Risk factors for Clostridioides difficile infection in children and adolescents with inflammatory bowel disease: a systematic review and meta-analysis

Sheng-Bo Fang · Yan-Qing Song · Chun-Yan Zhang · Li-Bo Wang

Abstract

Keywords Adolescent · Children · Clostridioides difficile infection · Inflammatory bowel disease

Introduction

Clostridioides difficile

is an important pathogen causing diarrhea acquired in healthcare settings [ 1]. Over the past few decades, the incidence of

Clostridioides difficile

infection (CDI) is rising in pediatric patients [ 2], especially in those with inflammatory bowel disease (IBD) [ 3]. It is reported that the prevalence of CDI in pediatric patients diagnosed with IBD ranges from 3.5 to 69% [ 3- 12].The impact of CDI on IBD has gained increasing attention. For one thing, IBD symptoms, which include watery or bloody diarrhea, abdominal pain and fever, can overlap with CDI [ 13]. Therefore, pediatricians are often faced with the challenge of distinguishing IBD exacerbation and CDI. A patient with disease flaring of his symptoms and concomitant infection with

C. difficile

may pose a difficult therapeutic choice to his physicians: whether withdraw immunosuppressant or administer antibiotics together with immunosuppressant therapy may be a tough choice in management. For another, IBD has risk of CDI exposure, immune suppression and altered gut flora. As IBD is a chronic disease with recurrent flares, patients with IBD have increased exposure to medical institutions, and the use of immunosuppressive drugs contributes an additional risk of CDI. Besides, it is still unclear whether the underlying inflammatory process renders the patients more prone to infections [ 14]. The gut microflora of IBD patients could be altered, resulting in less diversity hence reduced resistance to colonization to

C. difficile

and increased rate of CDI [ 15].

To date, most studies have demonstrated the risk factors of CDI with IBD in adult patients. In a meta-analysis study,antibiotic use within 30 days, colonic involvement in Crohn’s disease (CD) patients, and use of biologics were associated with increased CDI rates [ 16]. In the pediatric population,conflicting evidence existed when examining potential risk factors for CDI. Some studies reported that use of antibiotics was associated with higher CDI rates [ 17, 18], while others showed that there was no effect [ 4, 8, 19]. With respect to IBD therapy, there were studies reporting that use of steroid and anti-tumor necrosis factor biologics were risk factors for CDI[ 20], and use of immunosuppressants may increase the risk of CDI among IBD patients [ 17]. However, opposite evidence showed that there was no relationship between IBD therapy and CDI [ 4, 19]. Whether prior hospitalization is a risk factor is controversial, studies demonstrated that hospitalization within 30 days might be positively associated with CDI in IBD patients [ 4, 8] while others reported negatively [ 19].

The relationship between bowel surgery rate and CDI is also controversial. Studies by Hellmann et al. [ 20] and Kelsen et al. [ 8] found that compared to non-CDI IBD controls, there was an increased bowel surgical rate in CDI IBD patients, while another study demonstrated an inverse effect,with lower rate in CDI IBD patients [ 9].

The aim of this study was to summarize the current evidence and to perform a meta-analysis on the potential risk factors associated with CDI in pediatric IBD patients, and bowel surgery in IBD with CDI.

Methods

The meta-analysis and systematic review was registered in PROSPERO (registration number: CRD42021240384) and was performed in accordance with the guidelines established by the Preferred Reporting Items for Systematic Reviews and Meta-Analyses statement [ 21].

Literature search

We systematically searched PubMed, EMBASE and Cochrane Library databases from inception to February 24, 2021 using the following search terms: “inflammatory bowel disease”, “

Clostridioides difficile

” or the former name“

Clostridium difficile

”, “child” and “adolescent”. The full search strategy is detailed in Supplementary Table 1. There was no restriction on language or publication type.

Inclusion criteria

Included studies met the following criteria: (1) studies conducted in children and adolescents with IBD [CD,ulcerative colitis (UC), unspecified] (age range: 0-24 years old); (2) studies on risk factors associated with CDI in IBD patients compared to non-CDI IBD controls; (3) studies on bowel surgery rates in IBD patients with CDI compared to CDI-negative IBD controls; and (4) studies published in English or Chinese.

Exclusion criteria

Studies were excluded according to the following criteria:(1) if they lacked CDI-negative IBD controls; (2) studies conducted on adults subjects; (3) studies that were abstract only without full text; (4) studies that were case report,letters or editorials; (5) studies that lacked sufficient raw data for synthesis; (6) studies that were duplicates; and (7)studies that were ongoing or not yet finished.

Risk factors and outcomes

The primary outcome was to identify the risk factors and outcomes associated with CDI in pediatric IBD patients compared to non-CDI IBD controls. We investigated patient demographic data and disease-related risk factors to determine risk factors that could be included in the meta-analysis. Risk of bowel surgery in IBD patients with CDI compared to non-CDI IBD controls was also assessed.

Subgroup analyses were performed for risk factors that involved ≥ 5 studies, to identify potential sources of heterogeneity. Subgroups include: country where study was performed within USA vs. outside; high quality score[Newcastle-Ottawa scale (NOS) ＞ 6].

Data extraction

All potentially related articles were retained by two authors (FSB, SYQ) independently, and were strictly reviewed according to inclusion/exclusion criteria according to preset outcomes. Any disagreements were resolved by consensus or by consultation with a senior author(WLB). From the included studies, the following items were extracted: study characteristics (author, year of publication, country, study period, study design), patients characteristics (age, gender, type of IBD), risk factors[IBD type, disease activity, proton-pump inhibitor use,antibiotic use, IBD medications use (aminosalicylates,steroids, immunomodulators and biologic therapy)], and bowel surgery rates.

Methodological assessment

For quality assessment, NOS was used for case-control studies and cohort studies [ 22], including selection (four questions), comparability of study groups (two questions), and ascertainment of the outcome of interest (three questions).Based on scores, study quality was graded as low (0-3),moderate (4-6) and high (7-10), respectively. Two reviewers (FSB and SYQ) assessed study quality independently.The grade of evidence is shown in Supplementary Tables 2 and 3.

Statistical analysis

This meta-analysis was performed using STATA statistical software (version 16.0) with the “meta” package. The Hedges random effects model was used to calculate the weighted summary estimates, which were presented as log-ORs (odds ratio) with 95% confidence intervals (CI).

Study heterogeneity was evaluated using

I

2 value. An

I

2 value below 30% is considered as low heterogeneity;between 30 and 50% considered is considered as moderate heterogeneity; between 50 and 75% is considered as substantial heterogeneity and of 75% and above is considered as considerable heterogeneity [ 23]. Publication bias was determined by Egger’s regression test, with

P

＜ 0.05 indicating a high likelihood of bias [ 24].

Results

Search results

Details of the search strategy are summarized in Fig. 1.A total of 657 citations were identified through PubMed,EMBASE and Cochrane library. Fourteen studies that met inclusion criteria were included [ 4, 7- 10, 12, 17- 20, 25- 28].Of these, seven studies were case-control studies and seven were cohort studies. Four studies were performed prospectively, six studies were performed retrospectively and the other four studies were not explicit. Eight studies were published in North America, five in Europe and one in Asia.

Patient characteristics

Data were extracted from a total of 871 IBD patients with CDI and 16,243 IBD controls without CDI. The majority of studies investigated hospitalized patients (

n

= 10), three studies included patients from outpatient clinics as well and one study included only outpatients. Thirteen studies included both CD and UC patients, while one study included only CD patients. The baseline characteristics of these patients are summarized in Table 1.

Fig. 1 Flow chart of systematic review with meta-analysis

Clostridioides difficile infection diagnosis

Five studies diagnosed CDI by stool enzyme-linked immunosorbent assay or enzyme immunoassay (EIA), four studies diagnosed CDI by EIA or polymerase chain reaction (PCR),two studies diagnosed CDI by PCR only, two studies used International Classification of Diseases (ICD)9/ICD10 codes to identify CDI positive patients. The remaining one study used cell culture cytotoxicity assay.

Risk factors for Clostridioides difficile infection in pediatric inflammatory bowel disease patients

Risk factors that were included in this meta-analysis were type of IBD, disease activity, colonic involvement, prior hospitalization, and use of proton-pump inhibitor (PPI),antibiotics, 5-aminosalicylic acid (5-ASA), systemic steroids, immunomodulators and biologics. Disease activity was identified by pediatric Crohn’s disease activity index for CD, and pediatric ulcerative colitis activity index or Lichtiger colitis activity index or Truelove-Witts score for UC. Use of antibiotic and PPI was within the previous 1,2 or 4 months before CDI in different studies. The timeframe of the use of other medications was not explicit, but is being concurrently administered at the time of diagnosis of CDI. In the overall analysis, there was no association between CDI and IBD type (OR = 0.92, 95% CI 0.70-1.21)(

n

= 11,

I

2 = 0.00%) (Supplementary Fig. 1), disease activity(OR = 1.99, 95% CI 0.88-4.57) (

n

= 5,

I

2 = 51.83%) (Supplementary Fig. 2), or colonic involvement (OR = 1.72, 95% CI 0.55-5.31) (

n

= 8,

I

2 = 67.93%) (Supplementary Fig. 3). We also did not find any association between CDI and prior hospitalization in IBD patients (OR = 0.63, 95% CI 0.33-1.17)(

n

= 2,

I

2 = 0.00%) (Supplementary Fig. 4). Besides, use of PPI or antibiotics was unrelated to CDI with OR 1.13 (95%CI 0.59-2.14) (

n

= 4,

I

2 = 25.03%) (Supplementary Fig. 5)and OR 2.27 (95% CI 0.73-7.03) (

n

= 6,

I

2 = 81.25%) (Supplementary Fig. 6), respectively. Among medications used for IBD patients, patients who used 5-ASA were found to be more likely to develop CDI than those who did not (

n

= 3)(OR = 1.95, 95% CI 1.26-3.03) (Fig. 2 a), and this association was noted with low heterogeneity (

I

2 = 0.00%). There was no association between use of steroids (OR = 1.25, 95%CI 0.76-2.03) (

n

= 3,

I

2 = 0.00%) (Fig. 2 b), immunomodulators (OR = 1.13, 95% CI 0.74-1.73) (

n

= 4,

I

2 = 7.82%)(Fig. 2 c), or biologics (OR = 1.17, 95% CI 0.68-2.03) (

n

= 3,

I

2 = 0.00%) (Fig. 2 d) and CDI.

Outcomes of Clostridioides difficile infection

Only four studies reported bowel surgery rates. Of these,one study reported the rate of colectomy in UC; one study reported the rate of bowel resection surgery in CD was associated with CDI, with a mean follow-up time of 527(115-2254) days for CD patients tested positive for CDI compared with 1268 (158-2254) days for those who were tested negative. Two other studies did not specify type of IBD studied, and one was followed up for 6 months since the diagnosis of CDI. Overall, the bowel surgery rate in IBD with CDI was 3.8-57.1%, while those without CDI were 0-21.3%. The details are shown in Table 2.

Subgroup analysis

When studies were stratified by quality score, active disease was associated with CDI in studies in the high quality score group (NOS ＞ 6) (OR = 4.66, 95% CI 2.16-10.07) or those conducted outside USA (OR = 2.94. 95% CI 1.57-5.58). The conclusions of association between colonic involvement, or antibiotic use and CDI were not affected by location or quality of studies (Table 3).

Study quality and publication bias

A modified NOS was used for quality assessment (Supplementary Tables 2 and 3). The median quality score was 6.5(range 5-8). There was no publication bias on IBD type(Egger’s

P

= 0.41), colonic involvement (Egger’s

P

= 0.19),disease activity (Egger’s

P

= 0.79), use of antibiotic (Egger’s

P

= 0.67), 5-ASA (Egger’s

P

= 0.25), steroids (Egger’s

P

= 0.44), immunomodulators (Egger’s

P

= 0.19), or biologics (Egger’s

P

= 0.36). Publication bias could be detected on use of proton-pump inhibitor (Egger’s

P

= 0.01).

Discussion

To our knowledge, this study is the first meta-analysis and systematic review to explore the risk factors of CDI in pediatric IBD patients, and assess the impact of CDI on IBD regarding rate of bowel surgery, which is different from that of adults [ 16].

Initially, we identified that use of 5-ASA was associated with an increased incidence of CDI in pediatric IBD patients (OR = 1.96, 95% CI 1.26-3.03). The result could be attributed to several reasons and should be interpreted with caution. First, as 5-ASA has been recommended as the first-line therapy for the induction and maintenance of mild-to-moderate UC [ 29] and could be used for induction of remission with mild colonic inflammation in CD[ 30], hence 5-ASA was quite widely used in pediatricpatients with IBD. Second, it is reported that 5-ASA has immunosuppressive activity [ 31, 32], and when combined with other immunosuppressants, the risk of opportunistic infection would be significantly increased (OR = 4.37,95% CI 1.45-13.19) [ 33]. In another study, use of 5-ASA was found be more prevalent among CDI IBD cohort with two and more CDI episodes [ 34]. It is supposed that use of 5-ASA could alter colonic bacterial profile in patients with IBD, hence increasing the susceptibility to CDI [ 35,36]. Of the three studies included in the analysis, in one study 5-ASA was the only medication used in its cohort of patients, while in the other two studies, whether the patients with IBD and CDI were treated with 5-ASA therapy alone or in combination with other agents was not explicit. Third, among the studies included in the analysis,one study demonstrated that there was an increased risk of CDI in patients on 5-ASA therapy (OR = 2.29, 95% CI 1.36-3.90). This study was assigned 70% of the weight when calculating the effect size. The other two studies did not find any association between 5-ASA use and CDI.Lastly, there may be some confounding factors affecting the relationship, such as time period of the study, duration of 5-ASA use, and economic considerations as 5-ASA is one of the least expensive drugs available for treating IBD. Therefore, this issue still needed further researches to verify.

Characteristics of included studies

inflammatory bowel disease, type of inflammatory bowel disease, infection, enzyme immunoassay, polymerase chain reaction assay, International Classification of Diseases, colonic involvement, prior hospitalization, antibiotics use, proton-pump inhibitor, disease activity, not reported, - 5-aminosalicylic acid. a Include: Italy, Span, Israel, Denmark, Croatia

Author, year Country Study period Study design IBD, n CDI + IBD, n Non-CDI + IBD,n Method of CDI diagnosis Risk factors Outcomes Banaszkiewicz et al.[ 10]Chandrakumar et al.[ 12]Poland 2007-2010 Cohort 134 63 71 EIA CI, PH,ABU, PPI Canada 2011-2019 Cohort 216 20 241 EIA or PCR TIBD, CI Colectomy Hellmann et al. [ 20]USA 2006-2016 Cohort 75 14 61 PCR or toxinbased assay TIBD, ABU,PPI, CI,biologics,steroids Bowel surgery Hojsak et al.[ 17]Hourigan et al. [ 25]Croatia 2006-2011 Cohort 216 13 203 EIA TIBD,5-ASA,immunomodulators,USA 1993-2012 Case-control 2888 82 2806 ICD TIBD Kelsen et al.[ 8]USA 1997-2007 Case-control 238 111 127 EIA TIBD, ABU,PPI, CI,biologics,5-ASA,immunomodulators,steroids Bowel surgery Martinelli et al. [ 19]Multicountry a 2010-2011 Case-control 211 20 191 Enzyme immunocard DA, TIBD,PH, ABU,PPI, CI,immunomodulators Mir et al.[ 18]USA 2010-2012 Cohort 123 10 113 EIA or PCR ABU Pant et al. [ 9] USA 2000-2009 Cohort 12,610 447 12,163 ICD NR Bowel surgery Pascarella et al. [ 4]Wultańska et al. [ 7]Italy 2005-2007 Case-control 81 20 61 EIA or PCR DA, TIBD,CI, 5-ASA,immunomodulators,steroids Poland 2005-2007 Cohort 58 40 18 EIA or PCR DA, TIBD Li et al. [ 26] China 2015-2016 Case-control 30 6 24 PCR DA, TBID,CI Hyams et al.[ 27]USA 1982-1983 Case-control 44 3 41 Cell culture cytotoxicity assay TIBD, ABU,CI Lamousé-Smith et al.[ 28]USA 2006-2012 Case-control 145 22 123 PCR DA, TIBD

Fig. 2 Inflammatory bowel disease (IBD) medications and Clostridioides difficile infection (CDI) in IBD patients. a 5-aminosalicylic acid use; b systemic steroids use; c immunomodulators use; d biologics use. OR odds ratio, CI confidence interval

Table 2 Summary of bowel surgery rate

inflammatory bowel disease, infection, not reported, ulcerative colitis, Crohn’s disease. a Time for CDI + IBD group; b Time for non-CDI + IBD group

Study IBD type CDI + IBD, n (%) Non-CDI + IBD, n (%) Time from diagnosis to surgery (d)Chandrakumar et al. [ 12] UC 2 (10.0) 0 (0.0) NR Hellmann et al. [ 20] CD 8 (57.1) 13 (27.1) 527 (115 to 2254) a 1268 (158 to 2254) b Kelsen et al. [ 8] IBD 9 (8.1) 4 (3.3) Within 6 mon of CDI Pant et al. [ 9] IBD 17 (3.8) 1289 (11.9) NR

Table 3 Subgroup of risk factors that associated with CDI in IBD patients

inflammatory bowel disease, infection, Newcastle-Ottawa scale, odds ratio, confidence interval

Subgroups of risk factors Values LogOR (95% CI) I 2 (%)Disease activity High quality study (NOS ＞ 6) 3 1.54 (0.77, 2.31) 0.00 Non-USA patients only 4 1.08 (0.45, 1.72) 51.83 Colonic involvement High quality study (NOS ＞ 6) 7 0.74 (− 0.31, 1.79) 64.19 Non-USA patients only 5 0.47 (− 0.75, 1.68) 59.23 Antibiotics use High quality study (NOS ＞ 6) 4 0.69 (− 0.29, 1.95) 73.92 Non-USA patients only 2 0.82 (− 0.32, 1.95) 0.00

In the overall pooled analysis, we did not find the association between CDI and disease activity, but in subgroup analysis, active IBD seemed to have an increased risk of CDI. It is reported in more severe pediatric IBD there exist dysbiosis of intestinal bacterial flora, which may have an impact on reducing resistance to colonization and hence increased susceptibility to CDI [ 37]. Even so, it is difficult to ascertain the relationship between disease activity in IBD and CDI; as CDI could in turn increase disease activity of IBD [ 38]. These findings suggested CDI and disease activity in IBD may be positively correlated. Another reason for this result is due to the symptoms overlap, children with disease flare are disproportionately tested for

C. difficile

, and if

C.difficile

positive then are diagnosed with CDI.Traditionally, exposure to antibiotics, immunosuppression, acid reduction therapy, and hospitalization were considered as major risk factors for CDI in IBD patients [ 39].Antibiotics may destroy the normal gastrointestinal flora,resulting in reduced colonization resistance sufficiently to allow toxin-producing strains of

C. difficile

to become pathogenic. But there were evidence suggesting that use of antibiotics appeared to be a less important factor for CDI in IBD patients, being less frequent in IBD (42%) than non-IBD patients (69%) [ 40]. In our study, we did not find any association between antibiotics use and CDI. It indicated that although the gut flora of IBD had changed, yet the effect might not have affected the risk significantly. Theoretically, immunosuppressants, including corticosteroids and biologics, may be positively correlated with CDI. IBD is a chronic disease, and prolonged use of immunosuppressants may be required on a long-term basis, and high doses would be needed for the more active, severe cases. However,in a meta-analysis on adult patients, apart from biologics,use of corticosteroids and traditional immunosuppressants were not associated with CDI [ 16]. Use of biologics significantly increased the risk of CDI in IBD (OR = 1.65, 95% CI 1.18-2.30) [ 16]. Results showed that different immunosuppression regimens might carry different risk for CDI. For children, the results were in keeping with adult population except biologics. Biologics use was not identified to have increased the risk for CDI in our study though the sample size included in the analysis was small (two studies). As more and newer biologics are administered to children, further investigations may be needed to study their effect on the risk of CDI. In addition, mounting evidences suggested that PPI use increases the risk of CDI in pediatric populations[ 41, 42]. Use of PPI might result in intestinal dysbiosis with significantly increases in

Streptococcaceae

and

Enterococcaceae

, both risk factors for CDI, and decrease in

Faecalibacterium

, a commensal anti-inflammatory microbe [ 43]. In contrast to the evidence above, our pooled analysis showed no association between PPI use and development of CDI in IBD patients. It has been postulated that IBD patients may have changed their gut flora hence acid suppression would not change the sensitivity to CDI [ 16]. Furthermore,in a study performed on pediatric IBD patients, the authors suggest that IBD patients are more likely to develop community-acquired CDI while in comparison it is more likely for the general population to develop nosocomial infections[ 8]. We did not detect a significantly increased risk of prior hospitalization and development of CDI.

In adult studies, compared to CD, UC is more likely to acquire CDI [ 44, 45], and colonic involvement in CD also significantly increased the risk of CDI [ 16]. However, in children, no difference in risk was noted between IBD types in our study, and overall colonic involvement was not associated with CDI in pediatric IBD. One possible explanation is that pediatric patients have different patterns of IBD involvement, and have more colonic CD compared with adults [ 46].

Lastly, the risk of bowel surgery associated with CDI was also assessed. Overall, the bowel surgery rate varied considerably, ranging from 3.8 to 57.1% in IBD with CDI, and 0-21.3% in IBD without CDI. It is notable that one study conducted by Pant et al. reported a significantly lower rate of bowel surgery in IBD with CDI compared with IBD without CDI (3.8 vs. 11.85%) [ 9]. Sandberg et al. also found children and young adults hospitalized with CDI had decreased odds of the need to have colectomy (OR = 0.31, 95% CI 0.23-0.44)compared with those without CDI [ 47]. Similar results were identified in adults that the CDI IBD group was only half as likely to undergo bowel surgery as patients admitted to hospital for IBD alone (OR = 0.6) [ 48]. The reason for such an association is only conjectural. It is possible that the CDI IBD group of patients is too ill to undergo surgery for fear of serious complications such as infection, and that the patients might be hoping for good results from antibiotic administration hence refused surgery [ 9]. Other speculations are limitation of using management data sets, i.e., studies only calculated bowel surgery rates for the stipulated time frame and not subsequent hospitalizations. Moreover, when patients with IBD were ready for surgical treatment, and found be CDI positive, the planned surgery might be delayed pending CDI recovery hence the reduced number of surgeries [ 49].Therefore, this question needs further investigation with more well-designed clinical trials.

There are several limitations. Initially, there is a significant heterogeneity in the method of defining CDI. Most studies used enzyme immunoassay or PCR, but diagnostic codes and cell culture cytotoxicity assay were also used. Moreover, all the studies included were observational studies, some reported the data prospectively while the others used a retrospective approach, which may result in significant differences in clinical decision and confuse the sequence of exposures and outcomes.In addition, study populations were recruited from either hospitalized patients or outpatients or both, which might result in patient selection bias. Another limitation is substantial heterogeneity, with a lot of difference in characteristics of patients,CDI management, antibiotic use and IBD therapy.

In conclusion, 5-ASA use and active disease might be risk factors associated with CDI in pediatric IBD patients. Bowel surgery rates associated with CDI in IBD patients varied greatly. However, the sample size evaluated was small. In the future, it is still necessary to conduct larger-scale clinical trials to verify these findings, to determine the pathogenic risk factors and potentially protective factors, and hence to help develop evidence-based strategies of CDI management in IBD patients.

Supplementary Information

The online version contains supplementary material available at https:// doi. org/ 10. 1007/ s12519- 021- 00486-1.

Acknowledgements

Thanks for Yang-Yu Zhang who assisted with statistic assistance and the guide of Professor Ying kit Leung on the research direction.

Author contributions

FSB performed screening of papers, data extraction, study quality assessment, collected and analyzed the data, wrote and revised the manuscript, and performed the literature search. SYQ performed screening of papers, data extraction, study quality assessment, collected and analyzed the data. ZCY wrote and revised the manuscript. WLB developed the study concept and protocol, leading role of coordinating the project, senior authorship to resolve disagreements on inclusion, exclusion and quality assessment of papers and critiquing and editing the manuscript. All authors approved the final manuscript as submitted and agree to be accountable for all aspects of the work.

Funding

There is no fund in this article.

Data availability

All data used during the study are available from the corresponding author by request.

Declarations

Ethical approval

Not required for this meta-analysis.

Conflict of interest

No financial or non-financial benefits have been received or will be received from any party related directly or indirectly to the subject of this article. The authors have no conflict of interest to declare.