Development and validation of a nomogram to predict protein-energy wasting in patients with peritoneal dialysis: a multicenter cohort study

Introduction

PEW is a prevalent complication in chronic kidney disease patients with peritoneal dialysis, manifesting as a disorder of multiple metabolism and nutrition and a decrease of protein and fat (Fouque et al., 2008; Orozco-González et al., 2022). PEW results from inflammation, catabolic illness, and nutrient intake reduction. PD Patients with PEW develop a disequilibrium between energy intake and expenditure (Kittiskulnam et al., 2021; Tennankore & Bargman, 2013). According to the criteria of the International Society of Renal Nutrition and Metabolism (ISRNM), previous investigations reported PEW ranged between 14.1% and 83% in patients with peritoneal dialysis (Liu et al., 2015; Markaki et al., 2016; Harvinder et al., 2016; Zhou et al., 2017). PEW leads to lower quality of life and a higher hospitalization rate, morbidity, and mortality (Velasquez et al., 2013; Yasui et al., 2016; Chao et al., 2017). PEW is a crucial predictor of mortality in patients with peritoneal dialysis (Wang, 2011).

Peritoneal dialysis is an exchange between blood and a glucose-containing peritoneal solution. In this process, proteins and amino acids transport into the dialysate effluent. Patients lose amino acids of 1–3.5 g/d and proteins of 5–10 g/d during peritoneal dialysis, mainly albumin (Kopple et al., 1982; Krediet et al., 1986; Westra et al., 2007). A previous study presented that every 1 g/L elevation in serum albumin was independently related to an 8% reduction of all-cause mortality risk in patients with peritoneal dialysis. The subjective global assessment (SGA) score is a tool to estimate the nutritional condition. Every unit improvement in SGA score was associated with a 25% reduction of mortality risk in patients with peritoneal dialysis (Wang, 2011). Therefore, PEW prediction can prevent the loss of protein, energy, and mortality risk in patients with peritoneal dialysis. Nowadays, the ISRNM criterion is the only recognized method of PEW diagnosis. Several researchers supposed that this ISRNM criterion is complicated to diagnose PEW patients. Consequently, they aimed to develop more straightforward approaches to replace the ISRNM criterion for PEW diagnosis, such as SGA and Malnutrition Inflammation Score. However, these tools are not recognized methods for diagnosis applied in the clinical. In addition, they are merely used to estimate the nutritional status but hardly predict the risk of PEW (Harvinder et al., 2016). Predictive model development will guide clinicians to identify patients with a higher risk of PEW and prevent the occurrence.

Recently nomogram has been a technique for clinicians to predict the probability of disease or outcome. This study aimed to develop a PEW nomogram in patients with PD. According to the nomogram, clinicians can calculate PEW risk and identify the influence factors. Practical intervention approaches help to reduce PEW occurrence.

Methods

Results

A total of 369 patients were enrolled in this study. They were divided into development (n = 210) and validation (n = 159) sets. A total of 184 (49.86%) of 369 patients were PEW. Baseline characteristics are shown in Table 1. The average age was 57.44 years. A total of 48.24% of patients were women. The median dialysis duration was 17.5 months in PEW patients and 24 months in non-PEW patients. There was no significant variable difference between the training and validation sets (Table 2). Datasets can be analyzed to screen predictors and construct a nomogram.

Nomogram of PEW in patients with peritoneal dialysis

We included eight predictors to develop a nomogram by multivariate logistic regression (Fig. 1). According to the model, we can identify the PEW risk for an individual patient. For example, a 65-year-old patient who experienced peritoneal dialysis for 40 months showed a recent laboratory examination: Ccr 60 ml/min, Scr 500 mmol/L, serum calcium 2.4 mmol/L, glucose 6 mmol/L, TG 3 mmol/L, CRP 5 mg/L. The corresponding score of each indicator was 26 points, 23 points, 70 points, 52 points, 36 points, 68 points, 50 points, and 3 points. His total score was 328 points. The risk of PEW was 50%.

Nomogram evaluation

The Hosmer-Lemeshow test in development and validation sets was ${\chi }^2$ = 10.42, p = 0.404, and ${\chi }^2$ = 13.06, p = 0.220. This result presents the predicted outcomes were highly consistent with the actual effects. The ROCs in the development and validation sets performed a nomogram with an excellent discrimination ability (AUC 0.769, 95% CI [0.705–0.832], AUC 0.669, 95% CI [0.585−0.753]) (Figs. 2 and 3). We discovered a good correlation between predicted results and observed outcomes in the training and test sets (Figs. 4 and 5). The nomogram had a clinical utility from DCA (Figs. 6 and 7).

Discussion

In this study, the PEW incidence in patients with peritoneal dialysis was 49.86%. The rate was within the range reported in the previous study (Harvinder et al., 2016). For the first time, this study developed a nomogram to predict PEW in patients with peritoneal dialysis. Predictors were age, dialysis duration, Ccr, Scr, serum calcium, CRP, TG, and glucose. This nomogram performed an excellent discrimination function and practice utility. The predicted outcome highly correlated to the observed result.

Protein and amino acids transport to the dialysate in the peritoneal dialysis process. Patients with PD tend to suffer from protein catabolism, malnutrition, and inflammation. ISRNM suggested the status as PEW (Harvinder et al., 2016). The PEW incidence in patients with PD is higher than in hemodialysis (Harvinder et al., 2016; Kiebalo et al., 2020). PEW increases rehospitalization rate, cardiovascular morbidity, and mortality (As’habi et al., 2019). Therefore, PEW prevention improves the prognosis of patients with PD (Tsai et al., 2019). This study recognized hypocalcemia as a risk factor for PEW. Vitamin D modulates calcium absorption in the human body. Because vitamin D activation requires the 1,25(OH)2D in the proximal kidney tubule, renal injury weakens vitamin D activation to decrease serum calcium (Cannata-Andía et al., 2021; Zappulo et al., 2022). In CKD patients, hypocalcemia always results from vitamin D deficiency and indicates an occurrence of CKD-metabolic bone disease (CKD-MBD). Vitamin D and CKD-MBD are associated with PEW. Vitamin D has long been recognized as a contributor and involved in crucial molecular pathways for PEW (Garcia et al., 2011). Humans or experimental models with kidney injury presented a positive relationship between 25-hydroxyvitamin D and testosterone. This finding recommended that vitamin D may control muscle mass (Lee et al., 2012). Wasted muscle is one of the diagnostic criteria for PEW. Vitamin D insufficiency and CKD-MBD lead to bone loss and muscle decrease, which can predispose patients to PEW. Furthermore, hypocalcemia is a severe complication correlated to metabolic bone disease and bone loss, contributing to PEW (Li et al., 2022). Therefore, we should evaluate the potential vitamin D insufficiency and CKD-MBD in patients with hypocalcemia. Calcium supplements and CKD-MBD correction are effective methods to reduce PEW risk.

Although some variables’ odds ratios were very close to 1, such as SCr, age, Ccr, dialysis duration, and CRP, we still considered them PEW predictors according to previous research and clinical observation. The crucial impact of inflammation on the nutrition condition has been published (Carrero et al., 2007; Rattazzi et al., 2003). Inflammation reduces energy and stimulates protein consumption in patients with peritoneal dialysis. CRP elevation is one evidence of inflammation, which can predict wasting in dialysis patients (Landray et al., 2004; Kalantar-Zadeh et al., 2004). A previous study discovered a negative correlation between CRP and protein, serum albumin, and serum pre-albumin (Choi et al., 2010). They reported CRP as a helpful marker. Regular CRP assessment can improve malnutrition management in patients with peritoneal dialysis. This nomogram indicated that every 10 mg/L elevation in CRP level would increase the PEW risk by 9.4% for patients with a total score of more than 275. In the clinical, the expected level of CRP is less than 10 mg/L. Patients with peritoneal dialysis should prevent infection and other situations leading to high CRP of more than 10 mg/L. Medical intervention reducing CRP plays a significant role in PEW prevention for patients with infection or inflammation.

Uremia promotes protein catabolism, increases the generation of non-nutritional toxins, and decreases albumin levels (Garibotto et al., 2013). Toxins clearance plays a vital role in patients with peritoneal dialysis. Ccr indicates the clearance ability. Most patients with peritoneal dialysis still have residual kidney function to clear toxins. Furthermore, peritoneal dialysis is a direct approach to help them eliminate toxins. Therefore, we summarized the Ccr of residual kidney function and peritoneal dialysis to estimate a patient’s clearance ability. In this nomogram, every 10 ml/min reduction in total Ccr increased the PEW risk by 32% for patients with Ccr of less than 70 ml/min. Patients with Ccr of more than 70 ml/min prevent the PEW risk. However, ESRD patients will likely be confronted with several conditions reducing the residual kidney function and peritoneal dialysis performance. Maintaining a total Ccr level of more than 70 ml/min is challenging. In this study, the mean Ccr level is 61.44 ml/min. The number of patients with Ccr of more than 70 ml/min was 101 of 369 (27.37%). Identifying and correcting factors reducing the Ccr level is essential to prevent PEW. Aged patients with peritoneal dialysis often experience loss of skeletal muscle mass. The changes in protein metabolism of patients with uremia were related to age (Zhang & Ren, 2012), a risk factor for PEW in patients with peritoneal dialysis. Clinicians should pay attention to PEW risk assessment for aged patients.

This study discovered that patients with short-term peritoneal dialysis had a higher PEW risk. No researcher reported the association between dialysis duration and PEW risk before. In the clinical, we found that patients with short-term peritoneal dialysis tend to present higher ultrafiltration. Higher ultrafiltration stimulated peritoneal protein loss, resulting in hypoalbuminemia (Kiebalo et al., 2020). Therefore, we considered that peritoneal protein losses probably elevate the PEW risk for patients with short-term peritoneal dialysis. A previous analysis indirectly supported this conclusion. They explored whether peritoneal protein transport increased with peritoneal dialysis treatment duration by collecting 103 patients treated with peritoneal dialysis for more than 4 years (Goodlad & Davenport, 2017). Researchers found higher dialysate/serum total protein ratios and reduced serum albumin in patients with initial exposure to peritoneal dialysis. This result indicated faster transport in patients who have just started peritoneal dialysis. With the prolongation of peritoneal dialysis duration, peritoneal creatinine, and total protein transport will change. Large pore transport for protein loss will not sustain an increase. Another report also revealed a consistent result. They proved the impact of peritoneal performance on patient characteristics at the start of dialysis. During long-term peritoneal dialysis, the peritoneal function can remain fundamentally steady (Johansson & Haraldsson, 2006). Therefore, patients with initial peritoneal dialysis likely occurred peritoneal protein loss because of better ultrafiltration and faster transport. Protein supplement was crucial for this kind of patient to prevent PEW.

Creatine (Cr) is an energy precursor essential to cellular physiology. In patients with chronic kidney disease, skeletal and cardiac muscles are depleted of Cr with the dialysis duration. Cellular damage due to Cr depletion lead to a deterioration of musculoskeletal and neurological function and poor quality of life. Cr supplements improve patients’ musculoskeletal, brain, and peripheral nervous systems with chronic kidney disease. About 80–90% of total Cr exists in the human muscles. Approximately 2% of total Cr is converted to creatinine daily, which the kidney eliminates into the urine (Wallimann, Riek & Möddel, 2017). Serum creatinine is an indicator to measure the level of creatinine. PEW patients are always combined with muscle wasting, suggesting Cr reduction. Diminished conversion from Cr brought lower serum creatinine levels. Therefore, consistent with this nomogram, lower serum creatinine indicated a probability of muscle wasting, increasing the PEW risk.

Conclusion

We developed a predictive model for PEW in patients with peritoneal dialysis. Predictors were age, dialysis duration, glucose, CRP, Ccr, Scr, serum calcium, and TG. This nomogram provided valuable evidence for predicting PEW in patients with peritoneal dialysis. According to this nomogram, clinicians can identify factors and make specific intervention methods to prevent PEW occurrence in high-risk patients.

Supplemental Information