Study population and data source
This retrospective cohort study was conducted at a single academic hospital. The study cohort included patients who initiated PD treatment between April 1, 2007, and March 31, 2022. Patients aged < 20 years at baseline, those who did not receive PD treatment for at least a month, or those who switched from maintenance HD to PD were excluded from the study. All data were obtained from the electronic records of the Kumamoto University Hospital. Due to the anonymity of the patients included and the nonintrusive nature of the research, the requirement for written consent was waived. This study was approved by the Institutional Review Committee of the Kumamoto University Hospital (No. 2529).
Demographic, clinical, and laboratory measures
We obtained data on age, sex, height, weight, blood pressure, urine volume, information on the primary disease of ESKD, comorbidities, cardiovascular disease (CVD) events, CVD history or death, PD prescription, PD treatment period and peritonitis, HD treatment period, the use of renin–angiotensin–aldosterone system inhibitors, statins, diuretics, phosphate binders, vitamin D, and erythropoiesis-stimulating agents, and laboratory results from electronic medical records. Weight and height were used to calculate the body mass index (BMI).
Most laboratory parameters measured were obtained from the nearest day of PD initiation, which was within a week before or after initiation, including the levels of hemogram, albumin, urea nitrogen, creatinine, and electrolytes. Serum levels of ferritin, iron, total iron binding capacity, lipids, plasma brain natriuretic peptide (BNP), and intact parathyroid hormone (PTH) were measured within a month before or after PD initiation. Hemoglobin (Hb) A1c levels were measured within 3 months. The PNI was calculated using the following formula: (10 × serum albumin [g/dL]) + (0.005 × total lymphocyte count [/mm3)] .
The outcome of interest was the discontinuation of PD. Based on previous reports, the discontinuation of PD was defined as a switch to HD or hybrid therapy with HD, kidney transplantation, or death [12, 13]. Censoring was performed at another dialysis clinic or at the end of the study.
Data are described as proportion, mean (± standard deviation, SD), or median (interquartile range, IQR), as appropriate. We divided the PNI levels into two categories (< 40 and ≥ 40) according to previous reports [6, 8]. We used the t-test, Mann–Whitney U test, chi-square test, or Fisher’s exact test to assess patients’ characteristics between PNI categories. We conducted a Kaplan–Meier analysis to estimate the survival rates between the PNI groups, and the survival estimates were compared using the log-rank test. To assess the relationship between PNI and PD discontinuation, we used Cox proportional hazards regression models using PNI ≥ 40 as a reference. Plots of log [− log (survival rate)] against log (survival time) were generated to test the proportionality of this assumption.
In Cox proportional hazards regression models, three models were examined based on the level of multivariate adjustment: (a) model 1: a minimally adjusted model that included PNI (< 40 and ≥ 40), (b) model 2: PNI, age, and sex, (c) model 3: PNI, age, sex, presence of diabetes mellitus, CVD history, Hb, and urine volume. The selected variables were based on previous studies [14,15,16,17,18]. Associations between continuous PNI and PD discontinuation across the three adjustment levels were also modeled using restricted cubic splines with knots at 10, 50, and 90 percentiles of the PNI.
Most analyses were performed using JMP version 12 (SAS Institute Inc., Cary, NC, USA). Kaplan–Meier and spline analyses were performed using Stata version 17.1 (Stata Corporation, College Station, TX, USA). Statistical significance was set at P < 0.05.