The Association of Intra-Abdominal Adhesions with Peritoneal Dialysis Catheter-Related Complications.

This study included patients aged 18 years or older who underwent laparoscopic PD catheter insertion between November 1, 2011, and November 1, 2020, at one of eight institutions in Canada and the United States (list in Supplemental Material ). Only one catheter was used per patient. Embedded catheters were included. Patients were followed from the date of PD insertion (time zero) for a minimum of 6 months of potential follow-up 18 ; if the PD was terminated before 6 months due to a complication, they were included. Patients were followed until death, transplant, recovery of kidney function, termination of PD, loss to follow-up, or the end of the first year of PD. All data were collected using the web-based North American PD Catheter Registry. Baseline and outcome data were entered prospectively by study personnel every 90 days and reviewed by investigators (M.J. Oliver and R.R. Quinn). Patients were followed for catheter-related procedures and changes in treatment status ( i.e. , initiation of dialysis training, initiation of dialysis, changes in dialysis modality, death, transplantation, recovery of kidney function, or transfer out of the program). Procedure notes and changes in therapy were reviewed to determine whether they were PD-related complications (see Supplemental Material for data forms). Baseline variables collected at PD catheter insertion included age, sex, body mass index (BMI), presence of diabetes, cardiac disease, vascular disease, history of cancer, respiratory disease, liver disease, abdominal conditions, and a history of hernias. Prior abdominal surgeries were counted and classified by type and location. Surgery was defined as upper or lower on the basis of whether the incision scar was above or below the level of the umbilicus. The timing of PD catheter insertion (predialysis versus on hemodialysis; eGFR) was noted. Details of the surgical technique were recorded, including the anonymized operator ID, insertion point into the abdomen, location of the exit site, position of the deep cuff, and use of ancillary procedures, including adhesiolysis. The make and model of the PD catheter were recorded. Adhesions were recorded based on a review of operative notes by coordinators. All descriptions of adhesions were reviewed by investigators, and the location was arbitrated by an experienced surgeon (J.H. Crabtree), if required. Surgeons did not enter data directly into study forms. Adhesions were described by location using the umbilicus as a landmark to divide them into lower abdominal/pelvic or other. The other category included upper locations and other descriptions ( e.g. , right-sided, anterior abdominal wall). The primary outcome was the composite of never starting PD, termination of PD, or the need for an invasive procedure (radiologic manipulation, laparoscopic revision, or replacement) caused by flow restriction or abdominal pain. This outcome aligns with the definition of patency in guidelines. 6 Patients were observed for the primary outcome from PD insertion to the end of 12 months of PD therapy. Procedures were recorded prospectively. Individuals who never started PD or terminated PD may have also undergone an invasive procedure to address complications. Secondary outcomes were abnormalities found during PD catheter-related procedures. All procedure notes were reviewed retrospectively by investigators (M.J. Oliver and R.R. Quinn) to ensure abnormalities, and additional techniques were captured. Abnormalities included PD catheter tip migration; lumen occlusion; presence of adhesions (new or recurrent); wrapping of the PD catheter tip with the omentum, small bowel, or large bowel; fallopian tube; and the presence of hernias. Tip migration was defined when the tip was documented to have moved out of the pelvis or when the tip was in the pelvis, but it was repositioned deeper into the pelvis. The techniques included repositioning, tip mobilization, manual clearance of fibrin the catheter lumen, adhesiolysis, omentopexy, lower abdominal suture sling, and hernia repair. Categorical and continuous variables were summarized using counts (percentages) and medians (with interquartile ranges [IQRs]). Missing BMIs were input as median values according to biologic sex. Risk factors for adhesion were estimated using logistic models. Baseline relationships between location of surgery, adhesions, and adhesiolysis were compared using chi-squared tests. In the primary analysis, the primary outcome was summarized using cumulative incidence curves from the insertion to the first 12 months of PD to provide an accurate estimate of risk, accounting for competing risks (death, transplant, recovery of kidney function, PD never starting, or terminations for reasons unrelated to the primary outcome). The absolute risk of complications 6 months from insertion was also reported. We used a cause-specific Cox proportional hazards model to examine the association between the presence of adhesions and the primary outcome, censoring for competing events and calculated the unadjusted and adjusted hazard ratios. Covariates were selected based on clinical knowledge and empirical strategies. Age, sex, BMI, and presence of adhesions (categorized as yes or no and then by location [lower or other] and use of adhesiolysis [with or without]) were included in the model. The analysis was stratified by institution to account for potential center effects. Statistical and graphical diagnostics were conducted to check for model fit and violations of the proportional hazard assumption. All analyses were conducted using R version 4.2.2 ( R-project.org ).

One thousand three hundred and six PD catheter insertions from eight programs were enrolled; 383 (29%) were excluded because they did not undergo laparoscopic insertion. Of the remaining 923 individuals, 165 (18%) were excluded because the follow-up was less than 6 months (they were entered into the registry within 6 months of the end of the study and are still being followed), leaving 758 individuals. A total of 201 individuals (27%) had adhesions, whereas 557 (73%) did not (Figure 1 ). The median follow-up was 12 months (IQR, 7–13) for both groups. Of the 758 patients, reasons for exiting the study unrelated to the primary outcome included 13 (2%) deaths, 44 (6%) never started PD, 150 (20%) terminated PD, 3 (0.4%) recovered kidney function, 17 (2%) received a kidney transplant, 3 (0.4%) were transferred, and 503 (66%) reached the end of follow-up. We included 99 embedded catheters; 70 were used during follow-up, with a median start time of 5 months (IQR, 3–9); 15 were still being followed and had not started PD yet, and 14 never started PD (one primary outcome, one death, 12 other reasons). Reasons for exiting the study did not differ significantly by group. Cohort creation describing inclusion and exclusion criteria and sample size. PD, peritoneal dialysis. PD catheters were inserted during predialysis care for 527 (70%) patients, and the median eGFR at insertion was 8 ml/min (IQR, 7–10 ml/min). The insertion point was through the rectus muscle in 82% and midline through the linea alba in 14%, but the latter was predominantly from one center. Omentopexy was performed in 11% of the adhesion group and 8% with no adhesions group, and purse-string sutures around the deep cuff were placed in 58% and 56%, respectively. There were no significant differences in the laparoscopic surgical techniques between the groups. Adhesions were present in 201 individuals, of whom 117 (58%) had lower adhesions and 84 (42%) had adhesions in other locations. Individuals with adhesions were more likely to be older (median age 60 versus 59 years), female (54% versus 33%), and had a higher BMI (28.2 versus 26.7) (Table 1 ). The adjusted odds ratio for adhesion for women was 1.65 (95% confidence interval [CI], 1.12 to 2.41) and for BMI per five units was 1.16 (95% CI, 1.003 to 1.34). Age did not reach a statistical significance (hazard ratio 1.14, 95% CI, 0.998 to 1.31) in the adjusted model. The adhesion and no adhesion groups were similar in comorbidities and laboratory values. Characteristics of patients from Canada and the United States who had laparoscopic peritoneal dialysis catheters placed between November 1, 2011, and November 1, 2020 Percentages are column percentages except for prior surgeries, which are row percentages. BMI, body mass index; BSA, body surface area; IQR, interquartile range; PD, peritoneal dialysis. Missing data included body mass index: 13 (1.7%), hemoglobin: 1 (0.1%), albumin: 26 (3.4%) eGFR: 12 (1.6%), urea: 18 (2.4%), coronary artery disease: 1 (0.1%), cerebrovascular disease: 1 (0.1%), respiratory disease: 5 (0.7%), and liver disease: 2 (0.3%). eGFR based on laboratory reporting at the time of the insertion; reported only for predialysis patients. Includes patients with Pfannenstiel scars. Includes tubal ligation and cyst removal. Other surgeries include small bowel or colon surgeries ( n =8); gastric surgeries ( n =6); laparotomy ( n =4); aortic surgery ( n =2); ventral hernia repair ( n =2), esophageal surgery, prostatectomy, de-roofing of a hepatic cyst; thoracotomy but scar extends into the abdomen, peritoneal drain, hepatectomy, nephrolithotomy, suprapubic catheter, splenectomy, ureteric reimplantation, colpotomy, PEG tube placement, “gyne” surgeries—not specified, surgery for endometriosis. Includes patients with surgical scars noted by the surgeon in the preoperative assessment, but no surgical history was available. There was a significant association between prior abdominal surgery and the presence of adhesions (Table 2 ). Of the 230 (30%) individuals with prior surgery, 55% had adhesions, compared with only 14% without prior surgery. The adjusted odds ratio for adhesions with any surgery compared with no surgery was 8.34 (95% CI, 5.50 to 12.86). Adhesions were present in 58%–61% of individuals with multiple surgeries, but the odds of adhesions were not higher with a greater number of surgeries. The presence of adhesions varied according to the type and location of surgery performed. Common surgeries associated with adhesions were cholecystectomy (62%), cesarean section (55%), appendectomy (64%), hysterectomy (67%), prior PD catheter insertion (34%), and kidney transplant (71%). In the adhesion group, 117 (58%) of 201 were classified as having lower adhesions and 84 (42%) were classified as adhesions in other locations. Of the 96 individuals with lower abdominal surgery (with or without surgery in other locations), 57 (59%) had lower adhesions compared with only 10 (33%) of 30 individuals with only upper abdominal surgery ( P = 0.01) ( Supplemental Table 1 ). Association of prior abdominal surgery and type of surgery on the risk with adhesions Percentages displayed are row percentages. CI, confidence interval; OR, odds ratio. Odds ratios are adjusted for age, sex, and body mass index. The adjusted odds ratio for adhesions for women was 1.65 (1.12 to 2.41) and for body mass index per 5 units was 1.16 (1.003 to 1.34). Age did not reach statistical significance 1.14 (0.998 to 1.31). Individuals were included in the category if they had a history of this type of surgery, but they may also have had other surgeries. However, these categories are not mutually exclusive. Adhesions were significantly associated with PD catheter complications that required an invasive procedure or led to PD never starting or termination (Figure 2 ). Of the 120 individuals with a primary outcome, 14 never started PD (ten flow-related, four pain-related), 11 were PD terminations (eight flow-related, three pain-related), and 95 were invasive procedures (85 flow-related, ten pain-related). PD never started was dated as hemodialysis start ( N =9); PD catheter removal ( N =3) or the date training was stopped ( N =2). The cumulative risk of the primary outcome 6 months after insertion was 17% ( N =45) in the adhesion group compared with 10% ( N =58) in the nonadhesion group (unadjusted HR, 1.60; 95% CI, 1.09 to 2.35). The adjusted HR for the primary outcome 1.64 (95% CI, 1.05 to 2.55) compared with the no adhesions group. The adjusted HR for the primary outcome for lower abdominal or pelvic adhesions was 1.80 (95% CI, 1.08 to 2.98) compared with the no adhesions, respectively. Adhesions in other locations were not significantly associated with a higher risk. Cumulative risk of PD catheter complications by the presence of adhesions. PD catheter complications were defined as flow restriction or abdominal pain leading to PD never starting, PD termination, or the need for an invasive procedure to treat the complication. The cumulative risk of the primary outcome 6 months from insertion was 17% in the adhesions group compared with 10% in the nonadhesion group (unadjusted HR, 1.60; 95% CI, 1.09 to 2.35). CI, confidence interval. Adhesiolysis was performed in 86 (43%) of 201 patients with adhesions. ( Supplemental Figure 1 ) Adhesiolysis was performed in 55 (47%) of 117 patients with lower abdominal/pelvic adhesions compared with 31 (37%) of 84 in other locations ( P = 0.015). If adhesions were present and adhesiolysis was not performed, the HR for the primary outcome was 1.70 (95% CI, 1.02 to 2.82) compared with the no adhesion group, but if adhesiolysis was performed, the HR was 1.55 (95% CI, 0.86 to 2.80) compared with the no adhesions group (Table 3 ). Lower adhesions with and without adhesiolysis had a risk of 1.87 (95% CI, 1.02 to 3.44) and 1.72 (95% CI, 0.88 to 3.38), respectively, compared with the no adhesion group. Association of adhesions and adhesiolysis with peritoneal dialysis catheter-related complications HR was adjusted for age, sex, body mass index, and prior abdominal surgery. Invasive procedures were required in 26 (13%) and 47 (8%) of the adhesion and no adhesion groups, respectively (unadjusted HR, 1.60, 95% CI, 1.04 to 2.47) within 6 months of insertion. The abnormalities are presented in Table 4 . In those individuals with adhesions at the original PD catheter insertion, adhesions were observed during laparoscopic revision in ten of 21 (49%) and 5 (50%) underwent adhesiolysis. Among those without adhesions at the initial insertion, 15 of 38 (39%) had new adhesions observed during the revision and nine underwent adhesiolysis, indicating adhesions formed after PD catheter insertion. PD catheter tip migration was observed in 21 (64%) and 33 (55%) patients in the adhesion and nonadhesion groups, respectively. Repositioning of the tip was performed during radiologic manipulation or laparoscopic revision in 19 (58%) and 36 (60%) patients, respectively. PD slings and omentopexy, which can only be performed during laparoscopic procedures, were performed in 5 (24%) and 6 (16%) cases, respectively. Abnormalities found among individuals who had a radiologic manipulation or laparoscopic revision to treat peritoneal dialysis catheter-related complications Adhesion groups were defined at the time of the original peritoneal dialysis catheter insertion and not during an invasive procedure to treat complications. CI, confidence interval. Laparoscopic procedures include laparoscopic exploration where the original peritoneal dialysis catheter remains in place, removal of the original peritoneal dialysis catheter, and replacement with a new catheter. Other locations include central, out of pelvis, anterior to rectum, anterior abdominal wall, stuck in small bowel, coiled in bowel, and migration superiorly. Other abnormalities included catheter being too short ( N =4), deep cuff inside peritoneum ( N =3), limited pelvic space due to prior transplant, adherent to Merkel's diverticulum, adherent to cord lipoma, distend bladder displacing catheter ( N =4), bulky uterus displacing catheter, occluded by endometriosis ( N =2), catheter kinking ( N =2), catheter damage during exteriorization, adhered to appendix, catheter damage (hold in it), broken peritoneal dialysis sling, adherent to mesentery ( N =2), catheter in preperitoneal space, and rectal loading. Lumen occlusion by fibrin or debris, which can be detected during both radiologic and laparoscopic procedures, was found in 7 (21%) and 11 (18%) patients in the adhesion and no adhesion groups, respectively ( P = 0.74). Lumen clearance was performed in 6 (18%) and 9 (15%) patients. In three cases, the PD catheter was replaced, so the lumen was not cleared.

This large, prospective, multicenter study found that approximately one-quarter of the patients undergoing laparoscopic PD catheter had adhesions. Thirty percent of patients had prior abdominal surgery, which was the dominant risk factor for adhesions, but women and individuals with higher BMI were also at risk. Adhesions were associated with restricted flow through the PD catheter or abdominal pain. Adhesiolysis was performed in less than half of the patients with adhesions and was targeted to lower adhesions but only had modest effects. Procedures to treat complications found a wide variety of abnormalities, including the formation of new adhesions. Notably, most patients, with or without adhesions, did not experience complications, and most patients required revision procedures, not termination of PD. The prevalence of prior abdominal surgery in patients undergoing PD catheter insertion ranges from 22% to 59% in previous studies compared with 30% in this study. 13 – 16 , 19 – 21 Most of these studies do not report the prevalence of adhesions but rather the use of adhesiolysis, which ranges from 1% to 31% (mean 19%) compared with 11% in this study. Crabtree found that the number of prior surgeries and women were at higher risk for adhesiolysis. 13 Women may have undergone prior gynecologic surgery, which increases the risk of adhesions in the pelvis. We did not find that a greater number of surgeries were associated with a higher risk of complications. Our results showing a strong association between prior abdominal surgery and adhesions, but not escalating risk with a greater number of surgeries, may differ from previous studies because we provided adjusted risks for adhesions and adhesiolysis separately. Our study supports the importance of risk-stratifying patients before PD catheter insertion but does not support excluding them from PD for multiple prior surgeries. Patients with a history of abdominal surgeries should be informed that adhesions may be present at baseline and, if found during surgery, will increase their risk of PD complications. Resetting expectations early may reduce frustration if flow restrictions or pain occurs early in the course of PD. It is unclear from previous studies if adhesions are associated with a higher risk of PD catheter complications or if adhesiolysis mitigates this risk. Adhesions may cause intra-abdominal compartmentalization, reducing the pelvic space and effective surface area of the peritoneum. 16 The tip of the PD catheter may also adhere to adhesions with or without the involvement of the omentum, small bowel, colon, or fallopian tubes. Crabtree found that the adhesiolysis was associated with a higher risk of mechanical catheter obstruction, but it did not affect overall catheter survival. 13 By contrast, other studies have found no effect of adhesion or adhesiolysis on catheter outcomes. 14 , 15 Our results clearly show that adhesions are a risk factor for PD catheter complications, and adhesiolysis was associated with only a modestly lower risk despite being recommended by laparoscopic guidelines. 9 We found surgeons performed adhesiolysis selectively in only 43% of the patients with adhesions and targeted lower abdominal or pelvic adhesions. This study indicates that clinicians should carefully review operative reports to ascertain whether adhesions were present, their locations, and whether adhesiolysis was performed. Further studies are required to better understand the role of adhesiolysis in preventing PD catheter complications. One of the challenges in managing PD catheter complications is the wide variety of potential causes. Previous studies of radiologic manipulation found that 36%–50% have tip migration. 22 , 23 We also found that approximately half the patients undergoing procedures had tip migration, but lumen occlusion was also commonly reported. Abnormalities reported from previous studies of laparoscopic revision included omental wrapping, involvement of the small bowel, adhesions to the abdominal wall, and wrapping of the catheter tip by the fallopian tube. 24 – 26 These studies were generally single-center, small studies. This study provides a comprehensive view of the abnormalities found during both radiologic manipulation and laparoscopic revisions. Our results show that while radiologic manipulation may detect causes, such as migration and lumen occlusion, it cannot reliably detect other causes, so laparoscopic revision will often be required for definitive diagnosis and management. We also found adhesions formed after PD insertion, indicating it can be an inflammatory stimulus for some patients. This study has some limitations. We only reported data on patients who underwent laparoscopic PD catheter insertion because they could be assessed for adhesions. We have no information on patients excluded from PD due to prior surgery. Alternatively, lower-risk patients without prior abdominal surgery may have been directed to percutaneous PD catheter insertions. We did not use a standardized classification of adhesions or adhesiolysis, which might be used in future studies but would require strong engagement by surgeons. We included embedded catheters, which had longer periods from insertion to use, and some were not used during follow-up, which would lower the risk of complications compared with nonembedded catheters. Despite being the largest study to date, we had limited power to analyze subgroups on the basis of adhesion location and the use of adhesiolysis. We also focused on flow restriction and pain because, a priori , we thought these would be related to adhesions and it aligns with the definition of patency in guidelines. We did not examine a broader range of complications, such as infections, which may cause adhesions or be indirectly associated with adhesions. In conclusion, people with intra-abdominal adhesions undergoing PD catheter insertion were at higher risk for abdominal pain or flow restriction preventing PD from starting, PD termination, or requiring an invasive procedure. This information will help practitioners counsel patients with prior abdominal surgeries about the risk of complications, set expectations, and allow careful follow-up.

Peritoneal dialysis (PD) is considered equally effective for the treatment of kidney failure as hemodialysis but is less costly. 1 – 4 Individuals choosing PD require PD catheter insertion performed either percutaneously by an interventional radiologist or nephrologist or surgically using an open or laparoscopic method. Laparoscopic procedures allow direct visualization of the peritoneal space and detection of adhesions. 5 – 7 Adhesiolysis is recommended if the surgeon determines that the adhesions are likely to impede catheter placement or hinder its proper function. 8 , 9 Previous research has shown that the risk of adhesions averaged 54% among individuals with prior abdominal surgery, but it varied considerably by the type of surgery. 10 – 12 Crabtree found that individuals undergoing adhesiolysis were more likely to have mechanical obstruction, but catheter survival was high at 90% at 5 years. 13 Other studies found that adhesions are not associated with a higher risk of obstruction, revision, or catheter survival when adhesiolysis is performed. 14 – 17 These results may be less generalizable because they were conducted in single centers, and adjusted measures of risk were not reported. The objectives of this study were ( 1 ) to describe factors associated with the presence of intra-abdominal adhesions, ( 2 ) to describe the association of adhesions with PD catheter-related complications, and ( 3 ) to describe abnormalities found at the time of PD catheter revision procedures.