The Clock Starts Now: Does Early Start PD Improve Patient Outcomes?

Written by: Dr. Anant S. Pilawan

The increasing burden of chronic kidney disease worldwide necessitates the provision of cost-effective kidney replacement therapy (KRT). The underutilisation of peritoneal dialysis (PD) is a pressing issue that needs to be addressed. Despite the numerous advantages of PD over hemodialysis (HD), PD accounts for only 9% of all kidney replacement therapy (KRT) and 11% of all dialysis worldwide.

Efforts are underway to enhance the use of PD. Early start peritoneal  dialysis (ESPD), which involves starting PD within 14 days of catheter insertion, is being explored to improve patient outcomes. Current guidelines recommend waiting at least two weeks after catheter insertion before starting PD, known as conventional start peritoneal dialysis (CSPD). ESPD is gaining popularity internationally due to its potential benefits. ESPD offers a smoother transition to home-based therapy with a single definitive access, bypassing the need for bridging hemodialysis. Additionally, ESPD reduces the risks and complications associated with central venous catheter use, and the costs and infrastructure associated with hemodialysis from a resource perspective (Figure 1).

Figure 1: Comparison between ESPD and CSPD in PD initiation.

However, studies by Ranganathan et al., Htay H and Crabtree H have indicated that ESPD may be linked to a higher risk of mechanical complications than CSPD. A randomised controlled trial showed a greater risk of pericatheter leak with earlier commencement of PD, particularly in diabetic patients. Nevertheless, no definitive recommendation could be made for or against ESPD due to limitations in the available evidence.

Given the potential variation in the risk of mechanical complications based on the timing of PD commencement, further investigation into the practice of "urgent-start" PD, defined as PD initiation within three days of catheter insertion, is necessary. This approach has gained traction internationally, particularly for patients with urgent indications for dialysis who would otherwise receive hemodialysis until definitive dialysis access is established. However, studies done by Povlsen et al. and Nayak KS suggest that patients who start on urgent-start PD experience a higher frequency of mechanical complications, including pericatheter leak, compared to patients who start PD after a more extended break-in period.

The main objective of this study was to compare the outcomes of ESPD and CSPD for all patients who began PD using a newly inserted PD catheter. The study examined the clinical outcomes of new dialysis patients who started PD within 72 hours of catheter insertion compared to those who began PD after 72 hours but within 14 days of catheter insertion. Additionally, the study analysed the outcomes for patients who had their PD catheter inserted laparoscopically versus those who had it inserted using the modified Seldinger technique, an ultrasound-guided minimally invasive peritoneal dialysis catheterisation. It also looked at outcomes for patients who did and did not receive hemodialysis before starting PD.

Visual abstract by @cteodosiu

Study design:

A retrospective single-center study included 297 patients who started peritoneal dialysis (PD) using a newly inserted PD catheter between August 2019 and August 2022. The study included incident dialysis patients, patients transitioning to PD from HD, and patients restarting PD after catheter placement. A subgroup analysis of incident PD patients was conducted separately to assess urgent PD commencement. 

Early-start PD (ESPD) patients were divided into Urgent (U-ESPD), commencing PD within 3 days of catheter insertion and Elective (E-ESPD), commencing PD between 3 and 14 days. PD catheters were inserted either laparoscopically or using the modified Seldinger technique. 

The data was collected from electronic medical records, including patient demographics, kidney function, and PD-related characteristics. The outcomes assessed were infectious and mechanical complications, including PD-related peritonitis and PD catheter insertion–related peritonitis. PD exit site infection and PD catheter insertion-related exit site infection or tunnel infection were defined as infections occurring within 1 month of catheter insertion. Mechanical complications include immediate procedural issues such as failed catheter insertion, organ injury, hemoperitoneum, and postoperative surgical site bleeding. Other complications include pericatheter leak, pleuroperitoneal leak and catheter malposition. Additionally, the time to the first PD-related peritonitis episode and time to a permanent transfer to hemodialysis were assessed. At the end of the follow-up period, the current kidney replacement therapy modality and patient death events were recorded.

Results:

In a study of 297 patients, 225 were new to dialysis, 65 had previous hemodialysis, and 7 had catheter replacements. The average age was 58.7 years, with 70% being male. Diabetic nephropathy was the most common cause of kidney failure (43.1%). Catheters were inserted using laparoscopic techniques in 65.3% of cases (Table 1).

Of the patients, 130 (43.8%) started an early start peritoneal dialysis (ESPD) program, beginning a median of 11 days after catheter insertion compared to 26 days for conventional peritoneal dialysis (CSPD). ESPD patients were younger (56.6 vs. 60.3 years, P = 0.049) and exhibited worse kidney function (6.8 ml/min vs. 7.8 ml/min, P = 0.002). Additionally, they had a higher post-procedural admission rate (42.3% vs. 28.7%, P = 0.02). Within the ESPD group, 95 were new dialysis patients (Table 2), divided into urgent (U-ESPD) and elective (E-ESPD) groups. The median time to start PD was 1 day for U-ESPD and 12 days for E-ESPD. U-ESPD patients were more likely to have catheters inserted using the modified Seldinger technique (90% vs. 41.2%, P = 0.004) and experienced extended hospital stays (10.1 vs. 2.5 days, P = 0.001).

During an average follow-up of 27.9 months, 58.6% of patients experienced complications related to peritoneal dialysis (PD) (Table 3). Those starting early peritoneal dialysis (ESPD) had more pericatheter leaks than those on standard peritoneal dialysis (CSPD) (6.9% for ESPD vs. 0.6% for CSPD, P = 0.003). Most leaks occurred within 30 days of starting PD. ESPD did not increase other early complications. Rates of PD catheter insertion-related peritonitis, exit site infections, and surgical bleeding were similar between groups (e.g., peritonitis: ESPD 3.8% vs. CSPD 8.4%, P = 0.1).

For longer-term complications, mechanical issues like catheter malposition and pleuroperitoneal leaks were similar for both groups. There were comparable rates of infection-related complications, including PD-related peritonitis and exit site infections. A subgroup analysis compared unplanned early-start PD (U-ESPD) to early-planned PD (E-ESPD) (Table 4). While leaks were more common in the E-ESPD group, this was not statistically significant (U-ESPD 0% vs. E-ESPD 7.1%, P = 0.5).

A statistical model showed that patients starting ESPD were 16.3 times more likely to experience pericatheter leaks than those on CSPD (95% CI: 1.9–137.9, P = 0.01). Other factors like diabetes and BMI did not significantly affect leak rates. Management of leaks varied. Half of the patients continued on PD, while others paused treatment. Most who paused treatment resumed PD within 2 weeks, but one patient needed hemodialysis temporarily. By the end of the follow-up period, 68.7% of patients continued PD, 17.8% switched to hemodialysis, 5.7% received transplants, and 7.7% died. None regained kidney function. Kaplan-Meier analysis showed no significant difference between ESPD and CSPD in rates of PD-related peritonitis or transfer to hemodialysis (Figures 1 and 2).

A comparison of two methods for starting peritoneal dialysis, ESPD and CSPD, found no significant difference in the time it took for patients to develop peritonitis (HR: 1.0, P = 0.99). This study included deaths, transplants, or transfers to hemodialysis as risks to consider (Table 5).

Furthermore, the timing of when peritoneal dialysis began did not significantly impact the rate of transfers to hemodialysis (HR: 0.8, P = 0.5) when including death or transplant as competing risks (Table 6). Complications were similar for both laparoscopic catheter insertion and the modified Seldinger technique.

This study compared two types of PD in a large Australian unit, Early Start Peritoneal Dialysis (ESPD) and Conventional Start Peritoneal Dialysis (CSPD), and found that ESPD had more catheter leaks than CSPD. However, both groups had similar safety profiles and no impact on long-term outcomes like PD-related infections or the need for hemodialysis. Starting PD within 72 hours of catheter insertion was as safe as starting 3 to 14 days later. Previous hemodialysis treatment and different insertion techniques did not affect the outcomes.  The higher rate of leaks in the ESPD group aligns with other studies. While leaks can cause issues, only one patient (10%) required bridging hemodialysis. No patients needed surgery for persistent leaks. Analysis in this study showed that leaks did not significantly affect the timing of PD-related infections or the switch to hemodialysis. The leak rate in the ESPD group (6.9%) was similar to other studies and considered low compared to the benefits of early PD. Risk factors for catheter leaks include the dialysis method and patient characteristics like obesity and diabetes. There was no strong link between diabetes and leaks, contrasting with other findings. Although leaks were more common in obese patients, this was not statistically significant. It was also noted that physical strain and constipation might contribute to leaks. Future studies should focus on bowel care during PD initiation. On examining patients with and without previous hemodialysis, those from hemodialysis had lower pre-surgery urea levels, and rates of surgical bleeding and leaks were similar, suggesting hemodialysis may not provide a protective effect.

This study has limitations. It is observational so unmeasured factors may influence pericatheter leaks, and the known risk factors only account for a small portion of the overall risk. As a single-center study, selection bias may have affected the results since patients were chosen for either ESPD or CSPD and different catheter placement techniques. The small number of U-ESPD patients also makes it difficult to assess outcomes for this dialysis type. Furthermore, the varying experience levels of the interventionists, from seasoned professionals to trainees, could impact results.