Long-term impact of early non-infectious complications at the initiation of peritoneal dialysis

Abstract

Background:

Early non-infectious complications at initiation of peritoneal dialysis (PD) are a major burden with unknown long-term impacts on individuals.

Methods:

Prospective multicentre cohort study using univariable and multivariable Cox regression to identity mortality risk and PD discontinuation risk in those with and without non-infectious complications. All individuals commencing PD between 1 January 2014 and 31 December 2018, registered in the New Zealand Peritoneal Dialysis Registry (NZPDR) were followed up to 31 December 2020. Early non-infectious complications defined as functional, catheter-related, exit-site dialysate leak or anatomical leak complications occurring within 30 days of initiation of PD. Primary outcomes were patient survival and time on PD therapy. Secondary outcomes were peritonitis free survival, first PD catheter survival and catheter tunnel infection free survival.

Results:

Of 1596 individuals included in the study, 102 experienced an early non-infectious complication. Multivariable analysis demonstrated these complications were associated with higher risk of overall mortality (hazard ratio (HR) 1.71; 95% confidence interval (CI) 1.21–2.44), PD discontinuation (HR 1.84; 95% CI 1.41–2.41) and first catheter failure (HR 2.89; 95% CI 2.28–3.66). No difference was found for risk of development of first peritonitis episode or catheter tunnel infection. Mortality risk was associated with functional and exit-site dialysate leak complications and continued beyond 180 days. Risk of PD discontinuation and first catheter loss were associated with catheter and functional complications in the first 180 days.

Conclusion:

Early non-infectious complications are associated with long-term mortality risk. Further research in risk factors and causes of early non-infectious complications are required.

Keywords

Early catheter non-infectious complications peritoneal dialysis outcomes

Introduction

Since the introduction of continuous ambulatory peritoneal dialysis (PD) in the late 70s, it has become increasingly used as a home-based method of kidney replacement therapy. Current estimates worldwide suggest PD represents 11% of the dialysis population with a global growth rate of the number of patients on PD being 8% annually.¹ New Zealand has one of the highest prevalence of PD usage worldwide with 176 individuals per million population in 2019.^1,2 PD has been associated with increased patient quality of life and patient satisfaction compared to haemodialysis.^3,4 Studies examining survival differences between the two modalities however have mixed results.^5
–7

Complications of PD, either due to infectious or non-infectious causes, can lead to transfer to haemodialysis in up to 25% of individuals with more than 70% of them occurring in the first 2 years after initiation of PD, predominantly associated with peritonitis.⁸ PD discontinuations caused by non-infectious causes have been reported up to nine episodes per 100 patient-years at 6 months and seven episodes per 100 patient-years at 12 months in Australia and New Zealand.⁹ While there has been a focus on predictors of PD discontinuation, such as peritonitis rates, catheter types, insertion techniques and salvage techniques for peritoneal catheters,^{8

–16} there has been little focus on the consequences of early non-infectious complications following insertion of peritoneal catheters.

To the best of our knowledge, only one prospective study examined the optimum timing for catheter usage in regard to reducing risk of catheter leakage or primary catheter malfunction.¹⁷ Although prospective data do exist for predictors for PD discontinuation in PD, this does not focus on how early non-infectious complications at the initiation of PD affects either PD discontinuation rate or survival rate.⁸ Our study examined the impact of early non-infectious complications at the initiation of PD on individual’s long-term outcomes.

Materials and methods

Study design

This was a multicentre prospective cohort study using individual level data from the New Zealand Peritoneal Dialysis Registry (NZPDR).¹⁸ Permission to analyse the de-identified data was obtained from the NZPDR steering committee. NZPDR prospectively collected data on all PD patients from contributing centres and has ethical approval to collect and analyse the data.

Study population

All individuals who initiated PD between 1 January 2014 and 31 December 2018 and registered in the NZPDR were included in the study and were followed up to 31 December 2020. This included all individuals commencing PD, including those transferring from haemodialysis or returning to dialysis following a failed kidney transplant. All age groups fulfilling criteria were included.

Data collection

All data were collected prospectively by the NZPDR. Baseline characteristics at initiation of PD included sex, age, ethnicity, weight, presence of diabetes, primary cause for kidney disease, PD catheter type (coiled end/straight intercuff, straight end/swan neck intercuff, coiled end/swan neck intercuff, straight end/straight intercuff), operator who inserted catheter, year started PD, number of days post insertion prior to PD catheter use, small solute clearance (Kt/V), total kidney clearance from dialysis and residual kidney function and peritoneal equilibrium testing. Initiation of PD was defined as the first date of PD treatment. Non-infectious complications were grouped and defined as either catheter, functional, exit-site dialysate leak or anatomical leak complications. Catheter complications were defined as solely related to the catheter itself such as migration or damage to the catheter while functional complications were defined as complications with either inflow or outflow of dialysate. Exit-site dialysate leak complications were defined as the appearance of any moisture around the PD catheter exit site identified as dialysate. Anatomical leak complications were defined as loss of dialysate from the peritoneal cavity due to innate anatomical weakness or fistulae. Surgical, infective and other complications were not included in the definition of non-infectious complications. Infective complications were defined as either exit site, tunnel track or peritonitis infections, and surgical complications pertain to either intraoperative or post-operative complications associated with the insertion of the catheter including surgical wound infections. NZPDR complication codes and corresponding subgroups are provided as a supplementary file. Complications that occurred within 30 days of initiation of dialysis were defined as early complications. The first early complication an individual encountered assigned their complication type.

Clinical outcomes

The primary clinical outcomes were patient survival and time on PD therapy. Time on PD therapy was defined as number of days until transfer to haemodialysis therapy for more than 90 days, kidney transplantation or withdrawal from dialysis due to medical or psychosocial reasons. Secondary outcomes evaluated were peritonitis free survival, first PD catheter survival and PD catheter tunnel infection free survival. These secondary outcomes were defined as the period of time until the development of either peritonitis, removal of first PD catheter (whether resulting in or not resulting in PD discontinuation) or development of a catheter tunnel infection respectively.

Statistical analyses

Baseline characteristics that were in categorical form were presented as number of individuals and percentages, while those in continuous form were presented as median and interquartile range (IQR). Follow-up continued until the occurrence of an outcome (see above) or until became censored as appropriate for death, kidney transplantation or discontinuation from the registry. Time from initiation of PD to occurrence of each outcome (primary or secondary) was first compared between those with and without any type of non-infectious complications. This was done by using Kaplan–Meier plots and by estimating hazard ratios (HRs) using univariable Cox regressions. Then, multivariable Cox regressions were used to adjust for confounders. Underlying proportional hazard assumption was assessed by testing for nonzero slope in a generalised linear regression of the scaled Schoenfeld residuals on time, which is equivalent to testing that the log HR function is constant over time.¹⁹ These multivariable models included the covariates of sex, age, catheter type, operator who inserted the PD catheter and primary disease which caused end-stage kidney failure (ESKF). These adjusting covariates were chosen based on association with outcome of interest as per univariable analyses with p value <0.2 threshold and biological plausibility. Diabetes mellitus was not used as a separate adjusting variable due to perfect collinearity with ‘primary cause of ESKF’. Pairwise deletion was used in the multivariable analyses for missing covariate data. To assess longer term effects, sensitivity analyses were then performed using univariable and multivariable models with different two time intervals of 0–180 days and 181 days onwards since initiation of PD to estimate HRs for their respective outcomes.

Then, the same analysis was repeated to compare survival times between those who experienced each type of non-infectious complication (catheter, functional, exit-site leak) and those without that type of non-infectious complication. These latter analyses were confined only to the univariable level to avoid model overfitting due to smaller sample sizes.²⁰ Anatomical leaks were not analysed in subgroup analysis due to underlying pathological heterogeneity and small sample size. Sensitivity analyses were also performed using univariate models on their respective outcomes to assess statistically significant difference of HRs with two different time intervals of 0–180 days and 181 days onwards since initiation of PD.

Stata/SE16.0 (StataCorp, Texas, USA) was used in all analyses. The findings are reported following the STROBE statement.²¹

Results

Individuals

A total of 1596 individuals who were initiated PD between 1 January 2014 and 31 December 2018 and registered in NZPDR were identified (Figure 1). The demographic characteristics of these participants are presented in Table 1. Median age was 60 years (IQR 49–69) and 59.5% were male. Ethnic classification was 40.1% European, 27.8% Māori and 12.5% Pacifica. Diabetes mellitus was the aetiology of kidney failure for 43.4% of individuals. Median number of days after insertion of PD catheter prior to use was 24 (IQR 17–32).

Figure 1.

CONSORT flow diagram for multicentre prospective cohort study using registry level data.

Table 1.

Baseline characteristics and covariates of study participants.

Characteristic	All individuals no. (%) (n = 1596)	Non-infectious complication no. (%) (n = 102)	No non-infectious complication no. (%) (n = 1494)	Catheter complication no. (%) (n = 26)	Functional complication no. (%) (n = 39)	Exit-site leak complication no. (%) (n = 21)
Sex
Male	950 (59.5)	58 (56.9)	892 (59.7)	16 (61.5)	18 (46.2)	12 (57.1)
Female	646 (40.5)	44 (43.1)	602 (40.3)	10 (38.5)	21 (53.8)	9 (42.9)
Diabetes mellitus
No	904 (56.6)	53 (52.0)	851 (57.0)	10 (38.5)	20 (51.3)	13 (61.9)
Yes	692 (43.4)	49 (48.0)	643 (43.0)	16 (61.5)	19 (48.7)	8 (38.1)
Ethnicity
European	640 (40.1)	28 (27.5)	612 (41.0)	4 (15.4)	11 (28.2)	6 (28.6)
Māori	443 (27.8)	46 (45.1)	397 (26.6)	13 (50.0)	17 (43.6)	9 (42.9)
Pacifica	198 (12.4)	8 (7.8)	190 (12.7)	5 (19.2)	2 (5.1)	1 (4.7)
Asian	77 (4.8)	5 (4.9)	72 (4.8)	1 (3.9)	1 (2.6)	1 (4.7)
Other	238 (14.9)	15 (14.7)	223 (14.9)	3 (11.5)	8 (20.5)	4 (19.1)
Peritoneal dialysis catheter type
Straight end/swan neck intercuff	285 (17.9)	21 (20.6)	264 (17.7)	4 (15.4)	9 (23.1)	5 (23.8)
Coil end/straight intercuff	393 (24.6)	24 (23.5)	369 (24.7)	5 (19.2)	9 (23.1)	7 (33.3)
Coil end/swan neck intercuff	406 (25.4)	22 (21.6)	384 (25.7)	15 (15.4)	7 (18.0)	4 (19.1)
Straight end/straight intercuff	361 (22.6)	21 (20.6)	340 (22.7)	5 (19.2)	10 (25.6)	4 (19.1)
Not stated	151 (9.5)	14 (13.7)	137 (9.2)	8 (30.8)	4 (10.2)	1 (4.7)
Operator inserting peritoneal dialysis catheter
Surgeon	1065 (66.7)	46 (45.1)	1019 (68.2)	11 (42.3)	15 (38.4)	9 (42.9)
Nephrologist	400 (25.1)	48 (47.1)	352 (23.6)	11 (42.3)	22 (56.4)	11 (53.4)
Interventional radiologist	82 (5.1)	4 (3.9)	78 (5.2)	2 (7.7)	1 (2.6)	1 (4.7)
Not stated	49 (3.1)	4 (3.9)	45 (3.0)	2 (7.7)	1 (2.6)	0 (0)
Primary disease cause of end-stage kidney failure based on major groups
Glomerulonephritis	269 (16.9)	16 (15.7)	253 (16.9)	4 (15.4)	5 (12.8)	3 (14.3)
Vascular/hypertension	145 (9.1)	10 (9.8)	135 (9.0)	2 (7.7)	4 (10.3)	3 (14.3)
Autosomal dominant polycystic kidney disease	46 (2.9)	2 (2.0)	44 (3.0)	0 (0)	0 (0)	1 (4.8)
Diabetes mellitus	692 (43.4)	49 (48.0)	643 (43.0)	16 (61.5)	19 (48.7)	8 (38.1)
Other	251 (15.7)	17 (16.7)	234 (15.7)	2 (7.7)	6 (15.4)	5 (23.8)
Diagnosis not listed	193 (12.1)	8 (7.8)	185 (12.4)	2 (7.7)	5 (12.8)	1 (4.7)
	Median (IQR)	Median (IQR)	Median (IQR)	Median (IQR)	Median (IQR)	Median (IQR)
Number of days after insertion of peritoneal dialysis catheter prior to use	24 (17–32)	20 (15–25)	24 (18–33)	17 (12–24)	20 (17–27)	17 (14–24)
Age of patient when started PD (years)	60 (49–69)	62.5 (53–70)	60 (49–69)	60 (56–67)	63 (54–74)	66 (49–70)
Weight of patient at start in kg	77.1 (66–90)	76.5 (64–87)	77.5 (66.5–90)	75 (62–82)	72.4 (63–88)	75.5 (64.5–85.5)
First Kt/V (mL/min)	2.1 (1.7–2.5)	1.9 (1.6–2.5)	2.05 (1.7–2.5)	1.9 (1.5–2.2)	2.1 (1.7–2.7)	2.2 (1.4–2.5)
First total creatinine clearance – dialysis clearance plus residual kidney function (L/wk/1.73 m²)	75.2 (57.7–98.2)	67 (52.8–91.2)	75.7 (58.5–98.4)	64.4 (51.5–85.8)	71.6 (55.5–94.5)	76.8 (50.5–107.3)
First available peritoneal equilibrium testing	0.74 (0.65–0.82)	0.73 (0.65–0.80)	0.74 (0.65–0.82)	0.70 (0.56–0.77)	0.76 (0.59–0.87)	0.76 (0.68–0.81)

IQR: interquartile range.

Follow-up period during this study varied depending on the outcome measure and associated censoring. For the two primary outcomes of patient death and time on PD therapy, median follow-up days were 730 (IQR 369–1135)/756 (IQR 302–1182) (death/no death) and 450 (IQR 193–814)/983 (IQR 572–1375) (PD discontinuation/no discontinuation) respectively (Table 2).

Table 2.

Outcomes in patients with and without early non-infectious complications and in subgroups.

Outcome measure	All individuals		Non-infectious complication		No non-infectious complication		Catheter complication		Functional complication		Exit-site leak complication
Outcome measure	No (%)	Follow-up days median (IQR)	No (%)	Follow-up days median (IQR)	No (%)	Follow-up days median (IQR)	No (%)	Follow-up days median (IQR)	No (%)	Follow-up days median (IQR)	No (%)	Follow-up days median (IQR)
Patient survival
Death	465 (29.1)	730 (369–1135)	34 (33.3)	716 (294–1441)	430 (29.0)	731 (369–1129)	6 (23.1)	1120 (457–1550)	13 (13.3)	787 (543–1316)	12 (57.1)	418 (169–716)
No death	1131 (70.9)	755 (302–1182)	68 (66.6)	285.5 (60–677.5)	1064 (71.0)	795 (322.5–1195.5)	20 (76.9)	253 (104–550)	26 (66.7)	112 (34–521)	9 (42.9)	720 (471–1523)
Time on peritoneal dialysis therapy
PD discontinued	716 (44.9)	450 (193–814)	58 (57.0)	148.5 (38–467)	657 (44.0)	477 (212–840)	16 (62.5)	151 (88–378)	24 (61.5)	67.5 (32–426)	5 (23.8)	471 (94–525)
PD continued	880 (55.1)	983 (572–1375)	44 (43.0)	827 (438–1519)	837 (56.0)	991 (599–1368)	10 (38.5)	113 (798–1550)	15 (38.5)	828 (543–1739)	16 (76.2)	612 (251–1519)
Peritonitis free survival
Peritonitis	753 (47.2)	332 (148–635)	36 (35.3)	315 (95.5–523)	716 (48.0)	332.5 (148.5–637.5)	9 (34.6)	316 (49–766)	12 (30.8)	308 (129–371)	9 (42.9)	273 (86–461)
No peritonitis	843 (52.8)	543 (202–971)	66 (64.7)	148.5 (38–543)	778 (52.0)	574 (227–995)	17 (65.4)	144 (75–374)	27 (69.2)	71 (29–710)	12 (57.1)	345 (94–613)
Peritoneal dialysis catheter survival
Loss of first catheter	831 (52.1)	409 (144–738)	77 (75.5)	74 (29–374)	752 (50.3)	441.5 (197.5–771.5)	20 (76.9)	103.5 (32–331)	35 (89.7)	48 (25–163)	8 (38.1)	94 (39–693)
No loss of first catheter	765 (47.9)	964 (519–1325)	25 (24.5)	730 (211–995)	742 (49.7)	971.5 (534–1320)	6 (23.1)	877 (457–1550)	4 (10.3)	822 (385–1355)	13 (61.9)	522 (211–826)
Catheter tunnel infection free survival
First catheter tunnel infection	437 (27.4)	249 (70–538)	23 (22.1)	322 (70–521)	414 (27.7)	247 (69–543)	4 (15.4)	339 (228–369)	7 (18.0)	70 (27–521)	8 (38.1)	415 (71–735)
No catheter tunnel infection	1159 (72.6)	658 (273–1103)	79 (77.9)	291 (60–764)	1080 (72.3)	683.5 (292.5–1124)	22 (84.6)	331 (100–575)	32 (82.0)	323 (38–808)	13 (61.9)	399 (127–522)

IQR: interquartile range.

There were 1366 individuals who did not experience any early complication following initiation of PD compared to 230 individuals who experienced at least one complication. These included 26 catheter complications, 39 functional complications, 21 exit-site leaks, 16 anatomical leaks, 106 infective complications, 9 surgical and 13 defined as other.

For catheter complications, 14 had catheter tip misplacement/early migration, 4 had catheter break/tear/damage, 1 had cuff extrusion/pulled out and 7 had late catheter tip migration (>4 weeks post-surgery).

For functional complications, 8 had supine and upright flow obstruction, 1 had inflow obstruction, 18 had outflow obstruction, 7 had catheter obstruction and 5 had omental entrapment.

Impact of early non-infectious complications

Out of 1596 individuals, 102 experienced at least one early non-infectious complication. Of these 102 individuals, 34 died during the follow-up period, 58 ended up with PD discontinuation, 36 had at least one episode of peritonitis, 77 experienced first catheter failure and 23 had at least one catheter tunnel infection (i.e. total number of these outcomes exceeded 102 because multiple types of outcomes possible per person).

Based upon Kaplan–Meier plots comparing outcomes between those with and without non-infectious complications (Figure 2), and HRs from univariable Cox regression, there was higher risk among those with non-infectious complications in terms of subsequent patient mortality (HR 1.71; 95% CI 1.21–2.44), PD discontinuation (HR 1.84; 95% CI 1.41–2.41) and first catheter failure (HR 2.89; 95% CI 2.28–3.66) compared to those without such complications. There was no difference in risk in terms of developing first episode of peritonitis (HR 1.12; 95% CI 0.80–1.57) or first episode of catheter tunnel infection (HR 1.08; 95% CI 0.71–1.65).

Figure 2.

Kaplan–Meier plots for primary and secondary outcomes comparing early non-infectious complications versus those without early non-infectious complications.

The univariable findings did not vary by covariate adjusting (Table 3). Multivariable Cox regressions for each outcome measure using sex, age, catheter type, PD catheter inserter and primary cause for ESKF as adjusting covariates showed those with non-infectious complications had a higher risk of death (HR 1.62; 95% CI 1.11–2.37), PD discontinuation (HR 1.83; 95% CI 1.38–2.45) and first catheter failure (HR 2.70; 95% CI 2.10–3.47) compared to their counterparts without complications, while such a difference was not observed for risk of development of a first episode of peritonitis (HR 1.15; 95% CI 0.81–1.63) or a first episode of catheter tunnel infection (HR 0.99; 95% CI 0.62–1.56).

Table 3.

Multivariate (adjusted) hazard ratios comparing outcomes in patients with and without early non-infectious complication.

	Primary outcomes					Secondary outcomes
	No.	Patient survival		Time on PD therapy		Peritonitis free survival		First catheter survival		Tunnel infection free survival
	No.	HR (95% CI)	p Value^b	HR (95% CI)	p Value^b	HR (95% CI)	p Value^b	HR (95% CI)	p Value^b	HR (95% CI)	p value^b
Non-infectious complication^a
No	1273	Reference		Reference		Reference		Reference		Reference
Yes	90	1.62 (1.11–2.37)	0.013	1.83 (1.38–2.45)	<0.001	1.15 (0.81–1.63)	0.439	2.70 (2.10–3.47)	<0.001	0.99 (0.62–1.56)	0.951
Sex
Male	803	Reference		Reference		Reference		Reference		Reference
Female	560	1.05 (0.86–1.29)	0.627	0.79 (0.67–0.93)	0.006	0.98 (0.83–1.14)	0.754	0.87 (0.75–1.02)	0.080	1.14 (0.92–1.40)	0.227
Age^c
One year increase	1363	1.04 (1.03–1.05)	<0.001	0.98 (0.979–0.989)	<0.001	1.000 (0.990–1.010)	0.885	0.990 (0.980–0.990)	<0.001	0.99 (0.99–1.00)	0.129
Catheter type
Straight end/swan neck intercuff	272	Reference	0.198	Reference	0.161	Reference	0.120	Reference	0.831	Reference	0.007
Coil end/straight intercuff	362	0.75 (0.56–1.01)		0.97 (0.77–1.23)		1.11 (0.88–1.38)		1.07 (0.87–1.33)		0.63 (0.47–0.85)
Coil end/swan neck intercuff	389	0.95 (0.72–1.26)		1.05 (0.83–1.34)		0.83 (0.66–1.05)		1.11 (0.88–1.39)		0.85 (0.64–1.14)
Straight end/straight intercuff	234	0.96 (0.69–1.32)		0.82 (0.62–1.09)		1.06 (0.83–1.37)		1.08 (0.84–1.39)		0.68 (0.49–0.96)
Unknown/missing	106	1.15 (0.76–1.74)		1.26 (0.90–1.77)		0.99 (0.71–1.39)		0.95 (0.69–1.30)		0.52 (0.33–0.84)
PD catheter insertion proceduralist
Surgeon	913	Reference	0.735	Reference	0.568	Reference	0.463	Reference	<0.001	Reference	0.004
Renal physician	375	1.01 (0.78–1.31)		1.08 (0.88–1.33)		1.05 (0.86–1.27)		1.55 (1.29–1.86)		1.52 (1.18–1.95)
Interventional radiologist	75	0.82 (0.48–1.39)		1.17 (0.84–1.65)		0.83 (0.58–1.18)		1.30 (0.95–1.79)		1.03 (0.64–1.67)
Primary disease which caused ESKF
Glomerulonephritis	262	Reference	<0.001	Reference	0.033	Reference	<0.001	Reference	0.46	Reference	0.23
Vascular/hypertension	144	1.55 (0.99–2.41)		0.68 (0.49–0.95)		0.89 (0.64–1.23)		0.81 (0.60–1.10)		0.96 (0.66–1.42)
ADPKD	46	0.96 (0.42–2.19)		1.21 (0.79–1.86)		0.72 (0.43–1.22)		1.19 (0.78–1.80)		0.41 (0.18–0.95)
Diabetes mellitus	665	2.33 (1.62–3.35)		0.81 (0.66–1.00)		1.35 (1.09–1.68)		0.97 (0.80–1.80)		0.83 (0.64–1.09)
Other	246	1.84 (1.22–2.77)		0.97 (0.76–1.25)		0.98 (0.75–1.28)		1.04 (0.82–1.32)		0.87 (0.62–1.21)

HR: hazard ratio; CI: confidence interval; PD: peritoneal dialysis; ESKF: end-stage kidney failure; ADPKD: autosomal dominant polycystic kidney disease.

^a Number reduced to 1363 due to missing data in ‘primary disease which caused ESKF’ and ‘PD catheter insertion proceduralist’ variables.

^b p Value is testing the null hypothesis of HR = 1 for all groups in the variable.

^c Continuous variable.

As part of the multivariate analysis process, a number of additional observations were noted. The type of catheter inserted did not impact risk of mortality or PD discontinuation but there was a significant reduction in risk of developing a first catheter tunnel infection with the coiled end/straight intercuff (HR 0.63; 95% CI 0.47–0.85) and straight end/straight intercuff catheters (HR 0.68; 95% CI 0.49–0.96) when compared to straight end/swan neck intercuff catheters (Table 3). There were significantly higher risks in terms of failure of the first catheter (HR 1.55; 95% CI 1.29–1.86) and development of a first catheter tunnel infection (HR 1.52; 95% CI 1.18–1.95) among those catheters inserted by a nephrologist compared to surgical insertions (Table 3).

Sensitivity analyses with remodelling with two different time intervals of 0–180 days and 181 days onwards since initiation of PD reaffirmed the aforementioned higher mortality risk for those with non-infectious complications compared to those without such complications continues long-term (greater than 180 days), but not for PD discontinuation nor first catheter failure. Univariable analysis for the first 180 days did not show higher risk for patient mortality (HR 1.81; 95% CI 0.78–4.19) but was significantly higher for PD discontinuation (HR 3.96; 95% CI 2.69–5.81) and first catheter failure (HR 6.01; 95% CI 4.40–8.23). Beyond 180 days, however, higher mortality risk was significantly demonstrated (HR 1.96; 95% CI 1.16–2.50) but not for PD discontinuation (HR 1.12; 95% CI 0.76–1.66). Higher risk of first catheter failure just failed to reach significance (HR 1.46; 95% CI 0.99–2.17). Multivariate Cox regression resulted in similar observations with the first 180 days not demonstrating a higher risk of death (HR 1.37; 95% CI 0.54–3.49) but did demonstrate higher risk of PD discontinuation (HR 4.28; 95% CI 2.84–6.46) and first catheter failure (HR 6.14; 95% CI 4.4–8.57). Beyond 180 days, similar observations were demonstrated with higher mortality risk (HR 1.69; 95% CI 1.11–2.56) seen and no difference in either risk of PD discontinuation (HR 1.04; 95% CI 0.68–1.60) or risk of first catheter failure (HR 1.21; 95% CI 0.78–1.86) observed.

No statistical difference was seen between groups in both univariate and multivariate analysis for either time interval when comparing for outcomes for development of first peritonitis episode and first catheter tunnel infection.

Proportional hazard assumption was not found to be invalid in any of the models (both primary analyses and sensitivity analyses) except the univariable model for the first catheter failure outcome for the 0–180 days time interval. The assumption was met once adjusted for other covariates in the corresponding multivariable model.

Impact of each type of early non-infectious complication

Univariable analysis findings for each subgroup of early non-infectious complication are as follows. There was evidence of higher risk among those with catheter complications in terms of PD discontinuation (HR 1.94; 95% CI 1.18–3.18) and first catheter failure (HR 2.99; 95% CI 1.93–4.66) compared to those without catheter complications, but such a difference was not observed in terms of patient mortality (HR 1.14; 95% CI 0.51–2.56), development of first peritonitis episode (HR 1.13; 95% CI 0.58–2.14) or first catheter tunnel infection (HR 0.70; 95% CI 0.26–1.88). Sensitivity analysis with remodelling with the time interval of 0–180 days since initiation of PD demonstrated similar observations with higher risk of PD discontinuation (HR 3.69; 95% CI 1.88–7.22) and first catheter failure (HR 4.87; 95% CI 2.78–8.52) among those with catheter complications compared to those without. This was however not seen in remodelling for the time interval beyond 180 days for risk of PD discontinuation (HR 1.21; 95% CI 0.57–2.55) and first catheter failure (HR 1.75; 95% CI 0.83–3.70). No difference of risk was seen in either time interval for patient mortality.

For those with functional complications, there was a higher risk in terms of patient mortality (HR 1.74; 95% CI 1.00–3.02), PD discontinuation (HR 2.03; 95% CI 1.35–3.05) and first catheter failure (HR 4.59; 95% CI 3.27–6.45) compared to those with no functional complications. No difference was observed for development of first peritonitis episode (HR 0.93; 95% CI 0.53–1.65) or first catheter tunnel infection (HR 0.86; 95% CI 0.41–1.82). Sensitivity analysis with remodelling in the time interval of 0–180 days since initiation of PD similarly showed higher risk of PD discontinuation (HR 4.90; 95% CI 2.88–8.32) and first catheter loss (HR 9.65; 95% CI 6.44–14.46) but not in patient mortality (HR 0.81; 95% CI 0.11–5.88). Remodelling beyond 180 days, however, showed a higher risk of patient mortality (HR 1.92; 95% CI 1.08–3.41) but not in risk of PD discontinuation (HR 1.03; 95% CI 0.53–1.99) and first catheter loss (HR 1.69; 95% CI 0.84–3.40).

Exit-site leak complications showed evidence of increased risk of subsequent patient death (HR 2.14; 95% CI 1.14–4.01) but not in terms of PD discontinuation (HR 0.62; 95% CI 0.26–1.50), first catheter failure (HR 0.94; 95% CI 0.47–1.89), development of first peritonitis episode (HR 1.35; 95% CI 0.70–2.60) or first catheter tunnel infection (HR 1.76; 95% CI 0.88–3.55) when compared to those without leak complications. Sensitivity analysis showed similar risk with increased mortality in those with leak complications compared to those without in both 0 and 180 days time interval (HR 3.90; 95% CI 1.22–12.42) and beyond 180 days time interval (HR 2.11; 95% CI 1.09–4.08). No difference of risk was seen in either time interval for PD discontinuation (0–180 days: HR 0.93; 95% CI 0.23–3.75, 180+ days: HR 0.81; 95% CI 0.16–1.58) or first catheter failure (0–180 days: HR 1.93; 95% CI 0.80–4.68, 180+ days: HR 0.51; 95% CI 0.16–1.58).

When comparing the aforementioned specific complication groups with corresponding groups without that complication, sensitivity analysis did not identify a risk difference in either time interval for development of first peritonitis episode or first catheter tunnel infection.

Discussion

In this large prospective multicentre registry-base study, we examined the clinical impact of early (first 30 days) non-infectious complications following PD catheter insertion. We demonstrated that these early non-infectious complications were associated with a higher risk of overall subsequent mortality, PD discontinuation and first catheter failure. We demonstrated that mortality risk persisted long term and did not just pertain to higher risk in the short term unlike the risk of PD discontinuation and first catheter failure. This was demonstrated using a sensitivity analysis with remodelling for two separate time intervals with 180 days as a cut-off and the long median time of follow-up of greater than 700 days for patient survival in both with and those without non-infectious complications.

Early non-infectious complications were not associated with increased risks of development of first catheter tunnel infections or the first peritonitis episode. As peritonitis is a leading cause of mortality and PD discontinuation in individuals on PD,^22
–24 it is important to separate increased risk of peritonitis from early non-infectious complications on subsequent outcomes for PD.

The increased risk of mortality may be partially explained by the potential need for interim haemodialysis via haemodialysis catheters which is known to have associated increased risks of infection, cardiovascular events and mortality when compared to other forms of haemodialysis access.^25,26 The need for repetitive intervention for either haemodialysis access or re-establishment of PD access may also play a role due to the inherent associated risk of complications. Dialysis patients are more prone to complications and mortality associated with procedures in general. Studies examining adjusted mortality risks for chronic dialysis individuals report an adjusted risk ratio between 1.6 and 8.3 when compared to individuals with normal kidney function across different surgery types.²⁷ The increased risk of mortality in individuals with non-infectious complications however appeared predominantly confined to functional and leak complications only. This is likely due to differing underlying factors that predispose individuals to having certain complications and thus an underpinning a role in increased mortality. The lack of significance of higher mortality risk in the first 180 days is at least partly likely be due to the low outcome incidence and associated reduced power related with the small sample size, as evidence by the large confidence intervals. Only 6 out of 101 individuals who had an early non-infectious complication had the outcome of death in the first 180 days.

The risk of any early complications has been associated with diabetes, a history of glomerulonephritis, and has been seen more commonly in males. Previous abdominal surgery with abdominal wall complications and previous abdominal surgery in the setting of urgent start PD^28,29 have also been associated with early complications. Insertion technique of PD catheter (laparoscopic, open surgery and percutaneous insertion) may also play a role in obstruction and/or migration of the catheter risk with a tendency towards lower risk but not statistical significance in meta-analysis.¹⁴ Due to low numbers in the subgroups of non-infectious complications, adjusting for these underlying risk factors was not possible in current study. This however exemplifies the need for further research to identify risk factors for non-infectious complications related to PD catheters and insertion, as little is known other than that reported in clinical trials looking at insertion technique and catheter type.^14
–16

Reduced time on PD therapy following early non-infectious complications was most pronounced within the first 180 days. This temporally matches what is seen with risk of loss of first PD catheter correlating with need for intervention. Data on the specific nature of catheter and functional complications with their potential interactions such as catheter tip migration and flow obstruction are not captured at present by the NZPDR. The short-term nature of higher risk of either outcome due to non-infectious complications is both logical and plausible. These complications may result in PD discontinuation due to requiring intervention for causes which may not be fully amendable to intervention.^12,30 Although radiological manipulation for flow dysfunction has been reported having a 46–75% success rate and laparoscopic rescue procedures have reported 63–100% long-term success rate,³⁰ these data are not currently recorded by the NZPDR. Repetitive complications and need for intervention may also influence an individual to transition to haemodialysis. This may be the explanation for why the reduced time on PD therapy was driven by catheter and functional complications rather than tunnel leaks. Early leaks are usually managed conservatively and do not often require surgery or removal of the catheter.¹²

It was observed as part of the multivariate analysis that renal physician inserted PD catheters may be associated with higher risk of first catheter removal and first tunnel infection episode. This may be driven by how nephrologists inserted catheters, which was still a relatively new practice in New Zealand during this study period time, with the need to consider operator expertise development. An earlier study demonstrated similar first catheter 1 year and overall survival comparing peritoneoscopic (nephrologist) PD catheter insertion to surgical insertion but higher perioperative complications and poorer complication-free survival.³¹

The major strength of this study is the prospective multicentre approach for data collection with a large cohort over a significant period of time. Almost all individuals in New Zealand who commence PD during the study time period were included with only 60 individuals missing when compared to the Australia and New Zealand Dialysis and Transplant Registry over the same 5-year time period.³² Given the inclusion of those transferring for haemodialysis and following kidney transplantation failure, these results are representative of the New Zealand Peritoneal Dialysis population as a whole.

Major limitations include the NZPDR being not set up to record differences of local centre practices and guidelines. Due to this we are unable to examine how this may have affected the results. This study has highlighted some important gaps in the data that is collected as has been discussed above. Feedback to the registry will hopefully lead to improved data collection.

The low incident rate of specific complications did not allow multivariable analyses and adjustments for possible confounding in some of the analyses here. Further, possible contributing factors that may need to be adjusted for but were not available for multivariate analysis include body mass index, previous abdominal surgery, co-morbidity data and treatment interventions. However known factors associated with reduced patient survival on PD (diabetes, advancing age, reduced residual kidney function) and PD discontinuation (diabetes, younger age) were analysed and adjusted for in the primary analysis. Other known associated factors with subsequent mortality, including cardiovascular and peripheral vascular disease, and prior kidney replacement were not available but lack biological plausibility as risk factors for non-infectious complications.^8,9,33
–35

The NZPDR definition of time on PD therapy being defined as number of days until transfer to haemodialysis therapy for more than 90 days does not capture individuals with mild and reversible complications that resulted in temporary transfer to haemodialysis, but overall continuation with PD therapy.

In conclusion, we demonstrated that early non-infectious complications result in significant increased risk of hard outcomes of mortality long term and PD discontinuation short term. This stresses the need for further research in risk factors and causes of early non-infectious complications for risk stratification and prevention purposes.

Supplemental material

Supplemental Material, sj-pdf-1-ptd-10.1177_08968608221132647 - Long-term impact of early non-infectious complications at the initiation of peritoneal dialysis

Supplemental Material, sj-pdf-1-ptd-10.1177_08968608221132647 for Long-term impact of early non-infectious complications at the initiation of peritoneal dialysis by R Chan, RJ Walker, A Samaranayaka and JBW Schollum in Peritoneal Dialysis International

Footnotes

Authors’ note

Dr Chan is currently based at the Department of Renal Medicine, Wellington Hospital, Wellington, New Zealand.

Acknowledgement

The authors acknowledge the help of Dr Gerald Waters, the Director of the New Zealand Peritoneal Dialysis Registry, for his help with accessing the data for this study.

Author contributions

JS, RW and RC were involved in the concept of the study. RC was responsible for obtaining the data from the NZPDR. RC and AS undertook the statistical analyses. RC wrote the first draft of the manuscript. RW, AS, JS and RC were involved in the reviewing and editing subsequent versions of the manuscript. All authors approved the final version of the manuscript.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Ethical approval

Not required as it is an audit from the NZ Peritoneal Dialysis Registry. The use of de-identified data from the NZPDR was approved by the NZPDR scientific committee.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

RJ Walker

Informed consent

None.

Supplemental material

Supplemental material for this article is available online.

References

PK-T

Chow

Van de Luijtgaarden

MWM

, et al. Changes in the worldwide epidemiology of peritoneal dialysis. Nat Rev Nephrol 2017; 13(2): 90–103.

ANZDATA Registry. 43rd Report, Chapter 5: Peritoneal dialysis. Australia and New Zealand Dialysis and Transplant Registry, Adelaide, Australia. 2020. http://www.anzdata.org.au (accessed 19 July 2022).

Fink

Marsh-Manzi

, et al. Changes in quality of life during hemodialysis and peritoneal dialysis treatment: generic and disease specific measures. J Am Soc Nephrol 2004; 15(3): 743–753.

Rubin

Fink

Plantinga

, et al. Patient ratings of dialysis care with peritoneal dialysis vs hemodialysis. JAMA 2004; 291(6): 697.

Heaf

Lkkegaard

Madsen

. Initial survival advantage of peritoneal dialysis relative to haemodialysis. Nephrol Dial Transpl 2002; 17(1): 112–117.

Choi

J-Y

Jang

Park

, et al. Survival advantage of peritoneal dialysis relative to hemodialysis in the early period of incident dialysis individuals: a nationwide prospective propensity-matched study in Korea. PLoS One 2013; 8(12): e84257.

Thiery

Séverac

Hannedouche

, et al. Survival advantage of planned haemodialysis over peritoneal dialysis: a cohort study. Nephrol Dial Transpl 2018; 33(8): 1411–1419.

Jaar

Plantinga

Crews

, et al. Timing, causes, predictors and prognosis of switching from peritoneal dialysis to hemodialysis: a prospective study. BMC Nephrol 2009; 10(1): 3.

See

Johnson

Hawley

, et al. Risk predictors and causes of technique failure within the first year of peritoneal dialysis: an Australia and New Zealand dialysis and transplant registry (ANZDATA) study. Am J Kidney Dis 2018; 72(2): 188–197.

10.

Hung

C-C

Chang

C-T

Lee

C-C

, et al. Prognostic predictors of technique and patient survival in elderly Southeast Asian individuals undergoing continuous ambulatory peritoneal dialysis. Int J Clin Pract 2009; 63(2): 254–260.

11.

Chaudhary

. Peritoneal dialysis drop-out: causes and prevention strategies. Int J Nephrol 2011; 2011: 434608.

12.

Crabtree

. Rescue and salvage procedures for mechanical and infectious complications of peritoneal dialysis. Int J Artif Organs 2006; 29(1): 67–84.

13.

Peppelenbosch

Van Kuijk

WHM

Bouvy

, et al. Peritoneal dialysis catheter placement technique and complications. NDT Plus 2008; 1(Suppl 4): iv23–iv28.

14.

Zang

Qiu

Yang

, et al. Different techniques for peritoneal dialysis catheter implantation: a systematic review and network meta-analysis. Perit Dial Int 2021; 41(6): 522–532.

15.

Hagen

Lafranca

Ijzermans

JNM

, et al. A systematic review and meta-analysis of the influence of peritoneal dialysis catheter type on complication rate and catheter survival. Kidney Int 2014; 85(4): 920–932.

16.

Shrestha

Kumar

, et al. Advanced laparoscopic peritoneal dialysis catheter insertion: systematic review and meta-analysis. Perit Dial Int 2018; 38(3): 163–171.

17.

Ranganathan

John

Yeoh

, et al. A randomized controlled trial to determine the appropriate time to initiate peritoneal dialysis after insertion of catheter (Timely PD Study). Perit Dial Int 2017; 37(4): 420–428.

18.

www.pdregistry.org.nz (accessed 19 July 2022).

19.

Grambsch

Thereau

. Proportional hazards tests and diagnostics based on weighted residuals. Biometrika 1994; 81(3): 515–526.

20.

Hsieh

Lavori

. Sample-size calculations for the cox proportional hazards regression model with nonbinary covariates. Control Clin Trials 2000; 21(6): 552–560.

21.

Elm

E von

Altman

Egger

, et al. The strengthening the reporting of observational studies in epidemiology (STROBE) statement: guidelines for reporting observational studies. J Clin Epidemiol 2008; 61(4): 344–349.

22.

Troidle

Gorban-Brennan

Kliger

, et al. Renal Research Institute Symposium: continuous peritoneal dialysis-associated peritonitis: a review and current concepts. Semin Dial 2003; 16(6): 428–437.

23.

Chen

JHC

Johnson

Hawley

, et al. Association between causes of peritoneal dialysis technique failure and all-cause mortality. Sci Rep 2018; 8(1): 3980.

24.

Béchade

Guittet

Evans

, et al. Early failure in individuals starting peritoneal dialysis: a competing risks approach. Nephrol Dial Transpl 2014; 29(11): 2127–2135.

25.

Ravani

Palmer

Oliver

, et al. Associations between hemodialysis access type and clinical outcomes: a systematic review. J Am Soc Nephrol 2013; 24(3): 465–473.

26.

Ravani

Quinn

Oliver

, et al. Examining the association between hemodialysis access type and mortality: the role of access complications. Clin J Am Soc Nephrol 2017; 12(6): 955–964.

27.

Palamuthusingam

Johnson

Hawley

, et al. Perioperative outcomes and risk assessment in dialysis individuals: current knowledge and future directions. Int Med J 2019; 49(6): 702–710.

28.

Tiong

Poh

Sunderaraj

, et al. Surgical complications of Tenckhoff catheters used in continuous ambulatory peritoneal dialysis. Singapore Med J 2006; 47(8): 707–711.

29.

Liu

Dong

. Urgent-start peritoneal dialysis complications: prevalence and risk factors. Am J Kidney Dis 2017; 70(1): 102–110.

30.

Crabtree

Shrestha

Chow

K-M

, et al. Creating and maintaining optimal peritoneal dialysis access in the adult patient: 2019 update. Perit Dial Int 2019; 39(5): 414–436.

31.

Sun

TYT

Voss

Beechey

, et al. Comparison of peritoneal dialysis catheter insertion techniques: peritoneoscopic, radiological and laparoscopic: a single-centre study: PD catheter insertion technique comparison. Nephrology 2016; 21(5): 416–422.

32.

ANZDATA Registry. 42nd Report, Chapter 5: Peritoneal dialysis. Australia and New Zealand Dialysis and Transplant Registry, Adelaide, Australia. 2020. http://www.anzdata.org.au (accessed 19 July 2022).

33.

Kolesnyk

Dekker

Boeschoten

, et al. Time-dependent reasons for peritoneal dialysis technique failure and mortality. Perit Dial Int 2010; 30(2): 170–177.

34.

Kang

Cho

Park

, et al. Risk factors for mortality in stable peritoneal dialysis patients. Renal Failure 2012; 34(2): 149–154.

35.

Hayat

Saweirs

Predictors of technique failure and mortality on peritoneal dialysis: an analysis of New Zealand Peritoneal Dialysis Registry data. Nephrology 2021; 26(6): 530–540.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.08 MB