Introduction

Chronic kidney disease (CKD) eventually progresses to end-stage renal disease and requires renal replacement therapy. Although there is no sufficient evidence guiding the initiation of dialysis, the merit of early initiation of dialysis has not been recognized [1,2,3]. The Kidney Disease: Improving Global Outcomes (KDIGO) guidelines recommend that dialysis should be initiated when one or more of the following are present: symptoms or signs attributable to kidney failure (serositis, acid–base or electrolyte abnormalities, pruritus); inability to control volume status or blood pressure; a progressive deterioration in nutritional status refractory to dietary intervention; or cognitive impairment [4]. This often but not invariably occurs in the GFR range between 5 and 10 mL/min/1.73 m2. The United Kingdom Renal Association 2009 recommends renal replacement therapy initiation when the eGFR is less than 6 mL/min/1.73 m2, even if the patient is asymptomatic [5]. The Canadian Society of Nephrology (CSN) 2014 clinical practice guidelines suggest that chronic dialysis should be initiated when the eGFR drops to 6 mL/min/1.73 m2, even if there are no clinical indications [6]. Japanese Society of Dialysis Therapy (JSDT) recommends dialysis initiation when the eGFR is 4–6 mL/min/1.73 m2 [7]. These guidelines use eGFR as an indicator of renal function at the time of initiating dialysis because of its accuracy and availability. Moreover, eGFR is a good indicator of mortality once dialysis is initiated.

Besides eGFR, creatinine (Cr) clearance and levels of serum Cr, blood urea nitrogen (BUN), serum cystatin C, and serum β2 microglobulin are used for the evaluation of renal function. It is recognized that BUN is not an accurate index of renal function because it increases regardless of the decline in renal function. Excess of protein intake, body fluid depletion, heart failure, increased catabolism, and use of diuretics increase BUN levels. Increase in BUN levels out of proportion to serum creatinine levels reflects a critical condition.

However, there are few previous reports on the association between ratio of BUN to serum Cr (BUN/Cr) at the time of initiation of dialysis and mortality in these patients. Therefore, the purpose of this study was to examine the association between the BUN/Cr ratio at the time of initiation of dialysis and mortality during the maintenance dialysis.

Materials and methods

Subjects

The subjects were patients for whom dialysis was initiated at the 17 Aichi Cohort Study of Prognosis in Patients Newly Initiated into Dialysis (AICOPP) group centers from October 2011 to September 2013. Patients who were withdrawn from dialysis during hospitalization, who died during hospitalization, or who did not consent to be registered were excluded. The study included 1520 patients who were at least 20 years of age, had CKD, and provided written consent. We determined by a survey conducted at the end of September 2016.

Patient characteristics and data when dialysis was initiated (baseline)

Body mass index (BMI) was measured at the first hemodialysis session or the first exchange of peritoneal dialysate. Diabetes was defined as fasting blood glucose ≥126 mg/dL, random blood glucose ≥200 mg/dL, HbA1c (National Glycohemoglobin Standardization Program) ≥6.5%, use of insulin, or use of oral hypoglycemic agents. History of coronary artery disease (CAD) was defined as having a history of coronary artery intervention, heart bypass surgery, or ischemic changes on electrocardiogram. History of ischemic stroke was defined as confirmed diagnoses using computed tomography or magnetic resonance imaging. The period of nephrology care was established, based on the patients’ medical records, as the period from the referral to the nephrologist until the initiation of dialysis. Medication use referred to the drugs being taken at the time of dialysis initiation. Blood tests were performed on samples taken before the first hemodialysis session or the first exchange of peritoneal dialysate. Blood pressure was measured before the first dialysis session or the first exchange of peritoneal dialysate. Symptoms of heart failure were defined as the presence of subjective symptoms, pulmonary congestion, or pleural fluid on a chest radiogram, and hypoxemia. The following formulas were used to calculate eGFR by gender: for males, eGFR (mL/min/1.73 m2) = 194 × [age]−0.287 × [serum creatinine (mg/dL)]−1.094; and for females, eGFR (mL/min/1.73 m2) = 194 × [age]−0.287 × [serum creatinine (mg/dL)]−1.094 × 0.739. The following formulas were used to calculate Cr clearance by gender: for males, Cr clearance (mL/min) = [(140-age) × body weight (kg)]/[serum creatinine (mg/dL) × 72]; and for females, Cr clearance (mL/min) = [(140-age) × body weight (kg)]/[serum creatinine (mg/dL) × 72] × 0.85.

Parameters of renal function at dialysis initiation

We used blood urea nitrogen to creatinine ratio (BUN/Cr), BUN levels, serum creatinine levels, eGFR, Cr clearance or serum β2 microglobulin as parameters of renal function before initiating dialysis. Additionally, according to BUN/Cr ratio, patients were divided into four quartiles (Q1, Q2, Q3, and Q4). There were 380 patients in each group. We compared the association between these parameters and all-cause mortality.

Survey of survival prognosis

Survival prognosis as of September 30, 2016 was determined by surveying medical records. For patients who were transferred to other institutions, information was obtained by mailing out survey forms. Cardiovascular-related death was defined as death due to heart failure, acute coronary syndrome, fatal arrhythmia, aortic disease, stroke, and peripheral artery disease.

Outcomes

The study outcomes included: (1) comparisons of all-cause, cardiovascular-related, heart disease-related, infection-related, and cancer-related mortality rates in the four groups as categorized by BUN/Cr ratios; (2) extraction of factors, which included BUN/Cr ratio, BUN level, serum creatinine level, eGFR, Cr clearance or serum β2 microglobulin level, influencing all-cause mortality.

Statistical processing

SPSS statistics version 24 and the Easy R program [8] were used for statistical analyses. Comparisons of characteristics and baseline data between the four groups of patients were performed using the analysis of variance (ANOVA) for continuous variables and Chi square test for nominal variables. All-cause mortality rates were compared using the log-rank test for the Kaplan–Meier curves. Factors contributing to the all-cause mortality rates were examined using univariate Cox proportional hazard regression analysis. In addition to the four groups, BUN/Cr ratio, BUN level, serum creatinine level, eGFR, Cr clearance or serum β2 microglobulin level, factors that were significant in the univariate analysis (age, sex, BMI, duration of nephrology care, dialysis modality, diabetes, CAD, ischemic stroke, systolic blood pressure (SBP), diastolic blood pressure (DBP), heart rate, heart failure symptoms at the first dialysis session, levels of hemoglobin, serum albumin, adjusted calcium, serum phosphorus, serum CRP, serum PTH, use of angiotensin receptor blocker (ARB)/angiotensin converting enzyme inhibitor (ACEI), calcium channel blocker (CCB), loop diuretics, statins, erythropoiesis stimulating agent, and temporary vascular access at the first dialysis session and diabetes mellitus were used as explanatory variables for the multivariate Cox proportional hazard analysis. In stratified analyses using factors such as age, sex, dialysis modality, comorbidity of diabetes, and BMI, all-cause mortality rates were compared using Cox proportional hazard models adjusted for the factors used in the above-described multivariate analyses. Logistic regression analysis was used to examine the associations between the four groups, BUN/Cr ratio, BUN level, serum creatinine level, eGFR, Cr clearance or serum β2 microglobulin level and 180, 365, 730, and 1095 days of survival. Furthermore, receiver operative characteristic (ROC) curves were used to compare the sensitivity and specificity for all-cause mortality among BUN/Cr ratio, BUN level, serum creatinine level, eGFR, Cr clearance or serum β2 microglobulin level. A p value less than 0.05 was considered statistically significant.

Results

Comparison of patient characteristics and baseline data

Table 1 shows the patient characteristics and baseline data in the four groups. Significant differences between the four groups were observed in age, sex, history of CAD, ischemic stroke, dialysis modality, BMI, SBP, DBP, heart rate, heart failure symptoms at the first dialysis session, BUN level, serum creatinine level, eGFR, Cr clearance, levels of serum-adjusted calcium, serum low-density lipoprotein, serum triglyceride, serum CRP, use of ACEi/ARB, use of beta blockers, loop diuretics, thiazides, and corticosteroids.

Table 1 Baseline characteristics
Full size table

Comparison of all-cause mortality

Figure 1 shows the Kaplan–Meier curves for the cumulative survival rates of the four groups. There were 391 deaths during the follow-up period (Q1, 60 cases; Q2, 72 cases; Q3, 101 cases; Q4, 158 cases). Significant differences were observed between the cumulative survival rates of the four groups (p < 0.001).

Fig. 1
figure 1

Kaplan–Meier curves for the cumulative survival rates between the 4 groups. Significant differences were observed between the 4 groups’ cumulative survival rates (p < 0.001)

Full size image

Comparison of cause of death

Table 2 shows the causes of death among the four groups. Cardiovascular-related death including heart failure, acute coronary syndrome, cardiogenic sudden death, and stroke was the highest cause in each group. Figure 2 shows the Kaplan–Meier curves for the cumulative survival rates by causes of death. Significant differences were observed between the cumulative survival rates except for cancer-related death of the four groups (p < 0.001).

Table 2 Causes of death among the four groups
Full size table
Fig. 2
figure 2

Kaplan–Meier curves for the cumulative survival rates by causes of death between the 4 groups. Significant differences were observed between the cumulative survival rates except for cancer-related death of the four groups (p < 0.001). a CVD-related mortality, b heart disease-related mortality, c infection-related mortality, d cancer-related mortality

Full size image

Factors affecting all-cause mortality

The results of univariate Cox proportional hazard regression analysis are presented in Table 3. The Q3 and Q4 groups were associated with worse prognosis as compared to the Q1 group [hazard ratio (HR) = 1.71 and 2.97, 95% confidence interval (CI) 1.24–2.35 and 2.20–3.99, p = 0.001 and <0.001, respectively]. The increase in BUN/Cr ratio was associated with the survival prognosis (for increase in the ratio by 1, HR = 1.07, 95% CI 1.05–1.09, p < 0.001). Additionally, high mortality was associated with factors including high BUN level, low serum creatinine level, high eGFR, low Cr clearance, and high β2 microglobulin level. The results of multivariate Cox proportional hazard analysis are shown in Table 4. The Q4 group was associated with worse prognosis as compared to the Q1 group (HR 1.82, 95% CI 1.24–2.67, p = 0.002) after adjusting for factors, which were significant in the univariate analysis and comorbidity of diabetes mellitus. The increase in BUN/Cr was associated with the survival prognosis (for increase in the ratio by 1, HR 1.04, 95% CI 1.02–1.06, p = 0.002) after adjustment.

Table 3 Univariate regression analysis for all-cause mortality
Full size table
Table 4 Multivariate regression analysis for all-cause mortality
Full size table

Parameters of renal function affecting 180, 365, 730, and 1095 days mortality

Table 5 shows the odds ratio of all-cause mortality among the parameters of renal function. The Q3 and Q4 groups were associated with the lower survival rates of 365, 730, and 1095 days as compared to the Q1 group. The increase in BUN/Cr ratio was associated with the lower survival rates of 365, 730, and 1095 days. The increase in BUN level was associated with the lower survival rates of 180, 365, and 1095 days. In addition, the decrease in serum creatinine level and the increase in eGFR were associated with the lower survival rate of 180, 365, 730, and 1095 days.

Table 5 Odds ratio of 180, 365, 730, and 1095 days mortality
Full size table

ROC curve of parameters for all-cause mortality

Figure 3 shows the ROC curve of BUN/Cr ratio to examine the sensitivity and specificity for 180, 365, 730, and 1095 days mortality. The sensitivity and specificity ranged from 0.60 to 0.72 and 0.59 to 0.71, respectively. The cutoffs of BUN/Cr ratio were 12.0, 10.6, 10.8, and 10.8. Table 6 shows the comparison of area under curve (AUC) among renal parameters. The AUC of BUN/Cr ratio for 180, 365, 730, and 1095 days mortality was the highest.

Fig. 3
figure 3

The ROC curve for all-cause mortality using BUN/Cr ratio. The cutoff value (specificity, sensitivity). a 180 days mortality, 12.0 (0.71, 0.62). b 365 days mortality, 10.6 (0.59, 0.72). c 730 days mortality, 10.8 (0.62, 0.60). d 1095 days mortality, 10.8 (0.64, 0.60)

Full size image
Table 6 ROC curve of parameters for all-cause mortality
Full size table

Subgroup analyses

Table 7 shows the results of subgroup analyses. A significant association between BUN/Cr ratio and all-cause mortality was observed in people who were over 70 years of age, male, on hemodialysis, no comorbidity of diabetes, BMI <23.5, no heart failure symptoms at the first dialysis session, and no use of thiazide.

Table 7 Hazard ratio of BUN creatinine ratio for all-cause mortality by stratified analysis
Full size table

Discussion

The results of this study revealed that BUN/Cr ratio at the time of initiating dialysis was associated with mortality during maintenance dialysis. Furthermore, BUN/Cr ratio at the time of initiation of dialysis was better predictor of all-cause mortality than other factors including eGFR and Cr clearance. We would like to highlight the two following features of this study: (1) to our knowledge, this is the first study to examine the relationship between BUN/Cr ratio at the time of initiating dialysis and all-cause mortality among dialysis patients, and (2) BUN/Cr ratio, with 10.6–10.8 as the best cutoff points for all-cause mortality, could be a convenient marker for routine clinical settings. Some reports showed that BUN/Cr ratio was a predictor for prognosis in patients with acute heart failure and acute kidney injury [9,10,11,12]. Takaya et al. indicated that BUN/Cr ratio of 22 or more was associated with poor survival prognosis in patients with acute heart failure [13]. Body fluid depletion or heart failure increases reabsorption of urea at renal tubules and the BUN/Cr ratio as well. Gastrointestinal bleeding may increase the catabolism or increase the absorption of urea in the intestine, leading to an increase in BUN. In other words, higher BUN/Cr ratio indicated serious medical condition. In this study, the highest BUN/Cr ratio group presented the higher frequency of history of CAD and ischemic stroke, higher frequency of heart failure symptom, low BMI, and increase in CRP level at dialysis initiation, steroid use and diuretics use. Hence, we considered that in patients with history of cardiovascular disease chronic inflammation and malnutrition continued even after dialysis initiation. The assumption supported the higher cardiovascular- and infection-related mortality in addition to cardiovascular mortality in the highest BUN/Cr ratio group. Patients with CKD are often initiated to dialysis because uremia, dehydration, or infection exacerbate renal dysfunction. We reported in another paper that patients with rapid decline of renal function in the 3 months before initiating dialysis had poorer survival prognoses among the patients in this study patients [14]. Those patients also showed higher BUN/Cr ratio.

Previous studies have demonstrated the toxicity of urea to various organs through insulin resistance, free radical production, and apoptosis [15,16,17,18]. Furthermore, urea is the origin for the generation of cyanate [19], ammonia [20, 21], and carbamylated compounds [22, 23]. Urea is converted to ammonia by urease, which is expressed by intestinal bacteria. Ammonia is in part converted to ammonium hydroxide, and both compounds lead to an increase in the pH in the intestinal lumen and possibly are important causes of bacterial translocation [20, 24]. We thought that BUN/Cr ratio, which was hypothesized to adjust for muscle mass, was more likely than BUN level alone to predict prognosis in our study. The results of the subgroup analysis, which showed the association between BUN/Cr ratio and prognosis to be more obvious in older patients or those with lower BMI, supported the hypothesis.

Most studies concerning the association between renal function at the time of initiation of dialysis and prognosis used GFR in order to evaluate the renal function [4,5,6,7, 25]. Meta-analyses showed higher GFR at the initiation of dialysis was associated with higher all-cause mortality risk [26, 27]. Patients who are initiated with higher GFR were more likely to be comorbid of cardiovascular disease, malnutrition, and uremic condition. Therefore, the patients are needed to start dialysis in order to recovery from poor general conditions in spite of higher GFR. A national cohort study in Taiwan demonstrated that the group with the highest eGFR showed the highest mortality in 23,551 incident hemodialysis patients [28]. In the study, the highest eGFR group presented with higher BUN/Cr ratio. This finding was consistent with our results. In our study, eGFR was a good predictor of all-cause mortality. However, AUC of BUN/Cr ratio for 180 and 365 days mortality was higher than eGFR by analysis using ROC curve. Hence, we considered that BUN/Cr ratio at dialysis initiation was strongly associated with 1 year or less survival because high BUN/Cr ratio reflected an unstable condition.

The present study has the following limitations. First, this was an observational analysis, and there were some differences in the baseline characteristics and laboratory data among the four groups. Second, we did not use uniform criteria for initiating dialysis. Rather, it was left to the discretion of the attending physician. The timing for initiating dialysis changes over time depending on the available evidence, which is based on various guidelines of clinical management. Nevertheless, given that the nephrologists decided on the timing for initiating dialysis in all the subjects in this study and the decisions on when to initiate dialysis were made in the same historical period, major differences between institutions and attending physicians are unlikely. Third, GFR and Cr clearance were not measured actually but estimated. It was quite possible that there was discrepancy because body fluid imbalances or decrease in muscle mass influenced serum creatinine level and the resulting value calculated from it.

Conclusion

We conclude that the BUN/Cr ratio at the time of initiating dialysis was associated with all-cause mortality. To the best of our knowledge, this is the first study to analyze that relationship. Further work with regard to the underlying mechanisms is required to confirm this association.