Blood urea nitrogen is independently associated with renal outcomes in Japanese patients with stage 3–5 chronic kidney disease: a prospective observational study

Six hundred eighty-two consecutive Japanese patients with CKD stage 3–5 were admitted to our hospital, the National Hospital Organization Kyushu Medical Center, for evaluation and education about CKD between June 2009 and December 2017. Figure 1 shows a flow chart of the patient enrollment process. To decrease the effects of extrarenal factors on BUN levels, patients who used corticosteroids at a dose > 5 mg/day, or who had an active gastric ulcer or hemorrhagic gastritis, liver cirrhosis, or a left ventricular ejection fraction < 40% were excluded. Four hundred fifty-nine patients were analyzed. After discharge, patients were prospectively followed at our hospital for at least 6 months. Data were collected until July 2018. The composite renal endpoint was end-stage renal disease (ESRD) or death, whichever occurred first, and ESRD alone was added as an alternative outcome. ESRD was defined as renal dysfunction requiring renal replacement therapy such as maintenance hemodialysis or peritoneal dialysis or having undergone kidney transplantation. When the patients showed gradual worsening of kidney function during follow-up and were determined to require renal replacement therapy, they were diagnosed with ESRD and were re-admitted for renal replacement therapy. Time to ESRD was defined as the duration from baseline to the day of the first dialysis session or the kidney transplantation procedure. Death occurred before reaching ESRD. Information regarding patient death and definitions of censored patients are described in our previous report [13]. In total, 69 patients were censored; 63 were lost to follow-up and 6 required maintenance hemodialysis because of acute exacerbation of kidney function due to infectious diseases or congestive heart failure. As a result, 390 patients were completely followed up until July 2018.

Blood samples were obtained from each patient in the early morning after an overnight fast. Daily proteinuria was also measured. The eGFR (mL/min/1.73 m²) was calculated using the following new equation for Japanese patients: 194 × serum creatinine^− 1.094 × age^− 0.287 × 0.739 (if female) [14]. The cSosm (mOsm/kg) was calculated using the following formula: (2 × sodium) + (BUN/2.8) + (glucose/18) [15].

Clinical assessments at admission and the definitions of hypertension, diabetes mellitus, smoking, and dyslipidemia are described in our previous report [13]. An echocardiogram was performed to assess left atrial diameter and left ventricular ejection fraction.

Continuous data are expressed as the mean ± SD or median (interquartile range) depending on their distribution. Participant BUN levels were divided into quartiles. Categorical data were compared across BUN quartiles using the chi-squared test with the Fisher’s exact test used for groups containing fewer than five individuals in any given cell. An analysis of variance (ANOVA) was used to compare continuous variables that were approximately normally distributed, and the Kruskal–Wallis test was used to compare skewed continuous variables. Participant cSosm levels were also divided into quartiles. Spearman’s rank correlation coefficients were used to identify the association between two continuous variables.

To assess the robustness of the findings shown in Table 2, we performed sensitivity analysis, and these results are shown in Table 4. In a multivariable analysis, we included serum creatinine as a covariate, instead of eGFR. Higher BUN quartiles, but not cSosm quartiles, were independently associated with adverse renal composite outcomes. These findings were similar to those shown in Table 2.

This study had some limitations. First, patients who were admitted to a single regional hospital were recruited in this cohort; as a result, this study had relatively small sample size, old patients, and the number of male patients more than twice that of female patients. Accordingly, selection bias may exist in this population. In the stratified analyses (Fig. 4), higher BUN levels were independently associated with adverse renal outcomes in male subjects, but not in female subjects. In addition, a significant association was observed in the higher age group but not in the lower age group. Second, we did not examine uremic toxins, such as indoxyl sulfate and p-cresyl sulfate, which are associated with kidney disease progression [39, 40]. Third, we did not evaluate protein intake before admission or precise volume status at admission, which could have affected baseline BUN levels. A previous report suggested that left atrial diameter could be considered a marker of volume status in dialysis patients [41]. We assessed left atrial diameter in each patient and it was slightly increased in the higher BUN quartiles (Table 1). This suggests that higher BUN levels did not reflect volume depletion. Fourth, the follow-up period in the present study was relatively short. Finally, it might be considered that a single BUN measurement does not provide sufficient accuracy for predicting the renal outcome.