Low-Dose Metronidazole is Associated With a Decreased Rate of Endoscopic Recurrence of Crohn’s Disease After Ileal Resection: A Retrospective Cohort Study

Abstract

Background and Aims

Recurrence of Crohn’s disease after surgical resection and primary anastomosis is an important clinical challenge. Previous studies have demonstrated the benefit of imidazole antibiotics, but have been limited by adverse events and medication intolerance. We evaluated whether administration of low-dose metronidazole [250 mg three times per day] for 3 months reduces endoscopic postoperative recurrence rates.

Methods

We performed a retrospective cohort study of patients with Crohn’s disease who underwent ileal resection with a primary anastomosis and subsequently received care at our center. We compared the cases who received low-dose metronidazole for 3 months with control patients who did not receive this therapy. Data collected included demographics, risk factors for recurrence, and medications before and after surgery. The primary end point was the number of patients with ≥i2 [Rutgeerts] endoscopic recurrence by 12 months. Variables found to be predictive in univariate analysis at p < 0.10 were introduced in the Cox model for multivariate analysis.

Results

In all, 70 patients with Crohn’s disease [35 cases and 35 controls] met inclusion criteria. Risk factors for Crohn’s recurrence were similar between groups. The number of patients with ≥i2 endoscopic recurrence within 12 months following ileal resection was significantly lower in the metronidazole group [7 of 35 patients; 20%] compared with the number in the control group [19 of 35 patients; 54.3%] [p = 0.0058]. Eight participants [22.9%] in the metronidazole group experienced adverse events, and 3 of these patients [8.6%] discontinued the therapy.

Conclusion

Low-dose metronidazole reduces endoscopic recurrence of Crohn’s disease postoperatively and is well tolerated. This intervention should be considered as a therapy option following ileocolonic resection.

1. Introduction

Crohn’s disease [CD] is an idiopathic chronic inflammatory bowel disease, characterized by inflammation of the gastrointestinal tract. Due to progressive disease or stenosis and bowel obstruction, up to 70% of people with CD will require at least one surgical resection.^1–6 Despite resection of the affected segment of bowel, there is a significant risk of postoperative recurrence of the disease at or proximal to the surgical anastomosis. Therefore, identifying an intervention to reduce postoperative recurrence and the need for a repeat surgery is an important goal for the management of CD. Prevention of disease recurrence involves identifying the risk factors for recurrence as well as monitoring and instituting effective therapies.^7,8

Although there have been advances in prophylaxis therapy postoperatively, it remains a challenge to determine when to start therapy following resection, especially given that microscopic evidence of recurrence may be present within days after surgery.⁴ Initiating preventive therapy postoperatively has been limited by concerns for perioperative complications and uncertainty regarding timing of postoperative immune suppression.

One solution to reducing recurrence is using antibiotics immediately postoperatively. Antibiotics, in particular nitroimidazoles, have demonstrated postoperative benefit, with evidence of metronidazole and ornidazole succeeding in short-term prevention. In a prior randomized control study, Rutgeerts and colleagues randomized 60 patients who underwent ileal resection into a metronidazole [20 mg/kg] group or a placebo group for 3 months. After 1 year, there was significantly less endoscopic recurrence in the metronidazole group compared with the placebo group [13% vs 43%, p < 0.02]. However, this high dose of metronidazole was associated with significant side effects [n = 21] and patient dropout [n = 8], therefore limiting its subsequent use.⁹

Based on the above study, one of the members of our group instituted a lower-dose metronidazole regimen [250 mg three times daily for 3 months] in all of his patients undergoing ileocolonic resection. We hypothesized that low-dose metronidazole would be associated with decreased rates of endoscopic recurrence in CD patients after ileocecectomy compared with patients who have not received this therapy.

2. Materials and Methods

This study was approved by the University of Chicago Institutional Review Board. We performed a retrospective cohort study of patients with CD receiving care at the University of Chicago Medicine Inflammatory Bowel Disease Center [Chicago, IL]. We identified CD patients who underwent ileal resection and primary anastomosis who were prescribed low-dose metronidazole [250 mg three times daily] within 1 week of surgery for a period of 3 months. The control group did not receive low-dose metronidazole following surgery, but were otherwise managed similarly. Cases were matched 1:1 based on smoking status and disease duration and were excluded if there were lack of either clinical data or a matched control.

We identified patients who underwent ileal resection with primary anastomosis, defined as a resection of affected bowel, with unaffected bowel margins used for subsequent anastomosis. We included male and female patients 18 years of age or older with a confirmed diagnosis of CD. We excluded patients with [i] macroscopic evidence for CD proximally or distally to the site of resection, [ii] presence of a stoma, or [iii] previous intolerance or allergy to metronidazole.

2.1. Procedures and Measures

Data for this study were retrieved through GeneSys, our prospective digestive diseases database. A search of patients with IBD [using search terms ileocecetomy, ileal resection, and Crohn’s disease] was undertaken to retrieve the names and medical record numbers of those who fit the inclusion criteria. Patient information was de-identified and a spread sheet was created to collect information including: gender, age at diagnosis, age at time of surgery, exact type of anastomosis, previous resections, and endoscopic findings within 12 months after resection. All patients provided written informed consent.

The Rutgeerts post-operative grading score was utilized to assess the severity of lesions in the neoterminal ileum at the first endoscopic follow-up appointment.⁴ All colonoscopy reports and endoscopic images were reviewed and scored. At our center, it is routine practice for endoscopists to capture images of the terminal ileum and assign a Rutgeerts score following a colonoscopy. For this study, an independent observer, who was blinded to the patient’s group, then reviewed all these images to confirm the Rutgeerts score. Any discrepancies in the scoring were resolved by review by the senior investigator [DTR] and the original endoscopist.

Patients who received low-dose metronidazole postoperatively [cases] were evaluated using the same Rutgeerts score and compared with patients who did not receive antibiotics following ileal resection [controls]. Based on prior studies, we combined Rutgeerts i0 and i1 patients and Rutgeerts i2, i3, and i4 patients for comparison of ‘no recurrence’ and ‘clinically significant recurrence’ in order to dichotomize analysis.^5,6,10,11

2.2. Statistics

Sample size for the endoscopic recurrence rates at 12 months was calculated with the assumption of a 30% recurrence rate in the treatment group and a 70% recurrence rate in the control group. Using a power calculator, we determined that a minimum of 24 patients were needed in both groups with a power of 80%. To compensate for any possible dropouts, 35 patients were identified in the treatment arm and 35 patients in the control arm.

Data were analyzed by calculating p-values to determine statistical significance of recurrence between patients who were treated with low-dose metronidazole and those who were not treated with metronidazole. Continuous variables were dichotomized according to the median value or specific values known to be relevant in medical literature. Factors predictive of these events were studied. Variables that were found to be predictive in the univariate analysis at a p < 0.10 were included in the regression model for multivariate analysis. A p-value of <0.05 was considered significant. Relative risks were presented as mean estimates with a 95% confidence interval.

3. Results

Seventy patients were included in the study, with 35 participants in the low-dose metronidazole group matched to 35 participants in the control group. Characteristics, including race, sex, age at surgery, smoking status, number of previous resections, disease phenotype, and median time to endoscopic follow-up for the two groups were comparable [Table 1]. Thirty-two of the 35 patients who received low-dose metronidazole completed the course of therapy, with three patients discontinuing the study due to gastrointestinal intolerance [n = 2] or body rash [n = 1]. A total of eight participants in the metronidazole group experienced adverse events, including gastrointestinal intolerance, paresthesias, metallic taste in mouth, and rashes [Table 2].

Table 1.

Patient characteristics.

	Control	Low-dose metronidazole	p-value
Total number of patients	35	35
Median age at surgery [IQR]	34 [23–44]	39 [27–48]	0.10
Race			0.038
Caucasian	30	30
Hispanic	0	2
Black	1	2
Asian	0	1
Unknown, not Hispanic	4	0
Sex			0.227
Male	23	18
Female	12	17
Time to surgery [years]	9	10	0.204
Smoking status at surgery			1
Never	27	27
Former	6	6
Current	2	2
Number of previous resections			0.903
0	26	18
1	4	9
2	4	4
≥3	1	4
Phenotype			0.558
Penetrating	17	15
Non-penetrating	18	20
Median days to endoscopic follow-up [IQR]	192 [166–250]	184 [178–246]
Median days on metronidazole	–	90

IQR, interquartile range.

Table 2.

The number of patients with adverse events and the number of patients who discontinued therapy in the low-dose metronidazole group.

	Adverse effects	Dropouts
Total number of patients	8	3
Metallic taste	1	0
Gastrointestinal intolerance	4	2
Paresthesia	1	0
Rash	2	1

Endoscopic evaluation was performed in all patients within 12 months after surgery, with the median duration being 192 days [166–250] in the control group and 184 days [178–246] in the treatment group. Nineteen of 35 patients [54.3%] in the control group demonstrated endoscopic recurrence [Rutgeerts ≥ i2] in the neoterminal ileum within 12 months after ileal resection, compared with 7 of the 35 patients [20%] who took low-dose metronidazole for 3 months after surgery OR 0.27, 95% CI [0.09–0.7] p = 0.0096. Ten of the patients [28.6%] in the control group had no sign of recurrence [Rutgeerts i0], compared with 23 [65.7%] in the low-dose metronidazole arm [Figure 1].

Figure 1.

Percentage of patients with each Rutgeerts score after ileal resection with primary anastomosis

On multivariate analysis, only treatment with metronidazole was associated with a reduced rate of endoscopic recurrence, OR 0.21, 95% CI [0.065–0.610] p = 0.0058 [Figure 2]. There were numerically more patients in the control group on immunomodulator and biologic therapy; however, this difference was not statistically significant [Table 3]. Previous smoking was associated with postoperative disease recurrence on both univariate and multivariate analysis p = 0.0123 [Table 4].

Figure 2.

Percentage of patients without endoscopic recurrence [Rutgeerts score < i2] and with endoscopic recurrence [Rutgeerts score ≥i2] after ileal resection with primary anastomosis

Multivariate analysis: odds ratio [OR] 0.21, [95% CI: 0.065–0.610; p = 0.0058].

Table 3.

Medication at time of endoscopy.

	Low-dose metronidazole	Control	p-value
AZA/6MP	10	20	0.165
Methotrexate	4	3	0.219
Adalimumab	5	9	0.914
Cetolizumab pegol	0	4	0.922
Infliximab	1	1	0.996
Vedolizumab	1	0	0.997

P = 0.0058.

AZA = azathioprine.

Table 4.

Logistic regression analysis of endoscopic recurrence of Crohn’s disease.

	Univariate			Multivariate
Covariate	OR	95% CI	p-value	OR	95% CI	p-value
Metronidazole	0.267	0.094–0.708	0.0096*	0.211	0.065–0.610	0.0058*
Former smoker	5.333	1.371–26.53	0.0224*	7.402	1.696–41.98	0.0123*
Current smoker	2.000	0.226–17.74	0.5046	2.250	0.225–23.13	0.469

The asterisk indicates a statistically significant p-value in the multivariate analysis. OR = odds ratio; CI = confidence interval.

4. Discussion

This study demonstrated that CD patients who were given 3 months of low-dose metronidazole following ileal resection and anastomosis had significantly reduced endoscopic recurrence 1 year postoperatively, compared with those who did not receive this treatment [p = 0.0058]. Furthermore, 23 patients [65.7%] in the low-dose metronidazole group had no endoscopic ileal inflammation [Rutgeerts i0], and only 2 patients [5.7%] in that group had significant recurrence [Rutgeerts score of i3 or i4], demonstrating that this regimen of low-dose metronidazole may decrease both the incidence and severity of recurrent CD.¹⁰

In a prior prospective trial, Rutgeerts et al. randomized 80 patients to receive ornidazole 1 g/day for 1 year or a placebo for 1 year within 1 week of ileocolonic resection in patients with CD. That study found that ornidazole significantly reduced clinical recurrence [p = 0.0046] and endoscopic recurrence at 12 months [p = 0.037] postoperatively. However, there was a high rate of intolerance and patient dropout in the ornidazole group, limiting incorporation into standard treatment options.^9,10,12 D’Haens et al. followed up the aforementioned Rutgeerts et al. study with another randomized controlled trial examining whether metronidazole [250 mg three times per day] along with azathioprine [AZA] for three months lowered endoscopic recurrence in high-risk patients after 1 year. Their study found that significant endoscopic recurrence occurred in 14 of 32 [43.7%] patients in the treatment group compared with 20 of 29 [69.0%] in the placebo group [p = 0.048] 12 months postoperatively, further supporting the evidence that a regimen similar to this may decrease recurrence following ileal resection.¹³ Our study found that low-dose metronidazole had similar benefits to the high-dose metronidazole benefits seen in the Rutgeerts study, but had fewer side effects [70% of patients in the experimental group of the Rutgeerts study suffered adverse events, compared with 22.9% in this study]. Since low-dose metronidazole was both well tolerated and lowered endoscopic recurrence, we concluded that it is a promising therapy for patients in the clinical setting.

It is considered standard of care to use medications to prevent recurrent and progressive disease, with some medications providing more benefit than others. Probiotics, 5-aminosalicyclates [5-ASA] compounds, and corticosteroids have shown little to no benefit.^14–17 Tumor necrosis factor [TNF] antagonist therapies have shown the greatest promise in lowering rate of recurrence, with placebo-controlled studies demonstrating a significantly higher rate of 1-year endoscopic recurrence in the comparator arms compared with the anti-TNF arms.^1,11,14 However, these medications are expensive, and it is not clear whether they are needed in all patients who have undergone terminal ileal resection.

Since most patients receive medication following ileal resection, it was important to ensure that the significant decrease in endoscopic recurrence in the low-dose metronidazole compared with the control group was not due to confounding exposures to other therapies. It is of interest that there were more control patients than metronidazole patients on prophylactic immunosuppressive therapy, including biologic therapies [57.1% vs 28.6%]. If any bias were present, it would be in favor of the control patients, yet the benefit of metronidazole was still strongly significant.

More recent research has focused on the role the intestinal microbiome plays in the inflammatory process and the immune response in patients with CD. It is known that antibiotics alter the gut microbiota, and it is thought that the use of antibiotics, such as metronidazole, may either reduce the total volume of bacteria or shift the composition of the intestinal lumen away from the more pathogenic species and toward the advantageous bacteria. Specifically, studies have demonstrated that luminal bacteria are associated with an increased likelihood of recurrence of CD. Enterococcal species, Clostridium perfringens, and Bacteroides fragilis have been isolated more commonly in the mesenteric lymph nodes of patients with CD who underwent a surgical resection, compared with healthy controls. Metronidazole targets anaerobes such as Bacteroides fragilis, likely playing a role in decreased recurrence when using it postoperatively.^18,19

While the prevention of postoperative recurrence in patients with CD after ileal resection remains a difficult task, low-dose metronidazole is a promising drug therapy for future use in patients after ileal resection. Since our study only had three patients drop out in the metronidazole group, the side effects seem to be small and tolerable with the lower dose of the drug.⁸ Since recurrence of CD begins almost immediately after surgery, the concept of early medical preventive therapy is essential. Identification of the right treatment for the right patient, as well as identification of an affordable and safe therapy, has remained challenging. Given the effectiveness and safety of this regimen, we believe that, for many patients, it is reasonable to begin low-dose metronidazole immediately following surgery and monitor progress through a colonoscopy 6–12 months after surgical resection.^4,11,14

We used a Rutgeerts score of ≥i2 as signifying significant disease recurrence, in line with previous large prospective trials evaluating postoperative disease recurrence, and in line with the previous trials of the use of nitroimidazoles for the prevention of postoperative recurrence.^1,10–13 While there has been some debate about whether there is a differential risk of disease recurrence between the subclassifications of the Rutgeerts i2 score [i2a and i2b], a large study recently published [which included patients graded Rutgeerts i2a or i2b at their first postoperative colonoscopy] found no difference between these lesions with regards to risk of disease recurrence.^20,21

Although our study is consistent with previous studies, it is important to address the limitations of this study, including the fact that it was a retrospective review of medical records. Special care was taken, however, to avoid known risk factors—such as smoking status, duration of disease before surgery, phenotype, and number of prior resections—and endoscopists’ bias. Furthermore, all the participants in this study were treated at the tertiary University of Chicago IBD Center, possibly limiting generalizability. It is also important to address the limitation that more patients were on immunosuppressive therapy in the control group compared with in the experimental group. On one hand, given that the patient groups appeared comparable, it is possible that the increased exposure to immunosuppressive therapy in the control group may have blunted the difference between the low-dose metronidazole group and the control group. On the other hand, it is possible that patients in the control group were sicker in some way that we did not capture. However, we believe that this is unlikely, because we included variables that have previously been associated with recurrence of CD. Interestingly, we found that previous smokers but not current smokers had increased odds of disease recurrence. This may be due to the underpowered size of the study population, with only two active smokers in each group, or the differential effects of these factors on disease recurrence. In addition, some studies have suggested that recurrence rates may be >30% in patients treated with metronidazole, which means our study may have been underpowered to detect a difference in recurrence rates.^9,10 Nevertheless, the magnitude of the overall effect of low-dose metronidazole did demonstrate a strong statistical significance, and we remain confident in our primary findings. Moving forward, there is still a need for a prospective double-blind study to confirm the efficacy of low-dose metronidazole as a preventive strategy in post-operative CD surgery. While we looked at the rate of recurrence within the first year after operation, it is unclear whether the benefits of low-dose metronidazole extend past this time point. Furthermore, additional details about rates of histologic and clinical recurrence are of interest.

In conclusion, we showed that a lower dose of metronidazole [250 mg three times per day] reduced endoscopic post-operative recurrence of CD and was well tolerated. This intervention should be considered as a therapy option following ileocecectomy.

Glossary

Abbreviations:

AZA

azathioprine

CD

Crohn’s disease

LD

low dose; three times per day

TNF

tumor necrosis factor