Background

Peritoneal dialysis (PD) is a widely accepted modality in renal replacement therapy for end-stage renal disease. Management of PD catheter-associated infections is essential for the successful performance and continuation of PD therapy. To reduce infectious complications, such as exit-site and tunnel infections, of the PD catheter, Moncrief and Popovich developed a two-step implantation technique in 1993 [1]. This is a popular method of initiating the PD procedure in Japan [2]. First, during PD catheter implantation, the distal end of the catheter with the cuff is buried under the skin. Second, the distal end of the PD catheter is exteriorized on PD initiation. This method is believed to reduce complications due to infection and dialysis fluid leakage on PD initiation. When this method was used in PD patients, 10.7% of cases lacked patency on exteriorization because of a fibrin plug, kinking, or omental wrap; 1.6% developed incision-site and tunnel infections; 1.6% complained of exit-site leakage; and 0.8% experienced peritonitis [3]. Bowel perforation by the PD catheter is a serious complication that has been reported to occur in 1.3 to 1.6% of patients during PD [4].

Herein, we present a case of an unusual interaction between the PD catheter and the sigmoid colon through the side hole of the catheter, possibly caused by mechanical compression of the catheter implanted using the Moncrief–Popovich technique before PD initiation.

Case presentation

A 57-year-old woman with autosomal-dominant polycystic kidney disease (ADPKD) was treated at our hospital. As her renal function gradually declined, under laparoscopic guidance, a standard double-cuff, swan neck, straight PD catheter (JB-5(A), Hayashidera, Japan) was embedded using the Moncrief–Popovich technique. Nine months after the implantation, she was admitted to our hospital to initiate continuous ambulatory peritoneal dialysis (CAPD) therapy. When the distal end of the PD catheter was exteriorized, a brown fluid mixed with fluffy material was found in the catheter. She also complained of a small amount of watery diarrhea and abdominal pain around her pubis, and her body temperature increased to 38.3 °C after the exteriorization. Computed tomography (CT) without a contrast medium was performed (Fig. 1a), assuming that bowel perforation, bacterial peritonitis, and mechanical stress of the PD catheter occurred due to the exteriorization. This imaging showed that there were no abnormalities in her bowel and the tip of PD catheter was intrapelvic. She was started on PD therapy after sulbactam/ampicillin administration for 4 days. Immediately after 1 L of PD fluid flowed into her abdominal cavity, she complained of watery diarrhea. Thus, we decided to perform CT peritoneography with 1 L dialysate (Dianeal N PD-2, Baxter, Tokyo) containing 20 mL Omnipaque 300 (Daiichi-Sankyo, Tokyo) [5,6,7]. This diagnostic imaging technique revealed an outflow of the contrast medium into the luminal side of the sigmoid colon through the PD catheter’s side hole, suggesting an interaction between the PD catheter and the colon (Fig. 1b, c). An exploratory laparotomy conducted the same day revealed that the PD catheter had not penetrated the colon. Instead, the lateral side of the PD catheter (5 cm from the tip) had adhered to the serosal surface of the sigmoid colon, and a small orifice was observed at the corresponding location of the side hole of the catheter (Fig. 2b). This lesion was entirely surrounded by fibrous tissue, which prevented leakage of the bowel contents (Fig. 2a). Her small intestine was also attached to the PD catheter via a serosal erosion (Fig. 2c). These findings strongly suggested that this direct interaction of the PD catheter with the colon arose as a result of persistent mechanical compression of the PD catheter at one location on the colon surface accompanied by protective reactions during the period from implantation to externalization of the distal end of the PD catheter. After restoring the colon with a sigmoidectomy and colostomy, the patient was treated with hemodialysis. Seven months later, she underwent closure of the colostomy.

Fig. 1
figure 1

PD catheter and colon interaction diagnosed by CT peritoneography. a The tip of PD catheter (arrow head) was intrapelvic. The enlarged kidney occupied a major portion of the pelvic cavity (CT without a contrast medium, coronal view). b PD fluid containing contrast media injected from the PD catheter is found in the lumen of the sigmoid colon and the rectum (CT peritoneography, scout view). c PD catheter interacting with the sigmoid colon. Contrast media injected through the PD catheter (arrowhead) flows into the luminal side of the sigmoid colon (arrow) through a hole (top: CT peritoneography, horizontal image, bottom: magnification of lesion)

Full size image
Fig. 2
figure 2

Exploratory laparotomy reveals a small orifice in the eroding surface of the sigmoid colon formed where the side hole of the catheter is attached to the colon. a The PD catheter was wrapped in the omentum (arrow), and there was erosion on the sigmoid colon (arrowhead). The lateral side of the PD catheter was adhered to the sigmoid colon 5 cm from the catheter tip. b A small hole with oozing feces was found where the side hole of the catheter was attached to the colon. The lesion was entirely surrounded by fibrous tissues, which prevented leakage of the intraluminal contents. c The PD catheter was also attached to the serosa of the small intestine

Full size image

Discussion and conclusions

In this unusual case, bowel perforation by a PD catheter occurred asymptomatically during PD catheter embedding, which was ultimately diagnosed by CT peritoneography. Table 1 shows the clinical features of 33 case reports detailing delayed bowel perforation by PD catheters [4, 8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31]. The clinical manifestations of delayed bowel perforation by PD catheters were heterogeneous and included symptoms such as peritonitis, watery diarrhea, catheter protrusion from the anus, and feculent eluent from the catheter. In our case, the patient experienced watery diarrhea, abdominal pain, and feculent eluent, which was consistent with the results of previous reports of bowel perforation by PD catheters. Among 24 cases in which the duration from catheter insertion to bowel perforation was known, bowel perforations occurred within 12 months in 12 cases (50%) and 24 months in 19 cases (79%). The sites of bowel perforations ranged from the jejunum to the rectum, with 13 out of 17 colon cases (76%) occurring at the sigmoid colon.

Table 1 Literature review of delayed bowel penetration or erosion caused by peritoneal dialysis (PD) catheters
Full size table

We found two types of perforations reported. The first was erosion caused by the catheter side wall (n = 7), and the second was penetration by the PD catheter tip (n = 14) (Table 1). Interestingly, three ADPKD cases, including our case, were of the erosion type and only one case was the penetration type. We speculated that an enlarged kidney could be a risk factor for bowel erosion by PD catheters because the enlarged kidney occupied a major portion of the pelvic cavity, thereby compressing the PD catheter against the sigmoid colon through which solid stool passes. Further studies on bowel erosion by PD catheters are required.

CT peritoneography is a diagnostic imaging technique used to examine the interaction of the peritoneal cavity with other internal spaces. This technique is quite useful in detecting interactions between the peritoneal and pleural spaces through the diaphragm by infusing PD fluid containing an iodine contrast agent into the abdominal cavity [5,6,7]. In our case, this technique was helpful in detecting the patient’s bowel perforation before an exploratory laparoscopy was performed. Other methods, including exploratory laparotomy, colonoscopy, and contrast fluoroscopy, have been reported to be useful in the diagnosis of bowel perforations [16, 21]. Among them, CT peritoneography is considered superior to conventional methods in terms of the rapid and accurate identification of the perforation site with minimal invasiveness. Markel et al. [26] have reported that CT peritoneography could reveal an interaction between the PD catheter and the colon, which was not observed during laparotomy.

Bowel perforation typically occurs at the time of PD catheter insertion. This complication is uncommon during other procedures [12]. Delayed bowel perforation usually involves a dormant PD catheter [16, 21] and has also been reported during ongoing PD therapy (Table 1). Our case had several symptoms that would indicate a bowel perforation, and no inflammatory symptoms were apparent in the laboratory data from catheter implantation to PD initiation. Feculent eluent was observed in the PD catheter on exteriorization of the distal end when the patient complained of watery diarrhea and abdominal pain, which is consistent with symptoms previously reported after bowel perforation by a PD catheter [14].

An exploratory laparotomy revealed a small orifice on the eroded surface of the sigmoid colon that formed where the side hole of the catheter was attached to the colon. In addition, the PD catheter was wrapped in the omentum, and the erosion of the sigmoid colon was surrounded by fibrous tissues. This protective reaction is believed to require time, indicating that the lateral side of the PD catheter could have eroded the bowel wall, which resulted in a perforation due to the persistent mechanical compression of the PD catheter at one location over a long duration. This continuous pressure causes localized ischemia which eventually leads to the formation of a decubitus erosion, laceration, or frank perforation [32].

The risk factors for bowel perforation by PD catheters have been previously reported and include the use of immunosuppressants, the presence of diverticulitis [12], colonic amyloidosis [17], and a lack of fluid in the peritoneal cavity after cessation of continuous ambulatory PD because of an immobile catheter [16, 21]. A lengthy implantation of the PD catheter may be involved in this complication, although the optimal time interval from peritoneal catheter insertion to PD initiation has not been established yet. In our case, a lack of fluid in the peritoneal cavity and possible diverticulitis might have contributed to the patient’s bowel perforation although there were only a few complications of diverticulum. The patient’s enlarged kidney caused by ADPKD might also have been involved in the compression of the PD catheter into the sigmoid colon. Defecation control should be strengthened, or preoperative abdominal pressure/intra-abdominal pressure (IAP) monitoring should be considered when using Moncrief–Popovich technique with particular ADPKD patient.

In conclusion, a bowel perforation caused by an embedded PD catheter is a rare but possible complication even when using the Moncrief–Popovich technique. Watery diarrhea following a CAPD exchange is a particularly important clinical sign that indicates bowel perforation by a PD catheter. CT peritoneography may be helpful in identifying the bowel perforation site with minimal invasiveness.