Digestive Diseases and Sciences

, Volume 58, Issue 6, pp 1676–1682

The Effect of Rifaximin on Gut Flora and Staphylococcus Resistance


  • Mi-Sung Kim
    • Division of GastroenterologyCheongju St. Mary’s Hospital
  • Walter Morales
    • GI Motility Program, Division of GastroenterologyCedars-Sinai Medical Center
  • Andres Ardila Hani
    • GI Motility Program, Division of GastroenterologyCedars-Sinai Medical Center
  • Sharon Kim
    • GI Motility Program, Division of GastroenterologyCedars-Sinai Medical Center
  • Gene Kim
    • GI Motility Program, Division of GastroenterologyCedars-Sinai Medical Center
  • Stacy Weitsman
    • GI Motility Program, Division of GastroenterologyCedars-Sinai Medical Center
  • Christopher Chang
    • GI Motility Program, Division of GastroenterologyCedars-Sinai Medical Center
    • GI Motility Program, Division of GastroenterologyCedars-Sinai Medical Center
Original Article

DOI: 10.1007/s10620-013-2675-0

Cite this article as:
Kim, M., Morales, W., Hani, A.A. et al. Dig Dis Sci (2013) 58: 1676. doi:10.1007/s10620-013-2675-0



Rifaximin is a non-absorbed antibiotic relative of rifampicin. The location of effect and staphylococcal resistance are two recent potential concerns with rifaximin. In this study we evaluate the location of effect of rifaximin as well as the development of staphylococcal rifampicin resistance.


Rats were divided into three groups. Group 1 gavaged for 10 days with PBS, group 2 gavaged with rifaximin for 10 days, and group 3 gavaged with rifaximin for 10 days and housed for 30 days. In each group, stool was collected daily for quantitative culture of Staphylococcus spp. and coliforms. After euthanasia luminal bacterial counts were determined at multiple gut locations by qPCR. Rifampicin susceptibility was tested on Staphylococcus pre and post rifaximin.


At baseline, rats had a median of 2.90 × 106 cfu/ml Staphylococcus spp. in stool. After 10 days of rifaximin, this dropped to 1.20 × 105 cfu/ml (P < 0.01). With coliform counts, rats had a median of 1.86 × 104 cfu/ml at baseline which dropped to 2.2 × 103 cfu/ml (P < 0.01) after rifaximin. After cessation of rifaximin, coliform counts recovered within 3 days. When examining the total bacterial counts by qPCR, rifaximin reduced small bowel bacterial levels, but not colon. This reduction was sustained for 30 days. No colonies of Staphylococcus became resistant and only one colony was intermediate. The mean inhibitory concentration for rifampicin was not different before and after rifaximin.


Staphylococcal spp. fail to demonstrate resistance to rifampicin after rifaximin. The transient reductions in stool coliform counts recover while rifaximin appears to produce durable reductions in duodenal bacteria.




Rifaximin is a semisynthetic derivative of rifamycin [1] and was first approved in Italy in 1987. The pyridoimidazole ring of rifaximin is not inactivated by gastric fluids and the molecule is not water soluble, and thus the antibiotic is poorly absorbed from the gastrointestinal tract. Approximately 97 % of rifaximin is excreted into the feces as unchanged drug with a bioavailability of <0.4 % in the blood following oral administration [2, 3]. Rifaximin has broad spectrum activity similar to that of rifampicin against aerobic and anaerobic, gram positive and gram negative microorganisms [4].

Rifaximin is currently approved by the US Food and Drug Administration for the treatment of uncomplicated traveler’s diarrhea and hepatic encephalopathy [5]. However, rifaximin has recently gained attention in the treatment of irritable bowel syndrome (IBS) [6]. This is based on a mechanism that a subset of IBS subjects has small intestinal bacterial overgrowth. This mechanism has now been substantiated by two culture studies of small intestinal aspirates demonstrating elevated coliform levels in IBS compared to healthy controls [7, 8]. These elevations were seen even when compared to non-IBS subjects with gastrointestinal complaints. Up to 60 % of diarrhea predominant subjects have small intestinal bacterial overgrowth based on duodenal culture [9]. However, despite the efficacy of rifaximin in IBS, three issues have been raised in the treatment of IBS. The first is whether retreatment with rifaximin is effective. This has now been demonstrated in a retrospective study [10], in addition to an ongoing prospective study examining this question. The other two issues were the location of effect of rifaximin (small bowel or colon) and whether Staphylococcal spp. could develop resistance to rifampicin after rifaximin [5, 9]. The development of bacterial resistance to rifaximin appears to occur with a low frequency in vitro [11, 12]. However, concern has been raised that rifaximin might produce Staphylococcus spp. resistance to rifampicin in vivo [9]. In fact, some investigators report the development of bacterial resistance in Staphylococci during exposure to rifaximin in vitro and spontaneous rifaximin resistance, but its clinical importance remains to be fully defined [13, 14].

In this study we aim to evaluate the location of antimicrobial effect of rifaximin as well as examine the development of stool staphylococcal rifampicin resistance from rifaximin use in an in vivo animal model.


Baseline Measurements

All animal procedures were approved by the Institutional IACUC committee. Adult male Sprague–Dawley rats were acquired and quarantined for 5 days. During this time, chow was standardized to ensure all animals had identical feeding type.

After this equilibration period, fresh stool was collected from each rat (n = 30) by anal stimulation. The stool collected on day 0 was homogenized and plated by serial dilution with 1XPBS (phosphate buffered saline) on blood agar with phenylethyl alcohol (PEA; BD Diagnostics, Franklin Lakes, NJ, USA) to select for the presence of the Staphylococcal spp. Similarly, homogenized stool was serially plated on MacConkey agar to select for and determine coliform counts in the stool. All plates were incubated for 24 h at 37 °C. Based on serial stool dilution and colony counts, the baseline stool levels of Staphylococcus spp. and coliforms were determined.

Rifaximin Gavage and Stool Coliform Pattern

Rats were divided into three treatment groups and gavaged daily for 10 days with or without high dose rifaximin. Group 1 was gavaged with PBS alone, while group 2 was gavaged daily with rifaximin 200 mg in PBS, and then these two groups were euthanized. Group 3 was gavaged daily with rifaximin 200 mg in PBS and housed for 30 days following completion of rifaximin before euthanasia. The dose of rifaximin (200 mg/day) was based on similarity to human maximal doses used in clinical trial which equates to approximately 12–14 mg/g of stool in both humans and in these rats. During the rifaximin treatment, fresh stool was collected daily as described above. Again, stool was homogenized and plated by serial dilution on MacConkey and PEA agar to determine coliform and Staphylococcus spp. counts during the treatment. After the completion of 10 days of treatment, ten rats were euthanized for bacterial quantitation by culture and qPCR throughout the intestinal tract. The remaining ten rats were followed for 30 days after completing rifaximin. These rats had interval stool culture for quantitation of Staphylococcus spp. and coliforms and the determination of recovery time of stool flora if any.

Luminal Quantitation of Bacteria

For the examination of luminal bacterial counts, rats were euthanized and dissected after 10 days of PBS (n = 10; control), after 10 days of rifaximin (n = 10) and 30 days post rifaximin (n = 10). During the dissection, pre-specified segments of duodenum, jejunum, ileum, cecum and left colon were ligated and resected as previously described [15]. Luminal contents were extracted from each segment. For culture of coliforms, serial dilutions were again prepared and plated on MacConkey agar then incubated and counted. In addition, luminal contents were also used to determine total bacterial counts by qPCR.

To quantify bacteria in the luminal contents, qPCR was also used. DNA was isolated from luminal contents of the duodenal, jejunal, ileal, cecal and left colon samples using QIAamp Stool DNA Extraction kits (Qiagen, Valencia, CA, USA). Bacterial universal primers [16] were used to amplify the 16S rRNA gene from DNA using the CFX96™ Real-Time PCR Detection System (Bio-Rad Laboratories, Hercules, CA, USA) and optical grade 96-well plates. Samples were run in duplicate. The PCR reaction was performed in a total volume of 20 μl using the iQ™ SYBR GREEN Supermix (Bio-Rad Laboratories), containing 300 nM each of the universal forward and reverse primers. The reaction conditions were set at 95 °C for 3 min followed by 40 cycles at 95 °C for 10 s, 55 °C for 10 s and 72 °C for 30 s. Data analysis made use of CFX Manager software supplied by Bio-Rad. To generate standard curves, the Ct values were analyzed from ten-fold dilutions of lysed Escherichia coli cultures. E. coli from a number of laboratory strains were pooled and grown in LB media (Sigma Aldrich, St. Louis, MO, USA).

Rifampicin Susceptibility Testing

Rifampicin susceptibility was tested in Staphylococcus spp. isolated from stool of rats before and after rifaximin treatment for 10 days. From PEA plates, 30 random recognizable Staphylococcus colonies were picked for baseline cultures, suspended in PBS and spread on PEA agar plates to create a lawn. A rifampicin E-test strip (bioMérieux, Inc., Durham, NC, USA) was added to each lawn to detect rifampicin resistance. The range of rifampicin mean inhibitory concentrations (MICs) detected by E-test was 0.002–32 μg/ml. Based on manufacturers instruction, growth at <2 μg/ml was considered sensitive, growth at 2–4 μg/ml was considered intermediate resistance to rifampicin, and growth at >4 μg/ml was considered resistant to rifampicin. After 10 days of rifaximin, stool was again plated on PEA and another 30 staphylococcal colonies were picked for producing a lawn to analyze rifampicin resistance using the applied E-test strip. Since no significant Staphylococcus spp. colonies remained on day 30 post rifaximin treatment, sensitivity testing could not be assessed.

Data Analysis

To compare stool colony counts of bacteria or qPCR between groups, a Mann–Whitney U test was used and data were expressed as median due to data being non-normal. When comparing colony counts before and after rifaximin, a Wilcoxon rank-sum test for matched pairs was used. Significance was noted as a P value <0.05. In comparing trends in the counts from control to rifaximin 10 and 30 days post rifaximin, data were log transformed to normalize the data and compared by Kruskal–Wallis test.


Effect of Rifaximin on Stool Staphylococcus spp.

Overall, Staphylococcal species represented a significant component of stool flora as evidenced by the coliform counts below. At baseline, rats had a median of 2.90 × 106 cfu/ml (range, 0–1.96 × 106 cfu/ml) combined total Staphylococcus spp. (total of S. aureus and coag-Staphylococcus). Eighteen of 20 rats had detectable S. aureus in stool at baseline but all rats had coagulase negative Staphylococcus. After 10 days of rifaximin, the median total count dropped significantly to 1.20 × 105 cfu/ml (range, 0–8.6 × 105 cfu/ml; P < 0.01 by Wilcoxon rank-sum test for matched pairs; Fig. 1). On day 10 of rifaximin, five rats had detectable S. aureus in stool but at a low level of 2 × 103 cfu/ml (median). After 30 days there were insufficient Staphylococcus spp. colonies in any of the rats to analyze (Fig. 2).
Fig. 1

Staphylococcus spp. median colony counts at baseline and after 10-day rifaximin treatment

Fig. 2

Staphylococcus spp. colony counts by day after 200 mg rifaximin treatment

Effect of Rifaximin on Stool Coliforms

A reduction in stool coliforms was seen with rifaximin compared to baseline. At baseline, rats had a median of 1.86 × 104 cfu/ml. During the 10 days of rifaximin treatment there was a modest but significant reduction in coliform counts, down to 2.2 × 103 cfu/ml by day 10 of high dose rifaximin (P < 0.01, Wilcoxon rank-sum test for matched pairs; Fig. 3).
Fig. 3

Median change in stool coliform counts during and after rifaximin treatment

After the cessation of rifaximin, coliform counts were seen to recover rapidly. Within 3 days post rifaximin, the coliform count recovered (Fig. 3). This was due to PBS contamination at day 40 (Fig. 3). Figure 4 was redrawn to more closely demonstrate the reduction in counts as well as recovery to baseline during treatment and in the 7 days after cessation of rifaximin.
Fig. 4

Median change in stool coliform counts during and after rifaximin treatment (only 7 days after rifaximin)

Effects of Rifaximin on Luminal Bacteria

The luminal contents were then examined. Coliform counts in non-treated control rats were seen to increase in number with descent through the bowel with highest counts noted in the cecum and left colon. The duodenum was virtually free of coliforms in these rats similar to humans. After treatment with rifaximin there was no significant change in coliform counts by culture (Fig. 5).
Fig. 5

Total coliform counts by segment comparing PBS and two delivery systems of rifaximin. Wide variation in individual counts were seen and no significant difference was seen between groups (Mann–Whitney U test)

When examining total bacterial counts by qPCR, rifaximin was seen to reduce bacterial levels in the small bowel but not colon. When comparing control rats to all rats receiving rifaximin (n = 20; Fig. 6), there was a significant reduction in duodenal bacterial counts noted. While jejunal counts appeared to diminish as well, this did not reach statistical significance, even at n = 20. Beyond the jejunum, no differences were seen in bacterial counts in the ileum, cecum or left colon (Fig. 7).
Fig. 6

Total bacteria counts by qPCR in controls and any rifaximin-treated rats in small intestine

Fig. 7

Total bacteria counts by qPCR in controls and any rifaximin-treated rats in colon

Interestingly, rifaximin had a durable effect on total duodenal bacterial counts. A reduction in duodenal bacteria counts was observed after 10 days of rifaximin that was sustained 30 days after the completion of rifaximin, although this did not quite reach statistical significance (Fig. 8; P = 0.08 by Kruskall-Wallis test). A similar but non-significant trend was observed in the jejunum. There appeared to be no effect on colonic bacterial counts (Fig. 9).
Fig. 8

Total bacteria counts by qPCR in controls, immediately after rifaximin and 30 days after completion of rifaximin in small intestine in rats. *P = 0.08 by Kruskal–Wallis; after log transformation other comparisons were not significant

Fig. 9

Total bacteria counts by qPCR in controls, immediately after rifaximin and 30 days after completion of rifaximin in colon in rats. No significant differences were seen by Kruskal–Wallis test after log transformation

Rifampicin Susceptibility Test

Before rifaximin and at day 10 of rifaximin, staphylococcal species were picked and plated for testing rifampicin resistance. At baseline (before rifaximin), two colonies of Staphylococcus spp. were resistant and five colonies were intermediate to rifampicin (Table 1). However, after 10 days of rifaximin gavage, no colonies were resistant and only one was intermediate. In addition, the MIC for rifampicin was not different before and after rifaximin treatment. The MIC for rifampicin was 1.1 ± 1.6 μg/ml for baseline and 0.91 ± 0.52 μg/ml after 10 days of high dose rifaximin (P = 0.63).
Table 1

Summary of Staphylococcal resistance before and after rifaximin


Sensitive (<2 μg/ml)

Intermediate (2–4 μg/ml)

Resistant (>4 μg/ml)


22 (75.8 %)

5 (17.2 %)

2 (6.9 %)

Day 10 rifaximin

29 (96.6 %)

1 (3.3 %)


Values given as n (%)

aOnly 29/30 colonies formed a lawn for testing resistance. There was no statistically significant difference between groups


In this study, we demonstrate that a 10-day course of rifaximin has a modest but significant effect on stool coliforms and Staphylococcal spp. However, these counts recover within 3 days of cessation of therapy. Interestingly, rifaximin in vivo appears to reduce the number of resistant Staphylococcal spp. with no effect on MIC of rifampicin towards these bacteria. Using this in vivo model, the greatest effect of rifaximin appears to be on duodenal bacteria with a milder effect on colonic flora. This effect appears to last for at least 30 days after completion of the antibiotic.

Rifaximin’s non-absorbable property in intestine after oral administration and broad spectrum antibiotic activity makes this an ideal drug for intestinal disease. In fact, rifaximin’s clinical use has increasingly broadened, including Clostridium difficile-associated diarrhea, diverticular disease, hepatic encephalopathy, inflammatory bowel disease, and others [1723]. Rifampicin is an antimicrobial belonging to the rifamycin group, same as rifaximin [24], and is used to treat invasive infections such as methicillin-resistant staphylococcal infection (MRSA), mycobacterial infection, and meningococcal infection. It is well known that rifampicin resistance develops quickly during treatment of staphylococcal infection, so monotherapy should not be used to treat these infections [25, 26]. Rifampicin has a close structural and antibacterial spectrum with rifaximin. This leads to concerns about the possibility of cross-resistance among bacteria between these two substances.

The antibacterial activity of rifamycin-derived drugs such as rifaximin and rifampin is produced by binding the β-subunit of bacterial DNA-dependent RNA polymerase, inhibiting the initiation of chain formation in RNA synthesis [24]. Clinically, rifamycin-derived drugs are often used in combination with other antimicrobials because of the propensity to select for resistant mutants when used as a single agent [27]. As demonstrated in in vitro and in vivo studies, the development of rifaximin resistance is primarily caused by mutation in the rpoB gene, which encodes the target of rifamycins, the β subunit of RNA polymerase [28, 29]. This mechanism is different from the plasmid-mediated resistance of commonly acquired bacteria to aminoglycoside antibiotics such as neomycin [30]. Although the probability of spontaneous chromosomal point mutation with resistance to rifaximin is possible, given the large number of bacteria present, this type of resistance is less common than plasmid-based resistance.

There are some small early studies examining the development of resistance with rifaximin. Most of them demonstrated that the prevalence of bacterial resistance to rifaximin was none or low frequency [11, 12]. The resistant strains detected during treatment appear to be unstable and unable to persistently colonize the intestinal tract [12]. Previously, it was thought that there was no evidence of cross-resistance between rifampicin and rifaximin [5]. Some studies have demonstrated that there is no induction of resistance between rifampicin and rifaximin in Mycobacterium tuberculosis [3133]. But recently a study demonstrated rifampicin-resistant human skin staphylococci emerging after oral intake of rifaximin that persists after discontinuation of antibiotic. They assumed that rifaximin-resistance developed due to perianal skin staphylococci contact with rifaximin after rifaximin-containing stool was passed [9].

In this study, we found that rifaximin gavage for 10 days reduces the total staphylococcal colony counts in stool. Although Staphylococcus spp. persist after high dose rifaximin, none of the staphylococcus species demonstrate resistance to rifampicin. Interestingly, there were resistant colonies to rifampicin even before the administration of rifaximin. These were not seen after rifaximin, suggesting they may have been reduced by rifaximin. Thus rifaximin does not appear to select for rifampicin resistance in Staphylococci spp. after a 10-day course.

While the full effect of rifaximin on gut flora remains limited, existing data suggests that rifaximin has little effect on the normal gastrointestinal flora [12, 30]. In a study with healthy volunteers, after short-term (5 days) rifaximin treatment, the observed change in bowel flora returned to baseline level within 1–2 weeks. In another study of ulcerative colitis patients, rifaximin produced an initial decrease in the gastrointestinal flora, followed by normalization after 1 month. In our study, fresh collected stool from rats produced similar results. In the case of coliforms, there was a modest reduction during treatment followed by recovery of counts within 3 days of cessation of therapy.

On examining luminal bacteria in this study, coliforms were low in the small intestine. In the duodenum virtually no coliforms were seen. However, when total bacterial were examined, large numbers were evident. Based on total bacterial counts by qPCR, it appears that rifaximin reduces bacteria mostly in the small intestine. In fact, there was a nearly 70 % reduction in bacteria in the duodenum. In the study of IBS, the hypothesis is that small intestinal bacterial overgrowth may be important in this condition. Bacterial overgrowth is caused by elevated coliforms. Since there were almost no coliforms in the duodenum, where it appears rifaximin is most active, it is not surprising that we could not observe a reduction in coliforms from this low baseline. However, in a model of bacterial overgrowth, this may be important to re-examine.

An interesting observation in this study is that rifaximin produces a durable reduction in duodenal bacterial counts. In fact, 30 days after cessation of rifaximin, rats still demonstrated a reduction in total bacterial counts. This may support findings in clinical trials. In the study of IBS, a 14-day course of rifaximin appears to have beneficial effects long after the cessation of rifaximin [6]. Since in IBS, the hypothesis is a small bowel bacterial derangement, a sustained reduction of bacteria after initial treatment would be beneficial.

In conclusion, this in vivo study suggests that while Staphylococcal spp. are seen in the colon of rats, these organisms fail to demonstrate increasing resistance to rifampicin after a 10-day course of rifaximin. While producing transient reductions in stool coliform counts that quickly recover with 3 days of cessation of therapy, rifaximin appears to produce a durable reduction in duodenal bacteria counts lasting at least 30 days. With the hypothesis that IBS is caused by small intestinal bacterial overgrowth and rifaximin produces a durable clinical response in these subjects [6], this may be part of the mechanism of action. Further studies would include examining the effect of rifaximin in an abnormal model of flora such as small intestinal bacterial overgrowth where coliforms are elevated.


This investigator-initiated study was supported by a grant from Salix Pharmaceuticals. In addition, this work was further supported by a grant from the Beatrice and Samuel A. Seaver Foundation.

Conflict of interest

Cedars-Sinai has a licensing agreement with Salix Pharmaceuticals. In additions, Drs. Pimentel and Chang are consultants for Salix Pharmaceuticals.

Copyright information

© Springer Science+Business Media New York 2013