2023 Impact Factor
2023 Impact Factor
This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/4.0) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.The autoimmune disease systemic sclerosis (SSc) affects multiple organ systems, including the pulmonary, cardiac, renal and gastrointestinal (GI) systems and is characterized by abnormalities of the microvasculature with increased deposition of matrix proteins and collagen into the connective tissues, with subsequent fibrosis of skin and internal organs.1 SSc can be categorized into 2 major groups reflecting the extent of skin involvement; those with proximal involvement are addressed as diffuse cutaneous systemic sclerosis (dcSSc), whereas those with skin involvement mainly affecting the distal limbs (elbows or knees), with or without involvement of face and neck, are addressed as limited cutaneous systemic sclerosis (lcSSc).2
Most patients with SSc (~90%) develop GI manifestations and may present with a variety of symptoms ranging from dry mouth, dysphagia to fecal incontinence.3 Most relevant small intestinal manifestations in SSc include small intestinal bacterial overgrowth (SIBO), pseudo-obstruction, malabsorption, and pneumatosis cystoides intestinalis.4 SIBO is a disorder, associated with a variety of symptoms, when ≥ 105 colony forming units per millilitre (CFU/mL) of bacteria (typically found in the colon) are found in culture of jejunal aspirates. However, concentrations ≥ 105 CFU/mL are mostly seen initial studies investigating SIBO in patients with post-surgical anatomy, eg, stagnant loop syndrome. Healthy adults may have counts between 0 and 103 CFU/mL5,6 and more recently a bacterial concentration of ≥ 103 CFU/mL is considered as the cut off criteria for diagnosing SIBO.7 Traditionally, culture-based approach is the gold standard for the diagnosis of SIBO.8 However, besides being invasive, culture-based techniques for diagnosing SIBO have several limitations6 and are infrequently used in the standard clinical practice. As a consequence, breath tests (BT) have been developed. However, when culture-based methods are used as a standard, which as outlined above are clearly suboptimal tests for diagnosing SIBO, BT show poor sensitivity and limited specificity and have several methodological problems for SIBO diagnosis.9 All this has questioned their suitability as diagnostic tests in the clinical setting.10,11 Thus, one of the fundamental problems in diagnosing SIBO is the lack of validated and universally accepted diagnostic tests.
In SSc, intestinal hypomotility caused by the vasculopathy, smooth muscles atrophy, and subsequent fibrosis leading to small bowel stasis causes bacterial colonization and ultimately leading to SIBO. Thus, although SIBO can be considered a complication of SSc, SIBO potentially aggravates the clinical manifestation of SSc. The rates of SIBO prevalence in SSc and SSc-subtypes as well as the role of potential risk modifiers for SIBO such as treatment with proton pump inhibitors (PPI) therapy in the setting of SSc are uncertain.12 Moreover, the link between SIBO and GI symptoms in SSc is incompletely characterized and the efficacy of antibiotic therapy on elimination of SIBO and symptoms in SSc is poorly defined.
Against this background we performed a systematic review and meta-analysis (SRMA). As a primary endpoint of this systematic review and meta-analysis we aim to (1) determine the prevalence rates of SIBO in patients with SSc (and SSc-subtypes) and controls. The secondary endpoints are to (2) assess the role of diagnostic approach on differences in SIBO prevalence rates comparing various geographic regions, (3) explore the underlying risk modifiers for the occurrence of SIBO in SSc, and (4) analyze response to therapy (antimicrobial and octreotide) in relation to symptoms in SIBO positive patients with SSc.
This SRMA was conducted in accordance with the “Preferred Reporting Items for Systematic Reviews and Meta-Analysis” statement (PRISMA).13,14 Compliant with existing standards the study-protocol for this SRMA has been registered (PROSPERO, CRD42021274206).
Available electronic databases, (MEDLINE [Ovid], EMBASE, PUBMED, SCOPUS, Web of Science, and the COCHRANE Library), were searched from initiation (1966) until Jan 2022 for all studies evaluating the link between SIBO and SSc (and/or lcSSc/dcSSc). For the detailed search strategy please see the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) flow diagram (Fig. 1). With the expert support from our librarian the literature search was conducted. The search strategy utilized for the MEDLINE and EMBASE data bases is depicted (Supplementary Fig. 1A and 1B). For the initial search no restrictions in relation to languages were used. Subsequently, an advanced search step was used to identify “grey” literature with search engines such as Google or Google Scholar, and as a further step “Snowball” method was applied to find relevant articles.
Figure 1. Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) flow diagram.
Independently titles and abstracts were screened and publications with relevant information in relation to the research question (link between SIBO and SSc and/or lcSSc/dcSSc) were further assessed. Full texts of the relevant articles were subsequently assessed. Case-control studies, prevalence studies recruiting random (unselected) study subjects and controls meeting standardized diagnostic criteria for SSc and SIBO were included. For diagnosis of SSc the diagnostic standards of the updated American College of Rheumatology and the European League Against Rheumatism 2013 criteria or the former ACR 1980 diagnostic criteria15-18 were used. Studies without original data or including diverse populations of autoimmune disease or mixed connective tissue disorders without detailed extractable information on SSc, or without validated methods for the diagnosis of SIBO in SSc were excluded, as outlined in Supplementary Table 1. Information in relation to antimicrobial and/or PPI therapy were taken from the identified studies. Conference abstracts were also included if the required data were reported. The control group comprised of healthy subjects without symptoms or patients undergoing investigations for chronic or relapsing unexplained GI symptoms (patient controls). Table 1 summarizes eligibility criteria for study inclusion and studies that were not included are shown in Supplementary Table 1. Any disagreements between reviewers were jointly resolved by review of the original publication.
Table 1 . Eligibility Criteria for the Studies Included in Systematic Review and Meta-analysis
| Eligibility criteria |
| • Prevalence or case-control studies, published as full papers in peer reviewed journals or conference abstracts. |
| • Adults and children with a presumed diagnosis of Systemic sclerosis based on meeting specific diagnostic criteriaa. |
| • Non SSc control group, referred to as ‘controls’ included “healthy asymptomatic controls” as well as “patient controls” including subjects undergoing evaluation for unexplained gastrointestinal “syndromes” (eg, dyspepsia, abdominal pain, and diarrhea). |
| • Studies reporting on efficacy data after antibiotic treatment of SIBO in SSc-patients were also included. |
| • Clinically validated methods to diagnose SIBOb. |
| • Participants not specially selected. |
aAmerican College of Rheumatology/European League Against Rheumatism criteria15-18 or formerly known as the American Rheumatism Association Diagnostic and Therapeutic Criteria Committee.
bLactulose breath test, glucose breath test, or jejunal aspirate and culture (or any combination of these).
SSc, systemic sclerosis; SIBO, small intestinal bacterial overgrowth.
Two authors (V.P. and K.V) extracted independently the data and entered the data into a spreadsheet (Microsoft Excel, 2010 Professional edition; Microsoft Corp, Redmond, Washington, USA). For the purpose of this study, the following information was retrieved: year of publication, name of authors, design of the study, geographic region (country), type of controls, diagnostic methods for SIBO including details of tests (substrate used, cutoff values for SIBO diagnosis), age, gender, PPI use, fecal calprotectin (FC), and co-morbidities including prior surgery. Furthermore, the diagnostic criteria for SSc and subtypes, the treatment for SIBO in SSc-patients (antimicrobial compounds or other treatment modalities including octreotide) and response to treatment were recorded. We also captured the prevalence of subjects (both SSc-patients and controls) who had methane positivity on breath test or intestinal methanogen overgrowth.
For the quality assessment of prevalence studies, the Joanna Briggs Institute (JBI) critical appraisal tools19 was utilized, while the Newcastle-Ottawa scale (NOS)20 was used to assess the quality of case-control studies, further details in Supplementary Materials and Methods.
The numbers of patients with SSc and the respective controls in the various cohorts were assessed as the first step. Subsequently, the pooled prevalence of SIBO in SSc (with or without a control group) was calculated. In separate steps, the pooled odds ratio (OR) and the respective 95% confidence intervals (CI) for the prevalence of SIBO in SSc-patients and controls were calculated. Standardized mean difference and 95% CI were used to estimate the difference between GI symptom scores in SSc-patients compared to controls. Other data (eg, total GI symptom score) were also recorded as mean and standard deviation, further details in Supplementary Materials and Methods. Subgroup analyses with stratification of data by diagnostic modalities, geographic region, SSc-subtype (lcSSc/dcSSc), the PPI-use, proportion of subjects with methane-positive SIBO in patients with SSc were performed. Finally, we calculated the proportion of SSc-patients with SIBO, who responded to antibiotic therapy with regards to normalisation of positive breath test and improvement in GI symptoms.
Furthermore, descriptive analysis was used to assess the link between SIBO and SSc utilizing the Comprehensive Meta-Analysis software (version 3.3.070; Biostat Inc, Englewood, NJ, USA). In the results section we also provide the observed (or unweighted) number of positive cases and total tested per study in addition to the weighted pooled estimates of the OR comparing treated with control groups. Subsequently we calculated ORs and pooled prevalence estimates of disease utilizing a random effects model.21 This is intended to account for between-study variability and are reported such that an OR > 1 favors responses in treated patients. The Supplementary Materials and Methods provide additional details of statistical analysis.
Our search of the available literature (detailed in Fig. 1) identified 28 studies suitable for this systematic review and the subsequent meta-analysis. Eleven out of 28 studies22-32 were case-control studies while 1733-49 were prevalence studies. The specifics of these studies in relation to methodology for the diagnosis of SIBO and the characteristics of the patient cohorts are provided in Table 2 and Supplementary Tables 2 and 3.
Table 2 . Characteristics of the Studies Showing Mode of Diagnosis and Prevalence of Small Intestinal Bacterial Overgrowth in Systemic Sclerosis
| No | Author | Study year | Region | Type of study | Patients with SSc (n) | Patients with SSc sub-type (n) | Criteria for SSc | Controls (n) | Type of control | Mode of diagnosis of SIBO | SIBO in SSc (n [%]) | SIBO in SSc sub-types (n) | SIBO in controls (n) | ||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| dcSSc | lcSSc | dcSSc | lcSSc | ||||||||||||
| 1 | Adarsh et al48 | 2017 | India | Prevalence | 37 | 16 | 34 | ACR | NA | NA | GBT | 7 (18.9) | NA | NA | NA |
| 2 | Bae et al47 | 2013 | USA/France | Prevalence | 55 | 24 | 27 | ACR | NA | NA | LBT | 21 (38.2) | NA | NA | NA |
| 3 | Brown et al46a | 2008 | UK | Prevalence | 8 | 3 | 5 | Not stated | NA | NA | GBT | 8 (100.0) | 3 | 5 | NA |
| 4 | Chaudhary et al45a,b | 2010 | India | Prevalence | 35 | 9 | 26 | Not stated | NA | NA | BT | 12 (34.3) | NA | NA | NA |
| 5 | Cobden et al29 | 1980 | UK | Case control | 20 | NA | NA | Not stated | 18 | Not healthy | JAC | 4 (20.0) | NA | NA | 0 |
| 6 | Cruz-Dominguez et al44a,b | 2017 | Mexica | Prevalence | 68 | 27 | 41 | ACR | NA | NA | GBT | 44 (64.7) | NA | NA | NA |
| 7 | Di Ciaula et al28a | 2008 | Italy | Case control | 38 | 4 | 34 | ACR | 60 | NA | LBT | 0 (0.0) | 0 | 0 | 0 |
| 8 | Fynne et al43a | 2011 | Denmark | Prevalencec | 15 | NA | NA | ACR | 17 | Healthy | GBT | 3 (20.0) | NA | NA | NA |
| 9 | Garcia-Collinot et al42a | 2019 | Mexico | Prevalence | 74 | 32 | 43 | ACR/EULAR | NA | NA | LBT | 40 (54.1) | 18 | 22 | NA |
| 10 | Gemignani et al22 | 2013 | Italy | Case control | 50 | 18 | 32 | Not stated | 60 | Healthy | GBT | 9 (18.0) | NA | NA | 3 |
| 11 | Gough et al49a | 1995 | UK | Prevalence | 10 | 8 | 2 | ARA | NA | NA | LBT | 0 (0.0) | 0 | 0 | NA |
| 12 | Kaye et al23 1994 | 1994 | UK | Case control | 10 | NA | NA | ARA | 10 | Healthy | JAC | 4 (40.0) | NA | NA | 0 |
| 13 | Kaye et al24 1995 | 1995 | UK | Case control | 24 | 6 | 17 | ACR | 9 | Healthy | JAC | 8 (33.3) | 2 | 6 | 0 |
| 14 | Levin et al41 | 2021 | Canada | Prevalencec | 29 | 7 | 20 | Not stated | 20 | Healthy | GBT | 13 (44.8) | 2 | 10 | NA |
| 15 | Madrid et al40 2012a,b | 2012 | Chile | Prevalence | 30 | 9 | 21 | ARA | NA | NA | LBT | 16 (53.3) | 8 | 8 | NA |
| 16 | Madrid et al39 2020a,b | 2020 | Chile | Prevalence | 53 | NA | NA | Not stated | 37 | NA | LBT | 32 (60.3) | NA | NA | NA |
| 17 | Marie et al37 2009 | 2009 | France | Prevalence | 51 | 25 | 26 | ACR | NA | NA | GBT | 22 (43.1) | 8 | 14 | NA |
| 18 | Marie et al38 2015 | 2015 | France | Prevalence | 125 | 43 | 82 | ACR/EULAR | NA | NA | GBT | 44 (35.2) | NA | NA | NA |
| 19 | Owyang32a | 1994 | USA | Case control | 5 | NA | NA | ARA | 6 | Healthy | GBT | 5 (100.0) | NA | NA | NA |
| 20 | Parodi et al31 | 2008 | Italy | Case control | 55 | 18 | 37 | ARA | 60 | Healthy | LBT | 30 (54.6) | NA | NA | 4 |
| 21 | Polkowska-Pruszy´nska et al25 | 2020 | Poland | Case control | 39 | 6 | 33 | ACR/EULAR | 39 | Healthy | LBT | 19 (48.7) | NA | NA | 5 |
| 22 | Savarino et al26 | 2013 | Italy | Case control | 99 | 31 | 68 | ARA | 60 | Healthy | LBT | 47 (47.5) | NA | NA | 3 |
| 23 | Sawadpanich et al36 | 2019 | Thailand | Prevalence | 89 | 65 | 24 | ACR | NA | NA | GBT | 12 (13.5) | 9 | 3 | NA |
| 24 | Shindo et al27 | 1998 | Japan | Case control | 12 | 12 | 0 | ARA | 19 | Healthy | JAC | 7 (58.3) | 7 | 0 | 3 |
| 25 | Soudah et al35a | 1991 | USA | Prevalencec | 5 | 5 | 0 | ARA | 6 | Healthy | GBT | 5 (100.0) | 5 | 0 | NA |
| 26 | Tauber et al34 | 2014 | France | Prevalence | 37 | 14 | 23 | Not stated | NA | NA | GBT | 14 (37.8) | NA | NA | NA |
| 27 | Wegener et al33a | 1994 | Germany | Prevalence | 14 | 6 | 7 | ARA | NA | NA | GBT | 3 (21.4) | NA | NA | NA |
| 28 | Zou et al30 | 2019 | Canada | Case control | 25 | NA | NA | ACR/EULAR | 20 | Healthy | BT | 13 (52.0) | NA | NA | NA |
aIndicates studies which only measured hydrogen (and not methane) during breath testing.
bIndicates studies published as conference abstract.
cSIBO was measured only in cases (systemic sclerosis [SSc]-patients) and not in controls, hence classified as prevalence studies.
dcSSc, diffuse cutaneous systemic sclerosis; lcSSc; limited cutaneous systemic sclerosis; SIBO, small intestinal bacterial overgrowth; ACR, American College of Rheumatology, ACR/EULAR, American College of Rheumatology & European League Against Rheumatism; ARA, American Rheumatism Association; NA, not available; GBT, glucose breath test; LBT, lactulose breath test; BT, breath test; JAC, jejunal aspirate and culture.
Based upon 28 studies with 1112 SSc-patients the pooled prevalence for SIBO in SSc-patients was 39.9% (95% CI, 33.1-47.1) (Fig. 2). The primary analysis revealed considerable heterogeneity (I2 = 76.00%,
Figure 2. Forest plot of studies showing prevalence of small intestinal bacterial overgrowth in systemic sclerosis-patients (39.9% [95% CI, 33.1-47.1]; I2 = 76.00%; Nine out of 11 case-control studies in the final analysis included 347 adult SSc-patients and 335 controls.22-29,31 The remaining 2 studies did not report SIBO prevalence in controls, and were not included in further analysis.30,32 SIBO prevalence in SSc-patients was 7-fold higher at 36.9% (95% CI, 31.9-42.2) compared to 5.4% (95% CI, 3.2-8.3) in controls (Supplementary Table 4). In relation to the prevalence of SIBO, the pooled OR was significantly higher in SSc-patients as compared to controls (9.6; 95% CI, 5.6-16.5;
Figure 3. Forest plot of studies showing prevalence of small intestinal bacterial overgrowth (SIBO) in systemic sclerosis-patients and controls, stratified according to mode of diagnosis of SIBO (OR, 9.6; 95% CI, 5.6-16.5; The quality assessment of the studies is outlined in Supplementary Tables 5 and 6. Utilising the NOS, the majority (6/11, 54.5%) of the case-control studies were categorized as high-quality (NOS score of ≥ 6, Supplementary Table 5). With the JBI critical appraisal tool (Supplementary Table 6), 3 out of the 11 case-control studies had a low risk of bias, 4 had a moderate risk of bias and 4 had a high risk of bias. Furthermore, 6 out of 17 prevalence studies had low, 3 had moderate, and 8 had a high risk of bias.
Including all 16 high-quality studies, there was no significant difference for SIBO prevalence rates in SSc-patients (37.4%; 95% CI, 29.8-45.8; Supplementary Fig. 3). In addition, the analysis revealed considerable heterogeneity (I2 = 77.40%,
Healthy subjects as controls were included in 8 out of the 9, case-control studies Table 2. Subgroup analysis with studies that included healthy controls, the odds for SIBO prevalence in SSc-patients as compared to controls remained unchanged, (OR, 9.6; 95% CI, 5.5-16.6;
Twenty-four studies utilized BT (13 glucose breath test [GBT], 9 lactulose breath test [LBT], and 2 BT without specified substrates) and 4 studies (all case-control studies) utilized jejunal aspirate and culture to diagnose SIBO. When BT were used to diagnose SIBO, the pooled prevalence rate of SIBO in SSc-patients was 44.4% (95% CI, 38.0-51.1; Supplementary Fig. 4), considerable heterogeneity was noted in this analysis (I2 = 78.45%,
Four case-control studies utilizing jejunal aspirate, used a cut off ≥ 105 CFU/mL of bacteria for SIBO diagnosis. In these studies, the prevalence of SIBO in SSc-patients was 34.8% (95% CI, 23.3-46.3) as compared to 5.4% (95% CI, 1.1-14.8) in controls. Furthermore, the odds for SIBO prevalence in SSc-patients was 9.0 (95% CI, 2.7-30.4;
Utilizing jejunal aspirate and culture (all in case-control studies), the OR for SIBO in SSc-patients compared with controls was 9.0 (95% CI, 2.7-30.4;
Nine studies analysed the prevalence rates of SIBO in SSc-subtypes (Table 2). The odds of SIBO prevalence was not different in lcSSc as compared to dcSSc (OR, 1.01; 95% CI, 0.46-2.20;
Five studies analyzed the effects of PPI on the prevalence of SIBO in SSc-patients, (Supplementary Table 7). The prevalence of SIBO in 98/228 (43%; 95% CI, 36.5-49.6) SSc-patients on PPI was significantly higher as compared to 27/94 (28.7%; 95% CI, 19.8-28.9) in SSc-patients not on a PPI. PPI use in SIBO positive, SSc-patients was numerically higher than SIBO positive SSc-patients not on a PPI (OR, 2.3; 95% CI, 0.8-6.4;
Three studies explored the association between SIBO and disease duration in SSc-patients.34,36,37 All 3 found statistically significantly association between disease duration and SIBO prevalence in SSc-patients.34,36,37
Three studies evaluated the potential link between SIBO in SSc-patients and markers of inflammation and markers of malnutrition (Supplementary Table 8).34,36,37 Two studies found no significant association between prevalence of SIBO in SSc-patients and biochemical-markers of inflammation (erythrocyte sedimentation rate, C-reactive protein, leucocyte counts, serum concentrations of total protein, albumin, vitamin B12, serum ferritin, folic acid, or Vitamin D).34,36 In contrast, Marie et al,37 found that the SIBO positive SSc-patients in their study had significantly greater erythrocyte sedimentation rate, reduced serum total protein, albumin, and haemoglobin, and no differences in Vitamin B12, ferritin, and folic acid. Autoantibody screen tests (anti-Centromere antibody and anti-Scl 70 antibody)34,36,37,41 were similar in both SSc-patients with and without SIBO, except for one study,34 where only anti-Scl 70 was significantly associated with SIBO positivity in SSc.
Finally, all 3 studies found no association between other systemic manifestation of scleroderma (interstitial lung disease, digital ulceration, Raynaud’s phenomenon, and pulmonary arterial hypertension)34,36,37,42 and SIBO, except 1 study34 found significantly increased pulmonary arterial hypertension but not interstitial lung disease in SIBO positive SSc-patients. Two studies25,38 found significantly higher FC levels (> 200 μg/g) in SIBO positive SSc-patients as compared to SIBO negative SSc-patients (Supplementary Table 2).
Four studies reported on total GI symptoms, using different symptom assessment scales (Supplementary Table 9). Overall, the total GI symptom score in SSc-patients, with and without SIBO were not different (standardized mean difference, 0.28; 95% CI, –0.05-0.61;
Although numerically higher, we found no significant increase in the prevalence of constipation in SSc-patients with SIBO as compared to those without SIBO (OR, 2.6; 95% CI, 3.3-20.6;
Nine studies with 158 SIBO positive SSc-patients reported response to antibiotic treatment (Supplementary Table 10). There was significant symptom improvement reported by 60.4% (95% CI, 49.9-70.2) of these patients. Fifty-six percent (95% CI, 47.8-64.9) of SSc-patients treated with antibiotics had normalization of BT after treatment with antibiotics. Rifaximin22,25,31 was almost twice as effective as compared to rotating antibiotic therapy34,37,38 (77.8% [95% CI, 64.4-87.9] vs 44.8% [95% CI, 31.7-58.4]) in achieving normalization of the BT after treatment. Only 4 studies reported on antibiotic related adverse events, in 2 out of 4 studies,31,37 none of the patients developed any side effects. In one study,34 1 patient treated with rotating antibiotic developed pseudomembranous colitis, leading to antibiotic discontinuation while in another study42 12/26 (46.1%) patient reported mild GI symptoms.
Two studies evaluated the efficacy of antimicrobial therapy on SIBO in SSc-patients on FC.25,38 One study38 found a statistically significant reduction in FC post SIBO eradication with antibiotic therapy while in the other study25 the reduction in mean FC levels failed statistical significance.
Subgroup analysis stratified according to geographic regions, revealed the highest SIBO prevalence in SSc-patients in the USA (54.9% [95% CI, 49.5-60.2]), followed by the studies conducted in Europe at 36.1% (95% CI, 32.2-40.1) and lowest in studies from Asian countries 22% (95% CI, 16-28.8) (Supplementary Table 11). Moderate heterogeneity among the studies included in these analyses was found (I2 = 46.32%,
Twenty-eight published peer-reviewed studies (11 case-control and 17 prevalence) with 1112 SSc-patients and 335 controls from 13 countries were included in this systematic review and meta-analysis. With more than 1000 SSc-patients this is the thus far largest pooled analysis of case-control and prevalence studies focussing on the association between SIBO and SSc (and SSc sub-types) and potential risk modifiers for SIBO in SSc. Overall, the data reveal a strong link between SIBO and SSc with a 10-fold increased prevalence of SIBO in SSc-patients as compared to controls (OR, 9.6; 95% CI, 5.6-16.5). Conversely, no significant difference in SIBO prevalence rates between different SSc sub-types, namely lcSSc and dsSSc was found. Furthermore, antibiotic treatment targeting SIBO in SSc significantly improved symptoms suggesting that concomitant SIBO in SSc-patients aggravates GI symptoms.
Previous work by Grace et al,50 found diarrhea as the most common symptom in SIBO, which was followed by abdominal pain and bloating as the next most frequent symptoms. Notably, the symptom diarrhea was closely associated (OR, 5.9; 95% CI, 16.0-2.9;
We observed substantial heterogeneity and recurrent potential for publication bias among studies that were available for the primary and most secondary analyses. We thus performed a separate analysis according to study design. While prevalence studies yielded a high heterogeneity score, no measurable heterogeneity was found for case-control studies.
Furthermore, separate analysis stratified by type of diagnostic test used were conducted. This revealed numerically higher prevalence of SIBO when LBT was used for SIBO diagnosis instead of studies utilizing GBT or culture-based methods. Nevertheless, at least moderate heterogeneity was found for each subgroup analysis. Similar, subgroup analysis for cases-control studies was not possible due to the small number of studies utilizing different diagnostic modalities for SIBO diagnosis.
To explore heterogeneity of the primary analysis, we did an additional sensitivity analysis, by restricting the analysis only to only “high-quality” studies based upon NOS and the JBI appraisal tool. However, among the studies meeting this criterion, heterogeneity was also high and the potential for bias remained. Given this, the high heterogeneity scores and the obvious high risk of bias is most likely explained by extraneous but unreported features of the prevalence studies included in this systematic review and meta-analysis. It is noteworthy, that most prevalence studies included in these meta-analyses were based upon retrospective audits, of frequently poorly characterized study cohorts, with insufficient information regarding selection criteria or potential confounders (eg, PPI-use, previous antibiotic therapies, or probiotic use). This potentially could explain the increased SIBO prevalence in SSc-patients when prevalence studies are compared to case-control studies. Furthermore, 8 out of the 9 case-control studies used healthy (asymptomatic) subjects as controls, minimizing the risk of bias.
The lack of data on methane positivity during BT in patients with SSc is one of the limitations of this meta-analysis. Only 3 studies reported methane positivity (in addition to hydrogen) during breath testing to diagnose SIBO in SSc. Methane, produced by Archaea (and not bacteria), is believed to slow intestinal transit and is often associated with constipation.53-55 The importance of breath methane measurements in subjects with suspected intestinal dysbiosis, is emphasized by the guideline of the American College of Gastroenterology for SIBO.56 Indeed, the term, intestinal methanogen overgrowth has been coined to highlight the distinct importance of methane production by methanogens (Archaea) as compared to hydrogen positive SIBO caused by bacteria. Consequently, failure to measure methane, will result in an underestimation of the SIBO prevalence and is likely to influence the outcomes this meta-analysis.
SSc-patients frequently have severe motility disturbance of the esophagus, manifesting with heartburn, dysphagia, and regurgitation57 requiring treatment with PPI. A meta-analysis published by Su et al58 found that treatment with PPI and the subsequent chronic gastric acid suppression is linked with a moderate increase of SIBO in a variety of GI disease conditions (OR, 1.7; 95% CI, 1.2-2.4). However, a recent study showed that, PPI was not associated with an increased rate of SIBO, although modest changes were seen in the small intestinal microbiome in PPI users, including a notable reduction in relative abundance of the family Clostridiaceae.59 Thus, the effect of PPI on the small intestinal microbiome remains inconclusive. In the current systematic review and meta-analysis, although limited by a small sample size, PPI use was numerically (but not significantly) higher in SIBO positive SSc-patients (OR, 2.3; 95% CI, 0.8-6.4;
Increased duration of disease was significantly related with an increased SIBO prevalence rates in SSc-patients, suggesting that SIBO is potentially the consequence (and not the cause) from the worsening global GI dysmotility seen with disease progression, resulting in subsequent stasis of luminal contents promoting bacterial overgrowth in SSc.
This systematic review and meta-analysis revealed that a short course of antibiotic therapy or octreotide treatment were effective in treating SIBO in SSc-patients. In more than 60% of SSc-patients, antibiotic therapy resulted in a significant symptom improvement and in more than 60% of SSc-patients in normalization of breath tests. Overall, it appears that rifaximin was potentially twice as effective as rotating antibiotics for SIBO treatment in SSc, while our analysis could not account for potential confounders. Antibiotic therapy was well tolerated and only a small proportion of patients developed mild GI symptoms. Moreover, there was reduction in FC after SIBO eradication with antibiotic therapy in SSc-patients. Similarly, although only limited data is available,32,35 treatment of SIBO in SSc with octreotide, exerting a prokinetic effect,32 improved abdominal symptoms significantly and normalized hydrogen breath tests. Thus, antimicrobial therapy may be an effective therapeutic option to improve GI symptoms in SSc, similar to that reported for treating small intestinal dysbiosis and symptoms in FGIDs60 and IBD.61 Furthermore, only 1 open-label study reported on the efficacy of probiotics in treating SIBO in SSc-patients and found
Subgroup analysis for different countries and regions of the world, found the increased SIBO prevalence rates in the US, followed by the European countries and the lowest rates were from the Asian countries. In all these countries a combination of BT and/or small bowel aspirate and culture was used for SIBO diagnosis. It can be speculated that this variation in SIBO prevalence is caused by environmental factors such as diet or the background risk of GI infections.
The only earlier meta-analysis12 incorporated 14 studies as compared to 28 studies used for the primary analysis for this meta-analysis. This increased number of studies enabled detailed analyses of subgroups of the included studies and allowed to explore the heterogeneity inherent to these studies. The larger sample size also enabled additional analyses in relation to other predictors or risk factors for SIBO in SSc including PPI use, antibiotic therapy and octreotide, environmental factors like geographic region or the role of potential biomarkers such as FC, markers of inflammation and markers of malnutrition for SIBO in SSc. Furthermore, we assessed the impact of SIBO on GI symptoms and other systemic manifestations of SSc. Nevertheless, there are limitations of this systematic review and meta-analysis. The diagnosis of SIBO is always fraught by the absence of an appropriately validated and clinically accepted diagnostic tests. Furthermore, the various case-control studies included “healthy asymptomatic subjects” as well as diverse patient cohorts as controls. In addition, small sample sizes (eg, < 50 subjects per group) in some of the studies limited the statistical power of some of the sub-group analyses.
Collectively, this systematic review and meta-analysis observes an increased prevalence of SIBO in SSc-patients as compared to controls. However, there was no difference in SIBO prevalence between lcSSc and dcSSc. Diarrhea is strongly associated with SIBO in SSc-patients, while the risk of SIBO increases with longer disease duration and PPI use. While the data are limited, FC is increased in SIBO positive SSc-patients, but the role of other markers of inflammation and malnutrition, remains uncertain. Antibiotic treatment (or treatment with octreotide) of SIBO (predominantly rifaximin) in SSc-patients and/or octreotide results in a significant symptom improvement and in a significant proportion of patient’s the BT are normalized. On the other hand, moderate heterogeneity was found by the comparative analysis in addition to risk of bias. Furthermore, there is substantial “clinical heterogeneity,” which is most likely due to the absence of uniform criteria for selection of cases, a consequence of potential confounders and lack of validated tests to diagnose SIBO. Based upon this, the overall reliability of the evidence available must be considered low, and the results need to be interpreted with caution.
Note: To access the supplementary materials and methods, tables, and figures mentioned in this article, visit the online version of
The authors would like to acknowledge our Librarian, Mr Marcos Riba who has assisted with the literature search.
None.
Ayesha Shah, Michael P Jones, Mark Morrison, and Gerald Holtmann work for AGIRA (Australian Gastrointestinal Research Alliance).
Ayesha Shah, Veenaa Pakeerathan, Kate Virgo, and Gerald Holtmann: study idea, concept and design, data extraction and interpretation of data, and drafting of the manuscript; Ayesha Shah and Veenaa Pakeerathan share equal first co-authorship; Thomas Fairlie and Mark Morrison: drafting of the manuscript and review of final manuscript; Mike Jones: data analysis, drafting of the manuscript, and review of final manuscript; Uday C Ghoshal: critical input in the study and editing of the paper; and Purna C Kashyap: critical input in the study and editing of the paper.
 |
 |
