Introduction

The prevalence of obesity is increasing worldwide. According to the predictions of the World Health Organization (WHO), by 2025, one in five adults worldwide will be obese. Obesity is also increasing in the chronic kidney disease (CKD) population. The prevalence of obesity among patients undergoing dialysis in the USA is > 30% [1]. Although obesity is a well-known risk factor for end-stage renal disease (ESRD) [2], it has historically been considered a relative contraindication for peritoneal dialysis (PD) [3]. Although recent studies have shown that this myth has been dispelled and PD can be successfully carried out in obese patients [4, 5], observational studies in the CKD population have reported conflicting evidence about the relationship between obesity and mortality. There are also publications showing that it can provide a survival advantage for this patient population [6]. However, those with a body mass index (BMI) < 20 kg/m2 were associated with the highest mortality [7]. While higher BMI is associated with improved survival in dialysis patients, a lower BMI was associated with higher mortality [8,9,10,11]. This phenomenon of obesity paradox has greatly contributed to the growing confusion among nephrologists about whether to treat obesity in dialysis patients. Especially in the first year of PD, a BMI < 18.5 kg/m2 has been associated with mortality [12].

There are still conflicting reports about infectious complications such as peritonitis, which is known as one of the most important causes of morbidity and mortality for PD patients, and complications such as technique proficiency [13, 14]. In addition, obesity is considered among the modifiable risk factors for peritonitis [15]. Therefore, prospective data on these patients are insufficient, as PD is not recommended for many obese patients and patients with a BMI < 18.5 kg/m2 [13]. Our aim in this study is to present the data in our region to the literature by comparing the RRF, peritonitis, technique proficiency, and mortality rates of the patients we grouped according to BMI.

Materials and methods

The data of 404 patients whose PD were started by our clinic and who were followed up for at least 3 months between March 2005 and November 2022 were evaluated retrospectively. Patients aged 18 and over were included in the study. Approval was obtained from the local ethics committee. Since it was a retrospective observational study, patient consent was not required. At the start of the RRT, everyone’s BMI was calculated by dividing their weight in kilograms by the square of their height in meters, as recommended by the WHO. Patients were divided into 4 groups as Weak: BMI below 18.5 kg/m2, Normal weight: BMI 18.5 to less than or equal to 24.9 kg/m2, Overweight: BMI of 25 to less than or equal to 29.9 kg/m2, and Obese: BMI of 30 kg/m2 or greater. In addition, BMIs were recorded at the time of the patients’ withdrawal from PD. For the diagnosis of peritonitis, at least 2 of the following criteria are required according to International Society for Peritoneal Dialysis (ISPD) 2016 guidelines: clinical features consistent with peritonitis, peritoneal fluid leukocytosis (at least 100/mm3 white cell count, at least 50 polymorphonuclear neutrophilic cells) and positive culture of PD fluid. If the 24 h urine volume of the patients was less than 100 ml, it was considered as urinary cessation and was called complete loss RRF. In addition, demographic and clinical characteristics of the patients, echocardiography (ECHO) reports at 6-month periods, number of peritonitis, time to first peritonitis attack, Kt/V values, initial urine values, withdrawal urine values, and reasons for PD withdrawal were included in the analysis. Technique proficiency for PD patients was defined as failure to achieve the desired target in Kt/V (Kt/V < 1.7), catheter dysfunction, ultrafiltration failure, skin or genital leakage, and leakage into the third spaces outside the peritoneal cavity. The local institutional ethics committee approved this study (Local Ethics Committee. No: B.30.2.ATA.0.01.00/133).

PD modalities and selection

In our country, patients can choose the RRT method they want without any limiting factor except medical necessity. Patients make their choices after this educational model is applied. A predialysis education program (PDEP) is a specially prepared training kit that has been used in some centers in our country. Using visual and written cards, this training kit educates CKD patients and their relatives. It has 6 modules as summarized: Module 1: how do kidneys work? What is kidney failure? Which diseases cause kidney failure?; Module 2: why is diet important in kidney disease? The drugs used in kidney disease and the importance of exercise; Module 3: introduction to the treatment of renal failure and general information about RRT; Module 4: peritoneal dialysis; Module 5: hemodialysis; and Module 6: kidney transplantation.

At PD initiation, participants were assigned to one of the PD modalities; CAPD with a twin-bagged system or instrumented peritoneal dialysis (IPD). The prescription of CAPD was 4× 1.5−2.5 L (body surface area, RRF determined) exchanges if no sign of inadequate dialysis was observed. Dialysate fluids containing (i) 1.36%, 2.27%, or 3.86% glucose; (ii) amino acids; or (iii) icodextrin were used according to the clinical needs of patients. The prescription of IPD was to give 6–15 L of exchange fluid within 6–8 h.

Statistical analysis

SPSS 20.0 (IBM Cor., Chicago, IL, USA) program was used for data analysis. In the statistical analysis of the study, mean, standard deviation, frequency, and percentage values were defined. The findings in the data were analyzed by the Kolmogorov–Smirnov (K–S) test. The general characteristics and demographic characteristics of the groups were determined by frequency (descriptive analysis: frequency analysis for a single variable) analysis. The Pearson Chi-square test was used to determine the relationship between categorical variables. Kruskal–Wallis H test was used for the mean comparison of non-parametric values of multiple independent groups, and Tamhane’s T2 test was used for post-hoc analysis. One-way ANOVA (one-way analysis of variance) was used for the mean comparison of parametric values of multiple independent groups, and Duncan test was used for post-hoc analysis. In bilateral comparisons, paired sample t-test for parametric values and Wilcoxon test for non-parametric values was used to compare the single variable for 2 different situations.

The endpoint for the analysis of patient survival was death. The endpoint for the patients in peritoneal dialysis was conversion to hemodialysis or transplantation. Conversion to hemodialysis and transplantation were censored for patient survival, and mortality curves were generated using the Kaplan–Meier log-rank method. The endpoint of the patients’ RRF loss was a decrease in the amount of urine below 100 mL/day. The patient’s lack of RRF loss was censored, and graphs showing RRF loss were generated using the Kaplan–Meier log-rank method. The patients’ endpoint for the first peritonitis episode analysis was experiencing the first peritonitis episode. The patient’s absence of a peritonitis episode was censored, and graphs showing the peritonitis episode were generated using the Kaplan–Meier log-rank method. In the analysis of the survival of the technique, transition to hemodialysis and death were considered final events; functional dialysis and loss of follow-up data were censored. In the analysis of the survival of the technique, kidney transplant patients were not included in the analysis. Univariate and multivariate Cox regression analyses were performed to determine risk factors for mortality in patients. A p ≤ 0.05 value was considered statistically significant in the entire study.

Results

A total of 404 patients were identified as 44 underweight, 199 normal weight, 110 overweight, and 55 obese. The mean BMI at the beginning of PD was determined as 17.09 ± 1.29, 21.92 ± 1.73, 26.74 ± 1.29, and 33.92 ± 3.33, respectively. The ratio of female was high in the underweight and obese patient group. Among the groups, the underweight group had the lowest mean age and the highest mean age was in the overweight and obese group (Table 1). In terms of ESKD etiology, the rate of diabetes was high in overweight and obese patients (Table 1). Hypertension (HT) rate increased significantly as BMI increased. Other demographic and laboratory follow-up data are given in Table 1.

Table 1 Comparison of demographic, clinical, RRF and peritonitis rates of peritoneal dialysis patients according to BMI
Full size table

In the Kaplan–Meier analysis technique survey, no difference was found between the groups (p = 0.610, Fig. 1). Duration to the first peritonitis of the patients was determined as 23.79 ± 21.31 months in Underweight patients, 22.86 ± 22.582 months in Normal Weight patients, 19.45 ± 19.4 months in Overweight patients and 24.89 ± 23.61 months in Obese patients, and there was no significant difference between them (p = 0.445, Table 1). Peritonitis-free survival of the groups was similar in Kaplan–Meier analysis (p = 0.840) (Fig. 2). In addition, according to BMI, Methicillin-sensitive coagulase-negative staphylococci (MSCNS) were the most common peritonitis agent in the first peritonitis attack in all groups. The most common agents reproduced in tunnel infections were MSCNS and Methicillin-sensitive staphylococcus aureus (Table 1). There was no difference in urinary volume loss (residual renal function loss) between the patient groups undergoing PD (p = 0.896) (Fig. 3).

Fig. 1
figure 1

Kaplan–Meier analysis of technique survival: comparison of patient groups according to BMI

Full size image
Fig. 2
figure 2

Kaplan–Meier analysis of first peritonitis episode: comparison of patient groups according to BMI

Full size image
Fig. 3
figure 3

Kaplan–Meier analysis of probability RRF loss: comparison of patient groups according to BMI

Full size image

Of the 404 patients, 270 died. Their mean survey was 56,837 ± 2929 months. Among the groups, the underweight patients had the lowest mortality and overweight group had the highest mortality. Among the patient groups, the best survival by years was in the underweight patient group. Annual mortality rates are given in Table 2. In Kaplan–Meier analysis, the mortality of the overweight group was high (p = 0.029) (Fig. 4). Multivariate Cox regression analysis demonstrated that BMI did not affect mortality (HR 1.196 [95% CI 0.722–1.981] (p = 0.488)). Factors affecting mortality were loss of RRF, duration to the first episode of peritonitis, and mean albumin values (respectively, (HR 0.551) [95% CI 0.352–0.865] (p = 0.010); (HR 0.982) [95% CI 0.973–0.990] (p < 0.001); (HR 0.205) [95% CI 0.126–0.334] (p < 0.001)) (Table 3).

Table 2 Comparison of mortality according to BMI
Full size table
Fig. 4
figure 4

Kaplan–Meier analysis of survival: comparison of patient groups according to BMI

Full size image
Table 3 Univariate and multivariate Cox regression analyses showing influencing factors on overall survival
Full size table

The BMI of the patients at the time of PD withdrawal was significantly increased compared to the baseline BMI, except for the obese patients (Table 4). In the Kaplan–Meier Analysis performed according to the PD withdrawal BMI of the patients, no significant difference was found between the groups in the patient surveys (Fig. 5) (p = 0.292).

Table 4 BMI changes
Full size table
Fig. 5
figure 5

Kaplan–Meier analysis of survival: comparison of patient groups according to BMI in output

Full size image

Discussion

This study is one of the rare studies that evaluated not only the mortality of PD patients grouped according to BMI, but also peritonitis, the agents of the first peritonitis episode, tunnel infection, RRF, technique proficiency, and mortality according to withdrawal BMI. It has been shown that having a high BMI in hemodialysis has a protective effect on mortality [16]. This condition has conflicting data in PD patients. Although having a high BMI in PD patients had better mortality in the first year, this could not be demonstrated throughout the other years [17]. In addition, there are studies indicating that having a low BMI is a higher risk for death in PD patients [14]. On the contrary, high BMI was found to be associated with high mortality in some publications [4, 18]. The proportion of adipose tissue for the same BMI varies by ethnicity. The World Health Organization has proposed different BMI criteria to reflect obesity for Caucasians and Asians [19]. These data from different study regions suggest that ethnicity may influence mortality according to BMI. In addition, differences in mortality may be due to the fact that BMI does not distinguish between central obesity and generalized obesity. It has been reported that weight loss after the onset of PD increases mortality, and weight gain and maintaining a constant weight do not change mortality [20]. There are studies stating that weight gain with an increase in muscle mass improves mortality more than with an increase in adipose tissue [21]. It is difficult to determine this distinction in PD patients. The increase in BMI of the patients in our study after the onset of PD may have improved the mortality of the patients in the underweight patient group. Moreover, being underweight alone may not be an indicator of malnutrition. Therefore, the presence of conditions such as inflammation and arteriosclerosis together with low weight may increase mortality [22]. A neutral effect of BMI on survival has also been observed in several studies [23, 24]. In our study, we showed that BMI was not among the factors affecting mortality in the Cox regression analysis performed with baseline BMI and withdrawal BMI did not affect mortality. Hence, we think that BMI is not a determinant in the selection of PD.

There were limited studies comparing the relationship between BMI and peritonitis. Studies have reported that obesity increases the risk of peritonitis due to greater difficulties in exit site care and increased susceptibility to skin and soft tissue infection [14, 25]. There are previous data showing poor PD outcomes including poor solute clearance, higher risk of infectious complications, and early technique proficiency [4, 26, 27]. Recent data [28,29,30,31,32], like our results, report that there was no difference between BMI groups in PD patients in terms of complications such as technique proficiency, duration to first peritonitis attack, inadequate dialysis, and tunnel infection. In addition, we identified MSCNS as the most common peritonitis agent, similar to the other study [14].

In conclusion, we report that BMI alone is not an accurate parameter to predict mortality, peritonitis, RRF, and technique proficiency parameters in PD patients. In this respect, we believe that it is not appropriate for physicians and patients to evaluate BMI as a negative parameter in the selection of PD.

Study limitations

There were some limitations in our study. Due to the relatively small size of our patient groups and the single-center experience, racial differences could not be revealed. In addition, the nutritional adequacy assessment of the patients was just limited to albumin and BMI. Bioimpedance data of the patients were not available. Since it was a retrospective study, the causes of death could not be stated. Despite these facts, our study is the first to compare mortality, peritonitis, technique proficiency, RRF, and mortality according to PD withdrawal BMI in PD patients according to BMIs.