A preliminary study of bowel rest strategy in the management of Clostridioides difficile infection

Clostridioides difficile infection (CDI) is an important nosocomial infection and is the leading cause of infectious diarrhea in hospitalized patients. We aimed to assess the effect of bowel rest on the management of CDI. A single-center retrospective cohort study was conducted. The primary outcome was the composite of the all-cause mortality and CDI recurrence within 30 days. The main secondary outcome was switching from metronidazole to vancomycin. Of the 91 patients with CDI enrolled as the full cohort, 63 patients (69%) and 28 patients (31%) constituted the control group and the bowel rest group, respectively. After one-to-one propensity score matching, a total of 46 patients were included as the matched cohort. In the full cohort, the composite outcome occurred in 19.0% and 14.3% of the patients in the control and the bowel rest group, respectively (p = 0.768). In the matched cohort, it was 17.4% in each group. Although there was no statistically significant difference, the trend of switching was lower in the bowel rest group. The bowel rest may not affect the all-cause mortality and CDI recurrence within 30 days. However, in those prescribed bowel rest, switching from metronidazole to vancomycin may reduce.

www.nature.com/scientificreports/ cal characteristics were compared between the two groups: age, sex, comorbidities (diabetes, malignancy, and inflammatory bowel disease), initial treatment for CDI (metronidazole, vancomycin, or fidaxomicin), antibiotic use (pre-treatment and post-treatment), medications [proton pump inhibitor (PPI), histamine H2-receptor antagonist (H2RA), probiotics, immunosuppressants, chemotherapy, and vasopressor], tube feeding, surgery, intensive care unit (ICU) stay, and laboratory data (serum albumin, creatinine levels, and white blood cell count). Data of all the patients were extracted from electronic medical records in our hospital.
Outcomes. The primary outcome was the composite the all-cause mortality and CDI recurrence within 30 days. CDI recurrence was defined as the administration of metronidazole, vancomycin, or fidaxomicin for the treatment of clinically suspected CDI at least one day after the end of the initial treatment for CDI. The main secondary outcome was the percentage of patients who underwent continuous switching from metronidazole to vancomycin. The characteristics of the patients were compared between groups both before and after propensity score matching.

Statistical analyses.
To minimize selection bias in the bowel rest group, propensity score matching was done. For estimating the propensity score, a logistic regression analysis was performed using bowel rest as the independent variable. The dependent variables were age, sex, comorbidities (diabetes, malignancy, and inflammatory bowel disease), initial treatment of CDI, antibiotic use (pre-treatment), medications (PPI, H2RA, probiotics, immunosuppressants, chemotherapy, and vasopressor), tube feeding, surgery, ICU stay, and laboratory data (albumin, creatinine, and white blood cell count). Thereafter, one-to-one matching without replacement was performed using the nearest neighbor match based on estimated propensity scores of each patient with a caliper width set to 0.2 times the standard deviation of the logit of the propensity score. The C-statistics and sideby-side box plots were used to check the balancing of the matching. Clinical characteristics and outcomes were compared between the control group and the bowel rest group and separately in the full cohort and matched cohort. To check the significance of the difference in continuous variables and categorical variables, Mann-Whitney's U test and Fisher's exact test were used, respectively. Statistical analyses were performed in April 2020 using R (version 4.0.0) and EZR software 7 . A p-value < 0.05 was considered statistically significant.

Results
After the application of the inclusion and exclusion criteria, a total of 91 patients identified with CDI were recruited and they comprised the full study cohort. Among these, 63 patients (69%) were categorized into the control group and 28 patients (31%) into the bowel rest group. Metronidazole was the initial treatment for CDI in approximately 80% of the patients. Table 1 shows the clinical characteristics of the full study cohort. The two groups were comparable in all the parameters except sex which showed that the number of males was higher in the control group than the bowel rest group (p = 0.023). After propensity score matching (C-statistics 0.77), a total of 46 patients (23 patients in each group) were included in the matched cohort (Fig. 1). Figure 2 depicts the side-by-side box plots before and after matching. Table 2 shows the clinical characteristics in the matched cohort; there were no statistically significant differences between the groups.
The results of outcome analyses before and after matching are shown in Table 3. In the full cohort, the composite outcome occurred in 19.0% of the patients (95% CI 10.2-30.9) in the control group and 14.3% patients (95% CI 4.0-32.7) in the bowel rest group. In the matched cohort, the composite outcome occurred in 17.4% of the patients (95% CI 5.0-38.8) in each group. Both before and after matching, there were no statistically significant differences in the primary and secondary outcomes. However, the percentage of patients being switched from metronidazole to vancomycin showed a trend to be lower in the bowel rest group than in the matched cohort

Discussion
In the present study, we demonstrated two major clinical findings relating to bowel rest in the management of CDI. First, there were no statistically significant differences in both the composite outcome and its components (all-cause mortality or CDI recurrence) within 30 days. Second, the percentage of patients who were switched from metronidazole to vancomycin also showed no statistically significant difference between the two groups; however, it showed a lower trend in the bowel rest group in both the full cohort analysis and the propensity score-matched cohort. www.nature.com/scientificreports/ The first finding that there were no statistically significant differences in the all-cause mortality or CDI recurrence within 30 days suggests that bowel rest may not affect these outcomes which is in contrast to our hypothesis. Previous studies have reported all-cause mortality of 3.4-15.1% within 30 days in patients with CDI which is similar to that reported in our study (approximately 13%) 8 . Meanwhile, the CDI recurrence rate within 30 days was approximately 4% in our study; this was within the previously reported range of 3.3-27.3% 8 . Since CDI recurrence was defined as the repetition of treatment for CDI after the initial treatment in our study, we believe that CDI recurrence may have been underestimated compared to studies where the definition of CDI was defined by the recurrence of diarrhea alone. Based on the value of C-statistics and side-by-side box plots, our propensity score matching was considered well-balanced. Although the sample size may be insufficient to conclude, our data suggests that there were no significant clinical differences in the all-cause mortality or CDI recurrence within 30 days between the control group and bowel rest group.
The second finding that the proportion of switching from metronidazole to vancomycin had a trend to be lower in the bowel rest group suggests that those on bowel rest may have a lower vancomycin use. We believe that the cause of this trend may be because diarrhea is more likely to sustain in patients continuing oral or enteral feeding. As a result, the initial treatment with metronidazole may be judged as clinically failed. Although oral metronidazole is no longer recommended as the initial treatment for CDI under the current American guideline 5 , oral metronidazole is still used as the primary regimen for CDI in Japan 9 . Unnecessary use of oral vancomycin has the potential risk of developing highly resistant bacteria such as vancomycin-resistant enterococci, although this is debatable 10 . Thus, we suggest that the use of vancomycin can be reduced by giving bowel rest instead; this would prove beneficial towards the antimicrobial stewardship.
The present study has some limitations. First, this study was a single-center retrospective study and the number of analyzed patients was small. Furthermore, we did not examine all the factors that can affect the management or outcomes; e.g., the patients' activities of daily living, the severity of diarrhea, other comorbidities, reasons for hospitalization, changes of intestinal microbiota, and the strain of C. difficile. We assessed outcomes within 30 days; thus, long-term outcomes may have differed. Therefore, further well-designed investigations with a higher number of patients are warranted.