infection.7 An- ther study examined 75 individuals with AGE admitted o an infectious disease unit, of whom 25% had devel- ped new IBS when assessed 6 months after infection, an utcome confirmed by a more detailed study of mecha- isms 3 years later.8,9 A community-based study of 357 ndividuals with culture-positive bacterial AGE found hat 7% met Rome I criteria for IBS at 6 months, lthough 25% reported a persistent change in bowel abit.10 A further community survey evaluating only atients with Campylobacter gastroenteritis confirmed this ercentage, with 9% new IBS cases.11
More recently, a published case-control study using ome II criteria for IBS showed an incidence of new ostinfectious IBS of 16.7% over 6 months compared ith just 1.9% of controls (odds ratio, 10; 95% confi- ence interval [CI], 3–31).5
Abbreviations used in this paper: AGE, acute gastroenteritis; CI, onfidence interval; FGID, functional gastrointestinal disorder; IBS, rritable bowel syndrome.
© 2005 by the American Gastroenterological Association 0016-5085/05/$30.00
doi:10.1053/j.gastro.2005.04.012
A w e a n w f w p w
S l i t c a v w d w c s s t a m l p c h a
a s a c l t w t
v p p
i fi r o t
n i m n i
F p q
July 2005 DYSPEPSIA/IBS AFTER SALMONELLA GASTROENTERITIS 99
However, postinfectious dyspepsia has not been exten- ively examined and, to our knowledge, the only pub- ished data failed to show a statistically significant dif- erence in the incidence of dyspepsia between cases and ontrols at either 3 or 6 months postinfection.5
In June 2002, an outbreak of Salmonella enteritidis GE occurred in a Spanish village affecting 1243 people, hich provided a unique opportunity to evaluate the
volution of gastroenteritis in a large cohort of subjects fter a single outbreak of AGE by the same microorga- ism and of simultaneous onset and compare the results ith those of matched controls from the same village
ollowed up under the same conditions. Thus, our aim as to prospectively evaluate the development of dys- epsia and IBS during a 1-year follow-up in patients ith S enteritidis AGE.
Materials and Methods
Study Design
On June 23, 2002, an outbreak of AGE occurred in a panish village of 9004 inhabitants (Torroella de Mongrí) ocated in the Baix Empordà county (Catalonia) with 106,828 nhabitants. The cause was S enteritidis colonization of tradi- ional cream cakes, all made in the same baker’s shop to elebrate Saint John’s Eve. A total of 1243 persons were ffected; about 40% were Torroella residents, 40% from other illages of the same county, and 20% visitors. Information ith demographic (name, age, sex, and address) and clinical ata related to the outbreak of AGE (symptoms and treatment) ere available from the primary health center database of the
ounty. Thus, we decided to conduct a prospective cohort tudy only in Baix Empordà inhabitants, attended to in the ame county, because this was a unique opportunity to evaluate he evolution of gastroenteritis in a large sample of subjects fter a single outbreak of AGE of a single origin by the same icroorganism and of simultaneous onset and similar fol-
ow-up circumstances. A prospective cohort study was lanned; unexposed participants (ratio of approximately 2 ontrols per 1 case) were randomly selected from the primary ealth center database and matched for village of residence, ge, and sex.
A subject was considered to have AGE if diarrhea, fever, and bdominal pain appeared on June 24, 25, 26, or 27, 2002; in ome, but not all, patients, S enteritidis AGE was confirmed by
positive stool culture. However, given the epidemiologic ontext, stool culture was not required to diagnose salmonel- osis. Controls did not experience any of the previously men- ioned symptoms during the outbreak of salmonellosis, hether or not they had been in contact with the cream cake
hat originated the outbreak. Initially, 677 adults with AGE and 1201 unexposed indi-
iduals were detected; thus, 1878 questionnaires were sent to otential participants. A total of 1208 (481 AGE-exposed
ersons and 547 unexposed persons) returned the question- l
aires. Responses were reviewed and matched with the clinical harts in the database; 467 were confirmed as AGE-exposed ersons and 561 as having been unexposed. A flow chart of ecruitment is shown in Figure 1.
Follow-up was made at 3, 6, and 12 months to examine the requency of FGIDs, mainly dyspepsia and IBS. All subjects ere requested to complete a self-administered questionnaire
nd return it by mail. Each questionnaire was mailed twice; if o reply was obtained, subjects were contacted by telephone to ncourage participation. No personal interviews were held, ither with patients or controls, to avoid interference with esponses and potential clinician assessment bias.
Outcome Measure
All potential participants were requested to complete a elf-administered questionnaire at study entry and at 3, 6, and 2 months. The questionnaire was based on the modular Rome I questionnaire, which incorporates basic questions for clinical iagnosis of functional dyspepsia and IBS.12 It contained 5 uestions (a total of 8 items) related to dyspepsia and 4 uestions on intestinal symptoms (a total of 9 items). The uestionnaires referred to symptoms in the previous 3 months. On entry into the study, subjects were requested to specif-
cally exclude symptoms related to their recent infection. This rst questionnaire included a separate sheet with 9 questions egarding ingestion of the bakery product implicated in the utbreak of salmonellosis, gastrointestinal symptoms during heir AGE bout, and treatment received, if any.
Diagnosis of dyspepsia and IBS was based on the question- aire responses according to Rome II clinical criteria, accept- ng 25% (3 weeks) of the previous 3 months (12 weeks) as the inimum duration of symptoms. Thus, dyspepsia was diag-
osed when persistent or recurrent pain or discomfort centered n the upper abdomen were present and not exclusively re-
igure 1. Flow chart of recruitment. Response rates related to the otential participants in each group are shown in parentheses. Q, uestionnaire.
ieved by defecation or were not associated with onset of
c s o
W f c d d
a t g l i b c a t l w 1 S
i o e A c t a c d D
s w i
o S p s
p l o u o a h w N d t m 2
T
F
100 MEARIN ET AL GASTROENTEROLOGY Vol. 129, No. 1
hange in stool frequency or form. When patients had dyspep- ia and met criteria for IBS, they were considered to present an verlap between dyspepsia and IBS.
The study was approved by the ethics committee of the entro Médico Teknon and the research committee of Serveis e Salut Integrats Baix Empordà.
Data Management and Statistical Analysis
Two Microsoft Access (Microsoft Corp, Redmond, A) databases were created: one with the administrative data
or registry and control of ensuing mailings, and the other to ollect and validate the responses of each participant. Validated ata were transferred to the software package SPSS for Win- ows 10.0 (SPSS Inc, Chicago, IL) for statistical analysis. Bivariate analysis was performed with respect to age, sex,
nd prevalence of previous dyspepsia and IBS at each point of he study to assess the comparability of exposed and unexposed roups. The prevalence of each functional disorder was calcu- ated at each point of the study with the observed frequencies n exposed and unexposed groups to evaluate the association etween the AGE and dyspepsia or IBS. The following asso- iation measures were calculated at 12 months: relative risk, ttributable risk, and number of persons needed to be exposed o AGE for one case of dyspepsia or IBS to appear. For the atter analysis, participants with dyspepsia or IBS before AGE ere excluded and numbers of new cases of dyspepsia or IBS at 2 months of follow-up (cumulative incidence) were used. ignificance was assessed by calculating the 95% CI.
In a separate analysis, possible factors that might have nfluenced the development of dyspepsia or IBS after the utbreak of salmonellosis were evaluated in the group of xposed subjects. Participants with FGID symptoms before the GE were excluded from this analysis. Demographic and
linical data were compared between subjects with postinfec- ious dyspepsia or postinfectious IBS at the end of follow-up nd those without gastrointestinal symptoms at that time. The omparison was performed using Student t test for indepen- ent data and 2 test (using Fisher exact test when necessary). ifferences were considered significant when the P value was .05.
Results
Demographic and clinical data at study entry are hown in Table 1. Almost all (96.1%) of the patients ith AGE had eaten the infected cake; the origin of the
able 1. Demographic Data and Prevalence of Dyspepsia an
Patients (n
ge (SD)/range (y) 48.5 (17.3 ex (% women) 55. ad eaten cake (%) 96. revalence of dyspepsia before AGE (%) 2. revalence of IBS before AGE (%) 2.
nfection in the remainder could have been interpersonal 1
r another product from the same bakery also containing almonella. In the control group, 6.8% of subjects re- orted having eaten some type of cake bought at the ame bakery where the infected cake was made.
The prevalence of dyspepsia before AGE was similar in atients and controls (2.5% vs 3.8%), as was the preva- ence of IBS (2.9% vs 2.3%). At month 3, the prevalence f dyspepsia and IBS had not changed significantly in the nexposed group; however, a significant increase was bserved in patients who had had AGE (Figure 2). At 6 nd 12 months of follow-up, the prevalence of dyspepsia ad slightly decreased compared with that at 3 months, hereas the prevalence of IBS had slightly increased. evertheless, at all evaluated points, the prevalence of
yspepsia and IBS had increased more than 3-fold that of he initial prevalence. A significant number of patients et clinical criteria for both dyspepsia and IBS (between
3% and 36% at each point of follow-up; Figure 3). In
Before AGE in Patients and Controls
67) Controls (n 561) P
–93 49.8 (19.5)/18–95 NS 57.5 NS 6.8 .0001 3.8 NS 2.3 NS
igure 2. Changes in the prevalence of dyspepsia and IBS during the
d IBS
p 3 p 5 p g l a 3 o
t e l
d f s d a A I i o 3 e 8 6 p p l p w p
o w p g l u o I p
d s g a 5 e p y ( p (
F s I
July 2005 DYSPEPSIA/IBS AFTER SALMONELLA GASTROENTERITIS 101
atients with dyspepsia, the overlap with IBS was 35%, 3%, and 46% at 3, 6, and 12 months, respectively. In atients with IBS, the overlap with dyspepsia was 70%, 4%, and 62% at 3, 6, and 12 months, respectively. Two articipants in the exposed group and one in the control roup presented a chronic organic illness (ulcerative co- itis and a history of colon cancer in the exposed group nd ulcerative colitis in the control group). None of the participants presented with dyspepsia or IBS at the end
f follow-up. Excluding participants with previous FGID symp-
oms, at 1 year, dyspepsia had appeared in 46 of 267 xposed participants and 11 of 330 controls; this cumu- ative incidence is shown in Figure 4. According to these
igure 3. Overlap between ymptoms of dyspepsia (D) and BS during the 1-year follow-up.
igure 4. Cumulative incidence (new cases) of dyspepsia and IBS at
aach point of the study during the 1-year follow-up.
ata, the relative risk of developing dyspepsia at 1 year of ollow-up was 5.2 (95% CI, 2.7–9.8) and the number of ubjects needed to have S enteritidis AGE for one case of yspepsia to appear was 7.2 (95% CI, 6.9–11.3). The ttributable risk of developing dyspepsia to S enteritidis GE was 13.9 (95% CI, 8.9–18.8). Similarly, at 1 year,
BS had appeared in 31 of 266 exposed participants and n 5 of 333 controls; therefore, the relative risk of devel- ping IBS at 1 year of follow-up was 7.8 (95% CI, .1–19.7), the number of subjects needed to have S nteritidis for one case of IBS to appear was 9.9 (95% CI, .6–12.1), and the attributable risk was 10.1 (95% CI, .1–14.2). As expected in FGIDs, symptoms were not resent during the entire study period. Thus, 47% of atients with dyspepsia at the end of the 1-year fol- ow-up had had this diagnosis at least at one other revious point of follow-up. In the same way, for patients ith IBS at 12 months, the rate of concordance with revious diagnosis was very similar (48%). At 12 months, 33 exposed participants had symptoms
f dyspepsia; in 54% of cases, the predominant symptom as pain. The remainder of patients with dyspepsia resented with epigastric discomfort consisting of epi- astric bloating (73%), postprandial fullness (55%), and, ess frequently, nausea (18%); weight loss was not eval- ated. Twenty-four exposed participants had symptoms f IBS; diarrhea-predominant IBS was the most frequent BS subtype in 17 cases (70.8%), whereas constipation- redominant IBS was present in only 7 cases (29.2%). Some risk factors for development of postinfectious
yspepsia were detected. The following variables were ignificantly higher in the postinfectious dyspepsia roup: vomiting during AGE (71% vs 41%; P .003) nd days with abdominal pain during AGE (9.3 days vs .1 days; P .001). No statistically significant differ- nces were found between patients with and without ostinfectious dyspepsia with respect to age (46 vs 48 ears), emergency department attendance during AGE 16% vs 9%), or hospitalization (10% vs 3%). Female sex redominated in the postinfectious dyspepsia group 74% vs 55%; P .051).
No risk factors for development of IBS were found,
nd there were no statistically significant differences
b w 5 e 9
A h 9
m u t p t p S i a e a e b a c t
o 1 7 d s t v e s c T i t
f t 3 C d d 1
I n
o i f 1 d d fi o i f m d t f c o c y e t t
d s 1 p u o t
t f t W p s w t p o l d
f n t s s
102 MEARIN ET AL GASTROENTEROLOGY Vol. 129, No. 1
etween patients with and without postinfectious IBS ith respect to age (45 vs 49 years), sex (63% men vs 4% women), vomiting during AGE (61% vs 42%), mergency department attendance during AGE (13% vs %), or hospitalization beyond 3 days (8% vs 3%). Antibiotic treatment was administered in only 8% of
GE cases. The prevalence of dyspepsia and IBS was igher in this small subgroup (20% vs 12% and 17.6 vs .3%, respectively; P NS).
Discussion
Several previous studies related AGE to develop- ent of IBS. Between 6% and 17% of patients attrib-
ted their IBS to an attack of gastroenteritis.6 Prospec- ive investigations also found an increase in the revalence of IBS after AGE, but data vary depending on he severity of the infection (ambulatory or hospitalized atients) and infectious agent (with Campylobacter and higella carrying more risk than S enteritidis). Moreover, n prospective studies, no control group was included to scertain the spontaneous appearance of IBS, with the xception of those reported by Parry et al5 and Wang et l.13 We had an exceptional opportunity to evaluate the volution of gastroenteritis of a single AGE origin caused y the same microorganism and of simultaneous onset in large cohort of subjects and compare it with matched
ontrols from the same small village followed up under he same circumstances.
In our study, the cumulative incidence of IBS at 1 year f follow-up was 11.6% in postinfectious patients and .5% in controls, yielding a significant relative risk of .8. McKendrick et al7 found that 31% of patients eveloped new symptoms of IBS after an outbreak of almonellosis that were still…