Whether Antibiotic Prophylaxis Is Necessary in Peritoneal Dialysis Patients Undergoing Elective Colonoscopy With Postprocedural Peritoneal Lavage

Peritoneal dialysis (PD)−associated peritonitis is a major cause of technique failure and fatality and has been described as the Achilles heel of PD. One of the common causes of infection is transmural migration (enteric route), which is commonly related to constipation, diarrhea, and invasive intraluminal gastrointestinal (GI) procedures. A recent meta-analysis study demonstrated an increased risk of colorectal cancer in patients with end-stage kidney disease (ESKD).¹ Therefore, a recommendation of colorectal cancer screening has been advocated in such patients, beginning at age 50 years.² However, some concerns about the risk of postcolonoscopic peritonitis from mucosal tear or colonic microperforation from the therapeutic procedures (polypectomy or tissue biopsy) during colonoscopy, suggests the need for antibiotic prophylaxis. Thus, the International Society for Peritoneal Dialysis (ISPD) Peritonitis guidelines 2016³ and the 2015 American Society for Gastrointestinal Endoscopy guidelines⁴ recommended antibiotic prophylaxis before colonoscopy. However, the level of evidence was graded as “low” (2C). Nonetheless, the British Society of Gastroenterology⁵ and the European Society of Gastrointestinal Endoscopy⁶ did not mention this issue. The inconsistency of the recommendations results in a wide variety of current clinical practice in the use of antibiotic prophylaxis in PD patients undergoing colonoscopy.⁷ Therefore, an interventional study was conducted to demonstrate the evidence of microorganism translocation and the unnecessity of antibiotic prophylaxis during elective colonoscopy in patients undergoing PD for colorectal cancer screening with a “dry” abdomen condition and postprocedural peritoneal lavage.

Results

A total of 51 PD patients undergoing elective colonoscopy for colorectal cancer screening without clinical suspicion of colonic abnormalities besides positive heme tests were included in this interventional study (Table 1). Polypectomy was performed in 55% of the patients, and 1 tissue biopsy was performed at cecal ulcers. All of the polyps were benign. No intra- or postcolonoscopic complications were demonstrated (Table 1).

Table 1.

Baseline characteristics (N = 51)

Characteristic	Value
Age, yr	62 ± 10
Diabetes	15 (29)
Automated PD	26 (51)
Dialysis vintage, mo	8 (5–28)
Residual renal function, ml/d	400 (85–875)
Exit site classification
- Perfect/good	25 (88)
- Equivocal	5 (10)
- Acute exit site infection	1 (2)
Laboratories
- Hemoglobin, g/dl	10.5 ± 1.9
- Creatinine, mg/dl	8.6 ± 3.0
- Potassium, mEq/l	4.1 ± 0.7
- Albumin, g/dl	3.5 ± 0.4
Total weekly Kt/V	2.1 ± 0.6
Colonoscopy findings
- Normal	19 (37)
- Colonic polyp at different sites	29 (57)
- Diverticulum without diverticulitis	2 (4)
- Clean-based ulcer at terminal ileum and cecum	1 (2)
Colonoscopy procedure
- None	22 (43)
- Polypectomy	28 (55)
- Tissue biopsy	1 (2)
Colonoscopic complication (e.g., peritonitis)	0 (0)

CAPD, continuous ambulatory peritoneal dialysis; CCPD, continuous cyclic peritoneal dialysis; NIPD, nightly intermittent peritoneal dialysis; PD, peritoneal dialysis.

Data are presented as mean ± SD, median (interquartile range), or n (%) of patients.

A total of 102 PD effluent (PDE) cultures from 51 colonoscopies were obtained. No significant leukocytes (>100 cells per high-power field) were found in any PDE specimens, including those at 1 week after the procedure. All precolonoscopy PDE and blood culture results were negative. Two 2-hour postcolonoscopy PDE cultures from 2 nondiabetes patients undergoing continuous ambulatory PD (CAPD) were positive for coagulase-negative Staphylococcus (CoNS) and Corynebacterium spp. None of the participants had clinical signs or symptoms of peritonitis. One 2-hour postcolonoscopy PDE culture of the patient with ulcers at the terminal ileum and ileocecal (IC) valve was positive for Mycobacterium tuberculosis, whereas the precolonoscopy PDE culture for Mycobacterium was negative. The pathology of the terminal ileum, as well as ileocecal valve ulcers, was also compatible with M. tuberculosis−induced lesion (Table 2).

Table 2.

Characteristics and colonoscopic details of participants with positive postprocedural PDE and blood culture

Characteristic	Case 1	Case 2	Case 3	Case 4	Case 5
Age, yr	65	61	55	59	55
Sex	Female	Female	Male	Male	Female
Diabetes mellitus	No	No	Yes	Yes	No
Mode of PD	CAPD	CAPD	CCPD	NIPD	CAPD
Colonoscopy findings	Colonic polyp	0.3-cm sessile polyp at sigmoid colon	0.5-cm round clean-based ulcer at ileocecal value, 1.0-cm round clean-based ulcer at terminal ileum, and 0.3-cm sessile polyp at cecum	Normal	Normal
Procedure performed	Polypectomy	Polypectomy	Sampling lesion biopsy at ulcer and polypectomy	None	None
PDE culture
- Before colonoscopy	NG	NG	NG	NG	NG
- 2 h after colonoscopy	CoNS	Corynebacterium spp.	M. tuberculosis	NG	NG
Hemoculture
- Before colonoscopy	NG	NG	NG	NG	NG
- During cecal intubation	NG	NG	NG	NG	NG
- Immediately after colonoscopy	NG	NG	NG	S. anviculavis	NG
- 2 h after colonoscopy	NG	NG	S. hominis	NG	Microbacterium spp.
Clinical peritonitis (abdominal pain, fever, cloudy PDE)	No	No	No	No	No

CoNS, coagulase-negative Staphylococcus; NG, no growth; PD, peritoneal dialysis; PDE, peritoneal dialysis effluent.

Among 306 blood-culture specimens, there were 3 specimens from different participants having positive results for bacterial organisms, including Staphylococcus hominis, Staphylococcus anviculavis, and Microbacterium. One of these was positive in the blood specimens obtained during the procedure, and the other 2 were from blood specimens collected 2 hours after the procedure. Of note, 1 of the participants with positive blood specimen (S. hominis) taken 2 hours postprocedure was the case with an incidental finding of colonic tuberculosis and positive PDE specimen with M. tuberculosis. None of the participants developed septicemia or septic shock after the procedure. Table 2 shows participants’ characteristics and colonoscopic details of those participants with positive PDE culture and blood culture results.

Discussion

This interventional study demonstrated the existence of transmural migration of microorganisms during and immediately after elective colonoscopy in stable PD patients with “dry abdomen” condition and postprocedural peritoneal lavage (Figure 1). However, none of the patients with the presence of microbial translocation developed peritonitis or septicemia within 1 week postprocedure, despite the lack of antibiotic prophylaxis.

Figure 1.

Protocol for colonoscopy preparation, peritoneal dialysis fluid, and blood sampling. H/C, hemoculture; PDE, peritoneal dialysis effluent; PDF, peritoneal dialysis fluid.

Although the incidences of postcolonoscopy bacterial peritonitis among patients without preprocedure administration of antibiotic prophylaxis are reported to range from 6.3% to 8.5% depending on patient condition and concomitant therapeutic procedure during colonoscopy, none of the studies actively performed a postprocedural PDE culture from every single studied case.^8,9,S1 All of these studies collected PDE cultures only when the patients had clinical features suggesting peritonitis. It is intriguing that there were no peritonitis/septicemia episodes (0%) in the present study, although 3 patients (5.9%) and 2 patients (3.9%) had transient bacterascites and bacteremia, respectively. The discrepancy in an incidence of postprocedural peritonitis might be explained by differences in indications of colonoscopy. The majority of the enrolled patients in the previous studies^8,9 had clinical suspicion of bowel pathology, whereas all of our participants were asymptomatic and underwent elective colonoscopy for colorectal cancer screening protocol without clinical suspicion of bowel pathology. All peritonitis patients in the Hong Kong study had clinical suspicions of bowel pathology, including iron deficiency anemia (n = 3), bleeding per rectum (n = 1), and bloody effluent (n = 1).⁹

In addition, our study performed postprocedural peritoneal lavage and was strict about a “dry abdomen” condition as mentioned by Chaudary et al.S2 The U.S. group recently recommended keeping the abdomen dry for hours to days after colonoscopy as a means to improve host defenses.S2 However, the present study showed that emptying the abdomen for a short period during colonoscopy might be enough. Peritoneal lavage after the procedure aiming to eliminate pathogen contamination and inflammatory cytokines in the peritoneal cavity is widely used in both abdominal laparotomy and laparoscopy, with favorable outcomes in the reduction of postprocedural peritonitis and peritonitis-related death.S3^,S4 However, the use of postprocedural peritoneal lavage has not been encouraged in PD patients. Our finding might support the potential benefit of peritoneal lavage in removing the spilled-out pathogen in the peritoneal cavity and preventing postprocedural peritonitis, although transmural migration of the pathogens during colonoscopy was documented.

Based on the findings of the present study, the concept of giving antibiotic prophylaxis in every PD patient undergoing colonoscopy was not firmly supported, at least in the patients undergoing colorectal cancer surveillance protocol with “dry abdomen” and postprocedural peritoneal lavage. Our result supports the finding from Al-Hwiesh et al. in which ceftazidime prophylaxis before colonoscopy was not beneficial in the reduction of peritonitis episodes after the procedure (3 in 46 patients [6.5%] vs. 4 in 47 patients [8.5%], respectively, P = 0.279) in the prospective randomized controlled study.⁹

The strength of the present work is that the study collected both pre- and post-colonoscopy PDE and blood cultures from every patient. Of more importance, we could demonstrate the real evidence of bacterial translocation causing bacterascites after the intraluminal gastrointestinal procedure and also could demonstrate that all of the patients with bacterascites were asymptomatic and this bacterascites could resolve after peritoneal lavage. However, the results in the present study may not be applicable to all clinical settings. First, our study excluded patients with higher risks of peritonitis, including immunocompromised patients (i.e., those taking glucocorticoid and immunosuppressive drugs), a recent history of peritonitis, active GI symptoms, and advanced GI cancer. Second, our study had relatively few outcomes, and larger-scale multicenter prospective controlled trials are essential. The effects of postprocedural peritoneal lavage on peritonitis prevention require confirmation in future studies.

In conclusion, the risk of postprocedural peritonitis is low in asymptomatic PD patients undergoing colonoscopy for colorectal cancer screening with “dry abdomen” during the procedure and immediately postprocedural peritoneal lavage, regardless of the presence of transient bacterascites/bacteremia. Therefore, the decision to routinely prescribe antibiotic prophylaxis in these patients should be individualized.

Disclosure

All the authors declared no competing interests.