Introduction

The current global prevalence of Chronic Kidney Disease (CKD) is 13.4% [1]. However, the awareness rate is 9.5% and its mortality is 1.5% [2]. The condition is one of the top 20 causes of death globally. Deaths due to CKD increased by 34% worldwide between 1990 and 2013 [3], making CKD a “silent killer” [4]. Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation using creatinine measurement is used globally to estimate glomerular filtration rate (eGFR) [5]. An aging population is a global issue [6] with the number of individuals older than 60 years increasing at an annual rate of 2.6% [7]. Although there is no exact definition of “elderly” most developed countries define it as an age ≥ 65years [8] , while the standard is ≥ 60 years in most developing countries [9]. The structure and functionality of kidneys gradually decline at a rate of 6.3mL/min per 1.73m² every 10 years, with increase in age [10]. Renal efficiency has been shown to diminish with age [11] with an incidence rate of between 23.4% and 58.5% in the elderly [12-15]. As people age, renal uric acid excretion is reduced, hence becoming more susceptible to Hyperuricemia (HUA) [16]. The prevalence of HUA in American adults was 21.4% [17]. CKD has a higher prevalence in those suffering from hyperuricemia relative to the rest of the population [18]. Evidence indicates that the possibility of kidney function deterioration increases by 14% for each 1 mg/dl rise in uric acid level in the serum [19]. One study showed that for every 60umol/L rice in the levels of uric acid in the serum, there is a 17% increase in all-cause mortality [20] but other studies have concluded otherwise [21]. So far, only a few papers have examined the link between uric acid levels in the serum and death in aged CKD patients. Considering that CKD occurs more frequently, understanding the potential connection between these two will improve the clinical management of CKD patients. Herein, we enrolled aged CKD patients to delineate the impact of uric acid levels in the serum on mortality rates.

Materials and Methods

Patient recruitment and methods

From January 1, 2009 to July 31, 2016 536 inpatients were enrolled from the Third Affiliated Hospital of Sun Yat-sen University and were followed up until December 31, 2016. The inclusion criteria included: age ≥18 years, CKD and complete (99m) Tc-DTPA renal dynamic imaging. Patients with incomplete clinical data and initiation of renal replacement therapy were excluded from the study.

Baseline data collection and Biochemical Evaluation

Baseline data included demographic characteristics (age and gender) and measurement index (BMI and eGFR). The data also included blood Analysis Of Albumin (ALB), Fasting Blood Glucose (FBS), Total Cholesterol (CHOL), triglyceride, High Density Lipoprotein (HDL), Low Density Lipoprotein (LDL), Hemoglobin (HGB), pre-albumin, Potassium (K), carbon dioxide combining power (CO²CP), Calcium (Ca), Phosphorus (P) and Blood Urea Nitrogen (BUN). We applied the CKD-EPI equation to compute the urine routine including, medication status of uratelowering drug use and eGFR [22].

CKD-EPI equation for eGFR

GFR=141×min (Scr/0.9,1)^－0．4111×max (Scr/0.9,1)^－1.209×0．993^Age×1.159 (if black), if male

GFR=141×1.018×min (Scr/0.7,1)^－0.329×max (Scr/0.7,1)^－1.209×0.993^Age×1.159 (if black), if female

Definition

Hyperuricemia referred to serum UA ≥360 μmol/L in females and ≥420μmol/L in males [23].

Statistical analyses

Data analysis was completed with SPSS statistical software version 22.0. Normality of the data was determined with Kolmogorov-Smirnov test (P>0.05) Table 1). Continuous variables that met the normal distribution parameters were reported as the mean ± SD. Other variables were presented as median (interquartile range). Chi-square and Wilcoxon rank sum tests were employed to analyze continuous variables and compare the categorical variables, respectively. Analysis of Variance (ANOVA) was used to compare different treatments. Cox proportional hazards model was utilized for multivariate analysis. Confounding factors were adjusted by 7 models as described below. Variables with P<0.1 in univariate evaluations were subjected to multivariate assessments.

Model I: non-adjusted.

Model II: adjusted demographic data (gender & age).

Model III: adjusted demographic data (gender & age), eGFR and urate-lowering drug application.

Model IV: adjusted model 3 plus nutritional index (BMI, albumin, fasting blood glucose, triglyceride, CHOL, HDL, LDL, prealbumin, HGB).

Model V: adjusted model 3 plus kidney function indices (potassium, carbon dioxide combining power, calcium, phosphorus, urea nitrogen).

Model VI: adjusted model 3 plus proteinuria.

Model VII: adjusted all factors.

Results

Baseline characteristics of all the patients

Baseline features were presented based on quartiles of uric acid levels in the serum with the participants being classified into 4 groups according to SUA quartiles. The 1st, 2nd, 3rd and 4th quartiles were: 808-504.15μmol/L, 401.55-504.14μmol/L, 310.15-401.54μmol/L and 90.8-310.14μmol/L, respectively. The patient’s age was between 64-74 years, with the median age being 69 years. Out of the 536 patients, 285 (53.2%) were males and 251 (46.8%) females. The estimated GFR (CKD-Epi equation) was24.093 (22.195.26.573)/ min/1.73m². Deaths recorded were 82(15.3%) and 274 patients (51.1%) had hyperuricemia (Table 2).

Hyperuricemia versus all-cause mortality

In model I, hyperuricemia was positively correlated with all-cause mortality (β=2.319, P＜0.001). After correcting for confounding factors as per model II (β=2.37, P＜.001), model IV (β=1.793, P=.0027) and model VII (β=1.81, P=.0032), (the same correlation as in model I was observed. However, in model III (β=1.531, P=0.085), model V (β=1.59, P=.0069) and model VI (β=1.547, P=0.078) there was no correlation with regards to hyperuricemia versus all-cause mortality (Table 3A).

Serum uric acid levels versus all-cause mortality

The levels of uric acid in the serum was positively linked to all-cause mortality in; model I (β=1.003, P＜0.001), model II (β=1.003, P＜0.001), model V (β=1.002, P=.0047), model VI (β=1.002, P=0.022) and model 7 (β=1.002, P=.0036). No correlation was observed with regards to serum uric acid levels versus all-cause mortality in; model 3(β=1.002, P=0.053) and model 4(β=1.002, P=0.061) (Table 3B).

Serum uric acid quartiles versus all-cause mortality

Serum uric acid quartiles were grouped into 4. Uric acid levels in the serum from the highest to the lowest group were;808- 504.15μmol/L，401.55-504.14μmol/L，310.15-401.54μmol/L and 90.8-310.14μmol/L respectively. The third group was considered as the reference group. The lowest serum uric acid was protective for all-cause mortality in model 1 (β=0.398, P=0.009). In addition, the same correlation was observed in; the third serum uric acid quartile in model II (β=0.502, 0.423，P=0.047, 0.017) and the lowest serum uric acid group in model IV (β=0.478, 0.45, P=0.041, 0.039) and model VII (β=0.445, P=0.004). No correlation was observed with regards to serum uric acid quartiles versus all-cause mortality in models III, V and VII (Table 3C).

Discussion

After correcting for confounding factors, cox proportional hazards analysis revealed that hyperuricemia is the distinct risk factors of all-cause mortality in aged CKD patients, with 81% increment risk. The uric acid levels in the serum were positively correlated with all-cause mortality, with lower levels exhibiting significant protective effect for all-cause mortality. The results presented herein are similar to findings from other studies [24-27]. Nevertheless, hyperuricemia was also a protective factor of all-cause death but such results were only observed in hemodialysis patients [28,29]. Hyperuricemia may thus considered a marker in hemodialysis patients. Some studies have shown a J-shaped link in serum uric acid levels versus mortality [30,31]. Moreover,a clinical and cohort studies showed that hyperuricaemia was an independent risk factor for CKD progression in children and adolescents [32]. These studies however, did not target the aged population whereas this study majorly focused on an aged CKD population, a large sample size, complete followup and exclusion of confounding factors. Since the number of people suffering from hyperuricemia or CKD is constantly increasing with the aging of the population, aged CKD patients were the preferred study population. Furthermore, there was exclusion of most factors such as; demographic data, eGFR, urate-lowering drug application, nutritional index, kidney function indices and proteinuria.The results showed that hyperuricemia had a detrimental effect onall-cause mortality. Excellent medical care should therefore be accorded t to hyperuricemia in aged non-dialysis CKD patients, as it may offer a survival benefit when the serum uric acid is maintained at a low level. If not well- treated, hyperuricemia can negatively influence multiple mechanisms in the body such as; platelet function and subsequently the blood rheology [33], proliferation of vascular smooth muscle cells , reduce nitric oxide production , vascular endothelial cell dysfunction , activating the renin angiotensin system and many others [34-37].

Limitations of The Study

This research had a few limitations of limited sample size, being retrospective. Additionally, the serum uric acid was only assessed at baseline, without follow-up data.

Conclusion

This study showed that hyperuricemia is the independent risk factor of all-cause mortality in aged CKD patients with 81% increment risk. Nevertheless, a low level of uric acid in the serum has a significant protective effect for all-cause mortality. This insight will help improve diagnostic and treatment measures for hyperuricemia in aged patients with CKD, remarkably reducing the all- cause mortality in aged populations.