Association of Uric Acid and Left Ventricular Mass Index With Renal Outcomes in Chronic Kidney Disease

Background Hyperuricemia and left ventricular (LV) hypertrophy are prevalent in chronic kidney disease (CKD), but the association of uric acid (UA) and left ventricular mass index (LVMI) with renal outcomes in patients with CKD is unclear. We conducted a study to assess whether the combina- tion of UA and LVMI is associated with renal outcomes in patients with CKD of stages 3–5. This longitudinal study enrolled 540 patients, who were classified into four groups according to sex-specific median values of UA and LVMI. The study investigated the associations of the study groups with progres- sion to dialysis, rapid progression of decline in renal function (decline in estimated glomerular filtration rate (eGFR) > 3 ml/min/1.73 m 2 /year), and change in eGFR, using Cox proportional hazards modeling, logistic regression analysis, and linear mixed-effects modeling, respectively. < 0.04). conclusions Our findings show that the combination of a higher UA and LVMI is a risk factor for progression to dialysis and rapid progression of decline in renal function in patients with CKD of stages 3–5.

Hyperuricemia is highly prevalent in chronic kidney disease (CKD), and may result from the decreased renal excretion of uric acid (UA) when renal function declines. Hyperuricemia is also concurrently associated with various risk factors for CKD. [1][2][3] Some epidemiological studies have identified a role of hyperuricemia in adverse renal outcomes, [4][5][6][7][8] whereas others have failed to identify hyperuricemia as a risk factor. [9][10][11] Therefore, the status of hyperuricemia as a marker of risk for progressive decline in renal function remains controversial, and whether UA is merely a marker that reflects the integration of comorbidities and renal damage or a true risk factor for clinical outcomes is unclear.
The serum concentration of UA is also affected by or linked to various risk factors for left ventricular hypertrophy (LVH), such as obesity, insulin resistance, dyslipidemia, and hypertension. 2,12,13 In a study done by Iwashimaet al., serum UA was associated with LV mass index (LVMI) determined with echocardiography. 14 Left ventricular hypertrophy is also frequently encountered in patients with CKD because of pressure and volume overload. 15,16 Patients with LVH may have decreased myocardial contractility and impaired LV diastolic function, 17 which may have a progressive effect on renal function with an adverse outcome. 18,19 Iwashima et al. 14 had evaluated the effect of a combination of UA and LVMI on five cardiovascular outcomes in 619 patients with essential hypertension. They found that the combination of hyperuricemia and LVH was a predictor of cardiovascular outcomes. 14 However, studies of the association of UA and LVMI with renal outcomes in patients with CKD have been limited. Accordingly, the aim of the present study was to assess whether the combination of UA and LVMI is associated with progression to dialysis, a rapidly progressing decline in renal function, and a change in renal function in patients with CKD of stages 3-5.

suBjects and Methods study patients and design
This was a prospective observational study conducted in a regional hospital in southern Taiwan. From January 2007 to May 2010, we consecutively enrolled, from our Outpatient Department of Internal Medicine, 540 predialysis patients with CKD of stages 3-5 according to the National Kidney Foundation-Kidney Disease Outcomes Quality Initiative (K/ DOQI) guidelines . 20 We classified our patients with evidence of kidney damage lasting for more than 3 months into CKD stages 3, 4, and 5 on the basis of an eGFR (ml/min/1.73 m 2 ) of 30-59, 15-29, and < 15, respectively. Patients with significant aortic or mitral valve diseases and inadequate image visualization were excluded. The protocol for the study was approved by our Institutional Review Board, and all enrolled patients gave written, informed consent for their participation.

evaluation of cardiac structure and function
The echocardiographic examination of the patients was performed by two experienced cardiologists with a VIVID 7 cardiovascular ultrasound system (General Electric Medical Systems, Horten, Norway), with the participant respiring quietly in the left decubitus position. The cardiologists were blind to the participants' other data. Two-dimensional and two-dimensionally guided M-mode images were recorded from standardized views. The echocardiographic measurements included LV internal diameter in diastole (LVIDd), LV posterior wall thickness in diastole (LVPWTd), and interventricular septal wall thickness in diastole (IVSTd). Left ventricular systolic function was assessed from the LV ejection fraction (LVEF). Left ventricular relative wall thickness (LVRWT) was calculated as the ratio of 2 × LVPWTd/LVIDd. Left ventricular mass was calculated using the Devereuxmodified method (i.e., left ventricular mass = 1.04 × ((IVSTd + LVIDd + LVPWTd)3 -LVIDd3) -13.6 g). 21 Left ventricular mass index was calculated by dividing LV mass by body surface area. Left ventricular hypertrophy was defined as suggested by the 2007 European Society of Hypertension/ European Society of Cardiology guidelines. 22

collection of demographic, medical and laboratory data
Demographic and medical data including age, sex, and comorbid conditions were obtained from medical records or interviews with patients. The body mass index (BMI) was calculated as the ratio of weight in kilograms divided by the square of body height in meters. Laboratory data were measured in fasting blood samples using an autoanalyzer (COBAS Integra 400, Roche Diagnostics GmbH, Mannheim, Germany). Serum creatinine was measured by the compensated Jaffé (kinetic alkaline picrate) method in a Roche/Integra 400 Analyzer (Roche Diagnostics GmbH), using a calibrator traceable by isotope-dilution mass spectrometry. 23 The value of eGFR was calculated with the four-variable equation in the Modification of Diet in Renal Disease (MDRD) study. 24 Proteinuria was examined with dipsticks (Hema-Combistix, Bayer Diagnostics, Dublin, Ireland). A test result of proteinuria was classified as negative, 1+, or > 1+. Blood and urine samples were obtained within 1 month of enrollment of a patient in the study. In addition, information about patient medications used during the study period, including angiotensin converting enzyme inhibitors (ACEIs), angiotensin II receptor blockers (ARBs), β-blockers, calcium-channel blockers, diuretics, and hypouricemic agents (allopurinol and benzbromarone), was obtained from medical records.

definition of renal end point
The renal end point was defined as commencement of dialysis. In patients reaching the renal end point, renal function data were censored at the beginning of renal replacement therapy. Patients not reaching this end point were followed until February 2011. Dialysis was begun according to the regulations of the Taiwanese National Health Insurance system for dialysis therapy on the basis of laboratory data, nutrition status, and uremic symptoms and signs.

definition of rapid renal progression
The slope of the eGFR was defined as the regression coefficient of eGFR vs. time in units of ml/min/1.73 m 2 /year. At least three eGFR measurements after echocardiographic examination were required to estimate the slope of eGFR. Any reduction in eGFR of more than 3 ml/min/1.73 m 2 /year was considered to represent rapid progression of a decline in renal function. 25

statistical analysis
Statistical analysis was done with SPSS version 18.0 (SPSS Inc., Chicago, IL) for windows. Data are expressed as percentages, mean ± standard deviation (SD), and mean ± standard error of the mean (SEM), for continuous variables with approximately-normal distribution for the slope of eGFR, or median (25th-75th percentile) for triglyceride, number of serum creatinine measurements, and follow-up period.
The study patients were stratified into four groups according to sex-specific median values of UA and LVMI. Multiple comparisons among the study groups were performed with one-way analysis of variance (ANOVA) followed by the post hoc test adjusted with a Bonferroni correction. The relationship between two continuous variables was assessed through a bivariate correlation method (Pearson's product-moment correlation). Time to the beginning of dialysis and covariates of risk factors were modeled using the Cox proportional hazards model. Multiple logistic regression analysis was used to identify the risk factors associated with rapid progression of decline in renal function. The group with lower UA and LVMI, which was the lowest risk group for the outcomes, was used for reference. Continuous variables with a skewed distribution were log-transformed to attain normal distributions. A linear mixed-effects model analysis was used to identify the factors associated with a change in eGFR, with control for internal correlations and other covariates. A first-order autoregressive error structure accounted for intrapatient correlations. The adjusted covariates included age, diabetes mellitus (DM), hypertension, coronary artery disease, cerebrovascular disease, four study groups, mean arterial pressure (MAP), albumin, log triglyceride, total cholesterol, baseline eGFR, proteinuria (0, 1+, or >1+), LVEF < 50%, and medications including ACEIs or ARBs or both, diuretics, and hypouricemic agents. A difference was considered significant at P < 0.05.

results
A total of 540 nondialyzed patients with CKD were included in the study. The mean age of the patients was 66.3 ± 12.4 years, and there were 332 males and 208 females. The value of the slope of eGFR for the study group of patients was -1.51 ± 0.13 ml/min/1.73 m 2 /year. The number of serum creatinine measurements during the follow-up period was 8 measurements (25th-75th percentile: 5-12 measurements). The study patients were stratified into four groups according to sex-specific median values of UA (male, 7.9 mg/dl; female, 7.45 mg/dl) and LVMI (male, 134.7 g/m 2 ; female, 132.5 g/ m 2 ). The UA was positively correlated with LVMI (r = 0.173, P < 0.001). The comparison of clinical characteristics among study groups is shown in Table 1. Figure 1 shows the eGFR slopes of the four study groups.

risk oF progression to dialysis
The follow-up period in the study was 33.4 months (range, 19.8-39.6 months). During the follow-up period, 86 patients (15.9%) had initiation of hemodialysis. Table 2 shows the HR estimates for progression to dialysis, the OR estimates for rapid progression, and the unstandardized coefficient β estimates for change in eGFR after adjustment for age, DM, hypertension, coronary artery disease, cerebrovascular disease, four study groups, MAP, albumin, log triglyceride, total cholesterol, baseline eGFR, proteinuria (0, 1+, or > 1+), LVEF < 50%, and medications including ACEIs and/or ARBs, diuretics, and hypouricemic agents. The multivariate regression analysis showed that the combination of higher UA and LVMI was associated with a significant increase in progression to commencement of dialysis as compared with the combination of lower UA and LVMI (HR, 1.830; 95% CI, 1.007-3.326; P = 0.048). The interaction between higher UA and LVMI in the value of renal end point was statistically significant (HR, 2.147; 95% CI, 1.347-3.424; P < 0.01).
We also used the fasting value of blood glucose instead of the diagnosis of DM in our multivariate analysis, and found a weak association between the combination of higher UA and LVMI and progression to dialysis (HR, 1.802; 95% CI, 1.000-3.246; P = 0.05 vs. the combination of lower UA and LVMI).

relation of study groups to rapid progression of decline in renal function
In our investigation of the relationship between study group and rapid progression of decline in renal function (Table 2), the combination of higher UA and LVMI as compared with the combination of lower UA and LVMI was significantly associated with a rapid progression of decline in renal function in the multivariate logistic analysis used to assess this (OR, 2.231; 95% CI, 1.058-4.705; P = 0.04). The interaction between higher UA and LVMI in the rapidity of progression of decline in renal function was also statistically significant (OR, 2.431; 95% CI, 1.546-3.822; P < 0.001).
In our use of fasting glucose instead of the diagnosis of DM in the multivariate analysis for evaluating the relation of each of the study groups to rapid progression of decline in renal function, we found that this relation was attenuated as compared with the combination of lower UA and LVMI, but that it still remained significant (OR, 2.130; 95% CI, 1.014-4.475; P = 0.046). Table 2 shows the effect of study group on the change in eGFR in the linear mixed-effects model. The combination of higher UA and lower LVMI, the combination of lower UA and higher LVMI, and the combination of higher UA and LVMI were all associated with a significant decrease in eGFR over time (all P < 0.001). The decrease in eGFR over time was slower in the group with lower UA and LVMI than in the other groups (P < 0.001) and was more rapid in the group with higher UA and LVMI than in the other groups (P < 0.04). However, the decrease in eGFR over time did not differ in the group with higher UA and lower LVMI and the group with lower UA and higher LVMI (P = 0.54).

relation of subgroup analysis to renal outcomes
We further performed a subgroup analysis on patients not treated with hypouricemic agents (n = 450), and found that the combination of higher UA and LVMI, as compared with the combination of lower UA and LVMI, had no correlation with progression to dialysis or rapid progression of decline in renal function.

discussion
In the present study, we evaluated the association of UA and LVMI with renal outcomes in patients with CKD of stages 3-5. As compared with the combination of lower UA and LVMI, the combination of higher UA and LVMI was associated with progression to dialysis, even after adjustment for other risk factors. The combination of higher UA and LVMI was also associated with rapid progression of decline in renal function in our subjects. Additionally, the linear mixed-effects model showed a more rapid decrease in eGFR over time in the group with higher UA and LVMI than in the other groups.
In experimental and in vitro studies, the putative toxic mechanisms of hyperuricemia include mediating inflammation, inducing endothelial-cell dysfunction, and stimulating vascular smooth-muscle proliferation, all of which are known to induce cardiac hypertrophy. [26][27][28][29] In accord with the experimental model, hyperuricemia in humans has proven to be associated with LVH. 14,30 Our study also showed serum UA had a positive correlation with LVMI. Additionally, renal dysfunction increases serum UA and activates the reninangiotensin system, which may initiate or accelerate LVH. 31 Hence, there is a vicious cycle between UA, LVH, and CKD. In our study, as compared with the group with lower UA and LVMI, the group with higher UA and LVMI showed an Abbreviations: UA, uric acid; LVMI, left ventricular mass index; eGFR, estimated glomerular filtration rate; ACEI, angiotensin converting enzyme inhibitor; ARB, angiotensin II receptor blocker; LVEF, left ventricular ejection fraction; LVRWT, left ventricular relative wall thickness. The study patients were stratified into four groups according to sex-specific median values of UA (male: 7.9 mg/dl; female: 7.45 mg/dl) and LVMI (male: 134.7 g/m 2 ; female: 132.5 g/m 2 ). A test result of 1+ or more of proteinuria was defined as positive. * P < 0.05 compared with lower UA and LVMI; † P < 0.05 compared with higher UA and lower LVMI; # P < 0.05 compared with lower UA and higher LVMI.
increased risk for renal morbidity, such as a higher prevalence of hypertension, higher MAP, wider pulse pressure, lower serum albumin, higher fasting glucose, worse dyslipidemia, anemia, and lower eGFR. Even after adjustment for these confounding factors, the group with higher UA and LVMI was associated with progression to dialysis and rapid progression of decline in renal function. Hence, hyperuricemia and LVH might have a synergistic effect on adverse renal outcomes.
Many epidemiological studies have attempted to identify whether hyperuricemia is a risk factor for adverse renal outcomes, but achieved results in CKD that were limited and inconsistent. [9][10][11]18,32 Zoppini et al. investigated 194 patients with type 2 DM who had normal renal function and found that hyperuricemia was a risk factor for the development of incident CKD during a 5-year follow-up period. 8 Mok et al. evaluated the association between serum UA and CKD risk based on data collected over a 10.2-year period and found that a rank in the highest quartile of serum UA increased the risk of developing CKD. 7 One important finding in our study was that in a comparison of the two groups with a higher LVMI, only the group with a higher UA was associated with rapid progression of decline in renal function. This result might have been produced by higher triglyceride levels and greater diuretic use in the group with higher levels of UA, both of which factors were established risk factors for rapid progression of decline in renal function. 18,33 In the MDRD trial, during a median follow-up of 10 years, Madero et al. examined the effects of strict blood pressure control and dietary protein restriction on renal function in patients with predominantly nondiabetic stages 3 and 4 CKD. Their results demonstrated that hyperuricemia did not increase the risk of end-stage renal disease. 10 Our study evaluated a cohort of patients with CKD of stages 3-5 including diabetic and nondiabetic patients and verified that the combination of a higher UA and higher LVMI increased the risk for progression to dialysis, even after adjusting for demographic, clinical, and biochemical parameters. The inconsistency of the results the MDRD trial and our study might be related to a population with a wider range of CKD stages and the greater number of diabetic patients in our study.
In our study, patients with a lower UA and higher LVMI (21.1%) and those with a higher UA and lower LVMI (21.3%) received a similar percentage of hypouricemic agents. The reason for the high percentage of use of hypouricemic agents in the group with a lower UA and higher LVMI might Figure 1. Slopes of eGFR in the four study groups, of -0.35 ± 0.19, -1.40 ± 0.27, -1.76 ± 0.31, and -2.59 ± 0.24 ml/min/1.73 m 2 /year, respectively. The slope of eGFR was higher in the group with a lower UA and lower LVMI than in the other groups (P < 0.020) and lower in the group with a higher UA and LVMI than in the group with a higher UA and lower LVMI (P = 0.006). The study patients were stratified into four groups according to sex-specific median values of UA (male, 7.9 mg/dl; female, 7.45 mg/dl) and LVMI (male: 134.7 g/m 2 ; female: 132.5 g/m 2 ). *P < 0.05 compared with the group with a lower UA and LVMI; †P < 0.05 compared with the group with a higher UA and lower LVMI. Multivariate model: adjusted for age, diabetes mellitus, hypertension, coronary artery disease, cerebrovascular disease, 4 study groups, mean arterial pressure, albumin, log triglyceride, total cholesterol, baseline eGFR, proteinuria (negative, 1+, and > 1+), LVEF < 50%, and medications including ACEIs and/or ARBs, diuretics, and hypouricemic agents. The study patients were stratified into 4 groups according to sex-specific median values of UA (male: 7.9; female: 7.45 mg/dL) and LVMI (male: 134.7; female: 132.5 g/m 2 ). Abbreviations are the same as in Table 1. have been be related to a high prevalence of a concomitant diagnosis of gout (23.8%). The high percentage of use of hypouricemic agents might partly explain the lower serum levels of UA in this group.
Because the number and interval of serum creatinine measurements in our study varied for different patients, we applied the linear mixed-effects model directly to repeated measurements of eGFR to avoid a biased and inefficient estimation. In addition, because the qualitative assessment of proteinuria was not determined on first morning voided urines, the data for proteinuria in the study might have been influenced by hydration status.
Our results demonstrated that the coexistence of hyperuricemia and increased LVMI was associated with progression to dialysis and rapid progression of decline in renal function in patients with CKD of stages 3-5. Assessments of serum UA and LVMI may be useful in identifying a group of patients at high-risk for adverse outcomes of renal function in patients with CKD.