Long-term sonographic evaluation of the peritoneum in peritoneal dialysis patients: A retrospective cohort study

Long-term peritoneal dialysis (PD) causes morphological changes to the peritoneum. However, the sequential morphological changes of the peritoneum remain unclear due to the invasiveness and ethical dilemmas surrounding peritoneal biopsies. We aimed to evaluate these long-term morphological peritoneal changes using sonography, which was recently reported to be useful for morphological peritoneal evaluation. Methods We retrospectively identified 115 PD patients who underwent sonographic peritoneal membrane thickness (PMT) measurement. Univariate and multivariate linear regression analyses identified factors related to PMT at baseline (bPMT), at last measurement (lPMT), and the PMT change rate. Of the 115 patients, 42 patients had at least two PMT measurements, including a bPMT measurement. We evaluated the PMT change between bPMT and lPMT. We also evaluated the annual PMT change for 3 years before PD withdrawal in patients who discontinued PD due to peritoneal dysfunction. Clinical characteristics and parameters were analyzed according to PMT change rates (≤ 0 [n = 28] or > 0 [n = 20]). the morphological changes that occur in the peritoneum during PD treatment and examine whether these changes were related to peritoneal function and PD prognosis by long-term observation of the peritoneum using sonography.

evaluation of the peritoneum by sonography will enable the recognition of transition in peritoneal function in real time and allow for more appropriate PD management. Furthermore, the peritoneum was not necessarily thickened regardless of PD duration or cause of withdrawal.

Background
Peritoneal membrane dysfunction, including ultrafiltration and solute transport problems, is one of the major causes of withdrawal from peritoneal dialysis (PD) therapy [1,2]. Moreover, ultrafiltration failure not only shortens PD duration but also worsens patient survival [3]. The peritoneum is thought to sustain injury with long-term PD therapy, uremia, long-term exposure to bio-incompatible dialysis solutions, and frequent episodes of peritonitis [4][5][6]. These factors cause structural and functional changes to the peritoneum, leading to peritoneal membrane dysfunction. Histological studies have shown mesothelial denudation, interstitial fibrosis, and vasculopathy in the peritoneum of patients on long-term PD [5,7,8]. However, since peritoneal biopsies are invasive and difficult to repeat due to ethical issues, the sequential morphological changes of the peritoneum associated with PD treatment remain unclear.
Faller et al. demonstrated the effectiveness of abdominal sonography in the morphological evaluation of the peritoneum. They measured the peritoneal membrane thickness (PMT) of 131 healthy children using transabdominal ultrasonography and showed that the variation coefficient for the measurement was only 5%; the PMT was significantly correlated with age, weight, and height; and peritoneal thickening was observed in PD patients who had a history of peritonitis [9]. Subsequently, several studies have reported that PMT, measured with sonography, correlates with peritoneal permeability [10][11][12][13], and sonography is expected to be an alternative method for peritoneal evaluation. However, although sonography is a simple and non-invasive method, evaluation of the peritoneum with sonography is limited to cross-sectional studies, and there have been no reports on the sequential evaluation of the peritoneum over time during PD treatment.
The aim of this study was to clarify the morphological changes that occur in the peritoneum during PD treatment and examine whether these changes were related to peritoneal function and PD prognosis by long-term observation of the peritoneum using sonography.

STUDY DESIGN AND POPULATION
We retrospectively reviewed the clinical records of patients at Shonan Kamakura General Hospital between January 2008 and February 2019 and identified 115 PD patients who had a recorded sonographic PMT measurement. Patients who had received a kidney transplantation, hemodialysis treatment for more than one month, or underwent any abdominal surgeries before the start of PD treatment were excluded. Written informed consent was waived due to the retrospective nature of the study and anonymity of the data. However, participants were able to opt-out of the study on the hospital's website. This study complied with the standards of the Declaration of Helsinki and was approved by the Ethics Committee for Human Research of Shonan Kamakura General Hospital (approval number: TGE01288-024).

CLINICAL DATA COLLECTION
Clinical and laboratory data were obtained by reviewing medical records. Patient characteristics included age, sex, height, weight, body mass index (BMI), PD duration, etiology of end-stage renal disease (ESRD), comorbidities, and cause of PD discontinuation. Biological data (serum albumin, creatinine [Cr], estimated glomerular filtration rate, C-reactive protein, calcium, phosphate, uric acid, β 2 -microglobulin, low-density lipoprotein-cholesterol, and urinary volume), concomitant medication, dialysis prescription, transport characteristics, and PMT were evaluated at the start of PD treatment and annually until PD discontinuation or the start of combination therapy with PD and hemodialysis.
We evaluated the transport characteristics of the peritoneum using the standard peritoneal equilibration test, calculating the dialysate to plasma ratio of creatinine (D/P Cr) after a 4-h dwell.

PERITONEAL MEMBRANE THICKNESS MEASUREMENT
We examined the parietal peritoneum of PD patients using transabdominal ultrasonography. The peritoneal membrane was defined as the hyperechoic line surrounding the abdominal cavity ( Fig. 1).
We indwelled at least 1,000 mL of dialysate and measured the PMT on three different ventral windows with PD patients in the supine position, according to the study by Faller et al. [9]. Horizontal sections were made at the midclavicular line bilaterally in the upper and lower quadrants of the abdomen without the catheter port. We calculated the mean of the three measurements and used it in the analysis. All PMT measurements were taken with an 8.0 MHz linear probe (Canon Medical systems Corporation, Tokyo, Japan) and were performed at the start of the peritoneal equilibration test.

OUTCOMES AND STATISTICAL ANALYSIS
Data are shown as means ± standard deviations or medians (minimum-maximum) and percentages, as appropriate. To identify factors related to the PMT at baseline (measured within 4 weeks of the introduction of PD) (bPMT), at last measurement (lPMT), and the PMT change rate, univariate and multivariate linear regression analyses were performed. In the multivariate analysis, candidate variables, including age, sex, BMI, PD duration, having diabetes mellitus (DM), history of peritonitis, and D/P Cr, were selected by referencing previous studies [9][10][11][12][13][14][15][16]. Of the 115 patients, 42 patients had at least two PMT measurements, including a bPMT measurement. We evaluated the PMT change between bPMT and lPMT. We also evaluated the annual PMT change for 3 years before PD withdrawal in patients who discontinued PD due to peritoneal dysfunction. To compare the PMT at various points during PD treatment, the paired t-test and repeated-measures ANOVA were performed. The patients were divided into two groups according to the PMT change rate: PMT change rate ≤ 0 (n = 22) or > 0 (n = 20). We compared patients' characteristics and clinical parameters, including the PMTs of the two groups, using student's t-test (for normally distributed variables), the Mann-Whitney U test (for nonnormally distributed variables), and Fisher's exact test (for categorical variables). A p-value < 0.05 was considered to be significant. Statistical analyses were performed using R software version 3.5.2 (R Foundation for Statistical Computing, Vienna, Austria).

FACTORS CORRELATED WITH PERITONEAL MEMBRANE THICKNESS
In the univariate linear analysis, there was a significant correlation between D/P Cr and lPMT (r = 0.386, p = 0.004), but not bPMT (r=-0.114, p = 0.326) (Fig. 2, Table 2). BMI was significantly correlated with bPMT (r = 0.009, p = 0.047). Having DM tended to correlate with the bPMT (r = 0.057, p = 0.093). The duration of PD, use of a dialysate containing icodextrin and high-concentration glucose, and the peritonitis rate were not significantly correlated with lPMT or the PMT change rate (Table 2). In the multivariate analysis, D/P Cr remained the only independent predictor of lPMT (r = 0.478, p = 0.001) after adjusting for age, sex, BMI, PD duration, DM, and peritonitis rate (Table 3).  PMT CHANGE RATE Figure 3a shows the change in the PMT from baseline to the last measurement in all patients who had at least two PMT measurements, including a bPMT measurement (n = 42). The mean bPMT and lPMT measurements were 0.67 ± 0.15 mm and 0.69 ± 0.10 mm, respectively, with no statistical difference between these measurements (p = 0.49) (Fig. 3b). The mean D/P Cr at baseline and at last measurement were 0.72 ± 0.15 and 0.77 ± 0.09, respectively, with no statistical difference between these values (p = 0.11) (data not shown). Of the 42 patients, seven patients had discontinued PD due to peritoneal dysfunction. The mean bPMT and lPMT in these patients were 0.69 ± 0.10 mm and 0.67 ± 0.06 mm, respectively, with no significant difference between measurements (p = 0.68) (Fig. 3c). In patients who discontinued PD due to peritoneal dysfunction, the mean PMTs at 3, 2, and 1 year before PD withdrawal were 0.67 ± 0.13 mm, 0.66 ± 0.11 mm, and 0.67 ± 0.08 mm, respectively, with no significant differences between these measurements (p = 0.967) (Fig. 3d).

COMPARISON BETWEEN TWO GROUPS DIVIDED BY PMT CHANGE RATE
There were no significant differences in PD duration, the use of a dialysate containing more than 2.5% glucose or icodextrin, and the peritonitis rate between the groups divided by the PMT change rate (Additional File 1). The bPMT was higher in the group with a decreased PMT over time than in the group with an increased PMT over time (0.76 ± 0.14 mm versus 0.57 ± 0.09 mm, p < 0.001); however, there was no difference in the lPMT between the groups (0.67 ± 0.12 mm versus 0.70 ± 0.07 mm, p = 0.184).

Discussion
In this study, we showed that the PMT and peritoneal permeability, represented by D/P Cr, were significantly correlated, and PMT was not necessarily increased with an increase in PD duration. To our knowledge, this study is the largest study evaluating the peritoneum by sonography and is the first study in which the PMT was repeatedly measured over time during PD treatment.
Duman et al. measured the PMT of 42 PD patients who had undergone PD for > 12 months using sonography and showed, for the first time, that the PMT was positively correlated with D/P Cr [10]. Subsequently, the correlation between PMT and peritoneal permeability has been reported in several studies using sonography [11][12][13]. Honda et al. showed in their study, which evaluated peritoneal specimens of 253 patients before and during PD treatment, that the PMT increased and the lumen/vessel diameter ratio, which was reported as an indicator of vasculopathy in their study, decreased with an increase in PD duration [8]. They also showed that the PMT was significantly higher in patients with ultrafiltration failure than in patients with a preserved ultrafiltration capacity.
Plum et al. showed that patients characterized as being high transporters, according to the peritoneal equilibration test, had an increased submesothelial fibrous layer thickness [17]. Based on these pathological studies, PMT may reflect vasculopathy or fibrosis of the peritoneum.
Conversely, bPMT tended to correlate with BMI and the presence of DM, but not with D/P Cr in this study. A previous morphological study evaluating the peritoneum of PD patients with or without DM at the start of PD showed that the submesothelial connective tissue thickness and the number of capillaries were significantly greater in the DM group than in the non-DM group [16]. However, to date, there have been no studies evaluating the correlation between PMT and peritoneal permeability at the start of PD using either peritoneal specimens or sonography.
There were no significant differences between bPMT and lPMT in all patients. These measures were not significantly different even among patients who discontinued PD due to peritoneal dysfunction. Moreover, the annual PMT, measured for 3 years prior to PD withdrawal, was preserved, contrary to our hypothesis that the PMT would be higher in patients who had discontinued PD due to peritoneal dysfunction. Inconsistent with the findings of our study, several previous studies have shown that PMT increased with a longer PD duration [8,10,[12][13][14]; however, these studies had cross-sectional study designs. Williams et al. evaluated the peritoneal biopsy specimens of 130 PD patients and showed that the thickness of the submesothelial compact zone increased significantly with an increase in PD duration [5]. However, these findings were not observed in patients without PD-related problems or membrane failure. Similarly, Lee et al. showed, based on sonographic examination, that the PMT was not associated with PD duration [11].
In Japan, since around 2005, most PD patients have used a pH-neutral, low-glucose degradation product (GDP) dialysate, which is more biocompatible than conventional dialysates. Since then, several studies have shown that a pH-neutral low-GDP dialysate might prevent morphological and functional changes in the peritoneum [15,[18][19][20]. Tawada et al. compared the peritoneal biopsy samples of 54 PD patients treated with a conventional acidic dialysate to those of 73 PD patients treated with a pH-neutral low-GDP dialysate [15]. They showed that the PMT was higher and the lumen/vessel diameter ratio was smaller in the conventional dialysate group than in the pH-neutral dialysate group and that the lumen/vessel diameter ratio negatively correlated with PD duration and D/P Cr in the conventional dialysate group but not in the pH-neutral dialysate group. From these findings, they suggested that a pH-neutral low-GDP dialysate might prevent morphological and functional changes in the peritoneum. Among the 115 patients in this study, only seven patients started PD therapy before 2005, and most patients were treated with a pH-neutral dialysate from the onset of treatment. Based on these facts, one of the reasons as to why the PMT was not correlated with PD duration in this study might possibly be explained by the use of a biocompatible dialysate.
Several studies showed that inhibition of the renin-angiotensin-aldosterone system (RAAS) prevented the morphological deterioration of the peritoneum in animal models [21,22]; although, the effectiveness of RAAS inhibitors in protecting the peritoneal membrane function of PD patients has remained unclear [23]. In this study, a greater percentage of patients were taking RAAS inhibitors than that reported in a previous study [14], and this may have contributed to the preservation of the peritoneal membrane.
In this study, we were unable to identify factors related to the PMT change rate. The bPMT was higher in the group with a decreased PMT over time than in the group with an increased PMT over time. This might be partly explained by the percentage of patients with DM in both groups. We found that the complications of DM tended to correlate with the bPMT. Of those with a decreased PMT over time, 45.5% had DM, whereas 25.0% of patients in the group with an increased PMT over time had DM. However, this difference was not significantly different.
There were some limitations to this study. First, this study was a retrospective single-center study. Therefore, selection bias could not be completely eliminated. Second, the timing of the lPMT during PD treatment varied depending on the patients, because sonographic examinations were not necessarily conducted every year.
However, we evaluated the annual change in PMT for 3 years before PD withdrawal and found that the PMT was preserved over time. Third, the concentration of glucose in the dialysate was not adjusted for statistically, because detailed PD prescription data were not compiled in this study. However, the use of a dialysate containing more than 2.5% glucose was not correlated with lPMT or the PMT change rate in the univariate analysis.

Conclusions
Herein, we have demonstrated that the peritoneal thickness, measured by sonography, was positively correlated with peritoneal permeability. Repeated evaluation of the peritoneum using sonography may enable the recognition of the transition in peritoneal function in real time and allow for the more appropriate management of PD.
Furthermore, the peritoneum was not necessarily thickened regardless of the PD duration or cause of PD withdrawal, and this may be partly explained by the widespread use of a biocompatible dialysate. This study complied with the standards of the Declaration of Helsinki and was approved by the Ethics Committee for Human Research of Shonan Kamakura General Hospital (approval number: TGE01288-024). Written informed consent was waived due to the retrospective nature of the study and anonymity of the data. However, participants were able to opt-out of the study on the hospital's website.

Consent for Publication:
Written informed consent was waived due to the retrospective nature of the study and anonymity of the data.
However, participants were able to opt-out of the study on the hospital's website.

Availability of Data and Materials
All data generated or analyzed during this study are included in this published article [and its supplementary information files].  The change in peritoneal membrane thickness (a) Change in PMT between bPMT and lPMT measurements in all patients who had at least two PMT measurements, including a bPMT measurement (n=42). (b) Change in PMT among all patients (n=42). There is no significant difference (p=0.49). (c) Change in PMT in patients who discontinued PD due to peritoneal dysfunction. There is no significant difference (p=0.68). (d) Annual change in PMT 3 years before PD withdrawal in patients who discontinued PD due to peritoneal dysfunction. There are no significant differences between measurements (p=0.967). PMT, peritoneal membrane thickness; PD, peritoneal dialysis; bPMT, baseline peritoneal membrane thickness; lPMT, last measurement peritoneal membrane thickness Supplementary Files