Introduction
Patients with end-stage kidney failure essentially require renal replacement therapy. Three forms of renal replacement therapies commonly used are peritoneal dialysis (PD), hemodialysis, and kidney transplantation. There are two PD modalities, including continuous ambulatory PD (CAPD) and automated PD (APD). CAPD is performed by manual dialysate exchanges, and APD is performed by automatic dialysate exchanges using a machine. In contrast to CAPD, APD offers some advantages, such as a dry abdomen during daytime and short-time dialysate exchange. Despite concerns such as insufficient dialysis adequacy or maintenance of residual renal function (RRF), the proportion of APD is increasing owing to advances in APD technologies [1].
Icodextrin, a glucose polymer used in PD, helps improve fluid removal for patients with high peritoneal solute transfer rates [2]. Icodextrin also aids in maintaining stable ultrafiltration through long dwellings in the daytime in patients undergoing APD. Survival benefits associated with the use of APD or icodextrin have been assessed. Luo et al. [3] performed a meta-analysis of the survival benefits of APD and showed favorable outcomes in APD compared with those from CAPD, whereas Goossen et al. [4] evaluated the survival benefits of using icodextrin. However, in previous studies, the use of APD or icodextrin has been defined within a limited timeframe, such as the use of ≥90 days or at the time of baseline. In addition, the clinical benefits of icodextrin used during APD and CAPD differed. Therefore, despite favorable outcomes from previous studies, there is no conclusive evidence regarding the survival benefits of APD or the use of icodextrin. Hence, further studies are needed to confirm their efficacy based on the four groups based on the PD modality and icodextrin use. This study aimed to evaluate patient and technique survival among four groups divided based on PD modality and icodextrin use over 1 year.
Methods
This study received ethical approval from the Institutional Review Board of Yeungnam University Medical Center (No. YUMC 2020-01-025-001) and was conducted according to the principles of the World Medical Association Declaration of Helsinki. Informed consent was not obtained from the patients because their records and information were anonymized and de-identified before the analysis.
Study population
Our study employed a retrospective cohort design. The inclusion criteria were patients with prevalent PD who underwent a single PD modality for ≥1 year between November 2011 and October 2013. To ensure data accuracy, we excluded patients with insufficient information, those using non-icodextrin during daytime dwells while on APD, and those using icodextrin for <1 year after enrollment. Consequently, we excluded all patients with APD with daytime dwells using a glucose-based dialysate to focus on the independent effects of APD or icodextrin. None of the patients undergoing APD had a daytime dwell period except for icodextrin. Finally, 148 prevalent patients undergoing PD were included in our final analyses.
The participants were categorized into four groups for comparison: CAPD–ET, comprising patients on CAPD without icodextrin; CAPD+ET, including patients on CAPD using icodextrin; APD–ET, encompassing patients on APD without icodextrin; and APD+ET, comprising patients on APD with icodextrin.
Baseline characteristics
We obtained baseline data using a regular peritoneal membrane equilibration test on the day closest before the day of enrollment. Baseline data included age; sex; the presence of diabetes mellitus (DM); dialysis vintage (yr); body mass index (BMI, kg/m2); levels of serum albumin (g/dL), serum calcium (mg/dL), phosphorus (mg/dL), and hemoglobin (g/dL); RRF (mL/min/1.73 m2), weekly Kt/Vurea measure; levels of sodium (mmol/L), potassium (mmol/L), and chloride (mmol/L); dialysate-to-plasma creatinine concentration ratio 4-hour (DP4Cr); ultrafiltration volume (mL/day); urine volume (mL/day); and edema index. We also evaluated the etiology of end-stage kidney disease and the number of patients with coronary artery disease, heart failure, or cerebrovascular accidents.
All laboratory studies were performed following an overnight fast. The presence of DM was defined based on the patient’s history and medical records of DM diagnosis or medications. The RRF was calculated based on the creatinine and urea nitrogen excretion values from 24-hour urine collection as previously defined [5]. Weekly Kt/Vurea was calculated using 24-hour urine and dialysate, as previously published [6]. DP4Cr was evaluated using a 4.25% peritoneal equilibration test, and the level was calculated using the creatinine level of the drained dialysate 4 hours after injection as per the blood creatinine level. The edema index was defined as the extracellular water/total body water ratio obtained using bioimpedance measurements (InBody; InBody Co., Ltd.). The peritoneal equilibration test and edema index were performed annually in our center. The edema index, ultrafiltration volume, urine volume, RRF, and weekly Kt/Vurea were measured after 1 year of taking the baseline measurements.
We defined the starting point as the time of the first PD prescription for the same PD modality lasting ≥1 year. We collected data on baseline characteristics, including dialysis vintage (the time from the start of PD to the time of the first PD prescription for the same PD modality lasting ≥1 year). In addition, we performed a survival analysis using the duration from the starting point. Therefore, the patient and technique survival rates were 100% for the first year in all patients enrolled in this study; data on any changes in PD modality after the first year were also collected. Data were collected on volume status and analyzed compared to the data collected at the first PD prescription that maintained the same PD modality for 1 year afterward, including RRF, ultrafiltration volume, urine volume, weekly Kt/Vurea, and edema index. We evaluated both patient and technique survival. All patients were scheduled for follow-up visits in November 2021. Patient death was defined as an event, irrespectively the cause, marking the endpoint of the follow-up period. If the patient had kidney transplantation, transfer to hemodialysis (for ≥90 consecutive days), cessation of dialysis owing to renal recovery, loss of follow-up, or transfer to another hospital, the data were considered censored. Technique failure was defined as a patient’s death or transfer to hemodialysis for ≥90 consecutive days [7]. The data were considered as censoring, if the patient had kidney transplantation, shifted to hemodialysis consequent to the patient’s request without any medical indication, cessation of dialysis because of renal function recovery, loss of follow-up, or transfer to another hospital.
Statistical analysis
Data were analyzed using IBM SPSS (version 25.0; IBM Corp.). Categorical variables were expressed in terms of count (percentage) and were analyzed using the Pearson chi-square or Fisher exact test. Continuous variables were evaluated for distribution using the Kolmogorov-Smirnov test, and none of the continuous variables showed normal distribution. Therefore, the continuous variables were presented as the median (interquartile range [IQR]). Continuous variables were compared using the Kruskal-Wallis test, followed by Bonferroni post-hoc or the Mann-Whitney U tests for comparison between the two groups. Changes during 1 year were compared using the Wilcoxon signed-rank test. We used Kaplan-Meier analysis to plot the survival among the groups and the log-rank method to determine statistical significance. The start point of follow-up was defined as the time of the first PD prescription for the same PD modality lasting ≥1 year between November 2011 and October 2013. The end point of follow-up was November 2021. The total follow-up duration was calculated to include 1 year of consistent PD prescription and was defined as the period until survival, censoring, or death. Although the final follow-up point is more than 5 years from the start date, most patients were censored or died within 5 years, reducing the number of patients. Therefore, the Kaplan-Meier curve was illustrated for the period within 5 years. A p-value of <0.05 was considered statistically significant.
Results
Clinical characteristics of participants
The numbers of patients in the CAPD–ET, CAPD+ET, APD–ET, and APD+ET groups were 39, 35, 40, and 34, respectively (Table 1). The median (IQR) ages of the participants in the CAPD–ET, CAPD+ET, APD–ET, and APD+ET groups were 46 (39-56), 47 (41-54), 44 (34-54), and 49 (42-55) years, respectively (p = 0.41). The numbers of male patients in the CAPD–ET, CAPD+ET, APD–ET, and APD+ET groups were 28 (71.8%), 30 (85.7%), 30 (75.0%), and 30 (88.2%), respectively (p = 0.22). No significant differences were observed in age, sex, the presence of DM, dialysis vintage, BMI, serum albumin levels, calcium levels, phosphorus levels, hemoglobin levels, sodium levels, potassium levels, or chloride levels among the four groups. The APD+ET group participants showed greater DP4Cr than APD–ET participants. No significant differences were found among the four groups in the etiology of end-stage kidney disease, proportion of coronary artery disease, heart failure, or cerebrovascular accidents.
We evaluated the total dialysate volume, time, number of exchanges, and glucose concentration in the dialysate during APD, except during daytime dwell. The median (IQR) of the total dialysate volumes in the APD–ET and APD+ET groups were 9 (7.7–10) and 10 (5.3–10) L (p = 0.35). Moreover, the durations were 10 (9–10) and 10 (9–10) hours (p = 0.85), with 4 (3–4) and 4.5 (3–5) exchanges (p = 0.36), and glucose concentrations of 1,500 (1,500–1,740) mg/dL and 1,500 (1,500–2,000) mg/dL (p = 0.61), respectively. No significant differences were found between the two groups for these values.
Patient or technique survival according to groups
At 5 years of follow-up, 29 (74.4%), 29 (82.9%), 25 (62.5%), and 26 (76.5%) patients survived with PD; 5 (12.8%), 1 (2.9%), 7 (17.5%), and 4 (11.8%) died; 2 (5.1%), 4 (11.4%), 5 (12.5%), and 1 (2.9%) shifted to hemodialysis; and 3 (7.7%), 1 (2.9%), 3 (7.5%), and 3 (8.8%) had kidney transplantation in the CAPD–ET, CAPD+ET, APD–ET, and APD+ET groups, respectively (p = 0.46). The median (IQR) follow-up durations in the CAPD–ET, CAPD+ET, APD–ET, and APD+ET groups were 87 (54-128), 88 (63-105), 61 (36-86), and 68 (47-82) months, respectively. At 5 years of follow-up, the causes of death in the CAPD–ET group were cardiac arrest (n = 3), infection (n = 1), and unknown causes (n = 1). In the CAPD+ET group, one death occurred from cardiac arrest. Deaths in the APD–ET group were from cardiac arrest (n = 2), infection (n = 2), heart failure (n = 1), and unknown causes (n = 2). Deaths in the APD+ET group resulted from cardiac arrest, cerebral hemorrhage, infection, and unknown causes (n = 1, each; p = 0.85). No significant differences were observed in the cause of death among the four groups.
The Kaplan-Meier curves are shown in Fig. 1. The CAPD+ET group had a higher patient survival rate than that of the APD–ET group (p = 0.03). The CAPD+ET group also had a higher technique survival trend than that of the APD–ET group, despite no statistical significance (p = 0.08). Fig. 2 shows the Kaplan-Meier curves of the subgroups in the presence of DM. In patients with DM, there was no significant difference in patient survival among the four groups. Although no statistical significance was observed for patients without DM as well, the APD–ET group had a poorer patient survival trend than those of the APD+ET or CAPD+ET groups (p = 0.05 for the APD+ET group and p = 0.06 for the CAPD+ET group). Evaluation of the edema index based on the presence of DM in the four groups (Supplementary Table 1, available online) showed that the edema index of the APD–ET group improved over 1 year in patients with DM.
There was no significant difference in technique survival among the four groups of patients with DM. In patients without DM, the APD+ET group had a higher technique survival than the APD–ET group (p = 0.03). In addition, the APD+ET group showed a higher technique survival trend than did the CAPD–ET group, despite non-statistical significance (p = 0.07). The patients transferred to hemodialysis had peritonitis (n = 2) in the CAPD–ET group, peritonitis (n = 2) and inadequate dialysis (n = 3) in the CAPD+ET group, inadequate dialysis (n = 3) and gastrointestinal trouble (n = 2) in the APD–ET group, and peritonitis (n = 1) in the APD+ET group. The percentages of technique failure cases, excluding deaths, were 5.9%, 11.8%, 15.2%, and 3.3% in the CAPD–ET, CAPD+ET, APD–ET, and APD+ET groups, respectively (p = 0.34). Although the statistical significance was low, the proportion was the highest in the APD–ET group. The modality changes or additions of daytime dwells after 1 year of baseline are summarized in Supplementary Table 2 (available online).
Comparisons of factors associated with volume status
The measurements for RRF, weekly Kt/Vurea, edema index, ultrafiltration volume, and urine volume at baseline and after 1 year of follow-up are shown in Table 2. None of the variables except the edema index differed among the four groups significantly after 1 year of follow-up. The edema index after 1 year of follow-up was higher in the APD–ET group than in the other groups. No statistically significant differences were found in the baseline ultrafiltration volume between the CAPD+ET and other groups (p = 0.06 vs. APD–ET, p = 0.50 vs. APD+ET; p = 0.06 vs. CAPD–ET). The CAPD+ET group had a smaller ultrafiltration volume than the other three groups, but the differences were not significant. In addition, we found no significant difference among the four groups in the baseline urine volume compared to the baseline ultrafiltration volume.
The delta change of the edema index did not differ significantly in this 1 year (Table 3). Ultrafiltration volume increased during 1 year, while urine volume, RRF, and weekly Kt/Vurea showed decreasing trends. Although the increase in the ultrafiltration volume was lower in the APD–ET group than the other groups, particularly in those of the APD+ET or CAPD+ET groups; however, the difference was not statistically significant. The urine volume in the CAPD+ET group was maintained for 1 year.
We analyzed the relationship between the mean edema index over 1 year and patient or technique survival by dividing the patients into low, middle, and high groups based on the mean edema index over 1 year. The survival curves are shown in Supplementary Fig. 1 (available online). The patients with a high edema index were associated with poor patient and technique survival compared with the other two groups.
Discussion
We included 148 prevalent patients undergoing PD and evaluated the patient and technique survival rates and the indicators associated with volume status during 1 year. Our study showed that the patients in the APD–ET group patient had poorer patient and technique survival trends than patients with icodextrin. These trends were clearer in patients without DM than those with DM. The edema index at 1 year of follow-up was higher in the APD–ET group than in the other groups. The APD–ET group showed a lower trend in ultrafiltration volumes over 1 year than did the other groups, particularly those using icodextrin. Urine volumes were relatively stable in the CAPD+ET group over the 1-year study period.
We compared the patient and technique survival rates of APD and CAPD, and the results revealed that the two survival rates differed according to the use of icodextrin. The APD+ET group had similar patient and technique survival rates as those of the CAPD groups with or without icodextrin; however, the APD–ET group had poorer patient and technique survival trends than that of the CAPD+ET group. Previous studies have shown favorable or better patient or technique survival rates when using APD than when using CAPD. Beduschi et al. [8] included patients undergoing PD in Brazil and performed propensity-matching analyses. They showed better all-cause mortality associated with APD than that with CAPD and similar technique survival between the two modalities. Li et al. [9] studied patients undergoing PD in China and found better patient survival associated with APD than with CAPD. A study from Taiwan showed better technique survival with APD than with CAPD and similar patient survival between the two modalities [10]. However, Tang et al. [11] showed poorer patient and technique survival rates in APD than in CAPD using an old cohort analysis. A recent meta-analysis of 19 studies concluded that APD exhibited better associated patient survival but similar technique survival than those by CAPD [4]. Contrary to the favorable outcomes of APD from previous studies, our study indicated poor outcomes associated with APD. This may be attributed to two reasons. First, patients who underwent APD in most previous studies were categorized based on the modality at enrollment or having utilized APD ≥90 days. Some of these patients may have been subjected to CAPD as well, and the alternative use of APD and CAPD may have influenced the results. Second, subgroup analysis by using icodextrin in those studies may have yielded results comparable to ours. In our study, the APD+ET group had comparable outcomes with those of the CAPD groups and only the APD–ET group had a poorer outcome than that of the CAPD+ET group.
Poor outcomes of the APD–ET group may be associated with poor volume status. Patients undergoing APD perform dialysis for a relatively limited time compared with those undergoing CAPD. Patients with sufficient RRF would necessitate relatively minor ultrafiltration in those undergoing APD; however, RRF decreases as the duration of PD increases, which could be associated with the difficulty in maintaining APD without daytime dwellings. In our study, the APD–ET group had poor volume control than those of the groups using icodextrin owing to a limited increase in ultrafiltration volume, and the CAPD–ET participants also had a relatively limited increase in ultrafiltration volume compared with those of the patients using icodextrin. These findings reveal that incorporating additional daytime dwell, particularly with the use of icodextrin, at the optimal time in the APD–ET group could be beneficial for maintaining appropriate volume status. The use of icodextrin is useful in maintaining favorable volume status, which is associated with improved patient and technique survival compared to non-icodextrin users [12,13]. A recent meta-analysis of 19 studies also showed that patients using icodextrin had better ultrafiltration volume, fluid control, and patient survival than those patients not using icodextrin [4].
Our study found a tendency towards greater technique failure in patients with APD, particularly those who did not use icodextrin. Several studies comparing technique failures between APD and CAPD have shown inconsistent results and conclusions, with no clear mechanism identified. Therefore, various inconsistent and often contradictory hypotheses have been proposed to explain the study results. When APD is favorable, the results are explained by the relatively lower glucose absorption compared with CAPD or by the fewer comorbidities, younger age, and greater quality of life and social activity in APD patients [14–17]. Conversely, when CAPD shows a better prognosis than APD, the results are attributed to less solute removal (sodium, phosphorus, and fluid) and a faster decline of RRF in APD [18–20]. Limited by the sample size, our study did not statistically confirm the superiority of either CAPD or APD, emphasizing the need for caution, particularly in APD, for using icodextrin, which mediates volume effects.
In the CAPD+ET group, compared to the other groups, a relatively small ultrafiltration volume, similar urine volume, and stable volume status would be associated with a small fluid intake. The fluid intake in the CAPD+ET group increased similarly to other groups; presumably, the increased ultrafiltration volume occurred to maintain stable volume status. Similarly, an increase in ultrafiltration volume beyond the 1-year decrease in urine volume may have contributed to maintaining a stable volume in the APD+ET group.
APD without daytime dwell would be more applicable in patients with sufficient RRF than in the other groups. Our study included patients maintained on a single modality with a median of 2 to 3 years of dialysis vintage. Despite volume overload, several points should be considered for maintenance of APD without daytime dwell. First, the actual edema index values in patients with APD–ET may not align completely with clinical findings. Therefore, these patients may not have had clinically evident volume overload symptoms or signs that would necessitate daytime dwelling. Typically, an increase in PD is considered when clinical symptoms and signs are present. However, if only the edema index had been considered, a daytime dwell might have been added, regardless of symptoms. Second, despite having symptoms of volume overload, some patients did not receive additional daytime dwells for various reasons, including resistance to daytime dwells. As a result, the APD–ET group may have included patients with relative volume overload who needed an additional daytime dwell but did not receive it, possibly causing selection bias, an important limitation when assessing the independent effects of APD and icodextrin. A controlled prospective study involving patients with similar baseline volumes is necessary to address these limitations.
Subgroup analysis based on the presence of DM showed that among the patients without DM, the APD+ET group showed better patient and technique survival rates than the APD–ET group. However, no significant differences were observed between the two groups of patients with DM, possibly because of differences in the volume-control effect of APD or icodextrin, depending on the presence of DM. In patients with DM, the 1-year volume control in the APD–ET group was not poor. Patients with DM may have greater DP4Cr levels than those without DM, indicating that effective volume control is achievable with APD alone, possibly attenuating the effect of icodextrin. Conversely, in patients without DM, the volume-control capability of APD was less than that in patients with DM, suggesting that adding icodextrin resulted in more effective volume control. This result may explain why the effect of icodextrin was more pronounced in patients without DM who underwent APD.
Differences may occur in outcomes between patients with and without modality changes or the addition of a daytime dwell within a group. However, most patients in our study maintained the same modality during the follow-up period. Therefore, subgroup analyses using patients with modality changes or the addition of daytime dwells were difficult owing to an insufficient sample size. Comparison of outcomes between patients with and without a modality change or the addition of daytime dwells would be interesting; additional studies using sufficient data would help identify outcomes differences between groups.
The statistical significance in our study appears greater based on volume status rather than PD modality or icodextrin use. This suggests that volume control may largely influence the differences in prognosis due to PD modality or icodextrin. Thus, PD modality and icodextrin use may be seen as mediators influencing volume status rather than as main factors affecting prognosis. However, considering the impact of PD modality and icodextrin use on the peritoneal membrane, their metabolic benefits, and differences in quality of life, these factors should be considered through more diverse analyses rather than attributing their effects solely to volume control. This requires a larger number of patients or randomized controlled studies. Additionally, various subgroup analyses using patients with the same volume status could help identify volume-independent effects and their influences on PD modality and icodextrin use.
The study has several limitations. First, our study used a single-centered and retrospective design. Second, it included a small sample size, which limited the multivariate analyses using adjustment for covariates and subgroup analyses according to various clinical indicators except the presence of DM. Our study did not find statistically significant differences because of an insufficient number of patients, resulting in most findings indicating trends only. This result limits drawing robust conclusions and mandates caution during interpretation. Because this study was retrospective, the number of patients was limited to the unexpectedly few cases in which a single modality was maintained for ≥1 year. However, our study provides important information, such as the required sample size and volume status, for conducting future prospective studies on modality, icodextrin use, and treatment outcomes, positioning it as a pilot study. Third, it did not consider diet or fluid intake. Urine volume, RRF, or ultrafiltration volume can be influenced by diet or fluid intake, and the lack of data for these indicators makes it difficult to identify the independent efficacy of variables associated with volume status. Prospective studies, including multivariate analyses using more patients, are needed to overcome these limitations.
The present study demonstrated that patients undergoing APD without icodextrin had poor patient and technique survival trends, which may be caused by poor volume control. Therefore, the addition of icodextrin at optimal times in patients undergoing APD may be useful to maintain the proper volume status and improve clinical outcomes. However, our study has limitations due to the small sample size and limited follow-up time, making conclusive results difficult. Therefore, our findings must be interpreted with caution. Further studies with larger sample sizes and longer follow-up are necessary to provide additional evidence of the effects of icodextrin and APD observed in our results.
PDF Links
PubReader
ePub Link
XML Download
Full text via DOI
Download Citation
Supplement table 1
Print
