Effects of uric acid on kidney function decline differ depending on baseline kidney function in type 2 diabetic patients

Effects of uric acid on kidney function decline differ depending on baseline kidney function in... Abstract Background Most existing data regarding effects of uric acid (UA) on diabetic kidney disease have considered patients with preserved kidney function. We examined a hypothesis that there are differences in the effects of serum UA levels on the decline in kidney function depending on baseline kidney function in diabetic patients. Methods In this historical cohort study, 7033 type 2 diabetic patients were analyzed and classified into two groups as follows: nonchronic kidney disease (non-CKD), with an estimated glomerular filtration rate (eGFR) ≥60 mL/min/1.73 m2 (n = 4994), and CKD, with an eGFR <60 mL/min/1.73 m2 (n = 2039). The composite endpoint was a ≥30% decrease in eGFR from baseline or the initiation of renal replacement therapy. The hazard ratio (HR) of serum UA levels at baseline was estimated using multivariate Cox proportional hazards models. Results There was a significant interaction between UA levels and baseline eGFR with respect to the endpoint (P < 0.001). The HRs of 1 mg/dL increase in UA levels were 1.13 [95% confidence interval (CI) 1.05–1.22, P = 0.002] and 0.93 (95% CI 0.88–0.99, P = 0.02) in the non-CKD and CKD groups, respectively. When patients were classified by quintile of UA levels, the HRs of those in the 5th quintile (versus 1st quintile) were 1.64 (95% CI 1.23–2.18, P < 0.001) and 0.76 (95% CI 0.58–0.99, P = 0.05) in the non-CKD and CKD groups, respectively. Conclusions The effects of UA on kidney function decline might differ depending on baseline kidney function in type 2 diabetic patients. High UA levels are the prognostic factor only in patients with preserved kidney function. diabetic kidney disease, diabetic nephropathy, hyperuricemia, type 2 diabetes, uric acid INTRODUCTION Diabetic kidney disease (DKD) remains a major global health concern despite remarkable advances in the management of hyperglycemia and hypertension, which are two major established risk factors [1]. Uric acid (UA) has been demonstrated in basic research to stimulate inflammation, oxidative stress, endothelial damage and vasoconstriction [2]. The harmful effects of UA on human kidneys as well as on atherosclerotic cardiovascular disease and mortality have received a great deal of attention [2–4]. A number of studies, the majority of which considered patients with preserved kidney function, have shown an association between elevated serum UA levels and the development and progression of DKD [5–7]; however, the cutoff and target values remain unknown. Serum UA levels rise in patients with kidney insufficiency because excretion from the kidneys is responsible for much of the clearance of UA [8]. However, such increases might be mitigated by reduction of the biosynthesis of UA and upregulation of its secretion by the colon [9, 10]. The majority of studies in patients with kidney insufficiency have not reported an association between baseline serum UA levels and subsequent decline in kidney function [11–13], unlike studies in patients with preserved kidney function, suggesting that the effects of UA on the kidneys differ depending on kidney function. To our knowledge, information regarding the association of UA with the progression of DKD in patients with kidney insufficiency is limited. In the present study we aimed to examine a hypothesis that there are differences in the effects of serum UA levels on the decline in kidney function depending on baseline kidney function in patients with diabetes and, if any differences exist, to clarify the distinctions between groups classified based on kidney function. MATERIALS AND METHODS Study design This single-center, historical cohort study was designed in adherence with the Declaration of Helsinki and was performed as part of the Cohort Study Elucidating Factors Associated with the Pathogenesis and Prognosis of Diabetic Kidney Disease conducted in the Diabetes Center, Tokyo Women’s Medical University School of Medicine. The Ethics Committee of Tokyo Women’s Medical University School of Medicine approved the protocol. As this was a historical cohort study, not a prospective interventional study, the Ethics Committee approved a waiver of informed consent. Data used in the present study were obtained from the clinical information system (electronic medical records). Participants We initially identified 7430 Japanese ambulatory patients with type 2 diabetes who were ≥18 years of age; did not receive renal replacement therapy (RRT), which meant chronic dialysis or kidney transplant; and had available data on body weight, blood pressure, serum UA, hemoglobin A1c (HbA1c), lipid parameters, creatinine and urinary albumin simultaneously measured at a regular ambulatory visit in the Diabetes Center of Tokyo Women’s Medical University School of Medicine from 1 January 2004 to 30 April 2016. The earliest available data were set as baseline if there were multiple measurements available for subjects during the above-mentioned period. A flow diagram of the study population is presented in Figure 1. At baseline, pregnant patients (n = 52) or patients with malignant disease (n = 91) were excluded. Patients who had undergone removal of one kidney owing to cancer or transplantation before baseline (n = 8) or those with glomerular nephritis proven via biopsy before baseline (n = 1), acute kidney injury or postrenal failure at baseline (n = 2) or acute severe infection at baseline (n = 2) were excluded. Patients with no follow-up serum creatinine data (n = 234) or missing baseline profile values (n = 7) were also excluded. Overall, the data obtained from 7033 patients were analyzed and these patients were classified into two groups based on kidney function: nonchronic kidney disease (non-CKD), with an estimated glomerular filtration rate (eGFR) ≥60 mL/min/1.73 m2 (n = 4994), and CKD, with an eGFR <60 mL/min/1.73 m2 (n = 2039). FIGURE 1 View largeDownload slide A flow diagram of the study population. FIGURE 1 View largeDownload slide A flow diagram of the study population. Measurements Laboratory data were obtained from random spot blood samples and first morning urine samples. Serum UA, triglycerides and creatinine levels were determined using enzymatic methods. Levels of serum high-density lipoprotein (HDL) cholesterol were measured using polyethylene glycol–pretreated enzymes and levels of serum low-density lipoprotein (LDL) cholesterol were determined using enzymatic methods or the Friedewald formula (if triglycerides <400 mg/dL). HbA1c values obtained as Japan Diabetes Society (JDS) values were converted to National Glycohemoglobin Standardization Program (NGSP) values as follows: HbA1c (%) = 1.02 × HbA1c (JDS) (%) + 0.25% [14]. Levels of urinary creatinine were determined using enzymatic methods. Levels of urinary albumin were determined using the latex agglutination method and normalized by urinary creatinine concentrations. The GFR was estimated using the following formula, proposed by the Japanese Society for Nephrology: eGFR (mL/min/1.73 m2) = 194 × age (years)−0.287 × serum creatinine level (in mg/dL)−1.094 × (0.739, for women) [15]. Endpoints and follow-up The primary composite endpoint was a decrease in eGFR of ≥30% from baseline, established as a surrogate for the development of kidney failure [16], or the initiation of RRT, whichever came first. We also verified a ≥50% decrease in eGFR instead of a ≥30% decrease in eGFR as the secondary composite endpoint. We determined that the endpoint was reached when the above-mentioned criteria were met for at least 3 months to avoid capturing episodes of reversible acute kidney injury. Patients were censored when they initiated RRT, were deceased before the initiation of RRT, were lost to follow-up or reached the administrative censoring date (31 August 2016). Statistical analysis Categorical data were expressed as number (%). Continuous variables were expressed as the arithmetic mean [standard deviation (SD)] or median [interquartile range (IQR)], as appropriate. For statistical analyses, Fisher’s exact probability test, Student’s t-test, Mann–Whitney U test and Cox proportional hazards models were used as appropriate. The Kaplan–Meier method was used for an estimation of survival probability for endpoints and the statistical differences between the groups were compared using the log-rank test. Possible nonlinear associations between UA levels and the endpoint were examined using the multivariable-adjusted restricted cubic spline model [17], with the knots placed at the 5th, 35th, 65th and 95th percentile levels. We also used the Fine and Gray subdistribution hazards model instead of the traditional Cox proportional hazards model in a time-to-event analysis, as latter methods may overestimate the risk of events by undermining the competing risks [18, 19]. In the present study, death before reaching the endpoint was treated as a competing risk. Three sets of models were used for the calculation of the hazard ratio (HR) [95% confidence interval (CI)] of UA for the endpoint: univariate models (Model 1); models adjusted by age, sex and eGFR (Model 2) and models adjusted by the same variables as in Model 2 plus duration of diabetes, use of insulin, use of angiotensin-converting enzyme inhibitors or angiotensin receptor blockers, use of lipid-lowering drugs, use of urate-lowering drugs, smoking status (current or former), date of the baseline data (2004–2009 or 2010–2016), body mass index (BMI), systolic blood pressure, number of creatinine measurements during the follow-up period, HbA1c, triglycerides, LDL cholesterol and urinary albumin:creatinine ratio (Model 3). Triglycerides and urinary albumin:creatinine ratio levels transformed into common logarithm values were used in the analyses. P-values < 0.05 were considered significant. All analyses were completed using SAS version 9.4 (SAS Institute, Cary, NC, USA). RESULTS We analyzed 7033 patients, comprising 2493 women and 4540 men with a mean age of 61 (SD 12) years. Baseline characteristics for the overall and categorized cohort are presented in Table 1. During the median follow-up period of 5.5 (IQR 2.5–9.4) years, 1189 patients experienced a ≥30% decrease in eGFR from baseline, 529 experienced a  ≥50% decrease in eGFR from baseline and RRT was initiated in 440 patients. Prior to the initiation of RRT, 284 patients died. Consequently, 1250 patients, consisting of 1189 with a ≥30% decrease in eGFR and 61 who initiated RRT before experiencing a ≥30% decrease in eGFR, reached the primary composite endpoint; 561 and 689 were in the non-CKD and CKD groups, respectively. A total of 663 patients, consisting of 529 with a  ≥50% decrease in eGFR and 134 who initiated RRT before experiencing a  ≥50% decrease in eGFR, reached the secondary composite endpoint; 166 and 497 were in the non-CKD and CKD groups, respectively. Patients in the CKD group had greater incidences of both composite endpoints than those in the non-CKD group (both P < 0.001; Supplementary data, Figure S1). The median number of creatinine measurements used for the analyses was 18 (IQR 9–33) and 21 (IQR 10–37) for the primary and secondary composite endpoints, respectively. Table 1 Baseline demographic and laboratory data Characteristic Overall (n = 7033) Non-CKD (n = 4994) CKD (n = 2039) P-value (Non-CKD versus CKD) Age (years) 61 (12) 59 (12) 66 (11) <0.001 Gender (men), n (%) 4540 (64.6) 3144 (63.0) 1396 (68.5) <0.001 Duration of diabetes (years) 13 (9) 12 (9) 16 (10) <0.001 Diabetes therapy, n (%)  Insulin 2295 (32.6) 1393 (27.9) 902 (44.2) <0.001  Other medication 4220 (60.0) 3096 (62.0) 1124 (55.1) <0.001 ACE inhibitors or ARBs, n (%) 3135 (44.6) 1740 (34.8) 1395 (68.4) <0.001 Other antihypertensive drugs, n (%) 2415 (34.3) 1283 (25.7) 1132 (55.5) <0.001 Lipid-lowering drugs, n (%) 2607 (37.1) 1604 (32.1) 1003 (49.2) <0.001 Urate-lowering drugs, n (%) 763 (10.9) 228 (4.6) 535 (26.2) <0.001 History of ASCVD, n (%) 1443 (20.5) 734 (14.7) 709 (34.8) <0.001 Former or current smoker, n (%) 3774 (53.7) 2650 (53.1) 1124 (55.1) 0.120 Date of baseline data, n (%) <0.001  2004–09 4380 (62.3) 3215 (64.4) 1165 (57.1)  2010–16 2653 (37.7) 1779 (35.6) 874 (42.9) BMI (kg/m2) 25.0 (4.2) 25.0 (4.3) 25.2 (3.9) 0.036 SBP (mmHg) 136 (20) 135 (19) 140 (22) <0.001 DBP (mmHg) 76 (12) 77 (11) 74 (13) <0.001 Laboratory data  UA (mg/dL) 5.4 (1.4) 5.1 (1.3) 6.2 (1.4) <0.001  HbA1c (%) 7.6 (1.3) 7.7 (1.3) 7.4 (1.2) <0.001  Triglycerides (mg/dL) median (IQR) 125 (85–185) 120 (81–179) 136 (97–198) <0.001  HDL cholesterol (mg/dL) 55 (15) 56 (15) 52 (15) <0.001  LDL cholesterol (mg/dL) 115 (31) 116 (30) 111 (33) <0.001  Creatinine (mg/dL) 0.96 (0.73) 0.72 (0.14) 1.55 (1.13) <0.001  eGFR (mL/min/1.73 m2) 69.9 (23.0) 80.8 (15.6) 43.0 (14.8) <0.001  ACR (mg/g) median (IQR) 14.8 (7.5–68.2) 11.6 (6.9–29.8) 72.2 (13.1–931.8) <0.001 Characteristic Overall (n = 7033) Non-CKD (n = 4994) CKD (n = 2039) P-value (Non-CKD versus CKD) Age (years) 61 (12) 59 (12) 66 (11) <0.001 Gender (men), n (%) 4540 (64.6) 3144 (63.0) 1396 (68.5) <0.001 Duration of diabetes (years) 13 (9) 12 (9) 16 (10) <0.001 Diabetes therapy, n (%)  Insulin 2295 (32.6) 1393 (27.9) 902 (44.2) <0.001  Other medication 4220 (60.0) 3096 (62.0) 1124 (55.1) <0.001 ACE inhibitors or ARBs, n (%) 3135 (44.6) 1740 (34.8) 1395 (68.4) <0.001 Other antihypertensive drugs, n (%) 2415 (34.3) 1283 (25.7) 1132 (55.5) <0.001 Lipid-lowering drugs, n (%) 2607 (37.1) 1604 (32.1) 1003 (49.2) <0.001 Urate-lowering drugs, n (%) 763 (10.9) 228 (4.6) 535 (26.2) <0.001 History of ASCVD, n (%) 1443 (20.5) 734 (14.7) 709 (34.8) <0.001 Former or current smoker, n (%) 3774 (53.7) 2650 (53.1) 1124 (55.1) 0.120 Date of baseline data, n (%) <0.001  2004–09 4380 (62.3) 3215 (64.4) 1165 (57.1)  2010–16 2653 (37.7) 1779 (35.6) 874 (42.9) BMI (kg/m2) 25.0 (4.2) 25.0 (4.3) 25.2 (3.9) 0.036 SBP (mmHg) 136 (20) 135 (19) 140 (22) <0.001 DBP (mmHg) 76 (12) 77 (11) 74 (13) <0.001 Laboratory data  UA (mg/dL) 5.4 (1.4) 5.1 (1.3) 6.2 (1.4) <0.001  HbA1c (%) 7.6 (1.3) 7.7 (1.3) 7.4 (1.2) <0.001  Triglycerides (mg/dL) median (IQR) 125 (85–185) 120 (81–179) 136 (97–198) <0.001  HDL cholesterol (mg/dL) 55 (15) 56 (15) 52 (15) <0.001  LDL cholesterol (mg/dL) 115 (31) 116 (30) 111 (33) <0.001  Creatinine (mg/dL) 0.96 (0.73) 0.72 (0.14) 1.55 (1.13) <0.001  eGFR (mL/min/1.73 m2) 69.9 (23.0) 80.8 (15.6) 43.0 (14.8) <0.001  ACR (mg/g) median (IQR) 14.8 (7.5–68.2) 11.6 (6.9–29.8) 72.2 (13.1–931.8) <0.001 Data are expressed as mean (SD) unless stated otherwise. non-CKD: eGFR ≥60 mL/min/1.73 m2; CKD: eGFR <60 mL/min/1.73 m2. ACE, angiotensin-converting enzyme; ARBs, angiotensin receptor blockers; ASCVD, atherosclerotic cardiovascular disease; SBP, systolic blood pressure; DBP, diastolic blood pressure; ACR, albumin:creatinine ratio. Table 1 Baseline demographic and laboratory data Characteristic Overall (n = 7033) Non-CKD (n = 4994) CKD (n = 2039) P-value (Non-CKD versus CKD) Age (years) 61 (12) 59 (12) 66 (11) <0.001 Gender (men), n (%) 4540 (64.6) 3144 (63.0) 1396 (68.5) <0.001 Duration of diabetes (years) 13 (9) 12 (9) 16 (10) <0.001 Diabetes therapy, n (%)  Insulin 2295 (32.6) 1393 (27.9) 902 (44.2) <0.001  Other medication 4220 (60.0) 3096 (62.0) 1124 (55.1) <0.001 ACE inhibitors or ARBs, n (%) 3135 (44.6) 1740 (34.8) 1395 (68.4) <0.001 Other antihypertensive drugs, n (%) 2415 (34.3) 1283 (25.7) 1132 (55.5) <0.001 Lipid-lowering drugs, n (%) 2607 (37.1) 1604 (32.1) 1003 (49.2) <0.001 Urate-lowering drugs, n (%) 763 (10.9) 228 (4.6) 535 (26.2) <0.001 History of ASCVD, n (%) 1443 (20.5) 734 (14.7) 709 (34.8) <0.001 Former or current smoker, n (%) 3774 (53.7) 2650 (53.1) 1124 (55.1) 0.120 Date of baseline data, n (%) <0.001  2004–09 4380 (62.3) 3215 (64.4) 1165 (57.1)  2010–16 2653 (37.7) 1779 (35.6) 874 (42.9) BMI (kg/m2) 25.0 (4.2) 25.0 (4.3) 25.2 (3.9) 0.036 SBP (mmHg) 136 (20) 135 (19) 140 (22) <0.001 DBP (mmHg) 76 (12) 77 (11) 74 (13) <0.001 Laboratory data  UA (mg/dL) 5.4 (1.4) 5.1 (1.3) 6.2 (1.4) <0.001  HbA1c (%) 7.6 (1.3) 7.7 (1.3) 7.4 (1.2) <0.001  Triglycerides (mg/dL) median (IQR) 125 (85–185) 120 (81–179) 136 (97–198) <0.001  HDL cholesterol (mg/dL) 55 (15) 56 (15) 52 (15) <0.001  LDL cholesterol (mg/dL) 115 (31) 116 (30) 111 (33) <0.001  Creatinine (mg/dL) 0.96 (0.73) 0.72 (0.14) 1.55 (1.13) <0.001  eGFR (mL/min/1.73 m2) 69.9 (23.0) 80.8 (15.6) 43.0 (14.8) <0.001  ACR (mg/g) median (IQR) 14.8 (7.5–68.2) 11.6 (6.9–29.8) 72.2 (13.1–931.8) <0.001 Characteristic Overall (n = 7033) Non-CKD (n = 4994) CKD (n = 2039) P-value (Non-CKD versus CKD) Age (years) 61 (12) 59 (12) 66 (11) <0.001 Gender (men), n (%) 4540 (64.6) 3144 (63.0) 1396 (68.5) <0.001 Duration of diabetes (years) 13 (9) 12 (9) 16 (10) <0.001 Diabetes therapy, n (%)  Insulin 2295 (32.6) 1393 (27.9) 902 (44.2) <0.001  Other medication 4220 (60.0) 3096 (62.0) 1124 (55.1) <0.001 ACE inhibitors or ARBs, n (%) 3135 (44.6) 1740 (34.8) 1395 (68.4) <0.001 Other antihypertensive drugs, n (%) 2415 (34.3) 1283 (25.7) 1132 (55.5) <0.001 Lipid-lowering drugs, n (%) 2607 (37.1) 1604 (32.1) 1003 (49.2) <0.001 Urate-lowering drugs, n (%) 763 (10.9) 228 (4.6) 535 (26.2) <0.001 History of ASCVD, n (%) 1443 (20.5) 734 (14.7) 709 (34.8) <0.001 Former or current smoker, n (%) 3774 (53.7) 2650 (53.1) 1124 (55.1) 0.120 Date of baseline data, n (%) <0.001  2004–09 4380 (62.3) 3215 (64.4) 1165 (57.1)  2010–16 2653 (37.7) 1779 (35.6) 874 (42.9) BMI (kg/m2) 25.0 (4.2) 25.0 (4.3) 25.2 (3.9) 0.036 SBP (mmHg) 136 (20) 135 (19) 140 (22) <0.001 DBP (mmHg) 76 (12) 77 (11) 74 (13) <0.001 Laboratory data  UA (mg/dL) 5.4 (1.4) 5.1 (1.3) 6.2 (1.4) <0.001  HbA1c (%) 7.6 (1.3) 7.7 (1.3) 7.4 (1.2) <0.001  Triglycerides (mg/dL) median (IQR) 125 (85–185) 120 (81–179) 136 (97–198) <0.001  HDL cholesterol (mg/dL) 55 (15) 56 (15) 52 (15) <0.001  LDL cholesterol (mg/dL) 115 (31) 116 (30) 111 (33) <0.001  Creatinine (mg/dL) 0.96 (0.73) 0.72 (0.14) 1.55 (1.13) <0.001  eGFR (mL/min/1.73 m2) 69.9 (23.0) 80.8 (15.6) 43.0 (14.8) <0.001  ACR (mg/g) median (IQR) 14.8 (7.5–68.2) 11.6 (6.9–29.8) 72.2 (13.1–931.8) <0.001 Data are expressed as mean (SD) unless stated otherwise. non-CKD: eGFR ≥60 mL/min/1.73 m2; CKD: eGFR <60 mL/min/1.73 m2. ACE, angiotensin-converting enzyme; ARBs, angiotensin receptor blockers; ASCVD, atherosclerotic cardiovascular disease; SBP, systolic blood pressure; DBP, diastolic blood pressure; ACR, albumin:creatinine ratio. Association of serum UA levels with the incidence of the primary and secondary endpoint There was a significant interaction between UA levels and baseline eGFR with respect to the primary composite endpoint in the analysis adjusted by variables used for Model 3 (P-interaction < 0.001). The same result was obtained for the secondary composite endpoint (P-interaction = 0.004). In the non-CKD group, high levels of UA were a risk factor for both endpoints (Table 2). In the CKD group, the univariate analyses showed high UA levels to be a risk factor for the endpoint; however, in the multivariate models, low levels of UA were associated with the incidence of both endpoints (Table 2). When subjects were classified by quintile of UA levels within the non-CKD and CKD groups, those in the 5th quintile (versus the 1st quintile) in the non-CKD group had a significantly higher risk for the primary and secondary composite endpoint in all models (Table 3). On the contrary, although patients in the 2nd–5th quintile (versus the 1st quintile) in the CKD group had a greater risk for both endpoints in the univariate models, the HRs for those in the 5th quintile for the primary endpoint were significantly low in the multivariate models (Table 3). Finally, the restricted cubic spline curves (95% CI) of the association between UA levels and the composite endpoints are shown in Figure 2, following adjustment with variables used for Model 3. In both groups, a nonlinear association between UA levels and the endpoints was rejected. Table 2 HR of 1 mg/dL increase in serum UA levels for each composite endpoint Non-CKD (n = 4994) CKD (n = 2039) Models HR (95% CI) P-value HR (95% CI) P-value ≥30% decrease in eGFR or RRT  Model 1 1.11 (1.04–1.19) 0.003 1.30 (1.23–1.37) <0.001  Model 2 1.20 (1.12–1.30) <0.001 0.93 (0.88–0.98) 0.011  Model 3 1.13 (1.05–1.22) 0.002 0.93 (0.88–0.99) 0.022 ≥50% decrease in eGFR or RRT  Model 1 1.34 (1.18–1.51) <0.001 1.35 (1.27–1.43) <0.001  Model 2 1.36 (1.19–1.55) <0.001 0.90 (0.84–0.97) 0.003  Model 3 1.20 (1.05–1.37) 0.009 0.90 (0.84–0.97) 0.005 Non-CKD (n = 4994) CKD (n = 2039) Models HR (95% CI) P-value HR (95% CI) P-value ≥30% decrease in eGFR or RRT  Model 1 1.11 (1.04–1.19) 0.003 1.30 (1.23–1.37) <0.001  Model 2 1.20 (1.12–1.30) <0.001 0.93 (0.88–0.98) 0.011  Model 3 1.13 (1.05–1.22) 0.002 0.93 (0.88–0.99) 0.022 ≥50% decrease in eGFR or RRT  Model 1 1.34 (1.18–1.51) <0.001 1.35 (1.27–1.43) <0.001  Model 2 1.36 (1.19–1.55) <0.001 0.90 (0.84–0.97) 0.003  Model 3 1.20 (1.05–1.37) 0.009 0.90 (0.84–0.97) 0.005 Model 1: univariate. Model 2: adjusted by age, sex and eGFR. Model 3: adjusted by age, sex, eGFR, duration of diabetes, use of insulin, use of angiotensin-converting enzyme inhibitors or angiotensin receptor blockers, use of lipid-lowering drugs, use of urate-lowering drugs, smoking status, date of the baseline data, BMI, systolic blood pressure, number of creatinine measurements during the follow-up period, Hb A1c, logarithmically transformed triglycerides, LDL cholesterol and logarithmically transformed urinary albumin:creatinine ratio. Table 2 HR of 1 mg/dL increase in serum UA levels for each composite endpoint Non-CKD (n = 4994) CKD (n = 2039) Models HR (95% CI) P-value HR (95% CI) P-value ≥30% decrease in eGFR or RRT  Model 1 1.11 (1.04–1.19) 0.003 1.30 (1.23–1.37) <0.001  Model 2 1.20 (1.12–1.30) <0.001 0.93 (0.88–0.98) 0.011  Model 3 1.13 (1.05–1.22) 0.002 0.93 (0.88–0.99) 0.022 ≥50% decrease in eGFR or RRT  Model 1 1.34 (1.18–1.51) <0.001 1.35 (1.27–1.43) <0.001  Model 2 1.36 (1.19–1.55) <0.001 0.90 (0.84–0.97) 0.003  Model 3 1.20 (1.05–1.37) 0.009 0.90 (0.84–0.97) 0.005 Non-CKD (n = 4994) CKD (n = 2039) Models HR (95% CI) P-value HR (95% CI) P-value ≥30% decrease in eGFR or RRT  Model 1 1.11 (1.04–1.19) 0.003 1.30 (1.23–1.37) <0.001  Model 2 1.20 (1.12–1.30) <0.001 0.93 (0.88–0.98) 0.011  Model 3 1.13 (1.05–1.22) 0.002 0.93 (0.88–0.99) 0.022 ≥50% decrease in eGFR or RRT  Model 1 1.34 (1.18–1.51) <0.001 1.35 (1.27–1.43) <0.001  Model 2 1.36 (1.19–1.55) <0.001 0.90 (0.84–0.97) 0.003  Model 3 1.20 (1.05–1.37) 0.009 0.90 (0.84–0.97) 0.005 Model 1: univariate. Model 2: adjusted by age, sex and eGFR. Model 3: adjusted by age, sex, eGFR, duration of diabetes, use of insulin, use of angiotensin-converting enzyme inhibitors or angiotensin receptor blockers, use of lipid-lowering drugs, use of urate-lowering drugs, smoking status, date of the baseline data, BMI, systolic blood pressure, number of creatinine measurements during the follow-up period, Hb A1c, logarithmically transformed triglycerides, LDL cholesterol and logarithmically transformed urinary albumin:creatinine ratio. Table 3 HR (versus the 1st quintile) of other groups classified by quintile of serum UA levels for each composite endpoint Non-CKD (n = 4994) CKD (n = 2039) Models HR (95% CI) P-value HR (95% CI) P-value ≥30% decrease in eGFR or RRT  Model 1   2nd quintile 0.92 (0.71–1.20) 0.541 1.40 (1.06–1.86) 0.018   3rd quintile 0.84 (0.64–1.10) 0.208 1.75 (1.35–2.28) <0.001   4th quintile 0.93 (0.71–1.21) 0.586 2.29 (1.77–2.97) <0.001   5th quintile 1.39 (1.09–1.77) 0.009 2.67 (2.08–3.43) <0.001  Model 2   2nd quintile 1.00 (0.76–1.30) 0.973 0.82 (0.62–1.09) 0.175   3rd quintile 0.96 (0.73–1.27) 0.795 0.91 (0.69–1.19) 0.476   4th quintile 1.15 (0.87–1.52) 0.324 0.88 (0.67–1.16) 0.358   5th quintile 1.80 (1.37–2.36) <0.001 0.75 (0.57–0.98) 0.034  Model 3   2nd quintile 1.16 (0.89–1.52) 0.266 0.81 (0.61–1.09) 0.162   3rd quintile 1.09 (0.82–1.45) 0.544 0.84 (0.64–1.11) 0.219   4th quintile 1.22 (0.91–1.62) 0.184 0.90 (0.68–1.18) 0.442   5th quintile 1.64 (1.23–2.18) <0.001 0.76 (0.58–0.99) 0.046 ≥50% decrease in eGFR or RRT  Model 1   2nd quintile 1.13 (0.66–1.93) 0.651 1.64 (1.16–2.34) 0.006   3rd quintile 0.92 (0.52–1.62) 0.772 2.10 (1.51–2.92) <0.001   4th quintile 1.41 (0.85–2.35) 0.189 2.71 (1.96–3.75) <0.001   5th quintile 2.47 (1.55–3.93) <0.001 3.37 (2.46–4.61) <0.001  Model 2   2nd quintile 1.15 (0.67–1.97) 0.603 0.81 (0.57–1.16) 0.254   3rd quintile 0.94 (0.53–1.69) 0.844 0.83 (0.59–1.17) 0.284   4th quintile 1.46 (0.85–2.50) 0.168 0.79 (0.56–1.11) 0.178   5th quintile 2.56 (1.53–4.27) <0.001 0.69 (0.50–0.96) 0.030  Model 3   2nd quintile 1.44 (0.84–2.48) 0.189 0.81 (0.57–1.17) 0.258   3rd quintile 1.29 (0.71–2.32) 0.404 0.81 (0.58–1.14) 0.227   4th quintile 1.48 (0.85–2.60) 0.167 0.84 (0.60–1.19) 0.328   5th quintile 2.14 (1.25–3.69) 0.006 0.73 (0.52–1.02) 0.068 Non-CKD (n = 4994) CKD (n = 2039) Models HR (95% CI) P-value HR (95% CI) P-value ≥30% decrease in eGFR or RRT  Model 1   2nd quintile 0.92 (0.71–1.20) 0.541 1.40 (1.06–1.86) 0.018   3rd quintile 0.84 (0.64–1.10) 0.208 1.75 (1.35–2.28) <0.001   4th quintile 0.93 (0.71–1.21) 0.586 2.29 (1.77–2.97) <0.001   5th quintile 1.39 (1.09–1.77) 0.009 2.67 (2.08–3.43) <0.001  Model 2   2nd quintile 1.00 (0.76–1.30) 0.973 0.82 (0.62–1.09) 0.175   3rd quintile 0.96 (0.73–1.27) 0.795 0.91 (0.69–1.19) 0.476   4th quintile 1.15 (0.87–1.52) 0.324 0.88 (0.67–1.16) 0.358   5th quintile 1.80 (1.37–2.36) <0.001 0.75 (0.57–0.98) 0.034  Model 3   2nd quintile 1.16 (0.89–1.52) 0.266 0.81 (0.61–1.09) 0.162   3rd quintile 1.09 (0.82–1.45) 0.544 0.84 (0.64–1.11) 0.219   4th quintile 1.22 (0.91–1.62) 0.184 0.90 (0.68–1.18) 0.442   5th quintile 1.64 (1.23–2.18) <0.001 0.76 (0.58–0.99) 0.046 ≥50% decrease in eGFR or RRT  Model 1   2nd quintile 1.13 (0.66–1.93) 0.651 1.64 (1.16–2.34) 0.006   3rd quintile 0.92 (0.52–1.62) 0.772 2.10 (1.51–2.92) <0.001   4th quintile 1.41 (0.85–2.35) 0.189 2.71 (1.96–3.75) <0.001   5th quintile 2.47 (1.55–3.93) <0.001 3.37 (2.46–4.61) <0.001  Model 2   2nd quintile 1.15 (0.67–1.97) 0.603 0.81 (0.57–1.16) 0.254   3rd quintile 0.94 (0.53–1.69) 0.844 0.83 (0.59–1.17) 0.284   4th quintile 1.46 (0.85–2.50) 0.168 0.79 (0.56–1.11) 0.178   5th quintile 2.56 (1.53–4.27) <0.001 0.69 (0.50–0.96) 0.030  Model 3   2nd quintile 1.44 (0.84–2.48) 0.189 0.81 (0.57–1.17) 0.258   3rd quintile 1.29 (0.71–2.32) 0.404 0.81 (0.58–1.14) 0.227   4th quintile 1.48 (0.85–2.60) 0.167 0.84 (0.60–1.19) 0.328   5th quintile 2.14 (1.25–3.69) 0.006 0.73 (0.52–1.02) 0.068 Model 1: univariate. Model 2: adjusted by age, sex and eGFR. Model 3: adjusted by age, sex, eGFR, duration of diabetes, use of insulin, use of angiotensin-converting enzyme inhibitors or angiotensin receptor blockers, use of lipid-lowering drugs, use of urate-lowering drugs, smoking status, date of the baseline data, BMI, systolic blood pressure, number of creatinine measurements during the follow-up period, Hb A1c, logarithmically transformed triglycerides, LDL cholesterol and logarithmically transformed urinary albumin:creatinine ratio. Range of UA levels (mg/dL): non-CKD: 1st quintile (n = 990) 0.6–4.0, 2nd quintile (n = 956) 4.1–4.7, 3rd quintile (n = 1002) 4.8–5.3, 4th quintile (n = 1045) 5.4–6.1, 5th quintile (n = 1001) 6.2–11.6; CKD: 1st quintile (n = 415) 1.2–5.1, 2nd quintile (n = 393) 5.2–5.8, 3rd quintile (n = 433) 5.9–6.5, 4th quintile (n = 379) 6.6–7.2, 5th quintile (n = 419) 7.3–12.1. Table 3 HR (versus the 1st quintile) of other groups classified by quintile of serum UA levels for each composite endpoint Non-CKD (n = 4994) CKD (n = 2039) Models HR (95% CI) P-value HR (95% CI) P-value ≥30% decrease in eGFR or RRT  Model 1   2nd quintile 0.92 (0.71–1.20) 0.541 1.40 (1.06–1.86) 0.018   3rd quintile 0.84 (0.64–1.10) 0.208 1.75 (1.35–2.28) <0.001   4th quintile 0.93 (0.71–1.21) 0.586 2.29 (1.77–2.97) <0.001   5th quintile 1.39 (1.09–1.77) 0.009 2.67 (2.08–3.43) <0.001  Model 2   2nd quintile 1.00 (0.76–1.30) 0.973 0.82 (0.62–1.09) 0.175   3rd quintile 0.96 (0.73–1.27) 0.795 0.91 (0.69–1.19) 0.476   4th quintile 1.15 (0.87–1.52) 0.324 0.88 (0.67–1.16) 0.358   5th quintile 1.80 (1.37–2.36) <0.001 0.75 (0.57–0.98) 0.034  Model 3   2nd quintile 1.16 (0.89–1.52) 0.266 0.81 (0.61–1.09) 0.162   3rd quintile 1.09 (0.82–1.45) 0.544 0.84 (0.64–1.11) 0.219   4th quintile 1.22 (0.91–1.62) 0.184 0.90 (0.68–1.18) 0.442   5th quintile 1.64 (1.23–2.18) <0.001 0.76 (0.58–0.99) 0.046 ≥50% decrease in eGFR or RRT  Model 1   2nd quintile 1.13 (0.66–1.93) 0.651 1.64 (1.16–2.34) 0.006   3rd quintile 0.92 (0.52–1.62) 0.772 2.10 (1.51–2.92) <0.001   4th quintile 1.41 (0.85–2.35) 0.189 2.71 (1.96–3.75) <0.001   5th quintile 2.47 (1.55–3.93) <0.001 3.37 (2.46–4.61) <0.001  Model 2   2nd quintile 1.15 (0.67–1.97) 0.603 0.81 (0.57–1.16) 0.254   3rd quintile 0.94 (0.53–1.69) 0.844 0.83 (0.59–1.17) 0.284   4th quintile 1.46 (0.85–2.50) 0.168 0.79 (0.56–1.11) 0.178   5th quintile 2.56 (1.53–4.27) <0.001 0.69 (0.50–0.96) 0.030  Model 3   2nd quintile 1.44 (0.84–2.48) 0.189 0.81 (0.57–1.17) 0.258   3rd quintile 1.29 (0.71–2.32) 0.404 0.81 (0.58–1.14) 0.227   4th quintile 1.48 (0.85–2.60) 0.167 0.84 (0.60–1.19) 0.328   5th quintile 2.14 (1.25–3.69) 0.006 0.73 (0.52–1.02) 0.068 Non-CKD (n = 4994) CKD (n = 2039) Models HR (95% CI) P-value HR (95% CI) P-value ≥30% decrease in eGFR or RRT  Model 1   2nd quintile 0.92 (0.71–1.20) 0.541 1.40 (1.06–1.86) 0.018   3rd quintile 0.84 (0.64–1.10) 0.208 1.75 (1.35–2.28) <0.001   4th quintile 0.93 (0.71–1.21) 0.586 2.29 (1.77–2.97) <0.001   5th quintile 1.39 (1.09–1.77) 0.009 2.67 (2.08–3.43) <0.001  Model 2   2nd quintile 1.00 (0.76–1.30) 0.973 0.82 (0.62–1.09) 0.175   3rd quintile 0.96 (0.73–1.27) 0.795 0.91 (0.69–1.19) 0.476   4th quintile 1.15 (0.87–1.52) 0.324 0.88 (0.67–1.16) 0.358   5th quintile 1.80 (1.37–2.36) <0.001 0.75 (0.57–0.98) 0.034  Model 3   2nd quintile 1.16 (0.89–1.52) 0.266 0.81 (0.61–1.09) 0.162   3rd quintile 1.09 (0.82–1.45) 0.544 0.84 (0.64–1.11) 0.219   4th quintile 1.22 (0.91–1.62) 0.184 0.90 (0.68–1.18) 0.442   5th quintile 1.64 (1.23–2.18) <0.001 0.76 (0.58–0.99) 0.046 ≥50% decrease in eGFR or RRT  Model 1   2nd quintile 1.13 (0.66–1.93) 0.651 1.64 (1.16–2.34) 0.006   3rd quintile 0.92 (0.52–1.62) 0.772 2.10 (1.51–2.92) <0.001   4th quintile 1.41 (0.85–2.35) 0.189 2.71 (1.96–3.75) <0.001   5th quintile 2.47 (1.55–3.93) <0.001 3.37 (2.46–4.61) <0.001  Model 2   2nd quintile 1.15 (0.67–1.97) 0.603 0.81 (0.57–1.16) 0.254   3rd quintile 0.94 (0.53–1.69) 0.844 0.83 (0.59–1.17) 0.284   4th quintile 1.46 (0.85–2.50) 0.168 0.79 (0.56–1.11) 0.178   5th quintile 2.56 (1.53–4.27) <0.001 0.69 (0.50–0.96) 0.030  Model 3   2nd quintile 1.44 (0.84–2.48) 0.189 0.81 (0.57–1.17) 0.258   3rd quintile 1.29 (0.71–2.32) 0.404 0.81 (0.58–1.14) 0.227   4th quintile 1.48 (0.85–2.60) 0.167 0.84 (0.60–1.19) 0.328   5th quintile 2.14 (1.25–3.69) 0.006 0.73 (0.52–1.02) 0.068 Model 1: univariate. Model 2: adjusted by age, sex and eGFR. Model 3: adjusted by age, sex, eGFR, duration of diabetes, use of insulin, use of angiotensin-converting enzyme inhibitors or angiotensin receptor blockers, use of lipid-lowering drugs, use of urate-lowering drugs, smoking status, date of the baseline data, BMI, systolic blood pressure, number of creatinine measurements during the follow-up period, Hb A1c, logarithmically transformed triglycerides, LDL cholesterol and logarithmically transformed urinary albumin:creatinine ratio. Range of UA levels (mg/dL): non-CKD: 1st quintile (n = 990) 0.6–4.0, 2nd quintile (n = 956) 4.1–4.7, 3rd quintile (n = 1002) 4.8–5.3, 4th quintile (n = 1045) 5.4–6.1, 5th quintile (n = 1001) 6.2–11.6; CKD: 1st quintile (n = 415) 1.2–5.1, 2nd quintile (n = 393) 5.2–5.8, 3rd quintile (n = 433) 5.9–6.5, 4th quintile (n = 379) 6.6–7.2, 5th quintile (n = 419) 7.3–12.1. FIGURE 2 View largeDownload slide Multivariable-adjusted restricted cubic spline curves (95% CI) of the association between serum UA levels and the endpoint, which was adjusted by variables used for Model 3, and the histogram of UA levels. Reference: median of UA in each group (non-CKD, 5.1 mg/dL; CKD, 6.2 mg/dL). (A) and (C) ≥30% decrease in eGFR or RRT; (B) and (D) ≥50% decrease in eGFR or RRT. FIGURE 2 View largeDownload slide Multivariable-adjusted restricted cubic spline curves (95% CI) of the association between serum UA levels and the endpoint, which was adjusted by variables used for Model 3, and the histogram of UA levels. Reference: median of UA in each group (non-CKD, 5.1 mg/dL; CKD, 6.2 mg/dL). (A) and (C) ≥30% decrease in eGFR or RRT; (B) and (D) ≥50% decrease in eGFR or RRT. Sensitivity analyses We evaluated the robustness of the present results using sensitivity analyses. First, we used the Fine and Gray subdistribution hazards model, considering the presence of competing risks. There was a significant interaction between UA levels and baseline eGFR with respect to the primary and secondary composite endpoints in the analysis adjusted by variables used for Model 3 (P-interaction = 0.004 and 0.031, respectively). In the non-CKD group, similar findings to the above were obtained (Supplementary data, Tables S1 and S2, and Supplementary data, Figure S2). In the CKD group, the associations of low levels of UA with the endpoints were not significant (Supplementary data, Tables S1 and S2). The multivariable-adjusted restricted cubic spline model showed similar results to the above (Supplementary data, Figure S2). Second, we analyzed 6270 patients without a prescription for urate-lowering drugs at baseline. There was a significant interaction between UA levels and baseline eGFR with respect to the primary and secondary composite endpoints (P-interaction < 0.001 and 0.041, respectively). Similar results to the above were detected in the non-CKD group (Supplementary data, Tables S3 and S4). In the CKD group there was a significant association only in the analyses treating UA as a continuous variable (Supplementary data, Tables S3 and S4). Third, patients in the CKD group were further classified into two groups as follows: eGFR ≥30 but <60 mL/min/1.73 m2 (n = 1638) and eGFR <30 mL/min/1.73 m2 (n = 401). There were no associations in any multivariate model of UA levels with the incidence of either endpoint in patients with an eGFR <60 mL/min/1.73 m2 (Supplementary data, Tables S5 and S6 and Figure S3). Fourth, we used models adjusted by the same variables as in Model 3 plus the use of loop diuretics or thiazides because these drugs have possible effects on serum UA levels. A total of 694 patients were prescribed these drugs at baseline. Similar results were obtained in the models (Supplementary data, Tables S7 and S8). Fifth, some subjects might have had a transient decrease in eGFR consistent with acute kidney injury at baseline because baseline eGFR in the present study was defined by a single serum creatinine measurement. Therefore we excluded 56 patients who had a decrease in serum creatinine of ≥0.3 mg/dL from the baseline or a decrease to less than two-thirds of the baseline value in the second creatinine measurement during the follow-up period [20]. The 56 excluded patients belonged to the CKD group. Similar results were obtained in the models (Supplementary data, Tables S9 and S10). Finally, we added available data of blood glucose levels (n = 6722) or urinary sugar none, ± or 1+, and 2+ or over; n = 6842) to Model 3 as a covariate, considering the lowering effects of glycosuria on serum UA levels [21]. In the non-CKD group, high levels of UA were significantly associated with both endpoints in all analyses (Supplementary data, Tables S11–S14). The results of the analyses in the CKD group partly showed an association between low levels of UA and the endpoints (Supplementary data, Tables S11–S14). DISCUSSION This single-center, large, historical cohort study of Japanese patients with type 2 diabetes was the first to identify differences in the effects of serum UA levels on the decline in kidney function, assessed based on two composite kidney endpoints, depending on baseline kidney function. The robustness of this finding was strengthened by conducting various sensitivity analyses. In all analyses, high levels of serum UA were associated with the kidney endpoints in patients with an eGFR ≥60 mL/min/1.73 m2, and the association was likely to be linear. In patients with an eGFR <60 mL/min/1.73 m2, high levels of UA had no harmful effects on the kidney endpoints. This study confirmed the harmful effects of high levels of serum UA on kidneys in patients with type 2 diabetes with preserved kidney function, consistent with the findings of previous studies [5–7]. The positive and linear association in the restricted cubic spline analysis also suggested the following: (i) effects of UA on diabetic kidneys are independent of urate crystal deposition because its saturation point in blood is approximately 7.0 mg/dL and (ii) the lower the UA value, the better the kidney outcomes. Therefore it might be necessary to set the predictive cutoff values at fairly low levels. An important question that inevitably arises is whether UA has any causal effects, although previous basic studies have indicated that UA has a pathogenic role in kidney damage [22, 23]. An investigation using the Mendelian randomization approach from the Finnish Diabetic Nephropathy Study suggested that UA is not causally related to the development and progression of kidney damage in patients with type 1 diabetes [24]. However, several randomized controlled trials reported beneficial effects of urate-lowering therapy on human kidneys [25–27]; thus, further investigation is required to clarify this fundamental issue. It is important to understand why high levels of UA did not have harmful effects on kidney outcomes in patients with an eGFR <60 mL/min/1.73 m2 in the present study. UA-induced systemic and glomerular hypertension is considered to be one of the possible mechanisms by which UA causes kidney damage [21]; however, such hypertension is already apparent due to sodium and/or water retention in patients with CKD [28, 29]. Consistent with the present findings, several previous studies of nondiabetic patients with kidney insufficiency reported no association of high levels of UA with kidney outcomes [11–13]. Interestingly, a large cohort study of patients with diabetes from Canada reported that the increased risk of end-stage kidney disease associated with poor glycemic control was attenuated at a lower baseline eGFR, and the association disappeared in patients with severe kidney insufficiency [30], despite glycemic control being established as a major prognostic and interventional factor for early stage DKD. In light of these findings, together with the results of the present study, the effects of risk factors including UA for kidney damage in diabetes might be reduced with a decline in kidney function; in other words, it is possible that there is little scope for the risk factors to cause further harm in diabetic patients with kidney insufficiency. In the present study, part of the analyses in the CKD group showed an association between low levels of UA and the endpoints. We also proposed a likely explanation for the possible paradoxical association. Some research has shown the clinical utility of serum UA as a nutritional marker in patients receiving dialysis [31, 32]. Several studies have suggested that surrogates for malnutrition were associated with kidney outcomes in non-dialysis-dependent patients with CKD [33, 34]. Therefore the increased risk observed at fairly low levels of serum UA in the present study might reflect an association between these levels as a marker of poor nutrition and kidney outcomes (Figure 2 and Supplementary data, Figure S2). Loop diuretics or thiazides improve fluid overload and hypertension with possible increased effects on serum UA levels. Therefore the beneficial effects by these drugs might partly explain the paradoxical association, although the results after adjustment by use of these drugs remained unchanged (Supplementary data, Tables S7 and S8). The present study has several limitations. First, study participants comprised an ethnically homogeneous population at a single urban university hospital in Japan, therefore the generalizability of the present findings may be limited. Second, we could not clarify a causal relationship between UA and the kidney outcomes. Third, we could not show the continuous change of effects of UA on the endpoints according to baseline eGFR treated as a continuous variable because we could not develop a spline function including interaction of baseline eGFR and UA values treated as continuous variables in models for time-to-event data. Therefore, in the present study, the effects of UA were estimated with respect to each group categorized by baseline eGFR. Fourth, we did not evaluate information about alcohol intake, which has possible effects on serum UA levels. Finally, the present study did not evaluate time-dependent changes in serum UA, HbA1c, lipid profiles, blood pressure, BMI or medications, including urate-lowering drugs, during the follow-up period. In conclusion, the present study might provide evidence that the effects of UA on the decline in kidney function differ depending on baseline kidney function in patients with type 2 diabetes, suggesting that we need to consider the management of UA according to kidney function as therapeutic strategies for DKD. High levels of serum UA are the prognostic factor only in patients with preserved kidney function. SUPPLEMENTARY DATA Supplementary data are available at ndt online. ACKNOWLEDGEMENTS We would like to thank all members of the Diabetes Center, Tokyo Women’s Medical University School of Medicine for their helpful advice and discussions. Parts of this work were presented at the 77th Scientific Session of the American Diabetes Association, San Diego, CA, USA, 9–13 June 2017. AUTHORS’ CONTRIBUTIONS K.H. conceived the study, designed the protocol, contributed to data collection and preparation, analyzed all data, wrote the manuscript and contributed to the interpretation of the results. E.T., Y.N., T.M. and Y.Y. contributed to data collection and preparation and contributed to the interpretation of the results. Y.U. and T.B. designed the protocol and contributed to the interpretation of the results. T.B. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. All the authors have read the manuscript and have approved this submission. CONFLICT OF INTEREST STATEMENT None declared. REFERENCES 1 Alicic RZ , Rooney MT , Tuttle KR. Diabetic kidney disease: challenges, progress, and possibilities . Clin J Am Soc Nephrol 2017 ; 12 : 2032 – 2045 Google Scholar CrossRef Search ADS PubMed 2 Kanbay M , Segal M , Afsar B et al. The role of uric acid in the pathogenesis of human cardiovascular disease . Heart 2013 ; 99 : 759 – 766 Google Scholar CrossRef Search ADS PubMed 3 Volterrani M , Iellamo F , Sposato B et al. Uric acid lowering therapy in cardiovascular diseases . Int J Cardiol 2016 ; 213 : 20 – 22 Google Scholar CrossRef Search ADS PubMed 4 Zhu P , Liu Y , Han L et al. Serum uric acid is associated with incident chronic kidney disease in middle-aged populations: a meta-analysis of 15 cohort studies . PLoS One 2014 ; 9 : e100801 Google Scholar CrossRef Search ADS PubMed 5 Zoppini G , Targher G , Chonchol M et al. Serum uric acid levels and incident chronic kidney disease in patients with type 2 diabetes and preserved kidney function . Diabetes Care 2012 ; 35 : 99 – 104 Google Scholar CrossRef Search ADS PubMed 6 Takagi M , Babazono T , Uchigata Y. Differences in risk factors for the onset of albuminuria and decrease in glomerular filtration rate in people with type 2 diabetes mellitus: implications for the pathogenesis of diabetic kidney disease . Diabet Med 2015 ; 32 : 1354 – 1360 Google Scholar CrossRef Search ADS PubMed 7 Viazzi F , Piscitelli P , Giorda C et al. Metabolic syndrome, serum uric acid and renal risk in patients with T2D . PLoS One 2017 ; 12 : e0176058 Google Scholar CrossRef Search ADS PubMed 8 Edwards NL. The role of hyperuricemia and gout in kidney and cardiovascular disease . Cleve Clin J Med 2008 ; 75 : S13 – S16 Google Scholar CrossRef Search ADS PubMed 9 Hatch M , Vaziri ND. Enhanced enteric excretion of urate in rats with chronic renal failure . Clin Sci 1994 ; 86 : 511 – 516 Google Scholar CrossRef Search ADS PubMed 10 Vaziri ND , Freel RW , Hatch M. Effect of chronic experimental renal insufficiency on urate metabolism . J Am Soc Nephrol 1995 ; 6 : 1313 – 1317 Google Scholar PubMed 11 Madero M , Sarnak MJ , Wang X et al. Uric acid and long-term outcomes in CKD . Am J Kidney Dis 2009 ; 53 : 796 – 803 Google Scholar CrossRef Search ADS PubMed 12 Testa A , Mallamaci F , Spoto B et al. Association of a polymorphism in a gene encoding a urate transporter with CKD progression . Clin J Am Soc Nephrol 2014 ; 9 : 1059 – 1065 Google Scholar CrossRef Search ADS PubMed 13 Nacak H , van Diepen M , Qureshi AR et al. Uric acid is not associated with decline in renal function or time to renal replacement therapy initiation in a referred cohort of patients with stage III, IV and V chronic kidney disease . Nephrol Dial Transplant 2015 ; 30 : 2039 – 2045 Google Scholar CrossRef Search ADS PubMed 14 Kashiwagi A , Kasuga M , Araki E et al. International clinical harmonization of glycated hemoglobin in Japan: from Japan Diabetes Society to National Glycohemoglobin Standardization Program values . J Diabetes Invest 2012 ; 3 : 39 – 40 Google Scholar CrossRef Search ADS 15 Matsuo S , Imai E , Horio M , Collaborators developing the Japanese equation for estimated GFR et al. Revised equations for estimated GFR from serum creatinine in Japan . Am J Kidney Dis 2009 ; 53 : 982 – 992 Google Scholar CrossRef Search ADS PubMed 16 Coresh J , Turin TC , Matsushita K et al. Decline in estimated glomerular filtration rate and subsequent risk of end-stage renal disease and mortality . JAMA 2014 ; 311 : 2518 – 2531 Google Scholar CrossRef Search ADS PubMed 17 Durrleman S , Simon R. Flexible regression models with cubic splines . Stat Med 1989 ; 8 : 551 – 561 Google Scholar CrossRef Search ADS PubMed 18 Fine JP , Gray RJ. A proportional hazards model for the sub-distribution of a competing risk . J Am Stat Assoc 1999 ; 94 : 496 – 509 Google Scholar CrossRef Search ADS 19 Austin PC , Lee DS , Fine JP. Introduction to the analysis of survival data in the presence of competing risks . Circulation 2016 ; 133 : 601 – 609 Google Scholar CrossRef Search ADS PubMed 20 KDIGO clinical practice guideline for acute kidney injury . Kidney Int Suppl 2012 ; 2 : 1 – 138 CrossRef Search ADS 21 Johnson RJ , Nakagawa T , Jalal D et al. Uric acid and chronic kidney disease: which is causing which? Nephrol Dial Transplant 2013 ; 28 : 2221 – 2228 Google Scholar CrossRef Search ADS PubMed 22 Mazzali M , Hughes J , Kim YG et al. Elevated uric acid increases blood pressure in the rat by a novel crystal-independent mechanism . Hypertension 2001 ; 38 : 1101 – 1106 Google Scholar CrossRef Search ADS PubMed 23 Mazzali M , Kanellis J , Han L et al. Hyperuricemia induces a primary renal arteriolopathy in rats by a blood pressure-independent mechanism . Am J Physiol Renal Physiol 2002 ; 282 : F991 – F997 Google Scholar CrossRef Search ADS PubMed 24 Ahola AJ , Sandholm N , Forsblom C et al. The serum uric acid concentration is not causally linked to diabetic nephropathy in type 1 diabetes . Kidney Int 2017 ; 91 : 1178 – 1185 Google Scholar CrossRef Search ADS PubMed 25 Siu YP , Leung KT , Tong MK , Kwan TH. Use of allopurinol in slowing the progression of renal disease through its ability to lower serum uric acid level . Am J Kidney Dis 2006 ; 47 : 51 – 59 Google Scholar CrossRef Search ADS PubMed 26 Goicoechea M , de Vinuesa SG , Verdalles U et al. Effect of allopurinol in chronic kidney disease progression and cardiovascular risk . Clin J Am Soc Nephrol 2010 ; 5 : 1388 – 1393 Google Scholar CrossRef Search ADS PubMed 27 Liu P , Chen Y , Wang B et al. Allopurinol treatment improves renal function in patients with type 2 diabetes and asymptomatic hyperuricemia: 3-year randomized parallel-controlled study . Clin Endocrinol (Oxf) 2015 ; 83 : 475 – 482 Google Scholar CrossRef Search ADS PubMed 28 Rahman M , Yang W , Akkina S et al. Relation of serum lipids and lipoproteins with progression of CKD: the CRIC Study . Clin J Am Soc Nephrol 2014 ; 9 : 1190 – 1198 Google Scholar CrossRef Search ADS PubMed 29 Inaguma D , Imai E , Takeuchi A et al. Risk factors for CKD progression in Japanese patients: findings from the Chronic Kidney Disease Japan Cohort (CKD-JAC) study . Clin Exp Nephrol 2017 ; 21 : 446 – 456 Google Scholar CrossRef Search ADS PubMed 30 Shurraw S , Hemmelgarn B , Lin M et al. Association between glycemic control and adverse outcomes in people with diabetes mellitus and chronic kidney disease . Arch Intern Med 2011 ; 171 : 1920 – 1927 Google Scholar CrossRef Search ADS PubMed 31 Beberashvili I , Sinuani I , Azar A et al. Serum uric acid as a clinically useful nutritional marker and predictor of outcome in maintenance hemodialysis patients . Nutrition 2015 ; 31 : 138 – 147 Google Scholar CrossRef Search ADS PubMed 32 Park C , Obi Y , Streja E et al. Serum uric acid, protein intake and mortality in hemodialysis patients . Nephrol Dial Transplant 2017 ; 32 : 1750 – 1757 Google Scholar PubMed 33 Lu JL , Kalantar-Zadeh K , Ma JZ et al. Association of body mass index with outcomes in patients with CKD . J Am Soc Nephrol 2014 ; 25 : 2088 – 2096 Google Scholar CrossRef Search ADS PubMed 34 Kikuchi H , Kanda E , Mandai S et al. Combination of low body mass idex and serum albumin level is associated with chronic kidney disease progression: the chronic kidney disease-research of outcomes in treatment and epidemiology (CKD-ROUTE) study . Clin Exp Nephrol 2017 ; 21 : 55 – 62 Google Scholar CrossRef Search ADS PubMed © The Author(s) 2018. Published by Oxford University Press on behalf of ERA-EDTA. All rights reserved. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Nephrology Dialysis Transplantation Oxford University Press

Effects of uric acid on kidney function decline differ depending on baseline kidney function in type 2 diabetic patients

Loading next page...
 
/lp/ou_press/effects-of-uric-acid-on-kidney-function-decline-differ-depending-on-8wVHDOUlsp
Publisher
Oxford University Press
Copyright
© The Author(s) 2018. Published by Oxford University Press on behalf of ERA-EDTA. All rights reserved.
ISSN
0931-0509
eISSN
1460-2385
D.O.I.
10.1093/ndt/gfy138
Publisher site
See Article on Publisher Site

Abstract

Abstract Background Most existing data regarding effects of uric acid (UA) on diabetic kidney disease have considered patients with preserved kidney function. We examined a hypothesis that there are differences in the effects of serum UA levels on the decline in kidney function depending on baseline kidney function in diabetic patients. Methods In this historical cohort study, 7033 type 2 diabetic patients were analyzed and classified into two groups as follows: nonchronic kidney disease (non-CKD), with an estimated glomerular filtration rate (eGFR) ≥60 mL/min/1.73 m2 (n = 4994), and CKD, with an eGFR <60 mL/min/1.73 m2 (n = 2039). The composite endpoint was a ≥30% decrease in eGFR from baseline or the initiation of renal replacement therapy. The hazard ratio (HR) of serum UA levels at baseline was estimated using multivariate Cox proportional hazards models. Results There was a significant interaction between UA levels and baseline eGFR with respect to the endpoint (P < 0.001). The HRs of 1 mg/dL increase in UA levels were 1.13 [95% confidence interval (CI) 1.05–1.22, P = 0.002] and 0.93 (95% CI 0.88–0.99, P = 0.02) in the non-CKD and CKD groups, respectively. When patients were classified by quintile of UA levels, the HRs of those in the 5th quintile (versus 1st quintile) were 1.64 (95% CI 1.23–2.18, P < 0.001) and 0.76 (95% CI 0.58–0.99, P = 0.05) in the non-CKD and CKD groups, respectively. Conclusions The effects of UA on kidney function decline might differ depending on baseline kidney function in type 2 diabetic patients. High UA levels are the prognostic factor only in patients with preserved kidney function. diabetic kidney disease, diabetic nephropathy, hyperuricemia, type 2 diabetes, uric acid INTRODUCTION Diabetic kidney disease (DKD) remains a major global health concern despite remarkable advances in the management of hyperglycemia and hypertension, which are two major established risk factors [1]. Uric acid (UA) has been demonstrated in basic research to stimulate inflammation, oxidative stress, endothelial damage and vasoconstriction [2]. The harmful effects of UA on human kidneys as well as on atherosclerotic cardiovascular disease and mortality have received a great deal of attention [2–4]. A number of studies, the majority of which considered patients with preserved kidney function, have shown an association between elevated serum UA levels and the development and progression of DKD [5–7]; however, the cutoff and target values remain unknown. Serum UA levels rise in patients with kidney insufficiency because excretion from the kidneys is responsible for much of the clearance of UA [8]. However, such increases might be mitigated by reduction of the biosynthesis of UA and upregulation of its secretion by the colon [9, 10]. The majority of studies in patients with kidney insufficiency have not reported an association between baseline serum UA levels and subsequent decline in kidney function [11–13], unlike studies in patients with preserved kidney function, suggesting that the effects of UA on the kidneys differ depending on kidney function. To our knowledge, information regarding the association of UA with the progression of DKD in patients with kidney insufficiency is limited. In the present study we aimed to examine a hypothesis that there are differences in the effects of serum UA levels on the decline in kidney function depending on baseline kidney function in patients with diabetes and, if any differences exist, to clarify the distinctions between groups classified based on kidney function. MATERIALS AND METHODS Study design This single-center, historical cohort study was designed in adherence with the Declaration of Helsinki and was performed as part of the Cohort Study Elucidating Factors Associated with the Pathogenesis and Prognosis of Diabetic Kidney Disease conducted in the Diabetes Center, Tokyo Women’s Medical University School of Medicine. The Ethics Committee of Tokyo Women’s Medical University School of Medicine approved the protocol. As this was a historical cohort study, not a prospective interventional study, the Ethics Committee approved a waiver of informed consent. Data used in the present study were obtained from the clinical information system (electronic medical records). Participants We initially identified 7430 Japanese ambulatory patients with type 2 diabetes who were ≥18 years of age; did not receive renal replacement therapy (RRT), which meant chronic dialysis or kidney transplant; and had available data on body weight, blood pressure, serum UA, hemoglobin A1c (HbA1c), lipid parameters, creatinine and urinary albumin simultaneously measured at a regular ambulatory visit in the Diabetes Center of Tokyo Women’s Medical University School of Medicine from 1 January 2004 to 30 April 2016. The earliest available data were set as baseline if there were multiple measurements available for subjects during the above-mentioned period. A flow diagram of the study population is presented in Figure 1. At baseline, pregnant patients (n = 52) or patients with malignant disease (n = 91) were excluded. Patients who had undergone removal of one kidney owing to cancer or transplantation before baseline (n = 8) or those with glomerular nephritis proven via biopsy before baseline (n = 1), acute kidney injury or postrenal failure at baseline (n = 2) or acute severe infection at baseline (n = 2) were excluded. Patients with no follow-up serum creatinine data (n = 234) or missing baseline profile values (n = 7) were also excluded. Overall, the data obtained from 7033 patients were analyzed and these patients were classified into two groups based on kidney function: nonchronic kidney disease (non-CKD), with an estimated glomerular filtration rate (eGFR) ≥60 mL/min/1.73 m2 (n = 4994), and CKD, with an eGFR <60 mL/min/1.73 m2 (n = 2039). FIGURE 1 View largeDownload slide A flow diagram of the study population. FIGURE 1 View largeDownload slide A flow diagram of the study population. Measurements Laboratory data were obtained from random spot blood samples and first morning urine samples. Serum UA, triglycerides and creatinine levels were determined using enzymatic methods. Levels of serum high-density lipoprotein (HDL) cholesterol were measured using polyethylene glycol–pretreated enzymes and levels of serum low-density lipoprotein (LDL) cholesterol were determined using enzymatic methods or the Friedewald formula (if triglycerides <400 mg/dL). HbA1c values obtained as Japan Diabetes Society (JDS) values were converted to National Glycohemoglobin Standardization Program (NGSP) values as follows: HbA1c (%) = 1.02 × HbA1c (JDS) (%) + 0.25% [14]. Levels of urinary creatinine were determined using enzymatic methods. Levels of urinary albumin were determined using the latex agglutination method and normalized by urinary creatinine concentrations. The GFR was estimated using the following formula, proposed by the Japanese Society for Nephrology: eGFR (mL/min/1.73 m2) = 194 × age (years)−0.287 × serum creatinine level (in mg/dL)−1.094 × (0.739, for women) [15]. Endpoints and follow-up The primary composite endpoint was a decrease in eGFR of ≥30% from baseline, established as a surrogate for the development of kidney failure [16], or the initiation of RRT, whichever came first. We also verified a ≥50% decrease in eGFR instead of a ≥30% decrease in eGFR as the secondary composite endpoint. We determined that the endpoint was reached when the above-mentioned criteria were met for at least 3 months to avoid capturing episodes of reversible acute kidney injury. Patients were censored when they initiated RRT, were deceased before the initiation of RRT, were lost to follow-up or reached the administrative censoring date (31 August 2016). Statistical analysis Categorical data were expressed as number (%). Continuous variables were expressed as the arithmetic mean [standard deviation (SD)] or median [interquartile range (IQR)], as appropriate. For statistical analyses, Fisher’s exact probability test, Student’s t-test, Mann–Whitney U test and Cox proportional hazards models were used as appropriate. The Kaplan–Meier method was used for an estimation of survival probability for endpoints and the statistical differences between the groups were compared using the log-rank test. Possible nonlinear associations between UA levels and the endpoint were examined using the multivariable-adjusted restricted cubic spline model [17], with the knots placed at the 5th, 35th, 65th and 95th percentile levels. We also used the Fine and Gray subdistribution hazards model instead of the traditional Cox proportional hazards model in a time-to-event analysis, as latter methods may overestimate the risk of events by undermining the competing risks [18, 19]. In the present study, death before reaching the endpoint was treated as a competing risk. Three sets of models were used for the calculation of the hazard ratio (HR) [95% confidence interval (CI)] of UA for the endpoint: univariate models (Model 1); models adjusted by age, sex and eGFR (Model 2) and models adjusted by the same variables as in Model 2 plus duration of diabetes, use of insulin, use of angiotensin-converting enzyme inhibitors or angiotensin receptor blockers, use of lipid-lowering drugs, use of urate-lowering drugs, smoking status (current or former), date of the baseline data (2004–2009 or 2010–2016), body mass index (BMI), systolic blood pressure, number of creatinine measurements during the follow-up period, HbA1c, triglycerides, LDL cholesterol and urinary albumin:creatinine ratio (Model 3). Triglycerides and urinary albumin:creatinine ratio levels transformed into common logarithm values were used in the analyses. P-values < 0.05 were considered significant. All analyses were completed using SAS version 9.4 (SAS Institute, Cary, NC, USA). RESULTS We analyzed 7033 patients, comprising 2493 women and 4540 men with a mean age of 61 (SD 12) years. Baseline characteristics for the overall and categorized cohort are presented in Table 1. During the median follow-up period of 5.5 (IQR 2.5–9.4) years, 1189 patients experienced a ≥30% decrease in eGFR from baseline, 529 experienced a  ≥50% decrease in eGFR from baseline and RRT was initiated in 440 patients. Prior to the initiation of RRT, 284 patients died. Consequently, 1250 patients, consisting of 1189 with a ≥30% decrease in eGFR and 61 who initiated RRT before experiencing a ≥30% decrease in eGFR, reached the primary composite endpoint; 561 and 689 were in the non-CKD and CKD groups, respectively. A total of 663 patients, consisting of 529 with a  ≥50% decrease in eGFR and 134 who initiated RRT before experiencing a  ≥50% decrease in eGFR, reached the secondary composite endpoint; 166 and 497 were in the non-CKD and CKD groups, respectively. Patients in the CKD group had greater incidences of both composite endpoints than those in the non-CKD group (both P < 0.001; Supplementary data, Figure S1). The median number of creatinine measurements used for the analyses was 18 (IQR 9–33) and 21 (IQR 10–37) for the primary and secondary composite endpoints, respectively. Table 1 Baseline demographic and laboratory data Characteristic Overall (n = 7033) Non-CKD (n = 4994) CKD (n = 2039) P-value (Non-CKD versus CKD) Age (years) 61 (12) 59 (12) 66 (11) <0.001 Gender (men), n (%) 4540 (64.6) 3144 (63.0) 1396 (68.5) <0.001 Duration of diabetes (years) 13 (9) 12 (9) 16 (10) <0.001 Diabetes therapy, n (%)  Insulin 2295 (32.6) 1393 (27.9) 902 (44.2) <0.001  Other medication 4220 (60.0) 3096 (62.0) 1124 (55.1) <0.001 ACE inhibitors or ARBs, n (%) 3135 (44.6) 1740 (34.8) 1395 (68.4) <0.001 Other antihypertensive drugs, n (%) 2415 (34.3) 1283 (25.7) 1132 (55.5) <0.001 Lipid-lowering drugs, n (%) 2607 (37.1) 1604 (32.1) 1003 (49.2) <0.001 Urate-lowering drugs, n (%) 763 (10.9) 228 (4.6) 535 (26.2) <0.001 History of ASCVD, n (%) 1443 (20.5) 734 (14.7) 709 (34.8) <0.001 Former or current smoker, n (%) 3774 (53.7) 2650 (53.1) 1124 (55.1) 0.120 Date of baseline data, n (%) <0.001  2004–09 4380 (62.3) 3215 (64.4) 1165 (57.1)  2010–16 2653 (37.7) 1779 (35.6) 874 (42.9) BMI (kg/m2) 25.0 (4.2) 25.0 (4.3) 25.2 (3.9) 0.036 SBP (mmHg) 136 (20) 135 (19) 140 (22) <0.001 DBP (mmHg) 76 (12) 77 (11) 74 (13) <0.001 Laboratory data  UA (mg/dL) 5.4 (1.4) 5.1 (1.3) 6.2 (1.4) <0.001  HbA1c (%) 7.6 (1.3) 7.7 (1.3) 7.4 (1.2) <0.001  Triglycerides (mg/dL) median (IQR) 125 (85–185) 120 (81–179) 136 (97–198) <0.001  HDL cholesterol (mg/dL) 55 (15) 56 (15) 52 (15) <0.001  LDL cholesterol (mg/dL) 115 (31) 116 (30) 111 (33) <0.001  Creatinine (mg/dL) 0.96 (0.73) 0.72 (0.14) 1.55 (1.13) <0.001  eGFR (mL/min/1.73 m2) 69.9 (23.0) 80.8 (15.6) 43.0 (14.8) <0.001  ACR (mg/g) median (IQR) 14.8 (7.5–68.2) 11.6 (6.9–29.8) 72.2 (13.1–931.8) <0.001 Characteristic Overall (n = 7033) Non-CKD (n = 4994) CKD (n = 2039) P-value (Non-CKD versus CKD) Age (years) 61 (12) 59 (12) 66 (11) <0.001 Gender (men), n (%) 4540 (64.6) 3144 (63.0) 1396 (68.5) <0.001 Duration of diabetes (years) 13 (9) 12 (9) 16 (10) <0.001 Diabetes therapy, n (%)  Insulin 2295 (32.6) 1393 (27.9) 902 (44.2) <0.001  Other medication 4220 (60.0) 3096 (62.0) 1124 (55.1) <0.001 ACE inhibitors or ARBs, n (%) 3135 (44.6) 1740 (34.8) 1395 (68.4) <0.001 Other antihypertensive drugs, n (%) 2415 (34.3) 1283 (25.7) 1132 (55.5) <0.001 Lipid-lowering drugs, n (%) 2607 (37.1) 1604 (32.1) 1003 (49.2) <0.001 Urate-lowering drugs, n (%) 763 (10.9) 228 (4.6) 535 (26.2) <0.001 History of ASCVD, n (%) 1443 (20.5) 734 (14.7) 709 (34.8) <0.001 Former or current smoker, n (%) 3774 (53.7) 2650 (53.1) 1124 (55.1) 0.120 Date of baseline data, n (%) <0.001  2004–09 4380 (62.3) 3215 (64.4) 1165 (57.1)  2010–16 2653 (37.7) 1779 (35.6) 874 (42.9) BMI (kg/m2) 25.0 (4.2) 25.0 (4.3) 25.2 (3.9) 0.036 SBP (mmHg) 136 (20) 135 (19) 140 (22) <0.001 DBP (mmHg) 76 (12) 77 (11) 74 (13) <0.001 Laboratory data  UA (mg/dL) 5.4 (1.4) 5.1 (1.3) 6.2 (1.4) <0.001  HbA1c (%) 7.6 (1.3) 7.7 (1.3) 7.4 (1.2) <0.001  Triglycerides (mg/dL) median (IQR) 125 (85–185) 120 (81–179) 136 (97–198) <0.001  HDL cholesterol (mg/dL) 55 (15) 56 (15) 52 (15) <0.001  LDL cholesterol (mg/dL) 115 (31) 116 (30) 111 (33) <0.001  Creatinine (mg/dL) 0.96 (0.73) 0.72 (0.14) 1.55 (1.13) <0.001  eGFR (mL/min/1.73 m2) 69.9 (23.0) 80.8 (15.6) 43.0 (14.8) <0.001  ACR (mg/g) median (IQR) 14.8 (7.5–68.2) 11.6 (6.9–29.8) 72.2 (13.1–931.8) <0.001 Data are expressed as mean (SD) unless stated otherwise. non-CKD: eGFR ≥60 mL/min/1.73 m2; CKD: eGFR <60 mL/min/1.73 m2. ACE, angiotensin-converting enzyme; ARBs, angiotensin receptor blockers; ASCVD, atherosclerotic cardiovascular disease; SBP, systolic blood pressure; DBP, diastolic blood pressure; ACR, albumin:creatinine ratio. Table 1 Baseline demographic and laboratory data Characteristic Overall (n = 7033) Non-CKD (n = 4994) CKD (n = 2039) P-value (Non-CKD versus CKD) Age (years) 61 (12) 59 (12) 66 (11) <0.001 Gender (men), n (%) 4540 (64.6) 3144 (63.0) 1396 (68.5) <0.001 Duration of diabetes (years) 13 (9) 12 (9) 16 (10) <0.001 Diabetes therapy, n (%)  Insulin 2295 (32.6) 1393 (27.9) 902 (44.2) <0.001  Other medication 4220 (60.0) 3096 (62.0) 1124 (55.1) <0.001 ACE inhibitors or ARBs, n (%) 3135 (44.6) 1740 (34.8) 1395 (68.4) <0.001 Other antihypertensive drugs, n (%) 2415 (34.3) 1283 (25.7) 1132 (55.5) <0.001 Lipid-lowering drugs, n (%) 2607 (37.1) 1604 (32.1) 1003 (49.2) <0.001 Urate-lowering drugs, n (%) 763 (10.9) 228 (4.6) 535 (26.2) <0.001 History of ASCVD, n (%) 1443 (20.5) 734 (14.7) 709 (34.8) <0.001 Former or current smoker, n (%) 3774 (53.7) 2650 (53.1) 1124 (55.1) 0.120 Date of baseline data, n (%) <0.001  2004–09 4380 (62.3) 3215 (64.4) 1165 (57.1)  2010–16 2653 (37.7) 1779 (35.6) 874 (42.9) BMI (kg/m2) 25.0 (4.2) 25.0 (4.3) 25.2 (3.9) 0.036 SBP (mmHg) 136 (20) 135 (19) 140 (22) <0.001 DBP (mmHg) 76 (12) 77 (11) 74 (13) <0.001 Laboratory data  UA (mg/dL) 5.4 (1.4) 5.1 (1.3) 6.2 (1.4) <0.001  HbA1c (%) 7.6 (1.3) 7.7 (1.3) 7.4 (1.2) <0.001  Triglycerides (mg/dL) median (IQR) 125 (85–185) 120 (81–179) 136 (97–198) <0.001  HDL cholesterol (mg/dL) 55 (15) 56 (15) 52 (15) <0.001  LDL cholesterol (mg/dL) 115 (31) 116 (30) 111 (33) <0.001  Creatinine (mg/dL) 0.96 (0.73) 0.72 (0.14) 1.55 (1.13) <0.001  eGFR (mL/min/1.73 m2) 69.9 (23.0) 80.8 (15.6) 43.0 (14.8) <0.001  ACR (mg/g) median (IQR) 14.8 (7.5–68.2) 11.6 (6.9–29.8) 72.2 (13.1–931.8) <0.001 Characteristic Overall (n = 7033) Non-CKD (n = 4994) CKD (n = 2039) P-value (Non-CKD versus CKD) Age (years) 61 (12) 59 (12) 66 (11) <0.001 Gender (men), n (%) 4540 (64.6) 3144 (63.0) 1396 (68.5) <0.001 Duration of diabetes (years) 13 (9) 12 (9) 16 (10) <0.001 Diabetes therapy, n (%)  Insulin 2295 (32.6) 1393 (27.9) 902 (44.2) <0.001  Other medication 4220 (60.0) 3096 (62.0) 1124 (55.1) <0.001 ACE inhibitors or ARBs, n (%) 3135 (44.6) 1740 (34.8) 1395 (68.4) <0.001 Other antihypertensive drugs, n (%) 2415 (34.3) 1283 (25.7) 1132 (55.5) <0.001 Lipid-lowering drugs, n (%) 2607 (37.1) 1604 (32.1) 1003 (49.2) <0.001 Urate-lowering drugs, n (%) 763 (10.9) 228 (4.6) 535 (26.2) <0.001 History of ASCVD, n (%) 1443 (20.5) 734 (14.7) 709 (34.8) <0.001 Former or current smoker, n (%) 3774 (53.7) 2650 (53.1) 1124 (55.1) 0.120 Date of baseline data, n (%) <0.001  2004–09 4380 (62.3) 3215 (64.4) 1165 (57.1)  2010–16 2653 (37.7) 1779 (35.6) 874 (42.9) BMI (kg/m2) 25.0 (4.2) 25.0 (4.3) 25.2 (3.9) 0.036 SBP (mmHg) 136 (20) 135 (19) 140 (22) <0.001 DBP (mmHg) 76 (12) 77 (11) 74 (13) <0.001 Laboratory data  UA (mg/dL) 5.4 (1.4) 5.1 (1.3) 6.2 (1.4) <0.001  HbA1c (%) 7.6 (1.3) 7.7 (1.3) 7.4 (1.2) <0.001  Triglycerides (mg/dL) median (IQR) 125 (85–185) 120 (81–179) 136 (97–198) <0.001  HDL cholesterol (mg/dL) 55 (15) 56 (15) 52 (15) <0.001  LDL cholesterol (mg/dL) 115 (31) 116 (30) 111 (33) <0.001  Creatinine (mg/dL) 0.96 (0.73) 0.72 (0.14) 1.55 (1.13) <0.001  eGFR (mL/min/1.73 m2) 69.9 (23.0) 80.8 (15.6) 43.0 (14.8) <0.001  ACR (mg/g) median (IQR) 14.8 (7.5–68.2) 11.6 (6.9–29.8) 72.2 (13.1–931.8) <0.001 Data are expressed as mean (SD) unless stated otherwise. non-CKD: eGFR ≥60 mL/min/1.73 m2; CKD: eGFR <60 mL/min/1.73 m2. ACE, angiotensin-converting enzyme; ARBs, angiotensin receptor blockers; ASCVD, atherosclerotic cardiovascular disease; SBP, systolic blood pressure; DBP, diastolic blood pressure; ACR, albumin:creatinine ratio. Association of serum UA levels with the incidence of the primary and secondary endpoint There was a significant interaction between UA levels and baseline eGFR with respect to the primary composite endpoint in the analysis adjusted by variables used for Model 3 (P-interaction < 0.001). The same result was obtained for the secondary composite endpoint (P-interaction = 0.004). In the non-CKD group, high levels of UA were a risk factor for both endpoints (Table 2). In the CKD group, the univariate analyses showed high UA levels to be a risk factor for the endpoint; however, in the multivariate models, low levels of UA were associated with the incidence of both endpoints (Table 2). When subjects were classified by quintile of UA levels within the non-CKD and CKD groups, those in the 5th quintile (versus the 1st quintile) in the non-CKD group had a significantly higher risk for the primary and secondary composite endpoint in all models (Table 3). On the contrary, although patients in the 2nd–5th quintile (versus the 1st quintile) in the CKD group had a greater risk for both endpoints in the univariate models, the HRs for those in the 5th quintile for the primary endpoint were significantly low in the multivariate models (Table 3). Finally, the restricted cubic spline curves (95% CI) of the association between UA levels and the composite endpoints are shown in Figure 2, following adjustment with variables used for Model 3. In both groups, a nonlinear association between UA levels and the endpoints was rejected. Table 2 HR of 1 mg/dL increase in serum UA levels for each composite endpoint Non-CKD (n = 4994) CKD (n = 2039) Models HR (95% CI) P-value HR (95% CI) P-value ≥30% decrease in eGFR or RRT  Model 1 1.11 (1.04–1.19) 0.003 1.30 (1.23–1.37) <0.001  Model 2 1.20 (1.12–1.30) <0.001 0.93 (0.88–0.98) 0.011  Model 3 1.13 (1.05–1.22) 0.002 0.93 (0.88–0.99) 0.022 ≥50% decrease in eGFR or RRT  Model 1 1.34 (1.18–1.51) <0.001 1.35 (1.27–1.43) <0.001  Model 2 1.36 (1.19–1.55) <0.001 0.90 (0.84–0.97) 0.003  Model 3 1.20 (1.05–1.37) 0.009 0.90 (0.84–0.97) 0.005 Non-CKD (n = 4994) CKD (n = 2039) Models HR (95% CI) P-value HR (95% CI) P-value ≥30% decrease in eGFR or RRT  Model 1 1.11 (1.04–1.19) 0.003 1.30 (1.23–1.37) <0.001  Model 2 1.20 (1.12–1.30) <0.001 0.93 (0.88–0.98) 0.011  Model 3 1.13 (1.05–1.22) 0.002 0.93 (0.88–0.99) 0.022 ≥50% decrease in eGFR or RRT  Model 1 1.34 (1.18–1.51) <0.001 1.35 (1.27–1.43) <0.001  Model 2 1.36 (1.19–1.55) <0.001 0.90 (0.84–0.97) 0.003  Model 3 1.20 (1.05–1.37) 0.009 0.90 (0.84–0.97) 0.005 Model 1: univariate. Model 2: adjusted by age, sex and eGFR. Model 3: adjusted by age, sex, eGFR, duration of diabetes, use of insulin, use of angiotensin-converting enzyme inhibitors or angiotensin receptor blockers, use of lipid-lowering drugs, use of urate-lowering drugs, smoking status, date of the baseline data, BMI, systolic blood pressure, number of creatinine measurements during the follow-up period, Hb A1c, logarithmically transformed triglycerides, LDL cholesterol and logarithmically transformed urinary albumin:creatinine ratio. Table 2 HR of 1 mg/dL increase in serum UA levels for each composite endpoint Non-CKD (n = 4994) CKD (n = 2039) Models HR (95% CI) P-value HR (95% CI) P-value ≥30% decrease in eGFR or RRT  Model 1 1.11 (1.04–1.19) 0.003 1.30 (1.23–1.37) <0.001  Model 2 1.20 (1.12–1.30) <0.001 0.93 (0.88–0.98) 0.011  Model 3 1.13 (1.05–1.22) 0.002 0.93 (0.88–0.99) 0.022 ≥50% decrease in eGFR or RRT  Model 1 1.34 (1.18–1.51) <0.001 1.35 (1.27–1.43) <0.001  Model 2 1.36 (1.19–1.55) <0.001 0.90 (0.84–0.97) 0.003  Model 3 1.20 (1.05–1.37) 0.009 0.90 (0.84–0.97) 0.005 Non-CKD (n = 4994) CKD (n = 2039) Models HR (95% CI) P-value HR (95% CI) P-value ≥30% decrease in eGFR or RRT  Model 1 1.11 (1.04–1.19) 0.003 1.30 (1.23–1.37) <0.001  Model 2 1.20 (1.12–1.30) <0.001 0.93 (0.88–0.98) 0.011  Model 3 1.13 (1.05–1.22) 0.002 0.93 (0.88–0.99) 0.022 ≥50% decrease in eGFR or RRT  Model 1 1.34 (1.18–1.51) <0.001 1.35 (1.27–1.43) <0.001  Model 2 1.36 (1.19–1.55) <0.001 0.90 (0.84–0.97) 0.003  Model 3 1.20 (1.05–1.37) 0.009 0.90 (0.84–0.97) 0.005 Model 1: univariate. Model 2: adjusted by age, sex and eGFR. Model 3: adjusted by age, sex, eGFR, duration of diabetes, use of insulin, use of angiotensin-converting enzyme inhibitors or angiotensin receptor blockers, use of lipid-lowering drugs, use of urate-lowering drugs, smoking status, date of the baseline data, BMI, systolic blood pressure, number of creatinine measurements during the follow-up period, Hb A1c, logarithmically transformed triglycerides, LDL cholesterol and logarithmically transformed urinary albumin:creatinine ratio. Table 3 HR (versus the 1st quintile) of other groups classified by quintile of serum UA levels for each composite endpoint Non-CKD (n = 4994) CKD (n = 2039) Models HR (95% CI) P-value HR (95% CI) P-value ≥30% decrease in eGFR or RRT  Model 1   2nd quintile 0.92 (0.71–1.20) 0.541 1.40 (1.06–1.86) 0.018   3rd quintile 0.84 (0.64–1.10) 0.208 1.75 (1.35–2.28) <0.001   4th quintile 0.93 (0.71–1.21) 0.586 2.29 (1.77–2.97) <0.001   5th quintile 1.39 (1.09–1.77) 0.009 2.67 (2.08–3.43) <0.001  Model 2   2nd quintile 1.00 (0.76–1.30) 0.973 0.82 (0.62–1.09) 0.175   3rd quintile 0.96 (0.73–1.27) 0.795 0.91 (0.69–1.19) 0.476   4th quintile 1.15 (0.87–1.52) 0.324 0.88 (0.67–1.16) 0.358   5th quintile 1.80 (1.37–2.36) <0.001 0.75 (0.57–0.98) 0.034  Model 3   2nd quintile 1.16 (0.89–1.52) 0.266 0.81 (0.61–1.09) 0.162   3rd quintile 1.09 (0.82–1.45) 0.544 0.84 (0.64–1.11) 0.219   4th quintile 1.22 (0.91–1.62) 0.184 0.90 (0.68–1.18) 0.442   5th quintile 1.64 (1.23–2.18) <0.001 0.76 (0.58–0.99) 0.046 ≥50% decrease in eGFR or RRT  Model 1   2nd quintile 1.13 (0.66–1.93) 0.651 1.64 (1.16–2.34) 0.006   3rd quintile 0.92 (0.52–1.62) 0.772 2.10 (1.51–2.92) <0.001   4th quintile 1.41 (0.85–2.35) 0.189 2.71 (1.96–3.75) <0.001   5th quintile 2.47 (1.55–3.93) <0.001 3.37 (2.46–4.61) <0.001  Model 2   2nd quintile 1.15 (0.67–1.97) 0.603 0.81 (0.57–1.16) 0.254   3rd quintile 0.94 (0.53–1.69) 0.844 0.83 (0.59–1.17) 0.284   4th quintile 1.46 (0.85–2.50) 0.168 0.79 (0.56–1.11) 0.178   5th quintile 2.56 (1.53–4.27) <0.001 0.69 (0.50–0.96) 0.030  Model 3   2nd quintile 1.44 (0.84–2.48) 0.189 0.81 (0.57–1.17) 0.258   3rd quintile 1.29 (0.71–2.32) 0.404 0.81 (0.58–1.14) 0.227   4th quintile 1.48 (0.85–2.60) 0.167 0.84 (0.60–1.19) 0.328   5th quintile 2.14 (1.25–3.69) 0.006 0.73 (0.52–1.02) 0.068 Non-CKD (n = 4994) CKD (n = 2039) Models HR (95% CI) P-value HR (95% CI) P-value ≥30% decrease in eGFR or RRT  Model 1   2nd quintile 0.92 (0.71–1.20) 0.541 1.40 (1.06–1.86) 0.018   3rd quintile 0.84 (0.64–1.10) 0.208 1.75 (1.35–2.28) <0.001   4th quintile 0.93 (0.71–1.21) 0.586 2.29 (1.77–2.97) <0.001   5th quintile 1.39 (1.09–1.77) 0.009 2.67 (2.08–3.43) <0.001  Model 2   2nd quintile 1.00 (0.76–1.30) 0.973 0.82 (0.62–1.09) 0.175   3rd quintile 0.96 (0.73–1.27) 0.795 0.91 (0.69–1.19) 0.476   4th quintile 1.15 (0.87–1.52) 0.324 0.88 (0.67–1.16) 0.358   5th quintile 1.80 (1.37–2.36) <0.001 0.75 (0.57–0.98) 0.034  Model 3   2nd quintile 1.16 (0.89–1.52) 0.266 0.81 (0.61–1.09) 0.162   3rd quintile 1.09 (0.82–1.45) 0.544 0.84 (0.64–1.11) 0.219   4th quintile 1.22 (0.91–1.62) 0.184 0.90 (0.68–1.18) 0.442   5th quintile 1.64 (1.23–2.18) <0.001 0.76 (0.58–0.99) 0.046 ≥50% decrease in eGFR or RRT  Model 1   2nd quintile 1.13 (0.66–1.93) 0.651 1.64 (1.16–2.34) 0.006   3rd quintile 0.92 (0.52–1.62) 0.772 2.10 (1.51–2.92) <0.001   4th quintile 1.41 (0.85–2.35) 0.189 2.71 (1.96–3.75) <0.001   5th quintile 2.47 (1.55–3.93) <0.001 3.37 (2.46–4.61) <0.001  Model 2   2nd quintile 1.15 (0.67–1.97) 0.603 0.81 (0.57–1.16) 0.254   3rd quintile 0.94 (0.53–1.69) 0.844 0.83 (0.59–1.17) 0.284   4th quintile 1.46 (0.85–2.50) 0.168 0.79 (0.56–1.11) 0.178   5th quintile 2.56 (1.53–4.27) <0.001 0.69 (0.50–0.96) 0.030  Model 3   2nd quintile 1.44 (0.84–2.48) 0.189 0.81 (0.57–1.17) 0.258   3rd quintile 1.29 (0.71–2.32) 0.404 0.81 (0.58–1.14) 0.227   4th quintile 1.48 (0.85–2.60) 0.167 0.84 (0.60–1.19) 0.328   5th quintile 2.14 (1.25–3.69) 0.006 0.73 (0.52–1.02) 0.068 Model 1: univariate. Model 2: adjusted by age, sex and eGFR. Model 3: adjusted by age, sex, eGFR, duration of diabetes, use of insulin, use of angiotensin-converting enzyme inhibitors or angiotensin receptor blockers, use of lipid-lowering drugs, use of urate-lowering drugs, smoking status, date of the baseline data, BMI, systolic blood pressure, number of creatinine measurements during the follow-up period, Hb A1c, logarithmically transformed triglycerides, LDL cholesterol and logarithmically transformed urinary albumin:creatinine ratio. Range of UA levels (mg/dL): non-CKD: 1st quintile (n = 990) 0.6–4.0, 2nd quintile (n = 956) 4.1–4.7, 3rd quintile (n = 1002) 4.8–5.3, 4th quintile (n = 1045) 5.4–6.1, 5th quintile (n = 1001) 6.2–11.6; CKD: 1st quintile (n = 415) 1.2–5.1, 2nd quintile (n = 393) 5.2–5.8, 3rd quintile (n = 433) 5.9–6.5, 4th quintile (n = 379) 6.6–7.2, 5th quintile (n = 419) 7.3–12.1. Table 3 HR (versus the 1st quintile) of other groups classified by quintile of serum UA levels for each composite endpoint Non-CKD (n = 4994) CKD (n = 2039) Models HR (95% CI) P-value HR (95% CI) P-value ≥30% decrease in eGFR or RRT  Model 1   2nd quintile 0.92 (0.71–1.20) 0.541 1.40 (1.06–1.86) 0.018   3rd quintile 0.84 (0.64–1.10) 0.208 1.75 (1.35–2.28) <0.001   4th quintile 0.93 (0.71–1.21) 0.586 2.29 (1.77–2.97) <0.001   5th quintile 1.39 (1.09–1.77) 0.009 2.67 (2.08–3.43) <0.001  Model 2   2nd quintile 1.00 (0.76–1.30) 0.973 0.82 (0.62–1.09) 0.175   3rd quintile 0.96 (0.73–1.27) 0.795 0.91 (0.69–1.19) 0.476   4th quintile 1.15 (0.87–1.52) 0.324 0.88 (0.67–1.16) 0.358   5th quintile 1.80 (1.37–2.36) <0.001 0.75 (0.57–0.98) 0.034  Model 3   2nd quintile 1.16 (0.89–1.52) 0.266 0.81 (0.61–1.09) 0.162   3rd quintile 1.09 (0.82–1.45) 0.544 0.84 (0.64–1.11) 0.219   4th quintile 1.22 (0.91–1.62) 0.184 0.90 (0.68–1.18) 0.442   5th quintile 1.64 (1.23–2.18) <0.001 0.76 (0.58–0.99) 0.046 ≥50% decrease in eGFR or RRT  Model 1   2nd quintile 1.13 (0.66–1.93) 0.651 1.64 (1.16–2.34) 0.006   3rd quintile 0.92 (0.52–1.62) 0.772 2.10 (1.51–2.92) <0.001   4th quintile 1.41 (0.85–2.35) 0.189 2.71 (1.96–3.75) <0.001   5th quintile 2.47 (1.55–3.93) <0.001 3.37 (2.46–4.61) <0.001  Model 2   2nd quintile 1.15 (0.67–1.97) 0.603 0.81 (0.57–1.16) 0.254   3rd quintile 0.94 (0.53–1.69) 0.844 0.83 (0.59–1.17) 0.284   4th quintile 1.46 (0.85–2.50) 0.168 0.79 (0.56–1.11) 0.178   5th quintile 2.56 (1.53–4.27) <0.001 0.69 (0.50–0.96) 0.030  Model 3   2nd quintile 1.44 (0.84–2.48) 0.189 0.81 (0.57–1.17) 0.258   3rd quintile 1.29 (0.71–2.32) 0.404 0.81 (0.58–1.14) 0.227   4th quintile 1.48 (0.85–2.60) 0.167 0.84 (0.60–1.19) 0.328   5th quintile 2.14 (1.25–3.69) 0.006 0.73 (0.52–1.02) 0.068 Non-CKD (n = 4994) CKD (n = 2039) Models HR (95% CI) P-value HR (95% CI) P-value ≥30% decrease in eGFR or RRT  Model 1   2nd quintile 0.92 (0.71–1.20) 0.541 1.40 (1.06–1.86) 0.018   3rd quintile 0.84 (0.64–1.10) 0.208 1.75 (1.35–2.28) <0.001   4th quintile 0.93 (0.71–1.21) 0.586 2.29 (1.77–2.97) <0.001   5th quintile 1.39 (1.09–1.77) 0.009 2.67 (2.08–3.43) <0.001  Model 2   2nd quintile 1.00 (0.76–1.30) 0.973 0.82 (0.62–1.09) 0.175   3rd quintile 0.96 (0.73–1.27) 0.795 0.91 (0.69–1.19) 0.476   4th quintile 1.15 (0.87–1.52) 0.324 0.88 (0.67–1.16) 0.358   5th quintile 1.80 (1.37–2.36) <0.001 0.75 (0.57–0.98) 0.034  Model 3   2nd quintile 1.16 (0.89–1.52) 0.266 0.81 (0.61–1.09) 0.162   3rd quintile 1.09 (0.82–1.45) 0.544 0.84 (0.64–1.11) 0.219   4th quintile 1.22 (0.91–1.62) 0.184 0.90 (0.68–1.18) 0.442   5th quintile 1.64 (1.23–2.18) <0.001 0.76 (0.58–0.99) 0.046 ≥50% decrease in eGFR or RRT  Model 1   2nd quintile 1.13 (0.66–1.93) 0.651 1.64 (1.16–2.34) 0.006   3rd quintile 0.92 (0.52–1.62) 0.772 2.10 (1.51–2.92) <0.001   4th quintile 1.41 (0.85–2.35) 0.189 2.71 (1.96–3.75) <0.001   5th quintile 2.47 (1.55–3.93) <0.001 3.37 (2.46–4.61) <0.001  Model 2   2nd quintile 1.15 (0.67–1.97) 0.603 0.81 (0.57–1.16) 0.254   3rd quintile 0.94 (0.53–1.69) 0.844 0.83 (0.59–1.17) 0.284   4th quintile 1.46 (0.85–2.50) 0.168 0.79 (0.56–1.11) 0.178   5th quintile 2.56 (1.53–4.27) <0.001 0.69 (0.50–0.96) 0.030  Model 3   2nd quintile 1.44 (0.84–2.48) 0.189 0.81 (0.57–1.17) 0.258   3rd quintile 1.29 (0.71–2.32) 0.404 0.81 (0.58–1.14) 0.227   4th quintile 1.48 (0.85–2.60) 0.167 0.84 (0.60–1.19) 0.328   5th quintile 2.14 (1.25–3.69) 0.006 0.73 (0.52–1.02) 0.068 Model 1: univariate. Model 2: adjusted by age, sex and eGFR. Model 3: adjusted by age, sex, eGFR, duration of diabetes, use of insulin, use of angiotensin-converting enzyme inhibitors or angiotensin receptor blockers, use of lipid-lowering drugs, use of urate-lowering drugs, smoking status, date of the baseline data, BMI, systolic blood pressure, number of creatinine measurements during the follow-up period, Hb A1c, logarithmically transformed triglycerides, LDL cholesterol and logarithmically transformed urinary albumin:creatinine ratio. Range of UA levels (mg/dL): non-CKD: 1st quintile (n = 990) 0.6–4.0, 2nd quintile (n = 956) 4.1–4.7, 3rd quintile (n = 1002) 4.8–5.3, 4th quintile (n = 1045) 5.4–6.1, 5th quintile (n = 1001) 6.2–11.6; CKD: 1st quintile (n = 415) 1.2–5.1, 2nd quintile (n = 393) 5.2–5.8, 3rd quintile (n = 433) 5.9–6.5, 4th quintile (n = 379) 6.6–7.2, 5th quintile (n = 419) 7.3–12.1. FIGURE 2 View largeDownload slide Multivariable-adjusted restricted cubic spline curves (95% CI) of the association between serum UA levels and the endpoint, which was adjusted by variables used for Model 3, and the histogram of UA levels. Reference: median of UA in each group (non-CKD, 5.1 mg/dL; CKD, 6.2 mg/dL). (A) and (C) ≥30% decrease in eGFR or RRT; (B) and (D) ≥50% decrease in eGFR or RRT. FIGURE 2 View largeDownload slide Multivariable-adjusted restricted cubic spline curves (95% CI) of the association between serum UA levels and the endpoint, which was adjusted by variables used for Model 3, and the histogram of UA levels. Reference: median of UA in each group (non-CKD, 5.1 mg/dL; CKD, 6.2 mg/dL). (A) and (C) ≥30% decrease in eGFR or RRT; (B) and (D) ≥50% decrease in eGFR or RRT. Sensitivity analyses We evaluated the robustness of the present results using sensitivity analyses. First, we used the Fine and Gray subdistribution hazards model, considering the presence of competing risks. There was a significant interaction between UA levels and baseline eGFR with respect to the primary and secondary composite endpoints in the analysis adjusted by variables used for Model 3 (P-interaction = 0.004 and 0.031, respectively). In the non-CKD group, similar findings to the above were obtained (Supplementary data, Tables S1 and S2, and Supplementary data, Figure S2). In the CKD group, the associations of low levels of UA with the endpoints were not significant (Supplementary data, Tables S1 and S2). The multivariable-adjusted restricted cubic spline model showed similar results to the above (Supplementary data, Figure S2). Second, we analyzed 6270 patients without a prescription for urate-lowering drugs at baseline. There was a significant interaction between UA levels and baseline eGFR with respect to the primary and secondary composite endpoints (P-interaction < 0.001 and 0.041, respectively). Similar results to the above were detected in the non-CKD group (Supplementary data, Tables S3 and S4). In the CKD group there was a significant association only in the analyses treating UA as a continuous variable (Supplementary data, Tables S3 and S4). Third, patients in the CKD group were further classified into two groups as follows: eGFR ≥30 but <60 mL/min/1.73 m2 (n = 1638) and eGFR <30 mL/min/1.73 m2 (n = 401). There were no associations in any multivariate model of UA levels with the incidence of either endpoint in patients with an eGFR <60 mL/min/1.73 m2 (Supplementary data, Tables S5 and S6 and Figure S3). Fourth, we used models adjusted by the same variables as in Model 3 plus the use of loop diuretics or thiazides because these drugs have possible effects on serum UA levels. A total of 694 patients were prescribed these drugs at baseline. Similar results were obtained in the models (Supplementary data, Tables S7 and S8). Fifth, some subjects might have had a transient decrease in eGFR consistent with acute kidney injury at baseline because baseline eGFR in the present study was defined by a single serum creatinine measurement. Therefore we excluded 56 patients who had a decrease in serum creatinine of ≥0.3 mg/dL from the baseline or a decrease to less than two-thirds of the baseline value in the second creatinine measurement during the follow-up period [20]. The 56 excluded patients belonged to the CKD group. Similar results were obtained in the models (Supplementary data, Tables S9 and S10). Finally, we added available data of blood glucose levels (n = 6722) or urinary sugar none, ± or 1+, and 2+ or over; n = 6842) to Model 3 as a covariate, considering the lowering effects of glycosuria on serum UA levels [21]. In the non-CKD group, high levels of UA were significantly associated with both endpoints in all analyses (Supplementary data, Tables S11–S14). The results of the analyses in the CKD group partly showed an association between low levels of UA and the endpoints (Supplementary data, Tables S11–S14). DISCUSSION This single-center, large, historical cohort study of Japanese patients with type 2 diabetes was the first to identify differences in the effects of serum UA levels on the decline in kidney function, assessed based on two composite kidney endpoints, depending on baseline kidney function. The robustness of this finding was strengthened by conducting various sensitivity analyses. In all analyses, high levels of serum UA were associated with the kidney endpoints in patients with an eGFR ≥60 mL/min/1.73 m2, and the association was likely to be linear. In patients with an eGFR <60 mL/min/1.73 m2, high levels of UA had no harmful effects on the kidney endpoints. This study confirmed the harmful effects of high levels of serum UA on kidneys in patients with type 2 diabetes with preserved kidney function, consistent with the findings of previous studies [5–7]. The positive and linear association in the restricted cubic spline analysis also suggested the following: (i) effects of UA on diabetic kidneys are independent of urate crystal deposition because its saturation point in blood is approximately 7.0 mg/dL and (ii) the lower the UA value, the better the kidney outcomes. Therefore it might be necessary to set the predictive cutoff values at fairly low levels. An important question that inevitably arises is whether UA has any causal effects, although previous basic studies have indicated that UA has a pathogenic role in kidney damage [22, 23]. An investigation using the Mendelian randomization approach from the Finnish Diabetic Nephropathy Study suggested that UA is not causally related to the development and progression of kidney damage in patients with type 1 diabetes [24]. However, several randomized controlled trials reported beneficial effects of urate-lowering therapy on human kidneys [25–27]; thus, further investigation is required to clarify this fundamental issue. It is important to understand why high levels of UA did not have harmful effects on kidney outcomes in patients with an eGFR <60 mL/min/1.73 m2 in the present study. UA-induced systemic and glomerular hypertension is considered to be one of the possible mechanisms by which UA causes kidney damage [21]; however, such hypertension is already apparent due to sodium and/or water retention in patients with CKD [28, 29]. Consistent with the present findings, several previous studies of nondiabetic patients with kidney insufficiency reported no association of high levels of UA with kidney outcomes [11–13]. Interestingly, a large cohort study of patients with diabetes from Canada reported that the increased risk of end-stage kidney disease associated with poor glycemic control was attenuated at a lower baseline eGFR, and the association disappeared in patients with severe kidney insufficiency [30], despite glycemic control being established as a major prognostic and interventional factor for early stage DKD. In light of these findings, together with the results of the present study, the effects of risk factors including UA for kidney damage in diabetes might be reduced with a decline in kidney function; in other words, it is possible that there is little scope for the risk factors to cause further harm in diabetic patients with kidney insufficiency. In the present study, part of the analyses in the CKD group showed an association between low levels of UA and the endpoints. We also proposed a likely explanation for the possible paradoxical association. Some research has shown the clinical utility of serum UA as a nutritional marker in patients receiving dialysis [31, 32]. Several studies have suggested that surrogates for malnutrition were associated with kidney outcomes in non-dialysis-dependent patients with CKD [33, 34]. Therefore the increased risk observed at fairly low levels of serum UA in the present study might reflect an association between these levels as a marker of poor nutrition and kidney outcomes (Figure 2 and Supplementary data, Figure S2). Loop diuretics or thiazides improve fluid overload and hypertension with possible increased effects on serum UA levels. Therefore the beneficial effects by these drugs might partly explain the paradoxical association, although the results after adjustment by use of these drugs remained unchanged (Supplementary data, Tables S7 and S8). The present study has several limitations. First, study participants comprised an ethnically homogeneous population at a single urban university hospital in Japan, therefore the generalizability of the present findings may be limited. Second, we could not clarify a causal relationship between UA and the kidney outcomes. Third, we could not show the continuous change of effects of UA on the endpoints according to baseline eGFR treated as a continuous variable because we could not develop a spline function including interaction of baseline eGFR and UA values treated as continuous variables in models for time-to-event data. Therefore, in the present study, the effects of UA were estimated with respect to each group categorized by baseline eGFR. Fourth, we did not evaluate information about alcohol intake, which has possible effects on serum UA levels. Finally, the present study did not evaluate time-dependent changes in serum UA, HbA1c, lipid profiles, blood pressure, BMI or medications, including urate-lowering drugs, during the follow-up period. In conclusion, the present study might provide evidence that the effects of UA on the decline in kidney function differ depending on baseline kidney function in patients with type 2 diabetes, suggesting that we need to consider the management of UA according to kidney function as therapeutic strategies for DKD. High levels of serum UA are the prognostic factor only in patients with preserved kidney function. SUPPLEMENTARY DATA Supplementary data are available at ndt online. ACKNOWLEDGEMENTS We would like to thank all members of the Diabetes Center, Tokyo Women’s Medical University School of Medicine for their helpful advice and discussions. Parts of this work were presented at the 77th Scientific Session of the American Diabetes Association, San Diego, CA, USA, 9–13 June 2017. AUTHORS’ CONTRIBUTIONS K.H. conceived the study, designed the protocol, contributed to data collection and preparation, analyzed all data, wrote the manuscript and contributed to the interpretation of the results. E.T., Y.N., T.M. and Y.Y. contributed to data collection and preparation and contributed to the interpretation of the results. Y.U. and T.B. designed the protocol and contributed to the interpretation of the results. T.B. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. All the authors have read the manuscript and have approved this submission. CONFLICT OF INTEREST STATEMENT None declared. REFERENCES 1 Alicic RZ , Rooney MT , Tuttle KR. Diabetic kidney disease: challenges, progress, and possibilities . Clin J Am Soc Nephrol 2017 ; 12 : 2032 – 2045 Google Scholar CrossRef Search ADS PubMed 2 Kanbay M , Segal M , Afsar B et al. The role of uric acid in the pathogenesis of human cardiovascular disease . Heart 2013 ; 99 : 759 – 766 Google Scholar CrossRef Search ADS PubMed 3 Volterrani M , Iellamo F , Sposato B et al. Uric acid lowering therapy in cardiovascular diseases . Int J Cardiol 2016 ; 213 : 20 – 22 Google Scholar CrossRef Search ADS PubMed 4 Zhu P , Liu Y , Han L et al. Serum uric acid is associated with incident chronic kidney disease in middle-aged populations: a meta-analysis of 15 cohort studies . PLoS One 2014 ; 9 : e100801 Google Scholar CrossRef Search ADS PubMed 5 Zoppini G , Targher G , Chonchol M et al. Serum uric acid levels and incident chronic kidney disease in patients with type 2 diabetes and preserved kidney function . Diabetes Care 2012 ; 35 : 99 – 104 Google Scholar CrossRef Search ADS PubMed 6 Takagi M , Babazono T , Uchigata Y. Differences in risk factors for the onset of albuminuria and decrease in glomerular filtration rate in people with type 2 diabetes mellitus: implications for the pathogenesis of diabetic kidney disease . Diabet Med 2015 ; 32 : 1354 – 1360 Google Scholar CrossRef Search ADS PubMed 7 Viazzi F , Piscitelli P , Giorda C et al. Metabolic syndrome, serum uric acid and renal risk in patients with T2D . PLoS One 2017 ; 12 : e0176058 Google Scholar CrossRef Search ADS PubMed 8 Edwards NL. The role of hyperuricemia and gout in kidney and cardiovascular disease . Cleve Clin J Med 2008 ; 75 : S13 – S16 Google Scholar CrossRef Search ADS PubMed 9 Hatch M , Vaziri ND. Enhanced enteric excretion of urate in rats with chronic renal failure . Clin Sci 1994 ; 86 : 511 – 516 Google Scholar CrossRef Search ADS PubMed 10 Vaziri ND , Freel RW , Hatch M. Effect of chronic experimental renal insufficiency on urate metabolism . J Am Soc Nephrol 1995 ; 6 : 1313 – 1317 Google Scholar PubMed 11 Madero M , Sarnak MJ , Wang X et al. Uric acid and long-term outcomes in CKD . Am J Kidney Dis 2009 ; 53 : 796 – 803 Google Scholar CrossRef Search ADS PubMed 12 Testa A , Mallamaci F , Spoto B et al. Association of a polymorphism in a gene encoding a urate transporter with CKD progression . Clin J Am Soc Nephrol 2014 ; 9 : 1059 – 1065 Google Scholar CrossRef Search ADS PubMed 13 Nacak H , van Diepen M , Qureshi AR et al. Uric acid is not associated with decline in renal function or time to renal replacement therapy initiation in a referred cohort of patients with stage III, IV and V chronic kidney disease . Nephrol Dial Transplant 2015 ; 30 : 2039 – 2045 Google Scholar CrossRef Search ADS PubMed 14 Kashiwagi A , Kasuga M , Araki E et al. International clinical harmonization of glycated hemoglobin in Japan: from Japan Diabetes Society to National Glycohemoglobin Standardization Program values . J Diabetes Invest 2012 ; 3 : 39 – 40 Google Scholar CrossRef Search ADS 15 Matsuo S , Imai E , Horio M , Collaborators developing the Japanese equation for estimated GFR et al. Revised equations for estimated GFR from serum creatinine in Japan . Am J Kidney Dis 2009 ; 53 : 982 – 992 Google Scholar CrossRef Search ADS PubMed 16 Coresh J , Turin TC , Matsushita K et al. Decline in estimated glomerular filtration rate and subsequent risk of end-stage renal disease and mortality . JAMA 2014 ; 311 : 2518 – 2531 Google Scholar CrossRef Search ADS PubMed 17 Durrleman S , Simon R. Flexible regression models with cubic splines . Stat Med 1989 ; 8 : 551 – 561 Google Scholar CrossRef Search ADS PubMed 18 Fine JP , Gray RJ. A proportional hazards model for the sub-distribution of a competing risk . J Am Stat Assoc 1999 ; 94 : 496 – 509 Google Scholar CrossRef Search ADS 19 Austin PC , Lee DS , Fine JP. Introduction to the analysis of survival data in the presence of competing risks . Circulation 2016 ; 133 : 601 – 609 Google Scholar CrossRef Search ADS PubMed 20 KDIGO clinical practice guideline for acute kidney injury . Kidney Int Suppl 2012 ; 2 : 1 – 138 CrossRef Search ADS 21 Johnson RJ , Nakagawa T , Jalal D et al. Uric acid and chronic kidney disease: which is causing which? Nephrol Dial Transplant 2013 ; 28 : 2221 – 2228 Google Scholar CrossRef Search ADS PubMed 22 Mazzali M , Hughes J , Kim YG et al. Elevated uric acid increases blood pressure in the rat by a novel crystal-independent mechanism . Hypertension 2001 ; 38 : 1101 – 1106 Google Scholar CrossRef Search ADS PubMed 23 Mazzali M , Kanellis J , Han L et al. Hyperuricemia induces a primary renal arteriolopathy in rats by a blood pressure-independent mechanism . Am J Physiol Renal Physiol 2002 ; 282 : F991 – F997 Google Scholar CrossRef Search ADS PubMed 24 Ahola AJ , Sandholm N , Forsblom C et al. The serum uric acid concentration is not causally linked to diabetic nephropathy in type 1 diabetes . Kidney Int 2017 ; 91 : 1178 – 1185 Google Scholar CrossRef Search ADS PubMed 25 Siu YP , Leung KT , Tong MK , Kwan TH. Use of allopurinol in slowing the progression of renal disease through its ability to lower serum uric acid level . Am J Kidney Dis 2006 ; 47 : 51 – 59 Google Scholar CrossRef Search ADS PubMed 26 Goicoechea M , de Vinuesa SG , Verdalles U et al. Effect of allopurinol in chronic kidney disease progression and cardiovascular risk . Clin J Am Soc Nephrol 2010 ; 5 : 1388 – 1393 Google Scholar CrossRef Search ADS PubMed 27 Liu P , Chen Y , Wang B et al. Allopurinol treatment improves renal function in patients with type 2 diabetes and asymptomatic hyperuricemia: 3-year randomized parallel-controlled study . Clin Endocrinol (Oxf) 2015 ; 83 : 475 – 482 Google Scholar CrossRef Search ADS PubMed 28 Rahman M , Yang W , Akkina S et al. Relation of serum lipids and lipoproteins with progression of CKD: the CRIC Study . Clin J Am Soc Nephrol 2014 ; 9 : 1190 – 1198 Google Scholar CrossRef Search ADS PubMed 29 Inaguma D , Imai E , Takeuchi A et al. Risk factors for CKD progression in Japanese patients: findings from the Chronic Kidney Disease Japan Cohort (CKD-JAC) study . Clin Exp Nephrol 2017 ; 21 : 446 – 456 Google Scholar CrossRef Search ADS PubMed 30 Shurraw S , Hemmelgarn B , Lin M et al. Association between glycemic control and adverse outcomes in people with diabetes mellitus and chronic kidney disease . Arch Intern Med 2011 ; 171 : 1920 – 1927 Google Scholar CrossRef Search ADS PubMed 31 Beberashvili I , Sinuani I , Azar A et al. Serum uric acid as a clinically useful nutritional marker and predictor of outcome in maintenance hemodialysis patients . Nutrition 2015 ; 31 : 138 – 147 Google Scholar CrossRef Search ADS PubMed 32 Park C , Obi Y , Streja E et al. Serum uric acid, protein intake and mortality in hemodialysis patients . Nephrol Dial Transplant 2017 ; 32 : 1750 – 1757 Google Scholar PubMed 33 Lu JL , Kalantar-Zadeh K , Ma JZ et al. Association of body mass index with outcomes in patients with CKD . J Am Soc Nephrol 2014 ; 25 : 2088 – 2096 Google Scholar CrossRef Search ADS PubMed 34 Kikuchi H , Kanda E , Mandai S et al. Combination of low body mass idex and serum albumin level is associated with chronic kidney disease progression: the chronic kidney disease-research of outcomes in treatment and epidemiology (CKD-ROUTE) study . Clin Exp Nephrol 2017 ; 21 : 55 – 62 Google Scholar CrossRef Search ADS PubMed © The Author(s) 2018. Published by Oxford University Press on behalf of ERA-EDTA. All rights reserved. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)

Journal

Nephrology Dialysis TransplantationOxford University Press

Published: May 30, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off