Effectiveness of adjunctive use of antibiotics in acute severe ulcerative colitis: a propensity score matching analysis
Article information
Abstract
Background/Aims
The exact cause of ulcerative colitis (UC) remains unclear but likely involves an immune response to an unidentified component of gut microbiota. The use of antibiotics in treating acute severe UC (ASUC) remains controversial. This study aimed to assess the benefits of adding antibiotics to corticosteroids for managing ASUC.
Methods
This retrospective study included patients with ASUC who met the Truelove and Witts diagnostic criteria at admission and received intravenous (IV) corticosteroids, with or without adjunctive antibiotics. Key outcomes included the need for medical and surgical rescue therapy during hospitalization and UC-related re-hospitalization rates at 1 week, 3 months, and 1 year. Propensity score matching (PSM) was employed to minimize selection bias in evaluating the impact of adjunctive antibiotics on clinical outcomes.
Results
Of the 386 screened patients, 222 met the inclusion criteria: 75 (33.8%) received IV corticosteroids alone, while 147 (66.2%) received IV corticosteroids combined with antibiotics. After PSM (60 patients per group), the addition of antibiotics showed no significant effect on the need for medical rescue therapy during hospitalization, re-hospitalization, or clinical response based on the partial Mayo score after 1 week of hospitalization (68.3% vs. 60.0%; P= 0.341). Similarly, no significant differences were observed in the need for medical rescue therapy, re-hospitalization, or UC-related surgery rates at 3 months or 1 year (P> 0.05).
Conclusions
In patients with ASUC, treatment outcomes did not significantly differ between those receiving IV corticosteroids alone and those receiving a combination of IV corticosteroids and antibiotics, suggesting that adjunctive antibiotic use may offer no added clinical benefits.
INTRODUCTION
Inflammatory bowel diseases, including Crohn’s disease and ulcerative colitis (UC), are complex disorders characterized by dysregulated immune responses to intestinal microbes in genetically predisposed individuals. Antibiotics are proven effective in specific clinical scenarios, such as treating perianal fistulas or intra-abdominal abscesses in Crohn’s disease and managing acute pouchitis following total proctocolectomy [1,2]. However, their role in UC treatment remains unclear. UC is a chronic disease with no definitive cure, characterized by recurring periods of remission and flare-ups throughout a patient’s lifetime, often with the potential for life-threatening complications [3-5].
The 10-year risk of disease relapse in patients with UC is approximately 70% to 80%. Among these patients, 50% will require hospitalization, with a 5-year re-hospitalization rate of 50%. The risk of colectomy within 5–10 years is estimated at 10% to 15% [6-8]. Additionally, about 25% of patients with UC will experience at least one episode of acute severe UC (ASUC) during their disease course [9]. A meta-analysis of population-based studies over the past two decades estimates the mortality rate of ASUC at approximately 1% [10]. Without prompt and appropriate treatment, ASUC can rapidly progress to a life-threatening medical emergency, often requiring colectomy [10,11]. The diagnosis of ASUC is based on Truelove and Witts criteria, which include more than 6 bloody stools per day along with symptoms such as fever and anemia [12,13].
Alterations in the gut microbiota are thought to play a role in the pathogenesis of UC, providing a theoretical rationale for antibiotic use [14-16]. ASUC is often associated with elevated inflammatory markers, such as C-reactive protein (CRP), and fever in some patients. Some experts advocate for antibiotic use in its management, particularly in cases with signs of infection [17]. However, antibiotics disrupt intestinal microbiota, leading to significant microbial shifts [18]. As a result, their use in ASUC remains controversial because of their dual effects on inflammation and microbial balance. Although antibiotics may help control inflammation by targeting harmful bacteria, they can also disrupt gut microbiota, exacerbate dysbiosis, and potentially worsen disease progression.
Some studies suggest that combining antibiotics with intravenous (IV) corticosteroids lowers the likelihood of requiring additional treatment during hospitalization [19-21]. However, other studies report conflicting results, failing to demonstrate significant benefits from specific antibiotics. These discrepancies may result from variations in study design and patient characteristics. Overall, evidence supporting this practice remains limited, as previous studies had small sample sizes and significant variability in antibiotic type, dose, and duration. The Current European Crohn’s and Colitis Organisation guidelines recommend IV corticosteroids as the first-line treatment for ASUC [22]. However, clear evidence for the effectiveness of antibiotics as an adjunctive treatment in ASUC is still lacking. Few randomized controlled trials (RCTs) have been conducted in the Western population [20,21], while none have focused on Asian patients. Additionally, only a limited number of studies have utilized propensity score matching (PSM) analysis to minimize bias.
Therefore, this study aimed to evaluate the role of adjunctive antibiotics in the management of hospitalized patients with ASUC, assessing their potential benefits and limitations when used alongside IV corticosteroid therapy.
METHODS
1. Study Design and Participants
This retrospective, single-center PSM cohort study analyzed 386 patients hospitalized for ASUC at a tertiary referral hospital between November 2005 and July 2024. The inclusion criteria were: (1) adults aged 18 years or older and (2) patients hospitalized for first-line ASUC treatment who received IV corticosteroids. The exclusion criteria were: (1) age <18 years, (2) insufficient laboratory data or incomplete clinical information that precluded definitive assessment of ASUC diagnosis or clinical outcomes, (3) coexisting infections—such as sepsis, pneumonia, urinary tract infection, or Clostridioides difficile infection (CDI)—requiring antibiotic treatment, and (4) patients in whom corticosteroids had been administered as pretreatment prior to planned IV infliximab induction. To evaluate the clinical outcomes of adjunctive antibiotic use in ASUC, patients receiving both IV corticosteroids and adjunctive antibiotics were matched 1:1 using PSM with those who did not receive antibiotics. This study was conducted in accordance with the ethical guidelines of the 1975 Declaration of Helsinki and was approved by the Institutional Review Board of Severance Hospital (IRB No. 4-2024-1226). The requirement for written informed consent was waived.
2. Diagnosis, Treatment, and Assessment of Disease Activity in ASUC
ASUC is diagnosed based on the Truelove and Witts criteria, which include: (1) more than 6 bloody stools per day, along with at least one marker of systemic involvement, such as (2) tachycardia (heart rate exceeding 90 beats per minute), (3) fever (core temperature above 37.8°C), (4) elevated inflammatory markers—defined as erythrocyte sedimentation rate greater than 30 mm/hr or CRP levels above 30 mg/L, and (5) anemia, indicated by hemoglobin levels below 10.5 g/dL [22]. These criteria help assess disease severity, ensuring accurate diagnosis and timely therapeutic intervention. Current guidelines recommend IV corticosteroids as the first-line treatment for ASUC, with a dose of 0.8–1.0 mg/kg (methylprednisolone equivalent) for 5 to 7 days [23]. Clinical disease activity in UC is evaluated using the partial Mayo score, which assesses 3 parameters: rectal bleeding, stool frequency, physician’s global assessment. It does not include endoscopic findings. Each parameter is scored from 0 to 3, with a total score ranging from 0 to 12. A score of 3–5 indicates mild disease, 6–10 indicates moderate disease, and 11–12 indicates severe disease [24]. In accordance with the Korean Association for the Study of Intestinal Diseases (KASID) guidelines [25], all hospitalized ASUC patients undergo early flexible sigmoidoscopy to assess disease extent and severity, along with baseline laboratory testing and abdominal imaging (plain radiography or contrast-enhanced computed tomography). Endoscopic evaluation was performed cautiously by experienced endoscopists using minimal insufflation to reduce the risk of procedure-related complications. All hospitalized patients underwent routine testing for CDI at admission, in accordance with institutional protocols aligned with the 2017 Korean guidelines for UC. CDI screening was performed using stool toxin assays or polymerase chain reaction for toxigenic C. difficile. Screening for cytomegalovirus (CMV) reactivation was not performed universally at admission but was conducted selectively in patients with clinical suspicion, such as lack of response to IV corticosteroids or severe mucosal ulceration on endoscopy. CMV colitis was evaluated by immunohistochemical staining for CMV antigen on colonic biopsy. When performed, these diagnostic evaluations were conducted before or concurrently with corticosteroid initiation to ensure diagnostic accuracy and minimize the risk of inappropriate immunosuppression. Infectious etiologies such as C. difficile and CMV were routinely screened using stool toxin polymerase chain reaction and colonic biopsy-based immunohistochemistry, respectively, particularly in cases refractory to initial corticosteroid therapy. Laboratory markers, including CRP, hemoglobin, and albumin, were monitored daily during corticosteroid treatment. No endoscopy-related adverse events were observed during hospitalization.
3. Outcomes
The primary outcome was the requirement for in-hospital rescue therapy, including biologic treatment with infliximab or re-hospitalization within the first week of admission. The coprimary outcome was the achievement of first clinical remission, defined as a reduction in the partial Mayo score of at least 3 points and a decrease of at least 30% from baseline, along with either a reduction of at least 1 point on the rectal bleeding subscale or an absolute rectal bleeding score of 0 or 1 [26]. This was assessed by comparing measurements taken immediately before hospitalization and 1 week after admission. Secondary outcomes included the requirement for rescue therapy—such as immunomodulators (azathioprine, methotrexate), biologics (infliximab, golimumab, adalimumab, vedolizumab, ustekinumab), or small-molecule agents (tofacitinib, upadacitinib)—as well as the initiation of immunomodulators (azathioprine, methotrexate) as maintenance therapy, re-hospitalization, or UC-related surgery at 3 months and 1 year.
4. Statistical Analyses
Descriptive analyses were conducted to compare the demographic and baseline characteristics of patients receiving IV corticosteroids alone versus those treated with both IV corticosteroids and antibiotics. Continuous variables were reported as means with standard deviations and were compared using the independent two-sample t-test. Categorical variables were expressed as percentages and were analyzed using either the chi-square test or Fisher exact test, depending on expected frequencies. Survival analysis was conducted to assess the cumulative relapse-free survival of UC. The cumulative rate was estimated using the Kaplan-Meier method, and the survival curves for the 2 groups were compared using the log-rank test. The proportional hazards assumption was tested using the Kolmogorov-type supremum test and satisfied for all survival analyses (P=0.293, P=0.082, P=0.437). To minimize potential biases from baseline imbalances common in retrospective studies, PSM was applied to balance covariates between the groups. To evaluate outcomes conditional on early treatment response, landmark analyses were additionally performed by excluding patients who experienced clinical relapse before each landmark time point (1 week and 3 months) and analyzing subsequent relapse-free survival in the remaining cohort. All statistical tests were two-sided, with P-values < 0.05 considered statistically significant. Data analyses were performed using SAS 9.4 (SAS Institute, Cary, NC, USA).
RESULTS
1. Baseline Characteristics of the Study Population
A total of 386 patients with UC were screened for inclusion. After applying the exclusion criteria, 222 patients with ASUC met the inclusion criteria. Among them, 75 (33.8%) received IV corticosteroids alone, while 147 (66.2%) received IV corticosteroids combined with antibiotics. The study flowchart is presented in Fig. 1. After excluding unmatched individuals through PSM, 120 patients who received IV corticosteroids were included in the final analysis. The demographic and clinical characteristics of the study population are summarized in Table 1. After PSM, the 2 groups were well balanced in terms of age, sex, body mass index, disease duration, smoking status at diagnosis, disease extent, disease activity index (partial Mayo score), laboratory parameters, fecal calprotectin, concomitant medications (5-aminosalicylic acid, sulfasalazine, immunomodulators, biologics, and small molecular agents), corticosteroid dose and duration, and length of hospital stay (Table 1). According to the UC Montreal classification, proctitis, left-sided colitis, and pancolitis were observed in 4 (6.7%), 29 (48.3%), and 27 (45.0%) patients in the IV corticosteroid group, and in 3 (5.0%), 27 (45.0%), and 30 (50.0%) patients in the IV corticosteroid and adjunctive antibiotic group, respectively. The mean age was 46.6 ± 16.4 years, while the mean disease duration was 5.08 ± 5.98 years. The mean length of hospitalization and duration of IV antibiotics were 10.4 ±6.1 and 10.4 ± 5.9 days, respectively.
Flowchart of the study population. Among 386 patients with UC treated with IV corticosteroids, 222 met the inclusion criteria. Of these, 75 (33.8%) received IV corticosteroids alone, while 147 (66.2%) received IV corticosteroids with antibiotics. After 1:1 propensity score matching, 60 patients from each group were included in the final analysis. Patients were excluded primarily due to insufficient laboratory data or incomplete clinical information, which precluded accurate assessment of ASUC diagnosis or clinical outcomes. UC, ulcerative colitis; IV, intravenous; ASUC, acute severe ulcerative colitis.
Baseline Characteristics of the Study Population in Patients with ASUC Using PSM
Most patients (95%) received IV hydrocortisone (62.5–450 mg daily), while the remaining patients received IV methylprednisolone (40–75 mg daily). For analysis, these doses were converted, with 50 mg of methylprednisolone considered equivalent to 250 mg of hydrocortisone. The most commonly used IV antibiotics were a combination of third-generation cephalosporin and metronidazole (56.7%), followed by quinolone and metronidazole (25.0%). Monotherapy with a third-generation cephalosporin and quinolone was used in 10.0% and 6.7% of cases, respectively. Tazobactam was administered to only 1.7% of patients.
2. Clinical Outcomes at 1 Week, 3 Months, and 1 Year
After PSM, no significant differences were observed between the groups at 1-week post-hospitalization in terms of infliximab use (P>0.999) or re-hospitalization rates (P>0.999) (Table 2). The first clinical remission rate was 68.3% in the IV corticosteroid group and 60.0% in the IV corticosteroid combined with the adjunctive antibiotics group (P=0.341) (Fig. 2). The Kaplan- Meier analysis of relapse-free survival also showed no statistically significant difference between the 2 groups (log-rank P=0.947) (Fig. 3), suggesting that the use of adjunctive antibiotics did not significantly impact relapse-free survival at 1-week post-hospitalization.
Primary and Secondary Outcomes as Observed in Our Study Population with PSM
Comparison of clinical remission rates (defined as ≥3-point improvement in the partial Mayo score) at 1 week in patients with acute severe ulcerative colitis treated with intravenous corticosteroids alone versus adjunctive antibiotics (propensity score matching analysis).
Comparison of cumulative relapse-free survival rates at 1 week in patients with acute severe ulcerative colitis treated with intravenous corticosteroids alone versus those receiving adjunctive antibiotics (propensity score matching analysis).
At 3 months, the use of biologics, immunomodulators, and re-hospitalization rates were similar between the groups (P>0.05). However, the use of small-molecule agents was more frequent in the corticosteroid-alone group prior to PSM (P=0.038), though this difference was no longer significant after matching (P>0.999). UC-related surgery was rare, with no significant difference between the groups (P>0.999) (Table 2). The Kaplan-Meier analysis showed no significant difference in relapse-free survival at 3 months between the groups (Fig. 4).
Comparison of cumulative relapse-free survival rates at 3 months in patients with acute severe ulcerative colitis treated with intravenous corticosteroids alone versus those receiving adjunctive antibiotics (propensity score matching analysis).
At 1 year, there were no statistically significant differences between the groups in the use of biologics, immunomodulators, small-molecule agents, re-hospitalization rates, or UC-related surgery (P>0.05) (Table 2). The Kaplan-Meier analysis demonstrated no significant difference in relapse-free survival at 1 year between the groups (Fig. 5). Overall, adjunctive antibiotic use did not significantly impact clinical outcomes at 1 week (P=0.947), 3 months (P=0.749), or 1 year (P=0.969) in patients with ASUC receiving IV corticosteroids. For descriptive purposes, separate Kaplan-Meier plots based on landmark analysis showing the relapse-free survival distributions at 3 months and 1 year are provided in the Supplementary Figs. 1 and 2 (P=0.625 and P=0.863, respectively).
Comparison of cumulative relapse-free survival rates at 1 year in patients with acute severe ulcerative colitis treated with intravenous corticosteroids alone versus those receiving adjunctive antibiotics (propensity score matching analysis).
DISCUSSION
This retrospective study using PSM investigated the impact of adjunctive antibiotics in ASUC. The main finding was that there was no statistically significant difference in outcomes between treatment with IV corticosteroids alone and a combination of IV corticosteroids and antibiotics. This observation is particularly noteworthy given the limited supporting evidence for antibiotic use in ASUC, especially considering that adjunctive antibiotics are frequently prescribed for inflammatory exacerbations in various immune-mediated diseases despite guideline recommendations and expert opinion [27].
Several studies have explored the potential benefits of antibiotic use in ASUC treatment. A large retrospective cohort study and meta-analysis revealed that the requirement for inhospital medical rescue therapy was significantly lower in patients receiving antibiotics than in those treated with steroids alone (21.0% vs. 37.5%: adjusted odds ratio, 0.42; 95% confidence interval, 0.19–0.93; P=0.04) [28]. Similarly, a double-blind, placebo-controlled trial demonstrated that a 2-week course of amoxicillin, tetracycline, and metronidazole significantly improved clinical response rates (44.8% vs. 22.8%; P=0.001) and facilitated steroid withdrawal in active UC [29]. Although several prospective RCTs have been conducted to evaluate the role of adjunctive antibiotics in ASUC, most were limited by small sample sizes, heterogeneous antibiotic combinations, and inconsistent results regarding clinical efficacy. In contrast, consistent with our findings, other RCTs have demonstrated that adjunctive IV metronidazole or ciprofloxacin did not provide additional benefits for patients with ASUC [19,20]. Similarly, another study found no evidence supporting IV ceftriaxone and metronidazole use in these patients [21]. Additionally, randomized trials conducted in Western populations have demonstrated no clear benefit of antibiotics in ASUC [30,31]. A double-blind trial involving 40 patients with idiopathic colitis found no significant benefit of vancomycin, with similar improvement rates in both the vancomycin (89%) and placebo (89%) groups [30]. Likewise, a prospective double-blind trial evaluated IV metronidazole as an adjunct to corticosteroids in 39 patients with severe UC. After 5 days, remission rates were similar between the metronidazole (74%) and placebo (70%) groups, with no significant impact on colectomy rates [31].
ASUC is characterized by severe inflammation and an altered gut microbiome, with a loss of beneficial bacteria (Bifidobacterium, Lactobacillus) and an overgrowth of pro-inflammatory species (Enterobacteriaceae, Bacteroides, Fusobacterium varium). This imbalance contributes to increased gut permeability, immune activation, and epithelial damage [32-34]. Microbial diversity is closely linked to bacterial load, and bacteria are increasingly recognized for their significant role in mucosal inflammation [35,36]. Substantial evidence suggests that colonic luminal bacteria influence disease activity, indicating that targeted antimicrobial therapy, in combination with traditional immunosuppression, may be beneficial in UC management [29]. However, while antibiotics can help eliminate harmful bacteria, they also disrupt the microbial balance by reducing diversity, impairing the production of short-chain fatty acids—essential for mucosal healing—and promoting antibiotic-resistant pathogens, such as C. difficile [15,18,37]. As a result, the routine use of antibiotics remains controversial and should be approached with caution because of their potential to exacerbate gut dysbiosis.
This study has several clinical implications. First, our findings suggest that routinely adding antibiotics to IV corticosteroids in hospitalized Korean patients with ASUC does not confer short- or long-term benefits. Second, this study utilized PSM to minimize baseline differences between groups, thereby reducing potential confounding factors and strengthening the validity of the results. Although some short-term outcomes appeared numerically worse in the non-antibiotic group, this likely reflects confounding by indication, as antibiotics might possibly have been prescribed empirically to patients perceived to have greater inflammatory severity rather than reflecting a true therapeutic benefit of antibiotics. The absence of significant improvement in the antibiotic group argues against a clinically meaningful effect. Third, clinical outcomes were assessed at multiple time points (1 week, 3 months, and 1 year), providing a comprehensive evaluation of both short- and long-term effects of antibiotic use. Fourth, this study focused on hospitalized patients with ASUC who were managed according to standard clinical guidelines, enhancing the generalizability of its findings to real-world practice. Finally, the mean disease duration was 5.08 ±5.98 years, indicating that the majority of patients had an established diagnosis of UC prior to hospitalization. As the study was conducted at a tertiary referral center, many patients were referred after their initial diagnosis of UC. To establish more definitive clinical recommendations, further research—including larger, well-designed prospective studies and randomized trials—is warranted.
This study has several limitations. First, as a retrospective cohort study, inherent biases—such as unmeasured confounders—could not be entirely eliminated despite statistical adjustments. Notably, body temperature (fever), a key clinical determinant of empiric antibiotic use in ASUC, could not be reliably incorporated because of incomplete and inconsistent documentation in retrospective records. Accordingly, residual confounding related to clinical severity—including fever—cannot be fully excluded and may have contributed to indication bias. Second, although PSM helped reduce confounding, it inevitably reduced the sample size (60 patients per group), which may have limited the statistical power—particularly for detecting rare outcomes such as colectomy. Because this was an exploratory retrospective study without a predefined sample size, a post hoc power analysis was not performed, and the observed effect sizes may guide a future trial design. Third, variations in antibiotic types, dosages, and duration—determined at the discretion of the treating physician—may have introduced heterogeneity in treatment effects. This real-world variability may have attenuated regimen-specific effects, as our analysis focused on the overall impact of adjunctive antibiotic use rather than individual antibiotic strategies. Furthermore, antibiotics were administered empirically, without standardized or confirmed indications, which could further contribute to indication bias. Identification of risk factors for non-response to IV corticosteroids and assessment of whether adjunctive antibiotics influenced steroid responsiveness would be clinically valuable. However, the limited sample size after matching and the retrospective design precluded adequately powered multivariable or subgroup analyses. While CDI was routinely excluded, CMV reactivation screening was performed only in selected cases. Therefore, the possibility of undetected CMV colitis influencing treatment response—particularly in the steroid-only group—cannot be entirely ruled out. Fourth, we did not assess gut microbiota composition before and after antibiotic administration, limiting insights into the potential microbial mechanisms underlying our findings. Such microbiome-integrated approaches could elucidate host–microbe interactions and ultimately support more personalized therapeutic strategies in ASUC. Future studies integrating microbiome analyses are needed to identify subgroups that may benefit from selective antibiotic use. Finally, we cannot exclude the possibility that antibiotics might benefit specific subgroups of ASUC patients, which our limited sample size may have failed to identify.
In conclusion, our study suggests that adjunctive antibiotics cannot be routinely recommended, as they did not significantly impact early clinical remission, the need for additional rescue therapies, or long-term outcomes. Our findings contribute to the limited evidence base regarding antibiotic use in ASUC receiving IV corticosteroids. Additionally, future studies incorporating gut microbiome analysis might identify specific bacterial signatures that could guide selective antibiotic use in particular ASUC subgroups. Until such evidence is available, a more personalized therapeutic approach that carefully weighs the potential benefits against the risks of antibiotic-induced dysbiosis is warranted.
Notes
Funding Source
The authors received no financial support for the research, authorship, and/or publication of this article.
Conflict of Interest
Park J and Cheon JH are editorial board members of the journal but were not involved in the peer reviewer selection, evaluation, or decision process of this article. No other potential conflicts of interest relevant to this article were reported.
Data Availability Statement
The data underlying this article will be shared upon reasonable request from the corresponding authors.
Author Contributions
Conceptualization: Hyun HK, Cheon JH. Data curation: Hyun HK, Son NH, Cheon JH. Formal analysis: Hyun HK, Cheon JH. Investigation: Hyun HK. Methodology: Hyun HK, Son NH, Cheon JH. Project administration: Hyun HK, Cheon JH. Resources: Hyun HK, Park J, Park SJ, Park JJ, Kim TI, Cheon JH. Software: Hyun HK, Son NH. Supervision: Cheon JH. Validation: Hyun HK, Cheon JH. Visualization: Hyun HK, Son NH, Cheon JH. Writing–original draft: Hyun HK. Writing–review & editing: Hyun HK, Park J, Park SJ, Park JJ, Kim TI, Son NH, Cheon JH. Approval of final manuscript: all authors.
Supplementary Material
Supplementary materials are available at the Intestinal Research website (https://www.irjournal.org).
Supplementary Fig. 1.
Cumulative relapse-free survival at 3 months.
Supplementary Fig. 2.
Cumulative relapse-free survival at 1 year.