Effects of hyperuricemia on incident renal replacement therapy and all-cause mortality among patients with chronic kidney disease stages 3-5: a retrospective cohort study

ABSTRACT BACKGROUND: Findings regarding the effects of hyperuricemia on renal function and mortality have been inconsistent. OBJECTIVES: To investigate the effects of hyperuricemia on incident renal replacement therapy and all-cause mortality among patients with chronic kidney disease (CKD). DESIGN AND SETTING: Retrospective cohort study conducted in a medical center in Taiwan. METHODS: Patients with CKD in stages 3-5, without histories of renal replacement therapy, were consecutively recruited from 2007 to 2013. Their medical history, laboratory and medication data were collected from hospital records. The mean uric acid level in the first year of follow-up was used for analyses. Hyperuricemia was defined as mean uric acid level ≥ 7.0 mg/dl in men or ≥ 6.0 mg/dl in women. The primary outcomes were incident renal replacement therapy and all-cause mortality, and these data were retrospectively collected from hospital records until the end of 2015. RESULTS: A total of 4,381 patients were analyzed (mean age 71.0 ± 14.8 years; males 62.7%), and the median follow-up period was 2.5 years. Patients with hyperuricemia were at increased risk of incident renal replacement therapy and all-cause mortality, especially those with CKD in stages 4 or 5. Compared with patients with CKD in stage 3 and normouricemia, patients with CKD in stages 4 or 5 presented significantly higher risk of all-cause mortality only if they had hyperuricemia. CONCLUSIONS: In patients with CKD in stages 3-5, hyperuricemia was associated with higher risk of incident renal replacement therapy and all-cause mortality. Whether treatment with uric acid-lowering drugs in these patients would improve their outcomes merits further investigation.


INTRODUCTION
Hyperuricemia, defined as a uric acid level higher than 7.0 mg/dl in men or higher than 6.0 mg/ dl in women, 1 has been associated with cardiovascular risk factors 2-3 and cardiovascular diseases. 4,5 Hyperinsulinemia, a known consequence of insulin resistance, has been shown to reduce urinary excretion of uric acid. 6 Meanwhile, uric acid retention due to renal vasoconstriction is common in patients with hypertension. 7 In a meta-analysis on 25 observational studies with nearly 100,000 participants, 8 it was demonstrated that hyperuricemia independently increased the risk of incident hypertension. In another meta-analysis on 17 prospective observational studies with more than 160,000 participants, 9 a uric acid level that was one standard deviation higher was associated with an odds ratio for coronary heart disease of 1.07 (95% confidence interval, CI 1.04-1.10).
With regard to the risk of kidney disease, hyperuricemia has been associated with incident chronic kidney disease (CKD). In a prospective cohort study on more than 21,000 healthy volunteers, 10 a uric acid level ≥ 7.0 mg/dl at baseline was associated with a higher risk of incident kidney disease (defined as an estimated glomerular filtration rate [eGFR] < 60 ml/min/1.73 m 2 ) during a median follow-up duration of 7 years. Similar findings were noted in another prospective study 11 on more than 13,000 participants pooled from two community-based cohorts with a mean follow-up duration of 8.5 years. These results were consistent with the findings from a meta-analysis 12 on 15 cohort studies with a total of nearly 100,000 relatively healthy individuals and nearly 3,500 incident CKD cases (defined as eGFR < 60 ml/min/1.73 m 2 ).
However, among patients with CKD who were not on dialysis, the data regarding the effects of hyperuricemia on renal function and mortality have been limited and the findings have been inconsistent. Among patients with CKD in stages 3-4 who were on chronic allopurinol therapy for hyperuricemia, 13 withdrawal of allopurinol led to a significant increase in serum uric acid level and a significant deterioration in renal function, especially in those who were not on blockers of the renin-angiotensin system. In contrast, uric acid was not associated with the rate of decline in renal function in another cohort study 14 on patients with CKD in stages 3-5. With regard to risk of mortality, hyperuricemia was an independent risk factor for all-cause and cardiovascular mortality (but not kidney failure) in a cohort of patients with CKD in stages 3-4 during a median follow-up of 10 years. 15 In contrast, hyperuricemia was not associated with cardiovascular diseases and all-cause mortality in another cohort of patients with CKD in stages 3-4 during a median follow-up of 9 years. 16

OBJECTIVE
In this study, we aimed to investigate the effects of hyperuricemia on incident renal replacement therapy (RRT) and all-cause mortality, among patients with CKD in stages 3-5 who were not on dialysis at the baseline.

METHODS
This was a retrospective cohort study conducted in a medical center in central Taiwan. Patients with CKD in stages 3-5 with no history of RRT were consecutively recruited in a CKD care program from January 2007 to December 2013. The patients' medical history, laboratory and medication data were collected from hospital records. The primary outcomes of this study were incident RRT and all-cause mortality, and these data were collected from hospital records until the end of 2015. We included patients for whom information about their vital status (records of follow-up visits or mortality events) were available in the analyses. This study was approved by the Local Institutional Review Board in Taichung with no RRT). The patients' systolic and diastolic blood pressure and uric acid levels were collected from hospital records, and the mean blood pressure and uric acid levels in the first year of follow-up were used for analyses. Hyperuricemia was defined as a mean uric acid level ≥ 7.0 mg/dl in men or ≥ 6.0 mg/dl in women. The uric acid-lowering drugs used included allopurinol and benzbromarone.
RRT was defined as hemodialysis, peritoneal dialysis or renal transplantation. We classified patients' cause of death into cardiovascular or non-cardiovascular death 18 by reviewing the medical records.
Cardiovascular death included sudden cardiac death, death due to acute myocardial infarction, death due to heart failure, death due to stroke and death due to other cardiovascular causes. The codes for cardiovascular deaths that we used were International Classification of Diseases (ICD)-9 390-448. Non-cardiovascular death was defined as any death not covered by cardiac death or vascular death. 18 All causes of death were independently reviewed by two board-certified physicians of internal medicine. In the event of an inconsistency, an independent cardiologist reviewed the medical records and made the final classification.
All of the statistical analyses were performed using the SAS software (version 9.4; SAS Institute, Inc., Cary, NC, USA). Continuous variables were reported as the mean ± standard deviation (SD), while categorical data were given as numbers (with percentages).
The differences in clinical variables between groups were tested for statistical significance using the chi-square test for categorical variables; or using an independent-sample t test or one-way analysis of variance (ANOVA) for continuous variables.
We used Cox's proportional hazards model to compare the hazard ratios (HRs) of RRT, all-cause mortality and cardiovascular and non-cardiovascular death among patients with different CKD stages or uric acid levels. Since mortality was an event that competed with RRT, an extended Cox's proportional hazards model was used to calculate the subdistribution hazard ratio (SHR) as a sensitivity test. 19 For cause-specific mortality, we considered non-cardiovascular death to be an event competing with cardiovascular death, and vice versa. Therefore, an extended Cox's proportional hazards model was used to determine the SHR of cardiovascular and non-cardiovascular death among patients with different CKD stages or uric acid levels. The variables that were adjusted for in the analyses included factors that were known to be associated with the risk of mortality or with uric acid levels (age, sex, systolic blood pressure, smoking history, diabetes and use of uric acid-lowering drugs and statins). In all of the statistical analyses, a two-sided P-value < 0.05 was considered statistically significant.

RESULTS
A total of 4,381 patients were included in the analyses (mean age 71.0 ± 14.8 years; males 62.7%), and the median follow-up period was 2.5 years. Table 1 shows the baseline characteristics of the study patients according to CKD stages and uric acid levels. The patients who were at a later stage of CKD were younger, were more frequently female, had higher systolic and diastolic blood pressure, were less likely to have a smoking history, were more likely to have diabetes, had higher uric acid levels and were less likely to be on a statin, compared with those at CKD stage 3. The patients who had hyperuricemia (defined as a mean uric acid level ≥ 7.0 mg/ dl in men or ≥ 6.0 mg/dl in women) were younger, were more frequently female, had higher systolic and diastolic blood pressure, were less likely to have a smoking history, had lower eGFR and were more likely to be on uric acid-lowering treatment, compared with those presenting normouricemia.  The effects of hyperuricemia on incident RRT and mortality across the CKD stages are shown in Table 3. The higher risk of incident RRT associated with hyperuricemia was statistically significant among patients at CKD stage 4 or stage 5, but not among patients at CKD stage 3. With regard to risk of mortality, hyperuricemia was associated with significantly higher risk of all-cause mortality among patients at CKD stage 4. Patients with hyperuricemia did not have significantly higher risk of cardiovascular death across the CKD stages.
Hyperuricemia was associated with significantly higher risk of non-cardiovascular death among patients at CKD stage 4 or stage 5 ( Table 3).
When we stratified our patients according to their CKD stages and uric acid levels (  Table 4).

DISCUSSION
In this study, we investigated patients with CKD in stages 3-5 and found that patients with hyperuricemia (defined as a mean uric acid level ≥ 7.0 mg/dl in men or ≥ 6.0 mg/dl in women) were at increased risk of incident RRT and all-cause mortality, independent of traditional risk factors such as age, gender, blood pressure, smoking, diabetes etc. ( Table 2). Hyperuricemia was associated with worse outcomes mainly among patients at CKD stage 4 or stage 5 (significantly higher risk of incident RRT and noncardiovascular mortality; Table 3).  Per 1000 person-years. b Adjusted for age, gender, systolic blood pressure, smoking history, diabetes and use of uric acid-lowering drugs and statins. c Defined as a mean uric acid level ≥ 7.0 mg/dl in men or ≥ 6.0 mg/dl in women. CKD = chronic kidney disease; CV = cardiovascular; HR = hazard ratio; SHR = subdistribution hazard ratio.
Patients with advanced CKD and hyperuricemia were at substantially increased risk of incident RRT. Moreover, compared with patients with CKD in stage 3 with normouricemia, patients at CKD stage 4 or stage 5 were at significantly higher risk of allcause mortality only if they had hyperuricemia ( Table 4).
Our findings are consistent with those of previous studies [10][11][12]20,21 that showed that hyperuricemia was an independent risk factor for decreased kidney function and presence of end-stage renal disease in general populations. Our results suggest that hyperuricemia was a risk factor for decreased kidney function requiring RRT among patients with CKD in stages 3-5. Several studies have reported that hyperuricemia was not associated with any decline in renal function 14 or kidney failure 14,15 among patients with CKD in stages 3-5.
Certain factors may account for these inconsistent findings.
First, because uric acid is excreted by the kidney, a decrease in renal function is inevitably accompanied by an increase in serum uric acid level. This complicating phenomenon therefore makes it particularly challenging to study the role of uric acid in CKD. Second, patients with advanced CKD and/or hyperuricemia are at increased risk of mortality. [22][23][24][25][26] Thus, the lack of association between hyperuricemia and decreased kidney function in previous studies might be explained by the fact that the competing risk of mortality was not considered. In our study, we demonstrated that hyperuricemia was associated with incident RRT among patients with CKD ( Table 2), especially those at CKD stage 4 or stage 5 ( Table 3). Moreover, the risk remained significant after adjustment for multiple factors and consideration of the competing risk of mortality (Tables 2 and 3).
We also reported that hyperuricemia was associated with a higher risk of all-cause mortality (mainly non-cardiovascular death) among patients with CKD in stages 3-5 ( Table 2). It is well known that patients with CKD are at increased risk of mortality and cardiovascular events. [22][23][24] Although occurrences of hyperuricemia had previously been correlated with risks of mortality and cardiovascular diseases, 25,26 conflicting data regarding this association had been reported in the general population [26][27][28] and among patients with diabetes 29,30 or CKD. 15,16 Our data support an association between hyperuricemia and higher risk of mortality among patients with CKD. 15,31,32 Interestingly, less than 20% of the all-cause mortality comprised cardiovascular death among our patients ( Table 2). Although patients with CKD are at increased risk of cardiovascular events, [22][23][24] it is worth noting that the majority of patients with CKD have died due to non-cardiovascular causes. 33,34 For example, in a large population of non-dialysis-dependent CKD patients, 33 less than 35% of the all-cause mortality was due to cardiovascular death during a median follow-up duration of 2.3 years. In another Asian population with CKD, 34 less than 20% of the all-cause mortality was due to cardiovascular death among patients with baseline eGFR < 45 ml/min/1.73 m 2 , over the course of a median follow-up duration of 9.8 years. Thus, most patients with CKD died due to non-cardiovascular causes. 34 We found that hyperuricemia was associated with higher risk of all-cause mortality and non-cardiovascular death among patients with CKD in stages 3-5 ( Table 2).
In line with previous reports, 33,34 the leading cause of non-cardiovascular death in our patients was malignancy (n = 202; 56.7% of the all-cause mortality). We found that hyperuricemia was associated with higher risk of non-cardiovascular death ( Table 2), especially among patients with CKD in stage 4 or stage 5 ( Table 3).
More than three decades ago, it was hypothesized that uric acid might protect against carcinogenesis, owing to its antioxidant properties. 35 However, studies in more recent decades have suggested that hyperuricemia was associated with higher risk of incidence of cancer and mortality. 36 Thus, despite the possible antioxidative effects of uric acid, this substance may play a contributory role in carcinogenesis.
Since reactive oxygen species play a critical role in cell growth and survival, in both normal and cancer cells, 37 uric acid may promote cancer cell growth and survival by scavenging reactive oxygen species and reducing oxidative stress-induced apoptosis. 36,37 Another mechanism through which hyperuricemia may be related to higher risk of incidence of cancer and mortality is inflammation.
Uric acid has been found to induce the expression of several inflammatory mediators (such as monocyte chemoattractant protein-1 and C-reactive protein), 38 which may lead to a microenvironment with lowgrade inflammation, 36 thus favoring transformation into cancer cells.
Taken together, a growing body of evidence suggests that hyperuricemia is associated with a higher risk of mortality due to cancer, 36 which is the leading cause of non-cardiovascular death among patients with CKD. 33,34 Our results suggest that hyperuricemia was associated with higher risk of non-cardiovascular death among patients with non-dialysis-dependent CKD.
Our study had some limitations. First, the causal relationship between hyperuricemia and incident RRT and mortality among patients with CKD could not be confirmed in this cohort study.
Although several small studies have reported that treatment with the uric acid-lowering drug allopurinol slowed the progression of kidney disease and reduced cardiovascular risk in patients with CKD, 39,40 this effect needs to be confirmed in a large-scale prospective study. Second, the events of RRT and mortality among our patients were collected from our hospital records.
Third, this was a retrospective cohort study. Thus, we may have underestimated the event rates among our patients. In a large prospective cohort study 24 in which the mortality data were obtained from death certificate codes in Taiwan, the all-cause mortality rate among patients with CKD in stages 3-5 (n = 26,757) was 16.2% over a median follow-up period of 7.5 years. The all-cause mortality rate among our patients was 8.1% over a median follow-up period of 2.5 years. Although patients with CKD in stages [3][4][5] in the aforementioned study 24 were younger (mean age around 61.9 years) than our patients (mean age 71.0 years), the extent of underestimation of mortality events in our study was likely to have been small, and therefore probably did not confound the results.
Lastly, we investigated patients with CKD in stages 3-5 in this study. Whether our findings may be generalized to patients with early stages of CKD needs further investigation.

CONCLUSION
We demonstrated that in patients with CKD in stages 3-5, hyperuricemia was associated with higher risk of incident RRT and all-cause mortality (mainly non-cardiovascular death). Further studies are needed, to investigate whether treatment with uric acidlowering drugs in such patients would improve their outcomes.