Non–Clostridium difficile Bacterial Infections Are Rare in Patients With Flares of Inflammatory Bowel Disease

Accepted Manuscript Non–Clostridium difficile Bacterial Infections Are Rare in Patients With Flares of Inflammatory Bowel Disease Yuri Hanada, MD, Sahil Khanna, MBBS, Edward V. Loftus, Jr., MD, Laura E. Raffals, MD, Darrell S. Pardi, MD

PII: DOI: Reference:

S1542-3565(17)31232-6 10.1016/j.cgh.2017.10.008 YJCGH 55502

To appear in: Clinical Gastroenterology and Hepatology Accepted Date: 7 October 2017 Please cite this article as: Hanada Y, Khanna S, Loftus Jr EV, Raffals LE, Pardi DS, Non–Clostridium difficile Bacterial Infections Are Rare in Patients With Flares of Inflammatory Bowel Disease, Clinical Gastroenterology and Hepatology (2017), doi: 10.1016/j.cgh.2017.10.008. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT -1-

[Category: Original Article] [Subcategory: Alimentary Tract]

RI PT

Non–Clostridium difficile Bacterial Infections Are Rare in Patients With Flares of Inflammatory Bowel Disease Yuri Hanada, MD

Edward V. Loftus Jr, MD

Darrell S. Pardi, MD

M AN U

Laura E. Raffals, MD

SC

Sahil Khanna, MBBS

Author Affiliations: Division of Gastroenterology and Hepatology (Drs Khanna, Loftus, Raffals, and Pardi), Mayo Clinic, Rochester, Minnesota. Dr Hanada was a student, Mayo Clinic School of Medicine, Mayo Clinic College of Medicine and Science. Dr Hanada is

TE D

now with the Department of Medicine, Johns Hopkins Hospital, Baltimore, Maryland. Reprints: Darrell S. Pardi, MD, Division of Gastroenterology and Hepatology, Mayo Clinic, 200 First St SW Rochester, MN 55905 ([email protected] Phone: 507-284-

EP

2407 Fax: 507-284-0538).

Author Contributions: Y.H., S.K., D.S.P.: Study concept and design,

AC C

acquisition of data, analysis and interpretation of data, statistical analysis, drafting of manuscript, critical revision of manuscript. D.S.P.: Funding, study supervision. L.E.R., E.V.L.: Critical revision of the manuscript. Conflict of interest: None.

Text word count (includes references and legends): 3,150 Abstract word count: 427 No. of tables: 3 No. of figures: 1

ACCEPTED MANUSCRIPT Hanada et al -2- (RJF)

Running title: Non-CDI Bacterial Infections in IBD Publisher: To expedite proof approval, send proof via email to [email protected].

AC C

EP

TE D

M AN U

SC

RI PT

©2017 Mayo Foundation for Medical Education and Research

ACCEPTED MANUSCRIPT Hanada et al -3- (RJF)

Abstract Background & Aims: Clostridium difficile infection (CDI) causes flares in patients with

RI PT

inflammatory bowel disease (IBD). We investigated the frequency and outcomes of nonCDI bacterial enteric infections in symptomatic patients with IBD.

Methods: We performed a retrospective study of patients with ulcerative colitis (UC) or

SC

Crohn’s disease (CD) from whom stool samples were collected and analyzed by PCR or culture for bacterial pathogens (Campylobacter jejuni or C coli, Salmonella species,

M AN U

Shigella species, entero-invasive Escherichia coli, shiga toxin–producing E coli, or Yersinia species) from November 19, 2011, through June 30, 2014. Patients were excluded if they had nonbacterial infections or no symptoms. Data were collected from medical records on IBD duration, treatment, age at diagnosis, and presence of concurrent CDI. Patients were followed for 1 year after the date of infection resolution or until date

TE D

of last follow-up in the health record. Each patient with an enteric infection was matched with 2 patients with IBD flares and negative results from stool tests (non-infected control) and 2 patients with IBD and CDI (CDI control), adjusted for age (within 5 years

EP

at the time of stool test), sex, and IBD subtype. Outcome measures included IBD therapy

AC C

escalation and hospitalization.

Results: Of 9247 patients with IBD seen during the study period, stool samples were tested from 1345 patients (50% with UC and 50% with CD). There were 3 positive results (detection of bacterial pathogens) from 339 PCR analyses of stool samples from 296 patients with UC (0.88%) and 12 positive results from 486 cultures of stool samples from 418 patients with UC (2.5%). There was 1 positive result from 355 PCR analyses of stool samples from 311 patients with CD (0.28%) and 9 positive results from 496 cultures of stool samples from 413 patients with CD (1.8%). Of the 19 patients followed beyond

ACCEPTED MANUSCRIPT Hanada et al -4- (RJF)

infection, 9 patients required escalation of their IBD therapy (47%)—most commonly addition of an immunomodulator (5 patients) or a biological agent (3 patients)—

RI PT

compared with 34% of CDI controls and 66% of non-infected controls (P<.001). Higher proportions of patients with non-CDI bacterial infections were in remission 1 year after their infection (89%) than patients with CDI (55%) or negative results of stool tests

visits, or surgical interventions among groups.

SC

(63%) (P=.04). We did not observe differences in hospitalization, emergency department

M AN U

Conclusion: In a retrospective study of patients with an IBD flare, we detected non-CDI bacterial infections in fewer than 3% of those who were tested. Higher proportions of patients with non-CDI bacterial infections were in remission in the year after their infection than patients with CDI.

AC C

EP

TE D

KEY WORDS: Aeromonas; complication; relapse; rehospitalization

ACCEPTED MANUSCRIPT Hanada et al -5- (RJF)

Abbreviations CD, Crohn disease

RI PT

CDI, Clostridium difficile infection IBD, inflammatory bowel disease

ICD-9, International Classification of Diseases, Ninth Revision

MALDI-TOF, matrix-assisted laser desorption/ionization time-of-flight

SC

PCR, polymerase chain reaction

AC C

EP

TE D

M AN U

UC, ulcerative colitis

ACCEPTED MANUSCRIPT Hanada et al -6- (RJF)

Introduction Inflammatory bowel disease (IBD) is a chronic condition characterized by

RI PT

relapsing inflammation of the gastrointestinal tract. The pathogenesis of IBD is complex but is generally attributed to an inappropriate immunologic response to otherwise

commensal flora in a genetically susceptible host after exposure to environmental triggers (1). Subsequent IBD flares may occur in association with treatment adjustment,

SC

nonsteroidal anti-inflammatory drug use, and lifestyle modifications (2).

Gastrointestinal tract infections are also commonly implicated as precipitants

M AN U

of IBD flares. Specifically, the association between IBD flares and Clostridium difficile infection (CDI) has been well documented. IBD is an independent risk factor for CDI, and patients with IBD have a higher incidence of CDI than persons without IBD (3). Additionally, IBD patients with CDI have poorer outcomes, including longer length of hospital stay, higher colectomy rates, and increased mortality (4).

TE D

Flares of IBD have been described with various non-CDI bacterial infections, including Campylobacter, Salmonella, and Shigella infections (5). However, relatively few studies have described the overall frequency of these infections among IBD patients

EP

experiencing flares or the influence of these particular infections on patient outcomes. A prospective study in Spain (6) evaluated gastrointestinal tract infections in IBD patients

AC C

over a 2-year period and identified 5 non-CDI bacterial infections among 99 flares in 79 patients. Campylobacter jejuni was the causative agent in all 5, and these patients did not appear to have worse outcomes. A subsequent study in Italy (7) over a 3-year period identified 3 infections (all due to C jejuni) among 143 flares in 113 patients, and an earlier study (8) identified 5 non-CDI infections (3 with enteropathogenic Escherichia coli; 1, Salmonella typhimurium; and 1, C jejuni) among 64 patients over a 7-month period.

ACCEPTED MANUSCRIPT Hanada et al -7- (RJF)

The objectives of the present study were to establish the frequency of symptomatic, non-CDI bacterial infections in IBD patients at a large tertiary care practice

and patients with flares associated with CDI. Methods

RI PT

and to compare their outcomes to patients with IBD flares not related to enteric infections

Data were collected from the Mayo Clinic electronic health records of

SC

inpatient and outpatient IBD patients who had stool polymerase chain reaction (PCR) testing or culture for enteric pathogens performed between November 19, 2011, and June

M AN U

30, 2014, as identified through microbiology records. With methods described previously (9), the PCR panel tests for C jejuni or Campylobacter coli, Salmonella species, Shigella species, enteroinvasive E coli, shiga toxin–producing E coli, and Yersinia species. Culture testing involved conventional culture techniques, microscopy, biochemical testing, serologic agglutination testing, and, if required, matrix-assisted laser

TE D

desorption/ionization time-of-flight (MALDI-TOF) mass spectrometry and sequencing. Basic demographics were recorded for all patients with a concurrent diagnosis of ulcerative colitis (UC) (International Classification of Diseases, Ninth Revision [ICD-9]

EP

codes 556.0-556.6, 556.8, and 556.9) or Crohn disease (CD) (ICD-9 codes 555.0-555.2 and 555.9) and confirmed by manual health record review. Exclusion criteria included

AC C

age younger than 18 years, identification of only nonbacterial organisms on stool culture, and lack of concurrent UC or CD diagnosis. Patients (n=3) were also excluded if they had no clinical indication for a stool test (eg, patient was asymptomatic, and the test was obtained for clinical trial screening). While the clinical relevance of a positive stool test would be better estimated when comparing patients with and without symptoms, patients were not routinely screened with stool testing if they did not have active gastrointestinal symptoms; given the small number of asymptomatic patients, we decided to exclude these patients from the current analysis.

ACCEPTED MANUSCRIPT Hanada et al -8- (RJF)

Data collected on underlying IBD included subtype, disease duration, and IBD medication use before infection. Collected data on the infection included age at

RI PT

diagnosis, duration from initial symptom presentation to diagnosis, presence of concurrent CDI, and treatment. IBD-specific treatments administered during the flare, such as intravenous corticosteroids, were recorded.

Patients were followed for 1 year after the date of infection resolution or until

SC

date of last follow-up in the health record, whichever was earlier. Outcome measures included use of corticosteroids and IBD therapy escalation, defined as an increased dose

M AN U

for at least 1 baseline medication or a new medication added within 1 month after infection resolution. Outcomes in the year after infection included subsequent flares managed in the outpatient setting, emergency department, or hospital; surgical intervention; and death.

Each IBD patient with a non-CDI bacterial infection was matched with 2

TE D

IBD patients with negative stool tests (noninfectious flare) and 2 IBD patients with CDI by age (within 5 years at the time of stool testing), sex, and IBD subtype. Clostridium difficile testing was done by PCR as previously described (10).

EP

JMP version 10.0.0 software (SAS Institute Inc) was used to perform statistical analyses, including descriptive statistics, Kaplan-Meier estimates, and

AC C

Wilcoxon rank sum test for comparison of time-dependent outcomes among matched cases; t test, 1-way analysis of variance, or χ2 analysis was performed as appropriate for

comparison of time-independent characteristics among matched cases. P less than .05 was considered significant. The Mayo Clinic Institutional Review Board approved this study. All authors had access to the study data and have reviewed and approved the final manuscript.

ACCEPTED MANUSCRIPT Hanada et al -9- (RJF)

Results Out of 9,247 IBD patients who were seen at Mayo Clinic in Rochester,

RI PT

Minnesota, during the study period, a total of 1,345 symptomatic IBD patients (50% with UC; 50% with CD) underwent stool testing for non-CDI bacterial infections: 49% were male, with a median age of 42 years (range, 18-92 years) at the time of stool testing. Among the patients with UC, 339 stool PCR tests in 296 patients yielded 3 positive

SC

results (0.88%), and 486 stool cultures in 418 patients yielded 12 positive results (2.5%). After an initial negative test, 31 UC patients had at least 1 additional stool PCR test, and

M AN U

50 had at least 1 additional stool culture with none being positive. Overall, 47 UC patients had both stool PCR and culture testing, with 46 patients having concordant negative results and 1 patient with a negative PCR and positive culture (Salmonella). In 1 patient, the presentation of UC coincided with the bacterial infection (Salmonella). The other patients had an established diagnosis of IBD.

TE D

Among the patients with CD, 355 stool PCR tests in 311 patients yielded 1 positive result (0.28%), and 496 stool cultures in 413 patients yielded 9 positive results (1.8%). After an initial negative test, 37 CD patients had at least 1 additional stool PCR

EP

test, and 55 had at least 1 additional stool culture, with none being positive. Overall, 46 CD patients had both stool PCR and culture testing, with 45 patients having concordant

AC C

negative results and 1 having a positive PCR (shiga toxin–producing E coli) and a negative culture.

Clinical characteristics of the 25 patients with positive tests and their

identified organisms are summarized in Table 1. Overall, 68% (n=17) of these infections were treated. Common reasons cited for proceeding with treatment included prolonged symptoms, concern for immunocompromised state due to IBD medications or liver transplant, or severe symptoms requiring hospitalization. There was no significant difference between treated and untreated patients in terms of changes in IBD-specific

ACCEPTED MANUSCRIPT Hanada et al -10- (RJF)

treatment during the flare (P=.41) or post-flare period (P=.41). No differences were identified between treated and untreated patients in flare remission rates (P=.36),

RI PT

hospitalization rates (P=.82), emergency department visits (no patients with a non-CDI infection reported to the emergency department), or surgical interventions (P=.35).

Concomitant CDI was identified in 5 patients (20%), 3 of whom were treated for both CDI and non-CDI bacterial infection, and 2 of whom were treated for only CDI.

SC

Nineteen patients (76%) had follow-up after the infection-related flare, for a mean duration of 1.4 years (range, 0.25-2.8 years); 9 of these patients (47%) required

M AN U

escalation of their IBD therapy within 1 month of infection (Table 2). Clinical characteristics of the 76 matched patients with negative stool testing or CDI are described in Table 3. Other than IBD duration, there were no statistically significant differences between the cohorts. Baseline IBD treatment was similar across cohorts, as shown in Table 2.

TE D

The IBD patients with noninfectious flare patients were more likely to have changes in IBD-specific treatment than patients with non-CDI bacterial flares or CDI flares, including increased dosing of baseline medications or addition of a 5-

EP

aminosalicylic acid agent (Table 2). In the post-flare period, patients with noninfectious flares were more likely to have a 5-aminosalicylic acid agent added than patients with

AC C

non-CDI bacterial flares or CDI flares. There were no significant differences in other medication changes between cohorts in the post-flare period. Patients with a non-CDI infection were significantly more likely to remain in

remission in the next year compared with those with CDI or a noninfectious flare (P=.04, Wilcoxon rank sum test). In CDI patients, 50% of the subsequent flares were associated with a positive CDI test. No differences in hospitalization rates, emergency department visits, or surgical interventions were identified between the cohorts (Figure).

ACCEPTED MANUSCRIPT Hanada et al -11- (RJF)

Discussion The incidence of non-CDI bacterial infections in this cohort of IBD patients

RI PT

was low. Furthermore, these infections did not result in worse outcomes when compared with outcomes among IBD patients experiencing CDI-related or noninfectious flares. This study confirms the low incidence of non-CDI bacterial infections, which was previously reported as between 2% and 9% (11). Notably, the yield with stool

SC

cultures was greater than the yield with PCR testing. While this suggests that bacterial culture may be more sensitive in detecting a non-CDI bacterial infection, the overall

M AN U

incidence of positive stool cultures was still low, as has been shown in other studies, with rates of non-CDI bacterial infection identified by stool culture between 1.8% and 4% (1214). Additionally, the present study highlights the extremely low yield (0%) of additional stool testing after an initial negative test, suggesting a high negative predictive power that renders subsequent stool testing unnecessary.

TE D

The clinical significance of a positive non-CDI test was not always clear, however. Stool testing identified various organisms, unlike the findings in 2 European studies in which Campylobacter was the predominant, or only, non-CDI bacterial agent

EP

identified (6,7,12). Some bacteria recovered from stool cultures are not necessarily enteric pathogens in immunocompetent patients and typically do not warrant treatment.

AC C

In particular, the role of Aeromonas in the pathogenesis of gastrointestinal tract disease is controversial (15,16), although there is evidence that Aeromonas can be pathogenic in

IBD patients (14,15) perhaps because of the increasing use of immunosuppressive medications earlier and for longer durations in IBD (17). Indeed, in the present study, therapeutic immunosuppression was a common justification for initiating antimicrobial treatment. This practice is supported by data that suggest that antimicrobial treatment may decrease infection severity for specific infections such as salmonellosis (18). However, treatment of the enteric infection does not always achieve symptomatic

ACCEPTED MANUSCRIPT Hanada et al -12- (RJF)

remission, indicating that IBD activation does not always resolve with removal of the inciting trigger (6,19). In fact, in this study, approximately one-third of patients were not

RI PT

treated with antibiotics for the bacterial infection in the setting of uncertainty as to the pathogenicity of the identified organism, self-limiting symptoms, or perceived benefit with escalation of IBD therapy alone. While this study preliminarily found no differences in the outcomes of patients who were treated versus those who were left untreated, more

SC

studies are needed to adequately determine differences in outcomes between treated and untreated non-CDI enteric infections in immunosuppressed IBD patients who may be

M AN U

more vulnerable to severe infections.

The medical literature has limited data on subsequent outcomes in IBD patients after a non-CDI bacterial infection. This study is the only known report in the literature that directly compares patients with non-CDI bacterial infection to those with CDI and those without any infection to determine differences in clinically relevant

TE D

outcomes. We found that within 1 year after infection, patients fared similarly to those with noninfectious flares of IBD or CDI-related flares in their need for hospitalization, emergency department care, and surgical intervention. In a small study of 5 IBD patients

EP

infected with C jejuni, a comparison with IBD patients experiencing noninfectious flares showed no difference in requirements for “rescue therapies,” including cyclosporine and

AC C

infliximab, or surgical intervention, which is consistent with the findings in our study (6). However, a study of C jejuni infections in a slightly larger cohort of 21 UC patients

identified worse outcomes (although not statistically significant) in terms of mortality and need for surgical intervention 1 year after diagnosis, suggesting that particular non-CDI infections may portend more aggressive disease flares (19). Notably, a large majority of studies in the current literature on this topic (7,8,12), including the most recent study (6), were conducted in Europe. While Campylobacter is a common cause of gastrointestinal disease in both the United States

ACCEPTED MANUSCRIPT Hanada et al -13- (RJF)

and Europe, it occurs with much greater incidence in Europe (estimated at 29.9-13,500 cases per 100,000 people) compared to the United States (14.3 per 100,000) (20). Indeed,

RI PT

Campylobacter was typically isolated as the most common cause of non-CDI bacterial infection in these European cohorts. In contrast, in our cohort, Aeromonas and

Salmonella were the most common organisms, and in fact, a recently published study from Columbia University authors found Escherichia coli instead to be the most common

SC

in their cohort (21), suggesting that underlying environmental, and possibly host, differences were involved.

M AN U

The present study has several limitations, most of which are inherent in its retrospective nature. Our institution has no formal protocol for microbiologic testing of patients presenting with IBD flares, and there are no formal organizational guidelines (22-24). Therefore, tests performed were determined at the discretion of the treating provider. Similarly, few patients were tested with both PCR and culture, which limits our

TE D

ability to directly compare these 2 tests. Additionally, the incidence of non-CDI bacterial infections was possibly underestimated, because patients who underwent only PCR testing may have had infection with an organism that could have been identified through

EP

culture testing (eg, Aeromonas), as described in this study. The small number of patients with non-CDI bacterial flares limited our ability to fully estimate the magnitude of these

AC C

infections on outcomes compared with CDI-related or noninfectious flares. Finally, it is acknowledged that interpretation of these data is challenging with the presence of concomitant CDI in 5 of the patients included in the study cohort. These patients were included, however, because a non-CDI bacterial infection was also identified. In conclusion, non-CDI bacterial infections are uncommon triggers for IBD flares, and these infections did not have a significant negative impact on patient outcomes. Given these findings, routine testing for infections other than CDI is not

ACCEPTED MANUSCRIPT Hanada et al -14- (RJF)

recommended. Should such testing be desired, stool culture, with its higher yield, might

AC C

EP

TE D

M AN U

SC

RI PT

be preferred over PCR for identifying a potential causative agent.

ACCEPTED MANUSCRIPT Hanada et al -15- (RJF)

References 1. Khor B, Gardet A, Xavier RJ. Genetics and pathogenesis of inflammatory bowel

RI PT

disease. Nature. 2011 Jun 15;474(7351):307-17. 2. Singh S, Graff LA, Bernstein CN. Do NSAIDs, antibiotics, infections, or stress trigger flares in IBD? Am J Gastroenterol. 2009 May;104(5):1298-313.

3. Razik R, Rumman A, Bahreini Z, McGeer A, Nguyen GC. Recurrence of

SC

Clostridium difficile infection in patients with inflammatory bowel disease: the RECIDIVISM Study. Am J Gastroenterol. 2016 Aug;111(8):1141-6.

M AN U

4. Trifan A, Stanciu C, Stoica O, Girleanu I, Cojocariu C. Impact of Clostridium difficile infection on inflammatory bowel disease outcome: a review. World J Gastroenterol. 2014 Sep 7;20(33):11736-42.

5. Irving PM, Gibson PR. Infections and IBD. Nat Clin Pract Gastroenterol Hepatol. 2008 Jan;5(1):18-27.

TE D

6. Navarro-Llavat M, Domenech E, Bernal I, Sanchez-Delgado J, Manterola JM, Garcia-Planella E, et al. Prospective, observational, cross-sectional study of intestinal infections among acutely active inflammatory bowel disease patients.

EP

Digestion. 2009;80(1):25-9.

7. Antonelli E, Baldoni M, Giovenali P, Villanacci V, Essatari M, Bassotti G.

AC C

Intestinal superinfections in patients with inflammatory bowel diseases. J Crohns Colitis. 2012 Mar;6(2):154-9.

8. Weber P, Koch M, Heizmann WR, Scheurlen M, Jenss H, Hartmann F. Microbic superinfection in relapse of inflammatory bowel disease. J Clin Gastroenterol. 1992 Jun;14(4):302-8.

9. Cunningham SA, Sloan LM, Nyre LM, Vetter EA, Mandrekar J, Patel R. Threehour molecular detection of Campylobacter, Salmonella, Yersinia, and Shigella

ACCEPTED MANUSCRIPT Hanada et al -16- (RJF)

species in feces with accuracy as high as that of culture. J Clin Microbiol. 2010 Aug;48(8):2929-33. Erratum in: J Clin Microbiol. 2011 Oct;49(10):3725.

RI PT

10. Sloan LM, Duresko BJ, Gustafson DR, Rosenblatt JE. Comparison of real-time PCR for detection of the tcdC gene with four toxin immunoassays and culture in diagnosis of Clostridium difficile infection. J Clin Microbiol. 2008 Jun;46(6):19962001.

SC

11. Lobaton T, Domenech E. Bacterial intestinal superinfections in inflammatory bowel diseases beyond Clostridum difficile. Inflamm Bowel Dis. 2016 Jul;22(7):1755-62.

M AN U

12. Mylonaki M, Langmead L, Pantes A, Johnson F, Rampton DS. Enteric infection in relapse of inflammatory bowel disease: importance of microbiological examination of stool. Eur J Gastroenterol Hepatol. 2004 Aug;16(8):775-8. 13. Meyer AM, Ramzan NN, Loftus EV Jr, Heigh RI, Leighton JA. The diagnostic yield of stool pathogen studies during relapses of inflammatory bowel disease. J

TE D

Clin Gastroenterol. 2004 Oct;38(9):772-5. 14. Ihekweazu FD, Ajjarapu A, Kellermayer R. Diagnostic yield of routine enteropathogenic stool tests in pediatric ulcerative colitis. Ann Clin Lab Sci. 2015

EP

Fall;45(6):639-42.

15. Holmberg SD, Farmer JJ 3rd. Aeromonas hydrophila and Plesiomonas shigelloides

AC C

as causes of intestinal infections. Rev Infect Dis. 1984 Sep-Oct;6(5):633-9. 16. Lobaton T, Hoffman I, Vermeire S, Ferrante M, Verhaegen J, Van Assche G. Aeromonas species: an opportunistic enteropathogen in patients with inflammatory bowel diseases? A single center cohort study. Inflamm Bowel Dis. 2015 Jan;21(1):71-8. 17. Mill J, Lawrance IC. Preventing infective complications in inflammatory bowel disease. World J Gastroenterol. 2014 Aug 7;20(29):9691-8.

ACCEPTED MANUSCRIPT Hanada et al -17- (RJF)

18. Szilagyi A, Gerson M, Mendelson J, Yusuf NA. Salmonella infections complicating inflammatory bowel disease. J Clin Gastroenterol. 1985 Jun;7(3):251-5.

RI PT

19. Arora Z, Mukewar S, Wu X, Shen B. Risk factors and clinical implication of superimposed Campylobacter jejuni infection in patients with underlying ulcerative colitis. Gastroenterol Rep (Oxf). 2016 Nov;4(4):287-92.

20. Kaakoush NO, Castano-Rodriguez N, Mitchell HM, Man SM. Global epidemiology

SC

of Campylobacter infection. Clin Microbiol Rev. 2015 Jul;28(3):687-720.

21. Axelrad JE, Joelson A, Nobel YR, Lawlor G, Green PHR, Lichtiger S, et al. Enteric

M AN U

infection in relapse of inflammatory bowel disease: the utility of stool microbial PCR testing. Inflamm Bowel Dis. 2017 Jun;23(6):1034-39. 22. Kornbluth A, Sachar DB; Practice Parameters Committee of the American College of Gastroenterology. Ulcerative colitis practice guidelines in adults: American College Of Gastroenterology, Practice Parameters Committee. Am J Gastroenterol.

TE D

2010 Mar;105(3):501-23. Erratum in: Am J Gastroenterol. 2010 Mar;105(3):500. 23. Rahier JF, Magro F, Abreu C, Armuzzi A, Ben-Horin S, Chowers Y, et al; European Crohn’s and Colitis Organisation (ECCO). Second European evidence-

EP

based consensus on the prevention, diagnosis and management of opportunistic infections in inflammatory bowel disease. J Crohns Colitis. 2014 Jun;8(6):443-68.

AC C

24. Lichtenstein GR, Hanauer SB, Sandborn WJ; Practice Parameters Committee of American College of Gastroenterology. Management of Crohn’s disease in adults. Am J Gastroenterol. 2009 Feb;104(2):465-83.

ACCEPTED MANUSCRIPT Hanada et al -18- (RJF)

Table 1. Characteristics of IBD Patients With Non-CDI Bacterial Infections (n=25) Feature

Value

Male, %

RI PT

Demographics 48.0

Age at time of stool test, median (range), y

60.0

IBD duration at time of flare, median, y

Culture only PCR only Both culture and PCR

M AN U

Stool testing, %

6.3

SC

Ulcerative colitis, %

44.0 (18.0-92.0)

76.0 16.0 8.0

Non-CDI bacterial organism, % Aeromonas species

32.0 28.0

Plesiomonas species

12.0

Campylobacter jejuni or Campylobacter coli

8.0

Yersinia species

8.0

EP

TE D

Salmonella species

4.0

Escherichia coli

8.0

AC C

Arcobacter species

O157:H7

4.0

Shiga toxin–producing

4.0

Abbreviations: CDI, Clostridium difficile infection; IBD, inflammatory bowel disease;

PCR, polymerase chain reaction.

ACCEPTED MANUSCRIPT

Hanada et al -19- (RJF)

Non-CDI Bacterial

Noninfectious Flare (n=38)

P Valueb

36.8

34.2

≥.05

36.8

36.8

21.1

≥.05

16.0

31.6

26.0

≥.05

47.4

55.3

86.8

.002

36.8

47.4

57.9

≥.05

0

2.6

23.7

.003

10.5

10.5

36.8

.009

5.3

7.9

2.6

≥.05

5.3

18.4

5.3

≥.05

47.4

34.2

66.0

.02

Increased dosing of baseline medication

5.3

0

0

≥.05

Addition of 5-ASA agent

5.3

2.6

34.2

<.001

Addition of immunomodulator

26.3

26.3

15.8

≥.05

Feature Baseline IBD treatment, %

31.6

Immunomodulator Biological

Intravenous or oral corticosteroids

TE D

IBD treatment adjustment during flare, %c

Increased dosing of baseline medication

Addition of immunomodulator

AC C

Addition of biological

EP

Addition of 5-ASA agent

M AN U

5-ASA agents

Post-flare IBD treatment escalation, %c

CDI (n=38)

SC

Infection (n=19a)

RI PT

Table 2. IBD Treatment by Matched Cohort

ACCEPTED MANUSCRIPT

Hanada et al -20- (RJF)

15.8

15.8

RI PT

Addition of biological

13.2

≥.05

Abbreviations: CDI, Clostridium difficile infection; 5-ASA, 5-aminosalicylic acid; IBD, inflammatory bowel disease. Of the 25 patients with a non-CDI bacterial infection, 19 had follow-up after the diagnosis of infection.

b

Wilcoxon rank sum test.

c

Several patients had more than 1 treatment adjustment or post-flare treatment escalation.

AC C

EP

TE D

M AN U

SC

a

ACCEPTED MANUSCRIPT Hanada et al -21- (RJF)

Table 3. Characteristics by Matched Cohort Non-CDI

Feature

Infection (n=19)

Demographics Male, %

47.4 51.0 (18.0-68.0)

(range), y

IBD duration at time of flare, median, y Known IBD at time of presentation, %

52.6

Flare (n=38)

47.4

47.4

49.5 (19.0-72.0)

49.5 (18.0-73.0)

52.6

52.6

6.3

6.3

4.4

94.7

94.7

82.0

M AN U

Ulcerative colitis, %

CDI (n=38)

SC

Age at time of stool test, median

Noninfectious

RI PT

Bacterial

AC C

EP

TE D

Abbreviations: CDI, Clostridium difficile infection; IBD, inflammatory bowel disease.

ACCEPTED MANUSCRIPT Hanada et al -22- (RJF)

Legend Figure. Survival by Matched Cohort. Each patient with inflammatory bowel disease

RI PT

(IBD) and a bacterial enteric infection other than Clostridium difficile infection (CDI) (non-CDI bacterial flare) was matched with 2 IBD patients with negative stool tests

(noninfectious flare) and with 2 IBD patients with CDI. A, Survival free of relapse. B, Survival free of rehospitalization. C, Survival free of surgery. D, Survival free of

AC C

EP

TE D

M AN U

SC

emergency department visit.

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT